Principal investigator: Bruno Benevit
Original title: Behavioral Economics and Public Policy: A Pragmatic Perspective
Author Raj Chetty
Location of the Intervention: United States, Denmark
Sample Size: 1,2 billion person-years
Sector: Behavioral Economics
Primary Variable of Interest: Adherence to policies
Type of Intervention: Default, Nudge at
Methodology: Experiment
Summary
Behavioral economics has become a growing branch in economic literature, offering additional tools to the traditional neoclassical perspective. This article investigates how behavioral economics can enhance the analysis and design of public policies, focusing on pension programs, tax incentives, and residential choices. The study presents empirical approaches that combine natural experiments, administrative data analysis, and behavioral modeling in three policy cases. The results reveal that individuals frequently exhibit decision-making patterns inconsistent with traditional models, such as inertia in the face of tax incentives, undervaluation of long-term benefits, and sensitivity to informational presentation.default) of policies. This evidence indicates that policies that incorporate behavioral elements, such as automatic contributions to social security or nudges Informational interventions in housing programs tend to be more effective than interventions based solely on traditional economic incentives.
- Policy Problem
Behavioral economics has gained increasing relevance in the economic literature over the last few decades (CHETTY, 2015). This branch of economics emerges as a complement to the neoclassical economic approach, incorporating elements of psychology to understand economic decisions that deviate from perfect rationality. While neoclassical economics starts from assumptions such as optimization and rational expectations, the behavioral perspective recognizes that biases such as loss aversion, procrastination, and inattention can influence individuals' decision-making process when faced with real choices.
This integration allows for a more realistic analysis of public policies, focused not only on debating theories, but on solving practical problems, such as increasing savings or improving adherence to social programs. According to Chetty (2015), by adopting a pragmatic stance in the debate between these two approaches in the field of economics, it is possible to use the concepts of behavioral economics to adjust traditional neoclassical models, making them more accurate without completely abandoning their fundamentals.
The implications of behavioral economics for public policy can be organized into three main axes. The first concerns the development of new policy instruments, such as adjustments to default options or the format of incentives, which exploit cognitive biases to induce desirable behaviors. The second axis improves the ability to predict the effects of existing policies by incorporating factors such as behavioral inertia or inattention, which alter individuals' responses to economic incentives. Finally, the behavioral approach introduces new dimensions for assessing well-being by distinguishing between the utility perceived by the policymaker and the utility that guides individuals' decisions, often affected by distortions.
In this sense, this study explores three public policy cases to analyze these implications, verifying how behavioral economics can help in understanding them. Thus, considering behavioral factors not only expands the economic literature but also allows for improved design and evaluation of public policies.
- Policy Implementation Context
Behavioral economics offers alternative tools to influence behaviors, such as altering default options in retirement plans. A prime example of how new policy instruments can influence retirement savings is the Danish case. In 1999, the Danish government implemented a reform of the pension system, drastically reducing the tax benefit for contributions to type-specific accounts. capital pension (one-time redemption) for high-income contributors, while leaving incentives for type accounts unchanged. annuity pension (periodic payments). The system allowed workers to freely choose between these two types of accounts to direct their savings. This change created an abrupt difference in tax treatment between account types and between income brackets. This change allows us to verify the human tendency towards inertia and the potential of choices. default in public policy. Thus, this case can illustrate how behavioral biases can improve the design of more effective policies.
Regarding the impacts of existing policies, considering behavioral factors such as differences in knowledge about tax benefits can improve their prediction. The study considered the program Earned Income Tax Credit (EITC) in the US to examine how behavioral economics can improve the prediction of the effects of fiscal policies on labor supply. Implemented in its current form in 1996, the EITC establishes a system of tax credits with a non-linear structure, varying according to income, marital status, and number of dependents. The program design creates distinct incentives in different income brackets, with phases of accumulation, plateau, and reduction of benefits. This case is relevant because the beneficiaries' response to tax incentives can be influenced by difficulties in processing information about marginal benefits and biases in the evaluation of trade-offs between work and leisure.
Finally, the study considered the program. Moving to Opportunity (MTO) in the US to investigate how behavioral economics can aid in the well-being analysis of individuals' residential choices. Implemented in the 1990s, the program was an experiment that offered families living in housing complexes vouchers The study involved providing housing subsidies, allowing individuals to move to neighborhoods with lower poverty levels. The experimental design used randomization, creating treatment and control groups. This case is relevant because moving to better neighborhoods can bring significant benefits, breaking cycles of poverty. Children gain access to higher-quality schools and safer environments. In turn, adults begin to live in communities with more diverse social networks. Factors such as status quo bias, cognitive costs of change, and inaccurate assessment of future benefits can affect individuals' decision-making (CHETTY, 2015). In this sense, the consideration of concepts from behavioral economics such as... nudge – Small interventions that influence choices without restricting options can help in understanding family decisions, allowing us to understand why many families, even with... vouchersThey don't move to better areas.
- The case of retirement in Denmark
The study examined the Danish pension system, which offers two main types: capital pensions (with a single withdrawal upon retirement) and the annuity pensions (with installment payments). In 1999, the government implemented a reform that significantly reduced the tax benefit for contributions to capital pensions – from a 59% to a 45% deduction on income tax – exclusively for contributors in the 80th percentile of the income distribution (with annual income exceeding DKK 251.200). This change affected approximately 20% of adult contributors in the country, creating an ideal scenario to assess how people respond to changes in pension incentives.
The methodology adopted combined regulatory discontinuity analysis with detailed behavioral decomposition. Researchers compared the behavior of taxpayers marginally above and below the income threshold established by the reform. This approach allowed them to isolate the effect of the policy change, controlling for observable characteristics such as age, gender, and contribution history. To this end, they considered complete administrative records covering the entire adult Danish population between 1995 and 2009, encompassing 41 million observations.
According to the author, based on traditional neoclassical theory, all taxpayers affected by the reform would be expected to immediately adjust their savings strategies. Specifically, the model would predict: (i) a proportional reduction in contributions to capital pensions given the smaller tax advantage; (ii) partial reallocation of these resources to other savings vehicles (such as annuity pensions); and (iii) homogeneous behavioral responses across the affected population, with variations explained only by differences in objective parameters such as income elasticity.
Empirical evidence, however, revealed markedly different patterns. Although the reduction decreased aggregate contributions by almost 50%, the majority of individuals (80,7%) maintained their contributions unchanged, while only 19,3% adjusted their behaviors as predicted by traditional models. This revealed that strategies based on defaultsAutomatic contributions sponsored by employers, such as those that are more effective, are increasing total savings by about 85% without requiring active action from individuals.
- The EITC program
The EITC is the largest tax credit program for low-income workers in the U.S. In 2012, the program benefited 27.8 million taxpayers, at a federal cost of $63 billion. The program has a three-phase structure: (i) accumulation phase, where the credit increases with income (rates of 34-40%); (ii) plateau with constant benefit; and (iii) gradual reduction phase (rates of 16-21%). The optimal point at which beneficiaries maximize their tax credit occurs with annual incomes of $8.970 for families with one child or $12.590 for families with two or more children.
To understand the impacts of the policy, the methodology adopted explored the natural geographical variation in knowledge about the program. Using tax data from 78 million taxpayers (1996-2009), the researchers mapped income reporting patterns, especially comparing self-employed workers (with greater control over reported income) and salaried workers (with income verified via W-2). The analysis focused on “bunching"Anomatous concentrations of yields close to the optimal points of the program."
According to neoclassical theory, at least three assumptions underlie the context of this policy. First, all beneficiaries, once informed, would perfectly align their labor supply decisions with the program's marginal incentives. Second, responses would be homogeneous across different population groups, varying only according to objective parameters such as time elasticity. Third, regional differences would reflect only variations in observable characteristics of the population or the labor market. Thus, these models assume that workers perfectly understand and process the complex incentive structure of the EITC.
Upon analyzing the data, the author identified patterns inconsistent with neoclassical models. Self-employed workers in areas with high beneficiary density (such as South Texas) showed a strong concentration of incomes close to the optimal EITC values (5,2% reporting within ± US$500 of the optimal point), while salaried workers and regions with less awareness of the program (such as Kansas) showed significantly weaker responses (less than 0,5% of bunchingMigrants to areas with greater knowledge adapted their behaviors, with an increase of 2,1 percentage points (pp) in bunching. However, the reverse did not occur, indicating that the understanding of incentives spreads socially and asymmetrically. These patterns suggest that: (i) the understanding of incentives is gradual and socially mediated; (ii) cognitive costs limit perfect optimization; and (iii) interventions that simplify information can enhance the effects of fiscal policies.
- The MTO program
The MTO program was an experiment conducted between 1994-1998, providing vouchers housing for 4.604 low-income families in public projects in five cities in the United States. Participants were randomized into three groups based on their neighborhood of destination: (i) vouchers valid only for neighborhoods with less than 10% poverty; (ii) vouchers without geographical restriction; and (iii) control group (did not receive vouchersThe study data followed participating families for a period of 10 to 15 years after the intervention, and were supplemented with federal administrative records that tracked socioeconomic indicators at the national level.
Even considering the long-term benefits to the children of participants who used the vouchers (Chetty and Hendren, 2015), some families chose to remain in their localities. According to the authors, neoclassical models would explain these patterns through trade-offs rational. Specifically in this case, families would value other attributes more than long-term benefits for children, such as lower commuting costs (commute) or proximity to friends. Behavioral economics theories, however, suggest different explanations for permanence.
To analyze behavioral theories, the author adopted three strategies to evaluate housing policies from a behavioral perspective. First, they performed a direct measurement of experiential utility through... surveys This study analyzed subjective well-being and objective longitudinal data from the MTO, comparing residential choices with observed outcomes. Second, it applied a revealed preferences analysis in contexts of full information, using as a reference families that received full assistance for relocation. Finally, a structural model was created to identify behavioral bias, estimating parameters such as intertemporal discount factors. To incorporate uncertainty into the model, optimal policies were compared in two scenarios – “rational” agents (neoclassical perspective) and agents with behavioral biases (behavioral economics perspective).
The first two methods revealed that families who moved to less impoverished neighborhoods experienced significant gains in subjective well-being and better long-term objective outcomes (income and children's education). However, these improvements were found to be substantially greater than families had anticipated in their original decisions, highlighting a clear divergence between experiential and decision utility. Analysis in contexts with complete information showed that when families were fully aware of the benefits of the new neighborhoods, their choices aligned better with the actual outcomes, suggesting that information asymmetry was a key factor in suboptimal decisions.
Regarding structural modeling, an underestimation of future benefits and an overestimation of the immediate costs of change were identified. The estimated parameters indicated that families underestimated the long-term gains associated with better neighborhoods by approximately 30%. Simulations with these parameters demonstrated that policies combining nudges Informational interventions with moderate subsidies (approximately 40% of the change costs) could increase real well-being without compromising autonomy of choice. These results suggest that mild behavioral interventions may be more effective than traditional subsidies in contexts with underlying cognitive biases.
- Lessons in Public Policy
In this article, the author conducted several empirical studies to identify how the adoption of a 16-year-old IMCA (International Age for Alcohol Consumption) in Austria impacted alcohol consumption behavior among adolescents. The results indicate that legal access to alcohol at age 16 significantly increased the frequency and quantity of consumption, with particular emphasis on episodes of... binge drinking and hospitalizations due to alcohol poisoning. This effect was more pronounced among boys and adolescents from low socioeconomic backgrounds. Furthermore, it was identified that the perception of risk associated with excessive alcohol consumption decreased after the age of 16, while physical access to alcohol was already relatively easy even before that age.
The evidence presented in this article helps to understand the mechanisms behind the increase in alcohol consumption after the legal drinking age, providing relevant information for the formulation of public policies. The author emphasizes that, given the observed negative impacts, awareness campaigns about the risks of early consumption can be effective in mitigating these effects, especially for young people with a family history of alcohol abuse. Furthermore, promoting policies that reduce the availability of alcohol to adolescents and reinforce the perception of risk associated with its consumption has the potential to contribute to reducing problems related to alcohol abuse in this age group.
References
CHETTY, R. Behavioral Economics and Public Policy: A Pragmatic Perspective. American Economic Review, vol. 105, no. 5, 2015.
CHETTY, R, Hendren, N and Katz, L. Working Paper. “The Effects of Exposure to Better Neighborhoods on Children: New Evidence from the Moving to Opportunity Experiment.” 2015
