Teaching the Tax Code: Earnings Responses to an Experiment with EITC Recipients in the United States

Investigadores/as:
Trabajo de campo:
Ubicación:
Chicago Metro Area, United States of America
Muestra:
42,002 tax filers at 119 tax preparation offices
Línea de tiempo:
2007 - 2007
Grupo objetivo:
  • Families and households
Resultado de interés:
  • Earnings and income
  • Taxation
Tipo de intervención:
  • Financial literacy
  • Information
Número de registro del AEA RCT Registry:
AEARCTR-0001601
Socios Implementadores:

Many recipients of the Earned Income Tax Credit (EITC) in the United States do not fully understand how their earnings affect the size of their transfer. Researchers sought to determine whether providing information on the EITC could increase the amount individuals received by randomizing the information professional tax preparers supplied. They found that providing more information did not increase EITC transfers or annual earnings overall. EITC amounts did increase among clients who were given the exact information recommended by researchers, indicating that variation in how tax professionals deliver information influences its effectiveness.

Problema de política pública

Tax and cash transfer policies are often complicated. Flat-rate taxes and lump-sum transfers are rare. Instead, governments often vary transfer rates and amounts with an individual’s income in order to influence their decisions on whether and how much to work. One of the goals of the Earned Income Tax Credit (EITC) in the United States is to encourage low-income households to work. Past studies of the EITC have found that it has increased labor market participation: low-income households started to work in order to qualify for the transfer.

However, few studies have examined how the EITC affects the choices of individuals who are already in the labor market. Many American households on the lowest end of the tax bracket could increase the amount they receive through the EITC if they worked more. One explanation for why households do not maximize the size of their transfer is that they are not fully informed about the policy and the incentives it creates. This lack of knowledge is not surprising considering that the Internal Revenue Service (IRS) does not explain how it calculates transfers in a transparent, accessible way. For legal reasons, the IRS distributes comprehensive documents that cover all possible contingencies, making it difficult to highlight the information relevant for a given taxpayer. A simple fix may be to provide individuals with this information.

Contexto de la evaluación

The EITC is the largest cash transfer program for low-income families in the United States. In 2009, 25.9 million tax filers received US$57.7 billion in EITC payments. Eligibility for the program and the size of the transfer depend on the filer’s earnings and number of dependents. In order to preserve incentives to work for the lowest income earners while reducing transfers to higher earners, the size of the transfer initially increases with earnings, plateaus at a certain level of earnings, and then decreases with earnings until it is phased out completely. The figure below shows this pyramid shape. To claim the EITC, individuals must include an EITC form in their tax return, due April 15 of each year.

Most low-income families know about and claim the EITC, but very few understand that the way it is calculated changes with the level of earnings. They rarely know the levels of income where the plateau and the phase out begin. Many of these families also seek professional advice about filing their tax returns: in 2004, 74 percent of families with children receiving the EITC used paid tax preparers. The largest tax preparation company in the United States, and the partner in this evaluation, is H&R Block, which has about 13,000 offices around the country.

Detalles de la intervención

The researchers used a randomized evaluation to test whether supplying information about the EITC caused taxpayers to change their behavior in a way that increased the size of their EITC refund. They partnered with 119 H&R Block offices in the Chicago metro area during the 2007 tax season. Families (with at least one child) that had previously received the EITC were randomly assigned to either the group receiving additional information or the comparison group.

Individuals in the treatment group received a two-minute explanation about the EITC from an H&R Block employee. The employee told the client her EITC amount based on 2006 earnings, verbally explained how the transfer works, showed a graph with the EITC amount as a function of earnings, and presented a table with EITC ranges. The graph and table were customized based on the filer’s marital status and number of dependents, and the employee circled the client’s position on both. Individuals received a printout to take home and a letter summarizing the EITC information the following August. The 1,461 tax professionals involved in the study earned $5 for each client with whom they spent at least two minutes discussing the EITC.

Members of the comparison group saw the EITC amount they would receive, but did not receive an explanation of why they would receive this amount. About 72 percent of filers in both the treatment and comparison groups returned to H&R Block the following year, which allowed the researchers to use 2008 tax information to see how the additional information changed individuals’ behavior in 2007.

Resultados y lecciones de la política pública

Providing information about how EITC transfers are determined had no significant impact on the size of EITC transfers or total earnings. On average, people did not adjust the hours they worked and amount earned in order to receive a larger transfer. The researchers separated filers with incomes in the phase-in region of the EITC, who would receive a larger transfer if they earned more; filers in the plateau region, whose transfers would not change if they earned a small amount more or less; and filers in the phase-out region, who would receive a smaller transfer if they earned more. They did not find that the extra information had an impact on the transfers these subgroups received.

The researchers worried that differences in training and willingness to convey the treatment’s message might have created variation in the treatments received. Only 73 percent of individuals assigned to receive information about the EITC discussed it with their tax professional for at least two minutes. Additionally, some professionals felt that they should always encourage clients to work more, so they may have deviated from the official script when advising clients in the phase-out region of the EITC, even though the script already downplayed that earning more would reduce these clients’ transfers. In addition, many tax professionals reported that participating in the evaluation increased their own understanding of the EITC, so they may have offered better explanations to individuals in the comparison group.

The researchers found that clients who received the advice researchers recommended increased their EITC amount by US$90 more than those who heard a deviation from the original advice, but they increased their total earnings by US$420 less. In other words, professionals who followed the script encouraged their clients to increase their EITC transfer, but deviators encouraged clients to increase their total income, even if it meant receiving a slightly smaller transfer. These results indicate that variation in the delivery of information can affect whether individuals change their behavior.

This one-time provision of information about the EITC did not affect hours worked and earnings in the following year. Knowledge that develops organically and persistently—such as information shared by peers—may have a greater influence on behavior. Policymakers seeking to shape economic choices should seek to understand how information about such policies spreads.

Raj Chetty and Emmanuel Saez. 2013. "Teaching the Tax Code: Earnings Responses to an Experiment with EITC Recipients." American Economic Journal: Applied Economics 5(1): 1-31.