IGC-ISI India Development Policy Conference 2014

The 5th IGC-ISI India Development Policy Conference was held at Le Meridien, New Delhi, from 17 – 18 July 2014. The conference brought together researchers and policymakers to present and discuss the implications of research produced by the India Central Programme in Phase I and II of its operation, on various issues central to India’s growth and development, and to identify questions for future research.

The programme included research presentation sessions on Human Development, Governance, Macroeconomics and Finance, Energy, Enviroment, Infrastructure and Urbanisation.



Ideas for India 2nd Anniversary

The Ideas for India 2nd Anniversary Event was chaired by Prof. Ashok Kotwal, Editor-in-Chief, Ideas for India (I4I) and University of British Columbia. Prof. Banerjee, MIT, and Mukul Kesavan, Jamia Millia Islamia University, historian and contributor for the Telegraph, participated in the discussion.

I4I in the last two years has succeeded in the creation of national economic debate. In todays context, where the last national election led to a drastic change in the political panorama, I4I will represent even more a valuable platform for discussion on the political and economic national concerns and opportunities.

The discussion firstly verged toward the turnout of the last political election. As poverty and development indicators have vastly improved in the last ten years, voters gave a dramatic change the political direction of the country.

According to Prof. Kotwal, corruption, nationalism and frustrated aspiration of a fasting growing India are the likely causes for the Congress Party failure.

According to Prof. Banerjee, there are two possible causes that led voters’ disenchantment with the Congress Party. Due to the economic slowdown, Indian voters felt a strong sense of disappointment, seeking for a stronger leadership and success figure as Modi. A pervasive sense of struggling for basic needs and against rampant corruption, have also played a strong role in the will of political change, Prof. Banerjee stated.

Prof. Kesavan proposes pervasive corruption as one of the main reasons for the Congress Party’s fail. Also, the inability of the Congress Party to stop the concentration of provincial political powers and the sense of helpless indignation has led to the drastic election turn out.

The scholars also agreed that the lack of leadership in the Congress Party contributed largely to the political loss. The Congress Party, relying excessively on the dynastic nature of the party, blocked the emergence of local powerful leader proposing a less charismatic candidate.

On India’s future, according to Prof. Banerjee, Indian politics is going through a progressive era, passing from a highly corrupted society to a conventional democracy, where the power of local politicians is starting to be limited and impunity is not acceptable anymore. Although the rampant corruption, this transition was also due to the Congress Party that, in the last ten years, delivered a set of institutional reforms, leading the way to a new stage.

The discussion finally verged toward the role of elites in establishing extracting institutions and leading to poor governance and a failing state. Prof. Kotwal questioned whether local governments are turning into extracting institutions and what measures can improve local politics. According to Banerjee, historic reductionism proposed by the Robinson and Acemoglu theory of extracting institutions, cannot be helpful in analysing India’s current politics. The MIT Professor stressed instead the importance of the unfortunate feature of Indian politicians, who are often incapable of taking advantage of good economic opportunities.

Multiculturalism and Growth also were an important matter of discussion. According to Prof. Ashok, Indian ethnic heterogeneity still represents a challenge for collective action. BJP will need to coordinate and concertize with multiculturalism, having to decide whether being inclusive or exclusive in its political decisions.

According to Prof. Banerjee, diversity will not be a problem, since cultural and ethnic fractions are pervasive in India’s society, going well beyond the mere religious diversity. Instead, economic development could actually turn down any dangerous cultural revival.

Finally, the scholars agreed on the importance of further investments in Indian education, focusing on quality, accountability and inclusiveness of education at all different levels.

By Filippo Sebastio, Country Economist, IGC Bangladesh

Session 1: Environment, Energy, Urbanisation & Infrastructure (1)

The India Development Policy Conference’s first session, chaired by Sultan Hafeez Rahman (Country Director IGC Bangladesh), started with Professor Jacob Shapiro (Princeton) sharing updates and preliminary findings from his ongoing work with Oliver Vanden Eynde (PSE). The research team collected geospatial data on more than seven thousands new telecom towers (USOF), 350,000 km of new roads (PMGSY), three million projects providing households with drinking water (NRDWP), the Xth rural electrification plan (RGGVY) and 23,500 state-level infrastructure projects. While this dataset will soon be made publicly available to stimulate further policy-relevant research, it already allowed the researchers to answer a number of question that will affect the roll-out of future infrastructure programs. The available evidence shows that these programs were successfully targeted toward the regions most in need, but they encountered substantial implementation obstacles. First of all, Shapiro and Eynde showed that infrastructure provision cannot be successful without coordination between different programs due to complementarity between different services (e.g. Telecom towers and Electricity). Secondly, the efficient provision of infrastructure seems to encounter significant obstacles in areas where scheduled tribes account for a large share of the population and social unrest is more intense. Further work will shed light on issues of causality between conflict and infrastructure development, as well as the socio-economic impact of these infrastructure development programs.

The second presentation, delivered by Samuel Stolper (Harvard Kennedy School), looks into the issue of water pollution and it’s impact on health outcomes. This research was motivated by the evidence that Indian sewage capacity might fall short of the national needs by 70%, resulting in significant pollution of the country’s rivers which still represent one of the main source of water for a large share of the population. Using data from household surveys and monitoring stations covering 162 Indian rivers, the authors convincingly identified the impact of river water pollution on child mortality through IV estimation. Preliminary results show a strong impact of water pollution on child mortality, particularly significant on deaths occurring within the first month from birth. Moreover, this effect shows an high degree of persistence downstream, increasing the health cost of pollution’s externalities.

Arunish Chawla (Department of Expenditure, Ministry of Finance, Govt. of India) closed the session discussing the two ongoing projects and stressing the importance of further collaboration between researchers and policy-makers on the field of infrastructure development, environment and urbanization.

By Andrea Smurra, Country Economist, IGC Myanmar

Session 2: Environment, Energy, Urbanisation & Infrastructure (2)

Session 2 was chaired by Dr. Pronab Sen (Country Director IGC India Central) and Anil Jain (Planning Commission of India) was the discussant for the session. The presentations in the session include ‘How urban is India? ’ by Arindam Jana (Indian Institute of Human Settlements) and ‘Leakage in fuel subsidy: Evidence from Direct Benefit Transfer for LPG policy in India’ by Prabah Barnwal (Columbia University).

‘How urban is India?’ presents research findings on whether the parameters used in the Official Census to define urban settlements to capture the true extent of urbanisation in India. According to the official census, the three conditions that must be satisfied for settlements to be classified as urban are (a) Population greater than 5,000 persons (b) population density greater than 400 persons per square kilometre and (c) at least 75% of male main workers involved in non-agricultural pursuits.

The sensitivity tests of the parameters conducted by the authors show that population size and labour conditions have a limiting effect on declaring densely populated areas ‘urban’. In addition measurement errors are likely to arise due to sensitivity to process due to ex ante identification of “urban” settlements by Census.

The author asserts that the characterisation of extent and location of ‘urban’ India important, especially the part that is overlooked because these are important in urban policy and infrastructure finance discussions. There are important policy implications in case of large public sector programs like JNNURM, NREGA, etc. Moreover, small change in percentage of urban population would imply large changes in resource allocations. The research finds that the extent of “urban” India is sensitive to definition of urban settlement and the definition in itself is hard to defend.

The research is ongoing and the authors will undertake settlement level analysis for all India, estimate the metrics at various levels of geographical aggregation. The authors also propose ddefining rural-urban as a range between 0-1 rather than the existing binary: 0-truly rural, 1-truly urban, and all settlements that fall in the errors forming the range between 0 and 1.

Comments were received on the urban settlement criterion of economic activity in light of the vast improvement in connectivity in India which allows agriculture related occupation possible even when one is living in urban like settlements. In addition, authors must consider that most male hold more than two occupation on average when conducting the sensitivity test of parameters.

‘Leakage in Fuel Subsidies Evidence from Direct Benefit Transfer for LPG (DBTL) policy in India’ presents results of an evaluation of DBTL policy on how effective it was in controlling subsidy diversion, effect on recorded household LPG consumption and on black-market prices of LPG.

The identification strategy for estimating the causal effect of DBTL policy rests on phase wise rollout of the DBTL. The paper finds that introduction of DBTL policy is effective in reducing domestic LPG purchase (and thus the subsidy burden) by about 12 to 18% based on household transaction data. Since it includes impact on total domestic LPG refills (subsidized as well as non-subsidized), estimates indicate significant reduction in subsidy diversion. In addition, supporting evidence from survey data show that the impact on leakage is reflected well in LPG black market; DBTL roll back brings down black-market prices by about 15% in the DBTL districts.

The author concludes that introduction of DBTL policy is effective in reducing subsidy burden by controlling leakage. If effectively implemented, DBTL may save more than Rs.6000 Cr ($1 billion).However, frictionless implementation of DBTL must be paid due attention before government considers reintroduction.

Comments were received on the existence of three different price points for LPG (subsidised, non- subsidised and commercial end user price) is problematic as it provides an incentive for black market activity and subsidy diversion even with a DBTL policy in place. The large difference between subsidised and non-subsidised price may result in higher bargaining power of LPG delivery agents. The government is very keen to reduce subsidy leakage not only in LPG but also in kerosene and this research will provide insights to policy intervention in this regard.

By Farria Naeem, Country Economist, IGC Bangladesh

Session 3: Governance

The first presentation, from Karthik Muralidharan (University of California, San Diego) and Sandip Sukhtankar (Dartmouth College), focused on the use of biometric smartcards to deliver services to the poor in Andhra Pradesh. Two programs, the Mahatma Gandhi National Rural Employment Scheme (NREGS) and Social Security Pensions (SSP) were evaluated. Working with the government of Andhra Pradesh, researchers used a randomized rollout design to monitor the implementation of biometric smartcards across 8 districts, affecting a population of about 19 million people. The study, titled ‘Building State Capacity: Evidence from biometric smartcards in India,’ found that the new system paid for itself in saved beneficiary time (worth $4.44 million) and also reduced annual leakage by an estimated $38.7 million.

The discussants, Santosh Mathew (Ministry of Rural Development, Government of India) and Jayant Sinha (Member of Parliament, Lok Sabha), expanded the conversation to the political economy of ‘why this works where it works’ but fails elsewhere. Mathew asked: ‘Why are we struggling to make these additional investments to make sure some of our programmes work?’ Sinha highlighted the struggles of implementing projects where the ‘state doesn’t exist on the ground in many areas.’

In the second presentation, Lakshmi Iyer (Harvard Business School) presented a working paper on political identity and religious violence in India. Her work centres on the question ‘Does increasing minority political representation lower religious conflict?’ The paper is based on a new dataset of the religious identity of legislators and an expansion of the Varshney-Wilkinson database on riots to cover the period from 1980-2010. The Q & A session included a discussion of possible mechanisms through which political representation could affect levels of religious violence and directions for further analysis of these new datasets.

By Mari Oye, Country Economist, IGC Myanmar

Session 4: Macroeconomics and Finance

Chaired by Subir Gokarn from Brookings India, the session on Macroeconomics and Finance covered two research projects on capital flows in India, in particular foreign fund flows and borrowing in foreign currency. The first one, from Kiran Kumar (Indian Institute of Management, Indore) focuses on foreign fund flows and stock returns in India. Dividing firms in those with low and high innovation portfolios, the author finds that for high innovation portfolios returns reversal is less persistent, and returns are driven by differences in innovation in FII flows. This means that stocks with high innovations are associated with a coincident price increase that is permanent. On the other hand, stocks with low innovation are associated with a coincident price decline that is in part transient, reversing itself within two weeks. The results are consistent with a price “pressure” on stock returns induced by FII sales, as well as information being revealed through FII buys and sales. Moreover, the author concludes that there is a trade-off in the effect of FII flows on stock markets: while FII outflows contribute to transient volatility for stocks experiencing the outflows, trading by FIIs also generates new information. In addition, the price pressure effects are increasing in FII flow surprises and global “stress”. K.P. Krishnan (Dept of Economic Affairs, Ministry of Finance) stressed that FIIs are usually analysed in India as a single monolithic entity that centrally driven (partly as a result of data availability), which limit researchers’ ability to evaluate the effects of potential arbitrage opportunities among different subcategories. During the Q & A session, the issue on the differences between foreign and domestic flows was commented, as well as whether the efficient market hypothesis and the degree of risk sharing and consumption smoothing could be tested in India.

The second presentation, “Understanding foreign currency borrowing by firms”, Nirvikar Singh (University of California, Santa Cruz) assesses borrowing abroad and the degree of home bias in India estimating propensity score matching and a Tobit model using data of non-financial firms from 2001-2013 with sales greater than USD$1 million. Not surprisingly, but important, larger firms borrow a greater share of its total liability in foreign currency as they usually have better collaterals and liquidity position. At the same time, the percentage of firms naturally hedged through exports have increased to about 50% of the sample total, making foreign borrowing less risky on average; but inability to hedge limits the benefits of foreign borrowing for a sizable number of firms. Moreover, the researcher concludes that borrowing abroad has an impact on fixed asset growth, but this does not mean a sharp increase in output growth or export growth rates which would be expected in the cases of financially constrained firms. K. P. Krishnan commented on the policy implications this may have. If all firms were able to increase fixed assets by a few percentage points, similar to the growth rate achieved by firms borrowing abroad, this will translate in an overall increase in the country’s output and productivity. This raises the issue of facilitating external borrowing and opens the question of capital control from another perspective than the one normally discussed.

By Michelle Tejada, Hub Economist, IGC Hub

Session 5: Human Development - Health, Education & Labour Markets (1)

The first presentation by Seema Jayachandran (Northwestern University) tested the premise that gender attitudes are shaped in adolescence. The team is conducting an RCT in Haryana amongst 150 treatment schools and 164 control schools, surveying 14,800 students and 6,100 parents in total. The baseline survey was completed this year, and the team is going to measure the shifted attitudes of students and parents over the next two years. The experiment is using a number of tools to measure attitude shift in order to avoid social desirability bias including direct attitude questions, vignettes and implicit association tests. On average, the baseline survey found that while boys and girls tend to agree with pro-boy and pro-girl attitudes (respectively) in adolescence, girls seem to become more pro-male as they age. As for influence, the paper finds that parent’s gender attitudes affect children’s attitudes and mothers more than fathers. Adolescents are also equally influenced by their classmates. The main policy question here is: how do we ensure pro-girl attitudes persist into adulthood?

The discussant, Ratna Sudarshan (National University of Education Planning and Administration) noted that the goal of the study, to change attitudes to sex-selective abortion, is lofty and that any realistic change over a short period of time is impressive. She also challenged the assumption that richer people have more progressive gender attitudes.

The second presentation by Mehtabul Azam (Oklahoma State University) and Geeta Kingdon (University of London) asks which particular teacher characteristics improve educational achievement. They highlight that in India there have been rapid gains in enrolment and attendance but these have not resulted in skill gain, hence a new focus on teacher quality. Dr. Kingdon looked at two ways of measuring teacher quality: analysing qualifications and training (input based) or looking at their ability to consistently raise test scores of pupils (output based, or teacher value added). The authors used this latter definition and analysed a group of linked private schools in Uttar Pradesh, a total of 8,319 pupils and 191 teachers. Their findings show that there is no evidence that specific qualifications have a substantive impact on teacher quality, giving schools no indication as to who they should hire. However, teacher value added does have a substantive impact, which will help schools in deciding which teachers they should retain. However, Dr. Azam mentioned as a caveat that the study is only based on a small subset of private schools, so generalisability is limited.

The session concluded with Ratna Sudarshan asking the authors to analyse the impact of coaching on student achievement, as she believed too much emphasis was being placed on teacher quality vis a vis other factors that could improve performance.

The session was chaired by J.B.G. Tilak (National University of Education Planning and Administration).

By Helen Sims, Communications Coordinator, IGC Hub

Session 6: Human Development - Health, Education & Labour Markets (2)

The sixth session was chaired by Lakshmi Iyer (Delhi School of Economics). Presentations were by Diane Coffey (r.i.c.e. and Princeton University) who drew on her paper “Switching to Sanitation in South Asia: Qualitative Evidence on Health Technology,” followed by Rohini Somanathan (Delhi School of Economics) who presented “The Mixture as Before: Student Responses to the Changing Content of School Meal in India.” The discussions were lead by Sunil Jain (Financial Express).

Coffey asked why, in India, there is a very high rate of open defecation in rural areas, which has a serious impact on public health outcomes. Despite rising education levels, there is a very slow decline in the number of Indians defecating in the open, particularly in comparison to other less developed countries. 60% of people who defecate in open worldwide are in India. Poverty is not the explanation, as there is no apparent relationship between open defecation and GDP. Simple latrines are very inexpensive, but India has a low demand for them. Coffey conducted a rigorous qualitative study on four regions in India to look at if culture could explain why economic growth in India has translated into such a slow decline in open defecation.

Coffey’s study found that culture appears to shape the meaning of sanitation behaviours in people in rural India. Open defecation in these areas is culturally aligned with living a wholesome and healthy life, while a latrine implies ritual impurity and uncleanliness. This finding is supported by the fact that, when looking at religion, Hindus are more likely than Muslims to consider a latrine close to the house to be impure. People who do build latrines close to their homes make very expensive toilets to avoid the ritual pollution and social stigma. The cultural idea that cheap, simple latrines are polluting shuts poor people out of the latrine market.In addition, latrine use is perceived as an activity for weaker people, not the strong and healthy. Thus, there is no social pressure to encourage people to build simple latrines, and people who have them tend not to use them due to these cultural ideas. 1/3 of people in study who have a latrine don’t use it. 56% of households which have a latrine at least one person defecates in open.

Coffey concludes that, in order to stop open defecation, cultural ideas need to change. During the discussion mediated by Jain, several questions were raised. One issue was the topic of women’s safety and molestation, and Coffey indicated that her women respondents were less interested in latrines than you would think. This is partly because going to defecate is one of few chances they have to get out of the house, see their friends, etc. Another idea raised was that of community latrines to lessen the pollution stigma. However, if purity and health are wrapped up defecation, then people don’t want to defecate with their neighbours. Many issues arise of who cleans up communal messes, caste mixing, and fears of the pit filling and needing to be emptied. Coffey suggests that, to change cultural attitudes, a shift in policy from building latrines to latrine health education.

Somanathan then presented her paper “The Mixture as Before”, which looked at if the content of school meals affects school attendance in Delhi. Of particular note was that the changes were completely cost neutral, so she was able to see effects without worrying if they were worth the price. The Indian New Delhi school programme they observed is extremely cheap—about 3 cents per meal (USD), $3-6 per child per year. She looked at a natural experiment, where there was a phased roll out of changes in a New Delhi food for education programme. Schools switched over a period of time from giving snacks to giving out cooked lunches, which cost the same but were provided by different private vendors. The treatment was non-random, but she was able to look at attendance rates, menu data before and after transition, and some short term enrolment effects.

Somanathan found that school attendance went up with the transition to cooked meals, and the biggest effects were on the younger students. There appeared to be no intrinsic gender effect. She also looked to see if the type of food on the menu has an effect. By looking at the schools which offered two items (generally wheat and vegetable) as opposed to one item (such as porriage), she found a huge positive effect on youngest students’ enrolment immediately after programme roll out.

Somanathan qualified these promising results by cautioning that they likely vary by context. Food for education programmes need not be expensive, but she shows that design (timing, menus, providers) matter a great deal. During the discussion, Jain raised the issue that the study doesn’t look at educational outcomes and the effect likely wears out. However, it was heartening to see that, with the same costs, governments can be innovative and have a positive impact.

By Kyla Reid, Coordinator, IGC Hub