Riding the Momentum
Board: Berkshire Hathaway
The reason is, if they go up, then what? Are you supposed to sell them, even though holding them for 5 years longer will probably be immeasurably more profitable? There is no decision-making based on if they go up or down subsequently.
It depends on what kind of an investor you are. One approach is to pick a great stock, wait for it to be on sale, then wait even longer for it to bottom out and then start outperforming its value peers at least a little bit on at least some time frame, then buy it. At that point the odds of its continuing to fall in an unpleasant way are somewhat lower than if you buy it the first day it's cheap enough. Nobody buys right at the bottom, so one might argue that it's more possible to find an entry date which seems to be shortly after the bottom than it is to find an entry date which is shortly before the bottom. Then just keep it as you would any other deep-value-long-hold stock. In this approach you don't really need a sell condition, unless you think it's a good idea to bail out when something gets overvalued.
Another approach is a pure value momentum portfolio. You buy it when it's the best performing cheap-enough stock, and you sell it when it is no longer the best performing cheap-enough stock. Pick a number of stocks to hold and stick with it, which sets your sell rule. This is philosophically a very different way of investing. It has some great advantages and some great disadvantages. Not good if you pay a lot of capital gains tax and high commissions. But the sell rule is simple: you sell something when it's no longer top rated among its peers by both value and momentum. Most value stocks, most of the time, start underperforming their value peers in a fairly gradual way, not suddenly, so there is usually plenty of time to edge towards the exit before your gains are gone.
e.g., say you simply bought the 5 stocks in Berkshire's portfolio which have gone up the most in price recently. (as an example, take the average price in the last month, compare it to the average price between 6 and 7 months ago). Since 1985, this would have returned about 26%/year after allowing for trading costs, beating the average of the returns of the same set of stocks in the same period by about 7% a year. Just the top 3 by momentum is about 4%/year better still. (this obviously requires knowing in the past what Berkshire owns today, so the 26% has no meaning per se---the important thing is that it does better than a random or complete selection from the same set of stocks) From a value investing point of view this makes no sense, since you never buy anything at its cheapest, but it adds value if your trading costs and capital gains tax are low.
The big disadvantage of momentum investing is that it puts you into the wrong things for the period right after any big market reversal. For example, the things that did best from mid-2008 to Feb 2009 were those that were the most solid, but that was a bad guide to the rest of 2009 when the worst quality stuff did best. Naive momentum did horribly in Q2-Q4 of 2009 for this reason. And, unlike long term value investing, you're changing stocks frequently, so you don't just sit it out and wait for a rebound---those losses are fully realized. So, it doesn't work all the time, but the benefits are perhaps good enough most of the time that it more than makes up for the occasional huge reversals. It's the dark side of the force, for sure, but it doesn't work any smaller fraction of the time than does deep value long hold investing. Nothing works all the time. More than anything, it comes down to tax treatment. If deferred capital gains taxes give you a big benefit for long term holds, then that's where you should look first. If Berkshire didn't have that issue, you can bet your bottom dollar that Mr Buffett would have done a lot more position changes. He'd have sold Coke in 1999 and bought bonds. Deep-value-long-hold investing is, more than anything else, a very successful tax strategy.