92458 S pe So ci ut ResearchDigest al h Is As World Bank su ia e on VOLUME 9 L NUMBER 1 L FALL 2014 Village Governments, Gender, and Public Goods Provision in Tamil Nadu How knowledgeable and well invest more in public goods preferred IN THIS ISSUE connected are female village by women and that reservations seem Village Governments, Gender, and Public presidents in Tamil Nadu? And how to improve women’s electoral poten- Goods Provision in Tamil Nadu … page 1 tial. They also suggest that female GP do they fare in providing public Does gender affect the provision of public presidents may serve as role models, goods by village presidents in Tamil Nadu? goods? shrinking gaps between boys’ and girls’ T aspirations as well as between parents’ Innovative Health Insurance in oday 74 countries have quotas for aspirations for their sons and those Andhra Pradesh—Has It Worked as women in government specified for their daughters. On the other hand, Planned? … page 2 in their constitution or national one study shows that while women in The Aarogyasri program aims to make legislation. One motivation for reserv- reserved seats bring more credit pro- lifesaving inpatient care affordable for ing government seats for women is grams to the poor, they fare less well the majority of the population of Andhra to address their status in developing in providing employment programs, Pradesh. Has it succeeded? countries: compared with men, women constructing new concrete roads, and A Profile of Private School Participation are worse off in health, income, job raising nontax revenue. Another study in Pakistan … page 3 opportunities, rights, and political documents a negative effect of women’s Who goes to private school in Pakistan? A representation. Political reservations reservations on programs targeting SC new study uses household survey data to attempt to improve women’s status in and ST households. find out society and to ensure their participa- A recent paper by Gajwani and tion in government; this could lead to Zhang sheds new light on this issue. Does Involving Local NGOs Improve more pro-women policies, which could Using data from 144 GPs in Tamil Nadu Outcomes? … page 4 include pro-child, pro-family, and pro- state in 2005, this study includes two A new paper looks at whether involving local health policies. unique aspects of GP presidents: their NGOs in malaria control efforts in India India provides an excellent setting knowledge of village governance and leads to better outcomes for evaluating political reservations. their interactions with higher-level offi- Anatomy of a Failure: An Ethnography of a Following independence in 1947, lo- cials. The findings show that female GP Project in Rural India … page 5 cal governance existed in many states presidents in Tamil Nadu provide fewer but varied widely. Uniformity in de- of the public goods that require contact Quantitative data suggest that a trial program to deepen democracy in rural India centralization came in the early 1990s with higher-level government officials. failed. What do qualitative data show? with the 73rd and 74th Constitutional On a knowledge test of GP presi- Amendments. The 73rd Amendment fo- dents’ duties and the panchayat sys- Money or Ideas? Constraints to cuses on rural areas and requires three tem, female presidents scored much page 6 Entrepreneurship in Rural Pakistan … tiers of sub-state governance called lower than their male counterparts: Which has greater benefits for microfinance panchayats. It also calls for the reserva- only 28 percent of female presidents clients—larger loans or business training? tion of panchayat seats for women and answered at least 11 of 19 questions A field experiment explores the issue Scheduled Caste (SC) and Scheduled correctly, compared with 90 percent Tribe (ST) persons. In particular, at least of male presidents. Gender strongly How Does Financial Development Affect a third of village panchayats—known as predicts test scores even after control- Firm Life Cycle? … page 7 gram panchayats (GPs)—must reserve the ling for presidents’ age, term in office, India offers an ideal laboratory for testing position of president for a woman. education, income, caste, religion, po- the effect of institutions on firm life cycle, and Previous studies of India’s reserva- litical party, and motivation for running a new paper does just that tions find mixed results. They show that female GP presidents in some states (continued on page 8) 2 World Bank ResearchDigest Innovative Health Insurance in Andhra Pradesh— Has It Worked as Planned? A state-funded program in Andhra a staff that operates a complex pre- for differences between the two states Pradesh has increased per capita screening and verification process in potential confounders at both dates. inpatient admissions and slowed underpinned by a sophisticated infor- Two facts help make this comparison mation technology system; a system more robust: Maharashtra’s health growth in out-of-pocket spending of payment settlement with empan- sector innovations were very modest T eled hospitals that is electronic and over this period, and Andhra Pradesh’s he Indian state of Andhra swift, based on a fixed per-procedure other large initiative was an ambu- Pradesh has been a pioneer in price schedule; a call center that not lance program whose effects on the health reforms over the past only receives calls from patients but outcomes of interest are likely to have decade or so. Its largest and costliest also makes calls to check up on pa- been modest. There may have been innovation has been the Aarogyasri tients after discharge; and a network differential changes between the states health insurance program, aimed at of ancillary health workers known as that the analysis does not control for, making lifesaving inpatient care af- “Aarogyamitras” whose role is to help however. fordable for the majority of the state’s patients throughout their hospitaliza- The results of this differences-in- population. tion episode, from before their admis- differences exercise are consistent Notionally Aarogyasri targets sion to their convalescence at home. with Aarogyasri increasing per capita poor households, but because of the A new paper by Bergkvist, Wagstaff, inpatient and surgery admissions and state’s high poverty line the program Katyal, Singh, Samarth, and Rao pro- reducing the growth of out-of-pocket actually covers around 80 percent of vides an analysis of Aarogyasri and its spending per inpatient admission— the population. The program’s focus impacts. Using data from a household the latter to such an extent that growth on around 900 high-cost procedures survey, the authors find that 70 percent of total out-of-pocket inpatient spend- (such as coronary bypass surgery, renal of families in Andhra Pradesh know ing was also reduced. This change transplantation surgery, and surgery that they are covered by Aarogyasri—a seems to be driven by a reduction in for spinal cord tumors) reflects its aim high figure that reflects the efforts of the growth of large expenditures on in- of covering catastrophic illnesses that the state’s government to make people patient care in Andhra Pradesh, consis- could potentially wipe out a poor fam- aware of the program and of their tent with the objective of Aarogyasri. ily’s savings; this sets Aarogyasri apart rights. In 2012 Aarogyasri financed an The results also point to Aarogyasri from many publicly financed health estimated 11 percent of all hospital reducing the average length of stay for insurance schemes around the world, admissions in the state, equivalent to hospitalizations, a finding that could whose “benefit packages” usually focus around 520,000 admissions; of these, be due to people seeking care earlier on cost-effective (and often low-cost) an estimated 33 percent were in public or to improved efficiency in hospitals. interventions. hospitals, the rest in empaneled pri- Not all agree that tax revenues Aarogyasri levies no copayments vate hospitals. should be used to finance high-cost at the point of use for covered proce- While the survey revealed that most procedures in high-end hospitals rath- dures. The program relies entirely on people knew they were covered by er than low-cost (and possibly more general revenues (at the state level) Aarogyasri, it also showed that people cost-effective) procedures in primary for funding, and there is no enrollment were confused about what Aarogyasri care. The fact that the program steers fee or enrollment process; enrollment covers: as many as 60 percent of those public funds toward private hospitals is automatic for those with the below- who said they were covered by and has also been a source of controversy. poverty line (BPL) card. The program knowledgeable about the program said Yet without Aarogyasri, seriously ill has a sophisticated set of arrange- (incorrectly) that it “covers all health people in Andhra Pradesh would have ments under which hospitals, both issues.” The survey also revealed continued to face the awful choice: public and private sector, can apply to that, contrary to the program’s rules, forgo a high-cost surgical or other life- become “empaneled,” but not all those people did pay out-of-pocket during a saving procedure and risk death, or get that apply are empaneled; hospitals hospitalization that they thought was treated but risk ending up destitute. under a certain size threshold, for ex- covered by Aarogyasri. Knowledge of ample, cannot be empaneled. Once “Aarogyamitras” was also limited: only empaneled, a hospital can be “disem- 3 percent reported interacting with paneled” by Aarogyasri if its perfor- one. mance falls short of the standards set The authors assess Aarogyasri’s im- Sofi Bergkvist, Adam Wagstaff, Anuradha Katyal, out in its contract. pacts by comparing changes in Andhra Prabal V. Singh, Amit Samarth, and Mala Rao. Aarogyasri also has an elaborate Pradesh following the introduction 2014. “What a Difference a State Makes: Health set of governance arrangements: a of the health insurance program with Reform in Andhra Pradesh.” Policy Research board of trustees with representatives changes in neighboring Maharashtra Working Paper 6883, World Bank, Washington, from across the state government; over the same period while adjusting DC. World Bank Research Digest 3 A Profile of Private School Participation in Pakistan Pakistan’s private school system in India, as the two countries have a The authors’ examination of private educates a third of students. Its shared history prior to both achieving school participation over the 2000s continuing performance hinges on their independence in 1947. using rounds of household survey The authors’ examination of cur- data from 1998/99 to 2010/11 provides care in designing any regulations rent private school participation using three main findings. First, private P household survey data from 2010/11 school participation rates grew mark- rivate schooling in Pakistan has provides six main findings. First, pri- edly in Punjab, Khyber Pakhtunkhwa, received growing attention in vate school participation for children and Sindh provinces. Private school recent years. Researchers are in the age groups 6–10 and 11–15 is participation rates also grew markedly studying it, the popular media is re- high: a fifth of children in Pakistan go in all socioeconomic subgroups ex- porting on it, and the government and to private school. This translates into amined. Second, the growth in private international development agencies a third of all students, since a sizable school participation rates contributed are grappling with its implications for share of the country’s children do not more to the growth in overall school education policy, development and go to school at all. But Pakistan’s na- participation rates for boys, children reform, and service delivery. tional and provincial rates of private from urban households, and children Pakistan’s private school system school participation do not stand out from households in the highest wealth has a long history, with its roots when compared with corresponding quintile than for other socioeconomic in the preindependence era. In the rates in India and its states. subgroups. Third, the growth in private early decades after independence the Second, as expected, private school participation was neverthe- system remained modest, consisting school students tend to come from less equalizing, particularly in Punjab largely of schools supported by reli- wealthier, more urban, and more province, where the shares of private gious and secular nongovernmental educated households than do govern- school students from households in organizations. These catered mainly ment school students and, especially, the highest wealth quintile and urban to high-income families in major cit- out-of-school children. Third, among households fell. ies. Starting in the 1990s there was a Pakistan’s four provinces there are The evidence indicates the impor- boom in private schools, leading to a not only differences in private school tance of the private school system dramatic structural transformation in participation rates but also, at times, in Pakistan. Some provinces and ter- the overall school system. This trans- qualitative differences in the charac- ritories have introduced legislation to formation is still under way. teristics of private school students regulate private schools. These regula- Today the private school system compared with government school tions reach deep into many aspects consists largely of institutions that are students. The composition of private of school operations, though they do for-profit, fee-based, secular, autono- school students also differs across not appear to have been applied in a mous, unregulated in practice, and provinces, with the sharpest distinc- broad, systematic, or meaningful way. with no direct government support. tions between Punjab and Khyber There is growing demand for new, A large segment of the private school Pakhtunkhwa provinces, on one side, more intrusive regulations, however, as system is also highly affordable, with and Sindh and Balochistan provinces, well as for stringent implementation school fees generally within reach for on the other. and enforcement of existing regula- poor households. Fourth, private schooling is highly tions. If applied, these regulations A new study by Nguyen and Raju concentrated, with more than 50 per- could weaken the growth and general uses recent rounds of household cent of private school students resid- dynamism and performance of private sample survey data that are national ing in 10 of the country’s 113 districts. schools. Alternative legislation that in coverage and representative at a These 10 districts tend to be wealthier does not take the approach of micro- low administrative level to build a and more urban, and most are in north- regulating the private school system profile of private school participation ern Punjab. Fifth, most of the variation may be more effective in protecting at the primary and secondary levels in in school participation among children consumers and staff of private schools Pakistan. The authors disaggregate the is due to the variation in school partici- while preserving fair and effective data in multiple ways to reveal pat- pation among children across house- competition. terns across selected socioeconomic holds rather than among those within subgroups (female and male, rural and households. This pattern is even more urban, household wealth quintiles), pronounced for private school partici- administrative divisions and spatial pation than for government school par- units (country, province, district), and ticipation. Sixth, the spatial patterns in Quynh Nguyen and Dhushyanth Raju. 2014. children within households. They also private school participation often over- “Private School Participation in Pakistan.” Policy compare the extent of private school lap to a large degree with the spatial Research Working Paper 6897, World Bank, participation in Pakistan with that patterns in private school supply. Washington, DC. 4 World Bank ResearchDigest Does Involving Local NGOs Improve Outcomes? Evidence from India suggests that two endemic districts in the state of NGOs in the delivery of public services, having nonstate actors deliver public Orissa. It turns out that the effect of underresourced areas such as those services may not lead to better NGO involvement on patterns in the studied by the paper often have un- use of mosquito nets and the seeking derresourced local nonstate capacity. outcomes in very poor areas of medical care for fever—the main Involving nonstate actors therefore will I targeted outcomes of interest—differs not necessarily result in better out- n an effort to enhance the delivery significantly between the two districts. comes. This lack of local nonstate ca- of public services in Using either simple pacity in many underserved and under- areas where resources regression decompo- resourced areas speaks to the complex are scarce, governments In low-resource areas sition or propensity interactions between poverty and local and development agen- with few high-quality weighting, the paper characteristics. In highly impoverished cies often seek to involve examines three po- areas a combination of locally targeted existing local nongov- local institutions, tential causes for investments and antipoverty policies ernmental organizations there may be little the differences in may be necessary to improve public (NGOs) in implementing outcomes: differences service delivery and, consequently, tar- programs. The develop- alternative to greater between the two dis- geted welfare outcomes. ment literature has a state investment or tricts in the charac- limited understanding of teristics of the popu- the challenges of involv- state-led initiatives if lation, differences in ing local nonstate ac- the goal is to improve the characteristics of tors, however, and many health workers, and questions remain about public services differences in the when, where, and how to characteristics of the do so. Discussions often NGOs. The analysis highlight the potential benefits from shows that the differences in observed leveraging local capacity, including outcomes are due mainly to the first expanding the reach of basic services and third factors as well as to their and improving quality. But collabora- possible interaction. tion between state and nonstate ac- These findings raise important tors also has potential pitfalls, includ- questions about the ability of the ing mistrustful relationships between health system to benefit from limited governments and their nonstate coun- nonstate capacity in underresourced terparts. So, what might distinguish a areas. They also raise questions about successful government-NGO collabo- the extent to which evidence taken ration from a troubled one? from interventions implemented by A recent paper by Das, Friedman, heterogeneous actors can be general- and Kandpal takes advantage of an ized. The paper thus contributes to the experimental evaluation of a malaria small but important literature on the control intervention in India to exam- external validity of small-scale policy ine whether involving local NGOs in evaluations, especially those conduct- program implementation has a signifi- ed in the nonstate sector. Contextual cant influence on the individual health knowledge and implementation capac- outcomes targeted by the intervention. ity are surely key determinants of the The paper finds that involving local effectiveness of development interven- NGOs in program implementation tions, and there are many potential can indeed lead to positive outcomes, nonstate actors with varying degrees though success likely depends on the of capability. Without a better under- quality of the NGOs. And in certain standing of the context and of the underresourced areas there may be capabilities of the study participants, little choice in the selection of quali- as well as standardized measures of Ashis Das, Jed Friedman, and Eeshani Kandpal. fied local NGOs, thus diminishing the implementation capacity, it will be dif- 2014. “Does Involvement of Local NGOs Enhance likelihood of success. ficult to generalize findings from any Public Service Delivery? Cautionary Evidence from The malaria control intervention relatively small-scale evaluation. a Malaria-Prevention Evaluation in India.” Policy studied in the paper was conducted More narrowly, on the direct ques- Research Working Paper 6931, World Bank, simultaneously by three NGOs in tion of leveraging preexisting local Washington, DC. World Bank Research Digest 5 Anatomy of a Failure: An Ethnography of a Project in Rural India Qualitative data on a trial program agreement in which the bureaucrats participation, greater competition over to deepen democracy in rural India committed to providing funding and funds, and high levels of corruption. pointed to subtle impacts that were technical support for the projects in Within this tsunami of change, the the village action plan. During the intervention was unable to add signifi- hard to measure with survey data third phase, which was spread over cant value. P a two-year period, each village was •  Lack of a “sandwich.” The facilita- rograms promoting citizen par- monitored for about two days a month tors essentially operated on their own ticipation to improve the quality by the facilitators to track citizen initia- rather than as agents of government of government at the local level tives, implementation of the village with the backing of local bureau- are the focus of both large amounts of action plan, and the actions and proce- crats, and the bureaucrats were not funding and intense debate. A recent dures of the village council. integrated into the change process. study by Ananthpur, Malik, and Rao The ethnographic investigation was Consequently, higher-level bureaucrats combined a randomized control trial conducted in 10 treatment and 10 con- did not respond adequately to any of a two-year citizenship training and trol villages over the four-year period gains made by citizens. This allowed facilitation program in rural India with 2007–11. Each village was assigned a local elites to reassert their interests an in-depth, four-year ethnography qualitative field investigator who was and reverse gains made by citizens. of the intervention to understand the from the region and therefore familiar •  Variation in the quality of facilitation. underlying mechanisms of change. with the setting and the dialect. Once a Only 4 of the 10 treatment villages had The study had two objectives: First, to month the investigators sent in a 5- to high-quality, intrinsically motivated analyze the impact of the program in 10-page report on important changes facilitators. Facilitators in the other improving the quality of participation that had taken place in the village. villages were less than diligent, and and accountability in democratically They were instructed to record impor- there was much evidence that they had elected village governments in India tant local events, interview key actors missed obvious opportunities to make and thus contribute to the growing lit- in those events, investigate new village a difference. erature on local participatory develop- construction projects and the financing However, the qualitative work was ment. And second, to demonstrate the behind them, track electoral activities also able to observe subtle impacts added value of careful qualitative work and expenditures, examine changes in that were difficult to measure with in understanding the mechanisms the levels of local activism, and inves- survey data—such as improvements underlying the impacts measured by a tigate other issues relevant to the po- in the responsiveness of village gov- randomized trial. litical and economic life of the village. ernments to the demands of citizens. The program was randomly as- These village reports, supplemented This suggests a need to rethink the ap- signed to 100 villages in the very poor by regular field visits by the principal proach to evaluating complex, hetero- and arid region of North Karnataka investigators, provided the qualitative geneous interventions. Understanding in Southern India, with 100 villages data for the study. the mechanisms of change requires serving as controls. A baseline sur- The quantitative data reveal no taking qualitative studies just as seri- vey tracking a variety of governance impact from the intervention. While ously as econometric analysis of quan- outcomes, citizen knowledge and par- they show that there were consider- titative data. Using “thick description” ticipation, and availability of village able positive changes over time, the based on careful and detailed qualita- infrastructure and public services was differences in the changes between the tive work could add considerable value conducted in 2007, and a final survey treatment and control samples were to standard impact evaluations. was carried out at the end of the inter- not statistically significant. vention in 2009. Using the detailed qualitative data, The program had three phases. the study was able to analyze why the In the first phase facilitators spent a intervention “failed.” This analysis un- week in each village training citizens in covered several reasons for the lack of participatory planning and deliberative observed outcomes: decision making and disseminating •  Challenging context. Elite domina- information about the village council’s tion, low levels of literacy, and high budgets, programs, and procedures. levels of inequality and conflict made At the end of the week the facilitators working in this environment particu- held a village meeting to finalize citi- larly difficult. Moreover, village bud- Kripa Ananthpur, Kabir Malik, and Vijayendra zens’ priorities and list them in a vil- gets increased by 700 percent between Rao. 2014. “The Anatomy of Failure: An Ethnog- lage action plan. In the second phase 2005 and 2007 as a result of govern- raphy of a Randomized Trial to Deepen Democracy a meeting was held with local bu- ment programs, and this provoked a in Rural India.” Policy Research Working Paper reaucrats to reach an implementation series of changes—increased political 6958, World Bank, Washington, DC. 6 World Bank ResearchDigest Money or Ideas? Constraints to Entrepreneurship in Rural Pakistan Which has greater benefits for planning, marketing, and financial the business decisions and, indeed, microfinance clients—business management. After the training ses- women involved in business spend training or larger loans? A field sions the lottery was introduced: loan less time in managerial decision mak- requests were subject to the usual ing than their male counterparts. This experiment explores the issue screening, and amounts approved suggests low decision-making power S above the usual cycle limit were for- among female business owners. elf-employment accounts for as warded to NRSP headquarters, where The analysis also finds little evi- much as 70 percent of employ- the results of the lottery were main- dence of a technology-based poverty ment in developing countries tained. Lottery winners could borrow trap: lottery winners showed no signifi- and is especially important among the approved amount, while those cant improvements in household wel- low-income households. Most self- who lost the lottery could borrow up fare from the extra credit received. This employed people, however, operate to their maximum loan size, which de- may suggest that the loan size under enterprises that are small, without pended on the number of loans they the current limit already meets the de- paid employees, and often poorly run. had previously repaid successfully. mands of most borrowers. Donors, financial institutions, and The offers of training and larger Business training did improve governments devote considerable re- loans were made according to a 2x2 financial decision making and labor sources to attempting to improve the design, with clients assigned to one allocation. In particular, among men efficiency of these enterprises, though of four groups: those in the first group offered business training, those with with varying degrees of success. Their were offered both business training low entrepreneurial ability were less efforts typically target either lack of and the chance to take out a larger likely to borrow during the lottery and finance or lack of managerial capital. loan; those in the second group were devoted less time to the business. This Access to credit is critical for the offered business training but not a suggests that attending business train- creation and survival of a business, larger loan; those in the third group ing helped people realize how success- and lack of credit can reinforce poverty were offered the chance to take out a ful (or unsuccessful) they were as en- if the production technology provides larger loan but not business training; trepreneurs. This is consistent with the returns that quickly taper off at low and those in the fourth group were of- practice of many programs around the levels of investment and offers higher fered neither. This design made it pos- world that use training as a screening returns at levels of investment above sible to test whether business train- device for later providing additional some threshold. Removing borrowing ing and a larger loan had reinforcing services, such as credit or mentoring. constraints could then allow liquidity- effects. Despite the limited welfare impact constrained individuals to access the Analysis of the results finds that of the larger loans, these were profit- more productive technology, increas- offering business training leads to an able to the lender because they led to ing their income and reducing the level increase in household expenditures no increases in default or in the work- of poverty. An alternative view suggests of about 6 percent or $46 a year, an load of credit officers handling them. that the critical constraint to entrepre- increase of 0.07 of a standard devia- In contrast, the business training pro- neurship in poor countries is lack of tion in a business knowledge index, an gram did not increase disbursement business skills or of managerial capital increase of half of a business practice or repayment significantly and thus more generally. out of four studied, but no significant was not cost-effective for the lender A new paper by Giné and Mansuri increases in sales or profits. In addi- despite increasing household expen- reports on a field experiment that tion, the increase in household expen- ditures for (male) clients by more than takes both barriers seriously by of- ditures is concentrated among self- the cost of training. This may help ex- fering rural clients of the National employed households that can put the plain why so few microfinance lenders Rural Support Program (NRSP)—a training into practice. Business train- offer such business training programs microcredit provider that is the largest ing also enhances group cohesion and voluntarily. partner organization of the Pakistan improves the general outlook on life. Poverty Alleviation Fund—an eight-day These effects are concentrated business training course and access to mainly among male clients, however. a loan lottery allowing eligible clients Among men, business training also to borrow up to seven times the aver- leads to a 6 percent reduction in busi- age loan size. ness failure. Like men, women increase The project randomly offered business knowledge, but they show no Xavier Giné and Ghazala Mansuri. 2014. “Money the training to half of 747 groups of improvements in any other outcomes. or Ideas? A Field Experiment on Constraints to borrowers from five different NRSP As it turns out, some 40 percent of Entrepreneurship in Rural Pakistan.” Policy Re- branches in three districts. Training female entrepreneurs report that their search Working Paper 6959, World Bank, Wash- sessions focused on business (male) spouses are responsible for all ington, DC. World Bank Research Digest 7 How Does Financial Development Affect Firm Life Cycle? A study in India finds additional level of managers, and there is little in large-firm-dominated industries are evidence that state ownership in evidence of growth occurring so that larger at each stage of the life cycle, banking impedes the impact of informal firms eventually become large in both capital-intensive and labor- formal firms. intensive industries. finance on growth Turning again to the formal sec- The findings contrast with the lit- I tor, the authors find that despite erature that finds significant effects ndia has large, persistent differ- considerable differences in financial of within-country institutional differ- ences across regions in institu- depth across Indian states, the effect ences on firm performance. The likely tions, business environment, of financial development is marginal explanation for the differences in find- and income and thus offers an ideal in explaining life cycles in the broad ings is that the state-owned and state- laboratory for testing the effect of population of firms and in most of the controlled financial sector in India is institutions on firm life cycle. A new subsamples analyzed. These results not contributing significantly to firm paper by Ayyagari, Demirgüç-Kunt, are robust to a number of checks, in- growth. In such a system the regional and Maksimovic does just that. Using cluding looking at just the firms in the differences in financial development as detailed manufacturing census data right tail of the size distributions, us- measured by the financial depth of the from India, the authors examine the ing alternative indicators of financial banking system may be masked by oth- relationship between plant size, age, development, and comparing results er, potentially more important institu- and growth and ask several questions: for states with flexible labor market tional factors (such as infrastructure How does local financial development regulation with those for states with constraints) and firm-specific factors influence the size-age relationship? rigid labor market regulation. The au- (such as organizational form). In addi- Are there differences in the size-age thors also find no differential impact tion, the finance literature shows that relationship across different industry of financial development on life cycle in a large cross-section of countries characteristics and between the formal when looking at firms created after a higher level of government owner- and informal manufacturing sector— India’s financial liberalization in 1991 ship of banks is associated with slower and do these differences vary with the or looking at periods after industry de- subsequent financial development and extent of local financial development? licensing. And they find only marginal lower income per capita growth and Does the effect of local financial de- differences in the proportion of manu- productivity, as well as poorer access velopment on firm life cycle vary with facturing employment in old and new to finance for small firms. The authors’ major regulatory changes in India, firms across different levels of financial findings are also consistent with re- such as financial liberalization, chang- development. search that shows that financial depth es in labor regulation, and industry Moreover, the authors find that does not promote growth of provinces delicensing? the extent of industries’ financial in China, another country where the Overall, the authors’ analysis shows dependence does not predict the life banking system is largely state owned that 40-year-old firms in the formal cycles of firms across states in India. and controlled. sector in India are on average two to Classifying industries as financially de- The findings on India thus provide four times the size of firms less than pendent or independent depending on additional evidence that state owner- five years old. These results hold when the extent to which firms can support ship in banking impedes the impact the authors take sampling weights into their capital expenditures using cash of finance on growth. These findings account and look at the entire popula- flow from operations (based on the underline the importance of increasing tion of firms. Thus while firms in India experiences of U.S. firms in the same the private sector share of the banking may not be growing at the same rate as industries), the authors find that firms system through further liberalization, those in developed countries such as in financially dependent industries are increasing competition, and reducing the United States, where the size-age larger at all stages of their life cycle. state control and political interference ratio is eight times, older firms in the Thus there is no evidence that financial —all as ways of potentially improving formal manufacturing sector are clearly dependence affects the growth rates of the efficiency of credit allocation in larger than younger firms. But the re- established firms relative to the growth India. sults show stark differences in firm life rates of firms in industries that are not cycle between the formal and informal financially dependent. Furthermore, sectors: older firms in the unorganized the results do not show that firms in manufacturing sector in India employ financially dependent industries are fewer people than firms less than five larger or face different life-cycle effects Meghana Ayyagari, Asli Demirgüç-Kunt, and years old. This is consistent with other in financially developed states than Vojislav Maksimovic. 2014. “Does Local Financial literature showing that informal firms in financially underdeveloped states. Development Matter for Firm Lifecycle in India?” look very different from formal firms in Using a similar difference-in-difference Policy Research Working Paper 7008, World size, productivity, and the education setup, the authors also find that firms Bank, Washington, DC. 8 World Bank ResearchDigest (continued from page 1) for office. This is somewhat surprising: require coordination with higher-level Recent Policy Research Tamil Nadu is a leader in women’s em- officials—and streetlights. For all other Working Papers powerment in India and has a higher fe- public goods, there is generally no sta- male literacy rate (64 percent) than the tistically significant difference in provi- 6870 Converting Land into Affordable Housing country as a whole (54 percent) accord- sion by male and female presidents. Floor Space ing to the 2001 census. GP presidents These differences are unlikely to be Alain Bertaud 6899 Private Non-State Sector Engagement in in the study’s sample also have com- driven by gender differences in public the Provision of Educational Services at the parable education levels: 47 percent of goods preferences: among 270 citizens Primary and Secondary Levels in South Asia: An Analytical Review of Its Role in School male and 41 percent of female GP pres- randomly surveyed across 27 villages Enrollment and Student Achievement idents have at least a high school edu- in the study, there is no statistically Mahesh Dahal and Quynh Nguyen 6901 Strengthening Malaria Service Delivery cation, while 82 percent of male and 74 significant gender difference in these through Supportive Supervision and percent of female GP presidents have at preferences. Community Mobilization in an Endemic Indian Setting: An Evaluation of Nested least a middle school education. These findings highlight a need, in Delivery Models Compared with male GP presidents, at least some parts of India, to improve Ashis Das, Jed Friedman, Eeshani Kandpal, G.N.V. Ramana, R. K. Das Gupta, Madan M. female presidents also have substan- women’s preparedness for the position Pradhan, and Ramesh Govindaraj tially less contact with important of GP president and their connections 6909 Hybrid Survey to Improve the Reliability of Poverty Statistics in a Cost-Effective Manner higher-level government officials. To while in office. Targeted training for Faizuddin Ahmed, Cheku Dorji, Shinya fund projects, a GP president must female presidents may increase their Takamatsu, and Nobuo Yoshida 6919 Regional Diversity and Inclusive Growth in connect with higher-level officials; understanding of their role, potentially Indian Cities only very low-cost GP projects can be improving their performance while in Ejaz Ghani, William R. Kerr, and Ishani Tewari 6953 Parallel Systems and Human Resource undertaken without consultation with office. Indeed, many of the female pres- Management in India’s Public Health higher-level officials. idents in the study complained that Services: A View from the Front Lines Gerard La Forgia, Shomikho Raha, Shabbeer Of the 47 women-reserved GPs in training was inadequate or nonexistent Shaik, Sunil Kumar Maheshwari, and Rabia the survey, only one failed to comply and that they received an overwhelm- Ali with the reservation policy by seating ing amount of paperwork without 6961 How Survey-to-Survey Imputation Can Fail D. Newhouse, S. Shivakumaran, S. a male president. No women served as accompanying support in understand- Takamatsu, and N. Yoshida president in unreserved GPs. Using 10 ing it. While some findings in other 6975 Managing Quantity, Quality, and Timing in Indian Cane Sugar Production: Ex Post measures of public goods, the authors states show public goods successes for Marketing Permits or Ex Ante Production compare public goods provision across women in reserved seats, Tamil Nadu Contracts? Sandhyarani Patlolla, Rachael E. Goodhue, reserved and unreserved GPs from 1991 illustrates that these successes may not and Richard J. Sexton (before the 73rd Amendment) to 2005. happen automatically. 6987 Integrating Border Regions: Connectivity and Competitiveness in South Asia They find strong evidence that female Massimiliano Calì, Thomas Farole, Charles presidents in reserved GPs built fewer Kunaka, and Swarnim Waglé schools and roads—two public goods that require relatively more contact with higher-level officials. There is also Kiran Gajwani and Xiaobo Zhang. 2014. “Gender weak evidence that women provide and Public Goods Provision in Tamil Nadu’s Working Papers can be downloaded at http://econ.worldbank.org fewer household toilet and drinking Village Governments.” Policy Research Working To download the World Bank Research E-Newsletter, water connections—both of which Paper 6854, World Bank, Washington, DC. go to http://econ.worldbank.org/research_newsletter The World Bank Research Digest is a quarterly publica- The Research Digest is financed by the Bank’s Editorial Committee: Indermit S. Gill (managing editor), tion disseminating findings of World Bank research. Research Committee and managed by DECRS, the Aslı Demirgüç-Kunt, and Shiva S. Makki. Editor: Alison The views and interpretations in the articles are those research support unit of the Development Economics Strong; production: Roula Yazigi. For information or of the authors and do not necessarily represent the Senior Vice Presidency (DEC). The Research Digest is free subscriptions, send email to research@worldbank. views of the World Bank, its Executive Directors, or the not copyrighted and may be reproduced with appropri- org or visit http://econ.worldbank.org/research_digest. countries they represent. ate source attribution. The World Bank 1818 H Street, NW Washington, DC 20433, USA Printed on Recycled Paper