Dangers in data adjustment: The case of rights issues and returns

Abstract

In this paper we alert researchers to the potential for unrecognised errors in using adjusted price and daily return data. This problem is illustrated by considering the case of ex-rights price adjustments. We present five alternative adjustment procedures that would be expected to generate similar results. We show, however, that these procedures result in significantly different dilution factors and returns. Our investigations suggest that the problem is associated with the theoretical valuation of the rights. In a substantial proportion of cases, the standard textbook model is inappropriate because of the non-standard nature of the rights issue. Correcting for these non-standard cases is a non-trivial task since they constitute more than half of the issues. The extent of this problem does not appear to be well recognised. Deletion of non-standard rights issues eliminates extreme values in dilution factors, but statistically significant differences remain. Our moral is simple; uncritical acceptance of data 'as is' from computer data files may lead researchers to erroneous conclusions. It also seems noteworthy that the standard textbook model of rights pricing only applied to a minority of Australian rights issues over recent years. This result has implications for the calculation of EPS under AASB 1027. As a by-product, our analysis suggests that the ex-rights daily return is close to zero. © AAANZ, 2001.