The behaviour of stock returns in Amman stock market : a thin emerging market
In this thesis the behaviour of stock returns of firms listed on the Amman Stock Market is examined. The thin trading characteristic of the market is emphasised and its possible effects on empirical investigations are analysed. The first four chapters contain a review of the literature on the importance of stock markets, the Efficient Market Theory and the Capital Asset Pricing Model. The literature suggests that the allocative efficiency of funds via stock markets is related to the operational and pricing efficiency of these markets. In such an efficient market, the expected return on an investment is related only to its risk. Chapter 5 tests the weak form efficiency of the ASM with particular emphasis on the problem of thin trading. To achieve this, three alternatives for filling missing data gaps are examined. In particular, it was found that extrapolation, based on market movements, induces more dependence patterns. Yet, Examining the other two alternatives, using daily price changes, statistical inefficiencies were detected, on the one day level. Fewer dependence pa%erns were repored br \onger rnerva'1s. The reported first order positive serial correlation can coriseqnence. o'i 'p'icrng uirrxs imposed on trading in the market. Chapter 6 provides a database of individual stock and market returns. Compiling this database was a major contribution of this research. Chapter 7 investigates the effects of different return measurement and beta estimation approaches on tests of the CAPM. Specifically, the evidence indicates that the use of different return measurement approaches can affect the results of tests of this equilibrium model. Also, the adjustment of the trade-to-trade method, used for beta estimation, reduces heteroscedasticity resulting from using non equal time intervals when applying the market model. The first part of chapter 8 provides an investigation of the sensitivity of the results, of CAPM tests, to the length of the period used to estimate beta. The results suggest that the longer the period, used to estimate beta, the more are the reported deviations from the implied relationships of the model. The second part of Chapter 8 provides a test of the CAPM using pooled data, and employing four lengths of periods to estimate beta. The evidence was not consistent with the model. But, when specific attention was given to the problem of thin trading, by constructing sub samples of the most traded stocks, the validity of the model was established. However, this was only the case when beta is estimated using 24 months of past returns, suggesting that market risk in Jordan changes fairly rapidly. Chapter 9 investigates the power of some firm-specific variables in explaining the cross section of stock returns on the ASM. The evidence suggests that the book value, earnings, leverage and the firm size, do not help in explaining the cross section variation of firms listed on the ASM. This evidence is in accord with the CAPM.