Forecasting Models

Abstract

As outlined in the literature review section, Lo and MacKinlay (1990a) identify a leadlag relationship between the returns of small cap portfolios and the returns of large cap portfolios. The relationship is documented for return periods of up to four weeks. Thus, it seems an appealing approach to apply this finding in a forecasting model. In contrast, Bodoukh, Richardson, and Whitelaw (1994) argue that, even in a world in which large-firm returns are not more informative than the returns of small caps, there can be considerable cross-serial correlation. However, it cannot be ruled out that we live in a world where the returns of large caps (or other market segments) carry important information with respect to forecasting.