Michael Mauboussin is currently the Director of Research at BlueMountain Capital Management. Michael recently published a widely circulated paper discussing the merits and pitfalls of the usage of EV/EBITDA in valuation work.
Before his current role at BlueMountain, Michael was the Head of Global Financial Strategies at Credit Suisse and the Chief Investment Strategist at Legg Mason Capital Management. He has also authored multiple books, and a wide variety of articles in publications including the Harvard Business Review, The Journal of Applied Corporate Finance.
Kevin Harris from SumZero sat down with Michael to further discuss his thoughts on use of the EV/EBITDA multiple, value investing, and his career as an investor.
Kevin Harris, SumZero: In a recent interview with Business Insider, you commented that “multiples are shorthand for the valuation process.” Could you expand upon the blind spots caused by an over-reliance on multiples in the valuation process?
Michael Mauboussin, BlueMountain Capital Management: Let’s start at the very beginning. When we invest, we defer current consumption in order to consume more (after taking inflation into account) in the future. Most investment opportunities don’t guarantee that you’ll have more in the future than you do today, but that’s the motivating driver.
So it follows that the value of any financial asset is the present value of future cash flows. Say you buy a bond from a company. You give the company $1,000 and it is contractually obligated to pay you coupons in a timely fashion and to return your principal at maturity. The coupon reflects the risk of the investment and the maturity tells you about the timing.
The value of a stock is based on the same principle. The problem is that there is no coupon or maturity. Even dividends are at best a quasi-contract. The intellectually correct way to value stocks is similar to bonds, and that is a discounted cash flow (DCF) model.
The problem is that while the DCF model is analytically sound, it demands a number of judgments. And the model’s output varies greatly based on the inputs. Now there are ways to deal with this challenge, but most practitioners choose to avoid a DCF model altogether and instead use shorthands in the form of multiples.
Shorthands are helpful because they save time. But the tradeoff is that shorthands commonly come with blind spots. Popular multiples, including price-earnings (P/E) and enterprise value-to-earnings before interest, taxes, depreciation, and amortization (EV/EBITDA) have a number of limitations but the main one is that they fail to account for capital intensity—be it working capital, capital expenditures, or acquisitions.
That means that two businesses can have the same growth in earnings or EBITDA but very different capital needs. The company that needs less capital to grow will be more valuable because there will be more cash available to distribute to shareholders. The market tends to get this, but some businesses may screen cheap or expensive even when the multiple is an accurate representation of the underlying economics.
Harris: In several past papers, you’ve touched upon base rates as a powerful mental model you see as underutilized in investment management. How can investors best understand and implement the concept of base rates and mean regression in security analysis?
Mauboussin: This is one of the most powerful, and underutilized, ideas in forecasting in general and investing in particular. The basic idea is that there are a couple of ways of thinking about a problem or prediction. The first way, which is often what we do, is to gather lots of information, combine that with your own experience and input, and project into the future.
Ask a student when their term paper will be complete, a homeowner when their kitchen renovation project will be done (and what it will cost), or an investor how a stock will appreciate—you’ll find that most use this approach.
The second way is to consider what happened to others when they were in the same, or a similar, situation. This is the base rate. So rather than asking “what will this company’s sales growth rate be?” you can ask, “What is the distribution of growth rates for all companies of this size?” This is formally called reference-class forecasting.
Let me add quickly that using base rates is very unnatural for two reasons. First, you have to set aside the information and experience you’ve gathered. We tend to hold our own thoughts in high esteem. Second, you have to find the base rate, which may not be at your fingertips.
In order come up with an accurate forecast you want to intelligently combine your own view with the base rate. How you do that effectively also provides a great deal of insight into regression toward the mean.
We need to add the idea of persistence. Specifically, how we measure the correlation between one metric over two periods of time. correlated is a measure of performance over time. For example, if you measure year-over-year sales growth rates over time, you’ll see the correlation is about 0.30. That’s the base rate.
Positive correlations can range from zero to 1.0. Zero means there is no correlation (results are random) and 1.0 means the two measurements are perfectly correlated. When correlations are close to zero, luck tends to be the dominant force. When correlations are close to 1.0, skill tends to be the dominant force.
So here’s the really cool insight: The correlation tells you how much of the base rate, versus your own assessment, you should use for your forecast. If the correlation is zero, you place all of the weight on the base rate. If the correlation is 1.0, you can rely on your own assessment. If you think about this a bit, you can see that it is also telling you about the rate of regression toward the mean. Low correlations are consistent with rapid regression, and high correlations are consistent with slow regression.
Let me give you one example from the world of sports. The correlation between batting averages for players in Major League Baseball from one year to the next is about 0.4. The overall batting average for the league is around .250. So if a player hits .300 for a season, a good prediction for the subsequent year would be .270 (.250 x 0.6 + .300 x 0.4).
You can now take this mental model and apply it to investing. Gross profitability and operating profit margins tend to have high correlations. This, of course, varies by industry. Net income growth tends to have a very low correlation. Return on invested capital is in the middle of those. These base rates and the regression toward the mean they imply can be very helpful for an investor trying to model a business.
Harris: In your “Managing the Man Overboard” and “Celebrating the Summit” papers, you discussed exceptionally detailed frameworks to approaching rapid stock price inclines and declines. You’ve also mentioned your interest in Atul Gawande’s “The Checklist Manifesto” in multiple interviews. Do you have similarly detailed and structured checklists or heuristics for other parts of your investment process? If so, which do you rely on most?
Mauboussin: Checklists are demonstrably valuable in many fields, including aviation and medicine. Atul Gawande makes a very useful distinction between DO-CONFIRM and READ-DO checklists. DO-CONFIRM checklists ensure that execution is thorough. You basically do your job and stop periodically to make sure you’ve been methodical. READ-DO checklists are for emergencies. The checklist writer has anticipated certain types of problems and potential solutions. So you need only read the checklist and do what it says. “Managing the Man Overboard” is an example of a READ-DO checklist.
In my experience, many fundamental investors tend to shun checklists and feel they are constraining. Here’s my take: Being a good investor requires mastery of certain building blocks, including valuation, competitive strategy assessment, and evaluating a management team’s capital allocation skills. It’s like being a great tennis player. You need to have a solid forehand and backhand, footwork, and serve. The basics have to be rock solid. So, too, in investing. Checklists are valuable for those building blocks in investing. So our work in those areas always includes a checklist.
The actual tennis match includes combining those building blocks to win a match. Likewise, successful investing requires applying tools effectively in order to generate excess returns. I have found that the checklist is most effectively applied to the building blocks of an investment thesis. That allows for good process and permits the investor to exercise some judgment in the investment process.
Harris: A recent letter of Greenlight Capital’s aptly summarized the tension between rigidity and flexibility in investing: “we have been accused of being stubborn, but one person’s stubbornness is another person’s discipline.”. What is your advice for value investors balancing between flexibility and discipline in the investment process?
Mauboussin: This is very challenging, and to some degree is tied to time horizon. But the best advice is to become a good Bayesian updater. That is, be good at updating your prior view as new information reveals itself.
From a practical standpoint, there are some things you can do to make this process more transparent. For example, a good investment thesis means that your expectations for the future are distinct from what is priced into a security. If your expectations are different, you should be able to articulate what will happen to reshape market expectations. Call them signposts—events that indicate whether your thesis is on track.
You should write down, in advance, the signposts that are central to your thesis and make them statements that include probabilities and a time period. (There is a 70 percent probability that ABC Corporation’s operating profit margin will increase by 500 or more basis points in the fourth quarter but the market expects flat margins.)
As you pass these signposts, you have to be brutally honest as to whether the results are consistent with your thesis. And you have to avoid thesis creep, which is altering your justification for an investment after your original thesis didn’t play out.
The signpost is the prior, and the information prompts the update. When information disconfirms the thesis, you should change your mind. When it doesn’t, you can maintain your view. That is one way to manage stubbornness and discipline.
Harris: In a recent interview of yours featured in Graham and Doddsville, you detailed the two ways that doctors make mistakes – ignorance and execution. What are the most common execution related mistakes that investment managers make? How can these be avoided?
Mauboussin: I’ll mention two high level mistakes. The first is one I’ve talked about a lot: the failure to distinguish between fundamentals and expectations. Your job as an investor is to figure out when the market’s expectations are unduly high or low. One analogy is pari-mutuel betting at the horse race track. You don’t generate excess returns by picking winners; you win by figuring out which horse has odds that misspecify the horse’s chances of winning.
The big mistake is to focus too much on fundamentals. When things are good, people want to buy. When they are bad, people want to sell. But great investors separate what’s priced in from what’s going to happen. Almost all investors think they are doing this, but very few actually do.
The second mistake is a lack of congruence between actual process and espoused goals. There are a lot of ways to beat the market but you have to make sure that every aspect of your process is dedicated to what you are trying to do. For example, many money managers claim to have a long time horizon but the turnover in their portfolio is much higher than what would be consistent with that time framework. They say one thing and do another.
Investment managers should periodically step back and ask whether what they are doing is likely to generate excess returns over time and, if so, whether their process fully serves that goal. I think many firms suffer from a lack of congruence—they say they are playing one game but are actually playing another one.
Harris: A recent S&P report identified a representative sample of US companies as having missed self-projected EBITDA numbers on average by 29% and 34% in the last two years. What are your views on the rampant use of add-backs and inaccurate projections of EBITDA?
Mauboussin: Companies generally try to make their results look good. This is especially true when incentive compensation, or access to credit markets, is tied to certain metrics. There is not much new with this.
The antidote to this is to focus on free cash flow—the difference between a company’s earnings and the capital it needs to invest in the business to secure future growth. At the end of the day, free cash flow is the money available for distribution to the claimholders. And that is the lifeblood of value.
Harris: All else equal, should more leverage mean a higher EBITDA multiple, given that more debt vs. equity drives down the cost of capital? Or do you prefer to apply a risk discount for higher leverage to offset the benefit of the lower cost of capital?
Mauboussin: This is classic finance theory. John Graham wrote a highly-cited paper on this. More leverage lowers the cost of capital up to a point, after which the effect is reversed as the risk of distress looms larger. I would also add that the benefit of debt is more muted in a world with lower corporate tax rates.
Harris: Exhibit 9 of your recent “What Does an EV/EBITDA Multiple Mean?” paper demonstrates the strong correlation (r = .79) between EV/EBITDA and P/E multiples between the top 1500 US industrial companies. However, some companies with similar EV/EBITDA multiples have P/E multiples that diverge significantly. What plays into the divergence between the two, and when should analysts rely on one multiple or the other?
Mauboussin: There are a few reasons that these multiples can diverge. One example is assumed asset life. Imagine two competitors that make the same capital outlay but that assume different asset lives. The earnings of the company choosing a longer asset life will be higher than that of the company that chooses the shorter asset life even as the EBITDA is identical.
A company’s capital structure can also affect the P/E multiple. Consider the simple case of a debt-financed share buyback program, which serves to increase leverage by replacing equity with debt in the capital structure. The earnings per share impact of the buyback is a function of the P/E multiple and the after-tax cost of new debt. Whether a buyback adds to or detracts from earnings per share is independent of whether it adds to or detracts from value but it does affect the multiple.
Multiples can also vary as the result of different tax rates, which have an impact on enterprise value and earnings but not on EBITDA. Finally, some companies have unconsolidated businesses or cross holdings that may factor into a calculation of enterprise value but can distort earnings or EBITDA.
As always, you want to look through cosmetic differences or adjustments to truly understand the economics of the business.
Harris: Could you expand upon the limitations investors and management encounter using the EV/EBITDA multiple?
Mauboussin: EV/EBITDA can be useful but there are a number of pitfalls. The first is that there is not a proper reflection of the investment needs of the business, a point we’ve already touched on. The risk in using EBITDA is that it understates the capital intensity of the business and overstates the amount of cash a company can distribute.
The second pitfall is that multiples in general do not explicitly reflect business risk. Operating leverage, the percentage change in operating profit as a function of the percentage change in sales, is a useful measure of business risk. Higher risk justifiably leads to a lower multiple, but the risk is implicit.
The final problem has to do with taxes. Two companies with the same EBITDA and capital structures may pay taxes at dissimilar rates. As a result, the EV/EBITDA multiples will be justifiably different.
Harris: What are your biggest takeaways from having taught security analysis at Columbia Business School since 1993? How has teaching informed and impacted your investment career?
Mauboussin: In my experience, teaching is helpful in two very important ways. The first is that teaching compels introspection. Imagine someone saying to you, “Prepare 20 hours of lectures on what you do all day.” Your first reaction might be “That’s not so hard, I just have to watch myself and document what I do.” But if you are at all thoughtful, the question of “What am I doing?” becomes “Why am I doing it this way?” And that launches a quest to think about your process in a more rigorous and systematic way. Because my course and my day job overlap enormously, there is a great benefit for me to think a lot about how I approach my career.
The second benefit is that teaching compels clarity of thought. Many have said that you don’t understand a topic unless you have written about it, or taught it, effectively. I subscribe to that view. I have encountered many personal beliefs that I thought I understood but didn’t really understand until I put pen to paper or included them in a lecture.
Working with students who are bright, engaged, motivated, and critical is a great way to sharpen clarity of thought and to improve communication skills.
Harris: What advice would you have for managers or analysts at the beginning of their careers? What or who had the biggest impact on you?
A great quality, no matter what career you select, is curiosity. So the advice I would give is to find something that sparks your curiosity and start—and don’t stop—learning. In investing, a lot of learning comes through reading. Most of the great investors I know are avid readers. And they don’t limit their material to business books but rather expand it to other domains.
I have been incredibly blessed to have been exposed to some great thinkers who have had a deep influence on me. First I would mention Al Rappaport, a mentor, friend, and collaborator who has taught me a lot about investing and about how to live and think well. Bill Miller, a legendary investor, taught me a lot about how to learn, the importance of temperament, and how ideas for different fields can bear fruit in investing.
My association with the Santa Fe Institute has also been deeply influential. The combination of ideas and people, with the study of complex systems at the core, has enriched my thinking in almost all corners of my career. I would finally mention Charlie Munger, Vice Chairman at Berkshire Hathaway, who was one of the first investors to discuss biases and the value of a mental models approach to investing, business, and life.