This provides an additional motivation to examine the disincentive in response to the very large inheritances in the EITM data.
Table 8 presents estimates of changes in (log) earnings in response to the larger inheritances in the EITM.
Finally, the lack of controls (other than age) for wages in the EITM will produce a positive bias in the estimate of the disincentive if inheritance is correlated with earlier human capital transfers and, in turn, wage growth.
In addition, the results are very similar to those reported by Holtz-Eakin, Joulfaian, and Rosen (1993) who also examined the EITM and found that labor force participation declined as inheritances increased.
Although the inheritance disincentive is typically estimated to be significantly different from zero, all but one of the point estimates in both the PSID and EITM are smaller than the lower bound of the range of elasticities (-0.
To check this, we investigated the earnings responses by heirs of large fortunes in the EITM dataset.
There is potentially some measurement error in the EITM inheritances when decedents formed trusts because the beneficiaries of trusts are not identified.
The elasticity of earnings with respect to inheritance in the EITM is positive and significant when the year dummies are omitted.
Using men's wage information, which is unavailable in the EITM, does not greatly affect the (2SFE) disincentive estimate, although the estimated earnings increase three years after the inheritance is two-thirds the size and no longer significant.
The labor disincentives in the PSID and EITM data would be more clearly compared by presenting the estimates of regressions in the levels of hours and inheritance rather than the logarithms.
Among these, only the heirs with the two lowest inheritances had elasticities within a 90 percent, one-sided confidence interval based on the largest EITM estimates.