AbstractThis thesis investigates the weak-form of the Efficient Market Hypothesis (EMH) by examining the behaviour of equity returns in the Amman Stock Exchange (ASE). In particular, the 10 largest sectors in terms of market capitalisation and number of listed companies are considered. According to the weak-form of EMH, current share prices reflect all available historical information such that investors should not be able to beat the market or earn abnormal returns consistently by trading on the basis of historical price data. It is important topic for a country seeking to promote economic development as well as foster the financial and regulatory development that Jordan has sought to publicise over the past few years. Furthermore, while a number of studies have investigated the topic of stock market efficiency for ASE they have tended to focus on the whole market index, or used the old sectoral classification. To the best of the researcher’s knowledge no study has either used the new industry grouping or applied the multivariate General Autoregressive Conditional Heteroscedasticity (GARCH) model to test for time-varying variance and correlations between sectoral index returns in the ASE. This thesis tries to fill this gap in the literature by investigating market efficiency in the ASE using the new sectoral classification and finding the determinants of any interdependence between sectors.
In the first part of the analysis, the weak-form of EMH is tested by examining the risk-return relationship of the 10 ASE sectoral equity indices. Persistence in share volatility is investigated and the leverage effect is also studied by employing univariate symmetric and asymmetric GARCH models. A large sample of daily sectoral index data is used in the analysis over the 10-year period from 2003 to 2012. The results from estimating various GARCH models indicate that the returns for different sectors are characterised by different levels of volatility persistence and almost all reject the hypothesis that the ASE is weak-form efficient. This finding implies that share price changes in a sector may be predicted from its own historical information on return and volatility and a more up-to-date trading system is needed or more regulations concerning corporate disclosure are required.
To obtain a more in depth understanding of the share price formation process than that supplied in the first part of the analysis, the dynamic linkages between the 10 sectors of ASE in terms of both return and volatility are investigated in the second empirical component of the thesis. A Granger Causality test is employed to study the relationships between each pair of equity returns from the ASE sectors. The return spillovers between the 10 sectors of the ASE are investigated using a multivariate Vector Autoregressive (VAR) model, while the volatility spillovers in own as well as in other sectors’ returns and the dynamic conditional correlations among the sectors is examined using a multivariate Threshold GARCH model with Dynamic Conditional Correlation (DCC-MVTGARCH). In terms of return spillovers, the results indicate that the mean return of a sector is only affected by changes of historical share prices from other industries in a minority of cases. By contrast, evidence of interdependence in the volatility across the 10 ASE sectors is more evident; a large number of shocks and volatility spillovers are uncovered in the data. Furthermore, the results indicate that correlations between the sectoral equity returns are time-varying and not constant over the period of investigation. The results support the notion the news in one sector influences not only the return in that sector but also the variance of price changes in other sectors through their input-output linkages. These findings imply that the sectoral equity returns of the ASE are predictable from historical share price changes in their own series while their return volatility and interdependences are also predictable from the past volatility of the other sectors; this result calls the weak-form of the EMH into question since it suggests that investors can outperform the market by studying historic return and volatility information in the ASE.
Finally, building on the previous analysis the determinants of the time-varying conditional correlations between the different pairs of sectoral returns are investigated. Firm specific as well as macroeconomic variables are found to be significant determinants. In particular, 10 financial ratios and seven macroeconomic variables are investigated; a Principal Component Analysis (PCA) is used to narrow down the most relevant factors. Principal components are then extracted and used to construct the independent variables in the panel data analysis to explain the time-varying sectoral return correlations. The resulting findings show that profitability, aggregated demand and inflation are important in explaining time-varying sectoral return correlations between the ASE sectors. However, a further analysis indicated the effect on asset correlation of liquidity, profitability and stock market performance (growth) depends on aggregate demand (economic vulnerability).
|Date of Award
|German Jordanian University
|David Power (Supervisor) & Nongnuch Tantisantiwong (Supervisor)