Data suggest that over a five-year period, equity gives a premium of 3-4 per cent over debt. Over a 10-year period, this premium drops.
The buy and hold concept is simple. In a growing economy, corporate earnings outrun GDP and the broad equity market return outperforms other instruments. So an investor can hold a diversified portfolio over the long-term to his benefit. It's difficult to generate active equity returns that beat the broad market. So, most people are better off with systematic passive equity investments.
But the long term may be very long. For example, US stock markets were flat between 1966-1982. That seems extreme. But long periods of stagnation are not uncommon. Most non-oil economies saw negative returns between 1973-1979. Japanese investors have lost money since 1990. British investors lost money through the 1980s.
Post-liberalisation, Indian investors saw negative returns between 1994-2000 and 2000-2004. It appears that a five-year run of negative returns is not uncommon. Is there an optimal holding period with specific reference to the Indian market?
One way to get an inkling is via the examination of “rolling returns”. A rolling return assumes investor A bought equity on Jan 1, 1900 for example, while B bought on Feb 1, 1900 and C on March 1, 1900, etc. If each investment is held for exactly 10 years (From Jan 1, 1900 to Jan 1910 for our investor A) the return stream filters out short-term volatility at start and end points. So we can calculate compounded annualised return (CAR) for any given rolling period.
By comparing CAR for different rolling periods, we may be able to estimate the optimal holding period. The results are interesting. Let's start with the Nifty on July 3, 1990 which is the first date for which the NSE offers a back-calculated value of 279 points. (The Nifty index was actually launched in late 1994).
Starting July 1990, using daily data, we can calculate rolling returns for periods such as 1, 2, 3, 5, 7 and 10 years. Intuitively, one expects longer time periods to generate higher CAR. But the CAR of 1,2,3,5,7 and 10 years respectively are 11.6 per cent, 10.3 per cent, 10.7 per cent, 12.7 per cent, 12.3 per cent, and 10.8 per cent. There is no apparent explanation for the bulge at the 5 year and 7 year marks. Perhaps this is due to inherent economic cyclicality or there are similar bulges in interest rate yield curves.
The rolling returns may be risk-adjusted several ways. One is to just estimate the percentage of losing returns in each time-frame. On the now familiar ascending time-scale of 1,2,3,5,7,10 years, the percentage of losing returns is (all figures in per cent) 41, 41, 26, 14, 9, and 4 respectively. This validates buy and hold – the chances of losing returns drop significantly over longer time frames.
Another way to assess risk is to look at standard deviation of returns. Again, in ascending order of periods, this is (all figures in per cent) 34, 14, 17, 14, 12, 9. There is still an anomaly in 3 year and 5 year bulges. Nevertheless, standard deviation also does reduce as time period increases.
Despite the anomalies, the certainty of returns seems higher at the 10-year period where 96 per cent of readings offer positive returns over 2550 observations. In fact, the 10-year period is also the only one where the CAR is higher than the standard deviation.
A good deal more data basing and creative thought would be required to understand the anomalies. The results suggest two possible time-zones for long-term passive investors. If you don't mind some volatility, the risk: reward equation is reasonable for a holding period of 5 years. If you want to be extra safe, focus on 10-year holding periods.
There's no guarantee that future returns will resemble the past. But this statistical exercise does cover 20 years of turbulent economic history so it's a reasonable benchmark. The 20-year CAGR at 16 per cent is much higher than any of the rolling returns. But even the best index funds will have large tracking errors over 20 years since they must rebalance every time index stocks are replaced.
In practice, you should expect to make somewhere between 10-12 per cent on a passive buy-and-hold. Even that entails a continuous SIP taken over longer periods than most investors do. One last point could swing the focus in favour of 5 years over 10 years. The premium of equity returns over risk-free returns (interest yields) is on the order of 3 per cent at the 5-year mark, which is substantial but not extraordinary. The equity premium drops as the holding period approaches the 10-year mark. This, of course, excludes tax considerations.
