Session 6: Missing Markets

Quality upgrading to attain export premia is a key development strategy in many low and middle-income countries (LMICs). However, in many supply chains in LMICs, producers do not sell directly to the world market; instead, they sell through multiple layers of domestic intermediation. We study how these chains of intermediation affect incentives for quality production in the context of the Uganda coffee supply chain. We first map out the supply chain linking farm gate and export gate, documenting that coffee changes hands, on average, two times before reaching export markets. Next, leveraging high-frequency transaction-level data throughout the supply chain, including prices and lab quality assessments, we show that quality premium diminishes upstream. We build a model highlighting two key economic mechanisms that may drive these diminished quality premia upstream. First, barriers to entry to high-quality intermediation enable greater buyer power in the high-quality segment; resulting differential markdowns squeeze the quality premium upstream. Second, both producers and intermediaries engage in quality investments, with the degree of substitutability of these investments mediating the downstream’s demand for high quality from the upstream. Both mechanisms have distributional implications for farmer surplus, but only the former dampens aggregate quality production.  To separately quantify the two, we conduct an experiment that offers randomized coffee production contracts to induce quality-specific demand shocks throughout the supply chain to identify the key cost parameters of our model. We use the estimated model to examine the efficiency implications and distributional consequences of market power and productive substitutability, as well as how they interact to affect quality production and surplus distribution along the chain. Finally, we apply the model to investigate policy counterfactuals.

This paper introduces a method to study the impact of policy events on equilibriumselection in settings where strong complementarities may lead to multiple equilibria and coordination failures. Many industrial policies are rooted in coordination failures and ‘big-push’ theories, yet empirical evidence on their effectiveness remains limited since distin-guishing equilibrium shifts from direct changes in fundamentals is challenging. Leveraging tools from industrial organization and algebraic geometry, I develop an approach to study the coordination effects of policy without imposing strong assumptions on the distribution or responsiveness of economic fundamentals. The method identifies the ‘types’ of factual and counterfactual equilibria before and after the policy event through a three-step pro-cedure: model estimation and inversion, equilibrium enumeration, and type assignment. Types of factual equilibria may be used to examine how policies like urban infrastructure investment, subsidies, or trade liberalization, affect equilibrium selection. Types of counter-factual equilibria help identify the extent to which observed effects of policy are driven by fundamentals or coordination. I apply this method to study industrial zones in India. Using a newly constructed data on 4,000 industrial zones, I find that municipalities receiving a zone see a 60% increase in non-farm employment over 15 years, with significant spillovers to sectors and municipalities not targeted by the zones. Combining the type assignment methodology with event study designs, I find that industrial zones increase the probability of escaping a low-industrialization equilibrium by 38%, with coordination effects explaining roughly one-third of the observed change in outcomes.

We test whether improving customers' ability to discern product quality increases demand-side incentives for quality provision in consumers' markets in developing countries. While access to rich, sophisticated foreign markets has been shown to drive quality upgrading among exporting firms, this pathway is unavailable to the vast majority of micro and small enterprises in low-income countries. We conduct a series of experiments in Uganda's wooden furniture market—an archetypal sector featuring high agglomeration of small-scale producers, significant quality dispersion, a lack of effective quality signalling, and a flat price-quality gradient. We first use a framed field experiment in popular furniture outlets in the Greater Kampala area. We establish kiosks in the market and randomly provide prospective furniture buyers with a brief training on assessing vertical quality attributes, while a control group receives a placebo. Among the over 700 participants, we find that untrained individual consumers perform significantly worse than industry insiders at discerning quality differences at baseline. Our information intervention substantially improves consumers' ability to identify quality, with treated individuals closing the discernment gap with industry experts. Importantly, treated consumers also exhibit a significantly steeper price-quality gradient in their stated willingness to pay relative to the control, with quality premiums increasing by approximately 15 percent. We then conduct a field experiment with carpenters, providing a random subset with a skills training focusing on quality. Half of those receiving the skills training also receive a marketing poster aimed at improving consumer knowledge about quality and the importance of the skills those carpenters obtained. We show that the training increases the quality of key joints produced by the carpenters and increases the knowledge of consumers of those carpenters. Ongoing work uses detailed item-level quality and price data to examine the effect of the interventions on quality, price markups and price bargaining.

Despite the significant economic value of environmental data across a wide range of contexts, private markets for their collection and dissemination have largely failed to materialize. Amid the inherent spatial nature of climate conditions, I examine scope for social learning to explain these missing firms. In my model, sharing information with one’s peer can lead to free-riding yet also provide valuable precision given imper-fect measurement technology. I characterize the set of Nash Equilibria among the joint decisions to purchase environmental information and learn from one’s neighbors and derive conditions for market unraveling. I bring this model to the data in a series of ex-periments offering Bangladeshi rice farmers the chance to purchase soil salinity readings. Farmers recognize the high spatial covariance, exhibiting no higher willingness-to-pay for soil readings from their own plot above and beyond the local average. On aver- age, the free-riding effect dominates, though a substantial share of respondents have complementary demand in a private market with respect to their peers’ information set.

The modal firm in low- and middle-income countries has no non-family-member employees. While this fact is often attributed to informational or contractual frictions, an alternate view is that pressure to offer financial assistance to extended family in the form of employment may distort employers' hiring decisions. I design three field experiments---conducted with both urban firms and agricultural employers in Zambia---to test whether such pressure impacts firm hiring and examine its productivity implications. In the first experiment, urban firms are offered the chance to receive a 3-month subsidy to hire a full-time permanent employee. A subset of firms is then randomized into a treatment providing them plausible deniability in their hiring decision: receipt of a poster that suggests the firm may not have been eligible for the subsidy had it hired a relative. The treatment increases the probability that the firm chooses to hire a nonfamily employee, rather than a relative, by 11 p.p. (20%). In the second experiment, I show that it is socially very costly for employers to hire a nonrelative: When firms have plausible deniability, the subsidy required to induce a firm to choose a nonrelative over a relative falls substantially, to the equivalent of 4% of total monthly profits. This effect is concentrated among firm owners who report higher redistributive pressure at baseline. In the third experiment, I show that in maize shelling, a common agricultural job where productivity is measurable, a worker is likelier to shirk when working for a family employer, leading relatives to be 7% less productive than nonrelatives. These findings suggest that social pressure to hire relatives may distort the composition of employment and productivity in developing countries.

This paper examines how the introduction of formal financial institutions affects village cooperation– including public goods provision, informal risk-sharing, and social network structure – in rural India. We experimentally evaluate the expansion of brick and mortar bank branches into 50 service areas that were ran-domly selected across geographically matched pairs. Within two years of branch opening, 31% of households in villages with new banks had taken out loans from the bank. Moreover, this shift to formal loans significantly reduced the reliance on informal borrowing, with significantly fewer and smaller total amounts of informal loans observed in treated villages (Barboni et al., 2022). In order to examine the spillover effects of access to bank branches on village-level risk-sharing capacity and public goods provision, we make use of novel data that include full social network mappings and a census of individual public goods contributions and shocks to earnings capacity from 205 villages within these service areas. We first show that substitution away from informal financial intermediation was associated with fewer social network links and interaction time between village households. In terms of the structure of the village network, the opening of a bank branch leads to sparser and more clustered networks in treatment villages. Second, the insurance capacity of village networks – measured by the maximum amount that households state they could borrow in times of need from each of their links – becomes weaker. Finally, these changes are accompanied by a substantial decrease in public goods provision, with treated villages showing a 22% decrease in time contributions and a 44% decrease in monetary contributions to public goods. Additionally, households in treated areas are able to fully insure against the loss of a wage-earner. Together, these findings suggest that, while formal financial access can reduce the reliance on informal borrowing, it can also have unintended consequences on social capital and community cooperation.

This paper examines creation and distribution of surplus from global value chains (GVCs) in low- and middle-income country (LMIC) domestic supply chains. While GVC participation can enhance growth and productivity, low prices paid to small input suppliers raise concerns about the concentration of gains among large exporters. Many supply chain transactions are governed by bargained agreements rather than spot markets. Therefore, I enrich a Nash bargaining model to incorporate how both i) value creation through insurance-like agreement terms that mitigate spot market frictions and ii) value capture via buyers threatening to replace external suppliers with in-house production decrease prices paid to small, risk-averse suppliers. Using novel transaction data from an Indian garment manufacturer and its nearly 500 fabric suppliers, I estimate a structural model to decompose discounts into value creation and capture. Results indicate that discounts reflect value creation from profit-smoothing agreement terms rather than buyer capture. Counterfactual analyses highlight that increasing buyer competition has limited effects on supplier prices, whereas introducing profit insurance substantially increases prices for small, risk-averse suppliers.

Price control policies are often deployed in many important markets, such as health-care. In the presence of common market failures—imperfect competition or choice frictions—the impact of such policies on market outcomes and welfare is ambiguous. We provide novel analysis of these forces in the context of a large-scale pharmaceutical price control policy in India. We find the policy lowered the prices of branded drugs by 24% and increased their sales by 36%, with minimal impact on entry, exit, and dif-fusion of new molecules. The standard revealed preference framework implies that the policy increased consumer surplus by 23% because it corrected monopoly distortions without disrupting supply. However, we present novel evidence that choice frictions lead consumers to overvalue expensive brands relative to cheaper alternatives of similar quality. Once we account for these frictions, we find that the consumer surplus gains are 30% smaller and that total welfare declines, as the policy unintentionally steers consumers toward costly overvalued brands. Finally, we assess how to effectively set price ceilings and evaluate alternative nonprice regulations.

Air pollution within African cities is high but unevenly distributed. In principle, individuals could mitigate the severe health risk by working in the less polluted parts of the city. In practice, we show that pollution avoidance is challenging because firms locate on the busiest and most polluted roads searching for customer visibility. Both workers and entrepreneurs bear the cost of this pollution exposure, but the benefits are unequally distributed: profits are much higher in polluted areas, while compensating differentials in wages are minimal. An information experiment reveals limited awareness of pollution, suggesting that workers might be undercompensated for their exposure.

The Covid-19 pandemic represents one of the most rapid and severe shocks to hit global labor markets. We study how firms reacted to the heightened uncertainty and consequences for workers, in a developing country economy: Uganda. Our analysis is based on a panel of firms and workers, tracked from 2012 to 2022, including high frequency surveys during the pandemic. We find that firms' common response to the heightened uncertainty of the pandemic was to immediately lay off the highest earning workers, that is, the most experienced or skilled employees. We then study the differential impacts of such firm survival strategies on workers' labor market dynamics across the skills distribution, exploiting the fact that we randomly assigned individuals to the offer of vocational training in 2013. We find that high skill trained workers, who enjoyed better labor market outcomes pre-pandemic, were more likely to be laid off early in the pandemic given firm's survival strategies. However, trained workers recovered from this job loss and were resilient to the shock. Cumulatively over the pandemic, trained workers spend 61% more time employed than untrained controls, and earn 17% more. Hence, the returns to training survive the pandemic. The mechanisms driving the resilience of trained workers are the certifiability of their skills and their greater accumulation of sector-specific experience, both of which enable them to switch employers within the same sector during the crisis. Our findings provide new insights on firm responses to fast-moving aggregate shocks in low-income settings, consequences for workers' labor market trajectories, and drivers of the returns to training in good times and bad.