It is a truism that the long-term investor should always seek to buy at lower valuations. A downturn allows for reduced average cost of acquisition and, therefore, helps generate higher long-term returns. But, when we actually look at concrete numbers, it is astounding to realise how much of a difference lower valuations make.
The nominal value of the index does not seem to matter; it is the valuation which is crucial. Taking Nifty data over the past 10 years (since January 2005), we can calculate a three-year compounded return, session by session, and compare returns sorted on the price-to-earnings (PE) ratio of the index. The lower the valuation, the higher the return.
In the 10-year period, the Nifty has run through 2,700 sessions, and there is a three-year return available for about 1,900 of those sessions. The PE (past four quarters, as released by the National Stock Exchange) has moved between a maximum of 28.3 and a minimum of 10.7.
The frequency distribution conforms closely to the normal distribution with 69 per cent of observations lying within one standard deviation (3.04) of the average (19.3) and 94.9 per cent of observations within two standard deviations of the average. So, the distribution of index valuation by PE is reasonably predictable.
At the very least, an investor would hope to beat inflation (which averaged about six per cent in terms of the wholesale price index across this period) and would want to beat the safe return from long-term fixed deposits (FDs).
An investor who bought the index when it was trading at 10-16 PE received a fantastic averaged compounded annual growth rate (CAGR) of 24.4 per cent. The highest CAGR in this valuation bucket of 10-16 PE was 48 per cent. There was only one negative three-year return in the 347 sessions when the index was valued within those limits. Of course, the index traded into this zone of low valuations only 18 per cent of the entire time.
Investors who bought the index when it was trending between 16 and 19.5 PE (roughly within one standard deviation below the average) received an averaged CAGR 11.9 per cent. The highest return was 23.5 per cent and there were 56 sessions of negative return. That is, 9.5 per cent of all sessions in the range of 16-19.5 PE returned negative three-year returns. However, the overall average CAGR was quite good.
Investors who bought between 19.5 and 22.5 (roughly within one standard deviation above average) received averaged returns of only 6.5 per cent CAGR with a maximum CAGR of 15.5 per cent. This is no better than parking money in a long-term FD. In fact, the FD might do better.
Investors who bought between 22.5 and 25.5 received an averaged CAGR of only 2.1 per cent and negative returns in 12 per cent of sessions, while those who bought above PE 25.5 received averaged negative returns (minus 0.4 per cent) and suffered 46 per cent of negative sessions. The index ranged above PE 22.5 only 14 per cent of the time.
A systematic investor is by definition buying through that entire range of valuations. The blended three-year CAGR across that entire PE range of 10-28 is 10.7 per cent. That's a respectable return and it beats inflation by a comfortable margin. A five-year return shows similar patterns; the lower the valuation at acquisition, the better the long-term return. Obviously, this is not 'realistic' since a systematic investor will buy only once a month if he's using a standard plan.
But, the pattern indicates long-term returns can be improved by a systematic investor without relying on much discretion. Most of the time (two-thirds) the index will be between 16 and 22.5 PE and the investor should only maintain basic commitments without thinking.
The systematic investor could consider increasing commitments whenever the index is below the long-term average. He should definitely increase commitment if it falls below the average minus one standard deviation - that is, below PE 16. On the upside, consider cutting commitments if valuation is above the average and definitely minimise commitments above PE 23, when the index is more than one standard deviation above average.
It is possible to create a mechanical rule of this nature and put into practice without much trouble. Over a period of time, such a method should add considerably to the effectiveness of asset allocations. By the way, the data suggest market valuations were too rich (above PE 23) from mid-February 2015, and remained too rich till mid-August 2015. Even now, after correction, the index is trading close to the top end of the acceptable 16-22.5 range.
