Given the large changes in the level of ownership of shares and the differences across cohorts, it might be expected that the composition of the group of shareholders has changed over the period. In general, shareholders tend to be older and better educated than the rest of the population, but these differences have been getting smaller over time (as the results of the probit regression of share ownership in the next section will show more clearly).
At the start of the period, for example, the average age of shareholders was 45. This had fallen to 43 by the end of the period. Similarly, the proportion of shareholders with no further education was 28 % at the start of the period, but had risen to 39% by the end. Over the same period, the proportion of the whole sample with compulsory education actually fell from 61% to 54%. Table 1 shows that the biggest increases in share-ownership came among those with high incomes, but low levels of education. Over the period there has been a considerable narrowing of the difference in ownership rates between education groups. In 1978, households with college education were more than four times more likely to own shares than those in the same income quintiie with only compulsory education. By 1995, they were around twice as likely. If these changes in the composition of the group of shareholders imply differences in the consumption profiles of the group across the period, the implication of the consumption capital asset pricing market is that there will also be changes in asset market returns to preserve the equilibrium relationship. We explore this further below.