Julia Berazneva is an Assistant Professor at Middlebury College. Nathan Jensen is a Postdoctoral Associate at Cornell’s Dyson School who is working with the International Livestock Research Institute (ILRI). Linden McBride is a PhD candidate at Cornell’s Dyson School.
This is the third installment of our coverage of the National Bureau of Economic Research and the Feed the Future Innovation Lab for Assets and Market Access workshop on The Economics of Asset Dynamics and Poverty Traps. In this post we discuss two additional mechanisms posited to affect self-perpetuating poverty: Imperfect and Incomplete Financial Markets and Dynamics and Resilience in Natural Resources and Agriculture.
Presenters in the Imperfect and Incomplete Financial Markets session reviewed the empirical findings on the impacts of micro-financial interventions and assessed their theoretical implications and tried to identify solutions to the social protection paradox in which transfers to the poor fail to address vulnerabilities, capabilities, or incentives thus producing more beneficiaries over time.
In “Taking stock of the evidence on micro-financial interventions,” Buera, Kaboski, and Shin find that cash and in-kind grants to the poor, self-employed, and micro-entrepreneurs increase both capital and profits while in-kind grants to the ultra-poor (combined with other interventions) increase income and consumption but are only sometimes sustained. They find that offering microcredit access to new populations has mixed impacts on entry into small business, income, consumption, profits, and capital. These empirical findings suggest that the impacts of these interventions are heterogeneous across individuals—they vary by initial assets, ability, gender, and financial access. They also suggest heterogeneity across existing and new entrepreneurs. Little in the way of sustained impacts over time are identified. Finally, they conclude that grants are more effective than other types of microfinance interventions. Buera et. al. then develop a model in order to assess our theoretical understanding of the empirical findings they have identified. Their model, combined with their empirical analysis, suggests that 1) there are no widespread exits from poverty traps, 2) responses to interventions are heterogeneous and 3) general equilibrium and dynamics effects can lead to dissaving and an increase in wages (even for non participants).
Carter, Ikegami, and Barrett model the impacts of capital endowments and agent ability on long term asset holdings in the presence of uncertainty and poverty traps in “Poverty traps and the social protection paradox.” The model is used to compare the 50-year poverty implications of a limited budget social protection (cash transfer) program that targets the poorest with that of one that first prioritizes individuals who have recently fallen below the poverty threshold. The result of the latter approach is a targeting principle that favors higher ability individuals, which are more quickly able to graduate out of the program, so that greater resources can be brought to bear on the low-ability and more-needy individuals. This ‘triage’ approach to targeting has superior population-level poverty outcomes, but opens the door to moral hazard as the implicit insurance provided by the cash transfers for individuals just above the poverty threshold reduce the incentives to progress. The authors discuss the option of using a progressive insurance subsidy to mitigate moral hazard issues and to better align the social protection program with the political and ethical realities of implementation.
These papers suggest a lack of consensus in how we model welfare dynamics as well as raise questions for further research in this area. In particular, what can we account for in our models (and what is beyond their scope)? This question links to some of the other themes of the workshop including the role of cognitive functioning (attention, memory), emotion (depression), beliefs, and aspirations in poverty and poverty traps. A related question: what can/should be endogenous in such models? And how do we account for heterogeneity across individuals? Should the absorbing state of death be included in such models? Finally, once welfare dynamics have been observed/modeled, what are the appropriate policy responses?
In the session, Dynamics and Resilience in Natural Resources and Agriculture, presenters discussed the ways in which adverse states of nature can give rise to and/or exacerbate nonlinear wealth dynamics but that ability, resilience, and timely policy response can mitigate the effects of adverse shocks. However, one of the take aways of this session was that we need both better approaches to modeling these dynamic systems as well as further study of their mitigating factors beyond the limited settings currently available.
“Heterogeneous wealth dynamics the role of risk and ability,” by Santos and Barrett, examines the roles of risk and ability in the livestock dynamics of rural southern Ethiopia. The authors find evidence of nonlinear wealth dynamics that result in multiple dynamic equilibria only in adverse states of nature (drought years) and associated with herder ability, suggesting that the Shultz (1975) idea of the ability to deal with disequilibria is playing a role in this environment’s dynamics. They find that high ability herders face multiple equilibria while low ability herders face a single equilibrium. With heterogeneous ability, policy targeting becomes crucial, and the impact of interventions will critically depend on the mechanisms behind growth dynamics.
The paper, “Agro-ecosystem productivity and the dynamic response to shocks,” by Chavas, studies the nonlinear dynamic response to shocks, using a threshold quantile autoregression model in the context of Kansas agriculture over the period 1885-2012. The period studied includes the Dust Bowl, which was the product of major adverse shocks and poor agricultural management. The paper finds evidence of resilience in this agroecosystem with induced innovations in both policy (establishment of the Soil Conservation Service in 1935, for example) and management that followed the Dust Bowl.
Some additional insights from the discussion of these papers by Edward Barbier and the audience included the role that geography can play in creating poverty traps (e.g., severe biophysical constraints on production and limited access to markets), the dynamics of other assets (e.g., pasture biomass), and the disappearing option of migration to deal with adverse shocks. The session made a call for better tools and more empirical evidence to help understand the dynamics and resilience in natural resources and agriculture and to inform policy.
Both of these sessions highlighted the important and sometimes overlooked role of heterogeneity in our models—both heterogeneity of the interventions themselves as well as their impacts across individuals (either due to individual ability or other characteristics). The fact that current models are still limited in their ability to deal with such heterogeneity suggests a promising avenue for future research.
