What drives stock indices? A reality check: Economic Times 17th November 2010

The India shining story is often linked with the reflection seen in these numbers. At the same time, sceptics hold that the market is sentiment-driven and hence, at times, even good news on the economy front can be overwhelmed by random events. The market buzz on hiking of interest rates could lead to a decline in the market when economic conditions are otherwise sanguine. What really is the true picture?

The right way to go about this exercise is a regression analysis looking at changes in the Sensex and juxtaposing them with economic variables which prima facie have a bearing. The variables that can affect the market mood are growth in credit, industrial growth, capital issues, inflation, FII investments, exchange rate movements, changes in forex reserves and mutual funds investments.

The period chosen is from April 2006 to September 2010 and data has been reckoned on a monthly basis. This takes care of the day to day aberrations in the stock market movements. It has been noticed that even on a daily basis news affects the Sensex only momentarily which is generally mean reverting by the end of the day.
The multivariate regression model gives some interesting results. The first is that growth in credit, industrial production, exchange rate and changes in forex reserves do not really affect the Sensex. In fact, industrial production growth has a negative effect on the Sensex (though it is not significant, meaning the relation is spurious). The sign is also negative for the exchange rate suggesting that a falling rupee pushes down the market mood. Usually one would associate credit growth to mirror industrial buoyancy, but when it comes to the market, it doesn’t really matter.

How about the significant variables? The other four variables are significant. Higher inflation actually pushes back the Sensex and goes with a negative sign, which makes sense. But it questions the thought that the stock market buffers one from inflation. Primary issues have a negative relation with the Sensex, which is not what one would expect as the two are expected to move in consonance as IPOs become more visible when the market is on the rise. The other two important variables are net FII and mutual fund investments, which have coefficients of 0.003 and 0.00097. A net inflow of $1 million of FII funds leads to 0.003% increase in the Sensex or $1 billion implies 3% increase in the same. The impact of mutual funds is more muted with the coefficient being 0.00097. While these numbers by themselves are not sacrosanct, the major takeaway is that these four variables explain 48% of the variation in the Sensex, also called the coefficient of determination. What does this mean?

This implies that roughly half of the variation in the stock indices is driven by economic factors, while the rest is guided by sentiments that cannot really be explained. Therefore, when we talk of market sentiment which is positive or negative, it is really a collection of different trading thoughts that are not amenable to statistical calculation.
The implication is significant as it also means that we should not get swayed by this index in terms of being reflective of something dramatic in the economy. Maybe, this is why the stock market gyrations are often associated with the human emotion of exuberance — albeit irrational, when things move downwards.
The other important metric that can be examined within the world of econometrics is causation. While FIIs and mutual funds appear to be drivers of the market in statistical terms, is it possible to say that they cause the Sensex to move? The Granger Causality tests carried out do not show a relation either ways — FIIs or mutual funds causing the Sensex to move or the Sensex causing these flows to expand or contract. What all this means is that the stock market movements will remain an enigma for the statistically minded investor where economic fundamentals explain part of the story, while the rest will remain the proverbial mystery.