Working Paper: NBER ID: w1731
Authors: Steven F. Venti; David A. Wise
Abstract: Tax-deferred savings are potentially an important component of savings for retirement and could represent a very substantial increase in tax-free savings for many employees. IRAs may also have a substantial effect on national savings. Total IRA contributionsin 1982 were over 29 billion dollars. Despite the program's size and potential significance, little is known about the determinants of IRA contributions.This paper presents: (1) analysis of the effect of individual attributes on whether a person contributes, (2) analysis of the effect of individual attributes on how much is contributed,and (3) simulations of the effect of potential changes in contribution limits on the amount that is contributed to IRA accounts. Results of a similar analysis based on Canadian data are compared with results for the United States. Persons with low incomes are unlikely to have IRA accounts. In addition, after controlling for income, age, and other variables, persons without private pension plans are no more likely than those with them to Contribute to an IRA. The analysis of Canadian data yields similar findings, and indeed specific parameter estimates for the two countries are very similar. Simulations based on the estimates suggest that the current Treasury Department proposal would lead to about a 30 percent increase in IRA contributions.
Keywords: IRA contributions; tax-deferred savings; retirement savings; contribution limits
JEL Codes: H24; H55; D14
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Contribution limits (D64) | IRA contributions (H20) |
| Income (D31) | Likelihood of contributing to an IRA (D14) |
| Private pension coverage (H55) | Likelihood of contributing to an IRA (D14) |
| Contributor status (Y70) | Contribution amounts (D64) |