Teja

Shifting industries towards formality—reducing the dualism in the economy—constitutes another important form of structural transformation. Careful studies have documented large efficiency gaps between comparable manufacturing firms in the formal and informal sectors, implying large potential efficiency gains from growth of the formal economy ; Mazumdar and Sarkar 2008). Similarly, the service sector contains sharp distinctions in productivity levels between what are known as “modern” and “traditional” services. Modern services are technology-enabled, transportable, and tradable. They include financial intermediation, communication, computer services, business services and professional services.

Because of links to technology and trade, modern services perform much more like manufacturing: characterised by fast productivity growth and potential to leverage export markets for growth. In India, communications, finance, and computer-related services yield five or more times the output per worker than most traditional services.

• The modern/traditional distinction has been found across broad swaths of developing economies and in India in particular  although less distinct.
• Successfully reorienting India’s labour force towards higher productivity sectors would directly boost economic growth. Indeed, McMillan et al (2014) find that the main difference in the growth experience of Asia with that of Latin America and Africa has been due to Asia’s superior success at structural reform.
• The productivity gains imply large welfare gains for some of India’s poorest workers. The McKinsey Global Institute has estimated that an illiterate worker who moves from agriculture to light manufacturing can expect a wage increase of 40%. A worker with basic literacy can expect even better: a wage increase of 70% should he move from agriculture to heavy manufacturing (Gupta et al 2014).
• How should this structural transformation pulling labour into higher-productivity sectors occur? Economists have debated whether the best strategy for job creation in India lies in developing its service or manufacturing sectors.
• Green (2014) argues that the Indian manufacturing sector holds more growth potential in response to policy changes, and Ghose (2015) shows more employment potential for low-skilled workers in manufacturing. This study likewise explores the feasibility of boosting the manufacturing sector.
• Achieving this goal would require intra-sectoral shifts for driving faster expansion of labour-intensive activities, as well as reducing the dualism in the manufacturing sector to reap the benefits of productivity gains that are available from shifting activity and employment into the formal sector.
• This study analyses what could happen if India’s government took steps sufficient to achieve East Asia-style manufacturing growth. The vision for “Make in India” includes goals to increase the gross domestic product (GDP) share of manufacturing to 25% by 2022 and to create 100 million additional manufacturing jobs by 2022 (Department of Industrial Policy & Promotion 2016).
• Are these goals realistic? With assumptions about sector-level growth and employment elasticity, this study projects sectoral employment, productivity and output patterns over 20 years. The projection exercise presented here assumes a structural break in the manufacturing sector due to major policy changes. This implies two important shifts from the usual analysis of structural transformation in India.
• First, it implies that past patterns of the utilisation of labour (for example, labour intensity and skill intensity) will be broken, and therefore do not serve forecasts of the future.
• The experience of five East Asian economies that witnessed manufacturing-led growth booms provides a better benchmark for the parameterisation of the projections.
• Second, because of the minimal parametric restrictions and assumptions, the model can avoid the problem of false precision. Unlike other projections of sectoral employment—for example, Rangarajan et al (2007), Planning Commission (2012), Papola and Sahu (2012), Timmer et al (2014) and Gupta et al (2014)—that follows a detailed industrial classification into important subgroups,
• this paper breaks down manufacturing between informal and formal sectors, to distinguish between the fundamentally different segments of the economy that are often blended together. The projections provide a rough upper bound of possible outcomes from structural transformation, which may be informative for developing policies for structural change.

Developing the Projections

The core of the projection is a sector-wise GDP forecast. Employment figures then derive from an assumption of constant employment elasticity. Hence, the most important parameters are the assumptions of future growth and employment elasticity.

• Data: The employment data used here comes from various sources. The most comprehensive data on sector-wise employment at the four-digit level of India’s National Industrial Classification is provided by the National Sample Survey Office (NSSO). The National Sample Survey (NSS) employment data also breaks down the informal sector employment at the one-digit level. This study uses the NSS 68th (2011–12), 66th (2009–10) and 61st (2004–05) rounds.
• These are matched to the sectoral net value-added data from the national accounts such that detailed employment elasticities and productivity data can be constructed. Unfortunately, outside of manufacturing, the national accounts data only provides a formal/informal breakdown for the net value added. Hence, this study only explores the formal/informal difference for manufacturing.
• Consistent historical national accounts data are only available through 2014, so the projection begins in 2015. Formal sector manufacturing data on employment and value added also comes from the Annual Survey of Industries (ASI), which provides an alternative source to compare with for key parameters.
• The data for the East Asian countries comes from the Groningen Growth and Development Centre (GGDC) 10-sector database that has annual sector-level value added and employment data that match India’s sectoral breakdown fairly well. The main inconsistency is the inability to distinguish between the formal and informal sectors in the East Asian value-added data.6

Methodology:

The key units of observation are broad sectoral categories, namely manufacturing, other industry (construction and utilities), services and the primary sector (agriculture and mining). Manufacturing is further divided between formal and informal segments.

Services is divided between modern services (communications, financial and business services, and real estate) and traditional services (trade, transportation, public administration, hospitality, education, healthcare, entertainment, household services and other).

• A baseline scenario is constructed first to establish a “no change” scenario, in which current policies influencing sectoral transformation are held constant. It therefore relies as much as possible on parameters as currently observed in India. The International Monetory Fund (IMF) estimates India’s potential growth to be 7.75%, which underpins the baseline projection over the next 20 years (IMF 2018).
• A more difficult task is to match that growth rate to reasonable assumptions about sectoral growth. The approach here is to base sectoral growth rates on historical rates from 1994 to 2012.
• The reforms initiated in 1991 produced above-trend GDP growth starting in 1994, corroborated by the structural breaks in the growth rate found by Balakrishnan (2010). The high-growth period ends in 2012 when investor confidence and GDP growth collapses.
• In addition, the most recent available employment data comes from the NSS of 2011–12. By using 1994–2012 parameters, the baseline sets a high bar by presenting a sustained, high-growth period with which to contrast the alternate scenarios.
• The initial growth rate for each sector was taken from each sector’s compound annual growth rate (CAGR) from 1994–2012, which witnessed aggregate growth just above potential growth at 7.1%. To ensure the aggregate rate at the beginning of the projection equals 6.5% and sectoral rates are proportional to their historical pattern I trim each sectoral growth rate by 21%.
• After trimming, the growth rate for the construction and utilities and traditional services sectors lie below the rate of the general economy. However, industries like construction, trade and transportation tend to grow at the same pace as the overall economy. To account for this, a catch-up term is included in their growth projection to pull their growth rates towards the aggregate.
• For the baseline projection, their growth rate is adjusted by half of the distance between their growth rate in 2017 and the general economy growth rate.
• As time progresses in the projection, the faster-growing sectors occupy a larger share of the total economy. This means either the aggregate growth rate will climb over time, or the sectoral growth rates of faster-growing sectors must fall.
• The latter seems more realistic, given the torrid pace of growth during 1994–2012, and given that the baseline assumes no change in the policy mix to facilitate structural adjustment. The individual sectoral growth rates γit therefore decline each year by a factor δ, constant across time and sectors, which keeps the aggregate growth rate from exceeding 6.5% per year.

The annual growth rate for sector i therefore evolves according to the following process:

γit = γi0 (1– δt) …(1)

The exception noted above is for construction and utilities and traditional services, which evolve according to the equation for sector j:

γjt = [γj0 + λ(γt–1– γjt–1)](1 – δt) …(2)

where λ is the catch-up coefficient. For the baseline scenario,
λ = 0.5 and δ = 0.065%.

Sectoral Growth

With these parameters the sectoral growth rates average out to a level slightly lower than their initial rates, presented on lines 2.1 and 2.2 in Table 1.

• For the policy change scenario, the fundamental assumption is that India’s business climate for formal sector manufacturing alters sufficiently to ignite an East Asian-style growth spurt. Therefore, India’s historical pattern is not as relevant as that of East Asia.
• This study compares India with the experience of Korea, China, Indonesia, Malaysia and Thailand, five of the eight high-performing East Asian countries that experienced 20-year booms in manufacturing value added. Singapore and Taiwan were dropped due to their small population, and Japan due to its far more developed status at the time of its post-war boom.

The five countries examined all had large agrarian populations at the time that their manufacturing boom began.

The booms are measured to identify a 20-year period that followed a big bang of reforms comparable to what India might achieve. Therefore, this study matches the start to the time of major events, which admittedly can be somewhat arbitrary relative to a continuum of reform initiatives. However, the basic results are robust to small adjustments in the timing used. Korea’s boom is measured beginning with the election of Park Chung-hee in 1963.

China’s reform period begins under Deng Xiaoping in 1978. Indonesia begins with the major devaluation and banking reforms in 1978. Malaysia took major steps towards export-oriented industrialisation in 1985 and 1986, so this study uses 1985 as the start period. Thailand’s major reforms began in 1985 and continued into the next year.

• For comparison, India’s experience beginning in 1994 is included in Table 2 (p 40). Since this study involves a formal/informal breakdown, it presents India’s experience in the most recent 20-year period for all manufacturing activity as well as for just the formal sector.
• In terms of the initial share of manufacturing in the GDP, India’s full manufacturing sector falls in line with its East Asian peers. Even the formal sector does not have a smaller share of GDP than Korea in 1963. However, during the subsequent 20 years, the manufacturing sector in the other Asian countries gained on average 14 percentage points of GDP share, while India’s manufacturing only kept up with the overall GDP.
• For the projections, I assign India formal-sector manufacturing growth rates that match the country with the highest 20-year growth rate, Korea. Two reasons justify this choice. First, this study focuses on the formal manufacturing sector, which should grow faster than the overall manufacturing sector when structural reforms remove some of the barriers that previously forced firms into the informal sector.
• Since the other country data is for the overall manufacturing sector, the highest-growth country—percentage points above the average growth rate—provides a precedent for possible growth rates that India’s formal manufacturing sector might achieve.
• Second, the scenarios aim to present the potential impact of structural reforms on India’s manufacturing sector. Replicating the highest-growth country establishes a plausible upper bound of the impact on the manufacturing sector of sufficient reform treatment.
• Theoretical arguments can be made to support both positive and negative growth effects on the other sectors in response to big bang manufacturing-oriented reforms. These considerations are discussed in detail in Green (2015). With an array of possible sectoral responses to faster formal-sector manufacturing growth, the projection chooses the starkest point of contrast, assuming the remaining sectors follow their historical pattern, shown in line 3.1 in Table 1.

That set of first-year growth rates produces an aggregate growth rate of 9%, which is maintained for the full 20-year projection.

The informal manufacturing, modern services, and agriculture and mining sectors are assumed to grow according to equation 1. Because the construction and utilities and traditional services sectors are more likely to benefit from manufacturing growth, they are assumed to grow according to equation 2 with λ = 1. This value of λ means these two sectors grow at the same rate as the total economy, about 1.4% higher than their sectoral historical rates.

• The policy change scenario requires a higher level of δ than the baseline because it has two high-growth sectors. All the sectors are compressed by δ = 0.125% to ensure the entire economy’s growth rate remains constant at about 9% per year over the 20-year projection. Despite this limitation, the scenario is aggressive relative to historical growth experience.
• The overall rate at 9% slightly exceeds India’s highest five-year growth period 2003–08 and exceeds the 20-year growth rates of all the East Asian boom economies except China. The average growth rate for each sector appears on line 3.2 in Table 1.

Employment Elasticity of Growth

The economic growth rates combine with the employment elasticity of GDP to generate the core forecast of future employment. While GDP growth is quite commonly understood, the employment elasticity of GDP merits discussion to help apprehend the related assumptions in the projections.

• The employment elasticity of GDP is the percent change in employment for a 1% change in GDP, which is the inverse of marginal productivity, the change in aggregate productivity from adding one worker. Most often the elasticities are calculated from employment and GDP across several years, so they come close to the inverse of average productivity. The marginal/average distinction has three important implications.
• First, high-productivity industries will by definition have a lower elasticity than low-productivity industries. Hence, a low elasticity does not indicate a bad industry for job creation, since ultimately productivity growth lifts wages and living standards. If a high-productivity (low elasticity) industry grows fast enough it can provide a welcome source of high-quality jobs. Accordingly, very high elasticities can indicate problems with falling productivity.
• Second, it is always true that average productivity rises by adding new workers at higher marginal productivity. Because elasticity is the inverse of marginal productivity, a sector’s
  average productivity advances in the projections by adding new workers at lower elasticities. In fact, because it is adding new workers faster, a faster-growing sector will have greater productivity growth than a slower-growing sector even when both have identical elasticities.

Third, positive structural change means that higher-productivity (lower elasticity) industry output grows faster. The marginal productivity effect will cause average productivity to grow faster too.

Elasticity is typically measured as the ratio of growth rates of employment and output (arc elasticity) or as the coefficient of a log-log regression (point elasticity). For India, Misra and Suresh (2014), hereafter MS, use KLEMS methodology to construct an annual employment time series that matches GDP data frequency from 1994–2012 and perform log-log regressions for various sectors. They also use ASI data to perform industry-level panel log-log regressions to generate point estimates of employment elasticity in the formal manufacturing sector.

• Unfortunately for this study’s purposes, MS do not make the modern/traditional services distinction and do not address informal manufacturing. Their elasticities can be used for formal manufacturing, construction and utilities and agriculture and mining. Instead, for other sectors I have calculated the ratio of the CAGR for sectoral employment (from NSS data) and the CAGR for sectoral value added (from national accounts data) across the years 2005–12.
• The low frequency of employment data—every five years for NSS data—and occurrence of structural breaks in the economy hinder more precise methods. As a result, the ratio of growth rates methods used here is the most commonly used measure of employment elasticity for India (for example, Rangarajan et al [2007] and MS). These estimates are close to those of MS except for modern services, which in that study only comprises finance and real estate (Table 3).
• In recent years, formal manufacturing in India has witnessed high elasticity relative to other sectors—only construction and utilities is higher—justifying the policy focus on this sector for job growth. The MS estimate utilises industry-level data and, at 0.57, represents the middle value.
• This estimate is also close to the high end of the elasticities seen in most of the East Asian boomers (Table 4). However, those elasticities include the informal sector. When India’s informal sector is included, as in Table 4, India is an outlier in the other direction. The unusual amount of employment in the informal manufacturing sector in India pulls the overall manufacturing elasticity down.

Assuming a similar but smaller effect in the East Asian economies, a slightly high estimate for the formal manufacturing sector in India appears appropriate.

The policy change scenario assumes employment elasticity in formal manufacturing will rise slightly, to 0.7. Elasticity could theoretically rise through two channels. First, a lower effective cost of labour relative to capital because of labour market reforms could induce industries to raise their labour intensity. Second, reforms resulting in lower cost of labour or improvements in infrastructure could improve the comparative advantage of labour-intensive industries, giving them relatively higher growth. Shifting the rates of growth between industries in favour of labour-intensive ones can raise the overall sectoral elasticity. The historical range of elasticities among manufacturing industries is wide enough for the “between effect” to move the elasticity about +/- 0.2 without making unreasonable assumptions about long-run industry growth rates.

• Informal manufacturing has a lower elasticity than formal manufacturing, and formal manufacturing sector growth has been shown to lower the share of employment in informal manufacturing (Ghani et al 2013). Indeed, Unni (2003) finds that the growth of informal manufacturing employment after the 1991 reforms occurred because formal firms were restrained by labour laws.
• As the formal sector offers more jobs, disguised unemployment in the informal sector should decline. Of the two available estimates of informal manufacturing elasticity, the projections use the lower one of 0.15.
• The construction and utilities sector have high elasticity because construction is labour-intensive. Gupta et al (2014) argue that income from the Mahatma Gandhi National Rural Employment Guarantee Act (mgnREGA) programmes have generated a building boom in rural areas, meaning a greater share of construction takes place in low-wage, low-productivity areas.
• This has caused the elasticity of construction to rise. Notice the Rangarajan et al (2007) estimate, which pre-dates the mgnREGA programme, is the lowest. The mgnREGA-induced trend may not persist indefinitely, so the projection uses the MS estimate using log-log regressions, which at 0.99 is on the lower end of the range. This is not exceptionally low by the experience of the East Asian boomers shown in Table 8 (p 44).
• For modern and traditional services, this study’s estimates provide the only elasticities that distinguish between the two sectors appropriately. Most other estimates for modern services in India or East Asian economies are likely too high due to the exclusion of low-elasticity communications.
• Accordingly, the traditional services estimates are likely too low. Compared to the other studies, this study’s estimates appear appropriate. The East Asian boomers’ traditional services elasticities are much higher. Since this exercise assumes little change in this sector, estimates derived from India’s data are more appropriate.
• Agriculture and mining display declining elasticity for the same reason as informal manufacturing, namely shedding of surplus workers. Again, the East Asian boomers’ elasticities are much higher, but the Indian estimates are roughly declining over time, with the trend giving additional confidence that the pattern is not spurious.
• The baseline estimate uses the MS estimate. For the policy change scenario, a better supply of non-agricultural jobs will presumably pull excess workers out of agriculture faster, so the author’s estimate of -0.48 is applied.

Results

The simulation extends from 2014, assuming the employment situation is unchanged from the 2012 NSSO survey. The initial values for the projection are given in Table 5.

Running the simulation over 20 years produces significant differences between the baseline and the policy change scenarios (Table 6). Simple compounding of the assumed growth differential produces overall GDP that is 30% higher than what it might be without reform. Productivity (which should correlate with wages) also grows faster with reform.

• In the baseline scenario modern services follows its historical pattern of being the only high-growth sector (see Table 1). By the end of the projection, this produces a 45% share of GDP for services, which is approaching levels seen in advanced economies.
• In the policy change scenario modern services grow slightly faster, but because other sectors also grow faster, its share of GDP falls. Relatively higher growth rates in formal manufacturing cause it to grow 3.7 times larger than the no-reform scenario, yielding a substantial rise in its share of GDP.8
• Not only does the productivity in each sector expand, but because employment shifts towards higher-productivity sectors, aggregate productivity also expands faster than any individual sector. This inter-sectoral reallocation of labour is a form of structural change.
• Some productivity change due to the inter-sector labour shifts occurs in both scenarios (Table 7). However, the difference in the productivity growth between the two scenarios is mostly due to a higher degree of structural change in the policy change scenario.
• Perhaps most importantly, job growth would be substantially higher in the policy change scenario (Figure 1, p 44). Formal manufacturing employment would grow to exceed informal manufacturing (11% of employment versus 7%). The two together add 76 million new jobs by 2035 over their 2014 levels.
• Agriculture, on the other hand, is assumed to shed jobs faster in the reform scenario. Because so many work in agriculture, it takes ten years before the growth sectors overtake it for net gains over the baseline.
• Construction has a very high need for manpower, so employment in that sector would also expand rapidly. Agriculture sheds jobs, but the other sectors of the economy exhibit plenty of capacity to absorb those workers.
• Green (2014) calculates a need to create 10 million new jobs each year on top of what is needed to recoup manpower shedding in agriculture. Currently, India’s economy does not meet that mark, thereby creating a job gap that pushes people into fallback employment, underemployment, unemployment or out of the labour force.
• The baseline scenario does not reach a pace of creating 10 million jobs per year until 2030. This creates a backlog of workers (the cumulative historical gap) that do not find a job. The policy change projection hits a pace of 10 million new jobs per year by 2022, and completely covers the job gap backlog by 2027.

Alternative Specifications

• India’s economy is in a constant state of transformation, typical of developing economies with high growth rates. This makes any 20-year extrapolation risky. The data underpinning the projections is not perfect either. For instance, there may be short-run phenomena such as drought years that create misleading patterns. Employment is not a sharply defined concept, especially in an economy characterised by high rates of informality.
• Of course, almost any outcome can be achieved by selectively choosing high or low growth rates and employment elasticities. Therefore, this exercise has utilised mid-range parameters from recent estimates compared with historical patterns in India and East Asia. Nonetheless, it remains worthwhile to carry out some alternative specifications to explore how sensitive the projections are.

This section will focus only on the formal manufacturing sector—the main sector of interest—to limit the number of permutations explored.

The “Make in India” goals provide useful targets to structure alternative specifications around. These goals specify increasing the GDP share of the manufacturing sector (formal and informal together) to 25% by 2022 and creating 100 million additional manufacturing jobs by 2022. Structuring scenarios around meeting these targets allow a measure of the sensitivity of the projections to both growth and employment elasticity assumptions.

• Attaining 25% of GDP by 2022: The first alternative scenario asks what growth rate of formal manufacturing would be required for the overall sector to reach 25% of GDP by 2022, or within eight years of the start of the projection. In the original policy change scenario, manufacturing comprises 21% of GDP in 2022 and does not reach 25% of GDP until 2030.
• As discussed in the previous section, the growth rates in several other sectors may get pulled higher by faster growth in formal manufacturing. This would create headwinds for attaining a share-of-GDP target. For the purposes of simplicity, this scenario ignores such effects and assumes that growth rates in the other sectors—including informal manufacturing—remain identical to the policy change scenario.
• Making the single change of adjusting the formal manufacturing growth rate to meet the target, the projections indicate that the formal manufacturing sector would need to grow at 20% per year for the overall manufacturing sector to reach 25% of the GDP by 2022 (Table 8).
• This is about 6 percentage points higher than the growth rate assumed in the policy change scenario and 4 percentage points higher than the highest annual growth rate of formal manufacturing in the last 20 years.
• Further, the average annual productivity growth displayed by overall manufacturing in the first two scenarios is more than double the productivity of either the formal or the informal manufacturing sectors. This is a more extreme example of the inter-sectoral effect noted earlier, as the higher output growth rate in formal manufacturing shifts the proportions of economic activity from a low-productivity to a high-productivity sector.
• 100 million manufacturing jobs by 2022: The second scenario asks what parameters could yield 100 million new manufacturing jobs by 2022, compared to the 17 million as estimated from the same date in the policy change scenario.9 For illustrative purposes, the second scenario assumes that growth remains unchanged from the policy change scenario, so only the elasticity of formal manufacturing is allowed to adjust.
• In this case the elasticity would need to be 2.16. By almost tripling the elasticity from 0.70 in the policy change scenario, the projection produces a nearly sixfold rise in the number of new jobs created in the first eight years. By the end of the 20-year projection, the high elasticity yields a 38-fold increase, the product of compound growth rates of output.
• Such a rise in elasticity implies a completely unprecedented jump—by Indian or international standards—in labour intensity in formal manufacturing. As described above, theoretically this could occur either within the industries that constitute formal manufacturing, or between them as higher elasticity industries grow faster.
• However, the inter-industry channel has only limited range to impact elasticities. This would mean the labour intensity within industries would need to bear the burden of adjusting for the manufacturing elasticity to rise so fast. A rise in elasticity necessarily impacts average productivity. Applying such large quantities of labour on the same amount of output implies formal manufacturing productivity falling by more than 12% per year across the projection period.
• Combined goals: As a final exercise the two goals can be combined. If manufacturing reached 25% of GDP by 2022, what elasticity would formal manufacturing require to also reach the goal of 100 million new manufacturing jobs? This scenario repeats the assumption that the other sectors’ growth rates and elasticities remain the same as the original policy change scenario.
• In this case, the growth rate of formal manufacturing again reaches 20% per year to attain 25% of GDP for manufacturing. With that growth rate, a lower employment elasticity can achieve the same employment goal. Hence, the necessary elasticity falls to 1.6, still an unprecedented figure.
• If this growth rate and elasticity extended for the full 20 years of the projection, manufacturing would create 4,103 million new jobs, a 54-fold increase over the original policy change scenario. Formal manufacturing productivity would fall to 9.5% per year in this case.

Sensitivity Analysis

The alternative specifications, roughly demonstrate the range of possible outcomes from altering the key parameters of growth rate and employment elasticity. However, a fuller sensitivity analysis illustrates that even staying within previously observed values of GDP growth and employment elasticity generates a strikingly broad range of forecasts.

The sensitivity analysis begins by generating a variance–covariance matrix of the sectoral growth rates from the rates observed historically. For the baseline scenario, the matrix draws on India’s sectoral growth rates from 1994 to 2012. For the policy change scenario, the matrix is generated from the sectoral growth rates observed in the East Asian countries, which have been pooled together for comparison, during their growth boom years.10

• The model then runs 10,000 times each for the baseline and policy growth scenarios. While the growth rates applied in the earlier scenarios represent average values, the 10,000 runs use random draws of growth rates for each of the two distributions.
• Drawing from the covariance matrix provides some structure to allow sectoral growth rates that tend to co-move in the draws. The δ term will compress all sectors to keep the aggregate growth rate at the long-run level, so one sector’s draw will affect the other sectors’ final growth levels.
• The sectoral GDP illustrates the range of outcomes well, even if the levels have little intuitive meaning. The 95% bands of the GDP outcomes can be seen in Figure 2. While the median policy change scenario lies well above the median baseline scenario, the 95% confidence bands are wide. The range of experiences in East Asia was large enough so that the baseline median lies within the range of all policy change scenarios.
• The sensitivity of employment to parameter values also merits assessment. Employment depends on both the growth rate and elasticity, so both should be used to gauge the sensitivity of the projections. The two growth rate sensitivity analyses can produce employment levels, too.
• Applying the Indian and East Asian employment elasticities to the growth rates from the 10,000 draws of the baseline and policy change scenarios, respectively, generates a distribution of employment outcomes.
• Because employment elasticity should be measured across long time spans, India alone does not provide enough data points to construct a variance matrix and produce a sensitivity analysis. Even the data for East Asia is limited for this reason.
• For East Asia, the employment elasticity is measured across the two 10-year windows in each 20-year boom period of each East Asian economy. The random draws come from the variance measured across the observations for each sector, pooled across countries.
• Employment levels at the end of the projection period vary even more substantially than the growth rates. In Figure 3 (p 45), the two growth distributions produce employment levels that are proportional to the outcomes in Figure 1, but scaled by the employment elasticities. While the sectoral elasticities for India and Asia are not identical, the differences only substantially change the proportions for the agriculture and mining sector.
• The results using the distribution of the East Asia elasticities require some explanation. For formal manufacturing, other industry and agriculture and mining, all three distributions paint a similar picture. That is, the median policy change scenario lies well above the median baseline scenario, implying that the 95% confidence bands are wide.
• For the other three sectors, however, the East Asian elasticity results do not align with the scenarios using India’s elasticity. As noted earlier, the East Asian data does not have informal manufacturing broken out at all. Formal manufacturing is used instead, making the East Asian elasticity distribution for informal manufacturing a poor comparison.
• The East Asian elasticities for the two service sectors are much higher than the Indian elasticities. This elevates the top of the distribution far above the other two scenarios. In fact, for formal manufacturing, the mean of the baseline scenario is below the 95% confidence band for the East Asian elasticity distribution scenario.

Despite the indication of a significant difference between the baseline and reform scenarios, this sensitivity analysis, by and large, underscores the uncertainty involved in 20-year projections. The main projection results discussed in the paper use mean values, but a much wider set of outcomes might reasonably be expected to occur. That does not imply that projections based on mean values have no merit. They still provide useful central values for expectations of future outcomes.

Conclusions

This study has attempted to apply a rigorous approach for developing a 20-year projection of growth and employment in India. A realistic but ambitious parameterisation of a simple projection demonstrates the potential impact of an East Asia-style manufacturing boom in India. Growth, employment and productivity would all improve.

• This occurs because the central projection simulates the formal manufacturing sector growing to attain 27% of GDP in 2035 from the current 11%. Two implications of these results are worth noting.
• First, the policy change scenario forecasts that 15% of the workforce could be employed in high-productivity industries in the formal manufacturing sectors and modern services after 20 years. As a comparison point, Green (2014) estimates that almost half of India’s workforce will have finished high school by 2035, double the share today.
• Such a graduation rate would represent a dramatic improvement in worker quality over the current workforce. Compare this to the profile of the industries that are most likely to need workers with at least a high school education. Currently, 48% of the workers in formal manufacturing, 88% in the modern service sector and 60% in the traditional service sector have at least a high school education.
• Those three sectors employ 29% of the workers, while the remaining sectors utilise a much lower share of skilled labour.
• The potential rise in education levels above current industry need raises the question of where these workers will find work appropriate for their superior education. Another way to look at the potential mismatch is via Say’s Law that supply creates its own demand.
• Say’s Law suggest that businesses that can effectively utilise a better educated workforce will grow faster due to a growing skilled labour supply. Much better educational attainment may suggest that the projections presented here are not unrealistic.
• Second, the main policy conclusions of this study could be established with a more casual parameterisation, as the basic results are robust to a range of realistic assumptions. One point of rigorously parameterising the model is to rigorously rule out what is not realistic.
• The “Make in India” goals of the manufacturing sector reaching 25% of GDP and creating 100 million new jobs by 2022, while worthwhile for inspirational purposes, do not appear realistic. The latter does not even appear realistic in a 20-year time frame.

Notes

1. Author’s calculations based on the 14-industry aggregation in India’s national accounts. From largest to smallest, these three are chemicals and pharmaceuticals, basic metals, and transport equipment, according to the National Sample Survey (NSS) and national accounts data.
2. Author’s calculations based on NSS and national accounts data. Business services productivity stands out less clearly because high-productivity workers like call centre workers are far outnumbered by security guards and errand boys with productivity that compares more closely to workers in traditional service sectors.
3. This is a point missed by most evaluations of manufacturing versus service sector-led growth (Ministry of Finance 2015).
4. Rangarajan et al (2007), Planning Commission (2012), Papola and Sahu (2012), and Gupta et al (2014) break services into subsectors, but they follow a common national accounts breakdown in which communications—a modern service industry—is grouped together with transportation—a traditional service industry—hindering a modern/traditional distinction in their results.
5. The difference is consumption of fixed capital, akin to depreciation. All further references to value added indicate gross. Another concern with the formal/informal distinction outside manufacturing is a problem in the classification of non-manufacturing informal enterprises. Manufacturing enterprises with more than 20 employees (10 employees if power is used) must register with the government, and so are considered formal regardless of incorporation. Service sector formality derives only from incorporation. Hence, about 4% of unincorporated services firms that meet the employment threshold would be considered formal if they engaged in manufacturing, but instead are classified as informal.
6. The modern/traditional services split in the GGDC data suffers from the same problem noted in endnote 4, that modern-sector communications is aggregated with the traditional-sector transport and storage industries. Hence it is not strictly comparable to India’s data.
7. National accounts data includes a breakout for value added from services of owner-occupied dwellings, which is typically lumped with business services. Since these entail no employment component, they were excluded from value added attributed to modern services.
8. Further details are available in an online appendix (Green 2015).
9. For the exercises presented here “new jobs” means a rising headcount, net of replacing workers who leave the workforce.
10. The same limitations noted earlier about the East Asian GGDC data apply here, so formal and informal manufacturing get the same values in the covariance matrix.
11. Author’s calculations using data from Goldar (2014) and National Sample Survey Office (2011).

On 1 February 2019, a three-member delegation of Pakistani experts concluded an examination of the 1,000 megawatt (MW) Pakal Dul, 48 MW Lower Kalnai, 850 MW Ratle hydropower projects and the 900 MW Baglihar dam at the Chenab basin and found them to be operating according to the design. India also shared data about planned run-of-the-river.

hydropower projects at Balti Kalan, Kalaroos and Tamasha in the Jhelum and Indus basins. The last time India had shared such data was in 2013. Then came the terrorist strike on 14 February 2019.

Under the IWT, signed in 1960, India has control over water flowing in the eastern rivers—Beas, Ravi and Sutlej—while Pakistan has control over the western rivers of Indus—Chenab and Jhelum. Of the total 168 million acre-feet of the Indus basin, India’s share of water is 33 million acre-feet or just 20%. India uses nearly 95% of its share.

The deal was brokered by the World Bank after nine years of negotiation.

• Pakistan sought a guaranteed source of water, independent of Indian control. It interpreted sovereignty on the principles of maintaining status quo, or “prior appropriation.” In other words, since the people of Pakistan were already using this water, they have a claim on it and any curtailment by India would be an
  attack on its sovereignty.
• International water treaties do not recognise the validity of this principle. This doctrine implies “recognition of an international servitude upon the territory of one nation for the benefit of the other and would be entirely inconsistent with the sovereignty of the upper nation over its national domain.”
• The Government of India during the negotiations had adopted the position that the Indus dispute should not be settled using existing legal rights, but by accounting for potentialities of river development. India argued for an engineering, rather than a legal basis. Article XI(1) read,

“Nothing in this Treaty shall be construed as affecting existing territorial rights over the waters of any of the Rivers or the beds or banks thereof.” India has not relinquished its sovereign claims over the riverbeds.

• India’s first major concession was the agreement itself. Pakistan was allotted 80% of the Indus waters. Instead of any scientific rationale, sharing of water from the six rivers of the Indus river basin is based solely on the division of rivers.
• The second concession India made was to allow free-of-cost water flows into Pakistan. The 1948 Inter-Dominion Accord provided for Pakistan making payments for water based on the legal principle of India “owning” this water.
• The third concession was to relinquish 13 of the 16 pre-partition Punjab’s canal systems, though much of their river-heads were located in Kashmir, and to pay Pakistan $174 million for new works. The Indus Basin Development Fund Agreement specified the total cost of works in Pakistan to be $893.5 million.
• A consortium led by the United States (US) and the United Kingdom provided most of it as grants. Cold-war politics played a central role in heavily tilting the treaty in Pakistan’s favour. Pakistan was aligned with theUS while India tilted towards the Soviet Union.
• The international context is now fundamentally different. With the United Nations Security Council supporting India’s position on cross-border terrorism, the time is ripe to push for reciprocity. Climate variability is another new factor. Snowmelt and glacier melt comprise a significant portion of the water supply, and warming could increase variability of flows affecting seasonal requirements for agriculture as well as resulting in flooding.
• The fundamental flaw in the treaty is that Pakistan has not kept with its side of the grand bargain. The expectation that solving the Indus waters issue was a first step on the way to a Kashmir settlement has not been achieved.
• The treaty needs to be brought in line with other such bilateral treaties. It will not be easy as both countries have to agree, but a beginning should be made for shifting from “distribution” to “sharing” of the Indus waters.

Here in lies another disturbing aspect of India’s urban planning. Is it truly based on what is needed by the citizens and commuters, or is it motivated by political expediency and the greed of contractors?

The building of infrastructure with hardly any regard to commuter use—many of Mumbai’s sky walks and the monorail are prime examples—is a common phenomenon, and the media has time and again exposed how “blacklisted” contractors are hired for these projects.

• Is it not shameful that the vast engineering and architectural talent in the country often plays to the tune of politicians who have an eye on objectives other than the interests of citizens? Their engineering and architectural skills ought to be of service to the people and not for populist measures that help vested interests.
• Urban policies seem to encourage the lowest bidders rather than those with excellent professional credentials who can deliver the best services. This trend seems to be holding through in connection with most public infrastructure projects.
• Urban planning activists have time and again warned of the perils of policies that are automobile-centric, anti-public transport and geared towards the interests of private contractors. In fact, since the country’s rapid urbanisation has been chaotic, the urban infrastructure too tends to be random and piecemeal rather than mediated by context and public needs. Needless to say, corruption and lack of accountability both thrive in such a scenario.

The focus in Mumbai on metro systems and the coastal road project despite objections from urban transport and environmental experts is a familiar story with variations across the country.

On 3 July 2018, two persons were killed after a bridge collapsed onto the tracks near Andheri railway station in Mumbai. Immediately, a blame game followed and it transpired that while the railways had inspected the bridge between 2014 and 2017, there was no proper documentation to show for it.

• Before that, in 2017, a stampede on the Elphinstone railway bridge had claimed 23 lives and injured many others. The horrific stampede showed the complete failure of the authorities in harmonising land use with the massive and growing commuting population in this area.
• There are two other important facets to the recent bridge collapse. One is the post facto attempt made by the Shiv Sena to hold the growing population as the reason for the accident, which is a cynical attempt to shift responsibility.
• Mumbai, down the centuries, has attracted migrants and will continue to do so, given its employment potential. The city’s latest development plan itself speaks of creating eight million new jobs. Surely, urban planning should take into account the increase in commuter population that this will lead to? The other fact is the lack of any public outcry or protest campaigns over the third such tragedy in the past two years.
• It is as if citizens have simply accepted that these accidents and the connected deaths and injuries are the collateral damage of “development.” Governments have begun to simply ignore protests by citizen groups and carry on with controversial infrastructure projects.
• It is high time that infrastructure building be seen as a public service rather than as a distribution of largesse to contractors and other allied professionals. For this, it is imperative that citizens ask questions and demand answers persistently

And, yet, between them, the large internet platforms have suffered few, if any, consequences for their many misdemeanours. They have not been punished by the market (consumers and clients), they have not been swamped by competition, and they have certainly not been checked by governmental authorities. How did they get so powerful? Can they ever be held accountable for their actions? These are the pressing questions on the minds of academia and policymakers around the world. And, while Elizabeth Warren is today the most high-profile political proponent of a drastic solution, she may only be the first.

The Curse of Bigness

In calling for the likes of Facebook and Google to be broken up, Warren is echoing similar calls being made in academic circles and elsewhere. Prominent among these voices has been Tim Wu, a professor at Columbia Law School.

• Wu, famous for coining the term “net neutrality,” has, in his recent book The Curse of Bigness (2018), called for the adoption of a “neo-Brandesian” approach to the use of antitrust laws in the US, specifically in the context of internet platforms.1 Wu argues that the approach of the US Department of Justice has allowed internet platforms to amass economic power in a manner that could threaten the very future of democracy.
• He points out how in the past such concentrations of economic power, even by information technology companies (notably AT&T, IBM, and Microsoft), were effectively attacked using antitrust laws, resulting in the birth of the internet as we know it today (Kumar 2019).
• That is not the only argument that Wu makes in his book. He traces the political origins of the principal antitrust law in the US, the Sherman Act, 1896, and the underlying concerns which it was trying to address.
• He finds that the concerns were primarily political, that the amassing of such economic power as the world had never seen before in the hands of the robber barons of the so-called “gilded age” in the US were considered a threat to democracy.
• Election campaigns were fought and won on the pledge to break up the vast business empires of the likes of J D Rockefeller and J P Morgan. The speeches of the political leaders of the time, especially President Theodore Roosevelt, show that the concern was not just to address market failures, but also to stave off potential threats to democratic systems of governance posed by the so-called “trusts” which controlled businesses.
• This political aspect of antitrust, however, was lost since the 1980s as the so-called Chicago school of economists (led by Robert Bork) gained prominence. Wu argues that the Chicago school’s approach disingenuously tried to simplify antitrust law for lawyers and judges by reducing the whole field to the question of whether consumer welfare was being affected by a cartel or a monopoly.
• This approach also fit within the larger political move to shrink the state and give free rein to businesses, or the so-called “Reagonomics,” and this found favour with the governments and judges of the day. In Wu’s telling, this approach has allowed regulators in the US to ignore the harmful effects of internet platforms swallowing their competitors whole.

He points specifically to Facebook being allowed to acquire Instagram and WhatsApp without a murmur of disapproval from the authorities.

• Wu calls for a “neo-Brandesian” app­roach to the problem of tackling internet platforms’ dominance. While still short on specifics, The Curse of Bigness explains the broad outlines of the approach, harking back to Justice Louis Brandeis, a pioneering trust-busting judge.
• The approach calls for greater enforcement of existing laws to hinder the outright acquisition of competitors. Wu does not rule out the need to break up the existing tech companies to separate the various things to prevent them from getting access to more and more of our data.

For instance, he suggests a “de-merger” of Instagram and Whats­App from Facebook, allowing these platforms to compete rather than collude over users’ data.

Why Data Is Not the New Oil

Wu’s comparison of the near monopoly of present-day internet platforms with the Rockefeller-owned Standard Oil Trust might tempt one to conclude that “data is the new oil.” Inasmuch as data is a valuable resource and will continue to be so in the coming decades, it is true, but the comparison stops there. Data, unlike oil, is more valuable the more there is of it.

• While increased production of petroleum might lead to prices ­dropping, it is just the opposite with data. While petroleum can be mapped to jurisdictions and boundaries, data ­cannot. While petroleum, like all other natural resources, is finite, data is potentially infinite.
• As Wu points out, internet platforms owe their power to the network effect. With Google and Facebook offering their products free, there is little chance of a competitor being able to undercut them by price. Even when a competitor comes along with a better product, their accumulated capital allows competitors to be acquired swiftly, with little regulatory disapproval.
• Even if a competitor were to arise, they would be unable to compete on one key feature: data. Internet platforms probably know their customers better than they themselves do. The vast ecosystem of apps and devices which go along with the internet platforms means that incumbents will be virtually unassailable by entrants in the kind of service that they can provide their consumers.
• Seen in this light, Wu’s and Warren’s calls for antitrust action against internet platform companies cannot come too soon. They are of relevance for India too. While internet penetration in India is still low enough that it is possible for new entrants to compete with the incumbents (for example, the successful entry of Chinese apps into the Indian market) (Shaikh 2019), the concerns cannot be entirely brushed aside.
• The Competition Commission of India’s order in the context of Google Flights (John 2018), and the foreign direct investment in e-commerce policy limiting e-commerce companies from selling their own products (Badri Narayanan and Juneja 2019)2 are two instances of pushback on such concerns from regulators and concerned agencies.

The measures to prevent monopolisation of data in India could be both ex ante and post facto, depending on the situation and the powers of the regulator.

• Whether regulators in the US are alive to the fact or not, those around the world (especially in the European Union) are growing wary of the increasing economic clout of the internet platforms. The pushback has taken many forms, whether in the form of the European Commission levying one of its largest fines on Google for violations of EU competition law (Warren 2018), or India’s net neutrality regulations.
• If the US DoJ were to drop its laissez-faire attitude towards applying antitrust law and regulations on internet platforms, a new front may be opened in this long-running battle.

Notes

1. For the purposes of this column, the terms “competition law” and “antitrust” will be used interchangeably, though the latter is mostly used in the American context, while other jurisdictions discussed here use the former term.
2. However, the fact that this principle has not yet been extended to Indian e-commerce companies suggests protectionism rather than genuine concern over competition.

For decades, India’s statistical machinery has enjoyed a high level of reputation for the integrity of the data it produced on a range of economic and social parameters. It has often been criticised for the quality of its estimates, but never were allegations made of political interference influencing decisions and the estimates themselves.

Lately, Indian statistics and the institutions associated with it have however come under a cloud for being influenced and indeed even controlled by political considerations.

• In early 2015, the CSO issued a new gross domestic product (GDP) series (with the revised base year 2011–12), which showed a significantly faster growth rate for 2012–13 and 2013–14 compared to growth under the earlier series.
• These revised estimates were surprising as they did not square with related macroaggregates. Since then, with almost every new release of GDP numbers, more problems with the base year revision have come to light. In January 2019, for instance, the CSO’s revised estimates of GDP growth rate for 2016–17 (the year of demonetisation) shot up by 1.1 percentage points to 8.2%, the highest in a decade.

This seems to be at variance with the evidence marshalled by many economists.

• In 2018, two competing back series for varying lengths of time were prepared separately by two official bodies, (a committee of) the National Statistical Commission (NSC) and later by the CSO.
• The two showed quite opposite growth rates for the last decade. The NSC numbers were removed from the official website and the CSO numbers were later presented to the public by the Niti Aayog, an advisory body which had hitherto no expertise in statistical data collection.

All this caused great damage to the institutional integrity of the autonomous statistical bodies.

• In December 2018, the schedule for the release of results from the Periodic Labour Force Survey (PLFS) of the NSSO was not met. This was the first economy-wide employment survey conducted by the NSSO after 2011–12 and was therefore deemed important.
• Two members of the NSC, including the acting chairperson, subsequently resigned because they felt the NSSO was delaying the release of the report, though the NSC itself had officially cleared it. Subsequently, news reports based on leaks of the report showed an unprecedented rise in unemployment rates in 2017–18; this perhaps explained why the government did not want to release the report.

There have since been news reports that the PLFS of 2017–18 will be scrapped altogether by the government.

• In fact, any statistics that casts an iota of doubt on the achievement of the government seems to get revised or suppressed on the basis of some questionable methodology.
• This is the time for all professional economists, statisticians, independent researchers in policy—regardless of their political and ideological leanings—to come together to raise their voice against the tendency to suppress uncomfortable data, and impress upon the government authorities (current and future, and at all levels) to restore access and integrity to public statistics, and re-establish institutional independence and integrity to the statistical organisations.
• The national and global reputation of India’s statistical bodies is at stake. More than that, statistical integrity is crucial for generating data that would feed into economic policy making and that would make for honest and democratic public discourse.

Here is the most expensive infrastructure project of the country, with a record high unit cost of ₹ 1,200 crore per kilometre (km), but no functionality beyond electoral rhetoric. Whether this corridor can “decongest” the city roads is a black box given that the proposal of this project is not based on any extensive transport survey.

• If decongesting road traffic is the real intention, then why not first expedite the completion of the metro rail work across the city? The coastal road cannot be considered a silver bullet for the city’s infrastructural issues, which are not only varied but often mutually exclusive.For instance, if citizens must benefit from easy connectivity via the coastal roads, they must endure the degradation of their city’s inter-tidal ecology almost as a natural corollary.
• These hard choices, however, cannot be explained away simply as dilemmas of development. Evidences of coastal development in this country over the past two decades, and particularly in the last five years, indicate that such trade-offs are the result of an emerging partisan politics of welfare, which is characterised by a brazen display of corporate clientelism.
• From Gujarat to Kerala, vast stretches of the coastal lands are under corporate control through the state-abetted circumvention of regulations, especially in the name of special economic zone (SEZ), and the coastal regulation zone (crz) or the coastal management zone (CMZ) schemes.
• On the one hand, this encroachment has ousted traditional fishing communities from their ancestral lands, while on the other, a number of extractive industrial and construction activities in these zones are jeopardising their conventional livelihood.
• Simultaneously, it is hard to dismiss how the government policies, camouflaged by the rhetoric of “blue economy,” have “de-commonised” the sea and displaced the traditional (local) institutions of fisheries management.
• With the corporate-friendly policies framing the seas and the coasts as the new frontiers of economic opportunity and growth, private takeover of the marine resources is progressively squeezing out traditional fisherfolk from their native fishing grounds.
• Yet, be it the CMZ or the SagarMala, all schemes of the current government having to do with the fishery sector either make cursory or rhetorical references to the “blue economy.”
• Neither do these schemes or policy documents provide any comprehensive guidelines for the promised breakthrough, nor do they recognise the inherent heterogenity of the sector, for in doing so the government will have to face disconcerting questions on the eviction of the original inhabitants of the land in the name of “development.”
• In such a context, “rehabilitation” can gag those voices of concern that hold the government accountable for its failure to ensure the fundamental rights of its citizens.

The men are on death row for more than a decade, living solitary, tortured, and pitiable lives in jails. One of them turns out to have been a minor when the offence allegedly took place, but no heed is paid to that by any of the courts. When three of them find their death sentences commuted by the Bombay High Court, the Supreme Court overturns such commutation and awards the death penalty without hearing them at all. The ­Supreme Court then dismisses their review petitions without an open court hearing.

• It was only when the curative petition in the Supreme Court was placed before a totally different bench and the “error” was discovered that the wheels of justice finally began to turn.
• This case provides ample justification for the rule introduced by the Supreme Court that review petitions in death penalty cases must be heard in open court to bring in transparency to the ­process.
• In the end, the Supreme Court has acquitted all the ­accused, directed the state of Maharashtra to pay compensation to them, and attempted to hold the police accountable for this callous lapse. What it has not done is offer any sort of mea culpa for its own failings in this case.
• While the Supreme Court is fond of justifying the death penalty on the grounds that it must be applied only in the “rarest of rare cases” and after carefully weighing all mitigating and ­aggravating factors in a case, in reality it has been happy to discard the law at the drop of a hat. The judgment awarding the death penalty to the convicts in the “Nirbhaya” case, for example, is long on impassioned rhetoric and short on the law when it comes to awarding the death penalty.
• The very same bench that overturned the death penalty in the Ankush Shinde case also awarded the death penalty in another case on the same day, in a judgment that is empty of any analysis or reasoning on why the death penalty was merited in the matter (Khushwinder Singh v State of Punjab).
• To be fair to the Supreme Court, while it has commuted ­almost all death penalties in the recent past (11 out of 12 in 2018, for instance), the trial courts on the other hand seem to be more perversely enthusiastic about death penalties than ever.
• As the report by National Law University, Delhi on the annual statistics for the death penalty shows, 2018 paradoxically also saw the highest number of death penalties awarded by the trial courts since 2000. If the quality of the investigation, trials, and the judgments in the Ankush Shinde case are anything to go by, it is safe to assume that a large number of such death penalties have been awarded by the courts in cases with grievously faulty trials.
• And, yet, this hardly seems to inform bloodthirsty calls in the public for increasing the imposition of death penalty by the courts.
• Politicians of all stripes are happy to accede to these ­demands, which are ignorant or unconcerned about the ­vagaries of a broken criminal justice system that only ends up crea­ting more victims than punishing any hardened criminals. The courts at the lower level seem happy to go along with the popular mood, imposing the death penalty liberally despite the evident failings in the prosecution’s case or the ability of the ­defen­dant to mount a proper defence.
• It is then up to the appe­llate courts to try and remedy the injustice, but, as the Shinde case shows, they too are fallible. Even when they are eventually acquitted or find their death penalty commuted, the torment suffered by convicts as a result of the endless delays and flawed decision-making in a broken criminal justice system is rarely remedied.
• With the criminal justice system being what it is, it is hard to see the death penalty as anything but an institutionalised form of murder, one that unerringly chooses its victims from the ­oppressed and disenfranchised sections of society. With the political leadership unwilling to do the morally right thing, we can only hope that the Supreme Court will finally awaken to the reality of the situation and put in place an immediate moratorium on the death penalty.

What Is Epistemology?

The concept of epistemology, derived from the Greek word episteme (knowledge) and logos (reason) refers to the theory of knowledge. An important branch of philosophy, it is the study of the nature, origin and limits of human knowledge. The nature of knowledge is as important as the origin of knowledge in generating relevant epistemology.

• No scientific study can be evaluated or justified by the norms of faith or dictates of authority. For example, the discoveries of Copernicus (1473–1543) and Galileo (1564–1642) were epistemologically shocking to the College of Cardinals who had the monopoly of knowledge in the 16th century in Europe. Ultimately only scientific truth and not beliefs can promote and sustain progress.
• Kuhn’s (1962) view of the evolution of science as characterised by long periods of gradual “puzzle-solving normal science” followed by paradigm shifts offers an explanatory hypothesis about the nature of knowledge creation. Ola Olsson (2000: 254) argues that knowledge is created through convex combinations of older ideas and through paradigm shifts. We investigate in what manner this happened in economics.
• The nature of knowledge creation in the discipline of economics, has not been subjected to any in-depth analysis or interrogation. The almost unquestioned dominance (certainly during 1980–2008) of neoclassical economics in the academic profession and the rather pathological antipathy to Marxian epistemology and institutional economics has not been subjected to proper scrutiny.
• What I am concerned here is not Marxism as a creed but Marx’s unique contributions to the knowledge of understanding the dynamics of economic progress and the nature of the process of social history.
• I have provided a comprehensive analysis of economics courses in Indian universities based on an examination of curriculum of 26 Indian universities (Oommen 1987). This lengthy piece, inter alia, refers to the overemphasis on neoclassical economics and the deliberate de-emphasis of Marxian epistemology and economic history in teaching.
• The author was particularly struck by the dominance of Paul Samuelson’s Economics1 (this textbook first published in 1948, ran into several editions) which preferred “economic principles that display some of the logical beauty of Euclid’s geometry” as against that of Marx who defined economics as the science which studies how historically specific systems of economic relations originate, operate and change (Marx 1973: 852–53).
• Samuelson ignores the larger question of fairness in the distribution of wealth and income, but upholds “market efficiency” as the central theme of economics. An alternate textbook of the day imbued with considerable realism written by Joan Robinson and John Eatwell of Cambridge in 1974 titled An Introduction to Modern Economics rarely appeared in the syllabuses.
• My misgivings about the content of syllabuses and method of teachings, have persisted (Oommen 2004: chapter 1, 2012, 2017).2 Recent writings by Piketty (2015, 2017) gave the author hope and encouragement to raise the issue once again. After Piketty: The Agenda of Economics and Inequality, in which nearly two dozen scholars, economists and non-economists, have written, can be taken as a signal for rethinking and reflection on the epistemology of the discipline.
• It is high time that teachers and researchers of the subject ask fundamental questions about how the economy works and for whom it ticks. We may recall that most scholars of political economy in the 19th century placed the issue of distribution at the centre of their analysis, of course motivated by the great social changes they saw around them.

Neoclassical Economics

The neoclassical paradigm is treated as an immutable ingredient of social life, and many of the arguments are presented as non-falsifiable. But as Karl Popper rightly points out, all scientific propositions must be falsifiable. Further, so long as you treat prices, supply and demand, long-run equilibrium, full employment and so on, as an integrated set, the question of private ownership, vested interests, influence, power and such other factors that affect allocation and management of resources in real life, stand automatically ruled out.

• That the arbitrary whims and fancies of the top 1% or even a hardcore of 0.1% can alter the social priorities, needs and options of the rest of humanity itself and manipulate several such vital issues, are conveniently forgotten. Indeed the arc of neo-liberalism is not an innocent artifact but an engineered ideology to justify and perpetuate accumulation process and market society.
• If we accept that the ontological vocation of human beings is to transform the society in which they live, then no social science can be neutral between ends as postulated by the neoclassicals. Indeed economics is an applied social science “as medicine is an applied natural science. Biologists who do not see curing illness as their main job are not doctors even when associated with medical school” (Hobsbawm 1999: 128).
• Two historical events, namely, the Bolshevik Revolution of 1917 in Russia and the Great Depression of the 1930s checkmated the continued ascendancy of the neoclassical paradigm and influenced the growth of new economic ideas in the 20th century. The Bolshevik Revolution and the subsequent growth of industrialisation in the Soviet Union exploded the laissez-faire postulates and altered the political, economic and intellectual landscape of the world.
• The Great Depression also severely shook the prevailing faith in the free market capitalist economy. No wonder John Maynard Keynes’ General Theory of Employment, Interest and Money that sought to explain the Great Depression and to suggest remedies proved to be a spectacular success.The World War II, the post-world war rehabilitation, the growth of post-Keynesianism, the quest for social change in the postcolonial world, the emergence of development economics4 as a subdiscipline of economics, and the rise of innumerable development centres throughout the world (including India) gave a strong setback to the free market paradigm and to neoclassical economics which provided its ideological basis.
• However, with the collapse of Soviet Union, the fall of the Berlin Wall, the increasing questioning of the state spending on welfare and military hardware, the rise of Margaret Thatcher in the United Kingdom (UK) and Ronald Reagan in the United States (US) to power led to the re-emergence of the neoclassical paradigm of market-mediated growth.

During the last three decades prior to the financial crisis of 2008 that led to the downfall of Lehman Brothers and several others following the sub-prime lending, neoclassical economics influenced the world’s policy regime. Although challenged, this body of knowledge still rules the roost and continues to enjoy the patronage of most universities, notably in the US.

For those who embrace capitalism as an ideal, individuals are always rational and markets always work perfectly.

From the inception of the Nobel prize in economics by the Sveriges Riksbank of Sweden in 1969 till 2017, 49 prizes for 79 economists (39 of them are US citizens) were awarded. While the bank and the Royal Swedish Academy that give the awards have all the right to set the norms for the choice of awardees, even a cursory review of the citations and certainly the works of most of them show a pronounced bias in favour of neoclassical economics.

There are exceptions like Gunnar Myrdal, Amartya Sen, Paul Krugman and Joseph Stiglitz who could not be counted as neoclassicalists. Evidently, abstruse abstractions were chosen for research when the world faces abject poverty, growing inequality in wealth and income, intolerable inequity in the availability of opportunities, and such other acute societal problems. What Robert Solow (Nobel in 1987) while probing “inside the minds of 12 Nobel laureates” in his book Economics for the Curious observed recently (2014a) is worth citing:

The fundamental goal of economics as a discipline is to bring organized reason and systematic observation to bear on both large and small economic problems (and to have some intellectual fun on the way).

How can anyone say that economics is “for the curious” and the creation of “intellectual fun” as its avocation? Does the continued preference given to the neoclassical paradigm and its mathematical formulations provide the signal for future agenda for teaching and research? While so-called representative democracy is at best only a defender of capitalism (Dunn 1992), can the institution of social democracy or democratic values of social inclusion remain silent on questions like, for what and for whom the GDP (gross domestic product) is produced and how the claims on it are actually settled? Why economists shy away from questions of inequality and fairness in distribution? The Nobel prizes in economics do not seem to serve as a beacon for concerned social scientists to confront these and similar questions.

Richard Thaler of Chicago University, well known for his “nudge theory”

won the Nobel prize in economics for incorporating psychological assumptions into individual situations. While this is in refreshing contrast to the rational choice of neoclassical theories and opens up a rich area of study, the danger comes when policymakers like the World Bank use it as a theoretical legitimisation for inane policy options to fight poverty, inequality and the like. It is pertinent to read this author’s critique on World Bank’s “new set of development approaches” to contain poverty which to the bank is a behavioural distortion and that the poor alone have to be blamed for their plight.

• In this section, I wish to raise the question of why the Western academia as well as their colonial counterparts have treated Marxian epistemology and institutional economics with apathy while rewarding neoclassical economics and behavioural economics.
• The huge edifice of knowledge that Karl Marx created has been dismissed as poison and propaganda unsuitable for students. Can anyone consider Das Kapital and for that matter Marx’s writings in general as irrelevant knowledge? No one can refute the fact that Marx provides an excellent framework of analysis of capitalism. Mark Blaug, a noted historian of economic thought, points out that Marx wrote “no more than a dozen pages on the concept of social class, the theory of the state and the materialist conception of history” while he wrote “literally 10,000 pages on economics pure and simple”.
• Unlike the courses in macroeconomics and microeconomics that engage in abstractions, Marx does not see theory and practice or abstract and concrete as belonging to two distinct spheres. For Marx, what is abstract is closely related to what is concrete. In fact, the object of his epistemology was as much to interpret the world as to transform it. Those who love economics for its elegance and beauty are in a different moral universe altogether.
• As Thorstein Veblen (1857–1929), Clarence Ayres and other institutional economists argue, contemporary society has been considerably influenced by the folklores and mores of neoclassical economics (Ayres 1952, 1961). As noted above (and explained further in the next section) this paradigm is not an innocent artifact, but a strong ideological underpinning of the capital of the world to have their way and use all their might to perpetuate it as well exemplified in Piketty.
• Like Marxists, institutional economics founded by Veblen and enriched by a large number of economists rejected the neoclassical orthodoxy, including its most popular version by Samuelson, who attempted a grand synthesis of microeconomics (neoclassical economics) and macroeconomics (the Keynesian economics). Institutional economics proceeds on the assumption that human beings are interdependent and social and are influenced by social institutions.
• Institutional economics tries to show that the ability of a given society to use new problem-solving knowledge is limited by the patterns of social, political and economic dominations by the rich and powerful elites of that society. Here it is pertinent to quote two scholars on economic thought to show the indifference of the academic establishments towards these sources of knowledge.
• In many economics departments, the ideological domination of conservative neoclassical economists resulted in a situation in which the study of Veblen’s writings became personally, politically, and ideologically “unwise” as did the study of Marx’s writings. Evidence that a young economist took either Marx or Veblen seriously was often construed as evidence of intellectual incompetence.
• Consequently the institutionalist and Marxist schools of economic theory have remained small.
• Professional marginalisation of those who prefer socially relevant studies is a type of intellectual apartheid that must be fought for the sake of progress, alternate thinking and pursuit of relevant epistemology. It is in this setting that I wish to mention the arrival of Thomas Piketty.
• Several economists like Solow who got the Nobel prize in economics for their neoclassical scholarships have praised Piketty for the quality of his work and relevance of his findings (Solow 2014b). An empiricist par excellence, the historical evidence Piketty produced (covering 300 years and 20 countries) along with his reasoning and above all his accent on relevance rather than elegance should make many a Nobel pundit to sit back and reflect.

The Challenge Posed by Piketty

Social thinkers, economists, policymakers and ordinary people the world over have been shaken by Piketty’s book Capital in the 21st Century. That millions of copies have been sold in over 30 languages is something that the economics profession has to take note of. He has pushed forward our understanding of the working of the economy.

• His impact overflows into all social sciences. Economics has no special claim in explaining social reality and Piketty correctly treats economics as a subdiscipline of social sciences, alongside history, sociology, anthropology and political science
• Given the well-entrenched epistemological foundations of mainstream economics, it is very difficult to dislodge it.
• Boushey et al (2017) provides convincing evidence that show that the agenda for reforms of teaching, research and policy choices is formidable.
• Can the so-called “scientific” world of economists, close their eyes to the Gilded Age unfolding before them, ignore the reality of patrimonial capitalism, relentless “opportunity hoarding” and so on well underway in the world? I may mention a few lessons the discipline may do well to draw from the book:

1. history is the main source of knowledge of the economists and is the laboratory of the discipline;
2. price system knows no morality;
3. there is need for the democratisation of economic knowledge for continuously reinventing democracy. It is very important to note that economic rationality tolerates the perpetuation of inequality and in no way leads to democratic rationality (Piketty 2014: 551);
4. the destabilising growth of income and wealth is a writing on the wall and demands radical policy choices; and
5. in the context of “the past threatening to devour the future,” the call to the economics profession to bring back the political economy tradition of distribution into the centre of its discourse is very important.

Piketty has stressed the point that the profession needs a relook:

To put it bluntly, the discipline of economics has yet to get over its childish passion for mathematics and for purely theoretical and often highly ideological speculation, at the expense of historical research and collaboration with the other social sciences.

• Economists are all too often preoccupied with petty mathematical problems of interest only to themselves. This obsession with mathematics is an easy way of acquiring the appearance of scientificity without having to answer the far more complex questions posed by the world we live in.
• Hence they must set aside their contempt for other disciplines and their absurd claim to greater scientific legitimacy, despite the fact that they know almost nothing about anything.
• To conclude, I must say that Piketty’s work has upset the apple cart of mainstream economics. To be sure, the mighty defenders of the faith will fight back. Even so, he proved that the balancing forces of growth, competition and technological progress in the past have not succeeded in reducing inequalities and promoting social harmony.
• The world has before it incontrovertible evidence of widening inequality and its dangerous consequences (Oxfam 2014; World Inequality Lab 2018; Stiglitz 2012; Wilkinson and Pickett 2010; among others). Democratic societies and democratic values face serious threats. A global conversation and debate has to be initiated at the academic as well as the United Nations level sooner than later. Will the economics profession creatively respond? It is high time that we realise that pedagogy in graduate courses is sterile. The research agenda of universities also must radically change.
• If university departments are going to be funded by corporates, this can only be counterproductive. When the very large majority in the world has little access to quality education, good healthcare, good housing, quick transportation and no entitlements to participate in the market, why do the tribe of economists continue to chew the cud of neoclassical economics and call themselves “social” scientists? Certainly Piketty has not said enough about slavery, power, wealth dynamics, exploitation, and social justice, which should occupy the agenda of economics research and teaching to make it a relevant social science.