Portfolio Rule Two

For those that have e-mailed me about equity management, I will get back to you soon; I have been tied up in details of getting my assets management business going recently.

=-=-=–==–=-==–==-=-=-=-=-=-=-

If you have read me for a while, you have heard of my eight rules of stock investing.  Recently some people have been e-mailing me regarding my plans to start Aleph Investments, and one asked if I could write a series of pieces to explain how I manage stocks.  I thought it was a good idea, so this is the second of what is likely to be eight episodes.  Here’s portfolio rule two:

Purchase equities that are cheap relative to other names in the industry. Depending on the industry, this can mean low P/E, low P/B, low P/S, low P/CFO, low P/FCF, or low EV/EBITDA.

The first principle of value investing is having an adequate margin of safety.  Make sure that your downside is clipped, if things go wrong.  But that’s not today’s topic, I will cover it later in this series.

The second principle of value investing is buying the investment cheap to its intrinsic value.  That is an easy phrase to utter, but complicated to implement.  There is no one single simple metric that serves as a guideline for evaluating cheapness as a value investor.

Ideally, one would create an integrated free cash flow and cost of capital model, one like Michael Mauboussin did in his book Expectations Investing.  That is the correct general model to use; the only problem is that it is impossible for individuals or even small investment shops to implement.

So, when evaluating companies, rather than using a complex model for free cash flow and cost of capital, it makes more sense given limited time, to look at the most critical partial sensitivities of the true model.  What do I mean there?

I go back to what the best boss I ever had sent me regarding modeling: “80 to 90% of the valuable model can be encapsulated in the top 2-3 factors of the model.” What that implies to me regarding analysis of valuation is that simple ratios do still have punch.  But the challenge is selecting the ratios that are the most appropriate for companies in a given industry.

Here is the most basic thing I have learned so far about what valuation ratios to use: with financial companies use price-to-book, and with all other companies use price-to-sales.  How did I come up with this insight?  I spent a lot of time using one of Bloomberg’s functions [GE] and I found the tightest correlation of price movement to each of those variables.

But the intuition that I have received from that is not the end of the story.  Let’s stop for a moment and think about what various valuation ratios mean.

The classic ratio is the price to earnings multiple.  It makes a lot of intuitive sense because we want to know how much we are earning per dollar spent on the stock.  Earnings, and its cousin operating earnings, are the lowest figures on the income statement.  They are also the most manipulated numbers on the income statement.  But there are other ratios on the income statement, less manipulated, that exist higher up on the income statement, or over on the cash flow statement.

Operating income/earnings is a means of trying to look at the profitability the business before interest, taxes and nonrecurring elements.  Many analysts think that this is a better measure of earnings on a continuing basis than ordinary earnings, because it eliminates a lot of noise.  But you can go higher up on the income statement and move to price to sales.  Particularly in an environment like this where sales growth is so scarce, the price to sales ratio plays a larger role.  Exactly how cheap can a stock get on price to sales ratio basis?  Where would its valuation be stretched?

Away from that, on the cash flow statement, we have EBITDA, cash from operations and free cash flow.  EBITDA is earnings before interest, taxation, depreciation and amortization.  Cash from operations is simple to understand.  How much cash is the business generating?  So long as that does not involve the buildup of additional liabilities, or need for additional capital expenditure, that can be a useful figure for analysis.  The same is true of EBITDA.  Free cash flow goes further and subtracts maintenance capital expenditure.

Operating earnings and free cash flow are proxies for trying to understand what the income generating capacity of the business is on a normalized basis.

But then there are balance sheet measures like price-to-book and price to tangible book.  The idea is to ask how much assets are allocable to the equity investors.  In financial companies, where the cash flow statement is pretty meaningless, attempting to estimate the value of the firm, using book or tangible book has some power.  Why?  Because financial firms are in the business of shepherding scarce capital in order to produce returns.  Since capital is typically a constraint for earnings price-to-book is a useful measure for analyzing the valuation of the financial company.

The same is largely true of sales for industrial companies, sales are relatively hard to fake.  Now don’t get me wrong here, sales have been faked in many companies.  It helps to study revenue recognition policies, and it also helps to consider the buildup in accruals for accounts receivable.  Yes, sales can be faked, but cash from sales is very hard to fake.

One more metric worthy of consideration is enterprise value to EBITDA.  This metric is useful during times when there are a lot of mergers and acquisitions going on.  This metric measures in a crude way the amount of free cash flow that the assets of the business throw off.  It is the way many acquirers would look at a business.

Another way to think of it is, is the business more valuable in the hands of the current management, or the hands of new management?  If it is more valuable in the hands of new management, enterprise value to EBITDA is the better metric.  If the current management makes it more valuable, then price to sales is the better measure.

Now there are two more ideas that must be considered here: cost of capital, and reversion to mean.  In one sense we want to look at earnings, or free cash flow in excess of the cost capital employed.  The way I handled this is not to estimate the cost of capital but to look at the credit ratings, leverage on the balance sheet, volatility of the stock price, and volatility of earnings in order to get a feel for the riskiness of the company.

Mean-reversion is involved in value investing, in the sense that return on equity for firms tends to mean-revert over time.  What that implies is that it makes sense to pay attention to valuation, because firms in bad shape often clean up their act and get better, and firms in exceptionally good shape find their excess earnings competed away by other firms.  And that’s why value investing tends to work: companies with cheap valuations improve, and multiples expand.  Companies with high multiples tend to contract, because it is difficult to maintain superior growth over the long haul.

4 Comments

  • microcap says:

    I have had the same insight over the years David— P/B is everything for financials, of limited [but not zero] use for everything else.

  • > And that’s why value investing tends to work.

    One consideration, Mr. Merkel. You are not saying “when” it works… I have found with value investing that the stock of your choosing may remain depressed -sorry, I meant undervalued- for a long time. The fact that one can spot an undervalued issue does not necessarily mean everyone else will.

    But thank you for this extensive list. It has added a lot to my limited knowledge.

  • Joshua Chance says:

    “Depending on the industry, this can mean low P/E, low P/B, low P/S, low P/CFO, low P/FCF, or low EV/EBITDA.”

    If you’re in the request-taking mood then a rundown of which valuation ratios work best with which industry and why would be awesome. Haven’t seen it covered on the blogosphere yet, besides an exercise in datamining without the why of it all.

  • matt says:

    You mentioned cost of capital but did not elaborate. How do you model this?

2 Trackbacks