Economics

WHY GROWTH MATTERS

How Economic Growth in India Reduced Poverty and the Lessons for Other Developing Countries

__________________

JAGDISH BHAGWATI ARVIND PANAGARIYA

Ä Council on Foreign Relations Book

PublicAffairs

New York

Chapter 3

Reforms and Their Impact on Growth and Poverty

The economy ıs in cnsis... We are determined to address the problems of the economy in a decisive manner.... This government is committed to removing the cobwebs that come in the way of rapid industnalization. We will work towards making India internationally competitive, taking lull advantage of modem science and technology and opportunities offered by the evolving global economy.

We also welcome foreign direct investment so as to accelerate the tempo of development, upgrade our technologies and to promote our exports. Obstacles that come in the way of allocating foreign investment on a sizable scale will be removed. A time bound program will be worked out to streamline our industnal policies and program to achieve the goal of a vibrant econ omy that rewards creativity, enterprise and innovativeness....

Our vision is to create employment, eradicate poverty and reduce inequality. We want social harmony and communal amity. We want a more humane society As the twentieth century draws to a close, we cannot live with poverty and destitution among large sections ofour population. [Mahatma] Gandhi said that it was his ambition to wipe every tear from every eye. That is the vision which will inspire the work ofrny government Jai Htnd.

—Prime Minister P. V. Narasimha Rao in an address Io

the nation upon taking the office, June 22,1991

I I

2» WH'l GROWTH MATTERS

By the second half of che 1970s, the sorry plight of the economy, accentuated dramatically with the turn co socialist policies under Indira Gandhi, had become evident to those who did not wear ideological blinkers.

But the grip of socialist rhetoric on the national psyche was so overpowering chat few dared to challenge die policy framework itself. Therefore, the response was a very gradual, almost imperceptible process of unwinding the controls without disturbing the underlying framework.1

This process accelerated somewhat under Prime Minister Rajiv Gandhi, who took the reins of the government at the end of 1984 following the assassination of his mother, Indira Gandhi.

Through the 1980s, doubts concerning the controlled regime grew steadily even if only gradually. But as the decade ended, the doubts were greatly heightened and confidence İn the model India had followed was at a low point. It was further shaken by two external events: the success China had achieved after it turned outward—this was “learning by others’ doing”— and che demise of the Soviet Union—which was “learning by others’ undoing”—that had served as the model for Indian planners from the 1950s until at least the end of the 1970s. Therefore, when a balance-of-payments crisis in 1991 offered the opportunity to make more dramatic changes, newly elected Prime Minister P. V. Narasimha Rao, who came to the helm due to the assassination of Rajiv Gandhi during the election campaign and who had experienced the tyranny of central controls as the chief minister of Andhra Pradesh İn the early 1970s, did not hesitate.2 He launched a process of systemic reforms that firmly put India on a dramatically different course from the one Indira Gandhi had set for the nation.

These reforms posed the greatest challenge to the long-standing proponents of socialism, who now faced an existential threat. Unsurprisingly, their response was to take an offensive strategy designed specifically to undermine the credibility of the reforms. This required the creation of several new defensive myths. In this chapter, we address these myths, which relate to the alleged malign impact of the reforms on growth, poverty, and socially disadvantaged groups. Additional

Reforms and Thetr Impact on Growth nml Porerty 29 myths relating to the presumed deleterious impact of the reforms on inequality, education, health, and related issues are discussed in Part IL

Myth 3.1 : Reforms do not explain the faster growth in India since 1991.

The most surprising myth, surviving among a few economists, is that although growth did occur after reforms, it was not a result of the post- 1991 reforms and that instead it can be traced back to the 1980s.

• have seen that the command-and-control regime had peaked by the mid-1970s and a quiet process of loosening some of the controls began soon after, accelerating somewhat in the 1980s, especially under Prime Minister Rajiv Gandhi. It was this halting, partial process of reforms, introduced as it were “by stealth," that was replaced in 1991 by the reforms package that brought the liberalizing process into the open, made reforms more comprehensive across important issue areas, such as industrial licensing, and represented a fundamental shift in the policy framework.3
• we must ask: How can serious economists maintain that the reforms had no effect on the post-1991 acceleration of growth? Their argument takes one of two forms. They claim that first, the growth acceleration really started in the 1980s, and second, even that was a result not of the piecemeal reforms of the 1980s, but of “attitudinal changes,” which trumped the effects of any specific, concrete reform measures.

The economic historian Bradford DeLong (2003) was the first to argue that the post-1991 reforms followed rather than preceded the growth acceleration. But his claim that growth acceleration started in the 1980s did not automatically imply that reforms as such had nothing to do with the shift in the growth rate. Indeed, DeLong acknowledged that the reforms in the 1980s may have led to the acceleration in growth in that decade, and so one could not conclude that reforms and growth were not related. Going a step further, he also speculated that die 1980s growdi acceleration might have proven to be just “a short-lived flash in the pan" in the absence of more comprehensive reforms of the post-1991 variety (as is indeed true).

30 I WHY GROWTH MATTERS

The real problem lay with the separate assertion by Dani Rodrik (2003), in his editorial introduction to the volume carrying DeLong’s paper, that the 1980s growth had little to do with reforms in any case. H e argued that “ the change in official attitudes in the 1980s, towards encouraging rather than discouraging entrepreneurial activities and integration into the world economy, and a belief that the rules of the economic game had changed for good, may have had a bigger impact on growth than any specific policy reforms" (emphasis added)/

However, both DeLong’s statistical assertion about allegedly robust pre-1991 growth and Rodrik’s explanation for it are wrong.5

First, the claim that growth İn the 1990s was no higher chan in the 1980s carries what might be called a fallacy of aggregation. The acceleration in the early 1980s was in fact quite modest (see Figure 3.1), with the bulk of the growth back-loaded in the last three years of the decade. Once we exclude the years 1988-1991, growth in the remaining years—1980-1981 to 1987-1988—turns out to be just 4.6 per

Rejbrms and Their Impact on Growth and Poverty I 31

cent, which is closer to the 4.1 percent growth that had already been achieved between 1951-1952 and 1964—1965 and perceptibly lower than the 5.8 percent during 1988-2003 or 6.3 percent between 1992— 1993 and 1999-2000.6

Chetan Ghate and Stephen Wright (2008) have recently applied state-of-the-art techniques to detailed state- and industry-level data to identify the turning point of the economy. The authors’ careful and comprehensive work places this turning point at fiscal year 1987-1988, just as Panagariya’s (2004b) did.7

Second, the “super-high” annual growth of 7.2 percent during 1988- 1991 was preceded by significant, though partial reforms, especially in 1985-1986 and 1986-1987. It was also helped by significant deprecia- don of the rupee in the second half of the 1980s.8 But more important, this growth was also driven by fiscal expansion and external borrowing that were not sustainable. Unsurprisingly, the surge ended in a balance- of-payments crisis in June 1991.

Even if we ignore the differences in growth rates in the 1980s and 1990s, the 1980s growth could not have been sustained without the post-1991 reforms.

Third, the shift to 8.2 percent growth during the nine years between 2003-2004 and 2011-2012 represents a significant jump in the growth rate following the post-1991 systematic reforms. Surely attributing this latest acceleration to some vague “attitudinal” change in the 1980s strains credulity.

In feet, many of the structural changes since 1991 have a direct link to the liberalizing reforms. Could the trade-to-GDP ratio have risen from 17 percent İn 1990-1991 to more than 50 percent by the later 2000s without steady trade liberalization?9 Could foreign investment have risen from $100 million in 1990-1991 to more than $60 billion in 2007- 2008 without the liberalization of the foreign investment regime? Could the number of phones have risen from 5 million total at the end of 1990- 1991 to new additions of over 15 million every month without the liberalization of telecommunications? Could automobile production have risen from 180,000 in 1990-1991 to 2 million in 2009-2010 without delicensing of investment and opening up to foreign investment? The list

32 I WHY GROWTH MATTERS

goes on.10 Policies matter; “changes in bureaucratic attitudes” in the absence of policy changes are ephemeral.

Myth 3.2: There has been no reduction in poverty after the reforms.

Critics also complain that the reforms have done precious little for the poor. The complaint has taken different forms depending on the time and context. Initially rhe claim was simply that reforms had not helped bring down poverty. But as the reforms and growth progressed through the 1990s and 2000s and evidence on poverty reduction accumulated, these claims shifted first to asserting that reforms had not led to acceleration in poverty reduction from that in the pre-reform period and then to arguing that they had failed to reduce the absolute number of poor. We address each of these claims below.

Empirical evidence showing that poverty failed to decline during the heyday of socialism and that it has seen a steady decline in the postreform era is now incontrovertible. Figure 3.2 shows the evolution of

I —Poverty ratio —"Linear (Poverty ratio) |

Figure 3.2. The proportion of the population living below the official poverty line, 1951-1952 to 1973-1974

Source: Authors’ construction based on estimates in Dah, Guaiov, 1998. “Poverty in India and Indian States: An Update,” Discussion Paper No. 47, International Food Policy Research Institue (July)

Reforms aiul Their Impact on Growth and Poverty I 33

the proportion of those below the official poverty line, called the poverty ratio, at the national level from 1951-1952 to 1973-1974. The figure also includes the trend line for the poverty ratio during this period. It is evident from the figure that poverty fluctuated between 50 percent and 60 percent during this period with a slight upward trend. Because India had begun at a very low per capita income and grew at a very slow pace during the first twenty-five years of the development program, the country could make no dent in poverty whatsoever.

Figure 3.3, which shows the poverty ratio in rural and urban India and the country as a whole at various points between 1977-1978 and 2009-2010 presents a sharp contrast to Figure 3.2." Once the stranglehold of controls was loosened and reforms took root, growth accelerated and poverty fell in both rural and urban India and nationally. One can argue about the level of poverty since it depends on where precisely we draw the poverty line, but one cannot argue about the declining trend in it. Indeed, official estimates for 1993-1994, 2004—2005, and 2009- 2010 are at a higher poverty line and these also show a declining trend in both rural and urban India.

Recognizing the compelling nature of the evidence on the decline in poverty under reforms and accelerated growth, critics have shifted

J4 i WHÏ GROWTH MATTERS

ground. They now argue that the decline during the post-reform era has not accelerated and even slowed down relative to the pre-reform era. For this, they compare the annual percentage-point decline between 1983 and 1993-1994 and drat between 1993-1994 and 2004-2005. But this argument has two serious problems.

First, true comparison of pre- and post-reform performance İs not the one that critics draw but rather the one between the first three decades, spanning 1950-1980, and the past three decades, 1980-2010. Even though liberalizing reforms İn India became more systematic in 1991, they were already under way during the 1980s.

Second, and a tough nut for the critics to crack, is the poverty estimate based on the latest expenditure survey, conducted in 2009-2010, which captures the effect of the 8.7 percent growth during 2005 2006 to 2009—2010. Tliis estimate shows that poverty during these five years fell at a rate far exceeding that in any other period. Indeed, the acceleration during this period is so large that when we combine it with the earlier period and compute the average percentage-point decline per year during 1993 1994 to 2009-2010, this number turns out to be larger than during 1983 to 1993-1994, the period critics favor the most?-

But some critics reject the evidence of declining poverty by arguing that there has not been a major dent in the absolute number of people living below the poverty line and that this fact undermines the claim that poverty fell after the reforms. According to the official estimates provided by the Planning Commission, at the traditional poverty line, the India-wide absolute number of poor was 323 million in 1983, 320 million in 1993-1994, and 302 million in 2004—2005. Presumably, therefore, the number of poor has at best declined marginally?5

However, the practice, once popularized by the World Bank, of citing the absolute number of poor makes for a flawed method of measuring poverty in the face of a rising population?"1 This approach to measuring poverty will in fact downplay the decline in poverty because it does not distinguish between changes in the absolute and in the relative or proportionate number of poor.

The use of absolute number of poor biases the poverty measure upward. (One might cynically speculate that the biased measure was pop-

Rejoritis anil Their Impact on Growth ami Porertv 35

ular at the World Bank and diffused to the client nations, so as to increase die alarm over poverty and bolster the critiques of a reforms-ori- ented development strategy.)'5

Shifting, however, to the “proportionate” measure of poverty, we get a more meaningful idea of what happened in India. According to official poverty estimates, provided by the Planning Commission, the proportion of the population below the poverty line in India fell from 44.5 percent in 1983 to 27.5 percent in 2004—2005. During the same period, population rose by approximately 374 million. Unless you make the ludicrous assumption that the entire net addition of 374 million to the population was non-poor, a reasonable expectation would be that in the absence of an effective poverty-alleviation strategy, the poor would have been 44.5 percent of the additional 374 million people, or a further 166.5 million poor people?6 Adding this number to the original 323 million poor in 1983, this would have meant a total of 489.5 million poor people İn 2004-2005. But since the actual number of poor people in 2004-2005 was only 302 million, this suggests the exit of 187.5 million people from poverty.

This substantial decline is properly captured by what economists call the “poverty ratio,” which fell to 27.5 percent in 2004—2005 from 44.5 percent in 1983. This clearly demonstrates the absurdity of rhe contention that the unchanged absolute number of poor people is equivalent to no change in poverty.

A final form in which some critics make the argument that the reforms have not helped the poor is by citing continuing or even increased poverty among certain individuals or groups of individuals. But this will not work either since we discuss below that poverty has gone down among all broad-based groups, such as the Scheduled Castes and Scheduled Tribes, and across all states.

There is no doubt that there are many who were poor prior to the reforms and who remain so today. Nor can we rule out the possibility that reforms have impoverished some individuals, for example, when they are displaced from land to make way for alternative activities without proper compensation?' However, this approach to the criticism ofreforms is ill conceived since we can measure the efficacy of a set of policies only at some

36 I WHY GROWTH MATTERS

aggregate level, even when we disaggregate rhe effects by groups such as the Scheduled Castes and Scheduled Tribes (the impact on whom we examine in Myth 3.3 below). We know of no policy that makes everyone within each disaggregated group better off and hurts literally no one.

The reality for the critics of the reforms is stark: in regard to poverty reduction, the regime of socialist policies did far less good for (and indeed even inflicted harm on) the poor and the underprivileged groups than turned out to be the case under the regime of reformed policies. India ultimately moved away from the old policies precisely because those policies, with their deleterious effect on economic performance, had failed to deliver on poverty reduction and other social goals.

Myth 3.3: Reforms have bypassed, even hurt, the socially disadvantaged groups.

Some NGOs and journalists as well as international organizations argue that the reform-led growth may have reduced poverty overall but it has not helped bring down poverty among the socially disadvantaged groups, principally the Scheduled Castes and Scheduled Tribes but possibly also Other Backward Castes (OBC).

For example, a submission by the National Campaign on Dalit Human Rights to the House of Commons of the UK Parliament, published on January 14, 2011, stares, “In spite of high economic growth rate the poverty rate among excluded communities in India has increased, coupled with the insecurity of livelihoods.””

In a similar vein, writing in the Financial Chronicle (December 29, 2010), journalist Praful Bidwai argues, “Rising inequalities highlight what is wrong with India's growth trajectory, driven as it is by elite consumption and sectoral imbalances, which exclude disadvantaged groups from the benefit of rising GDP, while aggravating income disparities” (emphasis added).”

Most strikingly, even the World Bank has joined the chorus alleging that growth in India has done precious little for the tribal population of the country. A recent country brief by the bank puts the matter in these stark terms: “India is widely considered a success story in terms of

Re/anns tied The» hiipuct an Growth und Poverty l 3’

growth and poverty reduction. In just over two decades, national poverty rates have fallen by more than 20 percentage points, from 45.6 percent in 1983 to 27.5 percent in 2004-05. However it is widely acknowledged that growth has not touched everyone equitably and that many groups are left behind amid improving living standards. Among them are tribal groups identified by the Constitution as Scheduled Tribes.”2U

Once again, there is now irrefutable evidence that sustained growth alongside liberalizing reforms has reduced poverty not just among the better-off castes but across all broadly defined groups. It is true that the poverty ratios were and still remain significantly higher among the disadvantaged groups, reflecting historical injustices, but it is not true that these groups have not benefited from the recent growth. Indeed, evidence along all dimensions shows the Scheduled Castes and Scheduled Tribes gaining alongside the OBC and “forward” castes.

Mukim and Panagariya (2013) strikingly show that poverty fell for the Scheduled Castes and die Scheduled Tribes between every pair of successive surveys in rural and urban areas (see Figure 3.4). For the Scheduled Castes, the nationwide poverty ratio fell from 58.5 percent in 1983 to 48.9 percent in 1993-1994, 38 percent in 2004—2005, and 28.6 percent in 2009-2010. For the Scheduled Tribes, the ratio fell from 64.4 percent in 1983 to 51.2 percent in 1993-1994, co 46.3 percent in 2004-2005, and to 30.7 percent in 2009-2010. The authors also calculate the poverty ratios by states and find that they fell in all major states between 1983 and 2009-2010.

An extremely important recent development is the significantly larger decline in poverty among the Scheduled Castes and Scheduled Tribes relative to non-scheduled castes during the latest high-growth phase. Poverty for the Scheduled Castes fell by 9.4 percentage points and that for the Scheduled Tribes by a gigantic 15.3 percentage points relative to 6 percentage points for non-scheduled castes between 2004—2005 and 2009-2010. The disadvantaged groups by definition have had much higher poverty rates than other groups, but the gap has finally begun to be bridged decisively.21

There is an impression among some scholars that growth acceleration has not helped the Scheduled Tribes. This impression has derived partially

3« WHY GROWTH MATTERS

from the existence of the Maoist insurgency İn certain regions where the tribes are concentrated and partially from a sense that when displacement happens from projects involving mineral extraction, tribes are not adequately compensated. While the adverse impact of these factors on the tribes can scarcely be denied, the evidence of the decline in poverty among the Scheduled Tribes is unequivocal. In this regard, two factors must be kept in mind. First, while the beginning of the poverty decline can be traced to the early 1980s, as demonstrated by Figure 3.3, the Maoist insurgency and even the acceleration in mineral extraction activity are of more recent origin. And second, the states with the largest populations of Scheduled Tribes, which include Madhya Pradesh, Maharashtra, Rajasthan, and Gujarat, have nor been hotbeds of the Maoist insurgency.

Finally, we may also briefly mention the trends in poverty by religious groups. Here as well, estimates by Mukim and Panagariya (2013) show declining levels across all groups. In particular, they show poverty among Muslims declining from 52.2 percent in 1983 to 25.8 percent in 2009-2010, with the largest decline—9.7 percentage points—coming between 2004—2005 and 2009-2010.

Refonm and Their Impact on Growth and Poverty I 39

Dehejia and Panagariya (2012a) take a different route to analping the impact of liberalization on different social groups. They consider the status of entrepreneurship among the Scheduled Caste and Scheduled Tribe groups in the service sectors, using the survey data gathered by the National Sample Survey Organization in 2001-2002 and 2006-2007?2 •They find for entrepreneurship a pattern very similar to the one found by Mukim and Panagariya (2013) for poverty. The share of each of the Scheduled Caste and Scheduled Tribe groups according to the value added and the number of workers employed in the enterprises was and remains well below its corresponding share in the population, reflecting the historical injustices. But each social group experiences healthy growth in the value added and in the number of workers employed in the enterprises owned by its members?3

The Scheduled Caste and Scheduled Tribe entrepreneurs have experienced significant growth in their enterprises. Some anecdotal evidence is beginning to appear even on the rise of Dalits (untouchable castes tirat are included among the Scheduled Castes by the Indian Constitution) to large fortunes. In particular, newspapers have widely reported on thirty Dalit crorepatis (a crore equals 10 million and a crorepati refers ro someone having an accumulated wealth of 10 million rupees or more), almost all of them first-generation entrepreneurs, who were invited to a Planning Commission meeting specially organized for them in January 2011. The groups included Milind Kamble, who serves as chair of the Dalit Indian Chamber of Commerce and Industry, formed in 2005. According ro him, “Including mine, most of the big Dalit-owned businesses are 15 years old. With the emergence of globalization and the disappearance of the License-Permit Raj, many opportunities appeared and many of us jumped on them.”24 Referring to the meeting at the Planning Commission, he reportedly said, “The Planning Commission was stunned when they asked how many of us used government schemes to build their businesses. Only one entrepreneur from Mumbai raised his hand and described how he’d applied for $20,000, spent three years visiting government offices to chase his money and finally got $15,000.”

Thus, contrary to the general impression and especially the a priori fears of some critics, reforms and growth, and not governmental assistance,

40 I WHY GROWTH MATTERS

seem co have opened opportunities for the Scheduled Caste and Scheduled Tribe entrepreneurs to seize, in enterprises large and small.

Myth 3.4: The Planning Commission plays politics with poverty lines.

The suggestion that the Planning Commission plays fast and loose with the poverty lines couldn’t be farther from the truth.25 In setting the poverty lines, India has adhered to the highest standards of professionalism throughout its history.

The official poverty lines used until they were revised in 2011 were based entirely on the recommendations of the Lakdawala Committee of 1993.26 These poverty lines had been set such that anyone above them would be able to afford 2,400 and 2,100 calories’ worth of consumption in rural and urban areas, respectively, in addition to a subsistence level of clothing and shelter. A committee headed by Professor Suresh Tendulkar was likewise behind the revisions to the Lakdawala poverty lines, adopted in 2011 and reported to the Supreme Court by the Planning Commission. The integrity and qualifications of Tendulkar are beyond reproach.

• the second half of 2011, the media created the distinct impression that the Planning Commission, in its affidavit to the Supreme Court, had deliberately lowered the poverty lines to exclude many genuinely poor from the benefits reserved for the poor. The same impression was conveyed yet again when the Planning Commission releaseda repon in March 2012 showing acceleration in poverty reduction between 2004-2005 and 2009-2010 over that between 1993-1994 and 2004-2005. These allegations are false.
• the first case, the Planning Commission had actually raised the poverty line, while in the second case it had made no change. In 2011, the Planning Commission reported to the Supreme Court poverty lines based on the Tendulkar Committee recommendations of raising the rural poverty line from the original level while keeping the urban poverty line at its previous level. The Planning Commission had simply complied with the Tendulkar Committee recommendations.

Reforms and Their Impact on Growth and Poverty i 41

The claims of reductions in the poverty lines prominently surfaced yet again when the Planning Commission reported in March 2012 that poveri}' reduction had accelerated between 2004—2005 and 2009-2010 over 1993-1994 and 2004—2005. For instance, a headline on the NDTV website declared, “Planning Commission further lowers the [urban] poverty line to Rs 28 [from 32 rupees in the Supreme Court filing]." But once again, the Planning Commission had done no such dring. The 32-rupee line, reported to the Supreme Court, related to the year 2010-2011 and 28 rupees to 2009-2010, with the difference fully accounted for by the higher price level in 2010-2011.

One final accusation is that the Planning Commission has set the poverty lines at ultra-low levels so that it may exclude a large part of the population from benefiting from the government’s redistribution programs. While reasonable people may differ on whether it is desirable to further raise the poverty line, the subject is far more complex than is commonly appreciated. The guiding objective behind the poverty line in India and indeed worldwide has been to monitor progress in combating destitution. Therefore, poverty-line expenditures have traditionally been set at levels just sufficient to allow above-subsistence existence.

The dilemma in raising the poverty lines is best brought out by considering the implications of poverty lines that are significandy higher than those currently in use and are advocated by many of the current critics of the Planning Commission. Thus, for example, suppose we were to raise the rural poverty line to 80 rupees and the urban one to 100 rupees at 2009-2010 prices.

What would these lines imply? First, based on the expenditure survey of2009-2010, they would designate as poor 95 percent of the rural population and 85 percent of the urban population. But few analysts would suggest that all but 5 percent of the rural and 15 percent of the urban population live in destitution today. Even if we were to argue that poverty goes beyond the destitute, measuring progress at the 95<h percentile in rural and the 85th percentile in urban areas is unlikely to tell us very much about success in combating poverty.

Second, turning to the implications for redistribution, how much good to the bottom 30 percent or 40 percent who represent the truly

42 , WHY GROWTH MATTERS

destitute will we do if the tax revenues raised from the top 15 percent urban population were spread evenly over 95 percent of the rural and 85 percent of the urban population? With the tax revenues still relatively modest, significant redistribution in favor of the destitute requires limiting such redistributions to the bottom 40 percent or so of the population. Spreading them thinly over a vast population would give too little to the destitute to make a major dent in poverty.

To dramatize this argument, suppose we redistributed all expenditures, as reported in the 2009-2010 expenditure survey, equally across the population. Astonishingly, such redistribution would leave each individual with just 45 rupees per day İn expenditure. This level is well below even the lowest poverty line any critic of the Planning Commission has advocated during the recent debates on poverty lines.

Myth 3.5: Trade openness has exacerbated poverty.

The final myth we consider in this chapter folds into the criticism chat globalization is bad for the poor. It also links a specific policy reform— increased openness to trade—to an increase in poverty. It got a boost from a study by the International Monetary Fund economist Petiä Topalova (2007), who argued that enhanced openness adversely impacted poverty in India.27

However, several economists have successfully challenged her findings, showing chat increased openness has reduced poverty instead. Given the importance of this issue, and the Topalova Myth, below we summarize these studies (which can be skipped by readers not interested in the necessary technical arguments).

Topalova asked whether rural and urban districts, which are subject to different degrees of competition from imports depending on which goods they produce and in what quantities, experienced increased or reduced poverty as a result of trade liberalization in India. She measured the openness by tariffs correcting for the levels of employment in high- versus low-tariff sectors. Working at the district level, she found that increased openness had been associated with, and presumably had led to, increased incidence of poverty in the rural districts but had no statisti

Reforms «tırf Thetr impact on Growth and Poverty | 4Î

cally significant effect in the urban districts. She found no evidence in either rural or urban India that openness was associated with alleviating poverty. These were startling results because, as we argued earlier, trade openness in a labor-abundant economy stimulates growth in general and the expansion of labor-intensive industries in particular so that it can be expected to lower rather than raise poverty.

Hasan, Mitra, and Beyza Ural (2006-2007) have therefore revisited this question. They note that the analysis of poverty and trade openness at the district level poses several problems. For example, the data from the 1993-1994 expenditure survey by the National Sample Survey Organization do not readily allow the identification of urban districts. District boundaries also shift over time. There are also questions of randomness of the sample at the district level. Finally, sometimes the number of observations in a district is insufficient to yield a reliable estimate of poverty.

Therefore, these authors study the question ar the level of the state and (National Sample Survey Organization identified) regions within states. There being one or more regions within a state, regions are greater in number than states and therefore allow greater degrees of freedom. As such, their research offers an improvement over the Topalova approach: the focus on regions allows a tighter estimation of poverty than the district-focused approach and also allows for a tighter estimation of regression equations than a pure state-focused approach.

These authors also note that assigning zero tariffs to non-traded sectors in measuring openness, as done by Topalova, is erroneous. Many goods and services may be non-traded precisely because the barriers to trade are prohibitive. So they define openness as an employment-weighted sum of tariffs such that only exportable products are assigned zero tariffs, with non-traded sectors excluded from the calculation. These authors also take into account non-tariff barriers, which Topalova ignored.

In sharp contrast to Topalova’s claim that trade openness was not associated with reduced poverty, these authors’ superior methodology' failed to encounter even a single case in which reductions in trade protection worsened poverty at the state or regional level. Instead, they found that states more exposed to foreign competition had lower rural,

44 i WHY GROWTH MATTERS

urban, and overall poverty ratios, with this beneficial effect being more pronounced in states that had more flexible labor-market institutions. The authors also found that trade liberalization led to greater poverty reduction in states more fully exposed to foreign competition. Hie results held for overall, urban, and rural poverty with varying strengths and statistical significance.

Moreover, Jewel Cain, Hasan, and Mitra (2012) have also revisited the issue and reinforced the findings of Hasan, Mitra, and Ural, using data from the more recent round of the sample survey conducted in 2004-2005. They find that every percentage-point reduction in the weighted tariff rate led to 0.57 percent reduction in the poverty ratio on average. This implies that of the overall reduction in poverty during 1987-2004, on average, 38 percent can be attributed to change in the exposure to foreign trade. Since the authors control for time-fixed effects, and poverty has declined over time, they can infer that the greater exposure of the labor force to foreign competition speeded up poverty reduction. The magnitude of impact and its statistical significance naturally vary across rural, urban, and the two sectors considered together, as well as across the different tariff and non-tariff measures used. However, in no case do these authors find increased openness to result in increased poverty.

Finally, Mukim and Panagariya (2012) split the data by social group and analyze the impact of trade openness on poverty within each of the social groups in rural and urban areas. They find no evidence whatsoever in favor of the hypothesis that rising incomes or openness have adversely impacted poverty among any one of the groups. They also find that one or more measures of openness have had a statistically significant and favorable impact on poverty levels in the Scheduled Castes and non- Scheduled castes in rural and urban regions and in both regions taken together. As for the Scheduled Tribes, they find a statistically significant effect of openness on poverty in urban areas only.

Chapter 4

Reforms and Inequality

The surge of the growth rate during the eight years beginning in 2003-2004 to 8.5 percent from less than 4 percent until 1980 has meant the creation of substantial wealth. While there were no billionaires in dollar terms in India as recently as 2000, the 2007 list by Forbes reported as many as fifiy-five of them.

This new wealth has in turn led to claims that reforms have generated massive income inequalities and that India has entered a state similar to the American Gilded Age in the late nineteenth century. But while such claims may appear superficially plausible, they crumble in the face of close scrutiny.

Myth 4.1 : Reforms have led to increased inequality.

At the outset, we need to emphasize that what is an appropriate measure of inequality is not simply a technical issue—for example, whether the index of inequality should be the economists’ measure of what is called the Gini coefficient (explained below and more fully in Appendix 2), which is widely used by economists studying inequality in India and elsewhere. An appropriate measure of inequality must also reflect broader questions of relevance to the popular concerns.

Thus, for a measure to be relevant to the public-policy discussion, it must have political and social salience. For example, if incomes increase in Mumbai but not in the Ratnagiri district of Maharashtra, evidently inequality of income has increased between Mumbai and Ratnagiri. But

I 45 I

4t> i WHì GROWTH MATTERS

1. chose living in Ratnagiri are not comparing themselves co whatİs happening in Mumbai, why is this inequality measure of any relevance? So, measures of urban-rural inequality may have little relevance as well.
2. the other hand, when/w7Å/n-Mumbai inequality becomes more acute, the poor there are more likely to notice as they compare themselves with the rich in their own neighborhoods. Similarly, within our own university (Columbia University in New York), the inequality between the cop salaries—the president enjoys the highest salary—and the lowest salaries is a salient issue, bur (at least as of now) not the discrepane)' between our salaries and those on Wall Street.1 In short, an increase in inequality within one’s own village or institution is likely to raise hackles but not inequality between groups that have little relationship or contact with one another.

Then again, the political and social implications of any increase in appropriately measured inequality' would depend on the social context in which it occurs. Thus, if inequality increases and the rich spend money on conspicuous consumption, that could become socially explosive. But if mobility is high, the poor may react by celebrating the conspicuous inequality rather than resenting it, because they think that they too may someday “make it big” in that way.

Keeping these caveats in mind, consider some general economic arguments that bear on income distribution between the rich and the poor in an economy such as India’s. First, some forms of inequality can be expected to rise in a rapidly growing economy. Growth involves wealth creation. Insofar as a small number of entrepreneurs lead this wealth creation, and those creating wealth are unlikely to redistribute all of it in an act of altruism, disparity among the richest and the rest of the population in both income and expenditure is likely to rise.

Likewise, rapid growth is often led by the formation of a small number of agglomerations (i.e„ concentrations of economic activity' in limited geographical areas), which generally form in urban centers, leading to urban-rural as well as regional inequality'. On the other hand, in a labor-abundant economy, pro-growth policies are also expected to lead to specialization in the labor-intensive goods, which raises employment and wages of the poor. The poor can move from lower-paid jobs in the

Reforms and Inequality I 47

countryside to higher-paid jobs in rapidly growing urban agglomerations, thereby producing less inequality.

Against this background, what has been the Indian experience? As it happens, the evidence we discuss below shows that contrary to widespread impressions, inequality measures do not point to an unambiguously rising trend in inequality.

Thus, Krishna and Sethupathy (2012) have recently measured inequality in India, using the household expenditure survey data from the NSS rounds conducted in 1987-1988, 1993-1994, 1999-2000, and 2004-2005? They show that inequality between states and between urban and rural areas is dwarfed by inequality among households within each of these aggregates. For example, inequality within states accounts for more than 90 percent of the total inequality over the country (see Figure 4.1). Inequality between states accounts for less than 10 percent of the total inequality across the country. Likewise, inequality within rural and urban areas accounts for 90 percent or more of the total inequality across the nation.

Importandy, the overall inequality exhibits only modest variation over the period, rising slightly between 1988 and 1994 and again between 1994 and 2000, but by 2005 dropping to a level slightly above that in 1988 (see Figure 4.1). Inequality trends within states mirror the national experience: it rose between 1994 and 2000 and then fell between 2000 and 2005 in most states. Indeed, between 2000 and 2005, only four states—Mizoram, Maharashtra, Orissa, and Haryana—experienced significant increases in inequality. The picture is almost exactly the same for rural and urban areas within states; the vast majority experienced rising inequality between 1994 and 2000 but falling inequality between 2000 and 2005.

The results of other researchers, comprehensively surveyed by Weiss- kopf (2011), echo the basic message of Krishna and Sethupathy that, rather than exhibit secular trend, inequality has gone up and down during the growth process with at most a modest net rise since the 1980s. These researchers rely on the Gini coefficient, which usually varies by two or three percentage points, changing only rarely by five percentage points. What is interesting, however, is that some researchers have gone

48 I WHY GROWTH MATTERS

on to interpret these changes as representing “significant" or “pervasive” increase in inequality.

Thus, Deaton and Drèze (2002) report estimates that exhibit no clear trend in rural inequality and a small increase in urban inequality. But in the text of their paper, they surprisingly conclude, "To sum up, except for the absence of clear evidence of rising intra-rural inequality within states, we find strong indications of a pervasive increase in economic inequality in the nineties. This is a new development in the Indian economy: until 1993-94, the all-India Gini coefficients of per capita consumer expenditure in rural and urban areas were fairly stable” (p. 3740). Evidently their stark conclusion is not consistent with their statistical findings.

In almost an identical spirit, Weisskopf (2011) quotes Paria Topalova approvingly as staring that “all measures point to significant increase in overall inequality in the 1990s” (p. 46). Yet, the Gini coefficient calculated by Topalova and reported by Weisskopf changes from 31.9 in 1983-1984 to just 30.3 in 1993-1994 and 32.5 in 2004-2005. One

Reforms and Inequality I 49

would think that the difference between 31.9 and 32.5, which may not even be statistically significant, would hardly warrant the inference that there has been a significant increase in inequality.3

An extra dividend from Krishna and Sethupathy’s analysis is the finding that there is no correlation between the change in inequality across households within states and the change in state-level measures of tariff and non-tariff protection. Trade openness is not linked to increased inequality.

The critics of the reforms have also raised a different concern: that growth has been uneven across states and that this has resulted in increased inequality among them. This is correct since richer states have grown faster than the poorer states on average. This phenomenon may have political salience if it leads to resentment by the states lagging behind, which are poor to begin with. But four qualifying facts must be kept in view.

First, as Panagariya (2010a) and Chakraborty et al. (2011) show, the years 2003-2004 to 2010-2011 have seen nearly all states growing significantly faster than they did in any prior period. Therefore, the rise in interstate inequality does not reflect the poorer states’ remaining poor or being further impoverished. Instead, it represents the richer states growing faster than the poorer states in an environment in which all states are growing faster.

Second, two of the poorer states—Bihar and Orissa—are among the fastest-growing states today. Their success shows that when the national policies are conducive to growth (as they have been after the significant reforms began) and some of the states grow rapidly, the door to poorer states’ achieving similar success is also opened wider. As Bhagwati and Panagariya (2004) and Panagariya (2009a) have argued and Gupta and Panagariya (2012) have analyzed in detail, there will likely be a diffusion effect: when the rest of the economy is growing rapidly, the electorate in the poorer states will demand more from its leaders, prompting policy changes that increase prosperity. Both Bihar and Orissa have elected and reelected chief ministers who have performed well.

Third, faster growth in some states opens the scope for larger-scale redistribution programs in favor of poorer states. A program such as the

>0 I WHY GROWTH MATTERS

National Rural Employment Guarantee Scheme, which benefits the poorer states proportionately more, would not be feasible without certain states’ having large enough incomes to make the necessary revenues available.

Finally, labor is not immobile across states. It is well known that labor from Bihar has traditionally moved to Mumbai and Kolkata for jobs, as have people from the Punjab and from the hills. That means that faster growdi will attract migrants from the slower-growth states and regions, so that the prosperity (as distinct from growth) will diffuse to the slower-growth areas through the usual channels, such as remittances.

A further dimension of inequality relates to the socially disadvantaged groups. Once again, critics often assert that the income differences between the Scheduled Caste and Scheduled Tribe on the one hand and non-scheduled castes on the other have gone up during the years of rapid growth. But in a comprehensive analysis, Hnatkovska, Lahiri, and Paul (2012) show chat such claims are not supported by empirical evidence.

Using Employment-Unemployment Survey data from the NSS rounds conducted in 1983, 1987-1988, 1993-1994, 1999-2000, and 2004-2005, they show that the wages of the Scheduled Castes and Scheduled Tribes have been converging to those of non-scheduled castes since 1983. They demonstrate also that differential education levels of the two groups drive most of this convergence. Likewise, the occupation structure of the Scheduled Castes and Scheduled Tribes has also been converging toward that of non-scheduled castes. The Scheduled Castes and Scheduled Tribes have been able to take advantage of the rapid growth and structural changes in India during the post-reform period and have rapidly narrowed their huge historical economic disparities with non-scheduled castes and tribes.

Myth 4.2: Thanks to the reforms, India is nowin the Gilded Age that prevailed in the late nineteenth-century United States.

The emergence of billionaires and the exposure of a few mega-corruption cases have led some, especially Sinha and Varshney (2011), to argue that

Reforms am/ Inequality I 51

India has now entered a Gilded Age much like the United States in the late nineteenth century.4

During this period in the United States, four main strands of criticism were rampant. The first was that American business elites had succumbed to “gross materialism,” which was manifest in conspicuous consumption and crass displays of wealth. Second, this was made possible by the accumulation of great wealth by the likes of John D. Rockefeller and Andrew Carnegie, and indeed many others, while the vast masses toiled for minuscule wages. Third, these tycoons were not “captains of industry" to be admired but were rather “robber barons” who had built their fortunes on abusive business practices and high-handed suppression of attempts at unionizing labor. Fourth, (in modern terminology) there was a businesspolitical complex, such that these robber barons and corrupt politicians had greased one another’s palms and defrauded the nation.

Mark Twain’s 1873 novel with Charles Dudley Warner was titled The Gilded Age, a phrase that gained wide circulation.5 The book was a reaction to the excesses that accompanied the remarkable growth of the American economy as the production of iron and steel took off with rail transpon expanding rapidly to bring primary resources from the expanding western frontier to the east. Oil and banking expanded at an unprecedented pace as well, leading to massive fortunes for tycoons. The vignettes from this Gilded Age amply illustrate the criticisms that attended the extraordinary growth.

Lavish parties were a way of life for the nouveaux riches. An account of the time says: “Sherry’s Restaurant hosted formal horseback dinners for the New York Riding Club. Mrs. Stuyvesant Fish once threw a dinner party to honor her dog who arrived sporting a $15,000 diamond collar.”6

There was also a populist resentment of the extreme wealth contrasted with the tragic reality of slums and subsistence wages in the overcrowded tenements in the growing urban towns and cities. The general perception, reflecting that contrast, was that while the rich wore pearls, the poor were in rags. There was growing talk of retribution through emerging violence: fears grew of “carnivals of revenge.”

52 i WHY GROWTH MATTERS

Against this backdrop, labor began to organize against long hours and low wages, and the robber barons occasionally reacted by breaking the strikes brutally. Thus, even Andrew Carnegie, who professed sympathy for the poor, reacted to the Homestead Strike of 1892 by supporting his manager, Henry Frick, who locked out workers and hired Pinkerton musclemen to threaten the strikers. This was no isolated incident.

The offending titans also operated in a governance vacuum regarding business practices they used to gain monopoly control. Notorious for such business practices was John D. Rockefeller of Standard Oil Company, who in 1870 turned his company into one of the nation’s first notorious monopolistic trusts. Antitrust legislation came later: the Sherman Anti-Trust Act of 1890 and the later, tougher, and more effective Clayton Anti-Trust Act of 1914.

The era was also marked by corruption at the highest levels of government, including the president, but more typically at the levels of local governance where, as today, businesses and governments shared the spoils from local grants of cash subsidies and land gifts, presumably for a “social purpose" (such as constructing a railroad) but in fact for the sole purpose of defrauding the commonwealth.

Is India today in such a Gilded Age? There are superficial similarities, for sure. True, in the same way that fast growth in the nineteenth-century United States created the Vanderbilts, Carnegies, Rockefellers, and Morgans, it has created a large number of billionaires in India. Again, like the robber barons of the American Gilded Age, Indian billionaires have tilted the playing field to their advantage through securing mining and land resources, seeking regulations favorable to them, and blocking foreign entry. But the similarity ends there.7

The conditions characterizing the American Gilded Age and current- day India are vastly different. At the beginning of the Gilded Age, the dominant economic philosophy in the United States was laissez-faire. There was virtually no effective regulatory, labor, or social legislation at the federal level. Two key pieces of regulatory legislation—the Interstate Commerce Act of 1887, which aimed to limit the monopoly power of the railways, and the Sherman Act of 1890, which provided for antitrust

Reforms und friet ¡¡niitty i 51

action against businesses—were enacted during this period. Key laws providing protection to industrial labor, the poor, and the elderly came much later. With rare exceptions, only white males enjoyed voting rights.

In contrast, post-reform India and its growth explosion have been preceded by several decades of a command-and-control system complemented by stringent legislation in favor of industrial workers so that it is impossible to have business tycoons breaking strikes the way the American robber barons did. India, İt may be recalled, has had a longstanding national commitment to eradicating poverty and achieving universal adult suffrage since independence. The country' has all elements of a liberal democracy with the poor and the underprivileged having access to effective politics at the ballot box.

The economic reforms have allowed freer play to private entrepreneurs but can hardly be characterized as laissez-faire. Railways remain a public-sector monopoly and the government remains a major player in such key sectors as steel, coal, petroleum, and engineering goods. Despite private-sector entry in airlines, telecom, insurance, and electricity, the public-sector players have remained active. In banking, the role of domestic and foreign private players has been expanded but the public sector again remains dominant. And several sectoral regulatory agencies, topped by an all-encompassing Competition Commission, oversee business practices.

Whereas during America’s Gilded Age major sectors such as steel, oil, sugar, meatpacking, and the manufacture of agriculture machinery came to be dominated by “trusts,” the opposite is true in India today. There are multiple domestic firms within many sectors, competing against one another as well as with imports and foreign investors. Increased competitive pressures have led to reduced prices and improved quality of products and services in such diverse sectors as airlines, telecommunications, automobiles, two-wheelers, refrigerators, and air conditioners.

The treatment of industrial workers in India today stands in sharp contrast to that in late nineteenth-century America. During the Gilded Age, factory workers toiled sixty-hour weeks without pensions, compensation for job-related injuries, or insurance against layoffs. In contrast to

54 I WHY GROWTH MATTERS

die strikebreaking actions of the robber barons at the Homestead Steel Mill in 1892 and George Pullman’s railroad in 1894, labor laws in India provide a high degree of protection to industrial workers.

Finally, whereas the state provided no protection to those at the bottom of the income distribution, including farmers in the United States during its Gilded Age, the government in present-day India is sensitive to the fate of the poor. Indeed, growth and the social programs it has made feasible have helped bring down poverty significandy. The changes have also benefited the underprivileged groups, as we have already documented, and as we noted, there are now a handful of Dalit crorepatis.

But what about the corruption? How does today’s India compare to America’s Gilded Age? Critics contend that the post-reform era has been driven by crony capitalism. This implies that Indian entrepreneurs have accumulated wealth mostly by redistributing it in their favor through outright fraud İn collaboration with politicians, rather than by creating it.8 But the allegation is not persuasive. Unlike Mexico, for instance, where the billionaire Carlos Slim Held has used every conceivable means to generate monopoly profits for himself, most Indian entrepreneurs have become rich by creating wealth while operating in a highly competitive market. Recent empirical work by Alfaro and Chari (2012) also points to the existence of a highly competitive market in India with substantial entry of new firms on the margin. To be sure, one can find examples such as those of the Reddy brothers, who, according to their recent indictment, have accumulated wealth from illegal mining; but that is not the case with the vast majority of Indian entrepreneurs from the information technology, telecommunications, pharmaceuticals, or engineering goods industries.

So, while the American Gilded Age produced Rockefeller and Vanderbilt, India today has given rise to Narayana Murthy of INFOSYS, Azim Prernji of WIPRO, and Uday Kotak of Kotak Mahindra Bank. There is not a hint of corruption or shady practices among these successful tycoons. All of them are associated with extensive engagement with society, and have embraced corporate social responsibility and private social responsibility. Whereas Carnegie and Rockefeller gave away their fortunes on their death, the Indian tycoons have given away massive

Reforms and Inequality I 55

sums of money even as they have earned them. Besides, their lifestyles are simple, nor extravagant.

While the tycoon Mukesh Ambani has built a much-condemned high-rise in Mumbai, the display of personal wealth and gross materialism is far less rampant in India than in nineteenth-century America or in the New York of the 1970s prior to the current crisis and even after the financial sector’s recovery. Perhaps the worst displays take the form of flamboyant and unseemly weddings costing millions, but that is a long-standing cultural tradition.

While, therefore, India is not a throwback to the American Gilded Age, one might ask; What about China? Here the parallel is closer. Union rights are nonexistent; ostentatious displays of wealth are common; and there is little attempt at corporate and personal social responsibility. So here again India scores against China and handsomely at that.