Abstract
The concept of market efficiency is one with profound implications for asset returns. If studying past prices does not present opportunities for systematic excess returns to investors, then technical analysis will be an ineffective tool for asset allocation. In stronger forms of the efficient market hypothesis (EMH), full access to a company’s accounts or even having private information would not lead to a systematic advantage. Asset prices fully reflect all information and are expected to be random and unpredictable. Hence, efficient stock prices are said to follow a random walk process. The objective of the research presented in this thesis is to provide a detailed characterisation of UK stock market returns. It achieves this by (i) exploring the risk factors that determine excess asset returns, (ii) analysing asset price volatilities and quantifying their relationship with a group of macroeconomic variables, and (iii) testing the efficacy of momentum strategies. There is an extensive literature conducting similar tests using US data but there is a lack of similar work using UK data. The research presented here aims to fill that gap by running extensive tests using recent data and presenting the results in a unified framework. Following a review of the recent literature, the starting point is the application of the arbitrage pricing theory (APT) and a test of its special case, the capital asset pricing model (CAPM). Methodologically, the approach is the one established in the literature, i.e. running two-step regressions and forming beta-weighted portfolios. Departing from the literature, the innovations are generated from an autoregressive integrated moving average (ARIMA) model. Seven macroeconomic variables are found to be ‘priced’ using the APT model: real retail sales, industrial production, oil price, retail price index, money supply, long-term government bond yield and private sector bank lending. The evidence is rather damning for the CAPM. The volatilities of returns come in focus next. Appropriate generalised autoregressive conditional heteroscedastic (GARCH) procedures are applied and the relationship with the volatility of the macroeconomic variables—identified as significant in the previous chapter is explored in the context of ordinary least squares (OLS) regressions and vector autoregressions (VAR). The market variable is found to be an important factor in these models and also in the dynamic conditional correlation (DCC) model employed separately. By using a different type of portfolio sorting method, a momentum strategy is constructed and is found to be successful in predicting future returns using past share prices. In this case, a 12-month/12-month momentum strategy is found to be the most profitable. This finding suggests that the weak-type of EMH is rejected using UK data for the period 1985-2010. Finally, two macroeconomic variables are considered as potential determinants of momentum return and industrial production shows a significant effect.