The predictability of asset returns is one of the more controversial topics in financial economics. According to the Efficient Markets Hypothesis (EMH), all information in investors’ information set should be incorporated into asset prices, thus leaving asset returns unpredictable. However, researchers and practitioners have identified a number of variables that seem to predict returns, at least in-sample. Some examples of variables that have been shown to predict aggregate market returns are inflation (Fama and Schwert, 1977), the default spread and the term premium (Fama and French, 1989) and the dividend-price ratio (e.g. Campbell and Shiller, 1988; Fama and French, 1988). Other financial ratios, such as the book-to-market ratio (e.g. Kothari and Shanken, 1997) and the earnings ratio (e.g. Lamont, 1998) have also been shown to predict returns. Recently, other variables that are not price-based have also been found to predict the aggregate stock market, such as the output gap (Cooper and Priestley, 2008). Traditionally, these variables have been thought to capture time-varying risk, a concept which is not in conflict with a modern interpretation of informationally efficient markets. In addition there is a growing literature that abandons the idea of informational efficiency; this literature argues that behavioral biases induce stock return predictability. In one recent example Baker and Wurgler (2007) argue that investor sentiment predicts returns. Another recent example is Hong, Torous and Valkanov (2007) who find that certain industries lead the aggregate market; they argue that this is due to underreaction to information.
In this paper I examine how changes in the oil price affect stock markets. Oil is by far the most important commodity and as of September 2008 energy accounted for around 75% of the Goldman Sachs Commodity Index (S&P GSCI). Crude oil accounted for 40% of the entire index and Brent crude oil for another 15%, thus leaving oil products with a weight of around 55% of the entire commodity index. The individual commodities in the index are weighted by their respective world production quantities. Given its large weight, it is not surprising that changes in oil prices may be an important factor for fluctuations in the economy. It is also plausible that oil price changes are important for understanding changes in stock prices. However, no consensus seems to have been reached regarding the impact of oil price changes on stock returns. Two recent papers, Driesprong, Jacobsen and Maat (2008) and Pollet (2004), document that oil price changes predict stock returns. Driesprong et al. (2008) show that the predictability is strongest for developed markets and less pronounced for emerging markets.
Goyal and Welch (2008) show that most predictors, such as the dividend-price ratio, have performed poorly over the last 30 years. They examine a number of variables which have been suggested as predictors of the equity premium and show that the statistical significance of these variables comes from the years up to and including the oil shock of 1973-1975. This period is exactly when oil prices start fluctuating and hence my sample using oil as a predictor starts in this period. I show that oil, unlike the other known predictors, seemingly predicts well both in-sample (IS) and out-of-sample (OOS). Goyal and Welch (2008) write that the signal from a predictive regression could give an investor confidence in the signal if it showed (i) good IS and reasonable OOS performance over the sample period, (ii) an upward drift in both IS and OOS plots, (iii) an upward drift not only in special periods (such as the outbreak of a war etc.). I am particularly interested in the third point above; the oil price reacts sharply to e.g. military conflicts in the Middle East and OPEC crises. According to the advice from Goyal and Welch, an investor should not feel confident in the signal from the predictor if these military conflicts etc. are the only periods with good predictability. However, this is exactly what I find to be the case, thus casting doubt on the oil strategy’s virtue as an equity investment strategy.
A large literature has examined the macroeconomic effects of oil price shocks in the U.S. with a particular emphasis on the effects on growth and inflation, see e.g. Barsky and Kilian (2004) and Hamilton (2008) for a review. In particular, some of the largest fluctuations in the U.S. economy have been preceded by political events in the Middle East that have reduced the oil supply from this region. For instance from 1973 to 1975 the oil price doubled, and oil importing countries experienced both inflation and rising unemployment. The later oil price increases during the Iranian revolution and leading up to the Iran-Iraq war also caused stagflation. Turmoil in the oil market has also had positive effects on the economy; following the OPEC collapse in 1985 and 1986 the aggregate supply curve for oil was shifted to the right, oil prices fell dramatically, output grew, unemployment fell and inflation was low. However, the positive effects of the oil price decrease were not as large as the adverse effects following the oil price increases in the 1970s. Based on the above-mentioned events, it has been suggested that U.S. macroeconomic variables respond in a nonlinear fashion to oil shocks. Mork (1989) showed that oil price increases affect the economy whereas decreases do not. Another transformation has been suggested by Lee, Ni and Ratti (1995) who argue that ”an oil shock is likely to have greater impact in an environment where oil prices have been stable than in an environment where oil price movement has been frequent and erratic”. They control for this by scaling the oil price by the conditional volatility as measured by a GARCH model. The effect of the Lee et al. (1995) transformation is that a small shock that occurs in a calm period will be scaled up whereas a large shock in a volatile period will be scaled down. Elaborating on these ideas, Hamilton (2003) further examines the functional form for oil and the macroeconomy.
Download
PDF Ebook Oil Price Shocks and Stock Return Predictability
