Absence of conflict of interest.
Citation
Alpert, A. & Powell, D. (2012). Tax elasticity of labor earnings for older individuals. (Report no. 2012-272). Ann Arbor, MI: University of Michigan Retirement Research Center, University of Michigan.
Highlights
- The study examined the impact of tax changes on people’s employment, retirement, and labor income outcomes.
- The study uses a nonexperimental approach and the data from the Health and Retirement Study to estimate the effect.
- The study found positive statistically significant relationships between increases in after-tax labor income and the employment of men and women as well as the delay in retirement for men. The study also found a positive statistically significant relationship between increases in the marginal net-of-tax rate and earnings for women.
- The quality of causal evidence presented in this report is moderate because it was based on a well-implemented nonexperimental design. This means we are somewhat confident that the estimated effects are attributable to tax changes from the Economic Growth and Tax Relief Reconciliation Act (EGTRRA), the Jobs and Growth Tax Relief Reconciliation Act (JGTRRA), and their marriage tax penalty relief provision, but other factors might also have contributed.
Intervention Examined
Tax Changes
Features of the Intervention
The EGTRRA of 2001 lowered tax rates for tax filers in each bracket and especially for those with low incomes. The JGTRRA of 2003 lowered tax rates mainly for households with high incomes. This legislation included a marriage penalty relief provision in the EGTRRA and JGTRRA that pushed many households into a lower tax bracket by raising the threshold of taxable income for married filers. The legislative tax changes affect single and married tax filers differently depending on the filer’s initial income and marital status.
Features of the Study
The study used a nonexperimental instrumental variables approach to estimate two effects: (1) the effect of the marginal net-of-tax rate, or the amount of pay that a worker keeps for an additional $1 in earnings, on an individual’s decision about how much to work (as reflected by their labor earnings) and (2) the effect of after-tax labor income on an individual’s employment and retirement decisions.
A problem with estimating the effect of taxes on employment behavior is that an individual’s tax rate and tax liabilities are partly determined by employment behavior. Using data on income and marital status, the authors calculated approximately how much of the changes in the tax rate and tax liabilities variables might be attributable to the legislative tax changes and not people’s behavior, and they used these as the instrument variables. The authors used a statistical model to examine the relationships between employment, retirement, and labor income and the tax changes, accounting for how much the changes in tax rate and tax liabilities are attributable to legislative tax changes. The statistical model controlled for age and education of the individual and a spouse (if applicable), labor income, total income, and year.
Data are from individual panel data from the Health and Retirement Study from 2000 to 2008 and simulated data from the National Bureau of Economic Research’s Taxsim program. The analysis includes a sample of 16,443 men and women ages 55 to 74 who worked during the period observed.
Findings
Employment
- The study found positive statistically significant relationships between increases in after-tax labor income and the employment of men and women. The study found that a 10 percent increase in after-tax labor income was associated with a statistically significant 0.8 percentage point increase in employment for men and women.
- The study found a statistically significant relationships between increases in after-tax labor income and retirement for men. The study found that a 10 percent increase in after-tax labor income was associated with a statistically significant 0.7 percentage point decrease in retirement for men. Changes in after-tax income did not have a significant relationship with women’s likelihood of retirement.
Earnings
- The study found a positive statistically significant relationship between increases in the marginal net-of-tax rate and earnings for women. More specifically, the study found that a 10 percent increase in the marginal net-of-tax rate was associated with a statistically significant 30 percent increase in labor earnings for working women. Changes in the marginal net-of-tax rate did not have a significant relationship with men’s labor earnings.
Considerations for Interpreting the Findings
The authors demonstrated that their instruments are sufficiently strong and made a convincing case that the only way the instruments affect the outcome of interest is through the endogenous variable. Further, the study had the same number of instruments as there are endogenous explanatory variables.
Causal Evidence Rating
The quality of causal evidence presented in this report is moderate because it was based on a well-implemented nonexperimental design. This means we are somewhat confident that the estimated effects are attributable to tax changes from the EGTRRA, the JGTRRA, and their marriage tax penalty relief provision, but other factors might also have contributed.