Dollars and Jens
Sunday, April 04, 2004
Introduction to Valuation
There are two broad types of methods used to value shares of a company. I will refer to these as the "comparable" type and the "discounting" type.
Comparables are generally simpler and less precise. The general idea is to look at how similar companies are valued by the market, and figure out how the target company compares to them. The comparisons are usually based on simple ratios: price to earnings (P/E) is the most popular, but you can also look at price to sales, price to book, price to free-cash-flow, or various ratios involving "enterprise value" -- the market value of all of the shares outstanding, plus the value of the firm's debt, minus free cash -- rather than simply share price.
You might notice that most of the firms in the industry in which your target company operates have P/E ratios of 13-18, with the better companies having higher ratios and the worse companies getting lower ratios. If your target company seems to be better than most in the industry -- maybe it has better growth prospects and less risk -- you might argue that it should be worth 16-19 times earnings. Then you go have lunch.
Discounting methods rely on two fundamental concepts:
- A share of stock is a claim on the future cash generated by the company, and should be valued as such.
- Cash today is worth more than the same amount of cash tomorrow.
But what "cash flows" does one use? One can use dividends, or one can use free cash flow (i.e., cash generated by operations, minus cash used in investments and financing), or one can use earnings on the grounds that they represent a less-lumpy estimate of free cash flow. Or -- my personal favorite -- one can use "residual earnings", which I'll explain in a moment. As it happens, for any company that doesn't sit on its cash -- i.e., it either pays any excess cash back to shareholders or invests it -- all of these cash flows provide the same result1. The main factor in deciding which cash flows to use is which cash flows one considers easiest to estimate.
As I said, I like the residual earnings model. A company's residual earnings for a period are the earnings minus a charge for equity employed. For example, a company with a 12% cost of equity and a book value of one million dollars has to earn $120,000 per year to justify retaining its equity, rather than paying it out to shareholders. If it earns $150,000 in a year, the retained earnings for that year are $30,000. This is the amount of value added by the company not having sold off its parts at the beginning of the year with the proceeds invested elsewhere. The value of a company, then, is the total of its current book value and the present values of its residual earnings.
I like the residual earnings model for both practical reasons and psychological reasons. By "psychological reasons" I mean that it pushes the modeler to think properly about costs of capital, and sustainable competitive advantage. Most companies are worth more alive than dead -- to a first approximation, worth more than book value -- but one can't understand the company if one doesn't consider whether this is so and, if it is, why it's so. If the company has assets of $5 billion and a market value of $10 billion, why is it worth more than a start-up company with $5 billion in cash? Does it have a strong brand name (like Coke), intellectual property (like Pfizer, with its drug patents), a "network effect" (like EBay, which is more attractive to a seller every time a new buyer signs up, and vice versa), a product with replacement parts (like Gillette, which distributes razors that only work with Gillette blades), or at least a good customer list or stable top-notch management? How durable are its advantages?
But I also like residual income for practical reasons. As an industry matures, companies in that industry tend toward a state of competition -- in other words, residual income approaches zero. Free cash flows or dividends never have to stop, and the farther out you estimate them, the wronger you get. Like any model, garbage in will lead to garbage out, but the residual earnings model makes it easier to limit garbage than most models do.
If you want to read more, my brother has written up something related to, but different from, this entry.
1. Technically, one also has to assume "clean accounting". This doesn't mean a lack of fraud; this means that the retained earnings line on the balance sheet changes exactly by the reported earnings in a period minus dividends for that period. There are a few "comprehensive income" figures which don't follow this rule, but I feel comfortable ignoring them for now.