Thursday, August 23, 2012

Conventional Wisdom: The Falling Knife

I've talked about the opportunities that falling knives (aka large stock price drops) can create for value investors here. But I had no empirical basis for such statements. I was only making the case that buying a company at a dramatically lower price makes for smarter investing than paying the previous price; in other words, I was ignoring the price action itself. But it turns out the price action appears to be beneficial when it comes to falling knives!

Thanks to the site csinvesting, I recently came across a paper which analyzed the returns of "falling knives" both in US markets and around the world. The findings are unsurprising: stocks that fall big tend to outperform the market in subsequent periods. Interestingly, falling knives outperformed the market for each of the next three years, with the widest outperformances occurring in years 2 and 3; patience is key!

Despite this and other evidence (including mean regression), most traders and the media will continue quote conventional wisdom in warning you against buying falling knives. This is likely because they are focused on how buying a falling stock makes them feel, rather than employing a rational investing approach which focuses on buying when prices are lower than value.

As such, investors should ignore the conventional wisdom. The data suggests falling knives actually outperform.


balans said...

How do you reconcile this with James P. O'Shaughnessy: "What Works on Wall Street"
here is a paste from page 230 where he describes the strategy of buying last years losers:

If you’re looking for a great way to underperform the market, look no further. A $10,000 investment on December 31, 1951, in those 50 stocks from the All Stocks universe having the worst one-year price performance, was worth just $79,226 at the end of 2003, a compound return of 4.06 percent a year. Only those 50 stocks having the highest price-to-sales ratios did worse.
The standard deviation of return for the 50 losers was 32.63 percent, considerably higher than All Stocks’ 20.11 percent. With such abysmal returns, any
risk will wreak havoc with the Sharpe ratio, and here it’s a pathetic 10. Base rates are atrocious, with the 50 losers beating All Stocks in only 14 of the 52 years reviewed. The rolling five-year returns are even worse. The 50 losers beat All Stocks only twice in 48 five-year periods. And the magnitude of all rolling five-year losses was huge as well, with those 50 stocks having the worst performance on average almost 10 percent behind the average return
for All Stocks in any five-year period. But the booby prize goes to the 10 year returns, in which the losers never beat the All Stocks universe. The magnitude of the underperformance is staggering—over all 10-year periods, the
50 biggest losers had an average annual compound return that was 10 percent less than All Stocks. Tables 15-7, 15-8, and 15-9 detail the grim news.

Saj Karsan said...

Good question. I wonder if the difference is due to measuring differences or changes in market behaviour since the studies are of different periods.