Paper Round-up: Pacific Conference for Development Economics (PacDev) 2019

These short summaries of talks delivered at PacDev 2019 were written and curated by CEGA staff with an introduction by Vittorio Bassi, CEGA affiliate and Assistant Professor of Economics at the University of Southern California.

Credit: Mauricio Romero

PacDev is the largest conference on Development Economics on the West Coast, and one of the leading annual Development Economics events in the US.

This year’s conference took place on Saturday, March 16th at the University of Southern California and was organized by the Department of Economics and the Center for Economic and Social Research (CESR). The organizing Team was led by Professor John Strauss and included many faculty members from USC Economics and CESR, as well as a great team of Ph.D. students. Bringing together over 150 scholars, the conference featured a keynote speech by Professor Duncan Thomas of Duke University. Around 80 presentations were given at the conference, from a mix of Ph.D. students, junior and senior faculty members from leading research institutions in the US as well as overseas.

We hope the short summaries of each talk, organized by session below, are useful.

Enjoy!

Session 1A: Development Programs

• A new subway line increases rush-hour speed in nearby roads by 5%, and city-level data shows that there is a clear substitution between subway and bus. However, the benefits from the saved time are low compared to the cost of construction and operating a subway. (Gu et al.)
• What are the impacts when public policies are implemented simultaneously, rather than in isolation? Data from two programs in Brazil suggests that they had little if any impact on each other in the domains of land productivity, agricultural income, and child labor. (Costa et al.)
• “Land market misallocation reduces China’s overall welfare by 0.8%. This loss of welfare may be attributed to weakened firm agglomeration forces led by the distorted land prices from the land market misallocation. (Fei)”
• Studying the area near redeveloped mill sites in Mumbai shows there are substantial housing and agglomeration spillovers from the construction of modern, formal sector housing. Pro-formalization efforts might hurt the welfare of poor residents by reducing the stock of informal housing necessary for access to jobs in central locations. (Gechter and Tsivanidis)
• Data on water thefts in the Pakistani irrigation system show that short-run deficiencies in rainfall drive villages to take more than their allotted share of surface irrigation water. Long-run contamination of groundwater due to industrial pollution gives villages both time and incentive to establish better mechanisms for preventing overuse of surface irrigation water. (Muhammad Haseeb) *working paper forthcoming

Session 1B: Political Economy 1

• Voters in Brazil responded to information on the quality of service provision. In local elections vote shares increased by 2.15 percentage points for schools in the top 20% of the distribution of school quality, and the vote share decreased by 1.6 percentage points for schools in the bottom 20% of the distribution. (Dias and Ferraz)
• Response to monetary fines increases voter turnout in Peru. Significant heterogenous effects exist, but in general, monetary fines explain a small fraction of the aggregate effects of mandatory voting on turnout. (Gonzales et al.)
• Citizens in Pakistan responded to a low-cost and scalable technology called Interactive Voice Response (IVR). This intervention increased political communication in Pakistan and shows that constituents engage more with the democratic process when contacted by politicians using IVR. (Golden et al.)
• Politicians in India informally subsidize constituencies by tolerating manipulation of electricity bills. This manipulation causes frequent outages and reduces the quality of the service, which adversely affects consumer welfare. (Mahadevan)

Session 1C: Child Development

• Children born between 1995–2005 to mothers exposed in utero to the 1970s-era US bombing campaign in Cambodia have significantly lower weight-for-age z-scores. (Moyano)
• When children ignore parents’ unsolicited advice, the resulting Nash-equilibrium is Pareto suboptimal. Better information regarding the child’s need for parental involvement can improve the payoffs for both parents and children. (Tushar Bharati) *working paper forthcoming
• Data from Indonesia show that parents make reinforcing investments in their children’s health and education in early childhood, but do not continue this practice in later childhood. (Banerjee and Majid)
• Children exposed to the Bhopal gas disaster, the worst industrial disaster in India, while in-utero suffer adverse effects. These children are 7 percentage points more likely to be suffering from cancer compared to children 10 years or older at the time. (Arushi Kaushik) *working paper forthcoming

Session 1D: Labor

• Workers in the casual daily labor market in India are privately willing to accept agricultural work at wages below the market wage but almost never do so when this choice is observable to other workers. Consequences are substantial; workers will give up 26–49% of their weekly earnings to be seen as following social norms. Finally, consistent with the idea that social conformity could have aggregate implications, measures of social cohesion correlate with downward wage rigidity and business cycle volatility across India. (Breza et al.)
• A nationally representative survey of firms in urban Peru shows that (i) there is a substantial mismatch between requirements of the job and characteristics of new hire and (ii) employers rely substantially on networks and signals in their hiring process. ( Lee and Macaluso) *working paper forthcoming
• Access to information about matatu drivers location, productivity, and safety in Nairobi allows matatu owners to re-optimize their contract. Matatu drivers with re-optimized contracts exert greater effort, decrease risky behavior that damages the vehicle, and under-reporting revenue by less which leads to an overall increase in firm profitability. (Kelley et al.)
• Bandiera et al study the job search process of young workers in Uganda using a six-year field experiment. The main findings are that skills and beliefs about work prospects both shape how young workers search for jobs. Job search interventions that do not take this interaction into account can backfire by creating disappointment and distorting workers’ search strategies. (Bandiera et al.) *working paper forthcoming

Session 1E: Migration 1

• While current models of migration focus on existing network structures, this paper outlines a new model focused on stopping time and matching decisions. By combining a theoretical algorithm and lab tests, this model shows how uncertainty can result in inefficiencies in when and where to migrate. (Barnett-Howell)
• Kenyan migrants hide earning information from those remaining in the rural areas, which causes rural individuals to underestimate city earning and not migrate. Increased information increases the likelihood that individuals migrate to cities and report higher salaries. Learning about the average degree of hidden income increases planned migration to Nairobi by 13 percentage points. (Baseler)
• A migrant recruitment agency rating program in Sri Lanka helped reputable foreign employers in the Gulf countries match with highly rates Sri Lankan agencies, resulting in large reductions in reported physical harassment and abuse by Sri Lankan migrants. (A. Nilesh Fernando) *working paper forthcoming
• Cash incentives in Bangladesh increase seasonal migration, but do not increase permanent migration. Eight years after the treatment, agricultural households that received the incentive are less likely to have permanently left their village of origin. (Chowdhury et al.)

Session 2A: Demography 1

• Oil palm cultivation in Indonesia is associated with substantial reductions in fertility. These effects can best be explained by rising incomes at the farm level, an expanding non-agricultural sector, and increasing returns to education. (Kubitza and Gehrke)
• The season a child is born in China affects his or her outcomes later in life. There is a significant height disadvantage for girls born in winter, but the height disadvantage does not translate into deficits in skills in adolescence, which can be explained by compensatory educational investments by parents. (Maitra et al.)
• Why did the industrial revolution occur in Europe and not in Asia? Kumon et al take a unique approach to this question by exploiting a rich village-level dataset of Ancient Japan, showing empirically that Japan’s lack of industrialization may have been ironically due to high levels of equality, which caused the economy to be poor. (Kumon)
• The extremist group Boko Haram pulls young men into its ranks partially due to characteristics of the marriage market in rural Nigeria. Rexer et al use an instrumental variable approach to show that norms of paying a bride-price and polygamy lower marriage prospects for young men and result in them being more likely to join Boko Haram and increase the incidence of militant activity by the group. (Rexer)

Session 2B: Political Economy 2

• When state capacity is weak, non-monetary incentives may affect civil servant performance. In line with this hypothesis, an SMS campaign with personalized behavioral content proves to be a cost-effective strategy to enforce compliance with national policies among civil servants in Peru. (Dustan et al.)
• Do village council presidents favor their own households? Suggestive evidence shows that corruption takes the form of “performance pay” used to attract talented candidates, as winners of close elections receive 3 times as many days of labor as losers. (Shenoy et al.)
• Could adapting mobile phones improve “last-mile” public service delivery? An at-scale experiment in India shows that informing officials that program implementation would be measured via calls with beneficiaries led to a 3.9% increase in farmers receiving transfers on time, and a 1.5% increase overall. (Muralidharan et al.)
• Improvements are found in learning outcomes, teacher behaviors, and parental engagements for three teacher-focused interventions combining a community-based social accountability intervention with pay-for-performance schemes. The strongest improvements are found for the intervention that combines social accountability with a pay-for-performance scheme that focuses on a narrowly-defined indicator (presence) that can be objectively verified (using a tamper-proof camera). (Gaduh et al.)

Session 2C: Health 1

• Positive early-life income shocks increase children’s human capital, increasing the returns to both schooling and child labor in India. In places with low child labor, the increase in the returns to education leads to greater educational investment later on. In places with high child labor, the increase in child wages leads children to drop out early. (Bau et al.)
• Rates of return migration to Mexico increase in districts that gain access to free health insurance through Mexico’s Seguro Popular program. Effects of the program on return migration vary by age, sex, and education, with males, younger adults, and those with less than 12 years of education experiencing the strongest impacts. (Knox and Lara)
• A new premium subsidy scheme, in which relatively wealthier households had an increased fee and those with low socioeconomic status received full subsidies, is not an effective policy tool to promote take-up of insurance and universal health coverage. (Sjöholm)
• This RCT in Mali varied information about a discount to understand what drives overuse of antimalarials at public health clinics. We find that patient demand can cause overuse and lead to a worse match between treatment and cause of illness. By contrast, doctors do not use their information advantage to sell more powerful malaria treatment or increase revenue. (Lopez et al.)

Session 2D: Firms/Entrepreneurship

• A Diff-in-Diff approach finds that access to credit increases Total Factor Productivity (TFP) among High-productivity firms. However, it finds that those high-productivity firms are not more likely to receive the credit which the paper argues is an indication of misallocation of credit. (Banerjee et al.)
• What are the equilibrium effects of “de-reservation” of products that allows large firms to produce products that have been reserved for small firms? A diff-in-diff in India finds that an imperfect implementation of the policy increases the manufacturing GDP by about 0.4%, due to the size and scope of the firms. (Chiplunkar)
• Are supply-side constraints preventing small entrepreneurs from adopting digital payments in India? Data from Jaipur suggests that there is relatively low adoption (40%) of digital payment technologies among small-scale fixed store merchants, but these are not clearly explained by the supply-side constraints. (Trachtman et al.)
• What is the impact of access to randomized contracts to build schools and roads on future firm and labor outcomes? Authors find that access to the contracts increases future sales and revenue of firms substantially and revenue from private contracts happen with lags than those from public offers. (Matthew Pecenco) *working paper forthcoming

Session 2E: Migration 2

• What is the effect of living in a refugee camp versus self-settlement? Authors find that although living in a refugee camp does reduce household income, the gap is less than the rent saved by living in a camp. These camps can be an effective subsidy for refugees willing to opt out of urban areas. (Ginn)
• What consequences faced Cambodian refugees who emerged around the collapse of the Pol Pot regime (1975–1979)? Authors find substantial heterogeneity in the impacts of forced displacement on the educational and labor market outcomes between and within the two returnees from neighboring Thailand and Vietnam. (Masahiro Kubo) *working paper forthcoming
• By comparing conflict driven internal migration in Colombia in the 1990s against recent economic driven international migration of Venezuelans into Colombia, authors find that internal migrant inflows did not affect voting behavior, whereas international migrant inflows reduced the support for the incumbent party and increases that of right-wing candidates. This supports the idea that negative attitudes toward immigration result from “sociotropic” motives, where voters view migrants as a threat to local customs and social norms. (Rozo & Vargas)

Keynote: Duncan Thomas (Prof. of Economics, Duke University)

The Study of the Tsunami Aftermath and Recovery (STAR) showed that post-traumatic stress reactivity immediately after the disaster differed significantly among communities based on how badly they were affected by the tsunami, however, differences diminished across time. The data on the long-term effects on the communities’ health show that CRP (a measure of stress) and other characteristics went up at 2, 5, and 13-year follow-ups.

Prof. Duncan Thomas gives a keynote presentation on mortality by tsunami exposure. (Credit: CESR)

Session 3A: Gender

• In a context with a larger gender gap in political participation and limited female mobility, non-partisan mobilization only improves women’s turnout when it is targeted at men within households — targeting women alone is insufficient. (Asad Liaqat) *working paper forthcoming
• What is the relationship between having access to a high-quality communication network and economic participation for women in Pakistan? By being able to communicate with their households while traveling, data from the Pakistan Social and Living Standards Measurement survey shows that women in rural areas with good quality communication networks have 4–7% more economic participation and women in urban areas have 9–12% more economic participation, compared with bad quality communication networks. (Asad)
• Using data from multiple different contexts, it can be shown that women who inherit property are associated with a higher prevalence of cousin and arranged marriages, as well as lower economic participation and premarital sexual freedom. (Bahrami-Rad)

Session 3B: Public

• An evaluation of two Pakistani programs finds that both public disclosure and social recognition of top taxpayers causes a substantial increase in tax payments. These findings point to a cost-effective policy option for governments seeking to promote compliance. (Slemrod et al.)
• A field experiment in Ethiopia finds that donors reduce altruism/guilt-based transfers to recipients who don’t take up insurance and that these effects are larger for donors who hold a stronger ex-ante belief that the recipient would take up insurance. The model that investigates welfare implications highlights how the introduction of formal insurance may erode the norms of giving and lead vulnerable groups to face more volatile consumption. (Anderberg and Morsink)
• A large-scale RCT with administrative data of non-mandatory pension contribution in Mongolia reveals that their contribution rate can increase by knowing the mobile payment system or the presence of foreign aid to the pension office. (Junichi Yamasaki) *working paper forthcoming
• Using data from the worlds largest in-kind food subsidy program in India, this paper finds that in-kind food subsidies reduce household labor supply and raise the equilibrium wage, particularly on casual low skilled labor. This shows that in-kind food subsidies can improve welfare through a labor market effect in addition to their direct effect on food consumption and nutrition.(Shrinivas et al.)

Session 3C: Behavioral

• What’s the impact of enrolling in better college on cognitive, non-cognitive, and behavioral outcomes? Enrollment records from Delhi University in India show that risk preference increases more for women, conscientiousness reduces much more for men, and overconfidence reduces for both. (Dasgupta et al.)
• Data from lab experiments, with subjects from informal settlements in Nairobi, show that stressors do have statistically significant effects on behavioral measures of time discounting, but not on executive control and self-efficacy. These results may be due to the study setting since this population faces persistent stress every day. (Haushofer et al.)

Session 3D: Education 1

• This study experimentally evaluates a program that combines technology-based instruction with behavioral principles to improve math learning in Chile. Results indicate large gains in math achievement (0.27 of a standard deviation) on the Chilean national standardized examination. (Araya et al.) *working paper forthcoming
• This study provides evidence that arbitrary clock conventions — by generating large discrepancies in when the sun sets across locations — help determine the geographic distribution of educational attainment levels. When the sun sets later, children go to bed later; by contrast, wake-up times do not respond to solar cues. Sleep-deprived students decrease study effort. School-age children exposed to later sunsets attain fewer years of education. (Jagnani)
• Using the Cultural Revolution in Iran as a natural experiment, this study shows that when the demand for education is high, government interventions have little effect on higher education outcomes. (Majbouri)
• A randomized evaluation in Tanzania finds that despite the theoretical advantages of a Pay for Percentile incentive scheme for teachers, a scheme based on multiple proficiency thresholds was more effective at increasing learning and reducing grade repetition. (Romero)

Session 3E: Networks

• Experimental evidence on a new mechanism for spreading awareness about technology. Demonstration plots, where partnering farmers cultivate a new technology side by side with an existing one, increase awareness relative to a control where new technology is demonstrated on its own. Demonstration plots were also most effective for farmers who were the most isolated.(Sadoulet et al.)
• This study compares two ways to improve access to basic sanitation in India, price subsidies or loans. By creating a model based on household demand for sanitation, the author finds that price subsidies are most effective if the objective is to reach the maximum number of households. However, in villages with close to zero initial coverage, loans proved to be more effective.(Gautam)
• Examining an after-school program in India where participants are selected via popular vote, this study finds that students who were not selected experienced reduced educational aspirations and self-confidence. These negative effects are not found when participation is determined randomly. (Delavallade et al.)

Session 4A: Program Evaluation

• Although attrition is a common threat to internal validity in randomized experiments, a systematic review of the literature indicates that there is no consensus on how to test for attrition bias. We propose sharp tests of internal validity in the presence of attrition for either the respondent subpopulation or the study population as a whole — and also demonstrate that the most commonly used test in the literature is not an appropriate test in general. (Ghanem et al.)
• This study examines the demand for research evidence among local political officials (mayors) in Brazil and whether providing information on the impact of specific policies affects their take up and willingness to pay for implementation. Authors find that (a) mayors are willing to pay for studies with large sample size, and (b) information provision about policy effectiveness (and improved tax compliance) increases mayors’ tendency to adopt them. (Diana Moreira) *working paper forthcoming
• The study looks into how an alternative design for RCT’s, experiment-as-market (EXAM), can improve the welfare of research subjects. Compared to standard RCTs, EXAM substantially improves subjects’ predicted well-being while reaching treatment effect estimates with similar precision. (Narita)

Session 4B: Conflict

• Social benefits programs that disincentive formal employment induces young men in Medellin to join gangs. (Khanna et al.)
• Experimental evidence from Mozambique shows that information given to leaders increases elite capture and rent-seeking, while information given to citizens increases mobilization and decreases conflict. The authors conclude that information dissemination to communities can serve as a force against the political resource curse. (Armand et al.)
• A novel survey data in Eastern Congo shows that villagers tended to participate in the militia if they were attacked by an external ethnic group in the past, with a weak effect on household wealth, suggesting that intrinsic motivation is the main driver. (Raul Sanchez de la Sierra and Qihang ‘David’ Wu) *working paper forthcoming
• Estimates that rely on individual measures that aggregate across types of violence exposure or on geographic measures that aggregate across individuals reflect mixture distributions that depend on the underlying distributions of exposure to violence. Simple sampling differences can thereby generate the sort of variability of estimated effects that has been reported in the literature to date. (Rockmore et al.)

Session 4C: Health 2

• Data from Kyrgyzstan suggests that private management and ownership of land, as opposed to government controlled, improves child health and nutrition outcomes. (Kosec)
• A non-monetary commitment mechanism in the form of a public school pledge implemented in Indonesia to not smoke, is effective, with individual pledges increasing tobacco abstinence by 5 percentage points. (Triyana and White)
• Coal-fired power plants in India increase infant mortality rates near places where they are located but any economic benefits that arise from the placement of these plants are spread over space. (Barrows et al.)
• This paper finds a role for local media in halting the spread of a major epidemic in the context of the Ebola outbreak in Guinea: information campaigns aired in local radios lowered social resistance behavior, increased treatment uptake and led to a significant drop in Ebola. The findings suggest that local media serves as a coordination device to change health behaviors that are associated to cultural norms, such as burial practices, as opposed to private actions, such as hand-washing, more sensitive to other information sources. (Gonzalez-Torres)

Session 4D: Education 2

• Long-term follow-up data suggests that a school construction program in Indonesia improves labor and marriage market outcomes, as well as living standards. These benefits are passed on to children in the household and appear to be larger when the mother was exposed to the program. (Akresh et al.)
• Does free secondary education significantly improve individual labor market outcomes? Data from Ecuador suggests that while it does improve college participation, it does nothing to improve income. Also, the majority of the benefits go to people of higher socioeconomic status, failing to promote equality in access to secondary education. (Molina et al.)
• Seven years of data from a social promotion policy on learning outcomes in India shows reading scores improved by 2.5 percent, and math scores by 5 percent. On top of that, the strongest effects were seen in the lowest quartile of the treatment group, a probable outcome of improved motivation. (Ahsan et al.)
• Are graduation thresholds too low? Data from Mexico’s Oportunidades Conditional Cash Transfers program shows that for those below the threshold, losing this additional incentive had a negative effect on high school attendance. (Pfutze)

Session 4E: Household Finance

• A commitment savings product designed to help clients taking repeated overdrafts break their debt cycles was found to increase savings in general — both with the bank and other savings. The catch is that individuals with below-median overdraft level at baseline are more likely to benefit more than those with above-median overdrafts in the baseline. (Buehren et al.)
• How do informal taxation and public good provision respond to household income shocks when households are provided cash transfers? While transfers improved household welfare, they had no impact on the amount of informal taxes paid both by the recipient and non-recipient households nor on their likelihood to pay informal taxes. (Walker)
• By exploiting the timing of China’s ban on sex-selective abortions and interacting it with the gender of household’s first-born child as an instrumental variable, this study finds a stronger same-gender effect of intergenerational transmission of old-age family support, where parents influence same-gender children in the transmission. (Shi)
• Why do SMS help people save? This paper evaluates if it is because it brings savings to the top of the mind, if it personalizes the savings goal or if it combines both factors through a RCT conducted with Paraguayans at the bottom of the pyramid. Preliminary results suggest that salience, personalization and their interaction all influence the success of SMSs for savings but that they play a different role depending on the characteristics of the recipients. (Azevedo et al.) *working paper forthcoming

Session 4F: Trade

• As countries participate more in global agricultural value chains, modern-day agrarian economies are leapfrogging manufacturing to directly develop their services sector, which runs counter to conventional structural transformation narratives. (Sunghun)
• What was the impact of rising Chinese import competition on the gender-specific labor market outcomes of Peruvian workers? The data suggests that that exposure to Chinese imports reduced the employment rates of low-skilled female workers while not affecting the employment rates
  of male workers. (Mansour et al.) *working paper forthcoming
• Governments that ignore the implications of an imperfectly competitive market risk imprecisely predicting gains from selective policy interventions. Data from India shows that the elimination of firm-size restrictions in a single market affects productive firms across the supply-chain, as markets reallocate inputs to more productive firms. (Balasundharam)
• Enforcement of (safety-related) labor standards by multinational buyers can improve establishments’ compliance and safety in settings with weak state enforcement. Evidence from a randomized controlled trial with 84 establishments in Bangladesh shows that these improvements depend on establishments’ organizational capacity: Improvements are driven by better-managed establishments, with less-well managed establishments not significantly improving safety. (Boudreau)

If you would like to suggest a correction to any of the above summaries, please email CEGA Communications Associate, Dustin Marshall, at dmarsh1231@berkeley.edu.