1 Market efficiency and forecasting

The interest in predicting stock prices or returns is probably as old as the markets themselves, and the literature on the subject is enormous. Fama (1970) reviews early work and provides some organizing principles. This chapter concentrates selectively on developments following Fama's review. In that review, Fama describes increasingly fine information sets in a way that is useful in organizing the discussion. Weak-form predictability uses the information in past stock prices. Semi-strong form predictability uses variables that are obviously publicly available, and strong form uses anything else. While there is a literature characterizing strong-form predictability (e.g. analyzing the profitability of corporate insider's trades), this chapter concentrates on the first two categories of information. For a while, predicting the future price or value (price plus dividends) of a stock was thought to be easy. Early studies, reviewed by Fama (1970), concluded that a martingale or random walk was a good model for stock prices, values or their logarithms. Thus, the best forecast of the future price was the current price. However, predicting price or value changes, and thus rates of return, is more challenging and controversial. The current financial economics literature reflects two often-competing views about predictability in stock returns. The first argues that any predictability represents exploitable inefficiencies in the way capital markets function. The second view argues that predictability is a natural outcome of an efficient capital market. The exploitable inefficiencies view of return predictability argues that, in an efficient market, traders would bid up the prices of stocks with predictably high returns, thus lowering their return and removing any predictability at the new price (see, for example , Friedman, 1953; Samuelson, 1965). However, market frictions or human imperfections are assumed to impede such price-correcting, or 'arbitrage' trading. Predictable patterns can thus emerge when there are important market imperfections, like trading costs, taxes, or information costs, or important human imperfections in processing or responding to information, as studied in behavioural finance. These predictable patterns are thought to be exploitable, in the sense that an investor who could avoid the friction or cognitive imperfection could profit from the predictability at the expense of other traders. The 'efficient markets' view of predictability was described by Fama (1970). According to this view, returns may be predictable if required expected returns vary over time in association with changing interest rates, risk or investors' risk-aversion. If required expected returns vary over time there may …