Abstract

This paper shows that the widely cited empirical relation between equity misvaluation and the choice of merger currency (Rhodes-Kropf and Viswanathan, 2004; Rhodes-Kropfet al., 2005; Shleifer and Vishny, 2003) is spurious. We argue that before its abolishment in 2001, pooling accounting was the accounting method of choice for highly valued acquirers and that failing to control for this regulatory incentive leads to invalid inference (likely because the choice of pooling accounting is an endogenous omitted variable). We confirm these arguments with new empirical results: (i) the relation between acquirer valuation and the choice of mode of payment disappears in analysis of U.S. mergers with post-2001 data; (ii) this relation also fails in examination of data from Europe and Australia, environments where pooling was either not allowed or almost never used; (iii) this relation is absent even in analyses of pre-2001 subsamples of U.S. mergers that did not use pooling.