Relative Valuation (Part 3)

Relative valuation analysis is the most commonly used technique for which we try and deduce value. From real estate practitioners, to market analysts, sporting pundits, and grocery shoppers, its almost second nature to infer value from a set of comparable items. For example, if United beat Chelsea (let's say 2-1) and Chelsea beat Tottenham, then in mathematic terms (transitive relation) United must be better than Tottenham (and of course Chelsea :). If three different tuna brands sell between $1 and $1.20, then if we are looking at a fourth brand that is of similar quality and quantity then we would not expect to pay anything outside of this range. At one stage or another in our lives we have all used relative valuation to guide our investment/purchasing decisions.

The Easy Part

Relative valuation allows us to form an assessment about the value of an asset from a market of similar or comparable assets. In lectures, I typically spend no more than 5 minutes discussing the steps involved in relative valuation analysis because to do otherwise would be to insult your intelligence. The same applies in this forum. To perform relative firm/asset valuation, the following steps are typically followed:

1. Identify comparable firms and obtain market values for those firms.
2. Convert these market values into standardised values. (i.e. instead of comparing price, compare Price / Earnings; instead of comparing EV, compare EV / EBITDA)
3. Compare these standardised values against a series of comparable firms.
4. Evaluate using personal judgement whether the firm is over/under-valued.

The guiding principle in this type of analysis is arbitrage. What some like to term, the law of one price, suggests that two assets that have similar underlying characteristics should trade at a similar price. In fixed-income securities and currency trading this is a common type of strategy and for the former, because contracts have a limited life, convergence is ultimately a certainty. We must remember though, in equities, two assets can be similar in many respects and trade at substantially different premiums - remember the old coinage: the market can remain irrational longer than you can stay solvent! In light of this information, relative valuation seems a curious choice in investment decision analysis. Then again, many people still favour technical analysis, but this argument to be had on another day.

What's the difference between Intrinsic Valuation and Relative Valuation?

You will remember when I examined intrinsic valuation analysis that one of the biggest difficulties I spoke of was setting assumptions for the forecast horizon and terminal period. When most people first come across relative valuation, they believe for some reason that they are free from these assumptions in the same way politicians are free to skirt around the truth. A common phrase used by analysts goes something like this - "this is a fair valuation at 6*EBIT". I've never really known what this means (and i will discuss how flaky this comment is later), but because this is what the market is currently pricing the asset at, your position is what is typically referred to as defensible (this word is part of every investment bankers lexicon). We are all clever enough here to know, however, that just because our stock is trading at a particular price, that it doesn't necessarily represent fair value, or even a meaningful trading opportunity.

My point here is a very simple one. Relative valuation is based on assumptions about growth, risk and return on capital in the same way intrinsic valuation is. The difference is that they are explicitly made in an intrinsic valuation analysis, whereas with relative analysis they form part of the ratio that you select for your analysis. If you want to undertake relative valuation analysis properly, then there are two things you need to do:

1. Forget about rules of thumb (EV/EBITDA < 6 = cheap, P/E < 12 = cheap, P/Book < 1 = cheap).
2. Examine the drivers of the ratio in question and compare these drivers to other comparable assets.

The Ratios

There are literally hundreds of ratios that can be used to compute market values for an asset. To list a few:

1.Earnings ratios (P/E ratio, PEG ratio, EV/EBITDA, EV/EBIT),
2.Book ratios (P/Book, EV/Invested Capital),
3.Revenue ratios (EV/Sales, P/Sales), or even
4.Industry ratios (EV/Price of Ounce of Gold, P/No. of Customers)

Every year this list will grow, and every year particular ratios will move in and out of favour which begs the question, which multiple or group of multiples is relevant for our analysis? A good question, that we can only really answer after we have discussed the strengths and limitations of multiple analysis in more detail.

A Platform for Success (or at least an avenue to avoid failure)!

Every multiple/relative valuation analysis that is conducted should check off on the following things:

1. That the multiple is properly defined. Ratios should mean something and be consistently defined. Remember that this analysis is about calculating value. Therefore, if you are deciding on whether to use the ratio Price / No. of Users, then your first question is will no. of users translate into value for this company. Just because I have 10,000 users visit my site doesn't mean that it is going to (or will be able to) generate value for me. Being properly defined also means that the numerator and denominator of the ratio are defined in a consistent fashion. Therefore if we are dealing with equity cash flows in the numerator, then the denominator needs to represent a similar claim.

2. The multiple is appropriately described. Finding a comparable firm should be more involved than simply looking up the GICS industry code and picking a range of firms in the same industry. Ultimately selecting a comparable firm is a trade-off between similar companies and having enough data points to be able to find an "average".

3. The multiple is appropriately analysed. As I mentioned earlier in this post, ratios are driven by particular factors. It is imperative that we have some understanding of these factors if we are going to make buy/sell decisions.

There are two basic groups of problems with relative valuation analysis. The first group of problems can be summarised as economic problems and they include the fact that relative valuation subscribes to arbitrage principles and that it implicitly makes assumptions about risk and return. Although these problems are widely acknowledged the second class of problems are not. These are the statistical problems and the issue with not understanding these problems is that the conclusions you infer from the data can often be misleading.

I will discuss these problems as well expand on my post of how to undertake relative valuation in a more coherent and purposeful manner in a subsequent post.