What are the effects of the new U.S. immigration policy on the economy?

AbilityDrain.pdf

Home >Business & Finance homework help >Economics homework help >What are the effects of the new U.S. immigration policy on the economy?

econstor Make Your Publications Visible.

A Service of

zbw Leibniz-Informationszentrum Wirtschaft Leibniz Information Centre for Economics

Schiff, Maurice

Working Paper

Ability Drain: Size, Impact, and Comparison with Brain Drain under Alternative Immigration Policies

GLO Discussion Paper, No. 62

Provided in Cooperation with: Global Labor Organization (GLO)

Suggested Citation: Schiff, Maurice (2017) : Ability Drain: Size, Impact, and Comparison with Brain Drain under Alternative Immigration Policies, GLO Discussion Paper, No. 62

This Version is available at: http://hdl.handle.net/10419/157308

Standard-Nutzungsbedingungen:

Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.

Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.

Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte.

Documents in EconStor may be saved and copied for your personal and scholarly purposes.

You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public.

If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence.

www.econstor.eu

Ability Drain: Size, Impact, and Comparison with Brain Drain under Alternative Immigration Policies*

Maurice Schiff a

January 2017

Abstract

Ability drain’s impact seems economically significant, with 30% of US Nobel laureates since 1906 being immigrants, and immigrants or their children founding 40% of Fortune 500 companies. Nonetheless, while brain drain and gain have been studied extensively, has not. I examine migration’s impact on ability , education

, and productive human capital or ‘skill’ , , for source country residents and migrants under a) the points system ( ) which accounts for , and b) the ‘vetting’ system ( ) which accounts for (e.g., US H-1B program). Findings are: i) Migration reduces (raises) residents’ (migrants’) average ability, with an ambiguous (positive) impact on average education and skill, and net skill drain, , likelier than net ; ii) these effects increase with ability’s inequality or variance, are greater under than PS, and hurt source countries; iii) the model and two empirical studies suggest that, for educated US immigrants, average , with real income about twice home country income; iv)

holds for any , and also for a very small (7.4% of our estimate). Policy implications are provided. Keywords: Migration, points system, vetting system, ability drain, brain drain, brain gain

JEL Code: F22, J24, J61, O15

* Thanks are due to Lant Pritchett, Hillel Rapoport, two anonymous referees, and participants at the 2016 Western Economic Association International meetings, the 2016 Conference of the Society of Government Economists and a World Bank seminar for their comments.

a: Fellow, Institute for the Study of Labor (IZA). Email: [email protected]; Tel: 202-338-8079; Address: 3299 K St, NW, Apt 501, Washington, DC 20007, USA

1. Introduction

While the brain drain literature in the 1970s saw it as hurting migrants’ source countries

(e.g., Bhagwati and Hamada 1974, Bhagwati 1976),1 studies in the last two decades have

found that it has a number of positive effects, including on brain gain and growth (e.g.,

Mountford 1997; Vidal 1998; Beine et al. 2001, 2008), fertility (Beine et al. 2013),

institutions (Docquier et al. 2016), and many more. Two excellent surveys of brain drain

issues are Commander et al. (2004) and Docquier and Rapoport (2012).

This paper focuses on the migration of educated individuals. The studies above that

looked at the brain drain’s impact on source countries’ average level of education by

comparing it with the brain gain implicitly assumed that educated migrants are identical

and ignored an important source of heterogeneity, namely innate ability. 2 The latter

includes the ability to learn, adapt, communicate, motivate, work in groups, and attributes

such as entrepreneurship, creativity, responsibility, ambition, intelligence, leadership,

work ethic, and more, and which affect individuals’ migration and education decisions

(see Sections 3-5). With developed countries’ higher return to ability, migrants are

positively selected for it (Schiff 2006).

Some migrants’ performance may be (below) average, while others may become great

scientists or great entrepreneurs. For instance, 30 percent of all US Nobel laureates since

1906 (and a greater percentage since 1950) were foreign-born. And over 40 percent of US

Fortune 500 companies were founded by immigrants or their children (Partnership for a

New American Economy, 2011). Given the difficulty in measuring ability, its economic

significance has not been ascertained to date, though these examples suggest that the

“ability drain” may be important. Moreover, that a brain drain generates a brain gain while

an ability drain does not, raises the latter’s relative importance.

1 Bhagwati and Rodriguez (1975) provide an overview of a collection of studies dealing with pre-1970s and 1970s’ brain drain theory, evidence and policy. 2 Some innate ability might be further developed later in life, though at a cost (which may well be prohibitive for those poorly endowed with it or with complementary ones). I abstract from this possibility to provide a sharp contrast between this study and previous ones which have typically excluded heterogeneous ability.

The model developed here examines the impact of the points system, “vetting” system and

“new” points system. The points system – e.g., Canada’s pre-2015 immigration policy –

accounts for prospective migrants’ education (and some other factors, such as age and

fluency in the host country’s language), while the “vetting” system – e.g., the US H1-B

visa program (when properly run; see Section 7) – also accounts for their ability. The

“new points system” – such as those in Australia, Canada and New Zealand – consist of a

hybrid of the points and vetting systems.

The model shows, among others, that ability and brain drains are larger under the vetting

than under the points system, and that both increase with ability’s heterogeneity, with a

greater increase under the former than the latter.3 Combining the model with a study of

the gains for US immigrants from 42 developing countries suggests that the magnitude of

the average ability drain, , for those with a college degree or more is about the same as

that of the brain drain ( 1.074 . This result, together with an empirical study of

the brain drain’s impact on average education, suggests that average productive human

capital or ‘skill’ – a combination of ability and education – falls with migration, a result

that even holds for an ability drain that is only 7.4 percent of the levels obtained.

No statistical confidence levels or significance tests are available at this stage. However,

note that this does not diminish the potential importance of the results obtained, for two

reasons. First, as mentioned above, the conclusion that the brain drain’s average impact on

productive human capital or skill is negative at all brain drain levels is robust in the sense

that it holds even if the ability drain, , is only 0.079 or 7.4 percent of the average

value obtained. Second, this is as far as I know the first study that has attempted to put

some numbers on the level of the ability drain. Given the paucity or lack of data on

migrants’ and non-migrants’ average ability, this attempt should be viewed as an initial

effort that will hopefully lead to further work on this issue.

Though the ability drain and its impact may matter for both source and host countries, I

have only found three studies that use a direct measure of ability, or of some element of it,

to examine its relationship with migration. Miguel and Hamory (2009) find a higher rural-

3 The importance of heterogeneous ability and schooling quality for the brain drain has been emphasized in Haque and Kim (1995) and Haque (2007).

urban migration rate in Kenya for individuals with higher cognitive skills, i.e., for those

who scored higher on a primary school test. 4 Kleven et al. (2010) show that the migration

response to changes in European countries’ taxation rates is greater for the more

successful football players, i.e., they are more responsive to changes in incentives. As for

attitudes toward risk, Akgüҫ et al. (2015) and Dustmann et al. (2015) find for rural China

that those who are least averse to taking risks and better able to do so are the most likely to

migrate. These studies’ findings that more able individuals are more likely to migrate is

incorporated in the model in Section 2.

Other studies that infer some aspect of ability’s relationship with migration include Özden

(2006) and Mattoo et al. (2008). These studies examine highly educated US immigrants’

success, i.e., whether their occupation is commensurate with their education level or

whether they are overeducated. One finding is that migration distance has a positive

impact on migrants’ degree of success. As the cost of the migration project rises with

distance, its expected return must increase to make migration worthwhile, i.e., migrants’

ability must increase with distance.

Given the potential importance of the relationship between migration and ability, the

paucity of studies on this issue is unfortunate. Except for Clemens, Montenegro and

Pritchett (CMP, 2009), which focuses on low-skilled migrants, none of the studies

examined ability drain or its impact. This paper contributes to this fledgling literature i) by

developing a model to examine migration’ impact on average ability and education for

both source countries’ residents and migrants; and ii) by combining the model, empirics

and data in order to obtain a measure of the relative importance of the ability and brain

drain, and educated migrants’ impact on productive human capital or ‘skill’, a function of

both ability and education.

Immigration policies vary across countries and time. Three of them are examined here.

Under the points system, a policy that prevailed in Australia, Canada and New Zealand,

prospective migrants obtain points according to their level of education (and other criteria,

4 Hanushek and Woessmann (2008, 2009) find that cognitive skills strongly impact individual income, its distribution and growth, and Heckman and Rubinstein (2001) and Heckman and Kautz (2012) find that non- cognitive skill are important as well. None of these studies deal with migration.

e.g., age and fluency in the host country’s language). Under the US H1-B visa program,

prospective migrants must obtain a job offer and have at least a Bachelor’s degree or

equivalent in order to be able to immigrate. I refer to this type of policy as the “vetting”

system, given that employers are likely to thoroughly vet potential employees since they

benefit from good hiring decisions and pay the cost of bad ones. With points systems

leading to unsatisfactory employment outcomes, a number of countries, including Canada,

Australia and New Zealand moved to a ‘new’ points system, a hybrid of the (old) points

system and the vetting system, thus giving more weight to the labor market demand side.

The paper is organized as follows. Section 2 presents the model and closed economy case.

Sections 3 and 4 examine (and compare) the points and vetting systems. Section 5 briefly

looks at the new points system, while Section 6 provides a comparison of the size of the

ability and brain drains. Section 7 assesses the sign of migration’s impact on productive

human capital or ‘skill’. Section 8 presents policy implications and Section 9 concludes.

2. Model

Assume individuals’ productive human capital or skill can be observed and valued

properly by employers in both countries. This makes sense since, as mentioned earlier,

employers benefit from good hiring decisions and pay the cost of bad ones, and are thus

likely to thoroughly vet prospective employees in order to assess their skill level.

Denote individual ’s ability by , the source country or country of origin (destination) by

“0” (“d”), source country residents’ (migrants’) income by ( , and the immigration

probability by 0,1 . Skill , income in both countries, and expected income , are:

, , , 0, ,

1 .5 (1)

5 I selected as simple a model as possible in order to obtain results that are clear and make intuitive sense. For instance, with , there are no interaction effects between ability and education. Nevertheless, the optimal value of rises with , with the exact relationship depending on the host country’s immigration policy (see Sections 3 to 5). One could also specify as . This would complicate the model without affecting the qualitative results – though it would lead to a greater negative (positive) impact on home country residents’ (migrants’) average ability and education.

The cost of education is /2. Thus, (expected) utility or consumption is:

0. 6 7 (2)

Individuals maximize expected utility by selecting , subject to their innate ability and

the host country’s immigration policy. For comparison purposes, is such that the source

country’s expected migration rate is identical under the three policies examined, i.e.,

, where , denotes ’s probability density

function, and denote the points (vetting) (new vetting) system. Gross average

ability, , is the source country’s average ability before migration takes

place. Individuals take into account the fact that the migration probability depends on

education under the points system, i.e., , and depends on both education and

ability under the vetting system, i.e., , .8 Given that source countries have

both migrant and non-migrants, interior solutions are assumed throughout.

2.1. Closed Economy

Before turning to the points and vetting systems, results are provided for the ‘closed

economy’ immigration policy. In that case, the migration probability 0. Denoting the

variables in this case with subscript “c”, equation (2) becomes:

0. (3)

6 A large number of empirical studies show that investment in education exhibits diminishing returns. Given that income is a linear function of education in (1), assuming a quadratic education cost function results in diminishing returns to education (with a negative second derivative of with respect to education). 7 Of the 42 sample countries in the empirical analysis (provided in Section 6 and the Appendix), 55 percent are either low-income or lower-middle-income countries (defined by the World Bank for 2017 as countries with a per capita income below US $4,036 in 2015) and about two thirds of the sample countries had a per capita income of US $5,000 or less in 2015. A quadratic education cost seems reasonable for those countries as a constraint is likely to prevail on the number and qualifications of individuals teaching students who are completing a bachelor’s degree or more – which is the level of education for which the relationship between ability drain and brain drain is derived (as shown in the Appendix). 8 Thus, average education and skill levels are higher for migrants than for residents, i.e., migrants are positively selected for both ability and education . As Docquier and Marfouk (2006) show for education, the share of the highly educated in South-North migrants is three times that among the South’s residents, and the ratio is larger for poor, landlocked and island countries (e.g., 15 for Sub-Saharan Africa).

Maximizing with respect to , the values for , its average , average ability ,

skill , average skill and its variance , consumption and its average , are:

, , , , ,

, . (4)

3. Points System

Under the points system (e.g., Canada’s pre-2015 policy), applicants receive points for

education but not for ability. The immigration probability is , to which a constant,

, is added to ensure the average immigration probability or average migration rate is

identical under the points and vetting systems, i.e., , which is assumed for

comparison purposes (as shown later).

The immigration probability and consumption in this case are:

, 0, 0. (5)

Defining ≡ 1 2 and ≡ , , , , , and are: given

by:

, 2 ,

, , (6)

where 0 is the second-order condition, , where denotes the

average ‘gross’ level of education, i.e., the level that includes the brain gain (i.e.,

) generated by the points system, but before migration takes place, i.e.,

excluding the brain drain.

As shown in (6), implies , as the former are a multiple (by )

of the latter. This implies in turn that (compare (6) and (13) in Section 5)

and . Also, the brain gain is given by 2 . Since

1 2 , we have

1 1 1 . Thus, the points system

raises the variance of individual skills or skill inequality, relative to the closed economy

case. This is also apparent from the derivatives 1 and 1 .

The host country’s policy change from a closed economy to a points system raises the

expected return on education, with an impact on residents’ education and skill

2 0. However, residents’ average skill need not increase

because education increases with ability, which raises the migration probability. Thus, the

migration rate is higher (lower) at higher (lower) ability and education levels, which

reduces both average ability and average education.

Denote a variable ’s population-weighted average value by ≡ 1

for source-country residents, by ≡ for migrants,

by ≡ 1 for all natives, and by for the gross

(pre-migration) average skill. Solutions for , and , , are:

, , ,

, , . (7)

As shown in (7), the brain drain is times the ability drain, i.e., , the reason

being that enters into with coefficient (see (6)). Since 1 (footnote 8), the

ability drain is larger than the brain drain. And from (6), , i.e., the

variance of is greater than that of . Results are also shown in Table 1 below.

Another result from (7) is that residents’ (migrants’) population-weighted average ability,

education and skill levels fall (rise) with inequality in the source country’s ability

distribution, as measured by the variance of . Thus, the host country benefits from

greater inequality in ability as it raises the average skill level of its immigrants. And, as

shown above, the policy itself also raises inequality in migrants’ source country. Finally,

the variance of does not affect natives’ average ability, education or skill as its impact

on residents’ and migrants’ values cancel each other out. Table 1 presents the impact of

the points system on the ability, brain and skill drain and on the brain and skill gain,

relative to the closed economy case.

What is the policy’s impact on ability, education and skill, relative to a closed economy

policy ( 0 , i.e., ∆ ≡ , , ? Since there is no ability gain, average

ability declines. Moreover, given that , , 2 , and

, we have:

Table 1: Points System– Source Country’s Ability, Education and Skill: Net Gain or Net Drain? a

Ability

(1)

Education

(2)

Skill

(1 2

Ratio

2 / 1

Drain (i) . .

Gain (ii) -- 2 2 --

Net Gain

(i) + (ii)

. 2 ⋛ 0. 2

⋛ 0.

Variance 1

a: Results are relative to the closed economy case.

∆ ≡ 0, ∆ ≡ 2 1

⋛ 0,

∆ ≡ 2 1

⋛ 0, ∆ ∆ . (8)

Thus, the policy’s impact on ability (education and skill) is negative (ambiguous). Since

∆ ∆ ∆ ∆ 1

∆ , the policy’s impact on skill is more

likely to be negative than its impact on education. Some studies (e.g., Beine et al. 2008)

find that a net brain gain is more likely in larger source countries (∆ 0), in which

case ∆ 0 ∆ is a distinct possibility. They also find that most countries exhibit a

net brain drain (∆ 0), implying a larger net skill drain, i.e., ∆ ∆ 0. On the

other hand, the points system raises migrants’ ability, education and skill, with:

∆ ≡ 0, ∆ ≡ 0 2 0 2

∆ ≡ 2 0. (9)

Results for source country’s natives as a whole (denoted by subscript ) is:

2 0, 0. (10)

In other words, natives’ average education and skill levels are higher under the points

system than under a closed economy, while their average ability is unchanged.

Finally, note from (7) that both the brain and ability drain vanish ( 0 under

homogenous ability, i.e., for 0.

4. Vetting System

I refer to an immigration policy that takes both ability and education into account as a

“vetting system,” with variables designated by subscript ‘v’. One such system is the US

H1-B visa program, where employers’ hiring decisions determine whether or not

immigration takes place.9 Probability and consumption under this policy are:

9 This assumes a well-functioning visa program, which is not necessarily the case. See Section 8 for more on this issue.

0 , 0. (11)

Maximizing with respect to , the solutions for , , , and are:

2 , , 0 ,

1 ⋛ 0 ⇔ ⋛ ,

⋛ 0 ⇔ ⋛ . (12)

Thus, high- (low-) ability individuals attain a higher (lower) education level and have a

higher immigration probability under the vetting than under the points system, resulting in

greater education and skill inequality (or variance) under the former, as shown next.

From (12), 0 1 2 , i.e., the vetting system results in greater

residents’ individual skill relative to the no-migration case. From (10), we have:

1 2 . Since 1 , it follows

that , with 1, ] for 0, .5 . Moreover,

4 4 . Thus, inequality of residents’ skill is greater under the vetting

system. The solution for and is:

2 , , 0 . (13)

Solutions for residents, migrants and natives’ average ability, education and skill, are:

, , ;

, , ,

(14)

As and 2 1

, it follows that / 1/2 and, with

, one would expect / 1.10 The results are summarized in Table 2.

Residents (migrants’) average education, ability and skill levels decline (increase) with

inequality in the ability distribution under both the points and vetting systems, though

migration’s quantitative impact is greater under the latter. Denote ability drain by

under the vetting (points) system. From (14),

, and . Hence,

1 . Similarly,

migrants’ education, ability and skill gains are greater under the vetting than under the

points system.

Note also that, though affects both and , it has no impact on / under

either the points or vetting systems (see Table 2). Comparing average levels of education,

ability and skill, for residents, migrants and all natives, under the two systems, we have:

, ,

, . (15)

From (14), and from and (as shown in (4)), it follows that whether

the vetting system results in a net brain and skill gain or drain is ambiguous, though net 10 Proof that : Residents’ actual (as opposed to expected) consumption under the vetting system is

2 2 0, and thus

4 0 2 ,

which reaches a maximum, at 0, so that and , i.e., . QED.

brain drain and skill drains are more likely under the vetting than under the points system,

and the ability drain is greater under the former. Since by construction, it

follows that natives as a whole have the same average ability, education and skill levels

under the points and vetting systems.

The host country benefits from greater inequality in the source country’s ability as it raises

migrants’ average skill level, and more so under the vetting than under the points system.

Moreover, the policies themselves raise inequality, and more so under the vetting system.

Table 2: Vetting System – Non-migrants’ Ability, Brain and Skill: Net Gain or Net Drain? a

Ability

(1)

Education

(2)

Skill

(1 2

Ratio

(2)/(1)

Drain (i) . . . 2 1

Gain (ii)

2 .

Net Gain

(i) + (ii)

. 2 ⋛ 0. 2

⋛ 0.

Variance 4 1 2 4 1

a: Results are relative to the closed economy case.

5. New Points System

In order to attract immigrants with skills that better reflect labor market needs, various

host countries, including Australia, New Zealand and Canada, moved to a new points

system (denoted by subscript n), consisting of a combination of the (old) supply-driven

points system and the demand-driven vetting system. Denoting the weights of the old

points system and vetting system in the new points system by and 1 , respectively,

with 1 , it can be shown that the solutions are equal to the weighted

average of the solutions under the points and vetting systems, except for the variance. The

weighted average of the variance of (for the points system) and (for the vetting

system) is greater than the variance of (for the new points system), and the same holds

for the skill variance.

6. Comparing Ability and Brain Drain

This section examines the relationship between the ability drain, , and the brain drain,

. Educated immigrants typically enter the US under the H1-B visa program, i.e., they

must obtain a job offer and must have a Bachelor’s Degree or more in order to qualify. As

discussed in the Introduction and in Section 4, they are likely to be thoroughly vetted with

regard to both their education and their ability, as employers obtain the benefits of

judicious hiring decisions and bear the cost of hiring mistakes. Thus, a vetting system

policy is assumed in the analysis.

The analysis is based on the model and on empirical results in CMP (2009) who use a

database on PPP-adjusted wages and other characteristics for two million individuals in

the US and 42 developing source countries, based on the US Census in 2000 and

household surveys in the 42 source countries for 2000 or close to it, in order to obtain

estimates of the impact of migration on migrants’ income.

CMP find that correcting for selection on observables – i.e., for the difference between

migrants’ and non-migrants’ education level – reduces migrants’ average income, relative

to their income in the country of origin, from 7.99 to 5.11 0.64 , or by

0.36 (see Appendix). They then use several approaches, based on microeconomic and

macroeconomic evidence, to obtain an estimate of parameter in order to capture the

impact of selection on unobservable ability on the “place premium,” i.e., on the ratio of

migrants’ income in the US to their income back home, , where / .

The value of obtained by CMP is for migrants with nine years of education. As CMP

mention, selection on unobservable traits – i.e., ability – is unlikely to be strong for

immigrants with low or moderate education. On the other hand, immigrants who enter the

US under the H1-B visa program must have at least a Bachelor’s degree or sixteen years

of education, and the model is used to adjust the value of in order to reflect this

difference. The ratio of ability drain to brain drain, / , is derived for the range of

parameter values in order to obtain an overall average for it.

The main results for the 42 source countries’ average values are presented in Section 6.1.

The relationship between source countries’ income and the ratio / derived from the

model, as well as its relationship with the brain drain derived from data on the 42 source

countries and the US, are presented in Section 6.2. Derivation of the results is provided in

the Appendix.

6.1. Average for the forty-two developing source countries

The main findings for the 42 countries as a whole are:

i) The average value of / is 1.074.

ii) US immigrants from developing source countries with at least a Bachelor’s degree raise

their income on average by slightly over 102 percent.

iii) Education is endogenous and determined by ability (equation (12)). And, irrespective

of their relative size, ability’s heterogeneity is the cause of both the ability and the brain

drain – as 0 under homogeneous ability (i.e., for 0 (see equation

(14) or Table 2).

The result that the ability drain is 1.074 times the brain drain under the vetting system is

consistent with the findings of the model that the ratio / 1/2 is greater than or

equal to one (see Table 2).

6.2. Country groupings

As shown in equation (1), the income of source country individual is , where

parameter reflects the level of technology, institutional development, etc. Recall that

under the vetting system, 1

and , with the ratio

/ 1/2 , and with defined as ≡ 0

0 1 2 0

. Thus,

and /

. 0, i.e., increases relative to as source

countries’ income increases.

A contributing factor would be a negative impact of source countries’ income on the brain

drain. Data on the correction of US immigrants’ income for selection on education, i.e.,

for the brain drain’s impact on their income, are available for each of the 41 countries.

They show that countries and regions with the largest BD impact – i.e., with the highest

selection on ability – tend to be poorer than those with the smallest impact. For instance,

the five countries with the largest BD impact are Ethiopia (.666), India (.662), Sri Lanka

(.612), Nepal (.600) and Uganda (.584), with an average impact of .612. The five countries

with the smallest BD impact are Chile (.030), Jamaica (.045), Mexico (.056), Peru (.117)

and Argentina (.130), with an average impact is .076. Thus, the former group’s BD is eight

times that of the latter.

The region with the greatest BD impact is South Asia (.556), followed by Sub-Saharan

Africa (.415), South-East Asia (.400), the Middle East and North Africa (.328), the

Caribbean (.276), Central America (.219) and South America (.176). The average impact

for Latin America and the Caribbean – which includes the latter three regions plus Mexico

– is .209. These results suggest that the brain drain declines with source countries and

regions’ income.

7. Net Skill Gain or Net Skill Drain?

Beine et al. (2003, 2008) estimate the impact of migration of college-educated individuals

on the average level of education in their country of origin, i.e., they estimate the net brain

gain, or difference between the brain gain and the brain drain. The third-order

polynomial reduced-form relationship between the net brain gain (measured as the change

in the domestic proportion of college graduates) and the brain drain, depicted in Figure 1

of their paper, is 0.0788 0.4587 0.02746 , with 0

for 0.18. Thus, is positive (negative) for the larger (smaller) countries.

They also find that ’s global average is positive.

With and 1.0742 for migrants with at least a Bachelor’s degree, we have

1.0742 . Thus, the net skill gain, 1.0742 .

Assuming Beine et al.’s result holds for the 42 source countries as well, it follows that

.9954 0.4587 0.02746 0,∀ ϵ 0,1 . In other words,

migration results in an average net skill drain or a loss in the average level of productive

human capital or skill. Thus, while Beine et al. (2003, 2008) found a positive net brain

gain for 0.18, once ability drain is accounted for, the change in the average skill

level is negative for any positive .

Finally, note that a coefficient of , 0.0788 , equal to zero is a sufficient condition

for 0,∀ 0,1 . Thus, 0.0788 is a sufficient condition for 0,

∀ 0,1 . In other words, an ability drain that is as small as 7.5 percent of the value

obtained in this paper ( 1.0742) results in a net skill drain, for any positive .

8. Policy implications

Studies of the brain drain have found that a number of countries, particularly the larger

ones, experience a net brain gain (e.g., Beine et al. 2008). As migrants are also positively

selected for ability, migration results in an ability drain. Thus, countries might exhibit a

net brain gain together with a net skill drain, with a net skill gain requiring on average a

brain gain greater than twice the brain drain. The situation is obviously worse for countries

experiencing a net brain drain – including especially small poor island countries – as they

typically exhibit a high brain drain in addition to the ability drain.

The vetting system, such as the US H1-B visa program, was shown to generate a larger

ability drain and a larger net brain drain (or a smaller net brain gain) than the points

system, thereby further raising the likelihood of a net skill drain. The fact that several

immigration countries, including Australia, New Zealand and Canada, reformed their

immigration policy from the old to a new points system that includes elements of the

vetting system, raises the urgency for source countries of devising market-friendly policies

to minimize the skill drain and collaborating with host countries in order to raise both

countries’ migration benefits (see below).

The model shows that host countries benefit more from the ‘vetting system’ than from the

new points systems (assuming that vetting systems such as the H1-B visa program are

implemented as was intended; see below) and more from the new points system than from

its early versions Given the problems associated with the points system and the reforms

undertaken in a number of host countries, it is somewhat surprising that this system was

considered in the US Senate’s 2013 Immigration Bill.

Host countries concerned with source countries’ development could provide H1-B visas or

other skilled immigrant visas whose extension or conversion to permanent status would

require applicants to make some contribution to their home country, such as imparting

their acquired knowledge to home country individuals – whether by working there for

some period of time, regular visits of shorter duration, teaching via the internet, through

some business relationship, or other.11 One possibility is for source country universities to

allow joint appointments with host country ones. Such a policy has been successfully

pursued by Israeli universities, where top scientists and other academics often hold

positions in both Israel and the US or Europe.

Similarly, foreign students from developing countries often receive financial support from

some public or private agency in their home country (e.g., government agency, private

employer, university) or in the host country (e.g., university, foundation). Source and host

countries could cooperate to ensure that foreign students who obtain their degree and

apply for an immigrant visa spend some time in the source country (which is the case for

foreign students who enter the US with a J-visa) or engage in some other form of

interaction, such as cooperation with research institutions and scientists back home,

teaching, or other, which is likely to benefit both countries. Moreover, as Spilimbergo

(2009) has shown, foreign students who return after studying in advanced democracies

have a positive impact on democratic institutions in their home country, an outcome that

might also arise in the case of increased interaction between foreign graduates and their

home country.12

11 In addition to generating a direct benefit for migrants’ home country, such interaction would also likely raise bilateral trade and investment over and above the increases found in existing studies because of further reductions in information and transactions costs, thereby benefiting both countries (see Parsons and Winters’ (2014) excellent survey on migration’s impact on bilateral trade, and Javorcik et al. (2009) and Kugler and Rapoport (2007) on migration’s impact on bilateral investment). 12 As for agreements on expanded market access commitments for services, such as those delivered through the temporary cross-border movement of natural persons (a.k.a. Mode IV), would benefit both source and host countries – with the former supplying labor services in, say, construction, cleaning and hospitality, and the latter supplying, say, banking, insurance, and ICT. Such arrangements would reduce some of the

A more fundamental issue is that the market for talent is a global one and incentives

provided by the international market for the most talented are likely to be too powerful for

developing country governments to counter through some unilateral migration restrictions.

Retaining talented individuals and encouraging those who work or study abroad to return,

as well as competing in the global talent market, would require a wholesale change in

policy that goes way beyond migration policy. It is likely to require improvements in

governance, merit-based pay in countries where the public sector employs a large share of

the labor force and the wage distribution is compressed, provision of research facilities

and labs in order for professionals to maintain their skills and keep up with advances

elsewhere, appointment of the most talented rather than political appointees to head these

research facilities and labs as well as university and other education programs, liberalizing

domestic markets and trade, and more.13

As for host countries, some would benefit from enforcing existing skilled migration

policies. For instance, under the US H1-B visa program, skilled immigrants can be hired

for positions for which no Americans are available. However, as has been widely reported,

the policy has been ‘captured’ by a few large outsourcing firms that apply and obtain a

large share of the available visas, enabling some large corporations to replace US

professionals with younger and cheaper immigrants.14 These immigrants are typically less

experienced than the ones smaller enterprises that need someone with unique skills want

to hire but are unable to do so. In other words, the quality of H1-B immigrants is most

likely lower and they are likely to be closer substitutes to US natives – both of which are

less beneficial – than if the visa program worked properly. Moreover, many of the US

professionals who are being replaced end up in positions that pay less, another cost for the

concerns related to the brain and ability drain associated with permanent migration. So far, though, both sets of countries have limited the access to this mode of trading services. 13 Haque (2007) argues that human capital should be thought of exactly as financial capital, where the return of flight capital depends in large measure on the policies implemented by the country of origin. 14 A notable example is Southern California Edison, which replaced its IT employees with younger ones brought in through the H-1B program, with the original employees forced to train their replacements and sign nondisclosure agreements and gag orders. Salaries fell from $110,000 to $70,000 a year on average or by 36% (based on depositions in a Senate Judiciary Committee hearing spurred by complaints of the practice).

US. Avoiding such negative effects requires stricter enforcement of the rules of the H1-B

visa program.

9. Conclusion

A large number of studies have examined the determinants of migration and its impact on

education in source and host countries but have not done so in the case of ability. This

paper is an attempt to start filling this gap.

Based on the model and two empirical studies (CMP 2009; Beine et al. 2008), I find that:

i) The vetting system results in an average ability drain equal to 1.074 times the

average brain drain for individuals with at least a college degree (and an additional

year of education on average). Thus, the loss of average ability is slightly larger than

the change in average education in the 42 developing source countries (CMP 2009)

and can be negative even when the change in average education is positive.

ii) In fact, Beine et al.’s (2008) results imply that skilled migration reduces the source

country’s average skill level, including when the ability drain is only 7.4 percent of the

level obtained in this paper.

iii) Skilled migration results in an increase in average ability, education and skill levels

for source countries’ migrants.

iv) Source countries’ ability, brain and skill drains increase with inequality, as

measured by the variance of ability (and thus also by that of education). And the

policies’ positive impact on migrants’ average ability, education and skill also

increases with the variance in ability (and education). These effects are larger under

the vetting than under the points system. Thus, a host country obtains a greater benefit

from a vetting than from a points system, and from greater inequality in source

countries’ ability.

The paper’s findings suggest that the ability drain is likely to be important. Thus, policy

research and policymaking should focus on both education and ability, recognizing that

migration is likely to result in an ability drain and thus in a productivity loss, a loss that

must be the productivity change associated with a net brain drain or gain in order to obtain

a correct estimate of the impact of migration on income and growth. This issue has

essentially been ignored in the literature and in the policy debate.

Migration is also likely to result in higher average ability and education levels for

migrants because of its impact on the incentive for individuals to acquire more education

and because migrants comprise higher shares of high than low-ability individuals.

References

Akgüҫ Mehtap, Liu Xingfei, Tani Massimiliano, Zimmermann Klaus F. (2015) Risk Attitudes and Migration. IZA Discussion Paper No. 9347 (September)

Beine Michel, Docquier Frédéric, Rapoport Hillel (2001) Brain drain and economic growth: theory and evidence. Journal of Development Economics 64: 275-89

Beine Michel, Docquier Frédéric, Rapoport Hillel (2003) Brain Drain and LDCs Growth: Winners and Losers. IZA Discussion Paper No. 819, July.

Beine Michel, Docquier Frédéric, Rapoport Hillel (2008) Brain drain and human capital formation in developing countries: winners and losers. Economic Journal 118: 631-652. Beine Michel, Docquier Frédéric, Schiff Maurice (2013) International Migration, Transfer of Norms and Home Country Fertility. Canadian Journal of Economics 46 (4): 1406-1430.

Bhagwati Jagdish N. (1976) Taxing the Brain Drain. Challenge 19 (3): 34 – 38. Bhagwati Jagdish N., Hamada Koishi (1974) The Brain Drain, International

Integration of Markets for Professionals and Unemployment. A Theoretical Analysis. Journal of Development Economics 1: 19-42

Bhagwati Jagdish N., Rodriguez Carlos (1975) Welfare-Theoretical Analyses of the Brain Drain. Journal of Development Economics 2: 195-221.

Borjas, George J. (1989) Economic Theory and International Migration. International Migration Review 23 (3), Special Anniversary Issue. International Migration: An Assessment for the ’90s: 457-85 (Autumn).

Clemens Michael, Montenegro Claudio, Pritchett Lant (2009) The Place Premium: Wage Differences for Identical Workers Across the US Border. World Bank Policy Research Series Paper No. 4.

Commander Simon, Kangasniemi M., Winters L. Alan (2004) The Brain Drain: Curse or Boon? A Survey of the Literature. In: Baldwin R.E. and Winters L.A. (eds.). Challenges to Globalization: Analyzing the Economics. Univ. of Chicago Press, Chicago.

Docquier Frédéric and Marfouk Abdeslam (2006) International Migration by Education Attainment, 1990–2000. In: Özden Çaglar and Schiff Maurice (eds.). International Migration, Remittances & the Brain Drain, Palgrave Macmillan, New York, pp. 151–200.

Docquier Frédéric, Lodigiani Elisabetta, Rapoport Hillel, Schiff Maurice (2016) Emigration and Democracy. Journal of Development Economics 120 (May): 209-23.

Docquier Frédéric, Rapoport Hillel (2012) Globalization, Brain Drain and Development. Journal of Economic Literature 50 (3): 681-730.

Dustmann Christian, Fasani Francesco, Meng Xin, Minale Luigi (2015) Risk Attitudes and Household Migration Decisions. IZA Discussion Paper.

Haque, Nadeem Ul (2007) Brain Drain or Human Capital Flight. Lectures in Development Economics 11, Pakistan Institute of Development Economics.

_______, Kim Se-Jik (1995) Human Capital Flight: Impact of Migration on Income and Growth. IMF Staff Papers 42 (4): 577-607 (September).

Hanushek, Eric A., Woessmann Ludger. 2008. The Role of Cognitive Skills in Economic Development. Journal of Economic Literature 46 (3): 607– 668.

_______, ______. 2012. Schooling, cognitive skills, and the Latin American growth puzzle. Journal of Development Economics 99 (2): 497–512 (November).

Heckman James J., Kautz T. 2012. Hard Evidence on Soft Skills. Labour Economics 19 (4): 451-64.

_______, Rubinstein Yona (2001) The Importance of Noncognitive Skills: Lessons from the GED Testing Program. American Economic Review 91 (2): 145-49.

Isaksson, Anders (2007) World Productivity Database: An Overview. Research and Statistics Branch Working Paper 10/2007, UNIDO.

Javorcik, Beata S., Özden Çaglar, Spatareanu Mariana, Neagu Cristina (2011) Migrant networks and foreign direct investment. Journal of Development Economics 94 (2): 231-241.

Kleven Henrik, Landais Camille, Saez Emmanuel (2010) Taxation and International Migration of Superstars: Evidence from the European Football Market. NBER Working Paper No. 16545 (November).

Kugler Maurice, Rapoport Hillel (2007) International labor and capital flows: Complements or substitutes? Economics Letters 94 (2): 155-162.

Mattoo Aaditya, Özden Çaglar, Neagu Cristina (2008) Brain waste? Educated immigrants in the US labor market. Journal of Development Economics 87 (2): 255-69.

Miguel Edward, Hamory Joan (2009) Individual Ability and Selection into Migration in Kenya. Human Development Reports Research Paper 45. United Nations Development Programme.

Mountford Andrew (1997) Can a brain drain be good for growth in the source economy? Journal of Development Economics 53(2): 287-303.

Özden Çaglar (2006) “Educated Migrants: Is There Brain Waste?” In Özden C. and Schiff M. (eds.), op. cit., pp. 227-44

Parsons Christopher, Winters L. Alan (2014) Migration, Trade and Aid: A Survey. In: Lucas Robert E. B. (ed.). International Handbook on Migration and Economic Development, Edward Elgar: Cheltenham.

Partnership for a New American Economy (2011) The “New American” Fortune 500 (June).

Rothwell Jonathan, Ruiz Neil G (2013) H-1B Visas and the STEM Shortage: A Research Brief. Brookings, May 11.

Spilimbergo Antonio (2009) Democracy and Foreign Education. American Economic Review 99(1): 528-43.

Vidal J-P. 1998. The Effect of Emigration on Human Capital Formation. Journal of Population Economics 11 (4): 589-600.

Appendix

Clemens, Montenegro and Pritchett (2009) – referred to as CMP – use data on 41

developing source countries and the US, and examine various income ratios (denoted

here by lower-case letters), where the denominator is the average income of source

country residents, , where is unobservable ability and is

observable education. CMP’s objective is to obtain the average income ratio, / ,

of migrants living in the US who were educated in their home country, relative to the

income of home country residents with the same and levels, i.e., /

/ / . The problem with such comparisons is of course that

migrants self-select on both ability and education, whose levels are denoted by and

, respectively, and that their observed income is rather

than .

CMP find that for the 42 source countries, migrants’ average income ratio

/ / 7.99, i.e., migrants’ average income is 7.99

times that of source country residents. They first correct for migrants’ self-selection

with respect to observable in order to obtain / /

where, from equation (14), ∆ ≡ and, from equation (1),

∆ / . They find that 5.11 .64 for the 42 countries. Thus,

correcting for selection on observables (i.e., education) reduces migrants’ income

by ∆

2.88 0.36 , or a reduction in migrants’ relative income of 36 percent.

US immigrants and home-country residents may also differ in terms of non-observable

characteristics associated with migrants’ self-selection on ability. CMP correct for

migrants’ self-selection on ability, replacing / by /

, where ∆ ≡ , and ∆ / .

Two conditions make it possible to obtain the value of the ability drain, , brain drain,

, and their relative size, / , from the relationship between and :

i) the relationship between and is identical to that between ∆ and ∆ . From (14),

∆ ≡ , so that

∗ ∆ .15 Similarly, ∗ ∆ . Thus, / ∆ /∆ ; and

ii) ∆ and ∆ are multiplied by the same parameter, , to obtain the income changes

associated with the vetting system, so that / can be obtained from the difference

between relative incomes and .

CMP use various methods, based on both macroeconomic and microeconomic evidence,

to obtain an estimate of the impact on migrants’ average income of selection on (non-

observable) ability, , in ⁄ , where 1. They conclude that the degree of

positive selection on unobserved wage determinants results in a bias, , between 1.0 (no

bias) and 1.45 in the case of Peru, i.e., 1.0,1.45 . The average ability drain obtained

over these -values is 1.0742 , as shown below.

CMP obtained the range of values for workers with 9 years of education and state that selection on ability for less-educated workers is likely to be attenuated by the fact that they

tend to work in occupations without plausibly high returns to unobserved skill, a result

confirmed by the model.

Recalling that 1 represents 20 years of education, it follows that 0.45 for nine

years of education. From (12), we have 2 , or . For

individuals with nine years of education, we have 0.45 . Migrants who

enter the US via the H1-B visa program must have at least a Bachelor’s degree or a

15 Migrants’ average ability ‘gain’ is 1 / * and their average education ‘gain’

∆ 2 1 / * .

minimum of 16 years of education, i.e., a level of equal to 0.8 or higher, and the

equation for and the correction for selection on ability must take the difference in

education levels into account. Rothwell and Ruiz (2013) report that 90 percent of US

companies’ H-1B applications are for occupations that require high-level STEM (i.e., high-

level science, technology, engineering and math) knowledge. These typically require a

graduate degree or equivalent, which takes at least one year and often two years to

complete.16 Thus, it seems reasonable to assume that H-1B immigrants average one to two

more years of education. Assuming conservatively that they have one more year of

education implies that 0.85. Then, 0.85 , and the correction for

selection on ability, , becomes ′ 0.85 0 0.45 0

. Thus, for these individuals,

′⁄ and ability drain’s impact is 1 ⁄ 0.64 1 ⁄ .

With the correction for self-selection on education equal to 0.36 , we have:

. 1 ⁄ 1.778 1 ⁄ , ′

0.85 0 0.45 0

. (A.1)

Probability 1, with 1

, ∀ 0, , 0,1 , implying that

. Individual education is 2 0 2 0 1 2 0

1. Define ≡

0, so that 0 2 1 2

1, or 1 2 1 , ∀ 0, , i.e.,

1 2

. With 1, we have 1 2 1 2 . Thus,

I proceed now to ‘guess’ a solution for ⁄ , namely ⁄ 2 (this is verified below).

Then, from

, it follows that .2 and .4. Thus, .2 and

1 2 1 .4 . From and 1, we have .8 or 0, .8 .

16 This is consistent with their finding that H-1B visa holders earned on average 13.5 percent more than US native-born workers with a Bachelor’s degree.

I verify now whether the ‘guess’ that ⁄ 2 is correct. The average ability drain

relative to the brain drain, / , is obtained by averaging the / values obtained

for 1,1.45 and 0, .8 . For instance, take .2 and 1.25. Then, 1

.2 ∗ .4 0.92. Substituting the values for and into equation (A.1), the ratio /

1.0301, which means that the ability drain is 3 percent larger than the average brain drain.

The average value is / 1.0742. Thus, 1.0742 .3867 3.09, and

the impact of selection on unobservable traits is to reduce 5.11 by 3.09, so that

2.02 (or 1 percent above 2). Thus, developing country natives with a Bachelor’s

degree (or more) who migrate to the US would be expected to earn on average about twice

the real income they earned in their country of origin.17

17 Given that prices are typically lower in the home country, US immigrants gain more than 100 percent of the income earned back home if part of their income is transferred back home – say, through remittances – and consumed by family members there.