Posted by Kathryn Schwartz on June 25, 2014

The results suggest a number of extensions. Given the limitations in our sample period and the reliance on T-asymptotics to identify the parameters of interest, we have worked with very simple preference specifications. It would be interesting to work with preferences that are a more general the way demographic and labour supply factors are allowed to affect utilities. One possibility would be to estimate these effects using a longer time period and the Euler equation for a relatively safe asset and then check over the shorter period whether the orthogonality conditions hold, given the particular preference structure estimated. More generally, it would also be interesting to consider more flexible forms of preferences, including the non-expected utility preferences of the kind studied by Epstein and Zin (1989) and models with habit formation. The problem with the latter, however, is that they are very hard to study without longitudinal data.

The most important challenge, and the puzzle that our study leaves unresolved, is to explain the limited ownership of shares more structurally, particularly given the size of average excess returns. The descriptive evidence we present is suggestive and shows that the increase in ownership was quite widespread in the population. While it was probably triggered by the privatisations of the mid 1980s, and by the associated publicity, the trends cannot be explained only by that episode and/or by the ownership of shares in privatized firms. Understanding the factors, such as fixed costs, that still prevent ownership for large sectors of the population remains an important topic for future research.

Tags: , ,