IMPACT OF STOCK MARKET VOLATILITY(PART-I)

Impact of Stock Market Volatility (Part-I)

-----Dr.Debesh Bhowmik

The conventional finance theory suggests that the stock market (excess) return, being a forward-looking variable that incorporates expectation about future cash flows and discount factors, contains useful information about investment and future output growth. Empirical literature provides substantial evidence in favour of this proposition .It is also seen from a number of recent studies that increased stock market volatility depresses economic activity and output . Empirical results presented by Campbell et al. (2001), in particular, are very important in this regard. They show that each of stock market volatility and excess return, when considered singly (i.e. when only one of these variables - either return or volatility - is considered) after controlling the lag dependent variable, is significant in explaining future output growth. But, when both volatility and excess return are considered as regressors (in addition to lag dependent values), volatility drives out return in predicting future output growth. As per the existing literature, stock market volatility may affects output growth through several possible channels, such as, (i) its link with market uncertainty and hence economic activity, (ii) association between market volatility and structural change (which consumes resources) in the economy, (iii) link of volatility with cost-of-capital to corporate sector through expected return. It is, however, not clear to justify why volatility drives out return in predicting output growth as observed by Campbell et al.(2001). Guo (2002) has discussed major arguments put forward by the proponents of volatility effects on output and has reconciled the evidence provided by Campbell et al. (2001) with earlier empirical evidence on predictive power of the stock market returns and finance theory. Based on a small model he argues that volatility may influence output growth (or may drives out returns in predicting output) in some specifications possibly because of its influence on cost of capital through its link with expected return.

But if cost of capital is the main channel through which volatility affects output then returns should play more important role in forecasting output growth than volatility does. He also provides empirical results to support this hypothesis. He derives relevant results for three different time periods; one longer than (but covering), one identical with, and another adding more recent years but shorter in length than the Campbell et al. (2001) sample period. Interestingly, using Campbell et al. (2001) sample, he finds that

the volatility drives out returns in predicting output growth because of the positive relation between excess returns and past volatility; if this relation is controlled for, excess returns show up significantly in the forecasting equation. In the liberalisation era, volatility in Indian financial markets is believed to have

increased/changed and thus there is a need to assess the impact of financial market volatility on output growth. With this background, this article presents some preliminary observations regarding link between stock market volatility/excess return and future output growth.

Some recent studies have shown that elevated stock market volatility depresses output. As per the conventional finance theory, however, it is the stock market (excess) returns that should have impact on future output growth. Currently, the issue is important in India, as there has been a perception that the volatility in Indian financial markets has increased/changed during the liberalisation era. Empirical results show that stock market volatility is strongly influenced by its own past values – pointing to the presence of significant volatility-feedback effects in the stock market. The volatility is also quite strongly related (at least contemporaneously) to excess return in recent years. However, roles of stock market return and volatility in predicting future output growth are not clear. Because, coefficients of lags of both these variables in the equation for future output growth are generally insignificant and in some cases have wrong signs, though during April 1997 to December 2002 only the volatility shows quite strong influence on output growth. Thus, there is a need to undertake further in-depth research for understanding the relationship between stock market return/volatility and future output growth in the context of Indian economy.

Thus, stock market uncertainty can have large effects despite the fact that households’ direct stock market participation is rather limited. Similarly, if stock market volatility can be viewed as an indication of how uncertain firms regard future developments, it can have a large effect on investment even if only a small fraction of firms in an economy

are subject to financing conditions determined by stock price movements. We provide empirical evidence on the relationship between stock market volatility and the business cycle and review the existing literature.

The empirical observation that stock market volatility tends to be higher during recessions points toward a negative relationship between stock market volatility and output. Fig-1 shows a scatter plot of U.S. quarterly percentage growth of real GDP against implied U.S. stock market volatility together with a fitted regression line.

The negative relationship between volatility and output growth is clearly visible. Scatter plots using historical volatility or GJR-based volatility instead of implied volatility show a similar negative relationship.

Although the empirical evidence presentedin the previous section indicates a close relationship between stock market volatility and economic fluctuations, the evidence is only suggestive.However, several papers document similar linkages using more detailed empirical approaches. The empirical study of Romer (1990) deals primarily with the onset of the Great Depression. However, Romer also presents estimates of the relationship between stock market volatility and consumption in the U.S.A.

Table-1: U.S. Ouarterly Stock Market Volatility

in Periods of Expansion and Recession

Fig-1:U.S. Stock Market Volatility and GDP Growth

for the post-war period. Using annual U.S. data ranging from 1949 to 1986, she concludes that a doubling of stock market volatility reduces durable consumer

goods output by about 6%, whereas the effect on nondurables is essentially 0. This ordering of the magnitudes of the effects is consistent with the idea that stock market volatility is closely related to uncertainty about future real economic activity. This is

because non reversibility gives rise to a lock-in effect that is particularly pronounced

during periods of high uncertainty.

Consider for instance a consumer deciding to buy a durable consumption good. Given the durable nature of the good and the uncertainty about future income, it may turn out that the good is either too modest or too luxurious with respect to future income. However, if the consumer waits until uncertainty is resolved, it may be easier to choose an appropriate good.Thus, by postponing the purchase of the good, the lock-in effect can be avoided and the benefit of doing so increases with the level of uncertainty. It follows that decisions that are irreversible to a larger extent are postponed, resulting in particularly pronounced reactions of durable consumption expenditures and investment expenditures to increasing stock market volatility.

Since investment decisions are presumably the least reversible, one would expect that stock market volatility has the largest effect on investment spending, followed by durable consumption and nondurable consumption. Note that if households substitute away from durable consumption goods into nondurable consumption goods because of higher uncertainty, then nondurable consumption may even rise during periods of high stock market volatility. Raunig and Scharler (2010) evaluate the uncertainty hypothesis by estimating the influence of stock market volatility on durable consumption growth, nondurable consumption growth and investment growth. Their analysis is based on quarterly time series data for the U.S.A. Based on a number of different estimates of time-varying stock market volatility, Raunig and Scharler (2010) find that stock market volatility exerts an economically and statistically significant effect on aggregate demand.

Moreover, they find that the adverse effect of stock market volatility on aggregate

demand depends on the extent to which decisions are reversible. Based on their richest specification (Table 2), they find that an increase in volatility by one standard deviation reduces the quarterly growth of durable consumption by around –0.70 percentage points, whereas the effect on the growth of nondurable consumption is only –0.14 percentage points. Investment growth responds with a lag of one quarter and declines by 1.12 percentage points.

Table-2:Effect of an Increase in Stock MarketVolatility by one Standard Deviation

on U.S. Consumption and Investment Growth

Hence, the decline in the growth of durable consumption and investment is larger during periods of increased volatility than the decline in the growth of nondurable consumption, which is again fully consistent with the predictions of the uncertainty hypothesis. In addition to being statistically significant, the estimated effects are also substantial in an economic sense. Stock market returns, in contrast to volatility, have a quantitatively smaller and often statistically insignificant influence on consumption and investment.

This result is consistent with Lettau and Ludvigson (2004), who also find that returns exert only a limited influence on consumption. The reason is that although permanent shocks to stock prices have a strong effect on consumption, most fluctuations in prices are transitory and exert only small effects on consumption.

Alexopoulos and Cohen (2009) identify uncertainty shocks using vector autoregressive methods. To measure uncertainty, they use stock market volatility measures, as in Raunig and Scharler (2010) and Choudhry (2003), and also an index based on the number of New York Times’ articles on economic uncertainty. They find that uncertainty shocks play an important role for the business cycle. In particular, uncertainty measured by the New York Times’ index accounts for up to 25% of the short-run variation in employment

and output. Choudhry (2003) analyzes the influence of stock market volatility on GDP and the components of GDP using an error-correction framework. Under the assumption that volatility follows a non -stationary stochastic process, he estimates the short-run and long-run dynamics of GDP components using an error-correction framework. His results confirm that stock market volatility has adverse effects on consumption and investment.

A different, but closely related, issue is analyzed in Jansen and Nahius (2003) They analyze how stock market fluctuations influence consumer sentiment in a sample of eleven countries. They find that in the vast majority of countries under consideration, consumer sentiment and stock returns are positively related. They also find that causality runs from stock returns to consumer sentiment rather than vice versa. Moreover, they conclude that the correlation between stock returns and consumer sentiment mirrors expectations about future economic conditions. Therefore, the evidence presented in their paper also provides some backing for the uncertainty hypothesis, in the sense that stock market fluctuations give rise to uncertainty about future economic conditions.

Note that although the uncertainty hypothesis suggests that causality runs from stock market volatility to the business cycle, this need not necessarily be the case. Although the early literature on the determinants of stock market volatility finds only weak linkages between stock market volatility and macroeconomic variables, recent empirical research (e.g. Engle et al., 2008; Diebold and Yilmaz, 2010) establishes important linkages between macroeconomic fundamentals and stock market volatility. In particular, Arnold and Vrugt (2008) find a strong link between macroeconomic uncertainty and stock market volatility using survey data from the Survey of Professional Forecasters maintained by the Federal Reserve Bank of Philadelphia. The authors find that rising uncertainty about future macroeconomic developments increases stock market volatility. Thus, taken together with the evidence presented above, it appears that causality between macroeconomic outcomes and stock market volatility is bidirectional.