Capital decay and tax distortions: How to abandon exponential decay and benefit from it

The appropriate way of quantifying how taxation of a firm's income and capital can distort its optimizing conditions is a recurring issue in the literature on optimal taxation. Exponential decay, although empirically contested, is almost ubiquitous. In the present paper a generalized framework which allows for a general, non-exponential, decay pattern for both true and tax-permitted depreciation, is considered. Both convex and concave survival functions can be accommodated. Three capital concepts are involved, two of which coincide under exponential decay. The trade-of between various departures from neutrality is illustrated. Elements which contribute to non-neutrality are: (i) discrepancy between the definition of the tax-relevant accounting capital and true depreciation, (ii) mis-indexation of depreciation allowances, (iii) incomplete deductibility of interest costs, (iv) asymmetric treatment of interest costs and capital gains, and (v) taxation of the value of the capital stock. Finally, we show that substantial biases can arise in assessing the degree of non-neutrality if non-exponential depreciation schedules are forced, by 'approximation devices', to fit into the exponential decay schedule.