Buy when the market's fallen 20pc? First do some checks

For a couple of reasons, the 20pc rule hasn’t caught on. First, because it is too simple – if investing were really that simple then surely everyone would do it. Second, because it’s so easy to think of times when the rule would have let you down – what good is an adage that would have burnt you so spectacularly in the dotcom bust and during the financial crisis? In both of those cases, a 20pc fall was not a buy signal but a warning of worse to come.

I, too, was sceptical at first but intrigued enough to crunch the numbers. My conclusion is that there’s more to the advice than appears at first sight for three reasons. First, buying after a 20pc fall makes sense surprisingly often; second, it is interesting how often the market changes direction at close to a 20pc decline – this does seem to be a psychologically important level for stock market corrections; third, the differences between the circumstances in which the rule works and those in which it doesn’t are pretty clear. In other words you will probably know when to use the rule and when to ignore it.

The first thing I did was to look at 32 years of daily closing prices for the FTSE 100 since the index was launched in 1984. During that period there have been 13 occasions when the market has fallen meaningfully enough to prompt the question: is this a significant buying opportunity?

The occasions fall into two camps, pauses for breath in an ongoing bull market and clearly identifiable phases of bigger downturns.

During the late 1980s and throughout the 1990s there were five occasions when the FTSE 100 fell by 20pc or came within a whisker of doing so. The first is the most famous, the 1987 crash when the bear market trigger was reached two weeks after the cyclical high and the whole peak to trough journey took only a little over a month. The subsequent recovery from the bear market buy signal to exceeding the previous high took 10 months.

"Anyone who bought at the first bear market trigger in September 2008 would have had to wait four-and-a-half years to recover the previous cyclical high."

Tom Stevenson

There were four similar episodes before the end of the bull market in 1999. Between January and September 1990, the FTSE 100 fell by 19.2pc before regaining the previous high in just six months. In 1992, the index fell by 17pc and then reached the previous peak in three months. In 1994, when the Federal Reserve spooked investors with an unexpectedly aggressive rate hike cycle, the decline was 18.3pc with recovery taking a slightly more pedestrian 14 months. In 1998, the Russian default crisis saw the FTSE 100 slide by 22pc in six months, but the previous peak was reached after only three months.

• Explainer: How the UK entered the bear market trap
• FTSE 100 bear market: what does it mean for me?

All of these corrections were sentiment-driven corrections which took place against an otherwise relatively benign economic backdrop. But what about the times when a 20pc fall was not a signal to buy on the dip but a harbinger of worse to come? The biggest cluster of these was during the fallout from the technology bubble between 2000 and 2003. Then, on three separate occasions, a 20pc fall was a false signal and investors took years to recover their money – in the case of the 15-month slide between 1999 and 2001, the recovery took another 14 years. Five years on, anyone who bought at the first bear market trigger in September 2008 would have had to wait four-and-a-half years to recover the previous cyclical high.

Two things make these periods different from the bull market corrections. In the dotcom episode, the red flag was excessive valuation; in the financial crisis it was real economic dislocation. In both cases shares were not a buy again until ridiculously cheap valuations and widespread investor despair triggered a turn in the market.

So my conclusion is this: if peak valuations were stretched before a correction or there are significant economic dislocations you should tread carefully; if, however, valuations at the previous peak were reasonable and the real economy is doing OK, then a 20pc slide in share prices is a reliably good buy signal. The latest fall feels to me like one of these.

Tom Stevenson is an investment director at Fidelity International. The views expressed are his own. He tweets at @tomstevenson63.