This paper attempts to determine the relationship between conditional stock market volatility and conditional macroeconomic volatility based upon monthly UK data covering the period January 1967–December 1995. Conditional volatility is estimated using the well-known Autoregressive Conditional Heteroscedastic (ARCH), Generalised ARCH (GARCH) models. The macroeconomic variables used include industrial production, real retail sales, money supply, inflation, and an exchange rate variable, namely the German Deutsche mark/pound
For a number of years, much interest has been shown in the behaviour of stock market volatility (see Baillie & DeGennaro, 1990, French et al., 1987 and Poon & Taylor, 1992). Ever since the development of the Autoregressive Conditional Heteroscedastic (ARCH) models by Engle (1982), a model exists that allows the conditional variance, which is a measure of volatility, to change over time.1 Given that a model exists to measure this conditional volatility, the question that remains to be answered is that of why does conditional stock market volatility change over time.
Attempts have been made to examine the relationship between conditional stock market volatility and the release of information in the market place.2Fraser and Power (1997), examining the UK, US, and a number of Pacific Rim equity markets, found no substantial evidence to support the argument that the magnitude of information arrival in the market place has any impact upon conditional stock market volatility.3 The use of macroeconomic variables in explaining stock return volatility changes was examined by Schwert (1989) using US data. Using several macroeconomic variables, namely inflation, industrial production, and money, his findings showed weak evidence that macroeconomic volatility can help predict stock return volatility. A study by Liljeblom and Stenius (1997) based on Finnish data found that between one-sixth and more than two-thirds of changes in conditional stock market volatility was related to conditional macroeconomic volatility, namely inflation, industrial production, and money supply.
The objective of this paper is to examine conditional volatility so as to determine whether changes in UK stock market volatility through time, as measured by the conditional variance, can be explained by time-varying conditional volatility in a number of macroeconomic variables.4 The theoretical explanation for such a relationship comes from recognising that security prices are given by the discounted present value of expected future cashflows (Eq. (1)):
equation(1)
where Dt+k represents the capital gain plus any dividend paid to the shareholder in period t+1, and 1/(1+Rt+k)k is the discount rate for the period t+k based on information available to at time t−1. Et−1 denotes the conditional expectation. The conditional variance of security prices at time t−1, vart−1Pt will depend upon the conditional variance of expected future cash flows and future discount rates and on the conditional covariances between them. If one was to assume constant discount rates, then the conditional variance of both security prices and expected future cash flows should be proportional to one another. Since the level of corporate equity on the aggregate level should in turn depend on the health of the economy, it is plausible that a change in the level of uncertainty about future macroeconomic conditions would produce a proportional change in security return volatility.
The paper is organised as follows. Section 2 provides a description of the data used. Section 3 discusses the test strategy and methodology. In Section 4, the empirical results are presented. Section 5 provides the conclusion.
This study has examined the relationship between the conditional volatility in the UK stock market and a number of macroeconomic variables. In terms of the VAR estimation, a significant relationship was found between stock market and macroeconomic volatility with respect to the ability of macroeconomic volatility in predicting stock market volatility. Tests of the joint and simultaneous explanatory power of macroeconomic volatility were found to be disappointing since no significant relationship was found. It was found, however, that the inclusion of the conditional volatility of inflation in the conditional stock market volatility equation resulted in an improvement in the goodness of
fit based on the difference in the log-likelihood value. Stock market returns are constantly changing, and the volatility that exists within the stock market can be estimated using the ARCH models. This paper has looked at whether the changing conditional volatility in the stock market can be explained, in part, to the conditional volatility that exists within macroeconomic variables. Based on the macroeconomic variables used, from this study, one would have to conclude that the volatility in
the macrovariables selected does not explain the volatility in the stock market.
