The Bathtub Theory of Markets

When taking a bath one of the better ways to circulate newly added hot water is to swirl the water in the tub. One hand paddles backwards, the other goes forward and a giant wave soon circles the tub. At any given moment there is a limited volume of water in the tub. The goal is to circulate the water around to each point in the tub equally.

Our friend Sushil Kedia has posited that the Indian market leads the US market. Perhaps this is an application of the bathtub theory. To test this we might wish to look at simple correlations between INP, the Indian ETF, and SPY, the S&P 500 ETF. If in fact the Indian market leads we would expect that there wold be significant correlations at some lag. Following are the recent correlations of the INP with SPY at various lags:


Notably, five of six are strongly negative and the sixth is insignificant at 1.2%. So it would seem that the globe is a giant bathtub and that water does slosh around it in a non random manner. But if there is correlation then we know that a predictive regression model can be developed.

Using only lags 1, 3, 4, 5 & 6, our regression model has an overall correlation of 24.9%. All of the coefficients are negative, mostly in the neighborhood of -.04. The regression constant is -.001, reflecting the recent weak market behavior. Effectively what this says is that when money flows out of the Indian market it tends to go into the US market over the next six days or so. The converse is valid as well, in that if money goes into the Indian market it will come out of the US in the next six days. In any event the bathtub theory works reasonably well for the Indian - US markets, with India leading.