Assessing Alternative Methods of Estimating the Present Value of Future Earnings: A Fifteen-Year Update

Abstract

In the inaugural issue of this Journal, Dulaney (1987) offered an assessment of the forecast accuracy of four “conventional” methods of estimating the present value of future earnings. He tested the four methods over 15 rolling 20year future periods, with the first period covering 1953-1972 and the last covering 1967-1986. For each method and for each period, he first calculated the estimated present value of the future loss. He then compared this estimated present value to the actual present value,1 the amount of money that actually would have been needed to replace the lost future earnings as determined with the benefit of hindsight. These comparisons between the estimated and actual present values formed the basis for an assessment of the forecast accuracy of each method. In this update, these same four conventional methods are evaluated again over the next 15 rolling 20-year periods, with the first covering 1968-1987 and the last covering 1982-2001. Meaningful comparisons of Dulaney’s and some new results for the earlier periods with the results for the more recent periods are possible for three of the four methods. These comparisons indicate that the forecast accuracy of all three methods was much worse for the more recent periods. Furthermore, results for the fourth method indicate that it is no more accurate than the other three for the more recent periods. These findings are placed in the context of other recent efforts that have been made to assess the forecast accuracy of various estimation methods. Considering this entire body of work, it seems fair to say that there remains considerable room for improvement in the methods used by forensic economists to estimate the present value of lost future earnings.