Imperfect information and inattention to energy costs are important potential motivations for energy efficiency standards and subsidies. We evaluate these motivations in the lightbulb market using a theoretical model and two randomized experiments. We derive welfare effects as functions of reduced-form sufficient statistics capturing economic and psychological parameters, which we estimate using a novel within-subject information disclosure experiment. The main results suggest that moderate subsidies for energy-efficient lightbulbs may increase welfare, but informational and attentional biases alone do not justify a ban on incandescent lightbulbs. Our results and techniques generate broader methodological insights into welfare analysis with misoptimizing consumers.

It is often argued that people might take on too much high-cost debt because they are present focused and/or over-optimistic about how soon they will repay. We measure borrowers’ present focus and over-optimism using an experiment with a large payday lender. Although the most inexperienced quartile of borrowers under-estimate their likelihood of future borrowing, the more experienced three quartiles predict correctly on average. This finding contrasts sharply with priors we elicited from 103 payday lending and behavioural economics experts, who believed that the average borrower would be highly overoptimistic about getting out of debt. Borrowers are willing to pay a significant premium for an experimental incentive to avoid future borrowing, which we show implies that they perceive themselves to be time inconsistent. We use borrowers’ predicted behaviour and valuation of the experimental incentive to estimate a model of present focus and naivete. We then use the model to study common payday lending regulations. In our model, banning payday loans reduces welfare relative to existing regulation, while limits on repeat borrowing might increase welfare by inducing faster repayment that is more consistent with long-run preferences.

This paper tests costly attention models of consumers' misreaction to opaque taxes. We report an online shopping experiment that involves shrouded sales taxes that are exogenously varied within consumers over time. Some consumers systematically underreact to sales taxes whereas others systematically overreact, but higher stakes decrease both under- and overreaction. This is consistent with consumers using heterogeneous rules of thumb to compute the opaque tax when the stakes are low, but using costly mental effort at higher stakes. The results allow us to differentiate between various theories of limited attention. We also develop novel econometric techniques for quantifying individual differences.

What mental models do individuals use to approximate their tax schedule? Using incentivized forecasts of the U.S. Federal income tax schedule, we estimate the prevalence of the “schmeduling” heuristics for constructing mental representations of nonlinear incentive schemes. We find evidence of widespread reliance on the “ironing” heuristic, which linearizes the tax schedule using one’s average tax rate. In our preferred specification, 43% of the population irons. We find no evidence of reliance on the “spotlighting” heuristic, which linearizes the tax schedule using one’s marginal tax rate. We show that the presence of ironing rationalizes a number of empirical patterns in individuals’ perceptions of tax liability across the income distribution. Furthermore, while our empirical framework accommodates a rich class of other misperceptions, we find that a simple model including only ironers and correct forecasters accurately predicts average underestimation of marginal tax rates. We replicate our finding of prevalent ironing, and a lack of other systematic misperceptions, in a controlled experiment that studies real-stakes decisions across exogenously varied tax schedules. To illustrate the policy relevance of the ironing heuristic, we show that it augments the benefits of progressive taxation in a standard model of earnings choice. We quantify these benefits in a calibrated model of the U.S. tax system.

This paper shows that accounting for variation in mistakes can be crucial for welfare analysis. Focusing on consumer underreaction to not-fully-salient sales taxes, we show theoretically that the efficiency costs of taxation are amplified by differences in underreaction across individuals and across tax rates. To empirically assess the importance of these issues, we implement an online shopping experiment in which 2,998 consumers purchase common household products, facing tax rates that vary in size and salience. We replicate prior findings that, on average, consumers underreact to non-salient sales taxes-consumers in our study react to existing sales taxes as if they were only 25% of their size. However, we find significant individual differences in this underreaction, and accounting for this heterogeneity increases the efficiency cost of taxation estimates by at least 200%. Tripling existing sales tax rates nearly doubles consumers' attention to taxes, and accounting for this endogeneity increases efficiency cost estimates by 336%. Our results provide new insights into the mechanisms and determinants of boundedly rational processing of not-fully-salient incentives, and our general approach provides a framework for robust behavioral welfare analysis.

A common objection to “sin taxes”—corrective taxes on goods that are thought to be overconsumed, such as cigarettes, alcohol, and sugary drinks—is that they often fall disproportionately on low-income consumers. This paper studies the interaction between corrective and redistributive motives in a general optimal taxation framework and delivers empirically implementable formulas for sufficient statistics for the optimal commodity tax. The optimal sin tax is increasing in the price elasticity of demand, increasing in the degree to which lower-income consumers are more biased or more elastic to the tax, decreasing in the extent to which consumption is concentrated among the poor, and decreasing in income effects, because income effects imply that commodity taxes create labor supply distortions. Contrary to common intuitions, stronger preferences for redistribution can increase the optimal sin tax, if lower-income consumers are more responsive to taxes or are more biased. As an application, we estimate the optimal nationwide tax on sugar-sweetened beverages, using Nielsen Homescan data and a specially designed survey measuring nutrition knowledge and self-control. Holding federal income tax rates constant, our estimates imply an optimal federal sugar-sweetened beverage tax of 1 to 2.1 cents per ounce, although optimal city-level taxes could be as much as 60% lower due to cross-border shopping.

A corrective tax or subsidy is “well-targeted” if it primarily affects choices that are more distorted by market failures. Energy efficiency subsidies are designed to correct multiple distortions: externalities, credit constraints, “landlord-tenant” information asymmetries, imperfect information, and inattention. We show that three important energy efficiency subsidies are primarily taken up by consumers who are wealthier, own their own homes, and are more informed about and attentive to energy costs. This suggests that these subsidies are poorly targeted at the market failures they were designed to address. However, we show that “tagging” can lead to large efficiency gains.

We study how people’s beliefs about the economy covary with household-level events, utilizing a unique link between Danish administrative data and a large-scale survey of consumer expectations. We find that compared to actual inflation, people’s inflation forecasts covary much more strongly (and negatively) with both recently realized household income changes and measures of expected future household income changes. We formally establish that these findings are stark deviations from the Bayesian rational expectations benchmark. Similar results hold for perceptions of past inflation (“backcasts” ), suggesting that imperfect recall is a key mechanism for biased forecasts. Building on this, a series of additional tests, some of which utilize data on adverse health events, suggests that the forecast biases are at least partly due to affect-cued recall. That is, negative (positive) household-level events cue negative (positive) recollections, which lead to pessimistic (optimistic) forecasts.

We use natural experiments embedded in state-run lotteries and a new nationally representative survey to provide reduced-form and structural estimates of risk preferences and behavioural biases in lottery demand, and to explore the implications for optimal lottery design. We find that sales respond more to the expected value of the jackpot than to price but are unresponsive to variation in the second prize—a pattern that is consistent with probability weighting but is inconsistent with standard parameterizations. In the survey, we find that lottery spending decreases modestly with income and is strongly associated with measures of innumeracy, poor statistical reasoning, and other proxies for behavioural bias, which also decline with income. Regression predictions suggest that Americans would spend 43% less on lotteries if they were unbiased, while the remaining lottery demand is due to other factors such as anticipatory utility or entertainment value. We use these empirical moments to estimate a model of socially optimal lottery design. In the model, current multi-state lottery designs increase welfare but may harm heavy spenders.