Ask Colin

Am I correct in using Sector average PE ratios?

The detailed question was:

When evaluating a prospective stock’s PE ratio you mention in the book that you compare it to the ‘market PE’. In the few stocks that I have bought to date I have actually been comparing their PE ratio to their sector PE ratio as published in the AFR on Saturdays.  There is quite a difference between the All Ordinaries PE for the week and say Info Tech PE for the week. Am I right in doing this?

There are few absolute “right” and “wrong” answers in investing. In fact I try to avoid using these terms. This is partly because it all depends on what you are trying to do and for what purpose. It may also be that a particular method is more suitable in one situation than another.

The first thing to be careful of in this area is bogus precision.

Price earnings ratios are usually shown to two decimal places. If the ratio is based on historical earnings, then this is mathematically correct. However, the price of the stock fluctuates every day and with it the price earnings ratio. If the price earnings ratio is 12.36 one day, 12.11 the next day and 12.28 the following day, you will see that the precision is unwarranted. It is much better to round the ratio to the nearest whole number of 12. I usually do this mentally, but now that your question has reminded me, I should also start showing the rounded numbers in my stock investment journals.

Of course, if we were using an estimated price earnings ratio, then the precision is totally unwarranted, because the earnings number is a forecast and should never be seen as exact. I never use estimated price earnings ratios because I understand how bad we can be in forecasting earnings, but if I did use them, I would round the resulting ratio to a whole number.

I mention this here as a matter of general education, but also because sometimes the differences between sectors can be in the decimal places. However, you mention a greater difference than that, so I will go on.

The second thing to be careful of in this area is the use of averages.

Averages are dangerous things if we do not know what the composition of the average is. If the companies in a sector are quite different in the nature of their business (the Materials sector would be a good example), then the sector average is not going to tell us anything very useful and could lead us badly astray. So, before using any sector average, we should know exactly which companies make up the sector data and how comparable their business is to the company we are assessing by comparison to the sector average.

There is another important aspect to sector averages. The Australian stock market is not very large. It is divided into a relatively large number of sectors. As I understand it, the sector averages are capitalisation weighted and some of these sectors are dominated by one or two giant companies. The result is that the sector average is really just the price earnings ratio for the dominant company. Telecommunications may be such a sector.

There are also many situations where the company we may be assessing does not really have a business that is like the companies that are in its sector. The majority of the companies in the sector may be fairly similar, but there may also be exceptions, one of which may be the company we are assessing.

I have found that there are two more or less satisfactory approaches to this problem:

  1. The best solution is to compare the price earnings ratio of the company we are assessing to the price earnings ratio of a parcel of very similar listed companies.If the company we are assessing is a clothing retailer, we should compare it to other clothing retailers with a similar business, not to the retail sector which includes the giant grocers and large department stores. This is the best approach of the two, but it does mean a great deal of effort and skill. There are also some companies where there is no similar listed company.

  2. This brings us to the second solution, which is to use the market average price earnings ratio. Because the number of companies in it is relatively large (about 500), it does modify some of the distortions mentioned above. However, it is not perfect. It is simply a starting point from which we should go further along the lines of the first solution if necessary.

    I use the market average price earnings ratio as a crude measure of the relative valuation of the market. So, if the market average price earnings ratio is 10, I know that the market is undervalued and a stock with a price earnings ratio of that same number is therefore also likely to be on the side of being undervalued. However, if the market average price earnings ratio is 20, then I know that the market is overvalued and I should be very wary of companies with a price earnings ratio near or above that figure.

From this discussion I hope to have shown that we do many things in investing that are heuristics (rules of thumb) that have been shown to work well in a general way as a first cut when we are analysing companies. In many situations, there are so many variables and so many unknowns that the general heuristics serve us better than if we tried to be too precise, especially when that precision is based on forecasting the future.

Above all, I hope to have shown that we should never stop thinking about what it is we are doing and trying to fashion our methods so that they are as bullet-proof as possible. Precision is often not the way to achieve greater certainty. Investing is about managing uncertainty.