The familiar Hall-Jorgenson (1967) approach to measuring the cost of capital considers a perturbation from the optimal path of the capital stock, comprising a unit increment to physical capital in the current period and its disposal in the following period. The holding time for the incremental unit is therefore a single period. However, this pertutbation is not a uniquely feasible one (a similar point having been made by King (1974) in criticism of Stiglitz (1973)). In particular, because of institutional features of actual tax systems, the optimal holding time such an incremental unit of capital may, in general, be infinite. Conditions for this result to hold, and for the special case of indifference to holding time, are derived. It is shown that it is impossible to specify an optimal rule for recapturing excess depreciation allowances on disposal (independent of actual holding time) except under an expenditure tax.
The perturbation commonly considered in the literature may therefore provide misleading results. When the tax system is changing in a way that is foreseen, the difference in the resulting measure of the cost of capital can become highly significant. Failure to take account of this factor may have led to underestimation of the extent to which the radical changes to Corporation Tax in the UK in the 1984 affected the cost of capital over the transitional period to 1987. In particular, the size of the incentive to forestall planned investment by bringing it into an earlier tax year appears not to have been generally appreciated.
Download is not available
