Valuing Stocks: Multiple-Based Valuation
In last week's Discounted Free Cash flow article, I discussed the traditional and "textbook" method for valuing a stock, along with a few modifications to smooth out the inherent bumpiness in levels of cash flow. In this week's article, we'll take a look at another common way of valuing a stock, using statistical multiples of a company's financial metrics, such as earnings, net assets, and sales.
There are basically three statistical multiples that can be used in this kind of analysis: the price-to-sales (P/S) ratio, price-to-book (P/B) ratio, and price-to-earnings (P/E) ratio. All of them are used the same way in doing a valuation, so let's first describe the method and then discuss a bit about when to use the three different multiples, then go through an example.
The Multiple-Based Method
Valuing a stock in a multiple-based manner is simple to understand, but takes some work to get the parameters. In a nutshell, the object here is to come up with a reasonable "target multiple" that you believe the stock should reasonably trade at, given growth prospects, competitive position, and so forth. To come up with this "target multiple", there are a few things you should consider:
1) What is the stock's average historical multiple (P/E ratio, P/S ratio, etc.)? You should at least take a 5-year period, and preferably 10 years. This gives you an idea of the multiple in both bull and bear markets.
2) What are average multiples for competitors? How wide is the variance against the stock being investigated, and why?
3) Is the range of high and low values very wide, or very narrow?
4) What are the future prospects for the stock? If they are better than in the past, the "target multiple" could be set higher than historical norms. If they are not as good, the "target multiple" should be lower (sometimes substantially lower). Don't forget to consider potential competition when thinking about future prospects!
Once you have come up with a reasonable "target multiple", the rest is fairly easy. First, take current year estimates for revenue and/or earnings and multiply the target multiple against them to get a target market capitalization. Then you divide that by the share count, optionally adjusting it for dilution based on past trends and any announced stock buyback programs. This gives you a "reasonable price" valuation, from which you want to buy 20% or more under for a margin of safety.
If this is confusing, the example later in the article should help clear things up.
When to Use the Different Multiples
Each of the different multiples has their advantage in certain situations:
P/E ratio: The P/E is probably the most common multiple to use. However, I would adjust this to be the price-to-operating earnings ratio instead, where operating earnings in this case is defined as earnings before interest and taxes (EBIT - include depreciation and amortization). The reason for this is to smooth out one-time events that skew the bottom line earnings per share value from time to time. P/EBIT works well for profitable companies with relatively stable levels of sales and margins. It does *not* work at all for unprofitable companies, and does not work well for asset-based firms (banks, insurance companies) or heavy cyclicals.
P/B ratio: The price-to-book ratio is most useful for asset-based firms, particularly banks and insurance companies. Earnings are often unpredictable due to interest spreads and are full of more assumptions than basic product and service firms when you consider such nebulous accounting items as loan loss provisions. However, assets such as deposits and loans are relatively stable (2008-09 aside), and so book value is generally what they are valued on. On the other hand, book value doesn't mean much for "new economy" businesses like software and service firms, where the primary assets are the collective intellect of employees.
P/S ratio: Price-to-sales is a useful ratio across the board, but probably most valuable for valuing currently unprofitable companies. These firms have no earnings from which to use P/E, but comparing P/S ratio against historical norms and competitors could help give an idea of a reasonable price for the stock.
A Simple Example
To illustrate, let's stick with the same stock as in the last article: Lockheed Martin (LMT).
From doing some basic research, we know that Lockheed Martin is a established firm with an excellent competitive position in what has been a relatively stable industry, defense contracting. Furthermore, Lockheed has a long track record of profitability. We also know the firm is obviously not an asset-based business, so we'll go with the P/EBIT ratio.
Looking over the past 5 years of price and earnings data (which takes some spreadsheet work), I determine that Lockheed's average P/EBIT ratio over that period has been about 9.3. Now I consider the circumstances over the past 5 years and see that Lockheed has worked through some strong defense demand years in 2006 and 2007, followed by some significant political shakeups and a down market in 2008 and 2009, followed by a market rebound but problems with the important F-35 program early this year. Given the expected slow near-term growth of defense department spending, I conservatively theorize that 8.8 is probably a reasonable "target multiple" to use for this stock in the near term.
Once this multiple is determined, finding the reasonable price is pretty easy:
2010 revenue estimate is $46.95 billion, which would be a 4% increase from 2009. Earnings per share estimate is 7.27, which would be a 6.5% decline from 2009, and represents a 6% net margin. From these figures and empirical data, I estimate a 2010 EBIT of $4.46 billion (9.5% operating margin).
Now, I simply apply my 8.8 multiple to $4.6 billion to get a target market cap of $40.5 billion.
Lastly, we need to divide that by shares outstanding to get a target share price. Lockheed currently has 381.9 million shares outstanding, but usually buys back 2-5% a year. I'll split the difference on this and assume share count will decline 2.5% this year, leaving an end-of-year count of 379.18 million.
Dividing $40.5 billion by 378.18 million gives me a target share price of about $107. Interestingly, this is close to the discounted free cash flow valuation of $109. So, in both cases, I've used reasonable estimates and determined that the stock looks undervalued. Using my 20% minimum "margin of safety", I would only consider buying Lockheed at share prices of $85 and under.
Wrapping It Up
Obviously, you can easily plug the price-to-sales or price-to-book ratio in and, using the proper financial values, do a similar multiple-based valuation. This kind of stock valuation makes a bit more sense to most people, and accounts for market-based factors like the differing multiple ranges for different industries. However, one must be careful and consider how the future may differ from the past when estimating a "target multiple". Use your head and try to avoid using multiples that are significantly higher than historical market averages.