Poverty Traps and the Social Protection Paradox

In high-risk environments, such as the pastoral regions of Kenya, poverty traps can be particularly prevalent. Means-tested cash transfers have emerged as a popular instrument for addressing poverty, with programs such as the HSNP cash transfer in Kenya established in 2009. However, research has shown that these programs often have limited impact on poverty dynamics, particularly in environments where poverty traps are prevalent. One of the primary weaknesses of cash transfer programs is that they do little to enhance the capacities and incentives of poor households to accumulate and graduate from cash transfer dependence. In addition, cash transfer programs often fail to address the vulnerability of near-poor households, which can lead to a growth in the number of transfer-eligible households. This can result in a slow dilution of benefits to the poor if the social protection budget is fixed, or a budgetary sink if benefit levels are protected. In such environments, social protection based on contingent (insurance-like) transfers targeted at vulnerable, non-poor households may have a larger impact on poverty dynamics than purely progressive cash transfers. A stylized poverty trap model calibrated to the Northern Kenya environment suggests that households in this region face significant challenges in accumulating capital due to missing financial markets for credit and risk, non-convex production technology, and skill or technical efficiency heterogeneity. This can lead to the existence of two types of poverty traps: single equilibrium poverty traps that apply to households with low ability, and multiple equilibrium poverty traps that create the prospect for unnecessary deprivation among households with middle ability. The social protection paradox suggests that compared to a conventional cash transfer targeting, the extent and depth of poverty are lower in the medium term when a limited social protection budget is used to first target the vulnerable non-poor in preference to the poor. Anticipation of contingent transfers can promote upward mobility by the poor, but it can also reduce accumulation by others, creating a moral hazard problem. Implementing contingent social protection as a partially subsidized index insurance program can overcome this distortion. A dynamic model under financial autarchy, which assumes that households are unable to access external credit markets, highlights the challenges faced by households in accumulating capital. The model suggests that households may face a Micawber threshold, below which they rationally retreat to a low equilibrium and above which they will accumulate and strive to reach a high equilibrium. Risk accentuates this problem, making it more difficult for households to accumulate capital. Numerical analysis using a stochastic structure meant to mimic the reality of Northern Kenya supports the existence of two types of chronic poverty or poverty traps: single equilibrium and multiple equilibrium. The model suggests that households with low ability are more likely to be trapped in a single equilibrium poverty trap, while households with middle ability are more likely to be trapped in a multiple equilibrium poverty trap.