Working Paper: NBER ID: w4900
Authors: Orazio P. Attanasio; Thomas C. Deleire
Abstract: The effectiveness of tax-favored savings accounts in raising national savings depends crucially upon the willingness of households to reduce consumption in order to finance contributions to these accounts. The debate over the tax deductibility of IRA's has centered on whether IRA contributions represented new savings or reshuffled assets. We devise a test to distinguish between these two hypotheses where we compare the behavior of households which just opened an IRA account with that of households which already had an IRA account. Our test accounts for any unobservable heterogeneity across the two groups. We find evidence that supports the view that households financed their IRA contributions primarily through reductions in their stocks of other assets. Our results indicate that less than 20% of IRA contributions represented addition to national savings.
Keywords: IRA contributions; household savings; tax incentives; national savings
JEL Codes: H24; D91
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Opening IRA accounts (G23) | Decrease in non-IRA assets (D14) |
| New contributors (Y70) | Decrease in consumption (E21) |
| Old contributors (Y70) | No change in consumption (D19) |
| Tax incentives (H20) | Changes in consumption and asset behavior of new contributors (E21) |