writing an essay about "the effect of increasing minimum wage on employment rate in South Korea", 8 pages, economic related, must be original.

The Impact of the Minimum Wage on Older

Workers in South Korea

Min-a Jung

PhD Thesis

Department of Sociological Studies

The University of Sheffield

September 2019

Abstract

The purpose of this thesis is to explore the effects of the minimum wage on older

workers in South Korea and seek a better understanding of the effects in the context of

the minimum wage policy process. Using the KLIPS 2004-2013, this study examines

the effects of the minimum wage on the distributions of wages and family earnings,

employment, and exits from in-work poverty among older employees aged 55 or over.

Two government periods with the highest rates of increase in the minimum wage and

the lowest rates of increase were compared. The results show that the minimum wage

during the period with the highest growth rates raised wages, marginally reducing the

wage gap, but had insignificant effects on the distribution of family earnings and the

odds of individuals’ exits from in-work poverty. For the period with the lowest growth

rates, the minimum wage slightly improved the median of wages and of family

earnings, marginally reducing the wage gap, but decreased the chance to exit from in-

work poverty. An adverse employment effect was not found for both periods. This thesis

proposed a political economy framework and analysed the Minimum Wage Council’s

annual reports on the fixing process 2003-2012. The findings show that the effects for

the period with the highest growth rates were associated with the role of the minimum

wage preserving the monthly pay as a result of the reduction of working hours and the

exemption of and the subminimum for surveillance or intermittent work. The effects for

the period with the lowest growth rates were related to the minimum wage increases

slightly higher than annual pay awards, its role of preserving monthly pay consequent

upon the reduction of working hours, and the subminimum for surveillance or

intermittent work.

i

Table of Contents

Chapter One Introduction 1

Chapter Two The Minimum Wage 17

Chapter Three Older Workers in the Labour Market 35

Chapter Four Older Workers: Labour Market Status and Minimum Wages 53

Chapter Five Theoretical Framework and Research Methods 96

Chapter Six The Factual Context for the Case Study of South Korea 148

Chapter Seven The Effects of the Minimum Wage on Older Workers in South Korea 177

Chapter Eight Political Economy of the Impact of Minimum Wages 211

Chapter Nine Conclusion 258

Appendix 270

Bibliography 288

ii

List of Tables

Table 1.1 The Proportion of Employers and Workers Participating in

Workplace Pension Schemes in South Korea, 2015-2016 (%) ................... 7

Table 2.1 Minimum Wage Fixing Procedures among OECD Countries ........ 23

Table 2.2 Minimum Wage Fixing Criteria among OECD Countries ............. 27

Table 3.1 Part Time Employees Aged 55-64 and 65+ by Country in 2011 .... 44

Table 3.2 Temporary Employees Aged 55-64 and 65+ by Country in 2011 ... 46

Table 4.1 Pension Ages and Financial Work Incentives/Retirement Disincentives by Type of Scheme in OECD Countries ............................ 68

Table 4.2 Limits to Combining Work and Pensions ...................................... 71

Table 4.3 Studies of Minimum Wage Effects and Older Workers .................. 84

Table 6.1 Changes in Minimum Wage Rates in South Korea (KRW/GBP/USD; %; thousand persons) .............................................. 151

Table 6.2 Empirical Studies about the Effects of the Minimum Wage in South Korea.................................................................................................... 166

Table 7.1 Characteristics of 55+ Employees, 2004-2013 ............................ 179

Table 7.2 Estimated Models for Changes in the Percentiles of Log Hourly Wages among 55+ Employees, 2004-2008 and 2009-2013.................... 186

Table 7.3 Estimated Models for Changes in Employment & Unemployment among 55+ Employees, 2004-2008 and 2009-2013 ............................... 191

Table 7.4 Descriptive Statistics of 55+ Employees by Covariates, 2004-2008 & 2009-2013 ........................................................................................ 193

Table 7.5 Estimated Fixed-effects Models of Transitions from being Employed among 55+ Employees, 2004-2008 and 2009-2013 .............. 195

Table 7.6 Estimated Random-effects Models of Transitions from being Employed among 55+ Employees, 2004-2008 and 2009-2013 .............. 196

Table 7.7 Estimated Models for Changes in the Percentiles of Log Family Earnings among 55+ Employees, 2004-2008 and 2009-2013 ................ 200

Table 7.8 Descriptive Statistics of 55+ Working Poor by Covariates, 2004-

iii

2008 & 2009-2013................................................................................ 204

Table 7.9 Estimated Fixed-effects Models for Transitions from In-Work Poverty among 55+ Employees, 2004-2008 and 2009-2013 .................. 206

Table 7.10 Estimated Random-effects Models for Transitions from In-Work Poverty among 55+ Employees, 2004-2008 and 2009-2013 .................. 207

List of Figures

Figure 1.1 Average Normal and Effective Retirement Age in OECD Countries

by Sex, 1970-2015 .................................................................................... 1

Figure 1.2 Average Age of Retirement from Lifetime Main Job in South Korea, 2005-2017 ..................................................................................... 4

Figure 1.3 Effective Age of Labour Market Exit and Normal Pensionable Age in South Korea ......................................................................................... 5

Figure 1.4 The Share of Beneficiaries of the National Pension among Those Aged 60+ in South Korea (%) .................................................................. 6

Figure 1.5 Income Sources of Older People Aged 60+ in South Korea (%) .... 8

Figure 3.1 Labour Force Participation Rates by Age Group in OECD Countries, 1970-2011 (%) ....................................................................... 37

Figure 3.2 Labour Force Participation Rates of Workers Aged 55+ by Sex and Age Group in OECD Countries, 1995-2015 (%) ..................................... 37

Figure 3.3 Labour Force Participation Rates of Workers Aged 55-64 among OECD Countries, 1970-2011 (%) ........................................................... 38

Figure 3.4 Employment Rates by Age Group in OECD Countries, 1970-2011 (%) ......................................................................................................... 40

Figure 3.5 Employment Rates of Older Workers Aged 55+ by Sex and Age Group in OECD Countries, 1995-2015 (%) ............................................ 40

Figure 3.6 Employment Rates of Workers Aged 55-64 among OECD Countries, 1970-2011 (%) ....................................................................... 41

Figure 3.7 Incidence of Part Time Employment by Age Group in OECD Countries, 2000-2011 (%) ....................................................................... 42

iv

Figure 3.8 Incidence of Part Time Employment among Workers Aged 55-64 and 65+ by Sex in OECD Countries, 2001-2016 (%) .............................. 43

Figure 3.9 Incidence of Temporary Employment by Age Group in OECD Countries, 2000-2011 (%) ....................................................................... 45

Figure 3.10 Incidence of Temporary Employment among Workers Aged 55-64 and 65+ by Sex in OECD Countries, 2001-2016 (%) .............................. 45

Figure 3.11 Sectoral Empolyment Structure by Sector and Age Group in EU 25, 2011 (%) ........................................................................................... 47

Figure 3.12 Annual Employment Growth by Sector and Age Group in EU 25, 2008-2011 .............................................................................................. 48

Figure 3.13 Proportion of Low-Wage Earners by Age Group in EU Countries, 2010 (%)................................................................................................. 49

Figure 3.14 Changes in Proportion of Low-Wage Earners Aged 50+ in EU Countries, 2006-2010 (%) ....................................................................... 50

Figure 3.15 Gender Pay Gap among Older Workers Aged 55-64 and 65+ in EU Countries, 2010 (%).......................................................................... 51

Figure 4.1 Factors Influencing Labour Market Status of Older Workers ....... 59

Figure 6.1 Process of Minimum Wage Deliberation and Determination in South Korea .......................................................................................... 150

Figure 6.2 Growth in Minimum Wage, Annual Inflation (CPI) & GDP per Capita, and Minimum Wage Relative to Average Wage of Full-Time Workers in South Korea, 1988-2013 (%) .............................................. 152

Figure 6.3 The Share of Employees Paid Less than the Minimum Wage in South Korea, 2001-2013 (%) ................................................................ 153

Figure 6.4 Changes in Labour force Participation, Employment and Unemployment among Workers Aged 50 + in South Korea, 2000-2014 (%) ............................................................................................................. 155

Figure 6.5 Changes in the Growth Rate in Population, Labour force Participation and Employment among Workers Aged 50+ in South Korea, 2001-2014 (%)...................................................................................... 156

Figure 6.6 Changes in Employment Status of Workers Aged 50+ in South Korea, 2007-2013 (%) .......................................................................... 157

Figure 6.7 Changes in Industrial Composition among Workers Aged 55-79 in

v

South Korea, 2005-2014 (%) ................................................................ 158

Figure 6.8 Changes in Occupational Composition among Workers Aged 55-79 in South Korea, 2005-2014 (%) ............................................................ 159

Figure 6.9 Changes in the Ratios of Monthly Wage & Hourly Wage among Employees 50+ to among Those Aged 30-49 and Weerkly Hours Worked by Age Group in South Korea (%, hrs) ................................................. 160

Figure 6.10 Changes in the Proportion of Older Employees Earning Less Than 50% of the Median Wage in South Korea (%) ....................................... 161

Figure 7.1 Wage Distribution of 55+ Employees, 2004-2008 and 2009-2013 ............................................................................................................. 181

Figure 7.2 Distribution of Monthly Family Earnings among 55+ Employees, 2004-2008 and 2009-2013 .................................................................... 198

1

Chapter One Introduction

Overview of Topic

The era when early retirement was a norm in the workplace, in which retirees rarely

returned to the labour market and few workers were likely to be paid at or near a

minimum wage, disappeared almost twenty years ago (Fang and Gunderson, 2009). The

labour market exit age, which had dropped sharply mainly due to the maturing of

pension systems and early retirement schemes, has been gradually increasing since the

early 2000s, following the rise in the normal pensionable age (OECD, 2017, pp.48-49;

Figure 1.1). As the life after retirement was no longer being supported by pension plans

as before, the labour force participation of older workers and their employment have

increased since the mid-1990s, and a growing number of older workers have been

engaged in low paid jobs (see Chapter Three).

Figure 1.1 Average Normal and Effective Retirement Age in OECD Countries by Sex, 1970-2015

Source: OECD, Pension at a Glance 2017, p.49 Note: The normal retirement age refers to the average normal pensionable age; the effective retirement age is defined as the average labour market exit age which is estimated by OECD from observed changes in participation rates over a five-year period for successive cohorts of workers (by five-year age groups) aged 40 and over. For the effective retirement age, each indicated year points to the end year of each five-year period.

60 61 62 63 64 65 66 67 68 69 70

1971 1976 1981 1986 1991 1996 2001 2006 2011

Ag e

1971 1983 1989 1993 1999 2002 2010 2014 Normal 63.9 63.2 62.8 62.5 62.6 62.7 63.0 63.9

Effective 68.5 65.2 64.3 63.9 63.3 63.2 63.9 64.6

MEN

58

60

62

64

66

68

70

1971 1976 1981 1986 1991 1996 2001 2006 2011

Ag e

1971 1983 1989 1993 1999 2002 2010 2014 Normal 62.1 61.7 61.1 61.1 61.2 61.4 61.9 63.0

Effective 66.2 63.4 62.6 62.1 61.1 61.2 62.6 63.2

WOMEN

2

Along with the changes in older workers’ economic activity, research about their status

in the labour market has developed from pension-focused research on retirement to

multiple factors studies on retirement and further to studies on employment and age

discrimination in the labour market. However, despite a large amount of research with

regard to the status of older workers in the labour market, little attention has been paid

to the issue of their earned income. Considering the fact that financial need is one of the

most decisive factors for workers to remain in employment in their later life, the issue

of earned income among older workers is of importance. Further, the tendency among a

growing number of older workers to be hired in low paid jobs arouses interest in the

influence on older workers of policies for the low paid.

The minimum wage is the most popular policy for low paid workers, but its efficacy is

controversial. Twenty six out of 34 OECD countries have statutory minimum wages in

place, and in the eight OECD countries that do not have a statutory minimum wage, a

large portion of the workforce is covered by the wage floors specified by sector-level

collective agreements (OECD, 2015, p.1). Despite its prevalence, the effects of the

minimum wage have been very debatable, particularly in terms of employment and in-

work poverty. This has relevance mainly to two points. One is that the research on the

minimum wage has been focused mainly on younger workers who were regarded as the

group most affected by an increase in the minimum wage. The other is that the effects of

the minimum wage have been explained almost exclusively by economic theories of the

labour market. The first point implies that workers in other vulnerable groups, such as

women, disabled workers and older workers, were relatively less explored in the field.

In fact, Fang and Gunderson’s work (2009) on the minimum wage effects on older

workers’ employment is the only study focusing on this group. Based on studies that

have devoted attention mainly to younger workers, a common argument is that the

minimum wage is not a useful tool to improve household incomes among low paid

families and to enable the working poor to exit from poverty because ‘most of the

benefits of minimum wage increases goes to second or third earners living in

households well above the poverty line’ (Sabia and Burkhauser, 2010, p.595).1

However, in as much as workers in the groups less focused on by the research on the

minimum wage are more likely to be the primary or indispensable secondary earners in

their households, the argument should be examined further through research into other

vulnerable groups affected by the minimum wage. The second point, the dominance of

3

economic theory in this research field provides limited explanations of the conflicting

empirical evidence, particularly of employment effects of the minimum wage.

Considering that the minimum wage is a public policy in which conflicting interests

among major groups of society, such as government, employers and employees, are

involved and which is made through a political process, an alternative explanation

involving both economic and political features of minimum wage policy is needed for

fuller and contextual accounts of the minimum wage effects.

In recognition of problems presented above, this thesis mainly aims to explore the

effects of the minimum wage on older workers in the context of South Korea. Korea is

one of the countries where the population is the most rapidly ageing due to the

decreasing fertility rate and the increasing life expectancy. As the share of those aged 65

years and older was 14.02 per cent in August, 2017, Korea has become an ‘aged society’

17 years after it entered the ‘ageing society’ group in 2000 (Kim, S-M and Chang, H-T,

2017). According to the National Statistical Office of Korea [통계청] (2017, p.15), the

share is projected to be 41.0 per cent in 2060. The gender ratio is 57.5 per cent for

women and 42.5 per cent for men in 2007; however, the gap in the ratio is projected to

decline to 4.6 per cent by 2060 (The National Statistical Office of Korea [통계청], 2017,

p.16). Life expectancy at birth increased from 51.1 years in 1960 to 76.2 years in 2010

for men and from 53.7 to 82.9 years for women, and it is expected to be 82.9 for men

and 88.9 for women in 2050 (Choi, S-J, 2015, p.58).

Despite the significant increase in life expectancy, the actual age at which workers retire

from their lifetime main jobs is very low. The survey which was conducted in 2001 by

the Ministry of Labour [노동부] among firms with 300 plus employees showed that the

prevalent retirement age was 55 and the average was 56.7 years (Phang, H-N, 2011,

p.57). According to Supplementary Results (for the Old Population) of the

Economically Active Population Survey [경제활동인구조사 (고령층) 부가조사], the

actual age of retirement from lifetime main jobs among those aged between 55 and 64

was 52.2 in 2005, 51.5 in 2011 and 51.4 in 2017 for men and 48.1, 47.2 and 47.1 for

women, respectively (Figure 1.2). This is because mandatory retirement has been very

prevalently practiced among firms in Korea and the regulated retirement age,

particularly in the private sector, has often been set at a very low age ranged from 55 to

60 (Phang, H-N, 2011; Choi, S-J, 2015), relying on the absence of ‘law which governs

4

mandatory retirement age, except for the public sector’ (Choi, S-J, 2015, p.60). Very

recently, the mandatory retirement age should be set at the age of 60 years or over under

the Act on Prohibition of Age Discrimination in Employment and Elderly Employment

Promotion [고용상 연령차별금지 및 고령자고용촉진에 관한 법률], which has been

applied to firms with 300 employees and over and public organisations from 2016 and

to firms with less than 300 workers and state agencies and local governments from 2017

(The Ministry of Employment and Labor [고용노동부], 2016).

Figure 1.2 Average Age of Retirement from Lifetime Main Job in South Korea, 2005- 2017

Source: The National Statistical Office of Korea, Supplementary Results (for the Old Population) of the Economically Active Population Survey (every May in each year) [경제활동인구조사 (고령층) 부가조사 (각 연 도 5월)] Note: The average age of retirement from lifetime main job refers to the average age at which those aged between 55 and 64 retired or tranferred from their lifetime major jobs.

The actual age of retirement from a lifetime main job is well below the pensionable age

and effective age of exit from the labour market in Korea. As shown in Figure 1.3, the

normal pensionable age was originally set at 60 years for both men and women but has

been increasing by one year per five-year term from 2013 to 2033 under the National

Pension (Amendment) Act 2011 [국민연금법], reaching 61 between 2013 and 2017, 62

between 2018 and 2022, 63 between 2023 and 2027, 64 between 2028 and 2032 and 65

by 2033. The average effective age of exit from the labour market was 70.3 for men and

69.8 for women for the period of 2004-2009, increased to 72.9 and 70.6 respectively

over the period 2009-2014 and led to 72.0 and 72.2 for each over the period 2011-2016.

These facts suggest that Korean older people are likely to experience financial

difficulties and tend to remain at the workforce almost 20 years longer after retiring

from their lifetime main jobs. In terms of the labour force participation, ‘the old-age

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Total 50.0 50.3 49.6 49.6 49.5 49.3 49.2 49.3 49.4 49.4 49.0 49.0 49.1 Men 52.2 52.4 51.9 52.0 51.8 51.6 51.5 51.5 51.4 51.6 51.5 51.5 51.4 Women 48.1 48.3 47.5 47.4 47.4 47.2 47.2 47.4 47.6 47.4 46.9 47.0 47.1

44 45 46 47 48 49 50 51 52 53

Ag e

5

participation rates for both Korean men and women are slightly or considerably higher

than the average rate of OECD or European Union countries’ (Phang, H-N, 2011, p.55).

Such high participation and employment rates among the old age were due to retirees’

self-employment and those engaged in the agricultural and fishery industry; however,

the numbers of both have considerably decreased since the middle of the millennium

whereas the number of wage workers, particularly engaged in low paid jobs, has

significantly increased among older people (see Chapter Six).

Figure 1.3 Effective Age of Labour Market Exit and Normal Pensionable Age in South Korea

Source: OECD, Society at a Glance 2009 for the average effective age of labour market exit over the period 2002- 2007; OECD, Pension at a Glance 2011, 2013, 2015 & 2017 for the average effective ages over the period 2004- 2009, 2007-2012, 2009-2014 and 2011-2016, respectively; Article 21 of the supplementary provisions of the National Pension (Amendment) Act 2011 [국민연금법 부칙 21조] for normal pensionable ages Note: The average effective age of labour market exit was estimated by OECD from observed changes in participation rates over a five-year period for successive cohorts of workers (by five-year age groups) aged 40 and over.

Meanwhile, retirement income provisions have been long underdeveloped. As seen in

Figure 1.4, the state pension which was introduced in 1988 covered only 30.4 per cent

of those aged 60 years and over in 2008 and 40.9 per cent in 2016. Especially among

women, the share of the state pension beneficiaries has been much lower, staying at less

than 30 per cent until 2016. This low share among women can be understood mainly as

attributable to the fact that they have been more likely to be engaged in low-wage,

irregular jobs in which employers have not provided the four major public insurances,

including the National Pension, the National Health Insurance, Employment Insurance

and Industrial Accident Compensation Insurance and less likely to maintain their

subscription to the pension stable due to career discontinuity which is involved in

childbirth and child-rearing plus employment instability.

2002- 2007

2004- 2009

2007- 2012

2009- 2014

2011- 2016

1988- 2012

2013- 2017

2018- 2022

2023- 2027

2028- 2032 2033 +

Effective Age of Labour Market Exit Normal Pensionable Age Men 71.2 70.3 71.1 72.9 72.0 60 61 62 63 64 65 Women 67.9 69.8 69.8 70.6 72.2 60 61 62 63 64 65

0 10 20 30 40 50 60 70 80

Ag e

6

Figure 1.4 The Share of Beneficiaries of the National Pension among Those Aged 60+ in South Korea (%)

Source: The National Pension Service, National Pension Statistical Yearbook (each year) [국민연금통계연보 (각 연도)]for the number of beneficiaries of the National Pension; The National Statistical Office of Korea, Population Projections for Korea [장래추계인구]for the projected number of people aged 60+ Note: The share of beneficiaries was calculated by (the number of beneficiaries among those aged 60+/ the projected number of people aged 60+)*100 in each year.

Retirement allowance ‘which was introduced in 1953 under labour legislation’ and ‘has

provided retiring employees with an immediate lump-sum cash payment’ has played a

critical role of old-age income security (Klassen and Yang, 2010, pp.7-8). The

retirement allowance has typically become either ‘start-up capital’ for being self-

employed or ‘a financial cushion for the transition from the primary to invariably

precarious, secondary labour market’ (Klassen and Yang, 2010, p.8). But, as the idea of

‘a lifelong workplace’ which had been prevalent in Korea has disappeared under the

influence of labour market flexibilization and the bank interest rates has considerably

fallen with the financial reforms since the late 1990s, the retirement allowance has not

amounted as much as before. Since 2005, employers can replace the retirement

allowance with a workplace pension scheme under the Employee Retirement Benefit

Security Act [근로자퇴직급여보장법]. Table 1.1 shows the proportions of employers

and workers who participated in workplace pension schemes in 2015 and 2016. By

2016, only 10.9 per cent of employers with less than 5 employees and 40.3 per cent of

those with 5-29 employees instituted a company pension scheme while employers with

less than 30 employees comprised 93.6 per cent of the targeted establishments. For

workers, half of all those eligible were enrolled in a company pension scheme by 2016;

however, the enrolment decreased with age among those aged 50 years and over, and

there was a sizeable gap between men and women, except among those aged 60 years

and over.

2004 2008 2012 2013 2014 2015 2016 Total 20.8 30.4 37.1 37.7 37.9 39.2 40.9 Men 32.5 45.2 53.3 53.4 53.0 54.0 55.5 Women 12.3 19.2 24.6 25.5 26.1 27.6 29.2

0

10

20

30

40

50

60

7

Table 1.1 The Proportion of Employers and Workers Participating in Workplace Pension Schemes in South Korea, 2015-2016 (%)

Employers Workers

Size Participating Targeted

Age Total Men Women

2015 2016 2015 2016 2015 2016 2015 2016 2015 2016

Total 25.6 26.9 100.0 100.0 All ages 47.9 50.0 50.2 52.2 44.3 46.6

Less than 5 9.7 10.9 52.0 53.6 50 - 54 46.0 48.6 48.8 51.6 41.8 44.0

5 - 29 37.5 40.3 41.3 40.0 55 - 59 40.0 42.8 41.2 44.2 38.1 40.6

30-299 74.5 76.6 6.3 6.0 60 - 64 31.4 33.6 30.7 33.0 32.7 34.6

300 and over 88.2 88.1 0.4 0.4 65+ 22.1 25.9 23.3 25.9 19.6 25.7

Source: The National Statistical Office of Korea, (2017). Retirement Pension Statistics in 2016 [2016년 기준 퇴직 연금통계]

Along with the state pension, retirement allowance and workplace pensions, personal

pensions had begun in 1994 in Korea. But they tend to be regarded as a financial

product for preparation for retirement rather than as an old age social security system,

and in this regard, individuals show a propensity to join several insurance plans and not

to stay with them (Jeong, W-S, Kang, S-H and Lee, S-W, 2014, p.18).

These public transfers and private pensions account for less than 30 per cent of the total

income for Koreans aged 60 years and over (Figure 1.5). The rest mainly comes from

individuals’ own work and support from family and relatives; however, as shown in

Figure 1.5, the latter has been gradually eroded whereas the former has stayed above 50

per cent. The weakened family support is attributed to increase in women’s participation

in the labour market (Klassen and Yang, 2010), to fewer resources available as a result

of fewer children (Phang, H-N, 2011) and, most of all, to the change in both children’s

and old ages’ ideas and values relating to caring for the aged. This suggests the

importance of income from older people’s own work in Korea, despite the ongoing

development of the pension system.

Responding to the needs of work among older people and the expected increase in

social expenditures as a result of the population ageing, the Korean government has

developed a variety of measures to promote older workers’ employment, including

‘salary peak system’ in which employers cut salaries of their employees being a certain

age but, instead, ensure the employees’ employment until normal retirement age and

subsidies to employers who hire and retain older workers. ‘There are four types of

8

Figure 1.5 Income Sources of Older People Aged 60+ in South Korea (%)

Source: The National Statistical Office of Korea, Social Survey, every other year [사회조사, (각 조사 연도)] Note: Income from work includes earnings and income from business; income from assets includes profits from interests, dividends and real estates; pensions refers to both public and private ones; and welfare refers to resource- tested benefits and support from government or social groups.

subsidies under the Employment Insurance System [고용보험제도] which was

introduced 1995: subsidy to promote over-quota (6%) employment of older workers,

subsidy to promote newly employing older workers, subsidy to promote extended

employment of retiring workers and subsidy to promote older workers upon completion

of subsidised job training’ (The Ministry of Labor [노동부], 2009, quoted in Phang, H-

N, 2011, p.59). These can be promptly effective measures to improve employment

among older workers; however, they include, arguably, the nature of age discrimination

against older workers in terms of wages, or could make their target population more

vulnerable in employment in the long term unless they run parallel with a measure to

change employers’ perceptions of older workers, which seem to be more critical in as

much as the rates of employers taking up the government employment measures for

older workers are very low, and further, the main reason not to take up the ‘salary peak

system’ is the recent increase of the mandatory retirement age to 60 under the

amendment of the Act on Prohibition of Age Discrimination in Employment and Elderly

Employment Promotion [고용상 연령차별금지 및 고령자고용촉진에 관한 법률]

(Jung, S-Y, 2016).

The policy practices presented above shows that Korean government’s approaches to

the improvement of older workers’ status in the labour market have ruled out the

consideration of their wages, focusing on their employment or tends to have reflected

2009 2011 2013 2015 2017 Welfare 8.6 7.7 7.6 10.4 9.9 Family & Relatives 31.4 32.0 28.8 23.0 20.2 Installment Savings 6.1 9.1 9.9 6.3 7.3 Pensions & Retirement Allowances 24.5 27.6 24.9 27.6 28.1 Assets 13.3 12.4 12.2 11.7 10.3 Work 56.2 50.9 53.1 54.4 54.2

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

100%

9

employers’ perceptions of older workers and wider ‘age culture’ which is ‘the shorthand

description of social norms, values, ideals or perceptions in society that structure the

ideas of the age-work relationships’ (De Vroom, 2004, p.8) as the salary peak system

shows. But, allowing for the available income sources in old ages described above,

financial needs are likely to be a main driver for Korean older workers to remain longer

at the workforce; however, as the labour market has been dramatically changed since

the IMF bailout in 1998, and as a result, insecure, low paid jobs have exploded

throughout the first decade of the millennium, older workers have been more likely to

be engaged in low paid jobs due to their vulnerability in the labour market. This

disparity between their financial needs for work and status in the labour market,

particularly in terms of wages, have rarely been considered or not well addressed in the

context of a more common public policy practice for the improvement of wages among

vulnerable groups of workers.

The minimum wage is a typical wage-governing public policy for low pay workers in

Korea as in many other countries. The Minimum Wage Act was enforced in 1988 along

with the massive eruption of the labour union movement in the late 1980s (Yun, A-L,

2014). The minimum wage has been applied to all employers with one employee or

more since 2000, with the exception of domestic workers, seamen and those with a very

limited working capacity due to their physical and mental disabilities and has begun to

get attention from trade unions and the general public in the early 2000s. A

subminimum for older workers have been discussed within the Minimum Wage

Commission in the late 2000s, with the concern of likely disemployment among them.

However, the discussion was not based on any evidence of the minimum wage effects

on older workers, and the consideration of a subminimum for the group of workers was

finally discarded for the reason that it would be against the Act on Prohibition of Age

Discrimination in Employment and Elderly Employment Promotion [고용상 연령차별

금지 및 고령자고용촉진에 관한 법률]. The tendency that minimum wage discussion

has not been founded on research evidence, especially with regard to older workers, is

partly because research on the minimum wage in Korea has paid less attention to older

workers than other groups.

Based on these circumstances surrounding older workers presented so far, this thesis

explores the effects of the minimum wage on older workers in South Korea. How

10

minimum wages affect the distributions of wages and family incomes, employment, and

the exit from poverty among older workers are examined through a comparison between

two consecutive government periods which have a large gap in the rates of increase in

the minimum wage. The comparison is to figure out whether the differences in the rate

of increase in the minimum wage lead to different effects of the minimum wage among

older workers, on the one hand and to verify the relevance of theoretical explanations

for the effects, on the other hand. This thesis also aims to construct an alternative

theoretical explanation for the effects of minimum wages which takes account of both

economic and political attributes of the minimum wage as a public policy. Based on the

alternative theoretical framework, the empirical results of the minimum wage effects on

older workers will be further discussed, which involves a qualitative analysis of the

minimum wage fixing process.

Research Questions

Three closely related questions are addressed in this thesis. First, how the minimum

wage affects the distributions of wages and family earnings, employment and exits from

in-work poverty among older workers. In more detail, this includes how the minimum

wage changes the distributions of wages and family earnings of older workers, whether

increases in the minimum wage reduce older workers’ employment and whether

increases in the minimum wage raise the odds of exit from in-work poverty among

older workers. These questions aim to figure out the overall characteristics of economic

effects of the minimum wage on older workers. Second, whether different rates of

increase in the minimum wage have different effects among older workers. The second

question is to examine variations and uniformity in the effects of the minimum wage

among older workers by comparing two government periods when the minimum wage

was increased at distinctively different levels. This examination has special interests in

the effects on employment and on the exit from in-work poverty. With regard to the

employment effect, Fang and Gunderson (2009) argue that the positive employment

effect for older workers is a typical characteristic of the minimum wage in Canada,

compared to the fickle employment effects for younger workers. This thesis is interested

in whether non-adverse employment effects are more likely among older employees.

Also, it has been widely argued, based mainly on younger workers who are paid a

minimum wage but not the primary earners in their households, that the minimum wage

11

is a blunt tool to help working families in poverty. However, in as much as older

workers are more likely to be the primary earners in their households, whether higher

minimum wages tend to give better chances of leaving poverty than lower ones should

be examined among older employees. Further, this comparison of two government

periods with different rates of increase in the minimum wage is connected to the third

question of exploring the relation between the minimum wage effects on older workers

and the political factors involved. The third question, how the empirical results of the

minimum wage effects on older workers can be understood is to seek an alternative

explanation for the minimum wage effects in which both political and economic

attributes of the minimum wage as a public policy are taken into account. A political

economy theoretical framework for minimum wage effects will be proposed after the

review of the existing explanations; and the empirical results of minimum wage effects

on older workers will be further discussed through a contextual analysis which is

conducted based on the newly suggested theoretical framework.

Contributions

This study could help to fill the gap in the literature with regard to the impact of

minimum wages on older workers. As stated earlier, the subject of this thesis was

motivated by the vacuum in the research on older workers’ earned income. Particularly,

minimum wage research has excluded older workers as a research focus, as Fang and

Gunderson (2009, p.371) stated ‘to my knowledge, there are no studies that focus on the

impact on older workers’. Even after the first and only study, there has been little

subsequent work exploring the effects of the minimum wage focusing on older workers.

Considering the facts that a growing number of older workers have been working in

their later life for financial reasons, the number of older adults living in poverty has

been sharply growing, and the transformed labour market conditions are more likely to

provide low paid jobs to older workers, this study takes a significant step in the research

on low paid older workers’ earnings and, more specifically, in the one on the minimum

wage effects on older workers by extending the research subjects to wages, family

incomes, employments and poverty as well as by adding new empirical evidence.

Moreover, this study is the first try to explore the impact of minimum wages on older

workers in South Korea. Although the effects with regard to older workers have been

partially presented in some studies and a study examined the effect of the minimum

12

wage on labour demand in surveillance or intermittent work in which the majority of

workers were the aged, there are no comprehensive analysis that focused on older

workers. This study could provide a starting point for the study of the minimum wage

effects on older workers in Korea.

This thesis is also likely to broaden and deepen understanding of the minimum wage

effects. One of the main purposes of this research is to explore an alternative

explanatory framework for the minimum wage effects and to conduct a contextual

analysis of the minimum wage fixing process based on the framework for a fuller

understanding of the effects. Economic theories which have been exclusively dominant

in minimum wage research provide limited accounts with regard to some controversial

effects of the minimum wage, despite their strong explanatory power. The alternative

explanatory framework which will be constructed in this thesis as a result of reviewing

existing theories contains political attributes as well as economic ones of the minimum

wage as a public policy, in as much as the framework takes into account the policy

process relating to the minimum wage. The point that the framework is based on

minimum wage policy process indicates that the explanatory framework for the

minimum wage effects newly proposed in this thesis will be universally applicable,

though this study focuses on older workers and the minimum wage fixing process. The

contextual analysis of the minimum wage fixing process based on the theoretical

framework will provide a more concrete understanding of the resulting effects of the

minimum wage. This also calls attention to the importance of using both quantitative

and qualitative research methods in research field of the minimum wage.

Further, this study would provide important implications for minimum wage policy,

particularly with regard to the target groups of the policy. A growing number of older

workers in many countries are expected to have been under the influence of the

minimum wage, but they have supposedly been a low concern group in the minimum

wage policy process. This study calls the attention of decision-makers involved in the

minimum wage policy process to the necessity of considering older workers as a

vulnerable group in the labour market who are critically influenced in many ways by

minimum wage policy. Further, the discussion of the empirical results through the

contextual analysis of the minimum wage fixing process provides the decision-makers

with policy implication of how the interaction between the minimum wage and other

13

closely related policies, which is changed typically by a government’s policy orientation

and the dynamics of power among stakeholders, can be related to the minimum wage

effects on a specific target group. This is different from the typical emphasis on the

earned income tax credit (EITC) as an alternative measure for the increase in the

minimum wage to boost income among low paid workers.

The Structure of the Thesis

The remaining chapters of the thesis are broadly divided into four parts. The first part is

the basic background literature and data survey of the minimum wage and older workers

in the labour market (Chapter Two and Three). The second part is a comprehensive

review of existing literatures on the status of older workers in the labour market and on

the effects of the minimum wage with regard to older workers (Chapter Four). In the

third part, theoretical perspectives and research methods will be discussed (Chapter

Five). The last part is the Korean case study intended to examine the three main

research questions and a Conclusion (Chapter Six, Seven, Eight and Nine).

Chapter Two looks at the minimum wage systems among the OECD countries. Using

the ILO Working Conditions Laws Database of Minimum Wages, minimum wage fixing

procedures and criteria are analysed to identify the characteristics of the minimum wage

and to narrow down the research focus based on the identified purposes of minimum

wages. Chapter Three surveys the characteristics of older workers in terms of how a

significant proportion of older workers is active in the workforce, what sectors,

occupations and types of contract older workers are in and what their wage levels are.

Chapter Four consists of broadly two parts. In the first part of the chapter, the research

about the status of older workers in the labour market will be reviewed. The main

purpose of this review is to figure out what factors determine the status of older workers

in the labour market. This review shows that little attention has been paid to the issue of

earned income among older workers. For the second part of this chapter, research about

the effects of the minimum wage is reviewed, with a special focus on older workers.

This displays how little attention has been paid to older workers in the field of minimum

wage research.

Chapter Five also consists of broadly two parts. In the first part, a variety of theories of

the minimum wage, such as the neoclassical models, monopsony models, the Keynesian

14

approach and political economy approaches are discussed. This is ultimately for

exploring a plausible theoretical framework to explain the effects of the minimum wage

based on both political and economic attributes of the policy. A political economy

framework for an enhanced understanding of empirical results will be proposed at the

end of the first part. The second part describes research methods, data, and analyses

procedures used for this research. In this part, the necessity to employ a contextual

analysis with qualitative data as a supplementary method for understanding the

minimum wage effects is argued in association with the alternative theoretical

framework sought in the first part of this chapter.

Chapter Six, as the first chapter of the case study of South Korea, provides an overview

of the national minimum wage and the status of older workers in the labour market in

Korea. It includes brief history of the Minimum Wage Act [최저임금법], its current

system and changes in its level and influence and also surveys labour market

participation and employment among those aged 50 years and over, industrial sectors,

occupations and types of employment which they are involved in and their wage levels.

In the final section of this chapter, public debate and empirical studies about the

minimum wage in Korea are reviewed with a special interest in older workers. Chapter

Seven examines the first two research questions, how minimum wages affect the

distributions of wages and family incomes, employment and the exit from poverty

among older workers and whether the effects of the minimum wage on older workers

are differentiated by its level of increase. Using the Korean Labour and Income Panel

Study (KLIPS) [한국노동패널], linear least squares (OLS) and fixed-effects models are

estimated to identify the effects on the distributions of wages and family earnings and to

demonstrate the aggregate employment effect among older workers. Multilevel discrete-

time event history models for competing risks are also used to verify both transitions

from being employed and from being poor while working among older workers caused

by the increase in the minimum wage. All of the employed models compare two periods

of Korean governments, the Roh, Moo-Hyun government (2003-2008) which

maintained the highest rate of increase in the minimum wage, except the government

period in which it was introduced, and the Lee, Myung-Bak government (2008-2013)

which had the lowest rate of increase, in order to evaluate whether the minimum wage

has a typical economic effect on older workers, regardless of its level of increase.

Chapter Eight discusses how the results from the empirical analysis conducted in the

15

previous chapter can be understood with the political economy framework proposed in

Chapter Five. This involves a contextual analysis mainly of the Minimum Wage

Council’s annual report on the minimum wage fixing process. Chapter Nine, the

Conclusion summarises the main findings of this research and briefly reviews

contributions and policy implications, followed by limitations of and reflection on

research methods exploited.

16

Notes

1 As Sabia (2010, p.593) summarised, there are two main explanations with regard to the minimum wage increases ‘providing little more than symbolic support’ to the working poor. Stigler, G. (1946, The economics of minimum wage legislation, American Economic Review, 36, 358-365), Burkhauser, R.V., Couch, K.A. and Glenn, A.J. (1996, Public policies for the working poor: The earned income tax credit versus minimum wage legislation, Research in Labour Economics, 15, 65-109), and Burkhauser, R.V. and Sabia, J.J. (2007, The effectiveness of minimum wage increases in reducing poverty: Past, present, and future, Contemporary Economic Policy, 25(2), 262-281) argue that ‘the relationship between earning a low hourly wage rate and living in poverty is weak and has become weaker over time’ (Sabia, 2010, p.593). Neumark and Wascher (2002, Do minimum wages fight poverty? Economic Inquiry, 40(3), 315–33), Neumark, Schweitzer, and Wascher (2004, Minimum wage effects throughout the wage distribution, Journal of Human Resources, 39(2), 425–50; 2005), and Sabia (2008, Minimum wages and the economic wellbeing of single mothers, Journal of Policy Analysis and Management, 27, 848–66) contend that while some families of low-skilled workers in employment will be out of poverty by an increase of the minimum wage, other low-skilled workers will lose their jobs or have their hours cut, leading to reduction in their income and fall their families into poverty (Sabia, 2010, p.593).

17

Chapter Two The Minimum Wage

Introduction

The idea of a minimum wage was created in the late nineteenth century for the purpose

of alleviating the conditions of “sweated labour” working for very low wages

(Cunningham, 2007; Neumark and Wascher, 2008)1. It can be defined currently as the

wage floor ensuring by law or by collective bargaining that wage earners receive at least

a minimum of pay protection. The International Labour Organization (ILO) has

provided a definition of minimum wages in the same vein:

the lowest level of remuneration permitted … which in each country has the force

of law and which is enforceable under threat of penal or other appropriate sanctions.

Minimum wages fixed by collective agreements made binding by public authorities

are included in this definition (The ILO Committee on Employment and Social

Policy, 2009, p.2)2.

Minimum wages, which are used as a major labour market instrument by the majority of

countries the world over, present the main idea of the definition provided by the ILO

whether or not they have ratified the ILO’s Conventions on minimum wage fixing

procedures.3

In principle, the minimum wage is a very clear social policy intended to meet the

demands of social justice for low paid workers. However, in reality, things are much

more complicated as a variety of legislative texts and practices exist country by country.

The current expansion from minimum wages to a living wage in debates or in practice

not only reflects the complexities of minimum wages in effect but also adds to them.

Those complexities lead to difficulties in generalising about the issues of the minimum

wage, such as its employment effect, its efficacy as a tool of poverty reduction and its

effectiveness for protecting wages and reducing inequalities. A plausible starting point

for dealing with the issues of the minimum wage is to look at its systems implemented

across countries. The intricate array of all minimum wage systems can be captured to

18

some extent by reviewing motivations for the creation of minimum wages and

minimum wage fixing procedures and criteria held in legislative texts. Since minimum

wage systems also vary over time, the changes in the systems in response to social,

economic, and political flows should be taken into account for an enhanced

understanding of the operation of minimum wages. The first section of this chapter will

give an overview of motivations for the creation of minimum wages, basically on

reference to Starr’s work on it (1993, quoted in Cunningham, 2007). Then, the various

minimum wage fixing procedures and criteria will be reviewed in the following two

sections, using the ILO’s Working Conditions Laws Database of Minimum Wages 2012.

In these sections, we identify the classification of minimum wages, based on the work

accomplished by Eyraud and Saget (2005) and Nolte and Ghosheh (2010). Finally,

founded on the review of legislative texts, the present focus of minimum wage systems

around the world will also be discussed.

Motivations behind Minimum Wages

Motivations for introducing minimum wages reveal the two attributes of social justice

and economic adjustment at the same time. The types of motivations Starr (1993,

quoted in Cunningham, 2007, p.9) suggested - 1) protection of the most vulnerable, 2)

poverty reduction, 3) payment for inputs, 4) fair labour standards, 5) fair competition

and 6) macroeconomic objectives. The rationale for the creation of minimum wages

across countries explicitly or implicitly includes some of the motivations, and according

to time and place, the influence of pressure groups and the relation to other linked

systems, the motivations focus of a minimum wage moves from social justice to

economic adjustment and vice versa. In this regard, changes in the purpose of minimum

wages reflect the transition in main concerns with regard to the minimum wage.

Protection of the most vulnerable is the objective where low paid workers ‘with the

least bargaining power and the most inhuman living standards were the targets of the

policy’ (Cunningham, 2007, p.9). The early minimum wages set by law from the late

19th to the early 20th in New Zealand, Australia, the United Kingdom and the United

States resulted from the concern about “sweatshops” which were often thought of as

19

employing children, women and immigrants. The population addressed in the early

minimum wages worked under the most deprived working conditions at severely low or

even no pay, and furthermore, they did not have any organisational power to negotiate

the improvement of their working conditions including wages.

The aim of poverty reduction is in the same context of protecting the most vulnerable;

however, it identifies the most vulnerable as the poor and intends to protect workers in

all industries against unacceptably low wages by setting a generally applicable lower

rate, instead of fixing rates in line with individual industries and occupations. However,

greater flexibility is shown in practice through general minimum rates fixed at different

levels for various regions or broad industrial sectors, the exemption of a certain group of

workers, or sub-rates for apprentices, young workers and the handicapped; and also, this

purpose ‘presupposes that the influence of minimum wages on average wage

movements is marginal’ since their role is ‘confined to providing ‘safety net’ protection’

(Starr, 1933, pp.40-41). General rates with the basic floor concept are the most widely

used form of minimum wages in recent years, particularly among industrialised

countries.

Payments for input are associated with the interest in what should be paid by employers

rather than the concern about who should be targeted. The primary logic of this

objective is that employers are required to pay for the reproduction of labour and thus

‘wages should be sufficient to cover the cost of food, shelter, clothing and other

necessities that create labour’ (Cunningham, 2007, p.9). The original Wisconsin

minimum wage law in the United States, similar to most of other early minimum wage

laws (Arizona, Arkansas, California, Colorado, D.C., Kansas, Massachusetts,

Minnesota, Nebraska, North Dakota, Oregon, Puerto Rico, South Dakota, Texas, Utah,

Washington) before the end of the 1920s, provided for a minimum wage ‘sufficient for a

worker to maintain himself or herself under conditions consistent with his or her

welfare,’ defined to be ‘reasonable comfort, reasonable physical well-being, decency

and moral well-being’ (Thies, 1991, pp.717-719).

20

The motivation of fair labour standards begins with the contradiction between the idea

that all occupations and industries should have a “fair” wage which is ideally identified

by collective bargaining and the fact that there are some industries which are unable to

organise (Cunningham, 2007). In this regard, the objective of fair labour standards

views ‘a minimum wage set by government as the second-best solution’ (Cunningham,

2007 p.9) for ‘a ‘common rule’ to promote the application of the principle of equal pay

for equal work and to reduce areas of industrial conflict’ (Starr, 1993, p.24). The

Australian states of Victoria and Tasmania can be regarded as the cases of minimum

wages with this role. In those states, the wages board systems, although originally

intended for the protection of the most vulnerable workers at the turn of the nineteenth

century into the twentieth century, ‘have evolved into the predominant institutional

framework for determination of terms and conditions of employment’ in which

employer and employee representatives negotiated ‘the minimum rates for various

occupations in individual industries’ (Starr, 1993, p.28). Japan also has experience with

this role of minimum wages. Up to 1967 most minimum wages were determined on the

basis of inter-employer agreements and applied to only a part or all of an industry in a

given prefecture; since then minimum wage rates have been mostly fixed on the basis of

recommendations of minimum wages councils in the 47 prefectures of the country

(Starr, 1993, p.28).

The aim of fair competition is in line with the objective of fair labour standard;

however, it derives from a concern with unfair competition for factor inputs in

production. Among employers, ‘there is the desire to isolate wages from excessive

competitive pressures, the reasoning being that while producers should be free to

compete in matters of price, design, quality of product or service, it is unfair for

competition to be based on a bidding down of workers’ wages’; in this view, a minimum

wage set through a collective decision-making procedure is conceived of as a means of

preventing such unfair competitive pressures (Starr, 1993, p.24). In several provinces of

Canada, there were ‘special industry rates applied to a limited number of industries

where the problem of unfair competition with respect to wages was believed to be

particularly acute’ (Starr, 1993, p.29). In Ontario, in 1974, there were 84 schedules in

force concerned the clothing industry, the needle trades, construction and barbers’ shops

21

but have declined (Starr, 1993, p.29).

Macroeconomic objectives are related to the likely impacts of minimum wage on ‘the

entire wage distribution, which may lead to economic growth, inflation control, or

political gains’ (Cunningham, 2007, p.9). The use of minimum wages as a

macroeconomic policy instrument assumes that the minimum wage rates ‘will to a large

extent determine the wages actually paid by many workers’ either by their being fixed at

comparatively high levels or by of a comprehensive structure of industrial/occupational

rates linked to them and strengthen the purchasing power of wages which leads to the

increase in the level of demand for goods and services and accordingly, to avoidance of

serious unemployment or of the increased costs associated with the higher wages; ‘it is

also believed in this case that minimum wage fixing can be used to gain increased

government control over wage movements and wage structures without producing

major adverse effects on the resource allocation and incentive functions of wages’

(Starr, 1993, pp.47-48). Unlike most developed countries where minimum wage

systems intended to provide ‘safety net’ protection rather than to have a major influence

on prevailing wages, in a large number of developing countries, such as a number of

East African countries after independence, Colombia, Mexico and Costa Rica, minimum

wages have been often conceived of as the key instrument of wage policy (Starr, 1993).

The motivations behind a minimum wage can be traced by the minimum wage fixing

procedures and criteria employed in each system. Since minimum wage fixing

procedures are ostensibly involved in who makes decisions, they do not immediately

disclose the particular purpose of a minimum wage; instead, discussions and decisions

made in the procedures at a given time and place tell us about the specific purpose in a

certain case rather than a broad objective of a minimum wage. These insights may be

achieved by an in-depth analysis of the procedures. A variety of fixing procedures

across countries that are reviewed in the following section imply this point. In contrast,

each of the fixing criteria is more promptly linked to different aims (Eyraud and Saget,

2008); and hence, analysing the combination of the fixing criteria in a country may give

more concrete insights into what general objectives a certain country focuses on in

terms of its minimum wage system. This will be looked at in another section of this

chapter.

22

Minimum Wage Fixing Procedures

Different minimum wage fixing procedures exisitng around the world can be classified

by two major independent factors: the involvement of government or collective

bargaining as the key decision-makers and the number of minimum wage rates (Eyraud

and Saget, 2005). Based on consideration of those factors, countries can be sorted into

four categories in principle: a single rate for the whole country with the government as

the key decision-maker; multiple rates varying by region or by sector and/or occupation

with the government as the key decision-maker; a single rate for the whole country

determined by collective bargaining; and mutiple rates varying by region or by sector

and/or occupation determined by collective bargaining. These categories can be

compartmentalised according to the extent of government intervention and the

minimum wage fixing level. This means that whether government or collective

bargaining is involved as a key decision-maker is categorised again by the extent of

government intervention as 1) the government sets a minimum wage alone; 2) the

government determines a minimum wage following the direct consultation with the

social partners separately; 3) the government fixes a minimum wage based on advice

and/or recommendation of specialised committee; 4) a special body, mainly a tripartite

or bipartite committee, sets a minimum wage, and the government validate it; and 5)

minimum wages are set by collective bargaining without government intervention

(Eyraud and Saget, 2005; Nolte and Ghosheh, 2010); and whether a single rate or

multiple rates is applied can be divided again into five groups by the minimum wage

fixing level: 1) national level if minimum wages has universal coverage4, 2) regional

level if minimum wages are set by a regional wage-setting mechanism and vary across

the country; 3) national by sector and /or occupation if sectoral and/or occupational

minimum wages are set which apply to the whole country; 4) regional by sector and/or

occupation if sectoral and/or occupational minimum wages are set by regional wage-

setting mechanism; and 5) by sector and/or occupation if sectoral and/or occupational

minimum wages are set which do not automatically apply to all workers concerned in

the region or county (Nolte and Ghosheh, 2010).

23

Table 2.1 Minimum Wage Fixing Procedures among OECD Countries

Key decision-maker

Government¹ Collective bargaining

Government only

Tripartitie/bipartite involved

Government decision

following direct

consultation with the social

partners

Government decision based on recommendations

of specialized body (tripartite/

bipartite)

Special body’s decision

(tripartite/bipartite)

Number of

minimum wage rates

Single National Chile, Israel, Luxembourg, Netherlands, New Zealand, United States*

Czech Republic, Slovenia, Spain

Estonia, France, Hungary, Ireland*, Japan*, Republic of Korea, Portugal, Slovakia, United Kingdom

Australia*, Belgium*, Poland, Turkey

Multi- ple

Regional Canada*, United States*

Canada*, Japan*

Mexico*

National by sector

and/or occupat-

ion

Ireland * Australia*, Belgium*, Finland* Greece**, Mexico*

Austria*, Finland*, Iceland, Norway**

Regional by sector

and/or occupation

Canada* Canada*, Japan*

Austria* Switzerland

By sector/ occupa-

tion

Denmark, Germany, Italy, Sweden

Source: ILO, Minimum Wage Fixing Database 2012 Notes: 1. In many countries within this category, in certain sectors and/or occupations or at an enterprise level collective bargaining sets minimum wage rates that are higher than the minimum rates set by an government or an authority. These are not included in the table as the rates thus set are an addition to the basic statutory minimum. * Some countries have several types of fixing mechanisms. ** The categorisation of these two countries is not in accord with the previous analysis among which the most recent is conducted by Nolte and Ghosheh based on the ILO’s minimum wage fixing database 2009.

24

Table 2.1 shows a reclassification of the minimum wage fixing procedures among the

OECD countries in 2012, based on the vital factors identified by Eyraud and Saget in

2005 and by Nolte and Ghosheh in 2009, as explained above. In the majority of

analysed countries (64.7 per cent), the government sets a single minimum wage rate at

the national level. Among the countries belonging to this category, the most popular

system is the government determination of a single national minimum wage based on

the recommendation of a special body usually composed of employee, employer and

government representatives (26.5 per cent of the total); the second most commonly used

system among the OECD countries is where the government sets a single national

minimum wage alone without any obligation to consult the social partners or a special

body (17.6 per cent of the total); next is that of a tripartite committee fixing a single

national rate and the government simply confirming the decision (11.8 per cent of the

total); and the government determination of a single national rate following the direct

and separate consultation of the social partners is found in three among the analysed

countries (8.8 per cent of the total). At this point, it should be noted that, in practice, the

importance of consultation is recognised even in the case where the government is the

only key decision-maker, and the three types except minimum wage determination by a

special body are often very similar (Eyraud and Saget, 2005). Regardless of the

variation by the extent of the government intervention, a single national minimum wage

determination by the government is found among the most developed countries such as

France, Japan, the United Kingdom and the United States, and also among the countries

which have become OECD members since the 1990s, for instance, Chile, Czech

Republic, Estonia, Hungary, Israel, Republic of Korea, Slovakia and Slovenia. Indeed,

eight of the nine countries which have relatively recently joined the OECD have

adopted one of the systems in which the government plays the main role in setting a

single national minimum wage.

Meanwhile, minimum wage determination by collective bargaining is observed only in

multiple rates, in three fixing levels. This means that there is no case in which collective

bargaining fixes a single national rate or regional rates applying to all workers in a

region or area. In nine countries (26.5 per cent), collective bargaining determines

multiple rates in three fixing levels: in Denmark, Germany, Italy and Sweden, the social

25

partners directly negotiate wages through sectoral and/or occupational agreements, and

thus no minimum wage actually exists; in several countries including Iceland and

Norway, a national multi-sectoral agreement sets a single minimum wage rate for the

whole country; and in Switzerland, multi-sectoral agreements varying by region

determine a minimum wage rate for a region or area. All countries belonging to the type

of minimum wage determination by collective bargaining are well developed nations

with high per-capita income levels.

A minimum wage may also be set through decentralised collective bargaining in most of

countries adopting a nationl minimum wage determination by government, which

normally must not be less than the national minimum wage level. Apart from this, nine

of the thirty four OECD countries have multiple types of fixing mechanisms. Australia,

Belgium, Ireland, Japan, Mexico, and the United States have more than two different

fixing levels in a key decision-maker; and Austria, Canada and Finland have various

combinations of a key decision-maker and a fixing level. Eyraud and Saget (2005)

suggested two reasons for minimum wage rates to be set by region, the administrative

and political stucture of the country and allowance for differences in living standards.

Although they may not cover the whole range of countries, the two reasons can work as

explanation for the countries having multi-types of minimum wage procedures.

Minimum Wage Fixing Criteria

Minimum wage fixing procedures are related to who and how a minimum wage is set

and at what level it works. Minimum wage fixing criteria, meanwhile, are concerned

with what motivations or objectives a minimum wage has. Countries take different

criteria into account for minimum wage setting, and it means that the aims of minimum

wage vary by country according to the criteria considered. In this regard, as the

motivations behind minimum wages disclose the two properties of social justice and

economic adjustment as mentioned in the eariler section, the criteria are divided into

two broad categories: the first is of more social nature addressing the needs of workers,

and the other is more economically oriented, focusing on the country’s economic

conditions (Eyraud and Saget, 2005). Among the criteria mainly covered by the ILO’s

26

database 2012, the level of wages and income in the country, the needs of workers and

their families, social security benefits, inflation rate and the cost of living can be seen as

having more of social nature, while productivity, capacity of enterprises to pay and the

level of employment are deemed to be of greater economic concern. Economic

development can be regarded as cutting both ways depending on the economic situation

in a country.

Table 2.2 shows the criteria of minimum wage fixing taken into consideration among

the OECD countries. The criteria most frequently considered in these countries are the

level of wages and income and the cost of living, followed by economic development,

inflation rate and the need of workers and their families in sequence. Then both

productivity and the level of employment rank fifth, and social security benefits and the

capacity of enterprises to pay in order are at the bottom of the scale. The level of wages

and income in the country are taken into account for mostly three objectives when

determining the minimum wage rate: the equal pay for equal work, the prevention of the

wage gap expansion and/or the allowance for the existing regional difference in wages.

The purpose, the equal pay for equal work is not clearly stated in other countries, but in

Australia, the Fair Work Act 2009 specifies that Fair Work Australia, the minimum

wage fixing panels, must ‘have regard to the principle of equal remuneration for work

of equal or comparable value’ (Fair Work Act 2009 §284(1)(d), cited from the ILO legal

database). The object to avoid the expansion of the wage gap in a country is the most

expected outcome by the countries considering the level of wage for minimum wage

fixing. France, Ireland, Israel and the Netherlands explicitly determine it about the

criterion: in France, Labour Code states that ‘annual adjustment to the national

minimum wage rate should eliminate the imbalance between the national minimum

wage rate and the progression of general economic conditions and income’ (Labour

Code §L3231-9, cited from the ILO legal database); in Ireland, ‘if no agreement

between employer and employee representatives has been reached, the Labour Court

when making a recommendation to the Minister concerning the minimum wage rate has

to take the movement to earnings since the last minimum rate of pay into account’

(National Minimum Wages Act S.13(5), cited from the ILO legal database); and very

strictly enforced in Israel and the Netherlands, ‘the minimum wage is 47.5 per cent of

27

Table 2.2 Minimum Wage Fixing Criteria among OECD Countries

Social nature both Economic oriented

Level of wages

Social security benefits

Needs of

workers and their

families

Cost of

living

Inflation rate

Econom -ic

develop ment

Produc tivity

Capacity of

business to pay

Level of Employ -ment

No criteria

Australia √ √ √ √ √ √ √ √

Austria √ Belgium √

Canada √ Chile √ √

Czech √

Estonia √ √ √ √ √

Finland √

France √ √ √ √

Germany √ Greece √

Hungary √ √ √ √ √

Iceland √ Ireland √ √ √ √

Israel √ √

Italy √

Japan √ √ √ √

Korea √ √ √ √

Luxembourg √ √ √

Mexico √ √ √ √ √ √

Netherlands √ √ √ √

New Zealand √ Norway √ Poland √ √ √ √ √ √ √ √

Portugal √ √ √ √ √

Slovakia √ √ √ √ √

Slovenia √ √ √ √

Spain √ √ √ √ √ √

Switzerland √ Turkey √ √ √ √

UK √

US √ √ √ √

Source: ILO, Minimum Wage Fixing Database 2012 Notes: Austria and Germany are included in the ILO’s database 2012 on the criteria of minimum wages, with no information and thus regarded as with no criteria while Denmark and Sweden are excluded in the criteria part of the database and taken away from the table.

28

the average wage’ in Israel (Minimum Wages Act §1, cited from the ILO legal database),

which is automatically raised in accordance with changes in average wage levels, and

‘minimum wage increases are automatically indexed to estimated increases in average

negotiated wage rates in both the public and private sectors in the current year and

revised accordingly by the Ministry of Social Affairs and Employment’5 in the

Netherlands (Minimum Wage and Minimum Leave Allowance Act Art.14, cited from the

ILO legal database). The use of the level of wages and income when determining

minimum wages is found in countries which allows for the regional difference in wages,

seeking more affordable levels of minimum wages in each region. In Japan, regional

minimum wages are specified with consideration given to the wages of workers in the

region.

Considering reference to the cost of living has the main purpose of maintaing the

purchasing power among low pay workers. The inflation rate also concerns the same

object, and the needs of workers and their families can be regared as part of the

reference to the cost of living and inflation rate (Eyraud and Saget, 2005). Due to the

analogous attribute among the three criteria, they are deemed alternatives to each other

for the same purpose, keeping the workers’ purchasing power. Twenty two countries

take at least one of the three criteria into considertation when minimum wage fixing;

and seven countries (Australia, Estonia, Mexico, Poland, Portugal, Spain and Turkey)

allow for all three. However, in order to achieve the proposed goal, the maintenance of

workers’ purchasing power, how strictly each criterion is enforced would be more

critical than how many criteria are considered. Some countries more concretely stipulate

what they do for the consideration of the cost of living, inflation rate and/or the needs of

workers and their families: in Begium and France, minimum wage rates are indexed to

consumer price index (CPI); in France, in particular, ‘the increase in purchsing power of

the minimum wage rate must be equal to at least half the total increase in the purchasing

power of the average hourly wage as set forth in the quarterly inquiry carried out by the

Ministry of Labour’ (Labour Code §L3231-8, cited from the ILO legal database);

Poland and Slovakia take account of price index development and consumer prices,

respectively; in Australia, the relative living standards is required to be allowed for in

addition to the needs of the low paid; and in Mexico, ‘the National Commission on

29

Minimum Wages shall consider the investigation and studies carried out to set the

minimum wages according to the budget that a family may require to cover material

needs, such as accomodation, household goods, food, clothing and transport, social and

cultural needs, such as the assistance to live performances, the performing of sports, the

use of libraries and other cultural activities; and the needs related to the education of the

children’ (Federal Labour Act Art.90, art.562 II(a), cited from the ILO legal database).

Economic development as a minimum wage fixing criterion is typically for preventing

too high increases and thus seeking more viable levels of minimum wages; and

economic oriented criteria including productivity, the capacity of enterprises to pay and

the level of employment also have the same purpose. However, economic development

can be more positively taken into account for minimum wage setting. In the case that

Eyraud and Saget (2005) refered to, where the minimum wages are set not just for

covering the basic needs of the poorest workers and their families but also for enabling

them to benefit from growth in general in the same way that other workers do, the

criterion, economic development are considered with the aims of poverty reduction. As

to this, more discussion will be presented in the following section.

At this point it should be noted that the allocation of each criterion to either a category

of social nature or one of economic concern, even among other criteria which were not

included in the discussion above, is not clear-cut in practice because the same criterion

can be used in a different way or even the opposite one. The inflation rate is usually

taken into consideration in order to avoid the purchasing power of workers being

reduced; however, in Ireland, ‘the Labour Court must consider the likely impact that a

recommendation made would have on inflation in the economy’ (National Minimum

Wage Act S. 13(5)). Social security benefits, though they were not included above as a

universal criterion for minimum wage fixing, are also taken into account for quite

different reasons. In Hungary, Japan and the Netherlands where social security benefits

are included as a minimum wage fixing criterion, the consideration of social security

benefits is for not upsetting the social benefits budget and preventing the big gap of

income among people; but, in China, the minimum wage should be set higher than

unemployment benefits in order to encourage the unemployed to find jobs (Eyraud and

Saget, 2005).

30

Whereas the majority of the OECD countries adopts a part of the criteria discussed

above, some states including Austria, Canada, Germany, Greece, Iceland, New Zealand,

Norway and Switzerland do not specify certain criteria for minimum wage setting.

However, they maintain the highest level of minimum wage with more than 1,000 US

dollar as a monthly minimum rate as of 2009 (Nolte and Ghosheh, 2010). In this regard,

the key is the fixing procedures and government’s focus when using a minimum wage

as a tool for macroeconomic and/or social policy.

Present Focus of Minimum Wage Research

From the review of the minimum wage systems, particularly of the fixing criteria

among the OECD countries using the ILO’s legal database 2012, the present focus with

regard to the purposes of the minimum wage can be narrowed down to three topics,

poverty, wage inequality and employment which are also the most controversial issues

in recent minimum wage debates.

Firstly, the motivations behind minimum wages is related to poverty reduction. The

earliest minimum wage was based on the fact that workers were paid extremely low

wages no matter how much they produced and needed to be provided the least necessary

to reproduce the labour; however, the notion of protection from poverty has been

expanded with the idea of workers’ needs. What is included into workers’ needs makes

the view of protection from poverty vary by time and space. Eyraud and Saget (2005,

p.41) suggested three transition patterns of the variation in the needs of workers: ‘a

move from the notion of basic needs to the requirements for a decent standard of

living’; ‘an extension of the notion of a decent standard of living to include different

aspects of social protection outside the enterprise such as pensions, family benefits,

etc.’; and ‘an even wider extension, where the minimum implies not just covering the

needs of the poorest workers and their families, but also enabling them to benefit from

growth in general in the same way that other workers do’. A shift from the concept of

basic needs to the prerequisite for a decent standard of living refers that not only the

goods to meet a worker’s physical necessity but also the requirements for a decent

living such as education, health and social security for retirement are considered when

31

calculating workers’ basic needs. Mexico is a good example where legislative texts

proclaim the requirements for a decent standard of living. As mentioned in an earlier

section, in Mexico, social and cultural needs including the assistance to live

performances, the performing of sports, the use of libraries and other cultural activities

as well as expenditures for subsistence or survival such as food, clothing and shelter are

taken into account as the basic needs of workers. An extended notion of a decent

standard of living to include different aspects of social protection indicates that the

minimum wage functions as a ‘social floor’ by linking it to social security benefits. In

these systems, retirement benefits, disability payments, unemployment benefits and/or

maternity benefits are adjusted with a rise in the minimum wage so that the purchasing

power of the most vulnerable beneficiaries is preserved. However, this notion has

inherent limitations in practice because a rise in the minimum wage may potentially

cause a large increase in social security costs and therefore many governments are

reluctant to raise minimum wage levels. In the widest notion that a minimum wage

enables the poorest workers and their families to benefit from growth in general in the

same way that other worekrs do, the minimum wage functions as an instrument of

redistribution rather than a tool for reduction of absolute poverty, depending on the

country’s strong economic growth. France in 1970 and the Netherlands in 1968 had the

objectives behind the implementation of minimum wages that the ‘guaranteed’

minimum wage was replaced by the minimum ‘growth’ wage (Eyraud and Saget, 2005,

p.45).

Secondly, minimum wages have relevance to the objectives of reducing wage inequality

and discrimination. Increases in the minimum wage have an effect of narrowing the

wage hierachy if there is no similar rise in higher wages. Although there are few cases

in which legislative texts declare the reduction of wage inequality as an objective, many

countries use minimum wage policy in order to cut back the wage gap at some stage.

However, the minimum wage works as a wage floor to push the whole wage distribution

onto the higher level without curtailing the wage gap in some developed countires, and

in this case lessening the wage hierarchy is not easily achieved by the minimum wage.

Minimum wages can be also used to help to oppose wage discrimination. Beyond the

fact that more women than men are in low paid jobs, which informs the wage inequality

32

between women and men, women are often paid less than men for the same work. This

kind of wage discrimination is often found in some vulnerable age groups as well.

Australia, Hungary and Portugal specify the principle of ‘equal pay for equal work’ in

their minimum wage law in order to prevent the wage discimination. The National

minimum wage in the United Kingdom also addressed the gender pay gap when newly

introduced (Grimshaw, 2011).

Thirdly, the minimum wage is connected with employment or unemployment. Unlike

the relation with other objectives discussed above, the minimum wage itself does not

intend to raise the employment rate nor to decrease the unemployment rate. Due to the

potential negative effects of the minimum wage on employment, several economic

motivations such as hindering a rise in production costs, preventing out-of-hand

inflation and maintaining general economic situation are behind the relation between the

minimum wage and employment or unemployment rates. In other words, such

economic objectives are accomplished in the way of impeding the increase in minimum

wage rates or not encouraging the increase by taking into account of the level of

employment or unemployment rates when the minimum wage fixing while the

ostensible reason for the minimum wage is still kept. Regardless of the fact that whether

or not minimum wages have negative effects on employment is very controversial and

far from conclusive, the level of employment has been adopted as an criterion of

minimum wage fixing for the economic purposes in more and more countires, and eight

among the OECD countries (Australia, Hungary, Ireland, Mexico, the Netherlands,

Poland, Slovakia and Slovenia) currently take into account of it for minimum wage

setting.

Conclusion

This chapter reviewed minimum wage systems across different countries. The minimum

wage is a social policy intended to meet the demands of social justice for low paid

workers in principle, but is more complex in practice in terms of its purpose, fixing

procedures and criteria. The motivations behind minimum wages, which have been

addressed since the advent of the first minimum wage in the late nineteenth century, can

33

be summed up as protection of the most vulnerable, poverty reduction, payment for

inputs, fair labour standards, fair competition, and macroeconomic objectives. This

suggests that the motivations behind minimum wages display both the attributes of

social justice and economic adjustment at the same time, and their focus moves from

one attribute to the other with a variety of influences such as time and place, pressure

groups, and other related systems. The purpose of a minimum wage can be traced by its

fixing procedures and the criteria employed in each system. While the fixing criteria are

immediately linked to different aims so that the combination of the criteria in a country

informs us of a general objective that country focuses on in terms of its minimum wage

system, the fixing procedures which are involved in who makes the decision and at what

level a minimum wage operates would reveal the purpose through discussions and

decisions within them. The fact that the majority of the fixing procedures adopt a type

of multilateral consultation and even the case where the government is the only key

decision-maker recognises the importance of consultation in practice, indicates the

value of discussions and decisions held in the fixing procedures for understanding the

effects of minimum wages. This is the main basis for the new theoretical framework

advanced in Chapter Five and, subsequently, to understand the quantitative empirical

results discussed in Chapter Eight. From the review of the minimum wage systems,

particularly of the fixing criteria, the present focus with regard to the purposes of the

minimum wage is narrowed down to three topics, poverty, wage inequality and

employment. These topics are the main ones examined in this thesis.

34

Notes

1 The early minimum wages set by law in New Zealand in 1894, in Australia in 1896, and in England in 1909 were mainly intended to prevent employers from hiring children, women, apprentices, or wage workers in certain industries such as chain making, lace finishing, paper and cardboard boxing making, and tailoring at no or very low pay (Neumark and Wascher, 2008). 2 It is originally from ILO (1992). Minimum Wages: Wage-Fixing Machinery, Application and Supervision. International Labour Conference, 79th Session, Report III (Part 4B) (General Survey), Geneva: International Labour Office. 3 There are two international labour standards on minimum wages provided by the ILO, the Minimum Wage Fixing Machinery Convention, 1928 (No.26) and the Minimum Wage Fixing Convention, 1970 (No. 131). Convention No.26 asks its ratifying countries ‘to create or maintain machinery whereby minimum rates of wages can be fixed for workers employed in certain of the trades or parts of trades (and in particular in home working trades) in which no arrangements exist for the effective regulation of wages by collective agreement or otherwise and wages are exceptionally low’ (Minimum Wage Fixing Machinery Convention 1928 No.26, Art. 1). Convention No.131 adjures the member who ratify it ‘to establish a system of minimum wages which covers all groups of wage earners whose terms of employment are such that coverage would be appropriate’ (Minimum Wage Fixing Convention 1970 No.131, Art. 1). Both conventions commonly require the full consultation of the social partners while any of them does not impose a national statutory minimum wage (Eyraud and Saget, 2008). 4 Nolte and Ghosheh (2010) noted that this also includes countries which determine minimum wages at national level, varying by area or region, such as Thailand and Vietnam. However, among the OECD countries, there is no case of this. 5 The minimum wage rate in the Netherlands may be frozen at its current level in the two situations: ‘if the average wage rise is considered too high, and as a result an increase of unemployment is expected, or if the increase in the amount of social welfare benefits, which are based on minimum wage rates, increases to such an extent that a significant increase in premiums or taxes is necessary’ (Minimum Wage and Minimum Leave Allowance Act Art.14 Decree on the adjustment of the minimum wage per 1 July 2012 Explanatory note).

35

Chapter Three Older Workers in the Labour Market

Introduction

Whether the minimum wage is an influential policy in particular for older workers has

not been focused on much in comparison to younger age groups. There are several

reasons for this academic inclination that will be shown in Chapter Four, such as the

relatively small proportion of older employees as a result of the trend towards early

retirement throughout the 1970s and 1980s, the big differential across countries in the

proportion of older workers in the labour force even after the end of the trend since the

mid-1990s, the focus of attention on employment and inequality issues among younger

workers and female workers along with the sign of the economic turndown and the level

of the unemployment rate worsening, and the tendency that workers with higher

education and higer wage work longer in their later life. Linking the minimum wage to

older workers, at least three points with regard to the characteristics of older workers

should be clarified: how a significant proportion of older workers is active in the

workforce; what sectors, occupations and/or types of employment contract older

workers are in; and how they are spread in the wage distribution.

In this chapter, the labour force participation rate and the employment rate of older

workers, the types of contracts and sectors that they are involved in, and their wage

level are reviewed. The labour force participation rate is defined as the proportion of

people of a given age who are employed, self-employed or actively seeking jobs as a

percentage of the total population in that same age group; and the employment rate, as

the proportion of employed or self-employed people of a given age as a percentage of

the total population in that same age. The labour force participation rate and the

employment rate reflect both the extent of older people’s needs to be engaged in work

and the capacity of an economy to embrace the needs, respectively. In addition to these

rates, it is to be noted that the unemployment rate of an age group typically refers to the

number of people of a given age who are neither employed nor self-employed but

36

actively seeking work as a percentage of the total number of economically active people

- employed, self-employed, or actively seeking work - in that same age group. The

unemployment rate implies a limit in the capacity of an economy to embrace people’s

need to work. A high unemployment rate is associated with high involuntary exit from

the labour market. This tendency is more likely to be among older workers, but

unemployment is not easily distinguished from exit from the workforce in data. In this

regard, the unemployment rates used in Chapter Seven of this thesis were measured as

the proportion of people of a given age who were neither employed nor self-employed

or in non-wage family business work as a percentage of the total population in that same

age group. Meanwhile, the types of contracts and sectors which older workers are

involved in and their wage level would suggest to what extent older workers are

engaged in low-wage work and potentially influenced by the minimum wage.

Labour Market Participation

As a strong trend to early retirement throughout the 1970s and 1980s came to an end in

the mid-1990s and during the 2000s, the labour force participation rate of older workers

has crept up (OECD, 2011). The participation rate for those aged 55-64 in the OECD

countries has increased significantly from 48.5 per cent in 1995 to 57.8 per cent in 2011,

compared to the slightly increased one for their ‘prime aged’ counterparts aged 25-54

from 79.8 per cent to 81.3 per cent and contrasting with the decreased rate of those aged

15-24 from 52.9 per cent to 47.2 per cent in the same year. Figure 3.1 shows that the

rise in the participation rate is also found in the older age groups such as those aged 65-

69 and over 70 although the rates of increase among the groups are not conspicuous as

much as the one among those aged 55-64. Unlike the decline among younger workers

aged 15-24 and the stasis among the middle aged workers aged 25-54, the participation

rate of older workers display a tendency to increase even after the financial crisis. The

rate increased by 1.4 percentage points in both groups aged 55-64 and 65-69 and by 0.3

percentage points among workers aged over 70 between 2008 and 2011. The increase in

the participation rate is remarkable particularly among women aged 55-64. As shown in

Figure 3.2, the rate for the group has increased by 16.0 percentage points, from 36.2 per

cent in 1995 to 52.2 per cent in 2015.

37

Figure 3.1 Labour Force Participation Rates by Age Group in OECD Countries, 1970- 2011 (%)

Source: OECD, Labour Force Statistics database

Figure 3.2 Labour Force Participation Rates of Workers Aged 55+ by Sex and Age Group in OECD Countries, 1995-2015 (%)

Source: OECD, Labour Force Statistics database

The increasing pattern of the participation rates for older workers since the mid-1990s is

found in many OECD countries, in particular among those aged 55-64, despite the

difference in the steepness of each rise. However, as seen in Figure 3.3, deviating from

this pattern, the participation rates have fluctuated narrowly and thus remained at a

relatively similar level in Greece, Iceland, Japan, Korea and Mexico; and the increase

after the financial crisis is not shown in Ireland, Japan, Portugal, the United Kingdom

and the United States in which the participation rates of older workers aged 55-64 were

almost flat between 2008 and 2011. In Poland and Turkey, the participation rates have

declined until 2008 and henceforth increased to a great extent. Along with some

0

20

40

60

80

100 15 to 24 25 to 54 55 to 64 65 to 69 70+

1995 2005 2008 2015 55 to 64 62.6 65.6 67.1 70.5 65 to 69 22.4 27.0 29.2 32.5 70+ 4.6 5.8 7.4 8.7

0 10 20 30 40 50 60 70 80

MEN

1995 2005 2008 2015 55 to 64 36.2 43.5 46.0 52.2 65 to 69 11.7 14.5 16.3 19.6 70+ 0.9 1.6 2.2 3.5

0 10 20 30 40 50 60

WOMEN

38

Figure 3.3 Labour Force Participation Rates of Workers Aged 55-64 among OECD Countries, 1970-2011 (%)

Source: OECD, Labour Force Statistics database

0

20

40

60

80

100

19 70

19 73

19 76

19 79

19 82

19 85

19 88

19 91

19 94

19 97

20 00

20 03

20 06

20 09

Australia Canada

Japan New Zealand

United States OECD countries

0

20

40

60

80

19 70

19 73

19 76

19 79

19 82

19 85

19 88

19 91

19 94

19 97

20 00

20 03

20 06

20 09

Austria Belgium France Germany

Luxembourg Netherlands Switzerland OECD countries

0

20

40

60

80

100

19 70

19 73

19 76

19 79

19 82

19 85

19 88

19 91

19 94

19 97

20 00

20 03

20 06

20 09

Denmark Estonia Finland Iceland Ireland Norway Sweden United Kingdom OECD countries

0 10 20 30 40 50 60 70

19 70

19 73

19 76

19 79

19 82

19 85

19 88

19 91

19 94

19 97

20 00

20 03

20 06

20 09

Greece Italy

Portugal Slovenia

Spain OECD countries

0 10 20 30 40 50 60 70

19 70

19 73

19 76

19 79

19 82

19 85

19 88

19 91

19 94

19 97

20 00

20 03

20 06

20 09

Czech Republic Hungary

Poland Slovak Republic

OECD countries

0 10 20 30 40 50 60 70

19 70

19 73

19 76

19 79

19 82

19 85

19 88

19 91

19 94

19 97

20 00

20 03

20 06

20 09

Chile Israel

Korea Mexico

Turkey OECD countries

39

exceptions in the pattern of the participation rates, the rate varies by country from 33.0

per cent in Turkey to 84.1 per cent in Iceland as of 2011, for example. The average of

the participation rates for older workers aged 55-64 in the OECD countries was 57.8 per

cent in 2011, which is very close to the rates only in two countries, the Netherlands and

the United Kingdom. The participation rates for older workers exceeed 60 per cent in

Australia, Canada, Chile, Denmark, Estonia, Finland, Germany, Israel, Japan, Korea and

the United States, and even 70 per cent in five countries, including Iceland, New

Zealand, Norway, Sweden and Switzerland. At the other end of the spectrum, eleven

countries (Austria, Belgium, Czech Republic, France, Greece, Ireland, Luxembourg,

Mexico, Portugal, Slovakia and Spain) have the participation rates of 40 per cent or

over but below the OECD average, and the other five countries (Hungary, Italy, Poland,

Turkey and Slovenia) have less than 40 per cent of older workers active in the labour

force.

The employment rate of older workers is a more critical measure of their labour market

conditions in as much as the rate informs of the extent that they are embraced in practice

in an economy. Linking to minimum wages, the rate enables a rough assumption about

the population size of older workers who are likely to be affected by minimum wages.

The employment rates for older workers in the OECD countries have similarities with

the labour force participation rates for them in the countries in many ways. First of all,

the employment rate of older workers aged 55-64 in the OECD countries averaged 54.4

per cent in 2011. It is still low compared to 75.4 per cent, the average rate among ‘core-

age’ workers aged 25-54; however, considering the fact that the average rate for younger

workers aged 15-24 hit 39.5 per cent in the same year, their lowest employment rate

since 1970, the rate for older workers is not insignificant. Second, as seen in Figure 3.4,

employment rates have steadily increased since the mid-1990s in all three groups of

older workers aged 55-64, 65-69 and over 70. The rise in the employment rate is found

in both men and women, and the rate has sharply increased particularly among women

aged 55-64, as in the labour force participation rate, from 34.6 per cent in 1995 to 50.0

per cent in 2015 (Figure 3.5).

40

Figure 3.4 Employment Rates by Age Group in OECD Countries, 1970-2011 (%)

Source: OECD, Labour Force Statistics database

Figure 3.5 Employment Rates of Older Workers Aged 55+ by Sex and Age Group in

OECD Countries, 1995-2015 (%)

Source: OECD, Labour Force Statistics database

Finally, as shown in Figure 3.6, the employment rate and its pattern also differ by

country as the labour force participation rate and its pattern do. In 2011, the

employment rates ranged from 31.2 per cent to 79.5 per cent among the OECD

countries. In six countries, including Australia, Israel, Japan, Korea, Norway and

Switzerland, the employment rates for those aged 55-64 were over 60 per cent, and in

Iceland, New Zealand and Sweden, over 70 per cent of the counterparts were

employeed. However, sixteen countries had the employment rates for the age group

below the OECD average, and in eight of them, the rates were less than 40 per cent.

0 10 20 30 40 50 60 70 80 90

15 to 24 25 to 54 55 to 64 65 to 69 70+

1995 2005 2008 2015 55 to 64 59.0 62.4 64.2 66.8 65 to 69 21.7 26.3 28.3 31.4 70+ 4.6 5.8 7.3 8.6

0 10 20 30 40 50 60 70 80

MEN

1995 2005 2008 2015 55 to 64 34.6 41.7 44.3 50.0 65 to 69 11.3 14.1 15.9 19.1 70+ 0.9 1.6 2.1 3.4

0 10 20 30 40 50 60

WOMEN

41

Nevertheless, the employment rates are not dispersed from the average as much as the

labour force participation rates, and thus the gap between the average and the

employment rate in each country is not as wide as the one between the participation rate

in individual country and its average.

Figure 3.6 Employment Rates of Workers Aged 55-64 among OECD Countries, 1970- 2011 (%)

Source: OECD, Labour Force Statistics database

0

20

40

60

80

19 70

19 73

19 76

19 79

19 82

19 85

19 88

19 91

19 94

19 97

20 00

20 03

20 06

20 09

Australia Canada Japan New Zealand United States OECD countries

0

20

40

60

80

19 70

19 73

19 76

19 79

19 82

19 85

19 88

19 91

19 94

19 97

20 00

20 03

20 06

20 09

Austria Belgium France Germany Luxembourg Netherlands Switzerland OECD countries

0

50

100

19 70

19 73

19 76

19 79

19 82

19 85

19 88

19 91

19 94

19 97

20 00

20 03

20 06

20 09

Denmark Estonia Finland Iceland Ireland Norway Sweden United Kingdom OECD countries

0 10

20 30 40

50 60

19 70

19 73

19 76

19 79

19 82

19 85

19 88

19 91

19 94

19 97

20 00

20 03

20 06

20 09

Greece Italy Portugal Slovenia Spain OECD countries

0

20

40

60

19 70

19 73

19 76

19 79

19 82

19 85

19 88

19 91

19 94

19 97

20 00

20 03

20 06

20 09

Czech Republic Hungary Poland Slovak Republic OECD countries

0

20

40

60

80

19 70

19 73

19 76

19 79

19 82

19 85

19 88

19 91

19 94

19 97

20 00

20 03

20 06

20 09

Chile Israel Korea Mexico Turkey OECD countries

42

Non-Regular Employment and Sectoral Employment Structure

What types of employment and what sectors the growing number of older workers are

in could be an indicator to predict how many older workers will be affected by

minimum wages. In general, non-regular employees such as part-time and temporary

workers are more likely than their full-time and permanent counterparts to be low-wage

workers; and the employees in sectors of industry, typically including the retail trade,

hotels and restraunts, transport, social services and some areas of manufacturing are

regarded as being engaged in low pay work, which are highly expected to be influenced

by minimum wages.

Workers aged 55 and over are more likely than workers aged 25-54 to be employeed in

part-time jobs in the OECD countries. As shown in Figure 3.7, while the incidence of

part-time employees aged 25-54 has ranged from 10.8 to12.4 since 2000, the one of the

counterparts aged 55-64 has maintained over 16 per cent during the same period. Part-

time employment is the most popular among workers aged 65 and over. Although the

rate has decreased in the OECD countries since 2000, it still remained 36.9 per cent in

2011, which was incomparable to the rates for other age groups. Looking at part-time

employment among older workers by sex, women are much more likely to be engaged

in part-time work than men, and the incidence greatly rises at age of 65 or over for both

(Figure 3.8).

Figure 3.7 Incidence of Part Time Employment by Age Group in OECD Countries, 2000- 2011 (%)

Source : OECD, Labour Force Statistics database, ‘Incidence of Full-Time Part-Time Employment Based on Common Definition’

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

15 to 24 23.3 24.1 24.6 25.3 25.9 26.3 26.3 26.4 27.2 29.2 29.3 28.5

25 to 54 10.8 11.0 11.1 11.2 11.4 11.6 11.4 11.4 11.4 12.2 12.4 12.0

55 to 64 16.5 16.6 16.3 16.6 16.5 16.3 16.3 16.3 16.2 16.5 16.6 16.3

65+ 42.9 42.8 42.4 41.3 42.5 41.8 41.2 40.8 39.9 40.9 39.5 36.9

0 10 20 30 40 50

15 to 24 25 to 54 55 to 64 65+

43

Figure 3.8 Incidence of Part Time Employment among Workers Aged 55-64 and 65+ by Sex in OECD Countries, 2001-2016 (%)

Source: OECD, Labour Force Statistics database, ‘Incidence of Full-Time Part-Time Employment Based on Common Definition’

But the incidence of part-time employment in the two age groups differs significantly

across countries (Table 3.1). In Ireland, the Netherlands and Switzerland, over 35 per

cent of workers aged 55-64 were working part-time in 2011 while the rate was under 10

per cent in five OECD countries including Chile, Czech Republic, Hungary, Slovakia

and Turkey; and the share of part-time employment among those aged 65 and older was

more than 45 per cent in 17 OECD countries and over 70 per cent in Austria, Belgium,

Germany and the Netherlands while the rate was below 20 per cent in Chile, Greece and

Turkey. However, even in countries where the incidence of part-time employment was

much lower than other OECD countries in the two older age groups (Chile and Turkey),

the rates increased by from 3.2 to 12.9 pertange points between 2000 and 2011. It is

assumed that such small proportion of part-time employment for older workers in the

two countries is due to the composition of types of employment in each economy. In

fact, the rate of part-time employment for the total employees in Chile increased from

1.6 per cent in 2000 to 8.6 per cent in 2011, and the rate in Turkey, from 2.5 per cent to

4.6 per cent during the same period; these are very low compared to the OECD average

which was 13.5 per cent in 2000 and 15.3 per cent in 2011 (OECD Statistics, Labour

Force Survey database, ‘Incidence of Full-Time Part-Time Employment Based on

Common Definition’).

2001 2008 2011 2016 Men 7.7 7.6 7.9 7.6

Women 28.2 26.3 26.2 24.8

0

5

10

15

20

25

30

55-64

2001 2008 2011 2016 Men 35.6 33.2 33.0 30.6

Women 52.2 47.6 46.7 45.0

0

10

20

30

40

50

60

65+

44

Table 3.1 Part Time Employees Aged 55-64 and 65+ by Country in 2011

Part-time employees

aged 55-64

Country Part-time employee

aged 65+

Country

Above the OECD

average, 16.3% (x ≥ 20%)

Australia, Austria,

Belgium, Germany,

Ireland, Italy,

Netherlands, New

Zealnd, Norway,

Switzerland, United

Kingdom

Above the OECD

average, 36.9% (x ≥ 45%)

Australia, Austria,

Belgium, Czech Rep.,

Denmark, Finland,

Frnace, Germany,

Ireland, Luxembourg,

Netherlands, New

Zealnd, Norway, Poland,

Slovenia, Sweden,

United Kingdom

About the OECD

average, 16.3%

(10% ≤ x < 20%)

Canada, Denmark,

Finland, France,

Greece, Iceland, Israel,

Luxembourg, Mexico,

Poland, Portugal,

Slovenia, Spain,

Sweden, United States

About the OECD

average, 36.9%

(25% ≤ x < 45%)

Canada, Estonia,

Hungary, Israel, Italy,

Mexico, Portugal, Slovak

Rep., Spain, United

States

Below the OECD

average, 16.3% (x <

10%)

Chile, Czech Rep.,

Hungary, Slovak Rep.,

Turkey

Below the OECD

average, 36.9%

(x < 25%)

Chile, Greece, Turkey

Source: OECD, Labour Force Statistics database, ‘Incidence of Full-Time Part-Time Employment Based on Common Definition’ Note: The data are not available for both age groups in Japan and Korea and for 65+ group in Iceland and Switzerland.

In the incidence of temporary employment in the OECD countries, there is no

significant difference between workers aged 25-54 and those aged 55-64. The rate

ranged from 8.7 to 10.0 between 2000 and 2011 in both age groups (Figure 3.9). Also,

the difference between men and women is not consistent in the two older age groups,

55-64 and 65 and older, and the gap between the gender groups has declined from 2001

to 2016 (Figure 3.10). But, as shown in Figure 3.9, employees aged 65 and older were

more likely than those in the two age groups to work in temporary jobs during the same

period. The incidence of temporary employees aged 65 and older in the OECD countries

was 19.5 per cent in 2011, but it also varies by country. Over 30 per cent of workers

aged 65 and older in Belgium, France, the Netherlands, Poland, Slovakia and Sweden

and even over 50 per cent of their counterparts in Czech Republic, Korea and Slovenia

45

were working in temporary jobs while the rate was under 10 per cent in six OECD

countries and below five per cent in Austria and Estonia (Table 3.2). Furthermore, there

is no common tendency of increasing or decreasing number of older temporary workers

among the OECD countries. However, it should be noted that, as seen in Table 3.2, the

incidence of temporary employment among workers aged 65 and older is, at least, over

10 per cent in 20 OECD countries, and the rate in each country has been maintained a

significant level, despite a sharp or smooth fluctuation between 2000 and 2011.

Figure 3.9 Incidence of Temporary Employment by Age Group in OECD Countries, 2000-

2011 (%)

Source : OECD, Labour Force Statistics database, ‘Incidence of Permanent Temporary Employment’

Figure 3.10 Incidence of Temporary Employment among Workers Aged 55-64 and 65+ by Sex in OECD Countries, 2001-2016 (%)

Source : OECD, Labour Force Statistics database, ‘Incidence of Permanent Temporary Employment’

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

15 to 24 24.3 23.3 23.5 23.7 23.9 25.1 25.5 25.6 25.1 24.9 25.4 25.3

25 to 54 8.8 8.7 8.9 9.1 9.2 9.8 10.0 10.0 9.8 9.6 9.7 9.9

55 to 64 8.8 8.6 8.7 8.8 8.9 8.8 8.9 9.0 8.8 9.0 9.1 9.1

65+ 18.4 17.3 17.6 17.5 17.2 16.8 17.7 18.1 18.1 19.9 19.3 19.5

0

5

10

15

20

25

30

15 to 24 25 to 54 55 to 64 65+

2001 2008 2011 2016 Men 7.8 8.3 8.8 7.9

Women 9.7 9.6 9.9 7.9

0

2

4

6

8

10

12 55-64

2001 2008 2011 2016 Men 18.1 18.5 19.2 17.1

Women 16.2 17.5 19.8 17.8

0

5

10

15

20

25 65+

46

Table 3.2 Temporary Employees Aged 55-64 and 65+ by Country in 2011

Temporary

employees aged 55-

64

Country Temporary employee

aged 65+

Country

Above the OECD

average, 9.1% (x ≥ 15%)

Chile, Japan, Korea,

Poland, Turkey

Above the OECD

average, 19.5% (x ≥ 30%)

Belgium, Czech Rep.,

France, Korea,

Netherlands, Poland,

Slovakia, Slovenia,

Sweden

About the OECD

average, 9.1%

(5% ≤ x < 15%)

Canada, Czech Rep.,

Filand, France, Greece,

Hungary, Iceland,

Ireland, Italy,

Netherlands, Portugal,

Slovakia, Slovenia,

Spain, Sweden, United

Kingdom

About the OECD

average, 19.5%

(10% ≤ x < 30%)

Canada, Chile, Finland,

Ireland, Itlay, Japan,

Portugal, Spain,

Switzerland, Turkey,

United Kingdom

Below the OECD

average, 9.1% (x <

5%)

Austria, Belgium,

Denmark, Estonia,

Germany, Norway,

Switzerland

Below the OECD

average, 19.5%

(x < 10%)

Austria, Denmark,

Estonia, Germany, Greece,

Hungary, Iceland, Norway

Source : OECD, Labour Force Statistics database, ‘Incidence of Permanent Temporary Employment’ Note: The data are not available for Australia, Israel, Mexico, New Zealand, and the United States.

Based on the Labour Force Survey (LFS) provided by Eurostat, older workers aged 55

and over in 25 European countries, which are mostly OECD members, are mainly

employed in ‘agriculture,’ ‘manufacturing,’ ‘whole sale and retail trade,’ ‘education’ and

‘health and social work’ in 2011 (Figure 3.11). Agriculture and education are typically

regarded as traditional sectors in which older workers are over-represented. However,

workers aged 65 and up unequivocally outnumbered other age groups in 2011 when it

comes to their employment share in agriculture while those aged 55-64 were the most

highly engaged in education in the year. Along with this, the significant portions of both

workers aged 55-64 and those aged 65 and over were working in‘whole sale and retail

trade’ and ‘health and social work’ in 2011. Except for education, the sectors in which

older workers are chiefly employed are regarded as those offering a high proportion of

low-wage jobs.

47

Figure 3.11 Sectoral Empolyment Structure by Sector and Age Group in EU 25, 2011 (%)

Source : Eurostat, LFS, NACE Rev. 2

Looking at the changes in sectoral employment by age group (Figure 3.12), the annual

employment of workers aged 55-64 had a relatively steep decline in ‘manufacturing’,

‘construction,’ ‘whole sale and retail trade’ and ‘agriculture and fishing’ between 2008

and 2011 whereas it increased in ‘health and social work activities’ by 1.3 percentage

points during the same period. The employment of workers aged 65 and older also

showed a significant drop in ‘agriculture and fishing,’ ‘manufacturing’ and ‘wholesale

and retail trade’ while it increased the most in ‘health and social work activities’ and

‘transportation, storage and communication’ by 1.6 and 0.8 percentage points

respectively between 2008 and 2011. A remarkable change in employment for this age

group is that the rate in ‘agriculture and fishing’ dramatically declined by 3.9 percentage

points during the period, which was not comparable with other sectors of industry and

the other older age group. In contrast with the significant drop in ‘agriculture and

fishing’ for the both older age groups, employment in ‘education’, another traditional

sector where older workers were regarded as over-represented, showed rises of 0.2 and

0.6 percentage points for the group of 55-64 and the one of 65 years and over,

respectively. Although they cannot be compared equally due to the limitation of data,

the analogous pattern of sectoral employment of older workers is observed in non-

European OECD countries. In New Zealand, older workers (55-64 years old) are mainly

Agricultur e, fishing

Manufact uring

Constructi on

Wholesale , retail trade

Transport ation,

storage, communic

ation

Accommo dation, food

service activities

Financial, insurance activities

Real estate

Public administr

ation, defence

Education

Health, social work

activities

Other services

15-24 2.8 14.1 8.2 21.7 6.0 10.2 2.3 0.5 3.9 4.0 9.1 3.2

25-54 3.4 16.2 7.5 13.5 8.5 4.1 3.3 0.8 7.4 7.5 10.7 2.4

55-64 5.5 14.4 6.7 11.6 7.1 3.2 2.7 1.1 8.6 9.8 12.3 2.5

65+ 17.4 7.9 5.3 13.7 6.0 3.8 1.4 2.1 3.2 6.3 9.2 4.3

0

5

10

15

20

25

15-24 25-54 55-64 65+

48

employed in health and social work, manufacturing, education and whole sale and retail

trade (15.4, 14.2, 13.2, 13.0 per cent, respectively) in 2008 (Boyd and Dixon, 2009).

Figure 3.12 Annual Employment Growth by Sector and Age Group in EU 25, 2008-2011

Source : Eurostat, LFS, NACE Rev. 2

Wage Levels

The age-earnings profiles describe the growth of earnings over the life cycle (Thornton,

Rodgers and Brookshire, 1997), and the most common specification found in the

literature shows an inverted U-shaped pattern between age and earnings. This takes the

form of increases in earnings in the early years, a peak at around middle age and a

decline around the age of 50-55 (Luong and Hebert, 2009; Casanova, 2013). This

pattern of the age-earnings profile can be interpreted, with caution, to mean that the

older employees aged over 50 years are more likely to be the low paid than their

younger colleagues. Under the conditon that there are a lack of data and research on

wages and earnings of older workers in an international context, the review of the

proportion of low-wage earners among older workers would be a way to assume their

relative wage levels.

Agricultur e, fishing

Manufact uring

Constructi on

Wholesal e, retail trade

Transport ation,

storage, communi

cation

Accommo dation, food

service activities

Financial, insurance activities

Real estate

Public administr

ation, defence

Education

Health, social work

activities

Other services

15-24 0.1 -2.1 -1.5 0.0 -0.3 0.9 -0.1 0.0 -0.1 0.7 1.5 -0.1

25-54 -0.1 -1.2 -0.9 -0.1 0.0 0.3 0.0 0.0 -0.1 0.3 0.7 0.0

55-64 -0.4 -1.0 -0.8 -0.5 -0.2 0.1 0.2 0.1 0.2 0.2 1.3 -0.3

65+ -3.9 -1.0 0.0 -0.6 0.8 -0.2 0.2 0.2 0.2 0.6 1.6 0.2

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

1.0

2.0

15-24 25-54 55-64 65+

49

The definition of low pay most commonly used is a level equivalent to two-thirds of the

median wage for all employees in the economy (Grimshaw, 2011, p.3). The Structure of

Earnings Survey (SES) provided by the European Commission (EC) follows it, defining

low-wage workers as those employees earning two-thirds or less of the national median

gross hourly earnings (Eurostat, 2013). According to the survey, 14.5 per cent of EU

employees aged 50 years and over were low-wage earners in 2010. The highest

proportions of low-wage older earners were observed in Estonia (31.2%), Poland

(23.8%), Germany (20.8%), Slovakia (19.8%), Czech Republic (19.1%), Hungary

(18.7%) and the United Kingdom (18.3%), and the lowest were found in Turkey (0.2%),

Sweden (0.7%), Norway (1.3%), Denmark (3.7%), Finland (4.7%)), France (4.7%) and

Iceland (5.4%) (Figure 3.13).

Figure 3.13 Proportion of Low-Wage Earners by Age Group in EU Countries, 2010 (%)

Source : Eurostat, SES (earn_ses_pub1) Note: OECD member states only are presented, and the data are not available for Greece.

Low-wage earners aged 50 years and over in the EU 27 countries have increased by one

percentage point between 2006 and 2010 (Figure 3.14). This is a significant rise in

comparison to the increase of 0.2 percentage points in the proportion of total low-wage

earners in the EU countries in the same period (Eurostat, 2013). Except that noticeable

decreases in low-wage older workers were observed in Slovenia (-3.2 p.p.), Portugal (-

2.4 p.p.), Belgium (-2.3 p.p.) and Ireland (-1.0 p.p.) and France (-0.9 p.p.), and minor

EU 27 Austria Belgium CzechRep Denmar

k Estonia France Finland German

y Hungary Iceland Ireland Italy Luxemb

ourg Netherl

ands Norway Poland Portugal Slovenia Slovakia Spain Sweden Switzerl

and Turkey United Kingdo

m

Total 17.0 15.0 6.4 18.2 7.7 23.8 6.1 5.9 22.2 19.8 9.1 20.7 12.4 13.1 18.1 7.3 24.2 16.1 17.1 19.0 14.7 2.5 11.0 0.2 22.1

<30 30.4 24.3 15.3 21.1 27.8 21.3 12.4 13.2 38.1 23.1 22.5 39.4 25.0 22.7 46.1 26.2 31.3 25.0 24.3 20.9 23.3 9.5 21.2 0.2 40.6

30-49 13.7 11.9 4.7 16.6 5.6 19.1 5.0 4.2 18.0 19.4 6.6 17.4 11.6 11.0 8.0 2.3 21.8 14.2 16.4 18.0 13.6 1.0 8.9 0.2 14.7

50≤ 14.5 12.5 3.0 19.1 3.7 31.2 4.7 4.7 20.8 18.7 5.4 17.4 9.1 9.2 8.5 1.3 23.8 12.8 14.3 19.8 11.0 0.7 7.8 0.2 18.3

0 5

10 15 20 25 30 35 40 45 50

Total <30 30-49 50≤

50

declines, in Norway (-0.3 p.p.), the United Kingdom (-0.2 p.p.) and Luxembourg (-0.1

p.p.), increases in low-wage earners aged 50 years and older were found in fifteen

OECD member EU countries. Among the fifteen countries, Germany (+2.9 p.p.), Italy

(+2.6 p.p.) and Poland (+2.5 p.p.) recorded the highest rates of increase in the

proportion of low-wage earners aged 50 years and over between 2006 and 2010.

Figure 3.14 Changes in Proportion of Low-Wage Earners Aged 50+ in EU Countries, 2006-2010 (%)

Source : Eurostat, SES (earn_ses_pub1) Note: OECD member states only are presented, and the data are not available for Greece in 2010 and Switzerland in 2006.

A higher proportion of low-wage older earners is also found in non-European OECD

Countries. For example, 33 per cent of Canadian older workers over 65 earned less than

two-thirds of the national median hourly wage in 2011 (MacEwen, 2012). Also in the

United States, although a different definition of low pay is used, 34 per cent of workers

65 and older and 18 per cent of workers aged 55-64 were in low-wage jobs in 2010,

which were identified as those held by workers with hourly wages below the 20th

percentile of wages (GAO, 2011). According to a 2011 Government Accountability

Office (GAO) analysis of data from the Bureau of Labor Statistics (BLS) and Census

Bureau, there was no significant change between 2007 and 2010 in the percentage of

workers in low-wage jobs in any age group. South Korea has an even higher share of

low-wage older workers; 53.7 per cent of older workers aged 55-79 were the low paid in

2010, who were defined as those earning less than two-thirds of the national median

hourly wage (Jung, S-M, 2011). Although there was a minor decline (-0.3 p.p.) between

EU 27 Austria Belgium Czech Rep

Denma rk Estonia Finland France

Germa ny Greece

Hungar y Iceland Ireland Italy

Luxem bourg

Nether lands

Norwa y Poland

Portug al

Slovaki a

Sloveni a Spain

Swede n

Switzer land Turkey

United Kingdo

m

2006 13.5 10.9 5.3 17.7 3.4 29.6 3.7 5.6 17.9 6.7 17.6 5.1 18.4 6.5 9.3 7.1 1.6 21.3 15.2 18.7 17.5 10.2 0.4 - 0.2 18.5

2010 14.5 12.5 3.0 19.1 3.7 31.2 4.7 4.7 20.8 - 18.7 5.4 17.4 9.1 9.2 8.5 1.3 23.8 12.8 19.8 14.3 11.0 0.7 7.8 0.2 18.3

0 5

10 15 20 25 30 35

2006 2010

51

2006 and 2010, the figure is still high over the half of Korean older workers aged 55-79.

Apart from the main trend in the proportion of low-wage older workers, a large pay gap

between men and women is found among this group. Figure 3.15 shows the gender pay

gap among workers aged 55-64 and 65 years and older in twenty-three OECD member

EU countries. The gender pay gap here refers to the difference between average gross

hourly earnings of male paid employees and of female paid employees as a percentage

of average gross hourly earnings of male paid employees. Although the gender gap in

hourly earnings varied by country and a few exceptions were found, the rate of the

difference between men’s average hourly earnings and women’s relative to men’s

average hourly earnings ranged from 13.5 per cent to 36.3 per cent among workers aged

55-64 in 2010, except for the three extremely low or adverse cases of Italy, Poland and

Slovenia; and the rate for those aged 65 years and older, from 10.4 per cent to 45.1 per

cent in the year, except for the three cases of Poland, Slovenia and Slovakia in which

the rates were lower than ten per cent. This suggest that women are likely to form a

larger share of the low paid than men among older workers.

Figure 3.15 Gender Pay Gap among Older Workers Aged 55-64 and 65+ in EU Countries, 2010 (%)

Source: Eurostat, NACE Rev.2 (earn1_gr_gpgr2ag) Note: OECD member states only are presented, and the data are not available for thos aged 65 years and older in Belgium and Luxembourg.

Belgi um

Czech Repu blic

Denm ark

Germ any

Irelan d

Greec e

Spain Franc e

Italy Luxe

mbou rg

Hung ary

Neth erlan

ds

Austri a

Polan d

Portu gal

Slove nia

Slova kia

Finlan d

Swed en

Unite d

Kingd om

Icelan d

Norw ay

Switz erlan

d

55-64 17.6 15.2 16.5 27.5 20.4 19.1 23.7 22.7 3.9 21.6 19.7 24.4 36.3 3.5 9.7 -6.6 13.5 25.5 18.9 22.8 21.6 21.2 22.9

65+ 30.8 12.8 20.3 10.4 17.8 45.1 36.8 36.9 19.9 11.6 24.1 5.4 36.5 7.3 6.7 29.0 19.8 22.5 18.4 19.6 28.5

-10

0

10

20

30

40

50

52

Conclusion

This chapter reviewed older workers’ labour force participation rate and employment

rate, types of contracts and sectors that they were enagaed in, and their wage levels.

Despite the difference between countries in the level of labour force particiaption and

employment and in the steepness of their rises, the labour force participation rate and

the employment rate of older workers have visibly increased since the mid-1990s, in

genreal. However, older workers are more likely to be involved in part-time jobs and

employed in agriculture, manufacturing, whole sale and retail trade, education, and

health and social work. Except for agriculture in which the employment rate of older

workers have dramatically dropped and for education in which they are traditionally

over-represented, all the sectors in which older workers are mainly employed are those

with high proportions of low paid jobs. This is consistent with the fact that the

proportion of low-wage earners among workers aged 50 years and older has increased

significantly. In the following chapter, the review of the existing research about the

status of older workers in the labour market will be included to look at how the research

has developed and what has been focused on.

53

Chapter Four Older Workers: Labour Market Status and Minimum Wages

Introduction

As shown in Chapter Three, the labour market for older workers has been changing in a

direction that increases their vulnerability at the workplace. This raises the issue of

measures to relieve the expected disadvantages that older workers may face, and

accordingly, related policies, such as the minimum wage, and research on them are of

importance. However, with regard to the exclusive concern of this thesis on the

minimum wage, the combination of older workers and the minimum wage has been

rarely explored in research as mentioned earlier. This comes from the development of

different focuses in two research areas. This chapter reviews literature on the labour

market status of older workers and the effects of the minimum wage relating to them.

Through the review, it will be shown that the two research areas have long developed

apart from each other, and different research focus in each area has created an academic

vacuum between them. That is, research on older workers has focused exclusively on

employment issues, including retirement, and older workers have been almost out of

concern in research on the minimum wage effects which has concentrated on youth and

younger workers.

The Labour Market Status of Older Workers

Research Trend

Older workers as a group have been of research interest since the middle of the

twentieth century with the institutionalisation of retirement, a fairly new phenomenon

which characterised most industrialised nations and was reflected in the widespread exit

from the labour force within a relatively short age span and often prior to the loss of

indivduals’ physical or mental ability to remain gainfully employed (Szinovacz, 2003,

p.7). Although retirement existed in the pre-industrial peasant society, such withdrawal

54

from an active working life was confined, on the one hand, to the wealthy and powerful

who could afford to spend their later years in comfort and leisure (Macnicol, 2002,

p.20) and, on the other hand, to the disabled who were not able to work any more with

their physical incapacity. The age at retirement in the pre-industrial society was across a

wide range, and first and foremost, retirement had no place in traditional European

culture, where ‘the life cycle system consisted only of birth-marriage-death’ (Kohli,

1987, p.130). The notion of retirement emerged as industrial production became

noticeably more technology-intensive at the end of the nineteenth century and the small

family firms were replaced by the large bureaucratic unit of production (Macnicol,

2002). The technology-intensive production made labour market ‘tightened up with a

greater premium on youth, skills and adaptability, and thus older people were seen as

increasingly irrelevant to the labour process’ (Macnicol, 2002, p.5 & p.13). At the same

time, as large bureaucratic organisations ‘used more scientific management techniques,

including the rule of a fixed retirement age’ (Macnicol, 2002, p.19), older workers

began to be involved in involuntary retirement. Industrialism theory aptly addressed the

forces that had led to forced retirement of older workers in this early time period1, and

Pampel and Weiss (1983, pp. 354-355) briefly reviewed the components responsible for

this change: first, as developed countries transformed from a agricultural, self-employed

occupational structure to a wage and salary, bureaucratic one, workers who had been

able to choose how long they would remain at their job in the previous structure faced

formalised retirement rules based on chronological age in larger bureaucratic

organisations (Slavick, 1966; quoted in Pampel and Weiss, 1983); secondly, the

expansion of the education system in developed nations made older workers less

competitive in the labour market than younger workers who were likely to have

completed more years of schooling than earlier cohorts. Employers were likely to

require the retirement of older workers in order to take on younger workers with the

latest skills and knowledge (Clark, Kreps, and Spengler, 1978, quoted in Pampel and

Weiss, 1983); thirdly, given a higher proportion of aged persons in the population and

employers’ preference for younger workers, the increased number of older persons had

to compete for available jobs, and as a result, a lower share of older persons would have

been able to remain in the labour force (Cowgill, 1974; quoted in Pampel and Weiss,

1983); and fourthly, low demand for older workers reduced participation rates.

55

It is, however, only since the Second World War that retirement became the social norm

(Thane, 2006), which means retirement has been institutionalised and become a central

part of the life course in most developed, modern societies. This derived from the

development of state pension policy and social welfare although the normative

underpinning of retirement as an institution was also spurred by social and individual

attitudes on retirement and retirement age (Szinovacz, 2003). The pensions which were

initially introduced in Europe and Australasia in the late nineteenth and early twentieth

centuries provided minimal income to live on for some of those who had already been

forced by decrepitude to retire from gainful work, but they typically did not prohibit

earnings from work (Thane, 2006, pp.43-44). Before the Second World War, retirement

prior to physical incapacity was not a commonly anticipated phase of life-course;

however, as retirement income, mainly from pensions including postwar state and

occupational pensions, normally presumed and sometimes required that the pensioners

must retire from work, retirement at around the state pension age dramatically increased

in developed countries in the second half of the twentieth century (Thane, 2006, p.45).

The social policy explanation for the decline in the labour force participation of older

workers, in particular older males in developed countries, emphasizes four groups of

government policies which may have larger effects than economic growth variables in

the postwar period: it is likely that the higher the government expenditures for pension

and social insurance programmes per retired person, the lower the participation rate of

aged males; expenditures for nonretirement programmes, including medical welfare

payment, public housing, welfare programmes, income supplements for food and

energy, and various in-kind benefits, may also induce retirement; certain provisions of

pension programmes, such as lowering the age of eligibility for pensions or reducing the

benefits for those who continue to work after the age of eligibility, may increase the exit

from the labour maket among older workers, regardless of even controlling for the level

of pension and other government expenditures; and the number of years for which a

programme has been in existence and the extent to which the programme is recognised

as stable and reliable influence the retirement decision and the participation rate of older

workers (Pampel and Weiss, 1983, p.356). For female workers’s retirement behaviour,

there had been a long tendency that it was viewed based on established understanding of

men’s retirement. But, in relatively recent years, a variety of factors other than

56

government policies, including personal health, family and caring responsibilities,

financial issues, firms’ employment policies and discrimination, and institutional factors

that affect work and retirement decision, have been highlighted. This change is based on

consideration of women’s different career paths due to their status as an additional

earner in a household relating to marital status, ‘the different impact of family and

domestic responsibilies, and their relative underrepresentation at higher levels in

organisations’ (Duberley, Carmichael and Szmigin, 2014, p.71). The attention to

women’s more discontinuous and fragmented work histories has now led to stress on a

holistic research apporach to women’s work and retirement activities as Wong and Earl

(2011) argue. Since reseach on female workers’ retirement has been developed much

later than of male’s, it is more connected to the development of research on early

retirement and also on its reversal which are presented below.

The spread of retirement after the Second World War does not mean that it brought

about either social discouragement of working in old age or retirement-focused research

about older workers. The British government, for example, encouraged older workers to

stay on at work past the pensionable age until the early 1960s (Thane, 2006); and from

the late 1940s to the early 1960s, there were studies about older workers built around

theoretical models from social and occupational psychology and located in part in the

area of industrial gerontology (Phillipson, 2004, p.189) that focused on the problems of

older workers and conditions of work under which they can make optimum use of their

skills for higher productivity (Murrell, 1959, p.216). However, discussions in the 1960s

among most OECD countries identified older workers as a problem group at a time of

full employment. By the mid-1970s with demographic changes and the economic

recession in the years following the oil crisis, discrimination against older workers

became more pronounced by employers’ selective recruitment policies and selective

discriminatory dismissal practices (Casey and Bruche, 1983, p.2). Under these

circumstances, early retirement significantly started to be increased in most Western

societies and so did public and scientific attention to this issue (Hofacker, 2010, p.12).

Early retirement was widely supported by governments, trade unions, and employers ‘as

a means of coping with economic restructuring, workforce downsizing, and global

competition’ (Macnicol, 2008, p.581). However, the trend to early retirement resists

57

easy explanations that rely on social policy alone by constrasting ‘theories that focus on

the state as the key actor’ and explanations ‘that focus on actors in the economic sphere’

(Kohli and Rein, 1991, p.1). As implied in Kohli and Rein’s (1991) examination of the

evolution of early exit and the conditions, actors, and institutions which brought it

forth2, various factors, not from a single driving force but from multiple forcese in

difffernt levels, are indeed concerned in the explanation of early retirement. While very

few studies attempted to consider disparate causes at the same time, Walker’s study

(1985) about older workers and early retirement demonstrates the point well. He found

that ill health and redundancy were the major spur to early retirement; the demand for

older workers with specific skills and public attitudes such as official labour market

intermediaries’ in the job centres were crucial contributors to discouraging older

workers from searching for work and encouraging them finally to take early retirement

after prolonged unemployment; and the economic and social policy contexts influenced

taking early retirement. Furthermore, different factors at the individual, organisational,

institutional, and macroeconomic levels which explain the trend to early retirement

interact with each other, and their combined influence operates as ‘a social process of

exclusion’ (Walker, 1985, p.227) or potentially, as a process of integration of older

workers.

The full-fledged exploration of the position of older workers in the labour market, in

this regard, has begun with the attempts to explain early retirement but developed with

the transfer of research focus from retirement to employment. Since the mid-1990s,

early retirement has increasingly been considered as being financially unsustainable in

the face of the prospect of ageing in population and in the labour force, and older

workers’ participation in the labour market has gradually increased. Employment rates

among older workers have significantly risen since the turn of the millennium

(Hofäcker and Unt, 2013, p.164), and even after the global financial crisis in 2008, the

rates increased in some European countries. With reference to these trends, new

directions in research on older workers have emerged with the changing perspective

towards active ageing (Walker, 2006), bridge employment (Shultz, 2003, quoted in

Rocco and Thijssen, 2006), second careers (AARP, quoted in Lewis, 1996), and older

entrepreneurs (Minerd, 1999, quoted in Rocco and Thijssen, 2006). The concept of

58

active ageing which emerged in the early 1960s, stressing ‘the maintenance in old age

of the activity patterns and values typical of middle age’, ‘focused narrowly on the

production of goods and services’ in the 1980s and has developed ‘in the 1990s under

the influence of the World Health Organization (WHO)’ into one that emphasises ‘a

broad range of activities than those normally associated with production and the labour

market and inclusion of older people as full citizens’ (Walker, 2006, p.83-84). ‘Bridge

employment refers to the labour force participation patterns observed in older workers

between their career jobs and compete labour force withdrawal’ which allows them to

exit the labour force gradually (Shultz, 2003, p.215). Second career is in line with the

concept of bridge employment. But, it signifies a radical career change in later life

motivated not so much by money but by a strongly felt need to achieve a lifelong dream

or fulfil some inner goal (Lewis, 1996). The concept of older entrepreneurs stresses the

potential advantages for retirees to create their own business, such as hiring other

retirees and offering care services and organic products to younger generation. Although

all these concepts are not always confined to older people’s engagement in production,

they are chiefly based on ‘choices older workers make to remain within and not outside

of working life’ (Rocco and Thijssen, 2006, p.7). In the new directions of research on

older workers, factors accounting for retirement and early retirement explain older

workers’ employment, like the flipside of the same coin, as the polity encourages older

workers to remain in employment longer on the conceptual basis of those new

perspectives for reversing the trend to early retirement, and employers’ perceptions on

older workers gradually changes. This has broadened the spectrum of labour market

status of older workers from working full-time, working in a flexible form, looking for a

job but unemployed, tentatively retired without pensions to fully retired with pensions

or near pension age. Also, as the research focus has been transferred from retirement to

work, disadvantages and discriminations that older workers face in their working lives

seem to have been of more importance with regard to their status in the labour market.

This will be discussed in more detail later in this chapter.

59

Figure 4.1 Factors Influencing Labour Market Status of Older Workers

* De Vroom (2004, p.8) explains that ‘age-culture is the shorthand description of social norms, values, ideals or

perceptions in society that structure the ideas of the age-work relationships’.

What factors make a difference to the labour market status and experience among older

workers and how the factors are related to each other in influencing them are the main

questions in the research area of older workers in the labour market. Figure 4.1 shows

three types of critical factors which have been dealt with in this field, plus the influence

of changes in the macro-level context. The factors are interrelated within as well as

across supply-side, demand-side, policy-induced influence, and changes at the macro-

level and their combined influence typically regulates the status and experience of older

workers in the labour market, including employment, unemployment, earnings, and

mobility, as well as retirement.3 Note that older workers’ status and experience in the

labour market as a consequence of the practice in the labour market for older workers

potentially also affect the degree of influence of each factor affecting older workers’

status in the labour market and the dynamic between the factors in their combined

influence as indicated in Figure 4.1.

Demand-Side Factors Fixed costs & deferred pay Training Adaptability Productivity Employers’ attitude

Policy-Induced Factors Pensions & other benefits ‘Tax force’ Employment Protection Legislation (EPL)

Supply-Side Factors Health Financial situation Marital and family situation Tastes for leisure Attitudes of older workers to

employment

Work (full-time, flexible employment, unemployment, earnings, mobility)

or Retirement

Macro-Level Context Socio-Demographic Change, the State of the Economy & ‘Age Culture’*

60

Supply-Side Factors

The supply-side factors include individual characteristics, such as health status,

financial situation, marital and family situation, tastes for leisure, and attitudes of older

workers to employment. Numerous studies which have explored the personal

characteristics as major variables to influence older workers’experiences in the labour

market have shown that individuals’ health plays a decisive role in retirement and early

retirement (Heywood and Siebert, 2009, p.5). However, there is no apparent consensus

on the effect of self-reported health conditions on changes in employment status at older

ages. While Alavinia and Burdorf (2008) concluded in their analysis across European

countries4 that perceived poor health among persons aged 50-64 was strongly

associated with non-participating in the labour force due to early retirement, being

unemployed or being a homemaker, Blau and Shvydko (2011) showed in their US study

that among workers aged 51-72, 30 per cent of those whose health turned from good to

bad completely exited from the labour market while 15 per cent of individuals whose

health remained good did, but 69 per cent of exits from employment were individuals

whose health remained good whereas only 13 per cent of exits from employment were

associated with individuals’ decline in health from good to bad. Such inconsistency

suggests that health problems alone cannot explain ‘the abrupt transition to retirement’

(Heywood and Siebert, 2009, p.6), which, again, implies that health could be a push

factor for older workers to move toward another types of status in the labour market.

Zucchelli, Harris, and Zhao (2012) corroborates it in their findings for Australia that

health greatly influences cross-mobility between part-time, self-employment, and

inactivity, and both part-time and self-employment could be used as a bridge towards

permanent retirement by persons of ill-health even though health shock significantly

increases the probability of abrupt economic inactivity.

Financial situation, such as assets, pension wealth, and earnings, consistently predicts

retirement decisions as well (Gruber and Wise, 1999; Quinn, Burkhauser, and Myers,

1990; quoted in Wang and Shultz, 2010, p.185). In agreement with earlier research, De

Wind, Geuskens, Reeuwijk, Westerman, Ybema, Burdorf, Bongers, and Van der Beek

(2013) found in their qualitative study in the Netherlands that the financial situation and

expected income during retirement played a role in early retirement, highlighting the

61

different importance of financial factors between the different pathways to early

retirement in association with health status. Damman, Henkens, and Kalmijn (2011)

also showed in their Dutch study that men without a pension shortfall were more likely

to retire early, compared with those with one. However, the relationship between one’s

financial status and retirement decision is rather complicated (Wang and Shultz, 2010,

p.185). For instance, Wang, Zhan, Liu, and Shultz’s (2008, quoted in Wang and Shultz,

2010, p.185) longitudinal study which used the US data showed that retirees’ total

wealth was not able to predict the odds for retirees to take career bridge employment

against full retirement. The influence of financial situation becomes a little more

complex as well when it is combined with family care. If an individual is required to

take on a caregiving responsibility for a family member, this works as a push facor

toward retirement as explained below; but, the financial costs associated with a family

member’s illness or family financial obligations may preclude retirement. Damman,

Henkens, and Kalmijn (2011) found that the more financially dependent children older

men have, the less likely they retire early.

Marital status and family situation with regard to economic activity among older

workers have received relatively little attention in the literature on the labour market

participation of older workers (Lissenburgh and Smeaton, 2003; Szinovacz, 2003).

However, both are critical stimuli to changes in employment status of older workers,

providing a part of information on gender difference in retirement behaviour. Tanner

(1997) found from further analysis of the UK Retirement Survey that older women were

more likely than older men to say that they left employment before state pension age

because they wanted to retire at the same time as their partner. The result is consistent

with the findings from Szinovacz’s (2013) analysis of the US Health and Retirement

Study. She found that married women aged 50-64 were more likely than any other

marital status or men ‘to plan on stopping work in retirement and particularly less likely

to plan on working with reduced hours’ at old age (Szinovacz, 2013, p.1). But, Wang,

Zhan, Liu, and Shultz’s (2008) findings from the longitudinal US Health and Retirement

data do not support the predictive effect of marital status in differentiating full retirment

and general bridge employment. They explain that other aspects of family-related life,

such as spouse working status (Wang, 2007), may moderate the effect of marital status.

62

An alternative explanation may be that the predictive effect of marital status is not only

overshadowed by other predictors, including financial pressure or health considerations

(Barnes-Farrell, 2003, quoted in Wang, Zhan, Liu, and Shultz, 2008), but also

counterbalanced by no division of gender. On the other hand, family situation as a

variable that influences economic activity typically refers to the needs of family care.

Research on the relationship between work and care said that older women were most

likely to provide care for sick, disabled or elderly relatives, partners, children or

neighbours (Loretto, Vickerstaff, and White, 2005). According to the UK report by

Arrowsmith (2004), caregiving responsibilities are a key reason for the workless not

wanting a job or seeking work. It means that care is a contributor to discourage persons

who have once been out of employment with caring responsibilities to re-enter into the

labour market. However, comparing to marital status, family situation is associated

more with movement into flexible employment among older workers. Howard (2005)

and Mooney, Statham, and Simon (2002) suggested in their UK study that those who

combined employment and caring would welcome a range of flexible working options,

such as better access to flexible working hours, the opportunity to reduce working

hours, the right to time-off for caring responsibilities, and the ability to work from home

where feasible (quoted in Loretto, Vickerstaff, and White, 2005, p.43). Smeaton,

Vegeris, and Sahin-Dikmen (2009) also recognised in their UK analysis the likely

transition of older workers who had caring responsibilities to flexible working rather

than to complete withdrawal from the labour market; however, they pointed out that the

type of flexibility needed by carers for children and adults could differ significantly.

It is argued that the taste for leisure increases with age, and age per se is a strong

determinant of retirement controlling many other factors (Heywood and Siebert, 2009,

p.6). Scales and Scase (2000) and Phillipson (2004) suggested that people were

increasingly expecting to be able to enjoy their retirement, in which they would spend

pleasure time with their families, pursuit leisure interests, and leave behind the stress of

their working life (Loretto, Vickerstaff, and White, 2005, p.37). But, recent studies

conducted in the United States and the United Kingdom (McNair, Flynn, Owen,

Humphreys, and Woodfield, 2004; Smeaton, Vegeris, and Sahin-Dikmen, 2009; AARP,

2004; Lynch, 2006) found that the attitude of individual older workers to working for a

63

longer period was positive, and a considerable number of people, whether they were in

work or retired, wanted to continue to work in their later life, especially under a flexible

form. In the same vein, surveys which covered the attitude of older workers towards

working in later life in Bulgaria (Daskalova, 2007) and in Singapore (Lim, 2003)

observed that the majority of respondents preferred to continue working after

retirement. These incoherent results between the taste for leisure and preference for

working longer among older workers can be accounted for partly by the changes in law

which modify the taste. Gendell (2008) assumed that ‘the abolition of mandatory

retirement in the United States had a symbolic effect and confirmed the propriety of

continuing to work’ (quoted in Heywood and Siebert, 2009, p.6). Another possible

reason may be related to the facts that wages are considerably higher than pensions and

that longer employment enables workers to increase their pension as Daskalova (2007)

pointed out.

Demand-Side Factors

Although numerous studies have concentrated on supply-side factors that influence an

individual’s work/retirement decision, the labour market status of older workers is, in

practice, more restricted by demand-side factors, such as fixed costs, deferred pay,

training, adaptability, productivity, and employers’ attitude. Heywood and Siebert

(2009) give brief accounts and famous research examples of fixed costs and deferred

pay, and the main substance of the discussion here about each of the two factors is

borrowed from them. According to Hurd (1996, quoted in Heywood and Siebert, 2009,

p.6) who provided a comprehensive discussion of the fixed costs problem for older

workers, when firms hire, costs occur both when selecting and hiring workers to fill

jobs and when training the workers; and these costs are per head and make part-time

workers more expensive. Hurd (1996, quoted in Heywood and Siebert, 2009, p.7) also

addressed company health insurance premiums in addition to the fixed costs since the

premiums were charged on a per-worker basis rather than on a per-hour one. If older

workers wish to move into part-time work, they would have to accept a large reduction

in earnings to compensate for the job’s fixed costs; but, the reduction in pay may be

prohibited by anti-discrimination rules, and such wage inflexibility may lead to few new

64

part-time openings available (Heywood and Siebert, 2009, p.7). The problem of fixed

costs could explain a part of the reason why firms are put off older workers and of the

abruptness of the retirement transition (Heywood and Siebert, 2009, p.7). Deferred pay

in which employees are paid less than their marginal product earlier in the contract but

more later, and thus firms need to set a definite retirement date also deters firms from

hiring older workers because their shorter expected tenure makes them less motivated

by delayed compensation (Heywood and Siebert, 2009, pp.7-8). Several studies

‘support the predictions that deferred pay contracts give older workers less chance of

being hired and needs with mandatory retirement’ (Heywood and Siebert, 2009, p.8).

Hutchens’ famous early study (1986, quoted in Heywood and Siebert, 2009, p.8) found

that job opportunities for older workers were higher in occupations and industries in

which pensions, mandatory retirement, and tenure were lower, and delayed payment

contracts were not used. These results are confirmed in other US studies such as Scott,

Berger, and Garen (1995), Hirsch, Macpherson, and Hardy (2000) and Hu (2003)

(quoted in Daniel and Heywood, 2007). Daniel and Heywood (2007, quoted in

Heywood and Siebert, 2009, p.8) conducted a similar study for the United Kingdom and

found that ‘firms with pension provision, more steeply increasing wages, and with

longer tenure as expected with deferred compensation’ were less likely to hire older

workers. Comparable findings have been reported for Australia (Adams and Heywood,

2007), Germany (Heywood, Jirjahn, and Tsertsvardze, 2010), and Hong Kong

(Heywood, Ho, and Wei, 1999) (quoted in Heywood and Siebert, 2009).

Training is another demand-side factor influencing the status of older workers in the

labour market. Tikkanen, Lahn, Withnall, Ward, and Lyng’s (2002) case studies which

were added to WORKTOW, a multidisciplinary research project carried out in 27 small

and medium sized enterprises in the United Kingdom, Finland, and Norway showed that

work-based learning and training for older employees had the potential to improve

learning motivation, strengthen self-confidence and organisational commitment, and

improve the social climate in groups with mixed ages. Picchio and van Ours (2013) also

found, using data from the Netherlands, that on-the-job training significantly increased

future employment prospects even for older workers and argued that it suggested firm-

provided training may be an important instrument to retain older workers at work. The

65

problem is, however, that older employees to a lesser extent take part in training.

Wooden, VandenHeuvel, Cully, and Curtain (2001) identified six barriers to training for

older workers, including absense of paid work, learning capacity, education, uncertain

retirement age, employer discrimination, and self-discrimination. OECD (2011, p.74)

suggests not only that ‘employers and public employment services are less likely offer

training to older workers’ but also that ‘older workers are less willing to take up training

opportunities because the expected pay-back period on their investment in training is

shorter than for younger workers’.

Closely connected to skills and training, adaptability and productivity seem to be factors

that influence employers’ attitudes toward older workers rather than direct contributors

that affect the labour market status of older employees because studies of the two

factors have been based mostly on surveys of employers, managers, and foremen.

Adaptability of older workers is not a facet of employment which is amenable to

measure, but the general impression given by studies is that older workers have less

desire or ability to adapt to new methods of work (Makeham, 1980, p.21). In a survey

cited in the OECD (1967, quoted in Makeham, 1980, p.21), employers in the United

States rated their older employees less than satisfactory only on adaptability to change

among ten characteristics bearing on job performance. Aubert, Caroli, and Roger (2006)

asserted that older employees were far less likely to work in those positions or firms that

required the latest technological innovations. Meanwhile, given that older individuals

were less productive, earlier studies of relative productivity informed that despite a

slight decline in productivity with age, differences were greater within age groups than

between age groups. The US Department of Labor (1956; 1957) which studied eight

manufacturing establishments in footwear and men’s clothing during the pilot work and

twenty six establishments in footwear and household furniture during the extended

survey found that beyond the age of 55, performance declined by less than 10 per cent,

but there were considerable variations within each age group. The Canadian Department

of Labor’s (1959, p.3) study of retail trade also concluded that older employees’

performance appeared to ‘level off or decline slightly’. Findings from more recent

studies show that employers’ ratings on overall productivity of older workers varies by

country and stress differences in productivity at older ages by occupation. A

66

comparative survey which was conducted among employers from Greece, Spain, the

Netherlands, and the United Kingdom by van Dalen, Henkens, and Schippers (2009)

found that only employers from the United Kingdom perceived older employees’

productivity to be higher than that of young workers, and employers in the Netherlands

had the most negative expectations regarding productivity of older workers. Also,

Skirbekk (2004) concludes through his literature survey that although individuals’ job

performance often decreases towards the end of one’s career, in particular when

problem solving, learning, and speed are important, older individuals maintain a

relatively high productivity level in work taskes where experience and verbal abilities

matter more.

There is substantial evidence that employers consider older workers to be more reliable,

more committed to the organisation and have better interpersonal skills (Taylor and

Walker, 1994; McGregor, 2001; IRS, 2003, quoted in Loretto and White, 2006;

McGregor and Gray, 2002; van Dalen, Henkens, and Schippers, 2009) while older

workers are also perceived to be inflexible, resist training, and are resistant to change

(Chiu, Chan, Snape, and Redman, 2001; Redman and Snape, 2002, quoted in Loretto

and White, 2006; McGregor and Gray, 2002; van Dalen, Henkens, and Schippers,

2009). However, there is relatively little knowledge about how employers’ perceptions

of various attributes older workers have influences their recruitment and retention

(Loretto and White, 2006). Taylor and Walker’s (1998) postal survey of large employers

in the United Kingdom, which was a prominent attempt to examine a causal link

between employers’ attitudes and their discriminatory practices towards older workers,

indicated that perceived productivity, reliability, ability to adapt to new technologies,

interest in technological change, and flexibility were not associated with recruitment,

training, and promotion practice while perceived trainability, creativity, cautiousness,

physical capabilities, the likelihood of having an accident, and ability to work with

younger workers were related with employment practices. However, in Loretto and

White’s (2006) qualitative study which was inspired by Taylor and Walker’s research

and conducted in Scotland, reliability, productivity, and flexibility were among the most

frequently mentioned by employers as the attributes associated with their employment

behaviour. One reason for the difference observed by Loretto and White was that

67

different attributes were valued in orgranisations of different size.

The Role of Public Policy

Public policy makers have played important roles in influencing the participation of

older workers in the labour market, in both exit and retention. Public policy induced

early and on-time exit from the labour market during the two decades before the mid-

1990s. However, as the key theme of related policies has switched over in most

developed countries to a new line encouraging mature workers to remain in the labour

force, the impact of public policy on the status of older workers has entered another

phase. Public policies which exist in different countries and are involved in determining

the status of older workers in the labour market can be broadly grouped into two main

areas. One is old-age pensions and other parts of the tax and welfare system, which

mainly include public pension schemes, official early retirement schemes, implicit taxes

on continued work at older ages, de facto early retirememt schemes, and private pension

schemes (OECD, 2006a, p.52). The other is employment protection legislation (EPL),

including anti-age discrimination laws.

Pensions, Tax, and Other Welfare Benefits

Pension-related policies have provided financial incentives for older people to leave

work and disincentives for them to remain in or return to work through an ‘income

effect’ which refers to the tendency that ‘the higher are old-age pensions, the more

attractive retirement is’, and a ‘substitution effect’, which matters with regard to the

effect of ‘an extra year’s work to the flow of income both from earnings and from old-

age pensions and other benefits’ (OECD, 2006a, p.52). Recent pension reforms in many

countries, however, are seeking the dissolution or reversal of those effects and resulting

in lengthening working lives. Increases in the statutory pension age, improved provision

of financial incentives to work beyond pension age, and less or no early retirement

schemes have been largely taken up as measures for the aim (OECD, 2013b, p.23).

Table 4.1 which displays the current public and private, mandatory or quasi-mandatory

pension schemes in the OECD countries informs the specific strategies for the three

measures. In this part, the strategies and their effects will be briefly reviewed within the

68

scope of public pension schemes, private pension schemes, and early retirement

schemes.

Table 4.1 Pension Ages and Financial Work Incentives/Retirement Disincentives by Type of Scheme in OECD Countries

Schem e

Normal age

Incre- ase (%)

Early age

Reduct -ion (%)

Schem e

Normal age

Incre- ase (%)

Early age

Reduct -ion (%)

Austral ia

T DC

67 67

-

.. 60

-

Japan Basic/ DB

65 8.4 60 6.0

Austria DB 65 4.2 62 5.1 Korea DB 65 7.2 60 6.0 Belgiu

m DB 65 0 62 0 Luxem

bourg DB 65 .. 57/60 0

Canada Basic/ T

DB

67 65

7.2 8.4

.. 60

7.2

Mexico Min DC

65 65

0 -

60 any

age/60

0 -

Chile Basic/ T

DC

65 65/60

-

.. any age

-

Netherl ands

Basic 67 .. ..

Czech Republ

ic

DB 69* 6.0 64 3.6-5.6 New Zealan

d

Basic 65 .. ..

Denma rk

Basic/ T

DC

67 67

5.8 -

..

..

Norwa y

Min NDC/

DC

67 67

-

.. 62

-

Estonia Points DC

65 65

10.8 -

62 62

4.8 -

Poland NCD/ DC

67 - ..

Finland Min DB

65 68

7.2 4.8

63 63

4.8

Portug al

DB 65 4.0- 12.0

55 6.0

France DB DB(Oc

c)

67 67

5.0 0

62 60

5.0 4.0-7.0

Slovak Republ

ic

Points DC

67 67

6.0 -

65 65

6.5 -

Germa ny

Points 67 6.0 63 3.6 Sloveni a

DB 65 4.0 60 3.6

Greece DB 67 0 62 0/6.0 Spain DB 67 2.0-4.0 65 6.0-8.0 Hungar

y DB 65 6.0 .. Swede

n Min NDC DC

65 65 65

4.9-61

-

.. 61

55/61

4.1-4.7

- Iceland Basic/

T DB(Oc

c)

67 67

6.0

.. 65

7.0

Switzer land

DB DB(Oc

c)

65M/6 4F

65M/6 4F

5.2-6.3 4.5-5.0

63M/6 2F 58

6.8 6.35- 7.1

Ireland Basic/ T

68 .. .. Turkey DB 65 0 ..

Israel Basic/ T

DC

67M/6 4F 67

5.0 -

..

United Kingdo

m

Basic/ DB

68 10.4 ..

Italy NDC 67 - 62 - United States

DB 67 8.0 62 5.0/6.7

Source: OECD Pension at a Glance 2013, p.127 Note: DB = defined benefit; DC = defined contribution; Min = minimum benefit; NDC = notional defined contribution; Occ = occupational; T = targeted; .. = early retirement or deferral of pension is not available; - = benefits are automatically adjusted fore early or late retirement in DC scheme; Where pension ages for men and women differ they are shown as M/F. * The Czech Republic decided on an open-ended increase of pension age by two months per year (OECD, 2013b, p.9)

69

Public pension schemes

The main ways that public pensions affect individuals’ work/retirement decisions can be

summed up as the pension eligibility age, the generosity of pensions, and the degree of

flexibility in combining income from work and pensions (OECD, 2006a). In the last

decade, most of the 34 OECD countries have passed legislation to raise the pension age

or the contribution requirements that earn entitlement to full pension benefits (OECD,

2013b, p.23). As shown in Table 4.1, an official pension age of 67 is now becoming

more common, and some countries such as Czech Republic, Ireland and the United

Kingdom have gone even further, moving to 68 or 69 years (OECD, 2013b). The same

pension age for men and women has also come to be a clear trend across the OECD

countries, with exception only of Israel and Switzerland (OECD, 2013b).

Although there is little work on how a rise in the pension age affects the labour force

participation (Staubli and Zweimüller, 2013), increases in the pension eligibility age are

known to affect working longer. Gruber and Wise (2002) found from their analysis of

12 countries, Belgium, Canada, Denmark, France, Germany, Itlay, Japan, the

Netherlands, Spain, Sweden, the United Kingdom, and the United States with very

different social security programmes, labour market institutions, cultural histories, and

other social characteristics that raising the age of benefit eligibility by 3 years increased

the proportion of men aged 56-65 who were working by as much as 36 per cent over the

long run. They concluded that the state pension was clearly the key factor in enabling

and prompting people to retire, and a reform delaying benefit eligibility would reduce

the substantial proportion of men aged 56-65 out of the labour force. More recent

analysis that Mastrobuoni (2009) conducted in the US context showed that an increase

in the normal pension age by 2 months delayed effective retirement by around 1 month.

In response to the combination of lower mortality rate, higher life expectancy, and lower

fertility rate, many countries have also been cutting the benefits for the financial

sustainability of pension systems. Pension reforms in the OECD countries since the

early 1990s have reduced future public pension benefits on average by 20 per cent

(OECD, 2011). While some have done it by changing pension contribution rates or

financing mechanisms, such as building public pension reserves, others, by modifying

70

the calculation or indexation rules applicable to pension benefits under the existing

systems or even further transforming the entire systems, such as from a defined-benefit

(DB) scheme to a defined-contribution (DC) scheme or a notional defined-contribution

(NDC) scheme (Martin and Whitehouse, 2008). Overall pension benefit cuts may

prevent workers from retiring early and delay the timing of it to the official pension

eligibility age and over while more generous benefits work as a work disincentive.

Henseke (2011) observed among the OECD countries that, on average, countries with

relatively generous pension systems experienced lower employment rates among

persons aged 55-64, and more generous pension systems tended to lead to a more

noticeable decline in employment after the ages 55-59. Along with this, a large number

of the OECD countries have introduced benefit increments for deferral of pension take-

up and/or benefit penalties for retirement before the statutory or minimum pensionable

age as shown in Table 4.1 (OECD, 2013b). An extensive literature typically found that

changes in retirement benefits may have significant impacts on the timing of retirement

(Staubli and Zweimüller, 2013)5. However, compared to the benefit penalties for early

retirement, the pension increment may not have sizable influence on work incentives

when work and pension receipt after the normal pension age can be combined (OECD,

2013b). The financial penalties for early retirement and their effects will be discussed in

the section on early retirement schemes.

Apart from the pension eligibility age and the generosity of pension benefits, how other

institutional arrangements concerning pension receipts are set also influences working

after the normal pension age (OECD, 2006a, p.59). As shown in Table 4.2, a number of

countries have developed related rules to allow people to continue working and draw

pensions at the same time. However, some countries have limits to combining pension

receipts and working, paritcuarly during a period between a life-time job and full

retirement. In Poland and Slovenia, working while taking pension is simply not allowed

for workers below normal retirement age, and in Spain, pension benefits are reduced

according to the length of the working day if an individual below normal retirement age

combine working and pension receipt. In Austria and Germany, earnings from work

should be below a certain level for working individuals aged below 65 to receive full or

reduced pension benefits. These limits are typically not applied to workers aged 65 and

71

Table 4.2 Limits to Combining Work and Pensions

Country Rule Country Rule Country Rule Austria Below 65: above EUR

349.01/month the pension is fully withdrawn Above 65: no limit

Germany Means-tested benefit for the over 64s: reduced by 30% of income earned if > EUR 180/month and fully withdrawn if income > of the full means-tested (i.e. EUR 180); For those aged below 65 on early statutory retirement: amounts exceeding EUR 400/month are deducted from pensions

Netherlands No limits

Belgium If above EUR 21,436.5 (single) the pension is reduced by the amount beyond the limit; If earnings are 15% above the threshold, the pension is fully withdrawn

Greece Possible after age 55; Limited if monthly pension income is below EUR 733, the pension in this case is reduced by 70% for evey extra euro

Poland Lmits below normal retirement age

Bulgaria No limits Hungary Limit: 18 times the minimum wage (EUR 6,027)

Portugal No limit, but working in the same company as before retirement is not allowed for three years after pension

Cyprus No limits Ireland No limit for the state pension (contributory)*

Romania It is only allowed to combine work and pension if pension is lower than the gross average wage (EUR 463/year)

Czech Republic

No limit; Additional annual 0.4% receiving full pension; There are restriction for early retirement

Italy No limits; however, other social benefits (i.e. disability) or survivors’ pensions are cut progressively if annual income is above EUR 23,826.40 or EUR 18,229.77, respectively

Slovak Republic

No limits

Denmark Partial means test of supplementary pension

Lavita No limits; the pension contribution rate is lower when combining work and pensions rather than for pension deferral

Slovenia Lmits below normal retirement age

Estonia No limits Lithuania No limits Spain Under 65: pension reduced according to the length of the working day

Finland No limits Luxembou rg

No limits;contribution paid when working are refunded at the end of the year

Sweden No limits

France No limits for the over- 65s and for those aged between 60 and 65 who have contributed for at least 40 years

Malta No limits; contribution rate: 10% of wage until 65 (then contribution stop)

UK Pension credit is reduced by full income receipt as long as income is below EUR 168.77/week

Source: OECD (2014b), p.102 * In Ireland, there was a limit to earnings from work as EUR 38/week under the state pension (transition) payable between 65 and 66; however, the state pension (transition) has been abolished in January, 2014.

72

over. This suggests that the limits ostensibly do not act as a work disincentive, at least

for those over the normal pension age. However, the limits to combining work and

pension receipt in early old age can prevent those aged 65 and over from working in as

much as an individual would hardly find a job after the normal pension age if s/he

stopped working for pension benefits before reaching pension age.

Private pension schemes

Private pension schemes play a large role in pension provision in a number of countries,

and these have been getting more important in recent years as pension reforms have

reduced public pension entitlements (OECD, 2006a; OECD, 2013b). According to

OECD (2013b), private pensions are mandatory or quasi-mandatory in 18 OECD

countries and cover almost or more than 70 per cent of the working age population in

Australia, Chile, Denmark, Estonia, Finland, Iceland, Israel, the Netherlands, Norway,

and Switzerland and more than 40 per cent in Mexico and Slovak Republic; in further

eight OECD countries, such as Belgium, Czech Republic, Germany, Iceland, Ireland,

New Zealand, the United Kingdom, and the United States, voluntary private pensions

cover more than 40 per cent of the working age population. Because of this trend,

private pension schemes can exert an significant influence on work/retirement decisions

(OECD, 2006a).

The ways that private pensions can affect individuals’ work/retirement decisions are

related to the earliest age at which benefits can be accessed and the replacement rate

which shows the level of pension entitlement in retirement relative to earnings when

working. The earliest age accessible to pension benefits vary by country and across

schemes even within a country but is typically governed by some minimum age that tax

regulations set up (OECD, 2006a, p.60). Deferring the age can exert pressure not only

upon workers’ work/retirement decisions but also on employers’ utilisation of private

pension for sheding older workers. Australia and the United Kingdom are adopting a

higher minimum age for private pension benefits. Gradual rises of the age are scheduled

in Australia over the period 2015-2025, and the United Kingdom raised the age from 50

to 55 in 2010 and will make a further increase from 55 to 57 in 2028. Besides the direct

way of increasing the minimum accessible age to pension benefits, there are indirect

73

measures which may virtually put off the age of retirement. In Australia, the upper age

limits for private pension compulsory contributions has been removed, and in

Luxembourg, the rate of increase in pension savings has been lowered (OECD, 2013b,

p.24). OECD (2013b, p.24) forecasts that if workers are to gain pensions at pre-reform

levels under such arrangements, they will need to pay contributions for three extra years

or accept an average entitlement in 2050 which will be approximately 12 per cent less

than the present one.

Replacement rates also differ from country to country and from scheme to scheme even

within a country. Among thirteen OECD countries, which have mandatory or quasi-

mandatory schemes, replacement rates from the schemes for average earners range from

22 per cent to 39 per cent in eight; but, the rates are much higher than the range in

Demark, Iceland, Israel, and the Netherlands and considerably lower in Norway

(OECD, 2013b, p.136). Since private pensions have been developed in many countries

as public pensions have been relatively small, replacement rates in private pensions may

significantly influence individuals’ retirement income and thus their retirement

decisions. Studies analysing labour market exit in the United Kingdom agree that a large

portion of workers with private pension schemes, including occupational pensions,

retire at age 55 when relatively generous benefits become available while workers are

not encouraged to retire early by the flat rate state retirement pension scheme (Oswald,

1999). Meanwhile, the type of private pension scheme can affect the timing of relevant

pension wealth accrual. Defined-contribution schemes that have become more prevalent

in recent years arguably tend to require more years on the job for pension entitlement

with an acceptable replacement rate than defined-benefit schemes because the former is

based on an individul’s amount of contribution and its investment performance while

the latter is tied to workers’ earnings, tenure of service, and age. Friedberg and Webb

(2005) found in the United States that workers with defined-contribution plans retired

two years later on average, compared to their counterparts with defined-benefit plans,

and argued that the changes in pension structure from defined-benefit schemes to

defined-contribution plans can help explain the rise in American retirement age. But, it

should be noted that the defined-contribution scheme is promoted for improving

financial sustainability rather than for encouraging working longer.

74

Official early retirement schemes

Early exits from the labour market have occured mainly through provisions in the

pension systems, formal early retirement schemes, or other social transfer programmes,

such as disability or unemployment benefits (OECD, 2006a). Official early exit

provisions in the pension systems and formal early retirement schemes with relatively

low early pension eligibility age and generous entitlements were widespread across

countries in response to growing unemployment among younger people. But, their

effects which were below expectations with regard to umemployment have led to

reducing or abolishing the formal early retirement schemes, interlinking with rising

concerns about population ageing. As shown in Table 4.1, nine OECD countries,

including Denmark, Hungary, Ireland, Israel, the Netherlands, New Zealand, Poland,

Turkey, and the United Kingdom, do not allow early retirement in any mandatory part

of the pension systems while the majority of countries have replaced their arrangement

by increasing the early retirement age and reducing benefits. Meanwhile, the pathway

through disability or unemployment benefit systems, which is not an official form of

early retirement schemes, is still an effective route and will be dealth with in the section

of de facto early retirement schemes.

A series of previous studies on pension provision and retirement have agreed a sharp

increase in labour market exit at the age of first eligibility for retirement benefits

(Gruber and Wise, 1999). Given this empirical regularity, raising the early retirement

age is likely to be more effective in lengthening working life than increasing the normal

pension age. Using two pension reforms in Austria which increased the early retirement

age, Staubli and Zweimüller (2013) studied to what extent the increase in the early

retirement age turned out to be an effective tool to increase employment of older

workers and found that raising the early retirement age increased employment by 9.75

percentage points among affected men and by 11 percentage points among affected

women. Vestad (2013) who estimated labour supply effects of an early retirement

programme in Norway also had similar findings. If there had not been an early

retirement option, 50 per cent of early retired pensioners would be working at the age of

66.5, which is a little below the Norway retirement age, 67; and if the age limit for early

retirement had been 64 rather than 62, 70 per cent would be working at the age of 63.

75

On top of an increase in the early retirement age, the financial penalties for early

retirement are also likely to restrain workers from retiring earlier than, at least, the

normal retirement age. Hanel and Riphahn (2012) investigated how reduced benefits for

early retirement affected female workers’ retirement decisions after the 1991 reform of

the Switzerland mandatory retirement insurance and reported that reduction in benefits

by 3.4 per cent lowered the odds of retirement at age 62 from 46 per cent to 22 per cent.

‘Tax force’

‘Tax force’ is Gruber and Wise’s term referring to the ‘burden’ of implicit taxes on

continued work (Heywood and Siebert, 2009, p.9). If public retirement pension benefits

are raised, or eligibility ages, including early retirement/disability programmes, are

lowered, implicit taxes on working increases (Heywood and Siebert, 2009, p.9). Based

on their findings from the two consecutive analyses of eleven and twelve industrialised

countries, including Belgium, Canada, Denmark (included only in the second analysis),

France, Germany, Italy, Japan, the Netherlands, Spain, Sweden, the United Kingdom,

and the United States, for the period from the 1960s to the mid-1990s, Gruber and Wise

(1999; 2002) argued that there were a clear relationship between the public pension tax

on work and leaving from the labour force, which was largely causal. In a similar vein,

Duval (2003) estimated the implicit tax penalty of working an additional five years

among OECD countries and mainly found that changes in implicit tax rates and

standard retirement ages explained about one-third of the trend decline in older males’

participation in the OECD between 1970 and 2000 (Duval, 2003, p.22, quoted in

Heywood and Siebert, 2009, p.9). More specifically, the estimates suggested that in

several European countries, including Austria, Belgium, France, Finland, Germany,

Italy, and Luxembourg, there were substantial tax penalties to continuing to work after

the age of 60 and also that the tax penalties were closely interwinded with higher

probabilities of retirement before the age of 65 in those countries (OECD, 2006a, p.61).

Van Soest and Vonkova (2014, p.247) also found from their simulations with the Dutch

survey data that changing the rewards for delaying retirement from actuarially fair to 50

per cent of actuarially fair would lower the mean retirement age by 9.7 months.

76

However, Engelhardt’s (2012) analysis using the first two waves (2004/05 and 2006/07)

of the Survey of Health, Ageing and Retirement in Europe (SHARE) for eleven

European countries, including Austria, Belgium, Denmark, France, Germany, Itlay, the

Netherlands, Spain, Sweden, Switzerland, and the United Kingdom, did not find any

significant effect of implicit tax rate on the exit from the labour market. Moreover, it is

questionable whether no large implicit tax on working beyond the pensionable age and

delaying the receipt of benefits substantially encourages older adults to continue to

work. Although Duval’(2003) estimates indicated that Demark, Iceland, New Zealand,

the Netherlands, and Turkey had around or below five per cent implicit tax on continued

work for five more years in early retirement route as well as in regular old-age pension

system in both 2007 and 2009 (OECD, 2014a, p.96), OECD’s (2012) Pensions and

Labour Market Statistics showed that the average effective age of retirement in

Denmark and the Netherlands fell short of the official retirement age whereas the age

was reasonably higher than the official one in Iceland, New Zealand, and Turkey, during

the period of 2006-2011(OECD, 2013c, p.22). Even where there are ‘actuarially

neutral’6 increases in pension entitlements for working extra years, a marked spike in

retirement may occur around the earliest age at which workers can receive a pension

(OECD, 2006a, p.57). According to OECD (2006a), the United States is a representative

case, and Hanel and Riphahn (2012) explains it as the possibility of unobserved

heterogeneity in preference for retirement which might influence both (dis)incentives

and responses to them.

De facto early retirement schemes

The empirical regularity with regard to labour market exit at the age of first eligibility

for retirement benefits sheds light on the odds that the abolition of official early

retirement schemes can lead people to seek another possible way to leave work earlier

than their normal retirement age with affordable financial benefits. This implies larger

spillover effects of reforming early retirement schemes on other non-pension benefits.

OECD (2006a) pointed it out that any efforts to reform formal incentive systems for

early retirement could not make its purpose if they were weakened by other non-pension

benefit programmes which allowed people to exit from work prematurely. In pracitce,

disability or unemployment benefits are often taken by workers as alternative routes

77

into early retirement, especially where the benefits are relatively easy to access (OECD,

2006a; 2011). Duval (2003) found from his research across the OECD countries that

social transfer programmes, including disability and unemployment benefits, had

sizeable effects on the departure of older male workers aged 55-59 from the labour

force. Staubli and Zweimüller (2013) found from their study on the effects of increased

early retirement age on employment of older workers in Austria that raising the early

retirement age had large spillover effects on the unemployment insurance programme.

Registered unemployment among men increased by 12.5 percentage points and among

women by 11.8 percentage points. Spillovers to the disability programme were

relatively small; however, low-wage and less healthy workers, in particular, retired early

through the disability programme or waited for being the new early retirement age with

receiving unemployment benefits.

Although there are some examples of efforts to tackle those de facto early retirement

schemes by tightening eligibility or restricting replacement rates, such as in Canada,

Denmark, Finland, Italy, the Netherlands, Norway, Sweden, and the United Kingdom,

the majority of countries have not made much progress in tackling the utilisation of the

non-pension benefits for early retirement, mainly due to political resistance to change in

those provisions for the vulnerable (OECD, 2006a).

Employment Protection Legislation (EPL)

Employment protection legislation in general

While pension and tax-related policies are more related to the supply-side of the labour

market in as much as they affect individuals’ work/retirement decision, employment

protection legislation, including anti-discrimination law, concerns the demand-side of

the labour market because it has influence on employers’ employment/dismissal

decision. Strict employment protection legislation can operate as a double-edged sword

in labour market outcomes for workers. On the one hand, the legislation may lead to

greater retention of workers because it raises firing costs; at the same time, it may

reduce the number of hires since it increases hiring costs as well; and as a result, the net

impact on employment rates is uncertain (OECD, 2006a, p.70). Daniel and Siebert’s

(2005) analysis conducted mostly in the UK and the US contexts found that stricter

78

protection provisions tended to lower hiring rates at both ends of the age spectrum while

OECD (2004) suggested that the relationship between employment protection

legislation and labour market outcomes was less clear for older workers than for other

age and gender groups. Meanwhile, OECD (2011a) which used the index of the

strictness of employment protection legislation as set out in OECD (2004) and

compared the index with labour market outcomes for men aged 50-64, showed that

there was a strong negative relationship between employment protection and both

employment rates and hiring rates of older workers. However, OECD (2011a, p.73)

added that more rigorous studies which controlled for other factors affecting

employment rates of older workers, including OECD (2006b) and Dorn and Sousa-Poza

(2010), showed a much weaker relationship between employment protection and labour

market outcomes of older workers. Henseke (2011) also found that countries with

relatively rigid employment protection laws, along with relatively generous pension

systems, had lower employment rates among individuls aged 55-64, but a change just in

employment protection did not improve the employment of older workers as such.

Although stringent employment protection legislation reduces the odds of dismissal

among older workers, it, on the other hand, may raise unemployment duration of older

jobseekers because employers tend not to newly hire (Siebert, 2005; quoated in

Heywood and Siebert, 2009). While there are few studies on the relationship between

employment protection legislation and labour market outcomes of older workers,

research on the relationship between the legislation and unemployment has been

recently active. Daniel and Siebert’s (2005) work of the OECD countries showed the

prediction; and Bertola, Blau, and Kahn (2007) also found the same pattern (Heywood

and Siebert, 2009, p.11).

Anti-age discrimination legislation

Although legal provisons banning age discrimination have been introduced in a

considerable number of countries, such as Australia, Canada, Czech Republic,

Switzerland, the United Kingdom, and the United States, they are evaluated either not

very effective or insufficient (OECD, 2006a). Anti-age discrimination law is a form of

employment protection legislation, which also have two opposing effects on the status

79

of workers in the labour market (Heywood and Siebert, 2009). Lahey’s (2008) results

from the United States showed that older white male workers in states with anti-age

discrimination laws were 0.2 percentage points less likely to be hired, worked 0.8-1.3

weeks less a year, were 0.5-0.7 percentage points more likely to report being retired,

and were 1.6-3.0 percentage points more likely to report that they were not employed,

compared to counterparts in states without the laws. Neumark and Button (2014) also

examined in the US context whether stronger age discrimination protections helped

older workers weather the Great Recession and found very little evidence for their

positive effects. Rather, they concluded that stronger age discrimination protections

were combined with more negative effects of the Great Recession. Lahey (2010) added

in her further research the point that the US anti-age discrimination legislation have

more recently acted as a means of keeping down the costs of entitlement programmes.

According to her, European anti-age discrimination legislation, with less enforcement

and more exemptions than the US one, could have smaller effects on employment

among older workers.

Some countries have directly addressed the labour market to increase the participation

rates of older workers by abolishing or increasing mandatory retirement ages (OECD,

2013b). In Australia, Canada, New Zealand, the United Kingdom, and the United States,

for example, mandatory retirement is prohibited at any age as a way of combating age

discrimination against older workers on the job market; and in France, Japan, and

Sweden, minimum permitted mandatory retirement ages have been raised (Wood,

Robertson, and Wintersgill, 2010, quoted in OECD, 2013a, p.85). According to a

comparative review of international approaches to mandatory retirement conducted by

Wood, Robertson, and Wintersgill (2010), anti-age discrimination legislation, with no

other supporting policies to pull in older workers, did not have a significant impact on

the participation rates of older people in the labour market. They found that employers

often thought prohibiting mandatory retirement would increase their costs while it was

not clear whether the prohibition would bring about actually the case, with the

exception of countries where seniority wages were common. However, their findings

suggested that age legislation could make employers to provide more opportunities and

flexible conditions for older workers to work longer through phased transition to

80

retirement.

Age Discrimination in Working Life

The prevalence of early retirement brought about ‘an ‘early retirement culture’, which

involved a perception of early retirement as ‘normal’ practice incorporated into

individual retirement plans as well as companies’ policy, but also an increasing

emergence of negative age stereotypes’ (Hofäcker and Unt, 2013, pp.163-164).

Although early retirement has already been an outdated practice and working longer has

been encouraged by governments and become more acceptable in societies, employers’

perception and treatment of older workers has not changed very much. In this context,

disadvantages and discriminations that older workers face at the workplace can be an

index to inform of their status in the labour market.

There is a large volume of research in the field of age discrimination against older

workers at the workplace. Wood, Wilkinson, and Harcourt (2008) fully reviewed the

literature on this research area, and it would be useful to summarise their discussion for

an overview of it. They divided the existing research into three broad categories, ‘the

underlying causes and consequences of age discrimination at the workplace’, its ‘nature

and extent’, and ‘the effects of various governmental initiatives and policy options’

(Wood, Wilkinson, and Harcourt, 2008, p.426). Among the three, the review of research

on the effects of government initiatives and policy options is omitted in this section

because it was already discussed above. According to Wood, Wilkinson, and Harcourt

(2008), research which explores the causes and consequences of age discrimination at

the workplace is firmly based on theoretical perspectives, mainly neoliberal accounts

and political economy. There are three strands of neoliberal accounts: the first argues

that ‘older workers’ higher pay simply makes them less attractive to employers’; the

second suggests that older workers often return to the labour market to get insecure and

low paid jobs ‘as a lifestyle choice’; and the third argues that employers discriminate

against older workers due to their lack of information on older workers’ productivity

and reliance on ‘erroneous stereotypes’ (Wood, Wilkinson, and Harcourt, 2008, p.426).

Political economy accounts which focus on ‘the relationship between culture and,

81

particularly institutions, in reinforcing age-based inequality at the workplace’ (Wood,

Wilkinson, and Harcourt, 2008, p.428) contend that older workers tend to bear in the

context of industrial change a proportion of ‘costs of periodic structural changes under

capitalism’, such as flexibility, leaner organisation, the fragmentation of the

employment relationship, and the changing sectoral composition, particularly because

‘it is more socially acceptable, even to the older workers themselves’ and also because

‘they are more easily persuaded and pressured’ to take the costs as Taylor and Walker

(1997, pp.307-308) note (Wood, Wilkinsion, and Harcourt, 2008, p.429).

The second category of research which informs of the nature and extent of age

discrimination at the workplace shows that older workers are more likely to be

discriminated against in recruiting and hiring (Urwin, 2004; Walker, 1993; Sargeant,

2001, quoted in Wood, Wilkinson, and Harcourt, 2008), redundancy situation

(Arrowsmith and McGoldrick, 1997; Sargeant, 2001; Walker, 2005, quoted in Wood,

Wilkinson, and Harcourt, 2008), and promotion and training (Walker, 1993, quoted in

Wood, Wilkinson, and Harcourt, 2008), also more likely to be dismissed as a result of

the negative perception of older employees held by managers (Johnson and Neumark,

1997, quoted in Wood, Wilkinson, and Harcourt, 2008), but less likely to find jobs while

leaving the labour force (Rix, 2005; McGregor and Gray, 2001, quoted in Wood,

Wilkinson, and Harcourt, 2008) and hence, more likely to exit the labour force

permanently as Neumark (2003, quoted in Wood, Wilkinson, and Harcourt, 2008) points

out. More critical literature argues that older workers often face an unfavourable ‘choice

between poverty and workplace discrimination’ (Wood, Wilkinson, and Harcourt, 2008,

p.432) as the costs of social welfare have increased with the population ageing. This

wide range of discrimination against older workers is characterised as being more

related to stereotypes, prejudice, and age culture combined with changes in

demographics and economic structure in as much as findings from a rich body of

research on the relationship between age and productivity, which has been regarded as a

prominent factor to explain older workers’ disadvantages at the workplace, are

inconsistent. Another nature of discrimination against older workers is that the

discrimination ‘can be concealed in a range of ways’ (Wood, Wilkinson, and Harcourt,

2008, p.435), for example, relating to equal opportunity in hiring, pay discrimination,

82

and sectoral attributes, which force older employees out into retirement.

So far, research on older workers which illustrates their status in the labour market has

been reviewed. It suggests the fact that the research on older workers has focused

exclusively on employment issues, including hiring, dismissal, redundancy, retirement,

and discrimination, and also the one that older workers are likely to be vulnerable in the

labour market. This implies that there is a critical research vacuum in as much as older

workers’ vulnerability found in one dimension of their working life, such as

employment issues, is bound to lead to various dimensions, including wages and

earnings. That is, the need for research about the issues of wages and earnings and

policies relating to income from work among older workers, with a focus on their

vulnerability in the labour market is addressed. Studying the effect of the minimum

wage focusing on older workers may answer the need, in as much as the minimum wage

is a public policy which intervenes in wages and earnings for vulnerable workers in the

labour market. However, compared to an enormous amount of research about the effect

of the minimum wage, older workers tend to have long been ruled out in this research

field. This point will be shown in the following part of this chapter through a review of

studies which explored the effects of minimum wages focusing on older workers or

including them in their analysis. Note that the review in the following part of this

chapter is to provide an overview of the empirical studies and does not cover theoretical

literature which informs of major issues, assumptions, and hypotheses in the research

field of the effect of the minimum wage. The theoretical literature will be discussed in

Chapter Five where an alternative framework for understanding the effect of the

minimum wage is proposed alongside the research methods for this thesis. This

discussion will examine particularly the extent to which dominant theoretical accounts

deal with the issue of the minimum wage in terms of the ‘disruption’ made by its

introduction or uprating and presume actors’ rationality in the labour market.

83

Minimum Wages and Older Workers

Overview

The effect of the minimum wage has been intensively studied, particularly since the

renaissance of the minimum wage in the 1990s. The economy that the research covered

has also expanded from developed countries, including the United States and the United

Kingdom, to developing countries, especially Latin America, from that period.

However, the research about the effect of the minimum wage, which has been carried

out dominantly by economists, was mainly concerned with the aggregate effects for the

whole working-age population (aged 15-64) in a state, industry, or firm level, and a

significant number of them concentrated on youths and young workers under the age of

25. Fang and Gunderson’s (2009) work about the effects of minimum wages on

employment of older workers was the only study that focused on the impact on older

workers as they stated. It was recently followed by Lanot and Sousounis (2017) which

explored the degree of substitutability between labour inputs caused by the minimum

wage, paying special attention to older workers. Apart from these two that have

exclusive attention to older workers, a few studies linked the effects of the minimum

wage with older workers by including results for older workers as a part of their

analyses or through using advanced ages as a factor which may affect employment,

wage, income, and poverty. In this part, Fang and Gunderson’s (2009) and Lanot and

Sousounis’ (2017) works are first reviewed, and then the studies which linked the

effects with older workers will be surveyed (Table 4.3). This review aims to get an

overview of the scope and main findings of studies which have been done about the

effect of the minimum wage relating to older workers.

84

Table 4.3 Studies of Minimum Wage Effects and Older Workers

Study Minimum wage variation

Group Data Results in relation to older workers

Studies Focusing on the Effects on Older Workers Fang and Gunderson 2009

Provincial variation during 1993-1999

Older workers 50 years or older

SLID, 1993-1999 (Canada)

Statistically significant and positive effect on employment of older workers

Lanot and Sousounis 2017

National minimum wage rates during 1999- 2010

Various ages in low paying sectors

Quarterly LFS and ASHE, 1997-2010 (U.K.)

Statistically significant effect on changes to relative wages and relative wage bills but not to relative employment in the workforce age composition; No substitution between young workers aged 16-21 and older workers aged 55 and over

Studies Linking the Effects with Older Workers Employment Gramlich 1976

Federal minimum wage increases in 1961, 1967, and 1974

Teenagers and adult workers

U.S. Manpower Administration’s National Longitudinal Survey (quarterly), 1948-1975 (U.S.)

Rise in part-time employment for male workers aged 65 and over

Alpert and Guerard 1988

Federal minimum wage rates during 1965-1979

Nonwhites, teenageers, nonwhite teenagers and older workers aged 55 plus

Current Population Survey (monthly) 1965-1979 & 1980-1981 (U.S.)

Neither increase in unemployment nor decrease in employment for workers over age 55

Van Soest 1989

Simulated decline in statutory minimum wage

Various ages Socio Economic Panel (1st wave) 1984 (Netherlands)

Decrease in unemployment for both males and females aged 55-64 by 37% and 21%, respectively if minimum wage rate reduced by 10%

Dodson 2002

Kaitz minimum wage index (ratio of the federal minimum wages to the average hourly earnings for each industry at the county level)

Prime working age population (15-64)

Regional Economic Information System; West Virginia Bureau of Employment Programs, 1988-1995 (U.S.)

No significant correlation between the number of persons aged 65 plus and county-level employment

Phimister and Theodossion 2009

Introduction of national minimum wage in 1999

Men and women in various ages

BHPS 1992-1998, 1999-2005 (U.K.)

Significant increases in the high pay exit probability for men aged 45-56 while significant increase in unemployment probability for women aged 45- 56; reduction in expected duration of low pay spell for both

85

Study Minimum wage variation

Group Data Results in relation to older workers

Employment (Cont’) Comola and Mello 2011

Kaitz index (ratio of minimum-to- mean wage for formal-sector workers)

Indivuduals aged 15-65 in formal and informal sectors

National Labour Force Survey (Sakernas), 1996-2004 (Indonesia)

Significant and positive correlation between the population aged 56-65 and unemployment

Wage Mobility Smith and Vavrichek 1992

Federal minimum wage in 1981

Minimum wage workers 16 years or older

1984 & 1985 panels of Survey of Income and Program Participation, 1983-1987 (U.S.)

Older minimum wage workers aged 55 or older, less likely to experience a wage gain and more likely to experience smaller pay increase than young and middle-aged workers

Shannon 1996

Provincial minimum wages

Individuals aged 16-64 who held at least one paid job in 1986

Canadian Labour Market Activity Survey, 1986 (Canada)

Rise in wages with age up to age 55-64

Long 1999

Federal minimum wage in 1991 (or state minimum wages during 1991-1995, if higher)

Minimum wage workers 15 years or older

1992 & 1993 panels of Survey of Income and Program Participation, 1991-1995 (U.S.)

Older workers, least likely to experience real wage gains and move up the real earnings ladder

Income Inequality Volscho 2005 State minimum

wages in 1959, 1969, 1979, 1989, and 1999

Family income State and Metropolitan Area Data Books, 1979 1986, 1991, 1998

No significant effects of the percentage of the population aged 65 plus on family income inequality

Wu, Perloff and Golan 2006

Federal and state minimum wages during 1981-1997

Entire income distribution

March Current Population Survey, 1981-1997 (U.S.)

No significant effects of the percentage of the population aged 59 plus on income inequality in both urban and rural area

Poverty Sabia and Burkhauser 2010

State and federal minimum wage increases during 2003-2007; newly proposed federal minimum wage rate for next increase

Individuals aged 16-64 in household with income-to-needs ratio below 1.5

March Current Population Survey, 2004-2008 (U.S.)

No significant effect of the share of older individuals aged 55-64 on poverty

Campolieti, Gunderson and Lee 2012

Provincial minimum wage increases during 1997-2007

Individuals aged 16-64 in household with income-to-needs ratio below 1.5

Survey of Labour and Income Dynamics, 1997-2007 March LFS 2008 (Canada)

No significant effect of the percentage aged 54-64 on poverty

86

Studies Focusing on the Effects on Older Workers

The two studies which focused on the impacts on older workers of the minimum wage,

Fang and Gunderson’s (2009) and Lanot and Sousounis’ (2017) were mainly concerned

with their employment, even though the latter also includes the results of wages. Fang

and Gunderson (2009) carried out a study of the employment effects for workers aged

50 and over of minimum wage increases in different provincial jurisdictions. Using the

longitudinal data of the Survey of Labour and Income Dynamics in Canada for the

period of 1993-1999, they compared employment transition probabilities of workers

afftected by minimum wage increases with a range of comparison groups in which

individuals were in a juridiction that did not have a minimum wage increase but their

wage had one of a serious of increments above or shortfalls below the minimum wage

in their juridiction. The results showed that the minimum wage increases had

statistically significant and positive effects on the employment of older workers and the

effects were robust across various comparison groups and meaures of the minimum

wage increases. This finding was theoretically unexpected and against the adverse

effects which were widely found in the early research performed before the 1990s7.

Most of all, it was opposed to the negative impacts which were estimated for youths in

Canada based on the same period, specifications, and dataset in Campolieti, Fang, and

Gunderson’s (2005). Fang and Gunderson (2009) considered as likely reasons for the

unusual finding employers’efficient practices in other cost factors, employees’efficiency

wage response where the higher wages induced higher productivity, employers’

monopsonistic behaviour, publication bias, and the substitution effect, which, except for

publication bias, will be discussed in the theoretical literature review part of Chapter

Five. They argued that the substitution effect was the most credible because increases in

the employment of older workers could occur if employers substituted away from the

least productive workers towards the more productive when both were subject to a

minimum wage increase, in which case, the least productive were arguably teenagers

with few skills and little experience, and those more productive were arguably older

workers with considerable experience (Fang and Gunderson, 2009, p.385).

The substitutability between younger and older workers in low paying jobs, however,

was not supported by succeeding work which paid special attention to the effects on

87

older workers of the minimum wage. Lanot and Sousounis (2017) studied the effect of

the National Minimum Wage on the workforce age composition within the low paying

sectors of the British economy, using two pseudo panels which were constructed from

each of the UK Quarterly Labour Force Survey (LFS) and the Annual Survey of Hours

and Earnings (ASHE) over the period of 1997-2010. They estimated the effects of the

fraction of those affected by the minimum wage in each age band (16-17 year olds, 18-

20 year olds, 21 year olds, and 22-54 year olds) on the changes to wages, wage bills,

and employment of each age group relative to older workers aged 55 and over and then

measured the elasticity of substitution between the four younger age groups and older

workers. The results showed that the introduction and the regular upratings of the UK

National Minimum Wage had a significant effect on the changes to the relative wages

and to the relative wage bills but not to relative employment and also that the elasticity

of substitution was close to zero between young workers aged 18-20 and older workers

aged 55 and over and around 0.79 between those aged 22-54 and older workers. Lanot

and Sousounis (2017) concluded that the results implied significant complementarity

between younger and old employees. Unlike Fang and Gunderson’s (2009), Lanot and

Sousounis’ (2017) research does not provide a direct evidence of the employment effect

for older workers. However, their findings inform that the employment of older workers

is likely to be affected by the minimum wage independent of the employment effects for

other age groups.

In sum, the two empirical studies which focused on or paid special attention to the

impact of the minimum wage on older workers were mainly interested in the issue of

employment, and their findings were theoretically unexpected in as much as the positive

employment effect was not predicted by the dominant economic theory of the minimum

wage, on the one hand, and as the substitution effect which was a possible account

supported by the mainstream theory for the positive employment effect for older

workers was betrayed, on the other hand. This existing evidence may address the needs

of accumulating more empirical evidence, enlarging the subject matter other than

employment, and seeking an alternative explanation based on the review of establisehd

theoretical accounts.

88

Studies Linking the Effects with Older Workers

As mentioned earlier, there are a few studies which linked the effects of minimum

wages with older workers. Some of them incorporated the results for older workers as a

part of their analysis, and others included older people in their analysis for controlling

the demographic effect or testing age effects. Compared to the studies which focused on

the impact on older workers, the findings from the ones which linked the effects with

older workers cover relatively a variety of subject matter, including employment, wage,

income, and poverty.

Employment

Gramlich (1976) addressed the issue of whether prevailing estimates of disemployment

were high enough to make low-wage workers worse off from increases in the minimum

wage and examined the impact of minimum wages on the wage structure, employment

demands, and family income distribution, using the US Manpower Administration’s

National Longitudinal Survey for the period of 1948-1975. He mainly found that as a

minimum wage increased beyond the range of 40 to 50 per cent of the median wage,

more workers were likely to lose their jobs or have to take part-time jobs, and higher

income families were more likely to benefit from the increase. Particularly for older

workers, he indicated that increases in the minimum wage resulted in a rise in part-time

employment for male workers aged 65 and over. Meanwhile, Alpert and Guerard (1988)

investigated the relationship between employment, unemployment, and the minimum

wage, using the US Current Population Survey for the period of 1965-1981, and found

that an increase in the minimum wage neither increased unemployment nor reduced

employment for workers aged 55 and over in the same way as for other workers, except

male teenagers aged 16-19 years and non-white males over the age of 20. Phimister and

Theodossiou (2009) included the results for workers aged 45-56. They examined gender

differences in the duration of low pay employment spells prior to and after the

introduction of the UK National Minimum Wage in 1999. Their findings showed that

the effects of many covariates on the expected duration, including age and education,

were often less for women than men, and the differences generally declined after the

introduction of the minimum wage. Especially for those aged 45-56, it was reported that

89

among male workers, the high pay exit probability significantly increased,

unemployment probability was flat, and the expected duration of low pay employment

spell was much reduced after the introduction of minimum wage, whereas among

female counterparts, the high pay exit probability decreased, unemployment probability

increased, and the expected duration was reduced.

Other studies on the employment effects of minimum wages used the older population

aged 50 or older as a control indicator of demographic effects or an explanatory variable

for age effects. That is, advanced age was assumed to be a socioeconomic factor that

affected labour market outcomes rather than a target age group to be explored. Van

Soest (1989) analysed the impact of minimum wage regulations in the Dutch labour

market, following the two-equations model introduced by Meyer and Wise8 which was

composed of a wage equation and an equation to explain labour market participation. In

the model, the wage equation assumed that, in absence of minimum wage regulations,

the hourly before-tax rate equaled the worker’s marginal net productivity, which

depended on personal characteristics (p.280), and individuals’ age and elderly status

(age of 65 and older) were used as explanatory variables for the wage and employment

equations. The model was estimated with data from the first wave of the Socio

Economic Panel collected by the Dutch Central Bureau of Statistics (CBS) in 1984. The

estimation results from the two-equations model showed that, for older people,

productivity decreased with age, which thus had negative impact on their wage rates,

and also participation probability fell with age although there was a big gap in absolute

values between men and women. Then, the estimation results were used for simulating

the case that all minimum wage rates were reduced by 10 per cent. Van Soest (1989)

found that a 10 per cent reduction of before-tax minimum wage rates would lead to a 28

per cent decrease of involuntary unemployment which was explained by the gap

between productivity and the minimum wage rate. Particularly for older workers, the

reduction of minimum wage rates by 10 per cent would decrease by 37 per cent and

20.8 per cent of involuntary unemployment for males and females aged 55-64,

respectively and 5.5 per cent and 13.6 per cent for each counterpart aged 65 and older.

90

Dodson (2002) also used the older population aged 65 and older as an independent

variable in his study on the employment effect of minimum wages in West Virginia. He

reported that the number of persons aged 65 and older was not significantly correlated

with county-level employment. However, the estimation results for older workers were

not specified because the main concern of the study was the effects on the ratio of total

county employment to the county population aged 15-64. In Comola and Mello’s (2011)

research, the share of the population aged 56-65 was used as one of the controls. They

studied how decentralized minimum wage setting would affect unemployment, formal-

sector employment, and the incidence of informality in Indonesian urban areas, and

mainly found that an increase in the Kaitz index, the ratio of the minimum wage to the

average wage of the working population, destroyed jobs in the formal sector, but those

job losses were more than compensated for by the expansion of the informal sector.

Their findings included the results that, for the entire working-age (15-65 years)

population, the Kaitz index was negatively signed and statistically significant in the

unemployment equation while the share of the population aged 56-65 years is positively

significant in the equation, and further that the positive demographic effect on

unemployment was much greater in magnitude per unit change than the negative effect

of changes in the Kaitz index.

Wage Mobility

Smith and Vavrichek (1992) examined the earnings mobility of workers employed at the

minimum wage in the United States, using longitudinal data from the 1984 and 1985

panels of the Survey of Income and Program Participation for the period of 1983-1987.

They found that 63 per cent of workers who were earning the minimum wage in the

mid-1980s were earning higher wages one year later, with a typical increase amounting

to almost 20 per cent, but a significant number of workers who mostly did not have a

high school diploma or worked on a part-time basis did not advance beyond the

minimum wage during the observed period of time. More specifically relating to older

workers, results indicated that minimum wage workers aged 55 or older were less likely

to experince a wage gain during the mid-1980s and their pay increases were smaller

than young and middle-aged workers. Their findings are consistent with Long (1999).

Using data from the 1992 and 1993 panels of the Survey of Income and Program

91

Participation for the period of 1991-1995, Long (1999) also analysed earnings mobility

among persons employed in minimum wage jobs in the United States and confirmed

many of the findings reported by Smith and Vavrichek (1992). Results showed that

older workers aged 55 and over were the least likely to experience a real wage increase

whereas about 64 per cent of minimum wage workers were paid a higher wage in real

terms one year later, with the 30 per cent average increase in pay. In neither of the two

studies was there further discussion on why the wage growth among older minimum

wage workers was exclusively lower than the one of young and middle-aged workers

employed in minimum wage jobs. Instead, they commonly regarded elderly status as

one of the traits that affected the size of real wage gains within one or two year after a

minimum wage increase. Shannon (1996) also used age dummy, including age group

55-64, as a regressor in his study on the effects of provincial minimum wages on the

size and composition of the gender wage gap in Canada. He found from the analysis of

the Canadian Labour Market Activity Survey for 1986 that minimum wages reduced the

wage gap substantially for young workers aged 16-24 and less so for adults aged 25-64,

which was most attributable to the adverse employment effects of minimum wages, and

wages rose with age up to the age band 55-64.

Income Inequality

A minority of studies about the effects of minimum wages on income inequality

included the proportion of older population as an explanatory variable or a control of

demographic characteristics, but none specified the effects of minimum wages for older

persons. Volscho (2005) studied the effects of state minimum wages on family income

inequality in the United States and used the percentage of older persons aged 65 and

over as a control variable. Using decennial state data covering 1960-2000, he tested the

hypothesis that states with higher minimum wages had lower levels of family income

inequality. Results showed that, although the relation between the minimum wage and

income inequality was non-linear, the state minimum wage over a certain level reduced

family income inequality, and there was no significant effect of older population. With

regard to the effect of older population, Wu, Perloff, and Golan (2006) found the same

result. Using Current Population Survey data for the period of 1981-1997, they

examined the effect of income tax rates, the minimum wage, and all the major

92

government welfare and transfer programmes on the evolution of income inequality for

urban and rural areas by state in the United States. In the study, they used the percentage

of population aged 59 or older as an independent variable and found that the age

variable did not have statistically significant effects on income inequality in both urban

and rural areas while increases in minimum wages raised both pre-tax and post-tax

income inequality only in urban area.

Poverty

As with the studies about the effects on income inequality, a small number of studies

which explored the effects of minimum wages on poverty reduction used the share of

older individuals as one of time-varying socioleconomic controls, and none specified

the effects of minimum wages for older people. Using data from the March Current

Population Survey, Sabia and Burkhauser (2010) mainly found that state and federal

minimum wage increases between 2003 and 2007 had no effect on state poverty rates in

the United States, and a proposed increase in federal minimum wage from $7.25 to

$9.50 would be more poorly targeted to the working poor than was the last federal

increase from $5.15 to $7.25 when simulating the case. In this study, it was indicated

that the percentage of older individuals aged 54-64 had statistically insignificant effects

on poverty rates both of all individuals and of working individuals. Meanwhile,

Campolieti, Gunderson, and Lee (2012) estimated the effect on poverty for Canada

using data from the Survey of Labour and Income Dynamics for the period of 1997-

2007. Their main findings showed that minimum wages did not have a statistically

significant effect on poverty, and only 30 per cent of the net earnings gain from

minimum wage increases would go to the poor while about 70 per cent spill over into

the non-poor. However, this study was consistent with Sabia and Burkhauser (2010) in

terms of statistically non-significant effect of the share of older individuals aged 54-64.

Conclusion

This chapter reviewed two research fields, research about the status of older workers in

the labour market and research about the effect of the minimum wage, with special

attention to older workers, in as much as they dealt with one of the two axes in subject

93

matter for this thesis. The review suggests that the two research fields have developed

apart from each other with no apparent common interest. Research about the labour

market status of older workers has focused exclusively on employment issues, including

hiring, dismissal, redundancy, retirement, and discrimination. This informs of older

workers’ vulnerability in the labour market and, at the same time, addresses the need for

research about the issues of wages, earnings, and policies relating to income from work

among older workers, with a focus on their vulnerability. Research about the effect of

the minimum wage may answer the need in as much as the minimum wage is a public

policy that intervenes in wages and earnings for vulnerable workers in the labour

market. But, despite an enormous amount of work in research about the effect of the

minimum wage, older workers have long been ruled out in the research field. Very few

empirical studies focused on or paid special attention to the impact of the minimum

wage on older workers, having the main interest in the issue of employment. A minority

of studies incorporated the results for older workers as a part of their analysis, and a few

others included older people in their studies only for controlling the demographic effect

and testing age effects. The findings from the very few studies showed that the effects

of the minimum wage for older workers were theoretically unexpected and hard to

understand within the mainstream theory of the minimum wage.

As noted earlier, the existing empirical evidence on the effect of the minimum wage on

older workers addresses the need to gain more empirical evidence, enlarge the subject

matter other than the issue of employment, and seek an alternative explanation based on

the review of established theoretical accounts. Aiming at responding to this need

through the later parts of this thesis, the following chapter will discuss a variety of

theoretical perspectives of the minimum wage and then develop a research design for

this thesis.

94

Notes

1 The industrialism theory argues that the effect of public pension policy is limited, stressing the effects of economic growth. However, determinants to explain the labour force participation of older population may change over time. According to Pampel and Weiss (1983), economic growth is the best predictor of the labour force participation during the early time periods whereas for later time periods pension programmes may have larger effects on the participation than economic development variables. 2 The results of their examination pointed to the structural conditions of work organisation and economic actors as the main driving forces of early retirement. 3 The basic idea of Figure 4.1 was derived from Szinovacz’s (2003, p.23) ‘retirement decision process.’ 4 Austria, Denmark, France, Germany, Greece, Italy, the Netherlands, Spain, Sweden, and Switzerland were included in the analysis. 5 See Burtless, G. (1986). Social security, unanticipated benefit increases, and the timing of retirement. The Review of Economic Studies, 53(5), 781-805; Krueger, A. B. and Pischke, J. (1992). The effect of social security on labour supply: A cohort analysis of the Notch generation. Journal of Labour Economics, 10, 412-437; Borsch-Supan, A. and Schnabel, R. (1998). Social security and declining labor-force participation in Germany. American Economic Review, 88(2), 173-178; Coile, C. C. and Gruber, J. (2007). Future social security and entitlements and the retirement decision. Review of Economic Studies, 89(2), 234-246; Liebman, J. B., Luttmer, E. F. P., and Seif, D. G. (2009). Labour supply responses to marginal Social Security benefits: Evidence from discontinuities. Journal of Public Economics, 93, 1208-1223; and Manoli, D. and Weber, A. (2010). International substitution in labor force participation: Evidence from policy discontinuities. IZA Discussion Paper No. 5248. All were quoted in Staubli and Zweimüller (2013, p.18). 6 According to Queisser and Whitehouse (2006), ‘actuarial’ has been increasingly used in the analysis of pension systems and retirement incentives. They discuss two actuarial concepts, ‘actuarial fairness’ and ‘actuarial neutrality,’ and define them as in the following:

‘Actuarial fairness, which requires that the present value of lifetime contributions equals the present value of lifetime benefits. Actuarial fairness relates to the entire lifetime of contributions and benefits.’; ‘Actuarial neutrality, which requires that the present value of accrued pension benefits for working an additional year is the same as the year before (meaning that benefits increase only by the additional entitlement earned in that year). Conversely, retiring a year earlier should reduce the pension benefit both by the entitlement that would have been earned during the year and by an amount to reflect the longer duration for which the pension must be paid. Actuarial neutrality is a marginal concept, relating to the effect of working an additional year.’ (Queisser and Whitehouse, 2006, p.4)

‘Actuarial neutral’ in this thesis is loosely used for meaning no implicit tax on working beyond the pension eligibility age.

95

7 The first-generation studies are well described in Card and Krueger’s (1995), and the new generation ones published since the 1990s are comprehensively reviewed in Newmark and Wascher’s (2008). 8 Meyer, R. and Wise, D. (1983). The effects of the minimum wage on the employment and earnings of youth. Journal of Labor Economics, I, pp. 66-100.

96

Chapter Five Theoretical Framework and Research Methods

Introduction

Studies about the effect of the minimum wage rely primarily on mainstream economics.

The neoclassical paradigm has long dominated ideas about the effects, and orthodox

neoclassical models portray that the minimum wage entails adverse employment effects

and is an ineffective social policy tool for helping poor families (Neumark and Wascher,

2008, p.249). However, as the disemployment effect has not been supported by many of

empirical studies, monopsony models which modify the key assumptions embedded in

the neoclassical model have expanded the explanatory power of the economic theory,

opening the door for positive or neutral employment effects. Meanwhile, the Keynesian

paradigm, which supports government interventions in response to market failure in

contrast with the neoclassical perspective, has emerged in recent years as an alternative

economic approach for research on the effects of minimum wages. This perspective

argues that the minimum wage would stabilise the macroeconomic conditions,

preventing deflationary development, especially in period of economic crisis and under

weak union power, and helping to create a more equal income distribution with changes

in the structure of wages and the distribution within the working class (Herr,

Kazandziska and Mahnkopf-Praprotnik, 2009). However, both the neoclassical view

and the Keynesian approach rules out the political nature of the minimum wage from its

explanation by excluding political means from government interventions. Relatively

rare as they are, other alternative explanations can be sought in political economy

perspective. A problem of the political economy approach to the minimum wage is that

they are concerned with what other factors than economic ones make people support or

oppose the minimum wage rather than how the minimum wage affects workers.

The first part of this chapter discusses those existing theories of the minimum wage and

seeks an alternative theoretical framework for an enhanced understanding of the effects

of the minimum wage which contains both political and economic attributes as a public

97

policy. As mentioned in Chapter Four, the discussion of the existing theories aims at

being advised of the extent to which they deal with the issue of the minimum wage in

terms of the ‘disruption’ made by its introduction or uprating and assume its

stakeholders’ rationality in the labour market, and thus at providing leads for seeking an

alternative framework founded on practical limitations of the existing theories. Then,

the second part develops a research design outlining research methods, data, and

analyses procedures, which are employed in later chapters of this thesis. In this part, the

necessity to employ a contextual analysis with qualitative data is argued, in that the

alternative theoretical framework sought in the first part of this chapter allows for the

influence of political attributes of minimum wage policy on top of its economic ones.

Economic Effects of Minimum Wages in the Neoclassical Paradigm1

Distribution Effects

Compared to the elaboration for the employment effect which will be presented in the

following section, the theoretical account of the effect of the minimum wage on the

wage distribution has been relatively less advanced. However, economic theory

grounded in the neoclassical paradigm predicts that minimum wages bring about mainly

two effects in the wage distribution, a spike and spillovers. If it is adequately enforced

and complied with, the minimum wage will cut off or thin out the lower part of the

wage distribution (Neumark and Wascher, 2008, p.107). This causes wages initially

below the minimum wage either to disappear or to push them up to the new minimum

wage, creating a spike at the minimum (Dittrich, Knabe, and Leipold, 2014, p.792).

Also, an increase in the minimum wage can lead to changes in the wage interval above

the minimum wage for several reasons. The simplest neoclassical models which

assumes two types of labour, skilled labour and unskilled labour, or only one type of

skill with different degrees of it among workers, does not suggest a spike in the wage

distribution but predicts a wage boost for workers already paid above the minimum

wage through the substitution of skilled workers or those with more skill for unskilled

workers or those with less skill. However, in response to the spike of the wage

distribution observed at the minimum wage and the spillover found more noticeably in

98

wages only a little higher than the minimum, the model developed by Pettengill2

explains the spike either as a result of the fall in employment of workers directly

affected by the minimum wage until the marginal revenue product of workers in that

group increases to the point that it is equal to their wage or as a result of their upward

adjustment of effort for productivity improvement (Neumark and Wascher, 2008,

pp.109-110). This model assumes jobs with different sensitivity of worker productivity

to skill as well as a continuous distribution of worker skills and a labour market

equilibrium characterised by an upward-sloping wage curve that relates wages to skill,

and regards the sensitivity of productivity to skill as the reason for the spillover through

‘a greater degree of substitutability between workers whose skill levels are relatively

close than between workers whose skill levels are quite different’ (Neumark and

Wascher, 2008, p.109). A highly skilled worker in a job for which productivity is less

sensitive to skill would be only marginally more productive, but paid a lot more, than a

low skilled worker and thus tends to be hired in a job for which productivity is more

sensitive to skill. Accordingly, the minimum wage leads to the substitution of slightly

higher low skilled workers for the lowest skilled workers rather than of those with quite

different levels of skill, and raises the wages of the former by more than the ones of the

latter (Neumark and Wascher, 2008, p.109).

Monopsony models which assume the existence of frictions in the labour market which

allows employers to have potential market power over their workers (Manning, 2003,

p.4) also predict similar effects on the distribution of wages. In the original Burdett-

Mortensen model in which each employer posts a wage and employees randomly search

employers for a job paid at a higher wage (Burdett and Mortensen, 1998, p.258), spikes

are not expected because employers offer a wage slightly higher than the one paid by a

mass of firms (Manning, 2003, p.327; Neumark and Wascher, 2008, p.111). But,

Manning understands it as a result of ‘a discontinuity in the labour supply function

facing each firm at every wage that is paid by a mass of firms’ and suggests that taking

non-pecuniary firm features into account, such as ‘heterogeneity in the evaluation of

non-wage characteristics of jobs or mobility costs,’ allows to generate a spike in the

distribution of wages by eliminating the discontinuity in the labour supply function

(Manning, 2003, p.327; Neumark and Wascher, 2008, p.111). This equilibrium search

99

model predicts spillovers concentrated among jobs paying just above the minimum

wage as well (Neumark and Wascher, 2008, p.110) because reduction in the wage gap

between higher-wage firms and lower-wage firms caused by an increase in the

minimum wage leads to increases in the elasticity of the labour supply curve to an

employer and consequently, induces employers to pay higher wages, but this effect

declines as one moves up in the wage distribution (Manning, 2003, pp.328-329). A

structural search model, in which Flinn (2002; 2006, p.1021; 2010) assumes wage

bargaining as well as search frictions in the labour market, illustrates that the imposition

of a binding minimum wage constrains matches between employers and workers who

are worth being paid less than the minimum wage, and the constraint produces an

equilibrium wage distribution with a mass point at the minimum wage and continuously

distributed wages above the minimum under the condition that match values are

themselves distributed in succession. Meanwhile, an efficiency wage model developed

by Grossman (1983, p.361) assumes two types of labour, skilled labour which reduces

its effort following a relative wage deterioration and unskilled labour whose effort does

not vary with changes in relative wages, and explains that the spillover effects of the

minimum wage are generated through two channels, the equity effect and the

substitution effect. This is, employers directly raise other wages above the minimum

wage as it increases in order to maintain the effort of skilled labour and the optimal

level of productivity, on the one hand, and they lay off minimum wage workers and hire

more skilled ones, on the other hand (Grossman, 1983, p.366).

Alongside the effects on the distribution of wages, the effects on the distribution of

incomes and on poverty are another chief distributional concern of minimum wage

policy. As shown in Chapter Four, the focus of the research about the effects on the

distribution of incomes and on poverty shifts from individuals’ welfare to families’

economic well-being, but this subject matter has also been theoretically less developed.

The main reason would be that the implication of the effects of the minimum wage on

welfare is concerned not only with distributional effects but also with other issues of

minimum wage policy, including employment effects and the relationship between low-

wage workers and low-income families and between low wages and poverty.

Employment effects have been particularly paid great attention because of their

100

expected adverse impacts that would impede the distributional goal of minimum wage

policy. The simple neoclassical model which does not assume market imperfection

explains that minimum wages create distortion that particularly results from negative

employment effects and reduces welfare, and it suggests that the desired distributional

goals would be better achieved by tax-and-transfer system to avoid the distortion

(Gramlich, 1976, p.410; Neumark, Schweitzer, and Wascher, 2005, pp. 867-868;

Neumark and Wascher, 2008, p.142). Models that modify the assumptions of the

neoclassical model imply ambiguous welfare effects based on employment effects.

Flinn (2002; 2006; 2010) uses the value of unemployed search as the welfare criterion

in his structural search model. He argues that positive changes in the distribution of

wages in response to increases in the minimum wage can be misunderstood in disregard

of employment effects, and that neither an absence of employment effects nor

disemployment effects necessarily mean improvement or deterioration in welfare

respectively because changes in the distribution of wages can affect the value of the

unemployed search. Rebitzer and Taylor (1995, pp.253-254) also contend in their wage

efficiency model that a minimum wage can lead to economic well-being by increasing

employment in the long run as well as in the short run but can have ‘a seemingly

perverse positive employment effect’ when the efficiency wage a firm must pay

increases with employment for the reason that increasing a firm’s employment raises the

risk of unemployment for other workers under uncertainty as to product demand. In this

regard, they suggest that the implication of the minimum wage effects on welfare would

be better detected by empirical works rather than theoretical appeals with regard to

employment effects.

Meanwhile, a theoretical attempt which shifts its focus from employment effects to the

effects on the distribution of incomes or on poverty for measuring the effect of the

minimum wage on welfare is rare. Fields & Kanbur (2007) only provided a theoretical

framework based on income-sharing and demonstrated how four factors - the degree of

poverty aversion, the ratio of the minimum wage to the poverty line, the elasticity of

labour demand, and the extent of income-sharing – result in a rise, a fall, and no change

in poverty level. In the succeeding paper, they showed that wage differentiation,

employment in high-wage and low-wage jobs, and the elasticity of labour demand with

101

respect to the minimum wage determine the direction of the poverty change (Kapelyuk,

2014, p.8).

Employment Effects

As implied earlier, the effects of the minimum wage on employment have been the most

prominent issue in its policy evaluation, and hence much of the theoretical debate about

the minimum wage centres on this issue. The logic of the neoclassical standard model of

the minimum wage is quite simple with respect to the employment effects. It assumes

that labour and product markets are competitive, the labour covered by the minimum

wage is homogeneous, products are turned out with a combination of capital and labour

of which both are perfectly informed about the labour market, and all workers are

covered by the minimum wage (Neumark and Wascher, 2008). In such a perfectly

competitive labour market, the wage is set by the meeting point of the supply of and the

demand for labour, and the magnitude of employment is determined by the equilibrium

wage. If a minimum wage is fixed at above the equilibrium wage, a firm’s marginal cost

of production is raised, which gives rise to an increase in the price of products, a fall in

the demand for products, and finally a decrease in production. As a result, employers

adjust their businesses by reducing employment, in particular of low-wage workers,

who will be relatively overpaid compared to their contribution to the firms’ revenue

under a minimum wage higher than the equilibrium wage (Card and Krueger, 1995) and

by substituting capital for labour in the production process (Neumark and Wascher,

2008). In this simplified account, the market equilibrium wage is impersonally applied

to all firms, which means that firms do not choose wages paid to their employees,

workers are treated in the same manner as other inputs that employers buy, such as

machinery or electricity, are, and all employers purchase labour as much as they want at

the same market price (Card and Krueger, 1995).

The assumptions that this standard model postulated, however, ruled out many other

aspects of the operation of the labour market and the effect of the minimum wage, and

extended models to remedy shortcomings of the standard model have been developed.

The first question that an extended model addresses is whether all workers are covered

102

by the minimum wage. In less developed countries, the coverage of minimum wages is

often limited, and noncompliance is high. The two-sector models explain that the sector

not covered offers alternative opportunities to workers who cannot find the jobs in the

sector covered and thus the aggregate employment losses associated with an increase of

the minimum wage can be alleviated even though employment in the sector covered

decreases (Card and Krueger, 1995; Neumark and Wascher, 2008). Welch (1976) argues

that, with partial minimum wage coverage, jobs in the sector covered are restricted and

labour supply in the sector not covered increases as wages in the sector covered are

bound to above-equilibrium levels. The shift in labour supply could lead to a rise in

employment and a fall in wages in the sector not covered. Welch’s model assumes that

workers not chosen by employers in the sector covered work in the sector not covered if

the wage is higher than their reservation wage. Meanwhile, Gramlich (1976) and Mincer

(1976) presume that workers choose their sector based on the expected earnings in each

sector which refer to ‘wage times probability of having a job’ (Brown, 1988, p.136).

When the expected earnings in the sector covered are higher than wages in the sector

not covered, workers who do not get a job in the former flow into unemployment whilst

they work in the sectors not covered when wages in that sector are higher than the

expected earnings in the sector covered. Thus, in the case that displaced workers remain

unemployed until they can get a job in the sector covered, wage rise caused by an

increase in the minimum wage would not have a significant impact on employment and

wages.

The second question is whether the workers covered by the minimum wage are

homogeneous. An extended model assumes two skill groups, skilled workers and lower

skilled workers. If a minimum wage increases, skilled workers are substituted for lower

skilled workers as the employers’ demand for skilled ones rises; and as a result,

although the overall employment still declines due to the substitution effects of non-

labour inputs for lower skilled workers, the scale of the total drop in employment will

be smaller than the fall in lower skilled workers (Neumark and Wascher, 2008).

However, responding to the fact that the observed distribution of wages is apt to be

relatively continuous rather than discrete, another alternative model assumes that the

distribution of wages reflects the distribution of skills which is formed by different

103

amounts of human capital among different workers (Brown, Gilroy, and Kohen, 1982).

In this model, human capital is held together by education, experience, motivation, and

ability, the labour market pays for an ‘efficiency unit’ of human capital, and each

individual is a substitute in production (Card and Krueger, 1995). Under these

assumptions, the presence of the minimum wage indicates a rise in the market price of

human capital and unemployed status of workers with value of marginal product below

the minimum wage, which leads the entire distribution of wages, which has the left tail

truncated at the minimum, to rightward movement (Card and Krueger, 1995). In sum,

this model predicts that the larger the increase in the market price of human capital

induced by the minimum wage is, the smaller the elasticity of demand for human capital

in the labour market is, which brings about reduction in employment among low-wage

workers, and the higher wages for other workers are (Card and Krueger, 1995).

The third issue is that the standard model assumes an economy with only one industry.

Bearing the point in mind, Neumark and Wascher (2008, p.52) argue that extending the

model to include more than one industry with consideration for plausible substitution by

consumers between goods produced in different industries requires to allow for possible

general equilibrium cross-industry effects. If two products produced in separate

industries are substitutable for each other, and a minimum wage raises the cost for one

product more than for the other, the demand for the product which is produced with a

smaller share of minimum wage labour in costs will increase, leading to a growth in the

demand for less skilled workers in the industry producing the substitute (Neumark and

Wascher, 2008, p.52). Neumark and Wascher (2008) cited as an example the possibility

that, with an increase in the minimum wage, the demand for minimum wage labour

could shift from other types of restaurants to the fast food industry when the latter has

the lower cost share of minimum wage labour than the former. They noted that the

prediction of the standard neoclassical model was not made on an industry-by-industry

basis, and thus a failure by researchers to find a negative employment effect in an

industry should not be regarded as inconsistent with the theory (Neumark and Wascher,

2008, p.52).

104

The fourth problem is whether the minimum wage has the same effects on employment

and average hours worked per employee. In response to two different findings from

Card and Krueger’s (1995) and Neumark and Wascher’s (2000) studies on the effects of

the New Jersey minimum wage, Michl (2000) proposed an extended model that

distinguished the number of workers and their average working. Although labour input

and employment are often used interchangeably, labour input is the product of

employment and average hours worked per worker (Neumark and Wascher, 2008). In

the extended model, employers could cut back on labour inputs in response to an

increase in the minimum wage, either simply by dismissing workers or by reducing

working hours per employee (Michl, 2000, p.265). Particularly for the latter case, an

employer can hire more workers, in principle, with the margin associated with a

reduction in average hours worked per employee, though the firm’s total demand for

work hours declines (Michl, 2000, p.265). Strobl and Walsh (2008; 2011) further

expanded the model by allowing firms to choose combinations of the number of

workers and hours worked per employee in their production. In this case, the effects of

minimum wages on labour input are ambiguous and differ across firms and industries

because the changes in the two components of labour input, employment and average

hours worked per employee, are inversely related in response to the minimum wage

increase, and firms’ total work hours may even rise if there are small fixed costs for

hiring workers.

The final question is whether short-run and long-run effects of minimum wages on

employment are consistent. An extended model which distinguishes the two effects

claims that short-run and long-run adjustments to a change in the minimum wage are

quite similar because the high turnover among minimum wage workers makes hiring

and firing costs insignificant (Neumark and Wascher, 2008, p.52). Providing the

findings from their case study of fast-food restaurants, Card and Krueger (1995)

contended that a negative employment effect would be expected within a few months

after an increase in the minimum wage because employers hiring minimum wage

workers can adjust their staffing levels flexibly with no large costs of hiring and firing

due to the high turnover rates of workers, but the effect will decrease over time as the

real value of the minimum wage diminishes. However, recognising the possible gap

105

between short-term and long-term effects, Hamermesh (1995) argues that ‘sizeable

adjustment costs for capital or other inputs can result in lags in the adjustment of labour’

even if the costs of adjusting labour input are small (Neumark and Wascher, 2008, p.52).

Along the lines of Hamermesh’s argument, Neumark and Wascher (2008, p.53) define

the short-run effects as ‘the change in employment over a period during which the

capital stock is fixed’ and maintain that the decline in less-skilled labour will be

mitigated until firms can substitute capital for less-skilled labour in response to an

increase in the minimum wage.

Despite the theoretical insights the standard model provides and the considerable

modification throughout the extension of the standard model, the most obvious

difficulty with the neoclassical models of the minimum wage is that of explaining the

existence of neutral or even positive employment effects of minimum wages on affected

workers. Allowing for Fang and Gunderson’s (2009) findings shown in Chapter Four, it

is important to build an explanatory basis for neutral or positive employment effects,

particularly for the research that focuses on the impact on older workers. Although the

extended neoclassical models that add some supplementary conditions partially

acknowledge the likelihood of non-negative employment effects, the fundamental

inclination heads for adverse ones. A general feature of the neoclassical competitive

model is the assumption that firms are ‘price-takers in the labour market’ and ‘there is

no firm-specific component of wages’, but this assumption is an oversimplification

(Card and Krueger, 1995, p.369). Another major difficulty with the neoclassical models

is that their explanations are based on arithmetical calculations, which assumes that

employers and workers have all the information needed for their rational decision-

making. For example, provided that workers who do not get a job in the sectors covered

remain unemployed when the wages times probability of having a job in that sector are

higher than the ones in the sectors not covered, it is highly questionable whether all the

workers always make the same choice in practice under a certain combination of

conditions, on the one hand, and whether they are well informed of, at least, the

probability of having a job even if a wage rate is applied to all firms in a certain sector,

on the other hand. Moreover, the neoclassical models do not consider the interaction

between the demand change for low-wage workers and workers’ behavioural responses,

106

which result from a rise in the minimum wage. For instance, low-wage workers can

improve their productivity motivated by an increase in the minimum wage, and this

would lead to no demand change for the workers. Also, if required on-the-job-skills are

different between industries, low-wage workers might hesitate to shift from one industry

to another in accordance with the demand change for them in each industry.

As the adverse employment effect predicted by the neoclassical models has not been

firmly supported by empirical studies since Card and Krueger’s (1994) seminal work

was published, and theoretical and empirical debates on the effect of the minimum wage

have been intensely revived, the monopsony model became prominent, paving the way

for positive or neutral employment effects. This model discusses exceptional cases from

the viewpoint of the neoclassical labour market model (Herr, Kazandziska, and

Mahnkopf-Praprotnik, 2009, p.19). In the textbook monopsony model, individual firms

face an upward-sloping labour supply curve which mirrors homogeneous labour and

have some market power over wages to attract more workers. A single firm, a so-called

‘monopsonist’ sets a wage at the level that the marginal cost of labour and the marginal

revenue product of labour are even (Robinson, 1933, recited from Card and Krueger,

1995 and Neumark and Wascher, 2008). The monopsonistic market equilibrium wage is,

however, determined on the labour supply curve being lower than the marginal cost of

labour curve, and thus the gap between the wage and the marginal revenue product of

labour, what Pigou and Hicks referred to as ‘exploitation’ (Manning, 2003, p.49),

occurs. Also, the level of employment that results from the equilibrium wage is lower

than the one in the neoclassical competitive model because the marginal cost of labour

mounts more sharply than the wage as the monopsonist has to pay the higher wage to all

its workers (Ehrenberg and Stewart, 1997; Neumark and Wascher, 2008). Under these

circumstances, all cases of employment effects can be embraced. A minimum wage

fixed between the equilibrium wage and the point where the marginal cost of labour is

equated to the marginal revenue product of labour leads to an increase in employment as

the minimum wage reduces the marginal cost of labour. A minimum set at the point

where the marginal cost of labour curve meets the marginal revenue product of labour

curve has a neutral employment effect, and any further increase beyond the point brings

about a reduction in employment.

107

The point that a monopsonist faces an upward-sloping labour supply curve assumes that

the elasticity of labour supply is not perfectly infinite. This assumption is the most key

feature that clearly discriminates the monopsony model from the neoclassical model and

allows each monopsony model demonstrated in different ways. The traditional static

monopsony model postulates a single buyer of labour as an example of the imperfect

elasticity of labour supply. An early labour market example is the ‘company town,’

where a single employer dominates, such as coal mining communities in rural areas

(Ashenfelter, Farber, and Ransom, 2010, p.204). Ulsan city in the southeast part of

South Korea where Hyundai dominates with motor and shipbuilding industries could be

a present example. Under the static framework, some empirical studies validated the

assumption in several specialised labour markets, such as nurses (Sullivan, 1989;

Staiger, Spetz and Phibbs, 2010) and school teachers (Landon and Baird, 1971; Falch,

2010). However, neither is the company town common in the present labour market nor

is there any evidence of the limited elasticity of labour supply in low paid jobs. In this

regard, the traditional static monopsony model is regarded as just an intellectual

curiosity, not being relevant, particularly in the low-wage labour market. Critics of the

model suppose that the elasticity of labour supply to a certain firm is nearly infinite

unless the firm actually employs a considerable proportion of the total potential

workers, which belief is derived from the sphere of industrial organisation, where a

certain employer’s market power is reckoned to be correlated with the employer’s

market share (Card and Krueger, 1995, p.373). But, in the low-wage labour market, the

employers are normally considered as small firms which monopsonistic power is often

insignificant.

Making it more plausible in an industry with a large proportion of small firms, the

dynamic monopsony model, which was formulated by Burdett and Mortensen (1989)3

and advanced by Manning (2003), attributes the limited elasticity of labour supply to

search-related frictions in the labour market. The dynamic model assumes that, due to

the job search frictions, workers have imperfect information about the wages provided

by employers, and firms have monopsony power on wages even against many

competitors (Ashenfelter, Farber and Ransom, 2010). The basic idea of the dynamic

model is that the firm sets a wage at which the number of new hires balances the

108

number of turnovers (Card and Krueger, 1995), and thus the firm’s employment

corresponds to an equilibrium between the flow of workers in and those out of the firm

(Ashenfelter, Farber and Ransom, 2010). In this so-called general equilibrium model of

oligopsony, a binding minimum wage increase can lead to employment gains,

particularly at smaller, low-wage firms, by reducing turnovers and raising hires (Card

and Krueger, 1995, pp.380-381). This is possible not only because the minimum wage

increases the odds that some workers will be provided a wage offer higher than their

reservation wage, but also because it leads firms already paying above the minimum to

give their wage offer a raise as well (Neumark and Wascher, 2008, p.56). Card and

Krueger’s (1994) analysis of the effect of the New Jersey minimum wage on fast-food

restaurants provides empirical evidence, finding that employment increased among

restaurants initially paying the lowest wages and was stable among restaurants already

paying more than the new minimum wage.

Apart from the static and dynamic monopsony models driven by the assumption that

firms can attract and maintain more workers by paying a higher wage (Card and

Krueger, 1995), Rebitzer and Taylor (1995) developed an efficiency wage model in

which employers economise on monitoring costs by paying above-equilibrium wages

(Boal and Ransom, 1997, p.95). This model assumes that workers have some discretion

over the level of effort on the job, and firms can pay a higher wage to induce greater

effort of employees and threaten to fire employees who shirk (Card and Krueger, 1995;

Neumark and Wascher, 2008). Firms need to monitor employees’ efforts, but an

employer’s monitoring ability is supposed to decrease as a firm’s size is increased. If a

firm increases employment, it must raise wages to discourage shirking in lieu of its

weakened ability to monitor. This brings about an upward-sloping labour supply in the

long run (Boal and Ransom, 1997), and leads the marginal cost of labour to be higher

than the offered wage. Under these circumstances, imposing a binding minimum wage

slightly above the equilibrium wage induces firms to put less resources for monitoring

work effort and increase their level of employment instead (Neumark and Wascher,

2008, p.55). An implication of the efficiency wage model is that employers who are

concerned about low-wage workers’ shirking on the job will not offer subminimum

wages because workers being paid a subminimum would choose the desire to shirk on

109

the job rather than the value of keeping the job (Card and Krueger, 1995, p.382; 2000,

p.342) whereas those being paid over the subminimum would be concerned about losing

their jobs.

Empirical studies show that monopsonistic behaviour is found across the considerable

number of employers in a variety of countries and labour markets even though it is not

universal. With reference to the effects of the minimum wage, monopsony models have

critical implications by opening a door for theoretical underpinning of non-negative

employment effects. In addition, equity issues are always considered in the background

in the monopsony literature because the wage is regarded as an important factor of

economic welfare, specifically for low-wage workers (Boal and Ransom, 1997, p.87).

However, it may be asked whether the model can provide a universal explanation for the

employment effect of the minimum wage, regardless of the level of analysis.

Monopsony models before the general equilibrium model of oligopsony fundamentally

rely on firm-based modelling. It is obvious that a single result from a monopsonist

cannot be generalised in the labour market. Even in general equilibrium models, a

related problem remains. Noting that general equilibrium models distinguish between

the elasticity of labour supply to the labour market as a whole and to individual firms in

as much as the former determines an aggregate employment effect of the minimum

wage while the latter determines the gap between the marginal product of labour and the

wage, Manning (2003) argues that any measure of the impact on aggregate employment

must take firms’ heterogeneity into account because individual firms would otherwise

choose different wages and thus will have different employment effects in the

imposition of the minimum wage. But, most of the monopsony models, including the

general equilibrium models, do not incorporate heterogeneity in workers’ productivity

which would be a key factor to give an account of a firm’s heterogeneity, particularly in

terms of the level of employment. If all workers are assumed not to be equally

productive, monopsonistic firms with more productive employees than their competitors

can be supposed, for example. Faced with the imposition of a binding minimum wage,

those firms might not raise employment even if there is still a gap between the marginal

product of labour and the minimum wage. The firms, facing a minimum wage slightly

above the monopsonistic equilibrium wage, could even discharge some employees with

110

relatively less productivity in the hope that the rest of employees would improve their

productivity motivated by the increase in the minimum wage. However, the monopsony

models do not explain what the institutional forces are at play in practice or what the

role of wages is in the production process (Edwards and Gilman, 1999). Although the

latter is the focus of the efficiency wage model, higher wages as ‘a substitute for the

costs of supervision’ are criticised as unlikely and implausible (Edwards and Gilman,

1999, p.23). This is, ‘employees will avoid their duties if they can, and employers thus

face costs of supervision’ (Edwards and Gilman, 1998, p.9), regardless of the wage

level.

Economic Effects of Minimum Wages in the Keynesian Paradigm

Distribution Effects

Critical of the absence of a macroeconomic approach in the theoretical debate which has

been led mostly by the neoclassical view (Herr, Kazandziska and Mahnkopf-Praprotnik,

2009, p.5), the Keynesian model explains the effects of minimum wages in the context

of the interaction between the labour market and the national economy. The Keynesian

approach which was first advanced in the 1930s is set against a claim for market self-

regulation and instead focuses on systemic market failure to be resolved by government

intervention. The neoclassical approach recognizes the odds of market failure but

defines the failure as a bad fit between production and the wants of individuals with

purchasing power, which does not occur in an aggregate manner (Caporaso and Levine,

1992). By contrast, the Keynesian approach argues that the market mechanism

systemically fails to assure adequate purchasing power and underutilise the existing

productive capacity, and this aggregate demand failure results from the instability of the

process of reproduction and growth in a capitalist economy (Caporaso and Levine,

1992, p.100). The source of the instability is the self-reinforcing or cumulative nature of

the processes that the movements of output, investment, employment, and prices are

involved in ‘a leakage from the economic circular flow’ which does not stimulate any

comparable demand (Caporaso and Levine, 1992, p.103, p.118). The systemic failure of

aggregate demand requires government intervention, but a key point here is that the

111

intervention should adopt ‘administrative ways rather than political means’ (Caporaso

and Levine, 1992, p.101), and the government develops a stabilization policy that

affects the economic circular flow, typically using government spending and taxes, to

secure the macroeconomic conditions needed (Caporaso and Levine, 1992, pp.119-121).

According to the Keynesian explanation of the minimum wage effects which was

explored by Herr, Kazandziska, and Mahnkopf-Praprotnik (2009) and Herr and

Kazandziska (2011), minimum wages can perform two important roles in the Keynesian

paradigm, preventing deflationary developments, especially in a period of economic

crisis and weak union power, and changing the wage structure among wage earners. In a

closed economy, changes in profit rates, prices of natural resources, and taxes may

affect changes in the price level, but none of them can solely give rise to an inflationary

or deflationary process (Herr, Kazandziska, and Mahnkopf-Praprotnik, 2009, p.7). By

contrast, changes in the wage-price spiral can bring about those processes by itself, and

in this regard, nominal wages can be regarded as the nominal anchor which is a

precondition for the functioning of capitalist economies (Herr, Kazandziska, and

Mahnkopf-Praprotnik, 2009, pp.7-8). That is, when nominal wages increase by the wage

norm which refers to following the medium-term trend of productivity changes and the

inflation target of the central bank, they become the nominal anchor for the price level

and prevent deflation (Herr, 2002; Herr and Kazandziska, 2011). In the same vein,

minimum wage increases attuned to the wage norm support or become a nominal wage

anchor and avert a wage-price deflationary spiral by keeping down the erosion of

nominal wages (Herr, Kazandziska, and Mahnkopf-Praprotnik, 2009, p.8; Herr and

Kazandziska, 2011, pp.4-5). To achieve this goal, the Keynesian theory recommends

that the minimum wage should be adjusted annually with a rate high enough to give

benefits to the lowest wages and also affect a sufficient number of employees (Herr and

Kazandziska, 2011, p.5; Ghani, 2016, p.2).

In the Keynesian approach, changes in nominal wages cannot fundamentally change the

distribution of national income between wages and profits because wages affect the

price level (Herr, 2009, p.952). However, minimum wages have distributional effects

among wage earners (Herr, Kazandziska and Mahnkopf-Praprotnik, 2009, p.10). If all

wages change by the wage norm, the wage structure will not change, in which there is

112

no improvement in real income among low-wage workers. But, if minimum wages

increase faster than the average wage, the wage structure will be compressed from the

bottom, which increases real income and purchasing power among the low paid and

enhances the potential of minimum wage policy for combating poverty (Herr and

Kazandziska, 2011, p.5). In another case where minimum wages increase attuned to the

wage norm but the average wage level is raised higher than the wage norm, wage

dispersion will deepen, and hence increasing minimum wages based on the average

wage level is required (Herr and Kazandziska, 2011, p.5).

Employment Effects

Herr, Kazandziska, and Mahnkopf-Praprotnik (2009; 2011) argue that the minimum

wage does not have any systematic employment effects, and they remain theoretically

open in the Keynesian paradigm (Ghani, 2016, p.2). Changes in the wage level are

assumed to lead to changes in prices but not to directly induce changes in employment

in this approach (Herr, Kazandziska, and Mahnkopf-Praprotnik, 2009, p.9; Herr and

Kazandziska, 2011, p.3). The new structure of prices altered by an increase in the

minimum wage will change not only the technology and the structure of production but

also the structure of demand (Herr and Kazandziska, 2011, p.6). Higher minimum

wages may raise the level of consumption particularly because low-income households

show a higher propensity to consume compared to high-income households, which will

lead to a positive impact on aggregate demand, output, and employment (Herr and

Kazandziska, 2011, p.6). However, in some industries, such as security or cleaning,

increases in wages may refer to increases in price that leads to decrease in demand and

employment (Herr and Kazandziska, 2011, p.6). But, even in this case, those who

remain employed can consume more with higher income, and also, as products in one

industry can be inputs in another, the chain reaction of changes in prices will be

provoked with an obscure outcome in employment on a macroeconomic level (Herr and

Kazandziska, 2011, p.6).

With regard to the effects of the minimum wage, the Keynesian’s main concern is

whether the policy helps to stablize aggregate demand by enhancing purchasing power

113

among low-income households and to prevent deflationary development. A prerequisite

for this goal is that minimum wages should increase at least in line with the medium-

term productivity changes and the inflation target of the central bank. Increases faster

than the average wage level will bring about a compression of wage structure from the

bottom and a reduction in poverty. The Keynesian approach has little interest in the

mechanism whereby the minimum wage influences individuals. In this vein, it addresses

only three cases where the distribution of wages is differently adjusted in response to

the increase in the minimum wage and does not elaborate the theoretical relationship

between changes in wages and changes in employment. This makes it difficult to

understand a variety of effects of the minimum wage found in empirical studies. Further

to this, the Keynesian approach stresses only administrative intervention and disregards

the point that a minimum wage rate is essentially set through a political process,

although it takes several cases into account with regard to the level of the minimum

wage.

The Political Economy of Minimum Wages

What Makes Some Support and Others Oppose the Minimum Wage?

Both the neoclassical view and the Keynesian approach discussed so far rule out the

political nature of the minimum wage from their explanations. But, the minimum wage

as a public policy is a collective process, and its outcomes are shaped by both ‘the

domain of politics and the domain of markets as two sides of the same coin’ (Bickers

and Williams, 2001, p.x). Hence, its effects should be explained allowing for both

domains. This leads us to seek a political economy perspective for a more plausible,

real-world account of the effects of minimum wages. The existing theoretical

explorations underlying the political economy of minimum wages, however, focus

mainly on questions about what other reasons than economic ones make some support

and others oppose minimum wages, rather than on questions about how minimum

wages affect workers. These rely on voter behaviour or trade unions and are mostly set

in the US context.

114

West (1974) argues that the success of proposed legislation, including the minimum

wage, depends on the median voter group because politicians pay attention mainly to

the group’s preferences, but the group does not match up with those who are affected by

the minimum wage. The median voter group is not homogenous in terms of the

minimum wage. While some individuals, particularly who are affiliated to trade unions,

may be in favour of the legislation because their bargaining power will increase, the

majority in the median group are those whose real incomes may decrease with the

adverse price effects and small employers who are not typically wealthy businessmen

(West, 1974, pp.134-135). Some of the non-union median voters may have altruistic

virtue besides self-interest, based on the informed effects of the minimum wage (West,

1974, p.134). In this regard, West (1974) suggests in his theoretical model that the

probability of support for minimum wages is determined by the voting significance of

the group affected by the minimum wage, the proportion of non-union individuals and

small business employers, effects of the minimum wage on the median voter group, the

extent to which sub-groups in the median voting group are politically organised, and the

degree to which voters perceive the adverse effect of the minimum wage (West, 1974,

p.135).

More recently, Adam and Moutos (2011) constructed a political economy model to

show political support of median voters for the minimum wage in the context of

whether they would agree to replace the existing minimum wage with wage subsidies.

Their model assumes that there are a fixed number of households and many perfectly

competitive firms, and the majority of households earns income only from labour

(Adam and Moutos, 2011, p.171). This implies that most voters are workers whose

income is derived from the sale of their labour or from unemployment benefits, and the

imposition of a minimum wage will reduce employment but increase aggregate wage

income and the income of those who remain employment (Adam and Moutos, 2011,

p.172). Adam and Moutos (2011) predict that, if the fiscal cost of wage subsidies are

covered on by the increase in the wage income tax, median voters would not support the

replacement of minimum wage with wage subsidies, but, if the cost is paid by firms

through higher profit tax rates, wage subsidies can get broad political support from the

median voter (Adam and Moutos, 2011, p.173).

115

Meanwhile, attempting to recast the argument on support for the minimum wage,

Waltman (2000) argues that the reason why widespread public support for the minimum

wage does not lead to the American electorate’s support for its legislation is lack of

mobilisation which results not only from the fact that minimum wage workers, mostly

young, less educated, and poor, have a strong tendency not to vote and are unlikely to

maintain their political will for their own economic interests, but also from the fact that

there is the disparity between proponents’ scientific knowledge-based approach and

morals and ethics-based perception of the minimum wage among the public and the

potential mobilisers, such as politicians and trade unions. Waltman (2000, p.142)

suggests that minimum wage advocates need to recast the debate away from an

individualistic framework ‘toward a common good of society ’ as emphasised in ‘the

progressive era rhetoric of civic republicanism’ (Prasch, 2001, p.356) .

Compared with those who build their arguments upon voters’ behaviour, some social

scientists hypothesize that minimum wage policy has been driven by the pressure of

organised constituent groups that support or oppose the minimum wage in their own

interest (Neumark and Wascher, 2008, p.253). Trade unions are regarded as one group

able to organise a large bloc of voters in favour of the minimum wage because it gives

the incentive of rises in demand for higher skilled unionised workers to unions

representing workers in higher-wage industries (Neumark and Wascher, 2008, p.253) as

well as the incentive of increases in wages to unions representing those in low-wage

industries and also reduces competition in low-wage industries (Simons, 1944, pp.10-

11). Assessing the potential constituent groups in terms of the relative size of each and

the significance of minimum wage laws in their voting decision, Keech (1977) argues

that three important points enable the minimum wage politically viable. First,

individuals who benefit from an increase in the minimum wage are likely to consider

the minimum wage more heavily in their voting decision than those who do not (Keech,

1977, p.136). Second, unlike West’s (1974) argument presented earlier, the perception

of the adverse effect of the minimum wage would have little influence on most voting

decision. Third, workers who are paid slightly above the minimum wage, support the

increase in the minimum wage to reduce competition from lower wage counterparts,

and are represented by trade unions would have the greatest organisational power.

116

Cox and Oaxaca (1982) provide a more stylised theoretical model to predict the support

or opposition of the minimum wage among competing interest groups in the context of

an equilibrium model which assumes the unionised sector uses skilled labour and

capital as inputs while the non-unionised sector uses unskilled labour, skilled labour,

and capital as inputs (Neumark and Wascher, 2008, p.254). In this model, an increase in

the minimum wage raises wage costs in the non-unionised sector, which leads to

increases in the price of their products and the substitution of products produced in the

unionised sector for ones produced in the non-unionised sector. This product market

substitution drives an increase in the demand for union labour and its employment (Cox

and Oaxaca, 1982, p.552), and hence, trade unions have an incentive to support the

minimum wage for their own self-interest (Neumark and Wascher, 2008, p.254). By

contrast, an increase in the minimum wage reduces the real rental rate of capital and

harms the economic self- interest of capitalist, though the economic self-interest of

capital is not necessarily the same with the interest of corporation executives in the

minimum wage legislation (Cox and Oaxaca, 1982, p.542). Cox and Oaxaca’s

explanations are based on the assumption that politicians use their vote for maximising

utility of re-election and it ‘depends on the relative importance in the political

marketplace of competing interest groups’ (Cox and Oaxaca, 1982, p.552).

Meanwhile, Hayes (2007) identifies the minimum wage as a redistributive policy with

three different patterns of politics according to whether interest groups are organised or

not. The first pattern refers to the case that the issue of the minimum wage lies dormant

because low-wage workers, the primary beneficiaries of the minimum wage, fail to

organise themselves and thus effective demand for an increase in the minimum wage is

hard to be formed whereas there is constant, intense opposition to an increase in the

minimum wage from business groups. This case is particularly harmful to low-wage

workers ‘because the wage has never been indexed against inflation’ and thus ‘the real

value of the minimum wage declines’ (Hayes, 2007, p.469). The second pattern is the

case that the minimum wage issue constitutes ‘the ongoing struggle between labour and

business over shares of national income’ as trade unions take up the issue even if

beneficiaries fail to mobilise (Hayes, 2007, p.470). In this pattern of politics, a relative

power of labour and business is correlated with the magnitude and frequency of the

117

minimum wage increase4. The third pattern is the case that, along with the decline in

trade unions, proponents of the minimum wage increase take a new strategy, such as ‘a

combination of tax breaks and subsidies to business’ to offset the increase in wage costs

(Hayes, 2007, p.471). In this case, an increase in the minimum wage shifts the cost from

employers to taxpayers and consumers rather than redistributes income from employers

to low-wage workers.

The Level of the Minimum Wage and its Fixing Process

A different political economy explanation relating to the minimum wage has been

addressed by Brown (2009). Paying attention to the importance of the minimum wage

fixing process with regard to the level of the minimum wage, Brown (2009) analysed

the shifting internal dynamics of the Low Pay Commission and its negotiations over the

level of the minimum wage in the British context. He argued that confrontation between

Commissioners with diverse backgrounds was unavoidable on the issue of the level of

the minimum wage but the position of the Commissioners were complex, which led to

the negotiation ‘in a pluralist way’ (Brown, 2009, p.441). Within the Commissioners’

group with a trade union background, there were different interests relating to increases

in the minimum wage by sector, and the Commissioners’ group with a business

background also differed by firm size. Despite the ambivalent interests within each

group, the union Commissioners were consistent in the upward direction of their

influence ‘with the strong egalitarian ethos’ while the employer Commissioners were in

the downward direction of ‘their influence as one of restraint on the union

Commissioner’ (Brown, 2009, p.442).

The existing theoretical explanations of the minimum wage from the political economy

perspective give limited insight into how political power shape the effects of the

minimum wage on workers. Most of the political economy approaches in common sees

government as an interest group in pursuit of re-election, and this results in confining

the role of the state in the policy process to a ‘casting voter’ who makes a decision

based on a refined calculation of which group would be more helpful for their re-

election between employers and employees. Further, policy typology-based

118

explanations tend to be derived from Lowi’s early premise that ‘policies determine

politics’ (Lowi, 1972, p.299, see Lowi 1964, p.688), from which Lowi (1988, p.726)

later stepped back by acknowledging both directions of causal relationship between

policy and politics. This may lead to an explanatory vacuum of policy outcomes

produced or reproduced by the political process and overlook the changing nature of

policy outcomes which brings about variations in the effects of the policy. Unlike other

strands of the political economy explanations, Brown (2009) paid attention to the

importance of the minimum wage fixing process with regard to the level of the

minimum wage. However, he did not propose an explanation of the relationship

between the fixing process and minimum wage effects.

A Proposed Political Economy Framework for Understanding the Effects of

Minimum Wages

The above discussion of the existing theoretical accounts of the minimum wage tells us

that the mainstream economic theory provides predictions about changes in wages,

income, employment, and price, based on the view that the introduction or uprating of

the minimum wage is a ‘disruption to the normal labour market’ and the assumption that

employers and employees are rational actors who respond to the disruption towards

minimising losses and maximising gains. The modified models and the alternative

economic theory relax the view of the minimum wage as a disruption. However, they

keep leaning on the assumption of rational individuals in the labour market. Since

economic theories of the minimum wage, especially ones grounded in the neoclassical

paradigm, explain the fundamental mechanisms of how agents in the labour market

behave in response to the external stimulus, they are obviously indispensable for

exploring the effects of the minimum wage. But, as mentioned earlier, the predictions of

the mainstream economic theory are not supported by many empirical studies, and some

of the findings, including Fang and Gunderson’s (2009) presented in Chapter Four, are

not clearly understood even by modified models or alternative economic theory.

Accordingly, two significant questions may arise. One is whether individuals behave

rationally in the labour market, and the other is whether the effects of minimum wages

can be explained as the outcome of those repeated actions. The inconsistency between

119

theoretical predictions and empirical findings implies that agents in the labour market

might not repeat their actions when facing similar situations and that the effects of

minimum wages may be contextualised by reflective behaviours. Whether individuals

repeat rational actions in a similar labour market situation is not a focus of concern in

this thesis, which would be a merited question for the field of economics, in particular

for behavioural economics or labour economics. What this thesis pays attention to as

regards the questions and implications being addressed above is the potential of

observable reflective behaviours which may allow better understanding of the minimum

wage effects. At this point, it should be noted that the minimum wage is a public policy

produced through the political process. A set of policy outputs resulted from

stakeholders’ reflective actions in the minimum wage policy process would

contextualise the effects of the policy, in as much as they embrace a variety of intents

formed by debates and decisions in the process. In this regard, this section seeks a

theoretical framework which would complement the explanatory vacuum of existing

theories with a lens of putting more weight on the policy process behind the minimum

wage.

Minimum wage policy can be a political arena of keen power competition among the

key stakeholders in as much as it is involved not only in regulating the behaviour of

employers (Hayes, 2007, p.467) but also in redistributing profits from employers to

employees (Levin-Waldman, 2011). This suggests that, bearing on minimum wage

policy, power is neither equally distributed nor unchangeably structured among the key

stakeholders, and its dynamics dominate the policy process. The dynamics of power in

terms of the minimum wage rely on the extent to which each stakeholder, in particular

employees, can be collectively organised and mobilised for or against the minimum

wage (Eyraud and Saget, 2005; Hayes, 2007; Neumark and Wascher, 2008). Labour

unions enable employees to have the force to make their voice heard more effectively

and form a pivotal authority for obtaining favourable outcomes. However, the union

density of low-wage workers is typically very low, and they are seldom collectively

organised not only because they are directly and indirectly pressured by their employers

into not joining trade unions, but also because they are often physically scattered by

their workplaces (Hayes, 2007). Besides, some of trade unions speak for higher-skilled

120

workers with higher pay (Simons, 1944; Kau and Rubin, 1978), and this possibly leads

them to be indifferent to or even to oppose the minimum wage in their members’ own

self-interest. But it has long been recognised that labour unions representing higher-

skilled workers have an incentive to support the minimum wage because it may increase

labour demand for higher-skilled unionised workers (Neumark and Wascher, 2008) or

play the role of a wage floor to boost overall wage levels. In this regard, labour unions

typically spend a significant amount of time and resources on campaigning and

lobbying for minimum wage increases. They call for favourable public opinions and try

to have prominent political leaders on their side in order to get greater leverage for an

increase in the minimum wage.

Business groups have been effectively organised throughout the history of the minimum

wage (Levin-Waldman, 2001; Waltman, 2000, cited from Hayes, 2007, p.473). They

exert influence on the policy process, primarily using their instrumental power, ‘which

stems from political resources, such as informal ties to politicians and bureaucrats,

appointment or election to government office,’ including finance ministers or central

bank heads, and ‘core constituency relationship with political parties’ as well as from

‘organisation, money, technical expertise, and media access, all of which place business

in a stronger position to lobby, orchestrate collective action, command authority in

policy debates, finance campaign, and/or shape public opinion’ (Fairfield, 2015, p.420).

However, business interests in terms of the minimum wage can differ by the size of a

firm. Small businesses routinely oppose an increase in the minimum wage because they

typically hire more unskilled labour and thus are likely to be more directly affected by

the increase, whereas the position of big firms, of which many are only marginally

affected by an increase in the minimum wage, is theoretically ambiguous since they

may oppose higher minimum wages in order to have more leverage in wage bargaining

but could support higher minimum wages to ‘eliminate lower-cost rivals (Rustici, 1985,

p.120)’ (Dickson and Myatt, 2002, p.60; Neumark and Wascher, 2008). There is also a

gap in resources and organisational cohesion for political influence between small and

big business groups. Big business organisations can exert strong influence for their own

self-interests in the policy process, utilising the ample resources described above and

insignificant free-rider problems. By contrast, small firms’ associations ‘are a diffuse

121

group with free-rider problems’ (Dickson and Myatt, 2002, p.60) and may not have

sufficient resources to lobby effectively, organise collective action, dominate policy

debates, finance campaign, and shape public opinion, but ‘echo(ing) the free market

editorial writers’ (Waltman, 2008, p.27). In this regard, they often devote their efforts to

minimising an increase in the minimum wage rather than to blocking any increase in

and of itself (Waltman, 2008). These potential differences in interests and resources

between small and big businesses may create variations in the intensity of business

opposition as a whole to (or possibly support for) the minimum wage. Nonetheless,

‘business interests remained and remain adamantly and resolutely opposed to the whole

idea of a minimum wage’ (Waltman, 2008, p.3) because it interferes with ‘a basic

business decision (Waltman, 2008, p.3) and ‘their entrepreneurial liberty’ (Levin-

Waldman, 2001, p.105) and ‘legitimates governmental regulation in general’ (Waltman,

2008, p.3). ‘Much of the intensity of business opposition centres around the law’s

intrusion into the private relationship between employers and employees’ (Hayes, 2007,

p.467).

Government is also a key stakeholder of the minimum wage not only because it is the

final decision-maker and implementer of the policy, but also because it has its own

‘comprehensive interest’ in the policy, which is associated with a regime’s policy

orientation as well as its pursuit of political supports from the electorate. The capacity to

be organised rarely matters for a government, in as much as it is an organised entity

with power anchored in representative democracy and/or bureaucratic authority.

However, when a government deals with an issue, the low degree of coherence among

the relevant government departments may weaken its organisational power. When it

comes to government power as regards the minimum wage, a government’s

comprehensive interest to pursue both its own ideological orientation towards policies

and political support from the electorate for its re-election is critical. A regime’s policy

orientation is closely connected to its political interest in preserving the regime, but they

are often at odds with each other in practice. The extent and intensity of power that a

government exerts on a policy issue depend mainly on the interaction with other forces

which would have a strong influence on its policy orientation to be sustained or

obstructed. Public opinion and structural power are critical forces as regards minimum

122

wage policy.

Public opinion influences government decisions regarding a policy when an idea is

widely shared by the general public, and the support for the idea intensively converges

enough to mobilise extensive civic groups and/or to attract a key enabler, such as a

powerful political leader. The minimum wage tends to be widely supported by the

general public with moral sympathy and anticipation that it would do little harm to their

working life (Waltman, 2000). But it rarely comes up as a leading political agenda item

because the issue is not of direct interest to the majority of society and thus seldom

draws the mobilised political participation of civic society and/or strong support from

those with political leverage. Weakened trade unions and/or tepid support from the

unions representing more skilled workers would not be able to promote the mobilisation

of public opinion in favour of the minimum wage. Further, corporatist institutions

between labour, business, and government, which are adopted in many countries for

minimum wage fixing procedures as seen in Chapter Two, tend to deter the minimum

wage from being brought into the public political agenda by giving an impression that

‘government limits the power of business’ (Hill, 2009, p.56) and ‘depoliticising the

issue’ (Hill, 2009, p.59).

By comparison, structural power is a potent force that forms the base of the economic

power of business by coordinating the market in the capitalist economy. It constantly

influences government’s comprehensive interest regarding policies, including the

minimum wage. As Block (1992, quoted in Hundt, 2009, pp.23-24) defined, structural

power refers to the ability of business to prevent changes that conflict with the broad

interests of capital by threatening to withdraw their resources from the economy and

‘arises from the fact that, in market societies, states depend on private-sector agents to

invest in ways that generate growth, employment, and prosperity’ (Fairfield, 2015,

p.413). This is a specific variety of power that business is endowed with but differs from

the power that business holds as organisations, in as much as, unlike the organisational

power of business, the threat from structural power is ‘implicit … Government

anticipates the interests and reaction of capital without the need for interest group

representations’ (Marsh,1983, p.4, quoted in Hundt, 2009, p.23). However, structural

power can enhance the organisational power of business by facilitating government

123

and/or policymakers to embrace what business organisations demand, and at the same

time, organisational power also can strengthen structural power by reinforcing the

concerns of government and/or the general public on investment using its plentiful and

powerful resources (Fairfield, 2015). This suggests that the economic power of business

which is founded on both structural power and organisational power operates as a form

of political hegemony (Gramsci, 1971). The extent to which government policy

decisions are influenced by the economic power of business relies mainly on the extent

to which government policy orientation is market-oriented, the extent to which the

majority of voters supporting the government prefers pro-market interventions, and the

extent to which economic conditions and outlook a government faces afford to cope

with predictable and unpredictable risks that arise from the changes in policies. In terms

of the minimum wage issue, a government tends to stick to its own policy orientation

which enables it to maintain its foundation of political support unless radical changes in

macroeconomic conditions or a powerful mobilisation of public opinion against the

policy orientation occur. This has an important implication for the characteristics of

policy change as regards the minimum wage, which will be discussed in connection

with regime change below.

The facts that government is structurally in an intermediate position between labour and

business and that it is ‘not a neutral agent’ (Hill, 2009, p.40) but one with its own

interests, suggest that government plays a decisive role in the dynamics of power as

regards the minimum wage. In accordance with its policy orientation, a government can

reduce the potential prior imbalance of power between labour and business, give more

weight to one of the two conflicting power in an active or passive way, or remain a

spectator. These actions may not be consistent through all the years in office and/or all

the stages of the policy process at a certain time. For example, a government might

revise its position whenever a new minimum wage rate is fixed within its years in office

and/or could take a stance to balance labour and business in the stage of agenda setting

and policy formulation, take any sides in the stage of decision-making so as to pursue

their own interests, and remain a spectator in the stage of implementation in order to

offset risks from a policy decision. Different government actions lead to a variety of

power dynamics in the policy process and different types of equilibrium of power

124

among the key stakeholders of minimum wage policy, such as explicit consensus, tacit

consensus, influence, and domination.

The dynamic of power, which results in forming an equilibrium among the key

stakeholders at a particular time and stage of the policy process, assigns implicit

purposes and characteristics to the policy by producing a set of policy outputs.

Regardless of its legislative purpose and redistributive trait, a minimum wage has a tacit

aim and intent as a consequence of policy outputs. A typical example is that the

minimum wage is covertly characterised as symbolic when a rate is fixed at a much

lower level to achieve its original, explicit goal and/or when compliance with the law is

not properly supervised. This could be a case that the equilibrium of power is formed as

domination by business or implicit consensus between business and government at the

stage of decision-making and /or at the stage of implementation in the condition that

labour power and public opinion are weak, government policy orientation is more in

line with business interest, and/or government actions are strongly influenced by

structural power. However, this does not mean that a particular set of implicit purposes

and characteristics of a minimum wage are matched to a certain combination of

government actions and a form of equilibrium of power in the policy process. If a

minimum wage rate is fixed at a certain level for a specific objective, the minimum

wage comes to have an implicit purpose, for instance, an increase in employment of the

overall population or a part of it. In this case, government actions and the equilibrium of

power may take the same configurational forms with the case of the minimum wage

with implicitly symbolic characteristics. Changes in the implicit purposes and

characteristics of a minimum wage can be detected by the discourse created in the

course that a dynamic of power produces a set of policy outputs because a variety of

intents as regards the minimum wage are included in and excluded from the policy

outputs through debates and decisions in the process.

The implicit purposes and characteristics of the minimum wage shape its actual effects

because they provide a concrete direction toward which minimum wage policy works.

Unlike the legislative purpose of the minimum wage, the implicit one caught by the

discourse through its policy process is likely to be produced bearing more on a specific

issue addressed with more urgent and controversial concern at a particular point of time.

125

For instance, if a subminimum-related issue is discussed in the light of concern about

the employment level of a certain group of workers which is largely composed of

minimum wage workers, a sub-rate for the group may be explicitly introduced, or a

minimum wage rate for all could be newly fixed, taking into account the wage level that

would not make an adverse employment effect, particularly for the target group. In this

case, the minimum wage is intended as an employment policy rather than a wage policy,

at least for the group of workers concerned, and this becomes an explanatory basis of

minimum wage effects for the group, such as no negative employment effects but

insignificant increases in wages and in family income. Furthermore, when the implicit

purposes and characteristics of a minimum wage include explicit concern about a

particular group of workers, they can be reinforced by cultural prejudice against the

members of the group and their internalisation of it, which drives the intended purpose

more explicitly presented as an effect of the minimum wage. The aim of no employment

effect for a certain age group of workers, such as the young or older workers, would be

an example. Those workers who are typically disadvantaged in the labour market might

actively accept a lower wage than the minimum wage and/or oppose an increase in the

minimum wage in order to stay in employment. In the same context, they could favour

an age-differentiated subminimum. Such attitudes and responses to the prejudice and

disadvantaged practice in the labour market are likely to strengthen a tacit purpose of

the policy to stress the employment level for the targeted age group, regardless of the

general legislative purpose of the policy to improve wage and income levels for the

vulnerable. This would result in neither adverse employment effects nor increases in

wage and income for the group concerned.

A series of changes in what are involved in policy change of the minimum wage

discussed so far can come about at any time and/or any stage of the policy process in

principle. But, the tendency of government to stick to its own policy orientation under

neither radical change in macroeconomic conditions nor a powerful mobilisation of

public opinion against the policy orientation implies two points. One is that minimum

wage policy tends to be ‘punctuated by rare episodes of radical change’ (Pollitt and

Bouckaert, 2009, p.168) in practice, which assumes ‘a sharp distinction between long

periods of institutional stasis periodically interrupted by some sort of exogenous shock’

126

(Streeck and Thelen, 2005, p.1). The other is that government change would be the most

regularly observable shock to minimum wage policy. As long as there is no

socioeconomic shock which influences minimum wage policy, government policy

orientation becomes the key to determine a particular direction of government actions

and to shift an established equilibrium of power, which results in a new dynamic of

power among the key stakeholders and leads to the production of distinctive implicit

characteristics of the policy.

In sum, the theoretical framework proposed here argues that the implicit characteristics

of minimum wage policy are produced by a dynamic of power, which results in forming

an equilibrium of power among the key stakeholders at a particular time and stage of the

policy process, and the characteristics, which can be detected by the discourse

constructed through the policy process, shape the effects of the minimum wage.

Furthermore, minimum wage policy is likely to be punctuated mainly by government

change unless a shock in macroeconomic conditions and/or radical mobilisation of

public opinion occurs.

Research Methods

The theoretical discussion in the first part of this chapter has two methodological

implications for this research on the effects of minimum wages. First of all, the main

theoretical point is that the effects of minimum wages are not only produced by

economic mechanisms in the labour market but are also shaped by political

consequences in the policy process. This requires a methodology which allows an

understanding of the effects of minimum wages based on both the economic

mechanisms and the political consequences. However, the way each is involved in the

effects of minimum wages displays an epistemological distinction, which leads to the

mixed use of different types of data and research methods. That is, the economic

mechanisms in the labour market assume an objective reality which can be repeatedly

embodied by general rules regulating the market. Relying on this assumption, the

majority of studies on the effects of minimum wages have taken the stance that an

objective reality can be imperfectly and probabilistically explained by quantifiable data

127

and statistical methods. In contrast, the political consequences in the policy process

presume realities that are socially constructed and driven by the context of power

relations governing the process. This presumption suggests that realities can be detected

by a qualitative analysis of text materials and spoken and written discourse data.

The other main point of the theoretical discussions is that minimum wage policy would

rarely have a drastic change without a socioeconomic shock, and government change

would be the most regularly observable episode of radical change to the policy. This

suggests that the effects of minimum wages could be distinctive according to the term of

office of a particular government, and thus designing the research on the effects of

minimum wages by government would give a better sense of the effects. This

suggestion may allow us to explore whether different levels of increases in the

minimum wage have different effects, if the rate of increase in the minimum wage tends

to be kept at a similar level throughout the term of office of a government and is likely

to alter significantly as the policy orientation for the minimum wage shifts with a

change of government. Besides, exploring the question by government tenure and

comparing the results may enable us to check the practical relevance of monopsony

explanation, currently the most prominent account particularly in terms of non-negative

employment effects, and to go over the necessity of seeking a better understanding of

the effects based on the analysis on political consequences in the policy process.

Reflecting the two methodological implications from the theoretical discussion, this

thesis used both quantitative and qualitative methods to examine the effects of the

minimum wage under different governments and political consequences in its policy

process which were expected to shape the effects. Two research questions, how the

minimum wage affects the distributions of wages and family earnings, employment and

exits from in-work poverty among older workers and whether different rates of increases

in the minimum wage have different effects among older workers were examined

through a series of quantitative analyses using a set of secondary data and statistical

methods. The two questions were explored at the same time since two five-year periods

of two consecutive governments which had a big gap in the rate of increase in the

minimum wage were compared in each analysis of each subtopic. Theses quantitative

analyses were performed within the general assumptions of economic mechanisms in

128

the labour market. Meanwhile, the third research question, how the empirical results of

the minimum wage effects on older workers can be understood was explored through a

contextual analysis of texts and discourses produced particularly in the minimum wage

fixing process, after a comprehensive understanding of the empirical results from the

quantitative analyses based on economic explanations of the effects of minimum wages.

This qualitative analysis was conducted on the basis of the alternative political economy

framework for the research on the effects of minimum wages proposed earlier in this

chapter. Thus, the method chosen for the analysis followed the logical reasoning of

building the theoretical framework. The reason why the contextual analysis focused on

the minimum wage fixing process rather than embracing its whole policy process was

because this research assumed that comprehensive debates and decisions would be

made mostly in the fixing process and likely to regulate the whole of the minimum

wage policy process.

This research conducted a case study of South Korea on the effects of increases in the

national minimum wage on older workers, comparing two five-year periods of

consecutive Korean governments, ‘the Participatory Government’ under Roh, Moo-

Hyun (2003-2008) and the Lee, Myung-Bak government (2008-2013). Below are

described the specific data and methods used in each of quantitative and qualitative

analyses for this case study. This research did not take a quasi-experimental design in

the quantitative analyses, in which the causal impact of a minimum wage on target

population would be typically estimated by comparing before and after its introduction

or treatment and comparison groups when it increased in a region or a sector. This was

on account of the developmental and institutional peculiarities of the Korean national

minimum wage. As informed in the first section of Chapter Six, the national minimum

wage was introduced in the late 1980s in Korea but became widely known after it began

to be applied to all companies or workplaces with one employee or more in 2000 and as

the number of non-regular workers sharply increased in the mid-2000s influenced by

the International Monetary Fund (IMF)’ bailout programme with the 1998 Korean

financial crisis. This fact implies that the issue of noncompliance with the minimum

wage would be significant at least until the early 2000s. Further, since the Korean

national minimum wage has been implemented all over the country under the same

129

conditions since 2000, it is difficult to set up a comparison group for examining the

effects of the minimum wage. Internal validity of the inferential statistical analyses in

this case study was sought by using panel data which contained observations collected

over multiple time periods for the same individuals and also by allowing for regional

difference in the influence of an increase in the minimum wage. In line with the

consideration of those developmental and institutional peculiarities of the Korean

national minimum wage, the two Korean governments were chosen for this research

because the Roh, Moo-Hyun government maintained one of the highest rates of

increases in the minimum wage whereas the Lee, Myung-Bak government had the

lowest rate of increase, and they had contrasting political orientations. The big

difference in the rate of increase in the minimum wage and in political orientation

between the two governments was an ideal condition to embody the assumption of

government change as a radical impact on the minimum wage, and also to demonstrate

the methodology which this research has argued provides a better understanding of the

effects of minimum wages.

Quantitative Analyses

Using the Korean Labour and Income Panel Study (KLIPS), four subtopics - the effects

of the minimum wage on the distribution of wages, on employment, on the distribution

of family earnings, and on exit from working poor, among older employees aged 55 and

older - were examined for each five-year period of the two chosen Korean governments.

Considering the fixing point of a minimum wage rate by each government, its

application period, and the data collection time point of each wave (see Appendix 1),

wave 7 in 2004 to wave 16 in 2013 of the KLIPS were mainly included. However, some

variables required were derived from wave 6 in 2003 and wave 17 in 2014. For

example, family income for 2013 can be derived from wave 17 in 2014 because each

wave of the KLIPS provides household income for the previous year, not for the

concerned year; and also, the fraction of older employees affected by a minimum wage

rate newly applied in 2004 should be calculated with wave 6 in 2003. Except for the

kernel density estimates for descriptive analyses of the distributions of wages and of

family earnings, regression-based statistical methods for panel data analysis, such as

130

ordinary least squares (OLS) models and fixed-effects models, and the multilevel

discrete time event history model for competing risks were used as per the purpose of

each subtopic. The target age chosen for this research, as indicated in Chapter Six, was

grounded on the Korean context of the labour market that the actual age at which

workers retired from their lifetime main jobs has long been around 55 years.

Data

The KLIPS is an annual longitudinal panel study in South Korea which surveys the

economic and labour market activities of households and individuals residing in urban

areas across the country. It began in 1998 and is conducted annually by the Korea

Labour Institute (KLI) employing face-to-face interviews with the same set of questions

to the same set of households and their members. The sample was selected through a

two-stage stratified clustering method in which 951 areas were initially chosen from a

total of 21,675 sampling unit areas of the 1995 Korea Census, and then 5-6 households

were randomly opted for from each area. The original sample of 1998 included 5,000

households and 13,321 individuals therein to represent the adult population aged 15 and

over residing in urban areas and was complemented in the twelfth wave in 2009 by

adding 1,415 households and their members to redeem a significant number of dropouts

from the panel and to extend the sampling unit area to Jeju Island excluded from the

original sampling. The KLIPS consists of three datasets, household, individual, and

work history. The household dataset covers demographics, changes in household

members, family relations and financial resource exchanges between generations, types

of accommodation, children’s education and childcare, household income and

consumption, assets and debts, financial status, and burdensome spending. The

individual dataset includes the state of economic activity, earnings, working hours,

income and consumption, education and vocational training, and job-seeking activities.

The work history dataset contains information on individuals’ mobility in the labour

market.

The KLIPS is the best available dataset for this study about the effects of the minimum

wage on older workers in Korea. Firstly, it includes a reasonable panel size of older

employees aged 55 and over which represents the corresponding population in urban

131

areas, which means that it is feasible to conduct an inferential statistical analysis

focusing on older employees. Secondly, since the survey provides employees’ monthly

earnings and weekly working hours, hourly wage rates directly comparable to minimum

wage rates can be derived. Along with this, the KLIPS allows family earnings and

income identified per individual unit through merging the household dataset and the

individual dataset, and it facilitates an analysis of family income-related effects of the

minimum wage, including the effects on family earnings and on exit from working poor.

More importantly, the KLIPS enables the contextual limitation that a comparison group

which is not affected by an increase in the minimum wage cannot be set up to be

redeemed by providing the data from the same individuals obtained over time and also

by allowing the regional difference in influence of the minimum wage to be employed

in the analysis. Furthermore, since the KLIPS has all the strengths of the pre-existing

time-series data and cross-sectional data, and allows event history data with repeated

events and multiple types of event simultaneously to be constructed, the analysis of the

effects of an increase in the minimum wage on transitions in individuals’ status of

employment and/or of poverty, is rendered feasible.

Methods

The first concern with regard to the effects of the minimum wage on older workers was

its distributional effects on wages. The effects were examined by testing the two typical

predictions of economic theories on wages, spike and spillovers. As presented earlier in

this chapter, if a minimum wage is strongly enforced and complied with, it will raise the

pay of those who were previously paid below that rate, which truncates or thins out the

lower tail of the distribution of wages and creates a spike at the point of the minimum

wage (Neumark and Wascher, 2008; Manning, 2012). An increase in the minimum wage

may also raise the wages of those who were already paid slightly above the new

minimum rate, mainly due to substitution of workers with higher skill for those with the

lowest skill and wage differentials between workers with different levels of skill

(Neumark and Wascher, 2008). Following Alaniz, Gindling, and Terrell (2011), spikes at

the point of the minimum wage were assessed graphically per five-year period where

the rates each of the two concerned governments fixed were applied, by plotting the

kernel density estimates of log hourly wages minus log minimum wage for each

132

employee aged 55 and older in each year, where a zero indicated that the worker earned

the minimum wage rate of the year.

Then, how increases in the minimum wage caused changes in wages of employees aged

55 and older and relieved the wage gap among them were tested for each of the two

chosen five-year periods, using OLS and fixed-effects models. Following Jeong, J-H

(2011), an increase in the minimum wage was defined as two variables which were

shortened henceforth as the fraction affected and the fraction newly affected. The

fraction affected refers to the proportion of employees aged 55 and older who earned

less than a new minimum wage in the previous year to all employees of the same age

group, whereas the fraction newly affected refers to the proportion of employees aged

55 and older who were paid between an old and a new minimum wage in the previous

year to all employees of the same age group. The latter is more typical and therefore

suitable to stand for an increase in the minimum wage because the fraction represents

those directly affected by the increase. However, since a considerable proportion of

employees earned less than the minimum wage not only among mature workers aged 55

and over but also in the labour force as a whole in Korea, the former was also

incorporated as an indicator representing an increase in the minimum wage. Further,

both indicators of an increase in the minimum wage were computed by region

consisting of 7 metropolitan cities, 8 provinces, and one special self-governing

province.5 This was to redeem the limitation that a comparison group would hardly be

set up due to the institutional and developmental peculiarity of the Korean national

minimum wage, based on the assumption that regional difference in the fraction of

workers affected by an increase in the minimum wage would reflect the regional wage

differentials. The dependent variables were the changes in the 10th, 25th, 50th, 75th, and

90th percentiles of log hourly wages among employees aged 55 and older for each

region and the changes in ratios between the 90th and 10th, the 90th and 25th, the 90th and

50th, and the 50th and 10th percentiles of log hourly wages among employees aged 55

and older for each region. In the statistical models, the changes in the employment rate

of older employees by region were included as a control variable. This was defined as

the changes in the proportion of employees aged 55 and older who were neither self-

employed nor engaged in unpaid work to all individuals of the same age group for each

133

region. This variable was under the assumptions that regional difference in the rate

would reflect the different labour market conditions across regions and that the rate was

arguably the better labour market measure than the unemployment rate of older workers

in as much as those who had motivation to work but were not employed and those who

were completely left the labour market were hardly distinguished in practice for this age

group.

The employment effect of the minimum wage for older workers was explored on two

different levels. Using OLS and the fixed-effects models, the effect was first screened at

the regional level for each of the two five-year periods by regressing each of the

changes in the employment rate of older employees by region and the changes in the

unemployment rate of older individuals by region on both of the fraction affected and

the fraction newly affected, the same indicators of an increase in the minimum wage

used in the analyses of the effects on the distribution of wages. The changes in the

employment rate of older employees by region was also defined as the same with the

control variable in the analysis of the distributional effects on wages, and the change in

the unemployment rate by region referred to the changes in the proportion of those aged

55 and older who were neither employed, self-employed, nor engaged in unpaid work to

all individuals of the same age group for each region. For a valid causal inference,

regional differences in factors which would be involved in regional employment and

unemployment rates should be controlled for. For example, sectoral composition by

region may reflect different rates of growth in production, different elasticity of

employment with respect to output, different capital intensity, and different patterns of

technological change by regional economy which all will influence employment at the

regional level (Eurostat, 2014). However, since the number of individuals aged 55 and

over in the panel data used for this research was relatively few when taking into account

the number of industrial sectors classified in the data, it was questioned whether the

sectoral composition for the age group by region would have sufficient distinction to

reflect the regional economy, and no control variable was included in the models. In this

regard, the statistical tests for employment effects at the regional level aimed at getting

informed of the consistent correlation between an increase in the minimum wage and

the changes in employment or unemployment rates for older individuals, rather than at

134

determining the causal relationship between them.

The causal relationship between the increase in the minimum wage and older workers’

status in employment was explored at the individual level. Given the assumption that

older employees were more likely to be laid off in connection with labour market

interruptions and less likely to be re-employed once they were unemployed, whether

they would be discharged in practice as the minimum wage increases was the primary

concern for the group. In this regard, how an increase in the minimum wage affected the

transition from being employed among individuals aged 55 and older was tested for

each of the two five-year periods, using a multilevel discrete-time event history model

for competing risks. The model allows event history data with discrete-time nature and

complex structure to be handled (Steele, Goldstein, and Browne, 2004; Steele, 2011)

and requires a series of data preparations. The longitudinal record of individuals’

employment status collected every year in the KLIPS datasets was first converted into

event history data with an episode defined as a continuous period of being employed.

An episode began if an individual started to be employed and ended if the individual

moved into unemployment or other types of employment, such as self-employment and

non-wage family business work. The timing at which an episode ended was represented

by a form of time interval in which a transition occurred, and the exact timing of a

transition within a time interval was not important in a given case because the main

concern was to analyse whether or not an employee remained being employed along

with the annual increase in the minimum wage, based on the length of time spent in

being employed, not to study the timing of transition. This leads to a discrete-time

model which allowed us to analyse the duration of episodes (Steele, 2011). Discrete-

time models have flexibility to handle a complex structure including a hierarchical

structure and multiple possible ways in which an episode may end and also to facilitate

straightforward inclusion of time-varying covariates (Steele, Goldstein, and Browne,

2004). Transitions from being employed can be found several times for an individual

over the observation period, and the two types of transitions into unemployment and

other types of employment become competing risks at which an individual is of

experiencing a transition from being employed. These require a multilevel multinomial

model dealing with recurrent events within an individual and two competing risks, and

135

the model can be embodied within a discrete-time model by some initial data

restructuring.

The event history data of employment transition were restructured in discrete-time

format with one record per one-year time interval spent in an employment status, using

SPSS. These data were essentially in a long format where fewer cases contributed a

greater number of observations due to the year concerned. In this data, the case

indicated by an individual’s identifier was regarded as level 2, and the variable, year as

level 1. To construct the event variable which showed an individual’s transition in

employment status for each year, the original variable for employment status was

recoded as 0 for still being employed, 1 for becoming unemployed, and 2 for becoming

self-employed or a non-wage family business worker. The variable, duration which

indicated the duration of an episode defined was created, indexed as 1, 2, 3, … from the

starting year of being employed to the ending year with a transition to being

unemployed or in other types of employment, and it was reset to 1 at the start of second

time being employed. Since being employed was the only starting point in employment

status, the record of the following year(s) was not included as an observation if an

individual stayed unemployed or in other types of employment after a transition to one

of them. The variable, exposure which indicated the number of years of a one-year

interval for which an individual was at risk of a transition, was also created. In the

analysis, the exposure variable is typically used for giving observations weighted (The

University of Bristol, 2006). However, since the time interval used as the denominator

for weighting was one for this analysis, the value of the multinomial response weighted

was the same as the one before being weighted.

The discrete-time data procured through the first data restructuring presented above

were extended to have a set of binary responses for each multinomial response. This

construction was carried out automatically when a multinomial model was specified in

MLwiN (The University of Bristol, 2006).6 The event variable in the discrete-time data

was converted into four variables in the extended data, response (resp), response_index

(resp-ind), cons.unemployment (con.une), and cons.self-employment (cons.self). The

response_index variable indexed the binary responses as 1 for the unemployment

indicator and 2 for self-employment or non-wage family business work indicator. This

136

variable became the level 1 identifier when defining the multilevel structure, and the

‘real’ level 1 (time unit) was shifted up to level 2, with individual at level 3. When

response_index was 1, cons.unemployment was 1 and cons.self-employment was 0.

When response_index was 2, cons.unemployment was 0 and cons.self-employment was

1. The response variable was indexed as 1 only when the value of the event variable in

the discrete-time data was consistent with the value of the response_index variable in a

given year, and otherwise, the response variable was indexed as 0. The individual_long

variable and the atrisk_long variable were ‘long’ versions of individual_identifier and

exposure, respectively, in the discrete-time data. The duration.unemployment (dur.une)

variable was cons.une * a ‘long’ version of the duration variable. The con.* and dur.*

variables were created when cons and dur were added to the model as explanatory

variables. The coefficients of cons.une and cons.self became the intercepts in the

contrasts for unemployment and self-employment/non-wage family business work

versus remaining in employment while the coefficients of dur.une and dur.self were the

duration effects on each contrast. An example of the data expansion for episodes for two

individuals is as follows:

Discrete-time data

Individual Dur Event Atrisk 1 1 0 1 1 2 0 1 1 3 2 1 2 1 1 1

Extended data by MLwiN

Individual_long Resp_ind Resp Atrisk_long Cons.une Cons.self Dur.une Dur.self 1 1 0 1 1 0 1 0 1 2 0 1 0 1 0 1 1 1 0 1 1 0 2 0 1 2 0 1 0 1 0 2 1 1 0 1 1 0 3 0 1 2 1 1 0 1 0 3 2 1 1 1 1 0 1 0 2 2 0 1 0 1 0 1

137

The main explanatory variables included in the model which was set up through the

data preparation were duration spent in being employed and square of duration for the

effect of duration and the combination variable, the minimum wage application or not *

log hourly real minimum wage for the effect of the minimum wage. The duration

variable was centred by group defined by the variable, year in the model, which meant

that the annual mean of duration was subtracted from the corresponding individual

scores, and thus, after being centred, the value, zero indicated the mean of the centred

variable, duration, for each year (Hox, 2002). Centring duration was to avoid

multicollinearity issues possibly generated by including the transformed variable,

square of duration, in the model, which could affect model convergence problems

and/or inflated standard errors (Goldstein, 2015).7 The advantage of centring on the

variable, year is that the effect of duration can be interpreted under the control of the

year effect. Meanwhile, using square of duration for the effect of duration meant that

the quadratic relation between the duration spent in being employed and the transition to

unemployment or self-employment/non-wage family business work was tested,

assuming that the occurrence of a transition would increase or decrease with the

duration spent in being employed but at some point it could go the opposite way. This

assumption took into account older workers’ characteristics of the relatively high

proportion of temporary jobs and the plausibility of being in a critical transition period

from a major lifelong career. For the effect of the minimum wage, using the

combination variable, minimum wage application or not * log hourly real minimum

wage was to capture both the individual-varying feature and the time-varying one

within an individual, simultaneously. Minimum wage application or not in the

combination variable was indexed as 1 for the individual affected by a minimum wage

rate and 0 for those not affected, based on an employee’s hourly wage which was

derived from monthly pay and average weekly working hours one year earlier than each

concerned year. Some missing data of the derived hourly wage, which occurred due to

the absence of information about monthly pay and weekly working hours, were imputed

by expectation-maximization (EM) technique in SPSS for all employees each year,

considering the relationship with other variables, such as sex, age, the level of education

obtained, and the types of employment.8 In addition to the main explanatory variables,

sex, age, and the level of education obtained were included as additional controls in the

138

model. The education obtained variable was categorised as middle school and lower

and high school and higher since the proportion of those with higher education was very

small. All variables included in the model were time-varying, except for sex.

This multilevel discrete-time event history model for competing risks constructed for

the analysis of the minimum wage effects on employment among older employees aged

55 and older was examined in the fixed-effects model estimated by the 1st order

marginal quasi-likelihood (MQL) procedure and the random-effects model estimated by

Monte Carlo Malkov (MCMC) method, using MLwiN. Since MQL produced the most

rough approximation, leading to estimates biased downward, and predictive quasi-

likelihood (PQL) raised convergence problems in MLwiN due to its instability

(Rasbash, Steele, Browne, and Goldstein, 2016), whereas MCMC method produced

new estimates for unknown parameters from the last iteration by repeating the process

of combining a prior distribution into a posterior distribution for many times (Browne,

2016), the simulation-based estimation procedure was used as an alternative to

likelihood-based estimation procedures in the random-effects model which allowed for

unobserved heterogeneity at the individual level by enabling the coefficients of constant

variables for the two binary responses to vary randomly across older employees.

The third concern, the effect of the minimum wage on older workers’ family income

was explored by questioning how an increase in the minimum wage affected the

distribution of family earnings. Since the minimum wage can influence income from

labour only, family earnings were regarded as a relevant outcome variable for the

analysis. The distribution of family earnings of employees aged 55 and older was first

reviewed for each of the two five-year periods by plotting the kernel density estimates.

The estimates were calculated per year by subtracting the logarithm of 60 per cent of the

median annual family earnings per capita for all employees’ households from the

logarithm of annual family earnings per capita for each older employee’ household. The

zero in the density plot indicates that the older individual’s family gains 60 per cent of

the median annual household earnings for all employees’ households. Typically, total

earnings that all employed family members earn in a household is used for the study of

the distribution of household earnings. However, as the focal point of this analysis was

the changes in family earnings of older employees, the annual family earnings per capita

139

for each older employee were computed by dividing the total annual family earnings in

each individual’s household by the square of household size, an equivalence scale for

income estimates.9 Then the effect of the minimum wage on the distribution of family

earnings of older employees was tested for each of the two five-year periods, using OLS

and fixed-effects models. In the models, the same measures of an increase in the

minimum wage and of the changes in the employment rate of older employees by region

used in the analyses of the distributional effect on wages were employed as independent

variables and a control variable, respectively. The dependent variables were the changes

in the 10th, 50th, and 90th percentiles of log annual family earnings per capita among

employees aged 55 and over for each region and the changes in ratios between the 90th

and 10th and the 50th and 10th percentiles of log annual family earnings per capita

among employees aged 55 and older for each region.

The last concern as regards the effects of the minimum wage on older workers was

whether an increase in the minimum wage helped an older employee to exit from

poverty. This was examined for each of the two five-year concerned periods, using a

multilevel discrete-time event history model for competing risks, as in the analysis of

the minimum wage effect on older employees’ employment. Combining the individual

dataset and the household dataset of the KLIPS, event history data with an episode

defined as continuous period of being the working poor was constructed among

employees aged 55 and older. The working poor were defined as employees whose

household income was less than 60 per cent of the median annual household income for

all individuals surveyed.10 Note that household income rather than family earnings was

used for this analysis on the assumption that earned income would be the most elastic to

change a household’s finance among low-income families. Household income was

computed on a per-capita basis for figuring out an employee’s poverty status by

dividing the total annual amount of earned income, financial income, income from

estates, income from social insurance, transfer income, and other income in an

individual’s household by the square root of household size, an equivalence scale for

income estimates as in the calculation of family earnings per capita in the analysis of the

distributional effect on family earnings. Some missing data of the derived household

income per capita were imputed by expectation-maximization (EM) technique in SPSS

140

for all respondents each year, considering the relationship with other variables, such as

sex, age, the level of education obtained, marital status, householder or not, and the

number of family members, after the check for whether the missing data were small

enough and at random. An episode started if an individual began to be in in-work

poverty and ended if the individual was out of poverty while remaining employed,

became unemployed, or moved into other types of employment, such as self-

employment and non-wage family business work. Accordingly, there were three types

of competing risks at which an individual was of experiencing a transition from being

working poor, and three binary responses for the multinomial response were generated

in the extended discrete-time data in MLwiN.

The main explanatory variables were duration spent in being working poor and square

of duration for the effect of duration and the combination variable, the minimum wage

application or not * log hourly real minimum wage for the effect of the minimum wage.

By the same token in the analysis of the effect on older employees’ employment, the

duration variable was centred by groups defined by the variable, year in the model, and

the quadratic relation between duration spent in being working poor and the transition to

other states was tested with square of duration. Sex, age, the level of education

obtained, householder or not, and the number of family members were included as

control variables in the model.11 All variables were time-varying, except for sex. As in

the analysis of the employment effect at the individual level, the event history model for

the effects on older employees’ exit from being poor was examined in the fixed-effects

model using MQL estimation method and the random-effects model using the MCMC

estimation method.

Qualitative Analysis

Three components of the alternative political economy framework proposed earlier for

this research on minimum wage effects - government political orientation as regards

policies, the dynamic of power, and the implicit characteristics of the minimum wage -

were compared through a contextual analysis for the two chosen five-year periods in

order to better understand the empirical results from the quantitative analyses. The

141

method was chosen based on the logical reasoning of building the theoretical

framework. That is, the significant dissonance between empirical results and theoretical

predictions of the existing economic theories addresses an explanatory vacuum. This

evokes the fact that the minimum wage is fundamentally a public policy which is

framed and implemented through political policy processes and has multiple attributes

of labour and social policy as well as wage and economic policy. Although quantitative

methods are prevalent in public policy research as well, the main political elements

involved in the effect of minimum wages, such as government political orientation as

regards policies, the dynamic of power, and the implicit characteristics of the minimum

wage, cannot be quantifiable and analysed in a statistical model because they engage in

processes, contexts, and dominant political and social values.

The context and specifics of a case were assumed to be reflected in discourse which was

formed in the minimum wage policy process allowing for related political and economic

events occurred in a concerned period. In the same vein as Fischer (2003, p.45, quoted

in Hill, 2009, pp.80-81), a set of relationships between political factors concerning

minimum wage effects ‘can be constructed and discussed only through language’, and

as Hall (1993, p.289, quoted in Hill, 2009, p.76) argues, ‘the terms of political discourse

generally have a specific configuration that lends representative legitimacy to some

social interests more than others, delineate the accepted boundaries of state action …

and defines the context in which many issues will be understood’. This recognition of

the significance of discourse led to the use of an analysis of texts generated, particularly

in the policy process. However, the stress on the significance of discourse here does not

mean the negation of ‘the importance of the empirical testing of theories and hypotheses

(Hill, 2009, p.12)’ in the minimum wage research. Rather, as Hill (2009, p.12) argues, it

is ‘allowing for the possibility of alternative interpretations of evidence’.

Data and Material

The National Minimum Wage for the Year of OOOO: The Details of Deliberation and

Decision [OOOO년도 적용 최저임금 심의·의결경위] was chiefly employed for the

contextual analysis. It is the Korean Minimum Wage Council’s annual report on the

minimum wage fixing process. It has been published since 1988 by the Minimum Wage

142

Council which was established under the Ministry of (Employment) and Labour [고용노

동부] by Article 12 of the Minimum Wage Act [최저임금법] and is responsible for

deliberation on the level of the minimum wage and other related issues. Although there

are slight differences in the structure of contents every year, the report mostly consists

of an overview of the Minimum Wage Commissioners, the functions of plenary sessions

and sub-committees and the procedures of deliberation and decision, records of each

plenary session and meeting held by each sub-committee, the process of negotiation on

the rate of increase in the minimum wage, Employer Commissioners’ request proposal,

Employee Commissioners’ request proposal, a report on the Commissioners’ business

field investigation, results of annual evaluation survey of minimum wage effects, and

records of workshops held out of the fixing period. It provides detailed information on

the composition of and changes in the Minimum Wage Commission members, the

debate and events occurred in each plenary session and a meeting held by a sub-

committee during the fixing period, interactions among labour, business and public

interest representative members with regard to a specific issue addressed in the

Commission, and the whole process that a decision was made and an issue was shelved

for further discussion in the next fixing period. The report also includes, albeit in limited

form, interviews among employers and employees which were obtained from the

Commissioners’ annual field investigation. Since the report informs of all of the issues

addressed and discussed in the Commission, demands, opinions and behaviours

presented by each Commissioners group, and related events held in and out of the

Commissions, it provides a suitable textual data to look at the dynamic of power among

the Commissioner groups in the minimum wage fixing process and the implicit

characteristics of the minimum wage shaped by the dynamic, which were considered

earlier in the newly proposed political economy framework as critical factors to explain

the effects of the minimum wage. Despite the merit of the report as data, this was the

first attempt to use it for studying the effects of the minimum wage in Korea.

In addition to the Minimum Wage Council’s annual report, academic literature on

economic and labour policies and on the relationships between government, business,

and labour, government reports from the National Economic Advisory Council [국민경

제자문위원회], the Ministry of Strategy and Finance [기획재정부], and the Ministry of

143

(Employment) and Labour [(고용)노동부], reports from the Confederation of Korean

Industries [한국경영자총협회], the Confederation of Korean Government Employees’

Unions [대한민국공무원노동조합총연맹], and trade unions, and newspaper articles

were used, particularly to look at each government’s political orientation as regards

policy, which could not be covered by the Minimum Wage Council’s annual report, and

ultimately for constructing methodological triangulation of macro, meso, and micro

levels of analysis, which was embedded in the alternative political economy framework

proposed earlier.

Methods

Academic literature, reports from a variety of government departments and employers’

and employees’ organisations, and newspaper articles were comprehensively reviewed,

and the Minimum Wage Council’s annual reports were analysed by hand as below.

Because it was needed to figure out the flow of contents in several layers as presented

below, analysing it by hand was deemed more effective than using a data analysis

programme which basically relied on the categorisation of themes and relations between

them. Themes that emerged from the reports were first identified as the composition of

human resource among the public interest members of the Minimum Wage

Commission, roles and responsibilities of plenary sessions and sub-committees in the

Minimum Wage Council and Commission, minimum wage fixing criteria, issues

addressed and discussed by the representative members of the Commission, each

representative group’s involvement in the fixing process, and responses of frontline

government, employers and employees which were from field surveys during each

fixing period attached to the reports. Information from the data was rearranged under

each theme for each five-year period. In this rearrangement, the theme of minimum

wage fixing criteria was divided into debates on criteria and criteria practically used in

the fixing process, and issues addressed and discussed in the Commission were

reviewed by topic and graded importance on according to frequency of appearance and

intensity of discussion. Debates on fixing criteria and issues raised in the Commission

were also looked at by sub-committee in the Minimum Wage Council and Commission

in order to identify the changes in roles and responsibilities of plenary sessions and sub-

committees. Except for the composition of human resources among public interest

144

members in the Commission, all of the information presented under each theme and

sub-theme were revealed by each representative group.

Face-to-face interviews which are commonly used as a useful data collection method in

a qualitative analysis were not employed in this study. The main reason was that the

report provided the full descriptions of the relevant processes, which allowed

researchers to read the contexts of the rate fixing process and any continuity and

disparity between the rates of increase. A more practical reason was time and resource

restraints in completing this PhD research. As the last two waves of the KLIPS was

released in January, 2016, almost one year behind schedule, and the multilevel discrete-

time event history model for competing risks required extremely time-consuming

preparatory work with the data, supplementary data collection by face-to-face

interviews in addition to the analysis of the Minimum Wage Council’s annual reports

covering ten years was not feasible. The number of target interviewees and accessibility

to them in terms of time, distance, expenses, and personal information were expected to

be critical barriers for preparing and conducting interviews because they were the

former representative members in the Minimum Wage Commission which consisted of

27 representatives, 9 for each group of employers, employees, and public interests, of

whom each member served a three-year term. The limitation of this research which can

be raised by the exclusion of face-to-face interviews are discussed in the Conclusion

Chapter.

Conclusion

This chapter reviewed and discussed various theoretical perspectives on the minimum

wage. Mainstream economics, particularly neoclassical theory and the monopsony

model, have long dominated this field of research. These have attempted to improve

their explanatory power in many ways, but inconsistency in empirical evidence,

especially regarding employment and the exit from poverty, does not sustain their

predictions about what the effects of minimum wages would be like. In more recent

years, the Keynesian model has developed as an alternative account of the minimum

wage effects, claiming a macroeconomic approach in the theoretical debate. However, it

145

is hard to explain spillover effects in the distribution of wages and their influence on

wage dispersion within the Keynesian explanation. Setting the weakness of each branch

of economic theories aside, the first and foremost fact is that all the economic

explanations overlook the political attributes of minimum wages. Some political

economy explanations relating to the minimum wage can be found, but they focus on

what political factors as well as economic ones make people support or oppose the

minimum wage, rather than how the political attributes of minimum wages are involved

in shaping its effects on workers. Hence, an alternative political economy framework for

the research of the minimum wage effects was proposed. The framework argues that the

implicit characteristics of minimum wage policy are produced by a dynamic of power,

which results in forming an equilibrium of power among stakeholders at a particular

time and stage of the policy process, and the characteristics, which can be detected by

the discourse constructed through the policy process, shape the effects of the minimum

wage. Furthermore, minimum wage policy is likely to be punctuated mainly by

government change unless a shock in macroeconomic conditions and/or radical

mobilisation of public opinions occur. Following the logical reasoning of building the

theoretical framework, this thesis conducted both quantitative analyses using statistical

methods and contextual analysis with the review of the relevant literature and

documents.

146

Notes

1 This section incorporates not only theoretical models which strictly bear the fundamental elements of the neoclassical paradigm for the labour market economics but also ones that significantly modify the key elements and are named differently, such as monopsony models, search models, and efficiency wage models. 2 Pettengill, John S. (1981). The Long-run impact of a minimum wage on employment and the wage structure. In: MINIMUM WAGE STUDY COMMISION. Report of the Minimum Wage Study Commission, Vol. 6, 63-104. Washington DC: U.S. Government Printing Office. 3 It was originally written as a discussion paper at Northwestern University in 1989, published in 1998, and was presented at the LSE in 1990. 4 As Hayes summaries in his article, Sobel (1999) shows this by advancing a pressure group model of political influence based on Gary Becker’s earlier work (1983, A Theory of Competition Among Pressure Groups for Political Influence. The Quarterly Journal of Economics, 98(3), pp.371-400). In the model, the political strength of unions was measured by union membership as a share of non-agricultural employment (a), the political power of business interests, by the top marginal corporate income tax rate (b), and a relative union/business ratio was calculated by dividing (a) by 100 per cent minus (b). Sobel found that the relative union/business ratio was strongly correlated with the real value of the minimum wage over the entire historical period in the United States. 5 As Jeju Island, the special self-governing province was included as a sampling unit area into the KLIPS from 2009, the indicators of an increase in the minimum wage and the employment rate of older employees for the first five-year period were computed for 7 metropolitan cities and 8 provinces. 6 Macros which contains syntax to read ascii file into MLwiN were used to specify a multinomial response model, to create new variables for the model, and to set up an initial simple competing risks model, as presented in The University of Bristol’s (2006) training document for multilevel discrete-time event history analysis. An example of the macro used can be found in Appendix 2. Note that the reason why an initial simple model, rather than a full model for the analysis, was set up in the macro is that convergence problem often occurs when multiple covariates are included in a model at the same time. 7 In practice, the random-effects model for the second five-year period was not converged when the duration variable was uncentred or centred on its grand mean; and the model for the first five-year period was converged but produced highly inflated coefficients and standard errors with uncentred duration or centred one on its grand mean. 8 Adams & Neumark (2005) used imputed wages to identify a group affected by living wage and the other not affected in the analysis of employment effects. For our analysis, an imputation method which produces single values for cases with missing in order to rationally classify each case into a certain group is needed. EM imputation technique which imputes a value the most likely based on the values of other related variables is a plausible option when the rate of missing data is small, and the missing are completely at random (MCAR) or conditionally at random (MAR). The percentage of missing values of derived hourly wages for all employees in our datasets is below five per cent in each concerned year; missing data in some years are turned

147

out as MCAR by the Little’s MCAR test, and in other years there is no evidence that the probability of a missing value depends on the variable that is missing (NMAR). 9 A wide range of equivalence scales exist, and the square root scale has been used by recent OECD publications which compare income inequality and poverty across countries (OECD, 2013b). See OECD (2013d). OECD Framework for Statistics on the Distribution of Household Income, Comsumption and Wealth. OECD Publishing. for further details of household equivalence scales. 10 There are various ways to define and measure poverty depending on the purpose of research. The most commonly used approach is to compare each household’s income adjusted for family size to median income. OECD, EU and many other countries use 60 per cent of median household income as the poverty threshold. See OECD (2013e). The OECD Approach to Measure and Monitor Income Poverty across Countries. United Nations Economic Commission for Europe Conference of European Statisticians. Geneva. for further details of poverty thresholds. 11 The variable, marital status, were not included in the final model since other types other than ‘married’ rarely found among the cases included, which caused model convergence problem.

148

Chapter Six The Factual Context for the Case Study of South Korea

Introduction

This chapter reviews the national minimum wage system and the labour market status of

mature workers aged 50 and over in South Korea to provide background knowledge

before presenting the analysis results. It includes a brief history of the minimum wage

law, the current system, and the changes in its level and influence; and then, how

potentially Korean older workers can be affected by the minimum wage is assessed

through surveying their participation in the labour market, employment, unemployment,

types of employment, industry sectors and occupations they are involved in, their wage

levels and the level of income security guaranteed by their wage, and disadvantages

they face in the workforce. Then, empirical studies about the effects of the minimum

wage in Korea are also looked at in the last section, with a special interest in older

workers.

The National Minimum Wage

The statutory minimum wage in South Korea, with the purpose of ‘stabilising workers’

life and improving the quality of the labour force by guaranteeing a certain minimum

level of wages to workers, thereby contributing to the sound development of the

national economy (Minimum Wage Act Art.1)’ was enforced in 1988 under the

influence of the massive eruption of the labour union movement in the late 1980s (Yun,

A-L, 2014). The Labour Standards Act enacted in 1953 provided the legal grounds for

the statutory minimum wage, but it had not been implemented until the end of 1986

under the Korean government’s judgement at that time that the statutory minimum wage

was not acceptable to the national economy. The Minimum Wage Act was enacted in

December 1986 and came into effect in January 1987, covering only manufacturing

companies with ten employees and more. In 1988 the coverage of the minimum wage

rate was extended into the mining and construction sectors with 10 employees or more

149

as well as manufacturing. In the following year, the coverage was further extended to all

companies with ten employees or more; and after several amendments, it was finally

applied to all companies or workplaces with one employee or more in 2000. The

national minimum wage, however, became widely known to small and medium-sized

employers and employees as well as the public in practice as the number of non-regular

workers increased sharply in the mid-2000s influenced by the International Monetary

Fund (IMF)’s bailout programme following the 1998 Korean financial crisis. All

employees as defined in the Labour Standards Act, regardless of their employment

status or nationality, currently benefit from the statutory minimum wage, except

domestic workers, the seamen and those with a very limited working capacity due to

their physical and mental disabilities. But those engaged in an apprenticeship for less

than three months can be legally paid 10% less than the hourly rate of the minimum

wage. From 2007, the subminimum wage was introduced for those engaged in

surveillance or intermittent work who had been exempt from the minimum wage due to

their exclusion from the application of the Labour Standard Act under Article 63, Clause

3 of the Act.1

The minimum wage rate is fixed annually by the government based on the proposal the

Minimum Wage Commission [최저임금위원회] submits. The Commission consists of

27 representatives, 9 for each group of workers, employers, and public interests, and

should put forward a newly proposed rate to the government by the 29th of June every

year. The government should announce a new minimum wage rate by the 5th of August

in the same year, and a new rate takes effect from the 1st of January of the following

year (Figure 6.1).2 The Minimum Wage Commission is outwardly a tripartite

organisation with division of power among the three groups, but the public interest

members who are experts, mostly in economics or business administration, with

profound knowledge and experience, such as public officials, professors, and

researchers of an officially authorised research institute, actually take the lead in fixing

a new rate (Yun, A-L, 2014). Meanwhile, workers’ living costs, the wages of similar

workers, labour productivity, and income distribution ratio are stated in the Minimum

Wage Act as the minimum wage fixing criteria to be taken into account. However, in

many cases the main reasons for the low level of the minimum wage provided by the

150

government are economic conditions including the inflation rate and the capacity of

employers to pay especially in the case of small and medium-sized firms. This implies

that economic conditions are, in practice, more critical than the indicators stated in the

law.

Figure 6.1 Process of Minimum Wage Deliberation and Determination in South Korea

Source: The Minimum Wage Council (http://www.minimumwage.go.kr)

The current (2015) minimum wage rate is 5,580 KR won (£ 3.32 or $ 5.08 at the

exchange rate of February 2015). Compared to 462.50 KR won (£ 0.37 or $ 0.63 at the

then average yearly exchange rate) in 1988 when the minimum wage was introduced,

the wage rate has risen by twelve times during the past 27 years. The rate of increase

ranges from 4.9 per cent to 18.8 per cent excluding the three exceptionally low or high

rates in three years. The average rate is 9.7 per cent which is slightly higher than 8.3 per

cent, the median, owing to the exceptional cases (Table 6.1). Looking at the increases

sanctioned by the government, the average uplift rate was 13.8 per cent during the Roh,

Tae-Woo government (1988-1993) which represented the conservative ruling party but

ushered in a new ‘Democratic Era’ in Korea, 8.1 per cent during, so called, ‘Moon Min’

government, a civilian government in which Kim, Young-Sam took office (1993-1998),

9.0 per cent during the Kim, Dae-Jung government (Government of the People, 1998-

151

Table 6.1 Changes in Minimum Wage Rates in South Korea (KRW/GBP/USD; %; thousand persons)

Year

Hourly minimum wage rate* Increase

(%)

Number of workers covered

(A)

Number of beneficiary workers (B)

Influence rate** KRW

(₩) GBP(£) USD($)

2015 5,580 3.32 5.08 7.1 18,240 2,668 14.6 2014 5,210 3.00 4.95 7.2 17,734 2,565 14.5 2013 4,860 2.84 4.44 6.1 17,510 2,582 14.7 2012 4,580 2.57 4.06 6.0 17,048 2,343 13.7 2011 4,320 2.43 3.90 5.1 16,479 2,336 14.2 2010 4,110 2.30 3.55 2.75 16,103 2,566 15.9 2009 4,000 2.01 3.13 6.1 15,882 2,085 13.1 2008 3,770 1.87 3.42 8.3 15,351 2,214 13.8 2007 3,480 0.00 0.00 12.3 14,968 1,784 11.9

2005.09 ~ 2006.12 3,100 1.76 3.24 9.2 14,584 1,503 10.3

2004.09 ~ 2005.08 2,840 1.52 2.77 13.1 14,149 1,245 8.8

2003.09 ~ 2004.08 2,510 1.20 2.19 10.3 13,631 1,035 7.6

2002.09 ~ 2003.08 2,275 1.17 1.91 8.3 13,216 849 6.4

2001.09 ~ 2002.08 2,100 1.12 1.68 12.6 7,152 201 2.8

2000.09 ~ 2001.08 1,865 1.00 1.44 16.6 6,692 141 2.1

1999.09 ~ 2000.08 1,600 0.93 1.42 4.9 5,031 54 1.1

1998.09 ~ 1999.08 1,525 0.79 1.28 2.7 5,136 23 0.4

1997.09 ~ 1998.08 1,485 0.64 1.06 6.1 5,350 124 2.3

1996.09 ~ 1997.08 1,400 0.90 1.47 9.8 5,240 127 2.4

1995.09 ~ 1996.08 1,275 1.01 1.58 9.0 5,381 103 1.9

1994.09 ~ 1995.08 1,170 0.96 1.52 7.8 4,864 103 2.1

1994.01 ~ 1994.08 1,085 0.88 1.35 8.0 4,916 102 2.1

1993 1,005 0.83 1.25 8.6 5,045 228 4.5 1992 925 0.67 1.18 12.8 4,620 392 8.5 1991 820 0.63 1.12 18.8 4,556 393 8.6 1990 690 0.55 0.97 15.0 4,386 187 4.3 1989 600 0.55 0.89 26.3 3,053 328 10.7 1988 462.50 0.36 0.63 - 2,267 94 4.2

Source: Korea Minimum Wage Council (http://www.minimumwage.go.kr) Notes: Until the year 2002, the number of workers covered refers to regular employees only. * Hourly minimum wage rates in KRW were converted into GBP and USD at the average yearly exchange rate for

each year, except 2015 for which the average monthly exchange rate for February, 2015 was used. ** Influence rate = (B)/(A)*100.

152

2003), which was the first regime taken by the democratic party in Korean history of

politics, 10.6 per cent during the Roh, Moo-Hyun government, another democratic

regime (2003-2008), 5.2 per cent during the Lee, Myung-Bak government (2008-2013),

which claimed to stand for neoliberalism, and 7.4 per cent for four years of the Park,

Geun-Hye government (2013-2017) which showed a strong conservative inclination.

As shown in Figure 6.2, minimum wage rates have increased at rates which are mostly

higher than annual inflation rates and GDP per capita growth in South Korea. The rates

have been over 40 per cent of the median wage of full-time workers since 2007. This is

analogous to the level at which most countries set their minimum wages; however,

based on the mean wage of full-time workers, the minimum wage has still remained at a

little higher than 30 per cent. The fact that the level of minimum wages relative to the

mean wage is much lower than the level relative to the median implies the magnitude of

wage inequality. The minimum wage is also short of the minimum cost of living in

South Korea. While the minimum cost of living per month for households with two

members, which the Ministry of Health and Welfare set in 2015, was 1,051,048 KR won

(£ 625.3 or $ 956.9 at the exchange rate of February, 2015), the monthly earnings,

which a minimum wage worker was able to earn with working for 8 hours a day, 21

days a month, was 937,440 KR won (£ 557.7 or $ 853.5 at the exchange rate of

February, 2015), 89.2 per cent of the minimum cost of living.

Figure 6.2 Growth in Minimum Wage, Annual Inflation (CPI) & GDP per Capita, and Minimum Wage Relative to Average Wage of Full-Time Workers in South Korea, 1988- 2013 (%)

Source: Korea Minimum Wage Council (http://www.minimumwage.go.kr) & OECD.Stat (http://stats.oecd.org/Index.aspx); Figures in Appendix 3.

-20.0

0.0

20.0

40.0

60.0

19 88

19 89

19 90

19 91

19 92

19 93

19 94

19 95

19 96

19 97

19 98

19 99

20 00

20 01

20 02

20 03

20 04

20 05

20 06

20 07

20 08

20 09

20 10

20 11

20 12

20 13

20 14

Growth in minimum wage Change in annual inflation Growth in GDP per capita

153

The ‘influence rate’ which refers to the ratio of workers benefiting from the minimum

wage increase relative to the workers covered as a whole shows that the minimum wage

has obviously extended its influence, in particular, since the early 2000s (Table 6.1).

The rates remained at a very low level during the 1990s with the exception of the

several years after 1989 when the coverage of the minimum wage was extended from

companies in manufacturing, mining, and construction to all companies with 10

employees or more has been over 10 per cent since September of 2005. This is

attributed ostensibly to the further expansion of the coverage to all companies with one

employee or more in 2000; but it is most likely that the sustained increases in the

influence rates since the early 2000s is owed largely to the rise in the number of

precarious, low paid jobs caused by labour market flexibilization after the economic

recession in 1998.

Figure 6.3 The Share of Employees Paid Less than the Minimum Wage in South Korea, 2001-2013 (%)

Source: Minimum Wage Council, (2014), p.142

Under the Minimum Wage Act, an employer who fails to comply with the obligation to

pay at least minimum wage is expected to be punished by imprisonment of up to three

years or a fine not exceeding 20 million KR won (£ 11,899.2 or $18,208.3 at the

exchange rate of February 2015), or both. But there is neither the official figures of

noncompliance nor specific measures of a government crackdown on it, which implies

that noncompliance is rarely caught and sanctioned. Even when employers paying less

than the minimum wage are caught, they are scarcely ever penalised. The Ministry of

Employment and Labour caught 832 cases in which organisations paid less than the

4.3 4.9 4.9 5.8

8.1 9.4

11.9 10.8

12.8 11.5 10.8

9.6 11.4

0.0 2.0 4.0 6.0 8.0

10.0 12.0 14.0

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

154

minimum wage in 2014, but only 16 were fined (People's Solidarity for Paricipatory

Democracy [참여연대], 2015). According to People’s Solidarity for Participatory

Democracy (2015), a Korean citizen campaign group, although 33 businesses violated

twice, and 2 companies three times within 2014, only one among the habitual violating

organisations was subject to judicial action. The level of noncompliance can be weighed

up by the share of wage earners paid less than the minimum wage, and the share has

increased by 2.7 times, from 4.3% in 2001 to 11.4% in 2013 (Figure 6.3). According to

the employer representative’ proposal submitted to the Minimum Wage Commission,

98.7 per cent of wage earners paid below the minimum wage is employed in small and

medium-sized businesses with 300 employees or under, and 88.5 per cent of them works

for small-scale establishments with 30 employees or under, such as convenient stores

and restaurants (The Minimum Wage Council [최저임금위원회], 2014).

Older Workers in the Labour Market

Older workers are one of the most vulnerable groups in the labour market in South

Korea, particularly in terms of insecure employment, low pay, and a high poverty rate.

The fact that there is a big gap between 53 years old, the average age of retirement from

a lifetime major workplace (The National Statistical Office of Korea [통계청], 2013)

and 71.1 years old for men and 69.8 years old for women, the average effective

retirement age from the labour force which is the second highest among the OECD

countries (OECD, n.d.) suggests that many older people are economically active and

employed in a new workplace after retiring from their lifetime jobs, and at the same

time that it is highly possible that they are in low-quality, low paid jobs due to the

difficulty of career changes.

The labour force participation rate and the employment-to-population ratio among older

workers aged 50 and over are very high and have steadily grown, except during the

credit card lending crisis in 2003 and the financial crisis in 2008-09, while the

unemployment rate has remained low at around 2-3 per cent. As shown in Figure 6.4,

the labour force participation rates have increased from 72.5 per cent in 2000, 73.8 per

cent in 2005 to 78.8 per cent in 2014 among those aged 50-54, from 64.1 per cent in

155

2000, 64.9 per cent in 2005 to 72.6 per cent in 2014 in the 55-59 age group, from 54.3

per cent in 2000, 54.5 per cent in 2005 to 59.8 per cent in 2014 among the 60-64 age

group, and from 29.6 per cent in 2000, 30 per cent in 2005 to 31.9 per cent in 2014

among those aged 65 and older. The employment rates have also risen from 70.2 per

cent in 2000, 72.1 per cent in 2005 to 77.2 per cent in 2014 in the 50-54 age group, from

62.2 per cent in 2000, 63.1 per cent in 2005 to 70.8 per cent in 2014 in the 55-59 age

group, from 53 per cent in 2000, 53.4 per cent in 2005 to 58.3 per cent in 2014 in the

60-64 age group, and from 29.4 per cent in 2000, 29.8 per cent in 2005 to 31.3 per cent

in 2014 among those aged 65 and older. Such growth might be attributed to the rapid

population ageing and the growing portion of the older population. However, Figure 6.5

shows that the growth rate of the labour force participation and of employment among

those aged 50 and older surpassed the rate of increase in the population of the age

group, except the two financial crisis periods. This implies that there has been

substantial rise in the labour force participation and in employment among people aged

50 and older, especially since 2005.

Figure 6.4 Changes in Labour force Participation, Employment and Unemployment among Workers Aged 50 + in South Korea, 2000-2014 (%)

Source: The National Statistical Office of Korea, Economically Active Population Survey [경제활동인구조사], Each Year; Figures in Appendix 4.

156

Figure 6.5 Changes in the Growth Rate in Population, Labour force Participation and Employment among Workers Aged 50+ in South Korea, 2001-2014 (%)

Source: The National Statistical Office of Korea, Economically Active Population Survey [경제활동인구조사], Each Year, Author’s calculation; Figures in Appendix 5.

The high rates of the labour force participation and of employment among older

workers in South Korea have been largely due to the high proportion of the self-

employed and/or workers engaging in agriculture. But, in recent years the number of the

self-employed has significantly fallen while the number of wage workers has

appreciably risen among people aged 50 years and older (Figure 6.6). The self-

employed have decreased by 5.3 percentage points, from 36.9 per cent in 2007 to 31.6

per cent in 2013, among people aged 50-59 years old and also by 6.7 percentage point,

from 50.1 per cent to 43.4 per cent, among those aged 60 years and over in the same

period. On the contrary, employees have increased during the same period from 54.3 per

cent to 61.6 per cent among people aged 50-59 years old and from 37.1 per cent to 45.8

per cent among those aged 60 years and over. However, older workers are more likely to

engage in irregular employment. The proportion of irregular workers is around 40 per

cent among older workers aged 50-59 years old and over 65 per cent among those aged

60 years and over, which is much higher than around 30 per cent, the proportion of

irregular workers among those aged 15-64 years old. In addition, recalling 53 years old,

the average age of retirement from a lifetime workplace, the changes in the type of

157

employment implies that older workers who have once retired from a workplace and

seek a job are more likely to have an irregular one. While the proportion of regular

workers has risen from 57.4 per cent in 2007 to 62.9 per cent in 2013 and the percentage

of irregular workers has fallen from 42.6 per cent to 37.1 per cent in the same period

among those aged 50-59 years old, those percentages among those aged 60 years and

over have developed in reverse, being reduced from 34.0 per cent in 2007 to 32.5 per

cent in 2013 in regular workers and increased from 66.0 per cent to 67.5 per cent in

irregular workers in the same period.

Figure 6.6 Changes in Employment Status of Workers Aged 50+ in South Korea, 2007- 2013 (%)

Source: The National Statistical Office of Korea, Supplementary Results (by Type of Employment and for Non-wage Workers) of the Economically Active Population Survey [경제활동인구조사 (근로형태별, 비임금근로자) 부가조사], Every August in Each Year, Author’s calculation; Figures in Appendix 6.

Looking at the employment of older workers by industry, there are considerable changes

in the proportion of those engaging in agriculture and in business, personal, public

service and others, whereas the proportion of people engaged in wholesale and retail,

hotels and restaurants remians steady at around 20 per cent (Figure 6.7). The percentage

of workers aged 55-79 years old who are engaged in agriculture has dramatically

decreased from 33.6 per cent in 2005 to 19.2 per cent in 2014. In constrast, the

proportion of those engaging in business, personal, public service and others has

significantly increased from 25.3 per cent to 33.5 per cent during the same period. We

may not be able to affirm that these changes mean that the number of low-wage

employees has increased among older workers since business, personal, public service

and others cover a wide range of industries, including sewerage, waste management,

158

and environmental industry (E), real estate and leasing service (L), professional

scientific and technical service (M), business facilities management and support

services (N), public administration and social security (O), education service (P), health

and social work (Q), arts, sports, recreation related service (R), repair and other personal

services (S), undifferenciated employment & production activities of households for

own use (T), and activities of extraterritorial organisations (U). However, Jung, S-M

(2011) assumes, based on her analysis of the employment features among older workers

with the raw data of the Economically Active Population Survey in 2010, that the

increase in business, personal, public services and other services among older workers

since 2005 is attributed largely to the growth in business facilities management and

support services (N), public administration and social security (O), and health and social

work sector (Q), all of which provide mostly low-wage jobs.3

Figure 6.7 Changes in Industrial Composition among Workers Aged 55-79 in South Korea, 2005-2014 (%)

Source: The National Statistical Office of Korea, Supplementary Results (for the Old Population) of the Economically Active Population Survey [경제활동인구조사 (고령층) 부가조사 ], Every May in Each Year, Author’s calculation; Figures in Appendix 7.

159

In the same vein, the most remarkable change in occupational composition among older

workers is that the proportion of skilled agricultural, forestry, and fishing workers has

reduced from 31.3 per cent in 2005 to 17.8 per cent in 2014 while the percentage of

craft and equipment, machine operation and assembly workers has risen from 14.2 per

cent to 21.0 per cent during the same period, and the proportion of elementary workers

has been the most in the composition since 2009 (Figure 6.8). This implies the influx of

older workers into low-wage jobs. According to Jung, S-M’s (2011) cross analysis of

occupations and industries where older workers were in 2010, 82.8 per cent of older

workers in business facilities management and support services (N) was elementary

workers, and of those, 88.9 per cent was occupied with building cleaning or security

services. Further, the transportation (H) occupied the most (81.0 per cent) of the

category, electricity, transportation, communication and finance in industrial

composition among workers aged 55-79, and 83.0 per cent of older workers in

transportation sector was simple machine operation workers.

Figure 6.8 Changes in Occupational Composition among Workers Aged 55-79 in South Korea, 2005-2014 (%)

Source: The National Statistical Office of Korea, Supplementary Results (for the Old Population) of the Economically Active Population Survey [경제활동인구조사 (고령층) 부가조사 ], Every May in Each Year, Author’s calculation; Figures in Appendix 8.

160

The wage level of older employees aged 50 and over sharply falls with age. Nam, J-R’s

(2014) analysis of older workers in the labour market with raw data of the

supplementary results (by Type of Employment) of the Economically Active Population

Survey [경제활동인구조사 (근로형태별) 부가조사] provides the average wages of age

groups, which allows to find the relative level of wages among older workers (Figure

6.9). The average monthly wage of employees aged 50-59 years old was 2,408,000 KR

won (£ 1,353.0 or $1,912.4 at the then monthly exchange rate) in March 2014, 94.0 per

cent of the average monthly wage of employees aged 30-49 years old. However, the

average monthly wages among those aged 60-64 years old and 65 and over were 61.7

per cent and 34.9 per cent respectively in the same year. The average hourly wages of

older workers, which are calculated by dividing the average monthly wage by hours

worked because people rarely know about their hourly wage and monthly pay is

collected as information about wage in most of survey data in Korea, are comparable

with the average monthly wages. It should be noted that older employees aged 50-64

years old have worked longer hours than employees aged 30-49 years old until very

recently, and thus their hourly wages have been less than those aged 30-49 years old,

except since March 2013 among employees aged 60-64. On the other hand, the average

weekly hours for which employees aged 65 and older worked has been significantly

reduced, and the gap of weekly hours worked between employees aged 65 and over and

those aged 30-49 has been growing.

Figure 6.9 Changes in the Ratios of Monthly Wage & Hourly Wage among Employees 50+ to among Those Aged 30-49 and Weerkly Hours Worked by Age Group in South Korea (%, hrs)

Source: The National Statistical Office of Korea, Supplementary Results (by Type of Employment) of the Economically Active Population Survey [경제활동인구조사 (근로형태별) 부가조사], Each Year; Recalculated by author based on figures from Nam, J-R (2014), pp.15-16; Original figures in Appendix 9.

161

Figure 6.10 shows the percentage of older workers earning less than 50 per cent of the

median wage among wage workers aged 15 years old and over. It informs of the level of

income security among older workers which is generated from their wages. In 1987, the

proportion of low paid older employees aged 50-54 years old, 55-59 years old, and 60-

64 years old was 5.69 per cent, 6.43 per cent, and 6.64 per cent, respectively. They were

the same with or slightly below 6.64 per cent, the proportion of low paid employees

aged 15-64. However, the number of the low paid has grown rapidly, in particular,

among older workers. In 2012, the percentage of employees aged 50-54 years old, 55-59

years old, and 60-64 years old earning less than 50 per cent of the median wage was

13.17 per cent, 17.74 per cent, and 27.9 per cent, respectively, while the percentage of

employees aged 15-64 years old was 10.02 per cent. The proportion of low paid older

employees aged 50-64 years old has sharply decreased after the peak for the period from

2000 to 2007. However, it should be also noted that the gap in the proportion had been

much larger during the peak between older employees and employees aged 15-64, and

the gap after the peak still remains higher, except those aged 60-64 years old, than

before the peak. Moreover, since the data analysed in Figure 6.10 concern only

employees in organisations with more than 10 staff, it is likely that the proportion of

older employees earning less than 50 per cent of the median wage is larger in practice.

Figure 6.10 Changes in the Proportion of Older Employees Earning Less Than 50% of the Median Wage in South Korea (%)

Source: Survey on Labor Conditions by Type of Employment [고용형태별 근로실태조사], each year; recited from Ahn, J-K & Yoo, S-H (2015); Figures in Appendix 10.

0 5

10 15 20 25 30 35 40 45 50

19 87

19 88

19 89

19 90

19 91

19 92

19 93

19 94

19 95

19 96

19 97

19 98

19 99

20 00

20 01

20 02

20 03

20 04

20 05

20 06

20 07

20 08

20 09

20 10

20 11

20 12

50-54 years old 55-59 years old 60-64 years old 15-64 years old

162

This brief review of Korean older workers in the labour market so far does not provide

convincing information on the proportion of minimum wage workers among older

employees, the proportion of older employees among minimum wage workers, or

changes in those proportions.4 However, figures on their employment status, industrial

composition, and occupational composition indicate that older workers have become

more prevalent in low paid jobs, and the figures regarding their wages indirectly support

the assumption that more older workers are becoming affected by the minimum wage.

In particular, recalling the fact that the level of minimum wage rates has been about 40

per cent of the median wage of full-time employees, the wage level among older

employees and the change in the proportion of older employees earning less than 50 per

cent of the median wage imply that older employees not only occupy significant portion

of minimum wage workers but also contribute to the growth of the share of those paid

less than the minimum wage.

One thing to add as regards the rise of low paid older workers is that this may be

interlinked with the increase in the unemployment of young people and in the rise in

female labour force participation. As the number of high-quality jobs have significantly

decreased, influenced by the IMF bailout programme, young people who largely

received higher education, unlike their parent’s generation, began to delay entering the

labour market until they got a permanent job in a large firm or in government. Most

small and medium-sized firms which have no capacity to meet young people’s needs

have been confronted by labour shortages and tended to hire older people and women.

As the labour market has dramatically changed since the early 2000s, and men’s

employment has become more insecure, many women who were anxious about

financial shortages in their families started to engage in the labour market. This seems

to have brought about a significant change in the perception of adult children’s support

for their parents. Women’s participation in the labour market triggered by the concern

over family income provided the adult children with a good excuse to refuse financial

support and caregiving for their parents. Accordingly, it would be assumed that older

people tend to be placed in a situation in which they have no choice but to take care of

themselves and participate in paid jobs much more actively than their predecessors.

163

Further to the basic statistical information on Korean older workers, it would be useful

for understanding their status in the labour market to look at the scope and degree of

disadvantages and discrimination that they face. In Korea, the Act on Prohibition of Age

Discrimination in Employment and Elderly Employment Promotion [고용상 연령차별

금지 및 고령자고용촉진에 관한 법률] was enacted in 2008, amending the former

Employment Promotion for the Aged Act [고령자고용촉진법]. This Act prohibits

employers from discriminating against individuals on the basis of age regarding

recruiting, hiring, salary, education and training, placement, transfer, promotion,

retirement, and dismissal. Under the Act, an employee who has experienced

discrimination by age may file a petition with the National Human Rights Commission

[국가인권위원회]. The Commission investigates claims and recommend corrective

action to employers. If the employers fail to comply with the Commission’s advice

without a justifiable reason, and the discrimination is substantial, the Commission

notifies the cases to the Ministry of Employment and Labour [고용노동부] with non-

binding advisory opinions. The Ministry of Employment and Labour can issue an order

to employers to rectify the discriminatory treatment. Failure to comply with the

Ministry’s order may result in an administrative fine of up to 30 million Korean won

(about £ 20,000 or $ 25,000 at the exchange rate of June, 2019), and discriminatory

treatment in recruiting and hiring, in particular, may cause a criminal fine of up to 5

million Korean won (about £ 3,400 or $ 4,200 at the exchange rate of June, 2019),

regardless of the petition process through the National Human Rights Commission.

Since the Act on Prohibition of Age Discrimination in Employment and Elderly

Employment Promotion [고용상 연령차별금지 및 고령자고용촉진에 관한 법률] was

implemented, the number of petitions regarding age discrimination committed by

employers which were filed in the National Human Rights Commission has

significantly increased from 62 cases in 2008 to 138 cases in 2009, 196 cases in 2010,

and 151 cases per year on average between 2011 and 2013 (The National Human Rights

Commission of Korea [국가인권위원회], 2014). 1,214 cases of 16,912 petitions which

have been filed for discrimination with the Commission until 2014 were age

discrimination, and employment-related cases including recruiting (368 cases, 39.7%),

hiring (303 cases, 32.7%), wage (12 cases, 1.3%), non-wage reward (4 cases, 0.4%),

164

training (9 cases, 0.9%), placement (46 cases, 5.0%), promotion (16 cases, 1.7%),

retirement (38 cases, 4.1%), dismissal (65 cases, 7.0%) and others (26 cases, 2.8%)

accounted for 76. 5 per cent (928 cases) of all age discrimination cases (The National

Human Rights Commission of Korea [국가인권위원회], 2014).

Unfortunately, there is no data which allow us to overview the characteristics of age

discrimination in the Korean labour market. Moreover, studies of age discrimination

relating to employment and labour have recently emerged in the line of research on

ageism which started to be developed since 2010 (Kim, M-J and Yoon, K-Y, 2017).

Kim, D-S and Mo, S-H (2011) analysed 52 cases investigated by the National Human

Rights Commission [국가인권위원회] to find causes of age discrimination in

employment. They found that ageism, including stereotypes and discriminative

attitudes, is the most prominent and widespread factor to affect age discrimination. Kim,

D-S and Mo, S-H (2012) also surveyed 300 personnel managers in small and medium-

sized firms to figure out how personnel managers’ stereotypes of older workers affect

the managers’ discriminatory perceptions of older workers with regard to employment,

and found that the higher degree of stereotypes a personnel manager had, the stronger

his discriminatory perceptions of older workers were in terms of employment.

Empirical Studies about the Minimum Wage Effects

The effects of the minimum wage had rarely been explored in South Korea before the

end of the 1990s and began in earnest after the mid-2000s. A few empirical studies

include the effects of the minimum wage on older employees, particularly on their job

retention or new hires. Findings are not consistent as the ones in other countries are.

The existing studies about the minimum wage effects in Korea (Table 6.2) focus largely

on the employment effect, but there is no consensus on the issue among the research

findings. Kim, Y-S, Gwon, H-J, and Kim, J-J (2004), Lee, S-K (2007), Lee, B-H (2008),

Ahn, T-H (2009), Kang, D-U (2010), Kim, J-Y (2011), Kim, Y-S (2011), and Hwang, S-

J (2015) reported no significant disemployment effects or positive effects. Kim, Y-S,

Gwon, H-J, and Kim, J-J (2004) who conducted a time-series analysis using monthly

data from the Economically Active Population Survey [경제활동인구조사] for the

165

period of January 1988 to March 2004 found that the minimum wage had no significant

effect on employment and even increased employment among women and some age

groups depending on explanatory variable manipulation. Lee, S-K (2007) who

examined the change in employment by job and by organisation as a consequence of the

change in the real minimum wage, using the supplementary data of the Economically

Active Population Survey [경제활동인구조사 부가조사] for the period of 2000-2006

and the Workplace Panel Survey(WPS) [사업체 패널조사] for the period of 2003-2004

also found neutral or positive employment effects, based on a fixed-effects model as

well as a simple regression model. Similarly, Lee, B-H (2008), using the supplementary

data (by Type of Employment) of the Economically Active Population Survey [경제활동

인구 (고용형태별) 부가조사] and the Economically Active Population Survey [경제활

동인구조사] for the period of 2004-2005, concluded that the minimum wage did not

have disemployment effects, in particular, among women, younger workers, and older

workers, from his analysis with difference-in-difference of the change in employment

among employees who earned a minimum wage and those who made a little more than

the minimum wage. Ahn, T-H (2009) explored the odds of job retention among wage

workers after an increase of the minimum wage through a linear probability model with

fixed-effects using the Korean Labour and Income Panel Study (KLIPS) [한국노동패널

조사] for the period of 2001-2007 and did not find any considerable negative effects on

employment. Kang, D-U (2010) examined the effect of the minimum wage on the

employment rate of employees with disabilities using Survey on the Employment Status

of the Disabled in Business 2008 [2008년 사업체 장애인고용 실태조사]. He found

through a quasi-experimental analysis with dummy variables and time-series analysis

that if a minimum wage raises the wage of employees with disabilities by 1 percentage

point, their employment will decrease by 0.58 percentage point in the short term, which

was not a considerable amount, and moreover, no significant disemployment effect was

found in the long term. Kim, J-Y (2011), who conducted an analysis of the odds of job

retention among wage workers after an increase in the minimum wage using the KLIPS

[한국노동패널조사] for the period of 1998-2008 and probit models plus a linear

probability model with fixed-effects and random-effects, also found no statistically

significant disemployment effect. Likewise, Kim, Y-S (2011), who examined

166

Table 6.2 Empirical Studies about the Effects of the Minimum Wage in South Korea

Study Data sources Outcome measures/Method Main results Other remarks Employment Kim, Y-S, Gwon, H-J, & Kim, J-J 2004

Economically Active Population Survey (monthly) [경제활동인구조사 (월별)], Jan. 1988 – Mar. 2004

Time-series analysis: OLS (Ordinary least squares)

model, GLS (Generalised least squares) model, ARMA (Autoregressive- moving-average) model

No significant employment effect and even positive employment effects among women and some age groups, such as employees aged 25-54 years old and those 55 and older, depending on explanatory variables, the ratio of minimum wages to mean wages and logarithms of minimum wages

Lee, S-K 2007

Supplementary data of the Economically Active Population Survey [경제활동인구조사 부가조사], 2000-2006;

Workplace Panel Survey (WPS) [사업체 패널조사], 2003-2004

Fixed-effects models; Simple regression model

Neutral employment effects among the employees as a whole and even positive effects among the low paid depending on explanartory variables, logarithm of a real minimum wage divided by a real mean wage each year by job unit and logarithms of a real minium wage

Lee, B-H 2008

Economically Active Population Survey [경제활동인구조사] & supplementary data (by type of employment) of the Economically Active Population Survey [경제활동인구조사 부가조사 (고용형태별)], 2004-2005

Quasi-experimental analysis: Difference-in-difference

No statistically significant positive or negative effects on job retention and new hires among women, younger workers aged 15-24 years old and older workers aged 55 years old and over as well as the employees as a whole

Jeong, J-H & Lee, B-H 2008

Economically Active Population Survey [경제활동인구조사] & the Wage Structure Survey [임금구조기본통계조사], 2000-2006

Time-series cross-section analysis:

Poolded OLS model, Fixed-effects model, PCSE (Panel-Corrected

Standard Errors) model;

Quasi-experimental analysis: Difference-in-difference

Negative employment effects among younger workers aged 20-24 and older workers aged 55 and older compared to positive effects among those aged 25- 54 years old;

No statistically significant positive or negative effects

on job retention and new hires in any age group as well as among the employees as a whole

167

Nam, S-I 2008

Data on caretakers of 132 apartment complexes around the capital area, 2007

Reduced form model: Random-effects model

Increase in wage by over 10 % but decrease in employment by 3.5 % and working hours by 13.5% caused by the extended enforcement of the minimum wage to surveillance or intermittent work

Those engaged in surveillance or intermittent work only

Ahn, T-H 2009

Korean Labor and Income Panel Study (KLIPS) [한국노동패널조사], 2001-2007

Fixed-effects model No significant employment effect

Kim, W-Y 2010

Time-series data from the Occupational Employment Statistic (OES) [산업·직업별 고용구조조사], 2000-2008

Time-series analysis: Arellano-Bond difference

GMM(Generalised Method of Moments)

1.6% decrease in employmnet among female workers aged 15-24 years old ; 1.1% decrease among male workers aged 15-19 years old; no significant employment effect among younger workers as a whole

Younger employees aged less than 25 years old only

Kang, D-U 2010

Survey on the Employment Status of the Disabled in Business 2008 [2008년 사업체 장애인고용 실태조사];

Unofficial inside data of the Minimum Wage Council and the Ministry of Employment & Labor, 2000-2008

[Short-term effect] Quasi- experimental analysis:

Multiple regression model [Long-term effect] Time-

series analysis: Correlation between

employment rate among the disabled and increase rate in the minimum wage

In short term, 0.58% decrease in employment among the disabled by 1% increase in their wage, which is not considerable amount; but , in long term, no significant disemployment effect

Employees with disabilities only

Kim, J-Y 2011

KLIPS, 1998-2008 Quasi-experiment analysis: Pooled Probit model, Fixed-effects model, Random-effects model

No statistically significant employment effects

168

Kim, Y-S 2011

Economically Active Population Survey (monthly) [경제활동인구조사 (월별)], Jan. 1990 – Apr. 2010

Time-series analysis; OLS model, GLS

(Generalised Least Squares) model, ARMA (Autoregressive-Moving- Average) model

No significant disemployment effect and even positive effects on employment among those 15-19 years old and employees aged 25-54 (10% significance level) during the period of 1990-2011; positive employment effects after 2000 (5% significance level), except among those aged 55 and older

Kim, D-I 2012

Survey on Labor Conditions by Type of Employment [고용형태별 근로실태조사], 2008-2010

Time-series analysis: First-difference

6.6% reduction in new hires among the bottom 5% wage bracket; 3.8% reduction among of male younger workers aged below 29 years old at the bottom 5-15% wage bracket (5% significance level); 2.7% reduction among male employees aged 30-54 years old at the bottom 5-15% wage bracket (5% sig. level); 2.0% (10% sig. level) and 2.9% (5% sig. level) reduction among female workers aged below 29 years old at the bottom 5 % and 5-15% wage brackets, repectively; 34.7% reduction among female older employees aged 55 and older at the bottom 5% wage bracket (10% sig. level); 9.3% reduction in small companies with below 5 employees (10 % sig. level); 13.8% reduction in mining and manufacturing sector (5% sig. level); and 11.7% reduction in whole sale & retail and restaurant & hotels sector (1 per cent sig. level), all of which are caused by 1% increase in the minimum wage

Kim, M-S, Kim, Y-M, & Park, T-S 2013

KLIPS, 2000-2008 Probit model 3.7% decrease in employment among aged 24 years old or less with a high school deploma by a 10% increase in the real minimum wage

169

Kim, Y-M 2014

Supplemetary data (August) of the Economically Active Populaton Survey [경제활동인구조사 8월 부가조사], 2001-2014

Probit model Increase in wage among the employees as a whole by increase in the minimum wage, higher increase in service sector than in manufacturing;

Reduction in odds of becoming an employee on contract for one or more years, higher reduction in manufacturing than in service sector

Employees in manufacturing and service industries only

Hwang, S-J 2015

Wage Structure Survey [임금구조기본통계조사], 2009-2012 & 2002-2008

Fixed-effects model No statistically significant employment effects

Wage Mobility Kim, J-G 1999

Wage Structure Survey [임금구조기본통계조사], 1988-1995

Simulation analysis Less than 0.03% increase in the share of the bottom 20% in the wage distribution until 1993; no changes of the share after 1993

Kang, B-G & Sung, H-Y 2009

KLIPS, 2001-2007 Wage growth model Increase in both the average wages of the lowest, two wage brackets and the household earnings of, in particular, the lowest houshold earned income bracket

Jeong, J-H 2011

Survey on Labor Conditions by Type of Employment [고용형태별 근로실태조사], 2007-2009; KLIPS, 1998-2008

Comparison between minimum wage rates and hourly wages calculated by ordinary monthly wages and hours worked;

OLS model, Fixed-effects model

Spikes in the minimum wage and spillovers into the upper wage brackets with reduction in the ratio of the 10th and 90th percentiles of the wage distribution

Kim, Y-M & Kim, M-S 2013

Supplementary data (August) of the Economically Active Population Survey, 2004, 2010

Counterfactual wage distribution model

An ambiguous effect on female wage inequality Female employees only

Seong, J-M 2014

Local Area Labor Force Survey [지역별 고용조사], 2008-2012

OLS models, Fixed-effects models

Reduction in wage ineaulqity for lower and middle wage brackets with the size of reduction effects declined from lower to middle wage brackets

170

Family income & Poverty Jeong, J-H 2005

KLIPS, 2003 Ratios of beneficiary working households per household income bracket

More beneficieries among the higher household income brackets than among the bottom

Lee, S-K 2013

KLIPS, 2005-2010 Multi-level, multi-nominal logit model

More likely to lift the working poor out of poverty and not adversely related to their job retention

Seo, J-W & Jeong, J-O 2014

KLIPS, 2003-2008 Binary probit model with random-effects

Higer increase in minimum wage, less odds of falling down below the poverty line

171

employment effects for twenty years from January 1990 to April 2010 using a time-

series analysis with monthly data of the Economically Active Population Survey [경제

활동인구조사], reported that there have not been disemployment effects since the

minimum wage was introduced in South Korea, and rather that positive effects were

found after 2000. Most recently, Hwang, S-J (2015), who examined the change in

employment among low-wage workers by sex, age, education, organisation size, and

industrial sector, using the Wage Structure Survey [임금구조기본통계조사] for two

periods, 2009-2012 and 2002-2008, found that when the minimum wage increased, no

decline in employment was statistically significant.

However, Jeong, J-H and Lee, B-H (2008), Nam, S-I (2008), Kim, W-Y (2010), Kim,

D-I (2012), Kim, M-S, Kim, Y-M, and Park, T-S (2013), and Kim, Y-M (2014) all

reported adverse employment effects. Using panel data which were rebuilt with the

Economically Active Population Survey [경제활동인구조사] and the Wage Structure

Survey [임금구조기본통계조사] for the period of 2000-2006, Jeong, J-H and Lee, B-H

(2008) found that the minimum wage had negative effects on employment among

younger workers aged 20-24 and older workers aged 55 and older whereas it had

positive employment effects among those aged 25-54. They also examined the

minimum wage impact on job retention and new hires through a quasi-experimental

model using panel data which were rebuilt with the supplementary data of the

Economically Active Population Survey [경제활동인구조사 부가조사] and the

Economically Active Population Survey [경제활동인구조사] for the period of 2004-

2005, and reported that there was no statistically significant positive or negative effects

in any age group as well as among wage workers as a whole. Nam, S-I (2008) analysed

the employment effects of the introduction in 2007 of the minimum wage rate for those

engaged in surveillance or intermittent work. He used data about caretakers who worked

in 132 apartment complexes around the capital area and found that the newly enforced

rate for surveillance or intermittent work increased their wage by over 10 per cent but

decreased employment by 3.5 per cent. Kim, W-Y (2010) explored the effects of the

minimum wage on employment among younger employees aged 15-24 and those aged

15-29. He conducted a time-series analysis of the data which were rebuilt by the

researcher for 15 regions from 2000 to 2008 using the Occupational Employment

172

Statistic (OES) [산업·직업별 고용구조조사] and reported that a 10 per cent increase in

the minimum wage reduced employment by 1.6 per cent among female younger

workers aged 15-24 and by 1.1 per cent among male employees aged 15-29 while

disemployment effects were not found among younger workers as a whole. Using the

Survey on Labour Conditions by Type of Employment [고용형태별 근로실태조사] for

the period of 2008-2010, Kim, D-I (2012) explored whether an increase in the minimum

wage decreased the hiring of new employees among the low paid and found that an 1

per cent increase in the minimum wage was likely to restrain new hires among the

bottom 5 per cent wage bracket in the wage distribution by 6.6 percentage point at a 10

per cent significance level. He added that an 1 per cent increase in the minimum wage

reduced new hires of male younger workers aged below 29 years old at the bottom 5-15

per cent wage bracket by 3.8 per cent at a 5 per cent significance level, of male

counterparts aged 30-54 years old by 2.7 per cent at the same significance level, of

female older workers aged below 29 years old at the bottom 5 per cent wage bracket by

2.0 per cent at a 10 per cent significance level, of female counterparts aged 55 and older

by 34.7 per cent at the same significance level, of female employees aged below 29

years old at the bottom 5-15 per cent wage bracket by 2.9 per cent at 5 per cent

significance level, of small companies with below 5 employees by 9.3 per cent at a 10

per cent significance level, of mining and manufacturing sector by 13.8 per cent at a 5

per cent significance level, and of wholesale and retail and restaurant and hotels sector

by 11.7 per cent at a 1 per cent significance level. Kim, M-S, Kim, Y-M, and Park, T-S

(2013) examined employment effects of the real minimum wage by the size of region,

age, and education, using the KLIPS for the period of 2000-2008 and a probit model.

Their findings show that a 10 per cent increase in the real minimum wage reduced the

probability of employment among those aged 24 or less with a high school diploma by

3.7 per cent at a 1 per cent significance level. Kim, Y-M (2014) explored the effects of

the minimum wage on wage distribution and employment in manufacturing and service

sectors, using the supplementary data (August) of the Economically Active Population

for the period of 2001-2014 and a probit model. The findings show that an increase in

the minimum wage raised the wage level in both sectors with higher rise in service

industry than in manufacturing but reduced the odds of becoming an employee on

contract for one or more years with higher decline in manufacturing than in the service

173

sector.

Another research stream has focused on whether minimum wage policy has wage or

earnings distribution effects or poverty-alleviating effects as its own purpose. Kim, J-

G’s (1999) simulation analysis is an early study about the distribution effect of the

minimum wage. It used the Wage Structure Survey [임금구조기본통계조사] for the

period of 1988-1995 and found that the minimum wage increased the share of the

bottom 20 per cent wage bracket in the distribution of the total earnings by less than

0.03 percentage point, and the change in the share after 1993 was near to zero. He

concluded that the minimum wage made no significant distributional improvement in

wages, and the near-zero effect after 1993 was due to the rate of increase in the

minimum wage which fell into 8.6 per cent in 1993 from 18.2 per cent on average for

the previous four years and was kept on under the wage growth rate. However,

subsequent studies show different findings as regards the distributional effects. Using

the KLIPS for the period of 2001-2007, Kang, B-G and Sung, H-Y (2009) compared the

distribution effects of the Earned Income Tax Credit (EITC) and the national minimum

wage. In terms of the minimum wage effect, they found that the minimum wage

significantly increased both the average wages of the lowest, two wage brackets and the

household earnings of the lowest household earned income bracket. They concluded

that the importance of the statutory minimum wage in Korea as the means of income

support for the working poor was not weakened, particularly as long as the eligibility

and the subsidy rate of the EITC were highly limited. Jeong, J-H (2011) reported similar

findings and verified the existence of spikes in the minimum wage through a

comparison between minimum wage hourly rates and hourly wages calculated using the

information on employees’ ordinary monthly wages and hours worked from the Survey

on Labour Conditions by Type of Employment [고용형태별 근로실태조사] for the

period of 2007-2009 and also of spillovers into the upper wage brackets in the wage

distribution using the KLIPS for the period of 1998-2008. He found that increases in the

minimum wage reduced the proportion of the 10th and 90th percentiles of the wage

distribution and raised mainly the wage of the middle percentile. Seong, J-M (2014),

who used the Local Area Labour Force Survey [지역별 고용조사] for the period of

2008-2012, also found that the minimum wage reduced wage inequality for lower and

174

middle brackets of the wage distribution even though the effect declined as the wage

rank was higher from lower to middle.

Jeong, J-H (2005), Lee, S-K (2013), and Seo, J-W and Jeong, J-O (2014) focused on

family income distribution and poverty alleviation effects. Jeong, J-H (2005) explored

whether the minimum wage affects the working household income distribution, using

the KLIPS of 2003. He found that the percentage of beneficiaries from an increase in

the minimum wage among the bottom household income bracket was much lower than

the one of their counterparts among the higher income brackets and concluded that

minimum wage policy was a blunt instrument to improve income distribution among

the working household. However, subsequent studies provided conflicting results. Lee,

S-K (2013) examined whether an increase in the minimum wage alleviates poverty

among the working poor. Using the KLIPS for the period of 2005-2010 and the

multilevel multinomial logit model, he analysed the transition from the working poor to

the working non-poor, the unemployed, or the economically non-active and found that

an increase in the minimum wage was more likely to lift the working poor out of

poverty as well as not adversely related to job retention. Seo, J-W and Jeong, J-O (2014)

also studied the effects of minimum wage policy on reducing poverty rates using the

KLIPS for the period of 2003-2008. They concluded, based on their results from a

binary probit model, that a higher hourly minimum wage can reduce the odds of an

individual’s falling down below the poverty line.

Conclusion

This chapter has reviewed the national minimum wage and older workers in the labour

market in South Korea. The statutory minimum wage which was introduced in 1987 for

only manufacturing companies with ten employees or more has been extended to all

organisations with one employee or more since 2000 after several amendments. It is

adjusted annually by the government’s approval for a rate which the Minimum Wage

Commission, a tripartite consultation body consisting of workers, employers, and public

interest representatives proposes in the light of criteria stated in the Minimum Wage Act

and national economic conditions. The level of the minimum wage is controversial

175

because it varies considerably with reference wages, which implies wage inequality.

Meanwhile, older workers are a vulnerable group in the Korean labour market.

Although a growing number of older workers have been economically active, many of

them take part in irregular work, mainly in business facilities management and support

service, public administration and social security, health and social work, and wholesale

and retail and hotels and restaurants, mostly as elementary workers, service and sales

workers, or simple machine operating workers. The wage level of older employees aged

50 years old and over dramatically decreases with age, and the percentage of older

employees paid less than 50 per cent of the median wage significantly exceeds the one

of their counterparts aged 15-64 years old by 3.15 per cent at least and by 17.88 per cent

at most. All these facts refer to the high probability of the increase in the proportion of

older employees who are paid low and further likely to be affected by the minimum

wage. Despite their vulnerability in the labour market, older workers have not been

taken much attention in the research of the minimum wage effects. Some empirical

research includes the effects of the minimum wage on older workers, mostly on their

employment, in Korea but the findings are not consistent with each other.

176

Notes

1 Those engaged in surveillance or intermittent work were legally allowed to be paid 30 per cent less than the minimum wage in 2007, 20 per cent less from 2008 to 2011, and 10 per cent less from 2012 to 2014. As the curtailment clause applied to them was removed from the Minimum Wage Act, they have to be paid a normal minimum wage rate from 1st of January 2015. 2 From 1994 to 2006, a new minimum wage rate had been announced by the 5th of August every year and taken effect from the 1st of September in the same year. 3 In particular, the increase in public administration and social security seems to be due to ‘the Work for Hope [희망근로]’, a government-funded employment project which provides low- wage, temporary jobs to people aged 18 years old and over who are in a low-income family and vulnerable in employment. 4 In order to get the information on the portion of minimum wage workers among older employees or the percentage of older employees among the minimum wage workers, raw data of supplementary results (by the Type of Employment) of Economically Active Population Survey should be analysed. However, there is an access limit to raw data, and thus the review of Korean older workers in the labour market is based on data provided by Korean Statistical Information Service (KOSIS) and by previous study which was analysed with raw data.

177

Chapter Seven The Effects of the Minimum Wage on Older Workers in South

Korea

Introduction

This chapter presents the results of quantitative analyses which explored the two

research questions, how the minimum wage affects the distributions of wages and family

earnings, employment and exits from in-work poverty among older workers and whether

different rates of increases in the minimum wage have different effects among older

workers. As mentioned in Chapter Five, the two questions were simultaneously

examined by subtopic for employees aged 55 and older, using the KLIPS and statistical

methods and comparing the two five-year periods of Roh, Moo-Hyun government

(2003-2008) with 10.6 per cent of average rate of increase in the minimum wage and

Lee, Myung-Bak government (2008-2013) with 5.2 per cent of average growth rate.

After reviewing the basic characteristics of the cases selected from the KLIPS, the

empirical results are presented in each subtopic section. The statistical methods and

variables used for each statistical analysis are summarised in Appendix 11.

Basic Characteristics of Data

Table 7.1 shows the number and main characteristics of employees aged 55 and older

selected from wave 7 in 2004 to wave 16 in 2013 of the KLIPS. The number of cases

significantly increased since 2009 as a result of the supplementation of households in

twelfth wave. A remarkable feature of the sample is that the ratio of female older

employees has evidently grown while the ratio of male counterparts has declined since

2009. This corresponds with another feature that the proportion of temporary jobs has

risen while the proportion of permanent jobs fallen since 2009. These are largely due to

the combination of the replacement of irregular jobs for permanent ones and women’s

participation in the labour market, particularly among 40s and older having much

financial burdens in households after the global financial crisis in 2008 which

178

accelerated a big change in Korean economy and labour market. The mean and median

ages which were slightly higher for the second five-year period than for the first five-

year period reflect the gradual adjustment in the labour market to rapid population

ageing in South Korea. However, in practical terms, it is likely that older workers

flowed into non-regular, temporary jobs by government-driven public work projects

after the financial crisis have led the small increase in the average ages among older

employees aged 55 and older. In terms of the level of education obtained, the percentage

of those with a bachelor’s degree visibly decreased in the second five-year period

compared to the first five-year period while the ratios of those with high school diploma

and less, except no education, were kept relatively stable over the ten years. This could

mean that facing a career transition, older workers with a bachelor’s degree were more

likely to choose at a point in time a transition to self-employment or an exit from the

labour market based on their better financial preparation for later life, rather than to take

low paid jobs in the harsher labour market conditions after the global financial crisis in

2008. In contrast, those with qualifications below a bachelor’s degree seem to have been

more likely to seek being employed, even in low paid jobs, for financial reasons. Such

speculation is based on the assumption that low paid jobs are dominantly open for older

workers in a very transitional period of their careers. The large gaps between the mean

and the median of hourly wages and of household earnings per capita and the high

proportion of those affected by the minimum wage and of working poor, which are

possibly even underestimated with the loss of cases in calculation required to use two

different datasets for one year, support the assumption to some degree and imply the

financial needs prevailing among older employees. The majority of the older employees

were married. However, the proportion of married employees decreased whereas the

share of separate, divorced, or widowed counterparts increased, notably in the second

five-year period. The older employees were chiefly householders, and their proportion

was noticeably high in two metropolitan cities, Seoul and Busan, and Gyeonggi-do, a

large province nearest to Seoul, in which various small and medium-sized firms have

developed. The number of family members among the older employees’ households

also decreased slightly during the second five-year period.

179

Table 7.1 Characteristics of 55+ Employees, 2004-2013

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Case 457 463 486 521 579 774 828 897 1,001 1,070 Sex (%)

Male 67.0 68.7 68.1 64.3 64.9 62.0 62.0 59.5 59.4 59.3 Female 33.0 31.3 31.9 35.7 35.1 38.0 38.0 40.5 40.6 40.7

Age Mean 60.70 60.84 60.77 60.89 61.03 61.89 61.97 62.04 61.96 62.26

Median 60.00 60.00 59.00 59.00 60.00 60.00 60.00 61.00 60.00 61.00 Education (%)

No Education 9.4 9.1 8.0 6.7 6.2 6.3 6.0 5.2 4.3 3.4 Elementary School 33.5 32.4 29.8 31.3 29.0 32.6 31.4 31.4 28.3 27.9

Middle School 17.9 21.0 20.2 20.9 19.9 21.6 23.2 23.4 23.6 23.3 High School 26.5 24.2 28.6 26.9 29.7 28.0 27.3 27.4 30.4 31.1

2-year College 0.4 1.1 1.2 1.3 2.4 1.7 2.2 3.0 3.1 3.1 University 9.4 9.5 9.5 10.4 10.7 7.8 7.9 7.0 7.4 8.3

Graduate School 2.8 2.8 2.7 2.5 2.1 2.1 2.1 2.5 2.9 2.9 Marital Status (%)

Never Married 0.2 0.2 0.0 0.0 0.2 0.9 1.2 0.9 1.2 0.8 Married 80.3 80.1 79.0 78.5 78.4 74.4 74.6 73.9 74.1 75.1

Separate/Divorced/Widowe d

19.5 19.7 21.0 21.5 21.4 24.7 24.2 25.2 24.7 24.0

Householder (%) 78.6 79.0 80.0 79.1 78.8 78.2 77.8 76.3 75.4 74.7 N of Household Members

Mean 3.1 3.1 3.1 2.2 2.9 2.8 2.7 2.6 2.6 2.7 Median 3.0 3.0 3.0 2.0 3.0 3.0 3.0 2.0 2.0 3.0

Type of Employment (%) Permanent 60.8 58.5 58.4 57.6 55.4 49.1 47.8 46.2 48.0 47.3 Temporary 14.7 14.5 16.0 15.7 18.7 24.9 25.6 26.1 27.4 28.4 Daily Hire 24.5 27.0 25.5 26.7 25.9 26.0 26.6 27.8 24.7 24.3

Region (%) Seoul 26.3 26.8 28.0 25.7 27.3 24.7 23.7 22.3 22.7 22.2 Busan 11.2 11.0 9.7 9.6 9.3 9.3 9.2 9.4 10.1 9.1

180

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Region (%) Continued Daegu

5.9 6.5 5.6 5.6 4.7 4.7 6.2 6.2 5.0 6.0

Daejeon 3.7 4.1 3.9 4.4 4.0 3.1 3.4 4.1 3.7 2.9 Incheon 4.8 4.3 5.3 6.0 5.9 4.9 5.9 5.6 5.8 6.4

Gwangju 3.7 5.0 4.7 4.2 2.8 2.1 2.5 2.7 2.3 2.7 Ulsan 1.8 2.2 2.5 1.9 2.6 2.3 2.3 2.2 3.0 2.8

Gyeonggi-do 17.7 17.3 17.9 18.2 18.7 16.9 15.6 17.9 18.5 18.7 Gangwon-do 1.3 0.9 1.2 1.7 1.6 2.5 2.3 2.1 2.3 2.6

Chungcheongbuk-do 3.3 2.6 3.9 3.8 3.6 3.5 4.0 4.8 4.2 3.9 Chungcheonnam-do 2.0 2.4 2.1 3.3 3.5 4.4 5.4 5.5 5.1 5.3

Jeollabuk-do 6.3 5.0 4.3 5.0 3.8 4.3 3.7 3.8 3.8 4.1 Jeollanam-do 2.4 2.8 2.1 2.5 2.6 4.0 3.4 2.6 2.8 3.1

Gyeongsangbuk-do 3.9 3.2 3.3 1.7 2.9 4.5 5.0 4.1 4.3 3.9 Gyeongsangnam-do 5.7 6.0 5.6 6.3 6.9 8.8 7.2 6.5 6.2 5.7

Jeju - - - - - 0.1 0.2 0.2 0.3 0.5 Hourly Wage (\)a

Mean 6,864.64 6,941.23 7,367.03 8,532.00 8,059.75 7,475.22 7,817.29 8,261.84 8,615.62 9,273.03 Median 3,873.67 4,606.53 4,798.46 5,150.35 5,289.91 4,975.64 5,374.28 5,719.77 6,184.69 6,909.79

Yearly Household Earnings per Capita in a Household (\)a

Mean 16,534,977 19,091,334 19,932,451 22,331,819 19,275,246 21,298,915 21,804,941 24,749,189 24,470,673 25,850,291

Median 12,908,588 15,900,000 16,161,017 17,464,846 15,467,214 16,970,563 17,320,508 19,890,914 20,859,650 21,550,000 Proportion Affected by Minimum Wage b

19.7 24.9 25.1 27.8 23.1 20.8 25.9 26.7 25.1 24.7

Proportion of Working Poor c

25.4 24.6 28.8 25.3 27.8 26.9 28.6 27.2 23.9 21.9

Source : KLIPS data Wave6(2003) ~ Wave17(2014) Note: a. ‘\’ refers to Korean currency, won. b. This is the proportion of those who were employed in the concerned year and were paid in the previous year less than the minimum wage rate of the concerned year. c. The propotion is the percentage of those who were employed in the concerned year and whose yearly household income per capita calculated by the information collected one year later

was less than 60 per cent of its median for all employees.

181

Minimum Wage Effects on the Wage Distribution

The effects of the minimum wage on the distribution of wages among older employees

were examined with respect to three sub-questions, whether an increase in the minimum

wage creates a spike at the point of the minimum wage, whether the increase has a

spillover effect, raising wages of those who were already paid above a new minimum

wage rate, and whether the increase reduces the wage gap in the wage distribution

among older employees. A spike at the point of the minimum wage was graphically

assessed by plotting the kernel density estimates of log hourly wage minus log

minimum wage for each employee aged 55 and older for each year, where a zero

indicated that the employee earned the minimum wage rate of the year. The hourly wage

for each individual was derived with monthly pay and average weekly working hours

((monthly pay*10,000*12)/(weekly working hours*52.1)). Figure 7.1 shows that spikes

at or around the minimum wage were consistently found over the years in each five-year

period. This suggests that the minimum wage had some effects to raise the wages of

those who previously earned less than the minimum wage of the year unless the rate of

increase in the minimum wage was lower than the rate of increase in bargaining wages,

and if there was no clue as to a sharp rise in newly created jobs paying around the

minimum wage.

Figure 7.1 Wage Distribution of 55+ Employees, 2004-2008 and 2009-2013

Source: KLIPS data wave7(2004) ~ wave16(2013)