Global diversification is core to fund management

Business Standard, 16 August 2006

India is making tentative steps towards opening up to international diversification of local portfolios. This flies against a deeply held belief that India is a capital-scarce country, and that local capital should be forced to be in India so as to assist Indian growth. But globalisation means inter-dependence. The best path involves foreign investors holding Indian assets and Indian investors holding foreign assets.

We look at 16 years of weekly data for four stock market indexes: an index of large Indian stocks (NSE Nifty), an index of most Indian stocks (CMIE Cospi), an index of large US stocks (S&P 500) and an index of large UK stocks (FTSE 100). In terms of average returns, the Indian indexes have done extremely well. Nifty produced 0.288% per week, Cospi produced 0.312% per week. The S&P 500 did just 0.15% per week while the FTSE turned in an anaemic 0.108% per week.

But the Indian indexes had high volatility, reflecting the instability of the economy. Nifty had a weekly standard deviation of 3.88%. This roughly meant that in 95% of the weeks, Nifty returns ranged from -7.3% and 7.9%. For a comparison, in 95% of the weeks, the S&P 500 returns were between -3.9% to 4.2%. The weekly volatility for Cospi was 3.84% and for the FTSE it was 2.1%.

The correlations between the four indexes were often below 1, which suggests that gains from diversification could have been harnessed. Nifty and Cospi had a correlation of 0.93. Against the S&P 500, Nifty had a low correlation of 0.15 and Cospi was even better with 0.11. Nifty and the FTSE had a low correlation of 0.17.

What was the optimal strategy for an investor who knew the average returns, the volatility and the correlations between these four indexes? This question can be answered fairly sharply. The future is, of course, complex. Average returns, volatilities and correlations of the last 16 years might not hold in the future. However, if we can assume these facts, it is possible to compute optimal portfolios, which gives us an insight into global diversification for an Indian investor.

Suppose an Indian investor targets an above-Nifty return of 0.3% per week. We assume the investor has no access to borrowed money or short selling. In this case, the optimal portfolio consists of 92.3% invested in Cospi and 7.7% invested in the S&P 500. This delivered the target return of 0.3% at a cost of weekly standard deviation of 3.56%. The benefits from global diversification are already visible: this portfolio volatility was better than being purely invested in Cospi with weekly volatility of 3.84%.

Into this situation, we give the investor the power to short sell index futures on all four indexes, and the power to borrow or lend money. This is slightly unrealistic since futures on the Cospi are not available. However, in this case, the optimal portfolio becomes: +75.8% in Cospi, with a short position of 23.3% in Nifty, +86.2% in the S&P 500, a small 0.4% weight on the FTSE and the remainder financed by borrowing. This portfolio delivered the target return of 0.3% per week, while suffering a much lower volatility of 2.89%.

In three steps, we have seen great reductions in risk of the Indian portfolio. The investor who suffered a volatility of 3.84% for a pure Indian (Cospi) portfolio got a drop to 3.56% by simple global diversification, and a further decline to 2.89% when permitted short selling and borrowing. These are big gains, in response to going to only two countries (US and UK). The benefits from global diversification further improve as more and more countries are put into the calculation.

If a lower target return of 0.15% per month is sought -- roughly the same as the historical average return on the S&P 500 -- then the optimal portfolio works out to 37.9% in Cospi, 43.1% in the S&P 500, 0.2% in the FTSE-100, a short position of 11.7% in Nifty, and the residual financed by borrowing. This delivered the target return of 0.15% per week while suffering a volatility of only 1.45% per week. This globally diversified portfolio delivered the returns of the S&P 500 but at a lower risk, for the volatility of the S&P 500 was 2.09% per week.

These calculations suggest that international diversification is of profound importance for the Indian investor. International diversification is not a side show, where a few mutual funds will offer a few offshore portfolios, which will be tucked away on the side into household portfolios. Instead, households should aggresively seek to diversify assets globally so as to reduce their portfolio volatility caused by high Indian GDP growth volatility.

This is good for households -- who get a better risk/reward tradeoff. This is also good for India, since the returns from these globally diversified portfolios will smooth Indian consumption and help to reduce Indian GDP growth volatility.

From the viewpoint of mutual funds, every scheme needs to think international, for this is the only free lunch known in all economics. There is no mutual fund scheme in India today which delivers a reward/risk ratio (using weekly data) of 0.1. It is inefficient for any mutual fund scheme to be purely invested in India, i.e. to put all its eggs in one basket. A scheme that does so delivers an inferior risk/reward ratio as seen from the eyes of customers. A mutual fund which does not use foreign assets as "raw material" when making portfolios is as wrong as a factory that refuses to look at imported raw materials.

Finally, from the viewpoint of SEBI, recent guidelines have been put out to enable a tentative start for outbound investment by Indian mutual funds. These guidelines fail to appreciate how fundamental global diversification is for good fund management in India. They envisage separate international schemes, while the correct path is to internationalise every scheme. SEBI has done an unfortunate thing by having clauses where some kinds of outbound investment (ETFs) will only be permitted for mutual funds who have experience with outbound investment. It is like saying that industrial licensing is abolished as long as you already producing. Given that India is rising from a fifty-year autarkic slumber, most funds obviously have no experience with outbound investment, and will thus be disqualified.

Back up to Ajay Shah's 2006 media page
Back up to Ajay Shah's home page