In-Kind Transfers as Insurance

A central question in designing social protection programs is determining the optimal form of transfers to the poor. Historically, in-kind transfers were the primary anti-poverty program, and they remain prevalent globally. However, a recent shift among academics and policymakers favors unconditional cash transfers, driven by the success of programs like GiveDirectly and growing interest in universal basic income. The standard justification for cash transfers is that beneficiaries weakly prefer cash to in-kind. In-kind transfers are typically justified by the need to meet social objectives like pecuniary redistribution or targeting, or by the belief that beneficiaries would misallocate cash. However, this contradicts evidence showing beneficiaries often express a preference for in-kind transfers in various contexts, citing concerns about price instability. This paper investigates whether in-kind transfers can improve welfare relative to cash due to their inherent insurance value against commodity price risk, a significant factor in many developing countries with poorly integrated markets. The paper focuses on questioning the basic assumption that cash delivers higher expected utility, while acknowledging that a full welfare analysis should also consider the relative social cost of provision.

The paper reviews several relevant literatures. First, it addresses the ongoing debate regarding the optimal design of social protection programs, highlighting the benefits of unconditional cash transfers shown in recent studies. It contrasts this with existing rationales for in-kind transfers such as improved targeting, positive spillover effects on non-targeted households by reducing market prices, and increased efficiency in imperfectly competitive food markets. The authors note that while the insurance benefit of in-kind transfers has been mentioned in policy discussions, it remains largely unexplored in academic literature. Second, the paper engages with the literature on household exposure to price variability and its welfare consequences, comparing it to the effects of price stabilization. The authors highlight that stabilization policies are often considered expensive and ineffective. Third, the paper contributes to the literature on price shocks and food security, pointing out the mixed findings on the effect of food price shocks on nutrition. Finally, it provides new evidence on the impact of India's Public Distribution System (PDS), one of the world's largest in-kind transfer programs, whose impact on nutrition and the reduction of price sensitivity is understudied. The authors also acknowledge behavioral models that might lead to different allocations of cash or in-kind transfers, which are outside the scope of this paper.

The paper develops a simple model to analyze the welfare effects of cash versus in-kind transfers under price variation. The model shows that the optimal insurance policy involves price-indexed transfers equalizing marginal utility of income across different price states. However, as price-indexed transfers are often infeasible, the model compares price-invariant cash transfers and in-kind transfers. It demonstrates that households will prefer in-kind transfers if the covariance between the marginal utility of income and price is positive. The paper employs the National Sample Survey (NSS) data from India, encompassing over 500,000 households across 28 states and 10 years. Local rice prices are constructed using unit values from the NSS. Caloric intake and an indicator for meeting minimum calorie requirements (MCR) serve as proxies for the marginal utility of income. The authors justify their proxy for marginal utility by discussing the literature on the negative consequences of calorie shortfalls on health, human capital, and earnings, and the relatively low substitutability of calories. The paper empirically tests the model's prediction using the NSS data. To analyze the causal effects of India's PDS program, the authors leverage newly collected administrative data on state-level policy changes between 2003 and 2012. They use variation in mandated PDS prices and eligibility expansions to instrument for PDS value, defined as the quantity of rice obtained multiplied by the price difference between the market and PDS prices. They then estimate the effects of PDS expansions on caloric intake and the calorie-price relationship, controlling for various factors. The authors address potential endogeneity concerns by demonstrating flat trends in calories before eligibility expansions and perform robustness checks by controlling for political cycles and the generosity of other welfare programs.

The empirical analysis reveals that Indian households face considerable exposure to price risk, with a substantial portion of their budget allocated to food, particularly rice. A 10% increase in rice prices is significantly associated with a decrease in caloric intake, primarily impacting below-median socioeconomic status (SES) households. This negative calorie-price gradient supports the model's prediction that in-kind transfers are preferable to cash for poorer households. The analysis of the PDS demonstrates large positive effects of PDS expansions on caloric intake. A Rs. 100 increase in PDS value leads to a 10.7 percentage point increase in households meeting the MCR and a 6.4% increase in calories per capita. Overall, the PDS policy changes led to 40 million additional individuals meeting MCR thresholds. Importantly, PDS generosity significantly reduces the sensitivity of calories to market prices. A Rs. 100 increase in PDS value reduces this sensitivity by 73%, implying that the PDS provides substantial insurance against price risk. The results are robust to various sensitivity checks, including the exclusion of major rice-supplying states, inclusion of controls for political cycles and other welfare programs, and using either price or quantity variation in the PDS. The impact of PDS on market prices is too small to explain the findings.

The findings address the research question by demonstrating empirically that the condition for in-kind transfers to be welfare improving (positive covariance between marginal utility of income and price) is satisfied for poorer households in India. The significant negative relationship between rice prices and caloric intake, especially among poorer households, confirms the importance of price risk. The causal evidence from the PDS shows that in-kind transfers effectively mitigate this risk by increasing caloric intake and reducing price sensitivity. This suggests that the insurance value of in-kind transfers is a substantial factor in explaining their prevalence and beneficiary preference, particularly among the poor. The results challenge the implicit assumption of cash transfer superiority and highlight the importance of considering the insurance dimension in designing social protection programs. The significant impact of the PDS on nutrition underscores its vital role in ensuring food security and mitigating the effects of price volatility.

This paper contributes significantly to the ongoing debate on the optimal design of social protection programs by demonstrating the insurance value of in-kind transfers in mitigating price risk. The empirical findings from India strongly suggest that in-kind transfers, particularly the PDS, improve welfare among poorer households by boosting caloric intake and reducing price sensitivity. While a full welfare comparison requires consideration of the social cost of provision, the results highlight the substantial benefits of in-kind transfers in contexts with poor market integration. Future research could explore geographic and temporal variations in the relative benefits of in-kind and cash transfers and investigate the use of price indicators in designing cash transfer programs or exploring the potential of price-indexed cash transfers with advancements in technology.

The study's primary limitations stem from data constraints. The use of unit values as a proxy for prices might not perfectly capture local price variation. Additionally, the use of caloric intake and MCR as proxies for marginal utility of income involves some simplifying assumptions. While the authors address these limitations through robustness checks and detailed justification, these factors could potentially affect the precision of the estimates.