VALUE INVESTING
Odds On

By Stephen Bland (TMFPyad)
April 14, 2000

Something caught my eye over the last few days on the Qualiport board. Not a place where I would normally post, it's not my turf, but I read the messages from time to time. The style used, net present value or NPV, based around the supposed ideas of Warren Buffett I believe, does not appeal to me at all. I can see in theory why it is attractive, but in practice it requires the forecasting of earnings per share (EPS) many years ahead and I personally am distinctly unhappy with that, believing that this is impossible to any meaningful degree. I suspect that Mr Buffett does not actually do this at all but simply has a gut feel for a good share: something that can never be analysed and sold in books. You're either born with it or you're not, like any outstanding talent, artistic, sporting, business and so on.

I have enough difficulty with forcing myself to accept just one year's forecast EPS with my own approach, but several years ahead, to me anyway, is not something upon which I would be prepared to chance serious money. I would rather that, even with my style, I could afford to dispense with forecasts and rely on pure actuals. The whole value concept is constructed round fundamental ratios and facts. But I have learned that without rising EPS, the cheap share often stays just that, a cheap share. There has to be something to generate the re-rating that value players seek, and that something is a strongly rising EPS forecast. So I am condemned to living with that, regrettably, but just for one year. And even that one lousy year ahead can go wrong sometimes.

One of our regular readers was commenting upon the latest share bought for the Qualiport and referred to value investors, contrasting these with growth investors. He was referring to fundamental value investors, not the NPV style of the Qualiport, which is sometimes also, confusingly, described as a value approach.

I responded that I saw NPV investing as a rather long odds game. There is only a given chance of estimating future EPS to any particular degree of accuracy. Getting several years' figures correct within whatever tolerance you set becomes an increasingly longer shot the further out you project. If, say, you allow yourself to guess within 80% of each year's future EPS for your NPV model, then several years together becomes a progressively weaker chance of getting them all correct within your margin of error.

If each year's EPS is independent of any other, then the cumulative chance is the product of each year's chance. So at 80% accuracy for ten years you have a chance of only (0.8)^10 of getting all ten years right to within 80% each year. This works out at only around 11%. For every hundred attempts you are likely to get only 11 correct in forecasting each year's EPS of a company for ten years within 80%. Bad odds.

I am being a bit unkind here, because the above may be inaccurate if in fact there is some degree of a relationship between one year's EPS and the next. For example with a coin, if you throw twenty heads in a row, the probability of the next toss being another head is unaltered at 1/2, by what went before. However with company EPS, there is probably some correlation on average between successive years. And if that is so, you cannot simply use the above arithmetic to determine the overall chance of getting them all correct within the tolerance required. Whatever the actual relationship may be though, I think there is little argument that the longer ahead you try to forecast EPS, the less reliable it has to be and that was my point when I contrasted the styles of value and growth that our reader raised.

The more forecast figures there are in the investment selection, the greater the risk the investor has to take on board in buying it. I don't see how you can get away from that. And to me the converse is equally true. The less forecasting involved, the lower the risk of something going wrong.

One of my most basic concepts of all, how I started developing my approach in fact, is that to make money you have first of all to find ways of trying to avoid losing it. Shortening the odds in your favour, minimising the downside if you like. Asking first, not how much could I make on a share, but how much could I lose? The first step to making money is not losing it. No guarantees, of course, everyone has the occasional loss. Investing is about odds, likelihoods, not certainties. I try to turn these in my favour. I have to because I frequently risk all on one share. UnFoolish I know but there you go. Consequently I have no interest in seeking situations which lengthen the odds against me. Investing in individual shares is a difficult enough game from which to make money, without looking for ways to make it even harder.

Despite many adverse comments on my style, I will not weaken the approach in order to find more shares from which to choose. This leads to the boring situation where I can be in cash for long periods because I don't find the perfect share. Sometimes it's marginal whether a share qualifies or not. I don't want marginal, I want the whole works. If it ain't perfect, in my opinion, I don't want to know.

It can be tempting to weaken sometimes. Many shares that failed my tests, even just on smell despite all the numbers adding up, have gone on to perform outstandingly. So be it. I missed them. You can't win 'em all, and nor do I want to. I want just one kill at a time, I want it to be big and I'm always going to go to the utmost lengths to make sure it's the right one. And even with all my precautions I'll still make the occasional blunder.

Apart from my extreme caution on stock picking, one popular criticism has been that I get out too early on a strongly rising share. This is probably true. A good share will often go on performing long past the point where the value, as I see it, has gone out of it. Again, so be it, I might miss a lot of it. It's all about risk management, about not going out further than the point with which I feel comfortable. And my comfort level is well below other investors in the main. But by the same token, I will often be in a share very early on, because of the extreme contrarian nature of my approach. So I'm in early and out early, chopping out for myself just a particular part of the growth curve of the share. The low risk bit, as I see it. The next guy can have the rest.

In one sense the labels value and growth mean little. I mean, all investors are growth-seekers, obviously. But they do in fact mean something in helping to identify a person's approach. It is a question of the emphasis placed upon certain criteria. Those that look for strong and cheap fundamentals are typically value, those that are less concerned with such matters and look for rapidly expanding EPS over several years, regardless perhaps of yield or assets, would be growth. In between there are often mixtures of the two. Even at my end, the very deep value area, I insist on one growth criterion, rising EPS over a year. But it seems to me that growth players are more willing to entrust themselves to the future, and that is unknown territory which scares me. And yet undoubtedly there are people who can do this successfully.

Like a lot of styles which appear superficially to be based on numerical analysis and therefore capable of reproduction by others, the real skill may well not be all in the figures anyway, even if it appears that way. Perhaps smell, gut feel, intuition, whatever you call it, is more important than it may appear on the face of it. Which brings me round in a circular fashion to my comments on Warren Buffet in the opening paragraph.

Related Links

• Value Investing discussion board
• James Carlisle (TMFJimmyC) on risk