Charlie Rafkin

Job Market Candidate

Research Fields

Public Economics, Behavioral Economics

Contact Information

Job Market Paper

“Eviction as Bargaining Failure: Hostility and Misperceptions in the Rental Housing Market” (with Evan Soltas), Fall 2023. Online supplement

Abstract: Court evictions from rental housing are common but could be avoided if landlords and tenants bargained instead. Such evictions are inefficient if they are costlier than bargaining. We test for two potential causes of inefficient eviction — hostile social preferences and misperceptions — by conducting lab-in-the-field experiments in Memphis, Tennessee with 1,808 tenants at risk of eviction and 371 landlords of at-risk tenants. We detect heterogeneous social preferences: 24% of tenants and 15% of landlords exhibit hostility, giving up money to hurt the other in real-stakes Dictator Games, yet more than 50% of both are highly altruistic. Both parties misperceive court or bargaining payoffs in ways that undermine bargaining. Motivated by the possibility of inefficient eviction, we evaluate the Emergency Rental Assistance Program, a prominent policy intervention, and find small impacts on eviction in an event-study design. To quantify the share of evictions that are inefficient, we estimate a bargaining model using the lab-in-the-field and event-study evidence. Due to hostile social preferences and misperceptions, one in four evictions results from inefficient bargaining failure. More than half would be inefficient without altruism. Social preferences weaken policy: participation in emergency rental assistance is selected on social preferences, which attenuates the program’s impacts despite the presence of inefficiency.

Working Papers

“Legal Assistance for Evictions: Impacts, Mechanisms, and Demand” (with Aviv Caspi), November 2023. 

Abstract: We randomize provision of attorneys to tenants facing eviction in Memphis, Tennessee (N = 265 treatment, 753 control), who otherwise seldom have legal representation. Despite landlord-friendly eviction law, providing an attorney reduces tenant eviction judgment rates within 90 days by 27 percentage points (50%). However, attorneys’ effects persist only when they can connect tenants to other services. Once a concurrent emergency rental assistance program expires, effects on judgments at 90 days shrink by about 70% and are indistinguishable from zero. Attorneys have little effect on informal outcomes and bargaining. Incentivized surveys suggest tenants’ willingness to pay for an attorney is double attorneys’ price, and seven times attorneys’ implied impacts on tenants’ incomes via stopping evictions. This high willingness to pay does not appear to result from elicitation errors, misperceptions, or binding budget constraints. We contrast lawyers’ Marginal Value of Public Funds from using elicited willingness to pay (MVPF = 2.7 without rental assistance, ignoring impacts on landlords or general equilibrium) versus a standard calibrated approach (MVPF = 0.4). 

“Preferences for Rights” (with Aviv Caspi and Julia Gilman), November 2023. Survey instruments.

Abstract: Public discourse about in-kind transfers often appeals to “preferences for rights” — for instance, the “right to health care” or “right to counsel” for indigent legal defense. Preferences for rights are “non-welfarist” if the person values the right per se, holding fixed how the right instrumentally affects others’ utilities. We test for non-welfarist preferences for rights, and their relationship to redistributive choices, with incentivized online experiments (N = 1,800). Participants face choices about allocating rights goods (lawyers, health care) and benchmark goods (bus passes, YMCA memberships) to tenants facing eviction. We implement a share of choices. In two of three experiments, more than half of participants allocate rights goods in ways that are consistent with preferences for rights and dominated if preferences were entirely welfarist. Dominated behaviors are more common with rights goods than benchmarks. In a fourth experiment, those with preferences for rights also exhibit “anti-targeting,” where they redistribute lawyers and health care more universally than benchmark goods to recipients whose incomes differ. At least 26% of participants are non-welfarist, while at most 31% are welfarist.

“Self-Targeting in U.S. Transfer Programs” (with Adam Solomon and Evan Soltas), October 2023.

Abstract: Transfer receipt is voluntary and costly, generating “self-targeting” through selective take-up among the eligible. How does self-targeting select on need, and what are its policy implications? We show self-targeting is advantageous in eight U.S. transfers: On average, recipients have lower consumption and lifetime incomes than eligible nonrecipients with similar current incomes. Due to self-targeting, these transfers provide 50 to 75 percent more to the consumption-poorest and lifetime-poorest than would automatic transfers that are distributionally equivalent by income. Self-targeting makes automatic transfers undesirable: We estimate the social benefits of self-targeting are approximately six cents per transfer dollar, generally exceeding the social costs of ordeals.

“The Welfare Effects of Eligibility Expansions: Theory and Evidence from SNAP” (with Jenna Anders), November 2022. Conditionally accepted at American Economic Journal: Economic Policy.

Abstract: We study the U.S. rollout of eligibility expansions in the Supplemental Nutrition Assistance Program. Using administrative data from the U.S. Department of Agriculture, we show that expanding eligibility raises enrollment among the inframarginal (always-eligible) population. Using an online experiment and an administrative survey, we find evidence that information frictions, rather than stigma, drive the new take-up. To interpret our findings, we develop a general model of the optimal eligibility threshold for welfare programs with incomplete take-up. Given our empirical results and certain modeling assumptions, the SNAP eligibility threshold is lower than optimal.


“Intergenerational Mobility in India: New Measures and Estimates Across Time and Social Groups” (with Sam Asher and Paul Novosad). Forthcoming at American Economic Journal: Applied Economics.

Abstract: We study intergenerational mobility in India. We propose a new measure of upward mobility: the expected education rank of a child born to parents in the bottom half of the education distribution. This measure works well under data constraints common in developing countries and historical contexts. Intergenerational mobility in India has been constant and low since before liberalization. Among sons, we observe rising mobility for Scheduled Castes and declining mobility among Muslims. Daughters’ intergenerational mobility is lower than sons’, with less cross-group variation over time. A natural experiment suggests that affirmative action for Scheduled Castes has substantially improved their mobility.

“Optimal Regulation of E-cigarettes: Theory and Evidence” (with Hunt Allcott). NBER Working Paper #27000. American Economic Journal: Economic Policy, November 2022.

Abstract: We model optimal e-cigarette regulation and estimate key parameters. Using tax changes and scanner data, we estimate relatively elastic demand. A demographic shift-share identification strategy suggests limited substitution between e-cigarettes and cigarettes. We field a new survey of public health experts who report that vaping is more harmful than previously believed. In our model’s average Monte Carlo simulation, these results imply optimal e-cigarette taxes are higher than recent norms. However, e-cigarette subsidies may be optimal if vaping is a stronger substitute for smoking and is safer than our experts report, or if consumers overestimate the health harms from vaping.

“Mortality Change Among Less Educated Americans” (with Sam Asher and Paul Novosad). American Economic Journal: Applied Economics, October 2022. 

Abstract: Measurements of mortality change among less educated Americans can be biased because the least educated groups (e.g., dropouts) become smaller and more negatively selected over time. We show that mortality changes at constant education percentiles can be bounded with minimal assumptions. Middle-age mortality increases among non-Hispanic Whites from 1992 to 2018 are driven almost entirely by the bottom 10 percent of the education distribution. Drivers of mortality change differ substantially across groups. Deaths of despair explain most of the mortality change among young non-Hispanic Whites, but less among older Whites and non-Hispanic Blacks. Our bounds are applicable in many other contexts.

“When Guidance Changes: Government Stances and Public Beliefs” (with Advik Shreekumar and Pierre-Luc Vautrey). Journal of Public Economics, April 2021.

Abstract: Governments often make early recommendations about issues that remain uncertain. Do governments’ early positions affect how much people believe the latest recommendations? We investigate this question using an incentivized online experiment with 1900 US respondents in early April 2020. We present all participants with the latest CDC projection about coronavirus death counts. We randomize exposure to information that highlights how President Trump previously downplayed the coronavirus threat. When the President’s inconsistency is salient, participants are less likely to revise their prior beliefs about death counts from the projection. They also report lower trust in the government. These results align with a simple model of signal extraction from government communication, and have implications for the design of changing guidelines in other settings.