Abstract Consistent with the tax shield substitution effect, this study documents that firms that adopted LIFO in 1974 exhibit a negative association between the post-adoption change in their cost of goods sold and the post-adoption change in their interest expense. This negative association Is not observed for firms that retain FIFO. These results suggest that the capital structure and LIFO/FIFO choices are not independent of each other. A contribution of this study is that it investigates the response of tax shields to the change in a non-investment tax shield (i.e., cost of goods sold), thus providing evidence concerning the generalizability of the substitution effect.
Robert Trezevant (Mon,) studied this question.