Tuesday, May 3, 2016

DCF Myth 3.1: The Margin of Safety - Tool for Action or Excuse for Inaction?

In my last post on dealing with uncertainty, I brought up the margin of safety, the tool that many value investors claim to use to protect themselves against uncertainty. While there are certainly some in the value investing community who have found a good way to incorporate MOS into their investing process, there are many more who seem to have misconceptions about what it does for them as well as the trade off from using it. 

The Margin of Safety: Definition and Rationale

While the margin of safety has always been around, in one form or another, in investing, it was Ben Graham who brought the term into value investing in The Intelligent Investor, when he argued that the secret of sound investment is to have a margin of safety, with the margin of safety defined as the difference between the value of an asset and its price. The definitive book on MOS was written by Seth Klarman, a value investing icon. Klarman’s book has acquired a cult following, partly because of its content and partly because it has been out of print now for years; a quick check of Amazon indicates a second-hand copy can be acquired for about $1600. Klarman’s take on margin of safety is similar in spirit to Graham’s measure, with an asset-based focus on value, which is captured in his argument that investors gain the margin of safety by “buying at a significant discount to underlying business value and giving preference to tangible assets over intangibles”.

There are many reasons offered for maintaining a margin of safety. The first is that the value of an asset is always measured with error and investors, no matter how well versed they are in valuation techniques, have to recognize that they can be wrong in their judgments. The second is that the market price is determined by demand and supply and if it diverges from value, its pathway back is neither quick nor guaranteed. The proponents of margin of safety point to its benefits. By holding back on making investment decisions (buy or sell) until you feel that you have a margin of safety, they argue that you improve your odds of making successful investments. In addition, They also make the point that having a healthy margin of safety will reduce the potential downside on your investments and help protect and preserve your capital. 

The Margin of Safety: Divergence across Investors
As a concept, I not only understand the logic of the MOS, but also its allure, and I am sure that many investors adopt some variant of it in active investing, but there are differences in how it is employed:
  1. Valuation Basis: While MOS is often defined it as the difference between value and price, the way in which investors estimate value varies widely. The first approach is intrinsic value, either in its dividend discount model format or a more expansive DCF version. The second approach estimates value from accounting balance sheets, using either unadjusted book value or variants thereof (tangible book value, for instance). The third approach is to use a pricing multiple (PE, EV to EBITDA), in conjunction with peer group pricing, to estimate “a fair price” for the company. While I would contest even calling this number a value, it is still used by many investors as their estimated value.
  2. Magnitude and Variability: Among investors who use MOS in investing, there seems to be no consensus on what constitutes a sufficient margin. Even among investors who are explicit about their MOS, the follow up question becomes whether it should be a constant (say 15% for all investments) or whether it should be greater for some investments (say in risky sectors or growth stocks) than for others (utilities or MLPs).
The bottom line is that a room full of investors who all claim to use margin of safety can contain a group with vast disagreements on how the MOS is computed, how large it should be and whether it should vary across investments and time.

Myths about Margin of Safety
When talking about value, I am often challenged by value investors on how I control for risk and asked why I don’t explicitly build in a MOS. Those are fair questions but I do think that some of the investors who are most enamored with the concept fundamentally misunderstand it. So, at the risk of provoking their wrath, here is my list of MOS misconceptions.

Myth 1: Having a MOS is costless
There are some investors who believe that their investment returns will always be improved by using a margin of safety on their investments and that using a larger margin of safety is costless. There are very few actions in investing that don’t create costs and benefits and MOS is not an exception. In fact, the best way to understand the trade off between costs and benefits is to think about type 1 and type 2 errors in statistical analysis. If type 1 errors refer to the fact that you have a false positive, type 2 errors reflect the opposite problem, where you have a false negative. Translating this proposition into investing, let’s categorize type 1 errors as buying an expensive stock, because you mistake it to be under valued, and type 2 errors as not buying a bargain-priced stock, because you perceive it wrongly to be over valued. Increasing your MOS will reduce your type 1 errors but will increase your type 2 errors. 

Many risk averse value investors would accept this trade off but there is a cost to being too conservative and  if that cost exceeds the benefits of being careful in your investment choice, it will show up as sub-par returns on your portfolio over extended periods. So, will using a MOS yield a positive or negative payoff? I cannot answer that question for you, because each investor has to make his or her own judgment on the question, but there are simple tests that you can run on your own portfolios that will lead you to the truth (though you may not want to see it). If you find yourself consistently holding more of your overall portfolio in cash than your natural risk aversion and liquidity needs would lead you to, and/or you don't generate enough returns on your portfolio to beat what you would have earned investing passively (in index funds, for instance), your investment process, no matter what its pedigree, is generating net costs for you. The problems may be in any of the three steps in the process: your valuations may be badly off, your judgment on market catalysts can be wrong or you may be using too large a MOS.

Myth 2: If you use a MOS, you can be sloppy in your valuations
Value investors who spend all of their time coming up with the right MOS and little on valuation are doing themselves a disservice. If your valuations are incomplete, badly done or biased, having a MOS on that value will provide little protection and can only hurt you in the investment process (since you are creating type 2 errors, without the benefit of reducing type 1 errors). Given a choice between an investor with high quality valuations and no/little MOS and one with poorly done valuations and a sophisticated MOS, I would take the former over the latter every single time.

I am also uncomfortable with investors who start with conservative estimates of value and then apply the MOS to that conservative value. In intrinsic valuation, conservative values will usually mean haircutting cash flows below expectations, using high discount rates and not counting in growth that is uncertain. In asset-based valuation, it can take the form of counting only some of the assets because they are tangible, liquid or both. Remember that you are already double counting risk, when you use MOS, even if your valuation is a fair value (and not a conservative estimate of value), because that value is computed on a risk-adjusted basis. If you are using a conservative value estimate, you may be triple or even quadruple counting the same risk when making investment decisions. If you are using this process, I am amazed that any investment manages to make it through your risk gauntlets to emerge as a good investment, and it does not surprise me that nothing in the market looks cheap to you.

Myth 3: The MOS should be the same across all investments 
I have always been puzzled by the notion that one MOS fits all investments. How can a 15% margin of safety be sufficient for both an investment in a regulated utility as well as a money-losing start-up? Perhaps, the defense that would be offered is that the investors who use MOS as their risk breakers would not look at companies like the latter, but I would still expect that even in the value investing spectrum, different investments would evoke different degrees of uncertainty (and different MOS).

Myth 4: The MOS on your portfolio = MOS on individual investments in the portfolio
I know that those who use MOS are skeptics when it comes to modern portfolio theory, but modern portfolio theory is built on the law of large numbers, and that law is robust. Put simply, you can aggregate a large number of risky investments to create a relatively safe portfolio, as long as the risks in the individual stocks are not perfectly correlated. In MOS terms, this would mean that an investor with a concentrated portfolio (who invests in three, four or five stocks) would need a much larger MOS on individual investments than one who spreads his or her bets across more investments, sectors and markets.

Expanding on this point, using a MOS will create biases in your portfolio. Using the MOS to pick investment will then lead you away from investments that are more exposed to firm-specific risks, which loom large on an individual company basis but fade in your portfolio. Thus, biotechnology firms (where the primary risk lies in an FDA approval process) will never make your MOS cut, but food processing firms will, for all the wrong reasons. In the same vein, Valeant and Volkswagen will not make your MOS cut, even though the risk you face on either stock will be lowered if they are parts of larger portfolios. 

Myth 5: The MOS is an alternative risk measure
I know that many investors abhor betas, and believe it or not, I understand. In fact, I have long argued that there are replacements available for portfolio theory-based risk measures and that not only is intrinsic value robust enough to work with these alternative risk measures but that the discount rate is not (and should not) be the ultimate driver of value in most companies. That said, there are some in the value investing community who like to use their dislike of betas as a bludgeon against all financial theory and after they have beaten that straw horse to death, they will offer MOS as their alternative risk measure. That suggests a fundamental misunderstanding of MOS. To use MOS, you need an estimate of value and I am not aware of any intrinsic value model that does not require a risk adjustment to get to value. In other words, MOS is not an alternative to any existing risk measure used in valuation but an add-on, a way in which risk averse investors can add a second layer of risk protection.

There is one possible way in which the MOS may be your primary risk adjustment mechanism and that is if you use a constant discount rate when doing valuation (a cost of capital of 8% for all companies or even a risk free rate) and then apply a MOS to that valuation to capture risk. If that is your approach, you should definitely be using different MOS for different investments (see Myth 3), with a larger MOS being used on riskier investments. I would also be curious about how exactly you make this MOS adjustment for risk, including what risks you bring in and how you make the conversion.

Margin of Safety – Incorporating into a Strategy
I would not put myself in the MOS camp but I recognize its use in investing and believe that it can be incorporated into a good investing strategy. To do so, though, you would need to do the following:
  1. Self examination: Even if you believe that MOS is a good way of picking investments, it is not for everyone. Before you adopt it, you have to assess not only your own standing (including how much you have to invest, how risk averse you are) but also your faith (in your valuation prowess and that markets correct their mistakes). Once you have adopted it, you still need the effects it has on your portfolio, including how often you choose not to invest (and hold cash instead) and whether it makes a material difference to the returns you generate on your portfolio.
  2. Sound Value Judgments: As I noted in the last section, a MOS is useful only if it is an addendum to sound valuations. This may be a reflection of my biases but I believe that this requires intrinsic valuation, though I am willing to concede that there are multiple ways of doing it right. Accounting valuations seem to be built on the twin presumptions that book value is an approximation of liquidation value and that accounting fair value actually means what it says, and I have little faith in either. As for passing of pricing as value, it strikes me as inconsistent to use the market to get your pricing number (by using multiples and comparable firms) and then argue that the same market misprices the asset in question.
  3. A Flexible MOS: Tailor the MOS to the investment that you are looking at: There are two reasons for using a MOS in the first place. The first is an acceptance that, no matter how hard you try, your estimate of value can be wrong and the second is that even if the value estimate is right, there is uncertainty about whether the market will correct its mistakes over your time horizon. If you buy into these two reasons, it follows that your MOS should vary across investments, with the following determinants.
  • Valuation Uncertainty: The more uncertain you are about your estimated value for an asset, other things remaining equal, the larger the MOS should be. Thus, you should use a smaller MOS when investing in mature businesses and during stable markets, than when putting your money in young, riskier business or in markets in crises.
  • Portfolio Tailoring: The MOS that you use should also be tailored to your portfolio choices. If you are a concentrated investor, who invests in a four or five companies, you should use a much higher MOS than an investor who has a more diversified portfolio, and if you the latter, perhaps even modify the MOS to be larger for companies that are exposed to macroeconomic risks (interest rates, inflation, commodity prices or economic cycles) than to company-specific risks (regulatory approval, legal jeopardy, management flux).
  • Market Efficiency: I know that these are fighting words to an active investor, red flags that call forth intemperate responses. The truth, though, is that even the most rabid critics of market efficiency ultimately believe in their own versions of market efficiency, since if markets never corrected their mistakes, you would never make money of even your canniest investments. Consequently, you should settle for a smaller MOS when investing in stocks in markets that you perceive to be more liquid and efficient than in assets, where the corrections will presumably happen more quickly than in inefficient, illiquid markets where the wait can be longer.
  • Pricing Catalysts: Since you make money from the price adjusting to value, the presence of catalysts that can lead to this adjustment will allow you to settle for a lower MOS. Thus, if you believe that a stock has been mispriced ahead of an earnings report, a regulatory finding or a legal judgment, you should demand a lower MOS than when you invest in a stock that you believe is misvalued but with no obvious pricing catalyst in sight. 
Finally, if MOS is good enough to use when you buy a stock, it should be good enough to use when you sell that stock. Thus, if you need a stock to be under valued by at least 15%, to buy it, should you also not wait until it is at least 15% over valued, to sell it? This will require you to abandon another nostrum of value investing, which is that once you buy a great company, you should hold it forever, but that is not just unwise but is inconsistent with true value investing.
Would I prefer to buy a stock at a 50% discount on value rather than at just below fair value? Of course, and I would be even happier if you made that a 75% discount. Would I feel even more comfortable if you estimated value very conservatively. Yes and I would be delighted if all you counted was liquid assets. That said, I don't live in a  world where I see too many of these investments and when I do, it is usually the front for a scam rather than a legitimate bargain.  That is the reason that  I have never formally used a MOS in investing. I did buy Valeant at $32, because my valuation of the stock yielded $45 for the company. Would I have still bought the stock, if my value estimate had been only $35 or if it was a big chunk of my portfolio? Perhaps not, but I have bought stocks that were priced at my estimated fair value and have held back on investments that I have found to be under valued by 25% or more. Why? That has to wait for my coming post on simulations, since this one has run its course.

YouTube Video

Uncertainty Posts
  1. DCF Myth 3: You cannot do a valuation, when there is too much uncertainty
  2. The Margin of Safety: Excuse for Inaction or Tool for Action?
  3. Facing up to Uncertainty: Probabilities and Simulations
DCF Myth Posts
  1. If you have a D(discount rate) and a CF (cash flow), you have a DCF.  
  2. A DCF is an exercise in modeling & number crunching. 
  3. You cannot do a DCF when there is too much uncertainty.
  4. It's all about D in the DCF (Myths & 4.5)
  5. The Terminal Value: Elephant in the Room! (Myths & 5.5)
  6. A DCF requires too many assumptions and can be manipulated to yield any value you want.
  7. A DCF cannot value brand name or other intangibles. 
  8. A DCF yields a conservative estimate of value. 
  9. If your DCF value changes significantly over time, there is something wrong with your valuation.
  10. A DCF is an academic exercise.


Anonymous said...

Prof, I don't think investors should set a "required" MOS (say at 15%) when it comes to making investment decisions. MOS is meant to 'measure' the downside risk of an asset at the prevailing price, not the gap of price vs intrinsic value PER SE. By that definition, it's actually not quantifiable, because it is a probability game rather than a mathematical game.

With all due respect, I don't like it when the subject of MOS become too scientific. Risk should be a philosophical subject, imo.

Chee Fui

steve said...

1. Since i have a copy of Klarman's book i am glad to hear it is priced at $1,600. I can only wonder how the intrinsic value of this book has been arrived at- just kidding.
2. The Klarman book bashes the use of Beta, but highlights doing a DCF w/o specifying how you arrive at the discount rate- most curious to me.
3. Seems to me that you are investing with an MOS at times depending upon how close to your intrinsic value estimate is to the market price of the stock is. To me that is the appropriate way to use an MOS: do a DCF, examine the market price, and moreover the "price action" of how the market is trading which can provide you with a better guess as to an entry point.

Raffaele De Gennaro said...

Dear Professor,
another excellent post, but I have a question, thanks.

A fairly valued company it is not (always) a bad investment, but we know that we can reasonably get (only) a rate of return near the Cost of Equity.
Is it wrong asking for a MOS in order to gain a potentially higher rate of return?

But maybe the answer to this question will be discussed in your next post…

Yours sincerely,
Raffaele De Gennaro

Aswath Damodaran said...

If you have lots of choices, you should of course try to earn a return higher than your risk-adjusted required return. That is why I suggested one of the tests you run to see if it works for you is how much of your portfolio ends up in cash. If you have too much in cash, you are choosing not to invest at a required return and holding cash instead (when you don't have any good need for holding that cash).

Hugo Francisco Caycedo Godoy said...

Hi Professor Damodaran. very illustrative post. I've always considered the margin of safety as a tool to arrive to a "target price". Getting a "fair market value" (giving that the market is fair)is another issue that rely on different assumptions.

Professor, would it be possible to get betas from indexes based entirely on fundamentals rather than from indexes affected by behavioral biases?

steve said...

So, Prof D, do you believe that with the phenomena of globalization, better access to data, and the accelerated nature of news,etc that perhaps the efficient market hypothesis may be ringing louder than ever which may say it might be a better probabilistic bet to buy index funds, and perhaps some stocks around the edges where you think you might actually have some advantage such as specific industry knowledge?

Anonymous said...

Dear Prof. Damodaran,

I am currently doing my thesis on the limitations of valuation methods. Where in your blog I can find articles on this topic to get some inspiration?

Thanks in advance for your response.

Best Regards,


Larry said...

Dear Professor,

One variation on the Margin of safety that my company employs is to fix the discount rate at our corporate hurdle rate (12% - 15% depending on the target) when doing the DCF RATHER than trying to estimate the target's WACC. I'm curious to see what you think about this approach.

Thanks in advance!

Jay said...

I think you are wrong! Valuation and investing are two different things. Valuation is done from the perspective of a marginal investor who, unlike me, doesn't pay taxes. If I want the same return on my investment after taxes as he does, I have to pay a lot less for any asset I buy. Or have I missed something here?

Anonymous said...

I admire your posts Prof. They are all very informative. That is very generous of you.

However, I have some thoughts on MOS and what I call its counter part, i.e. the discount rate.

We have opposite arguments from two of the renowned value investors.

Warren Buffett rubbishes the argument that using higher discount rate can reduce risk of an investment. Yet, I wonder how MOS does the job better.

You argue that there isn’t another way to value an investment without using a risk-adjusted discount rate. You clearly defy Buffett.

I find (http://valueanalyst.blogspot.in/2016/02/capm-and-margin-of-safety-two-sides-of.html) more acceptable although it does not offer any conclusion.

The question I ask is how an MOS can reduce risk that higher discount rate cannot. Apparently, Buffett’s use of MOS is also based on intuition. It could be 10% or 50%.

In fact, I agree with you on the fact that value of a business is all cash flows discounted at the risk-adjusted discount rate.

However, your value of the business is for the marginal investor, not for you and me. It could also be the right cost of capital for the business. That does not help us as investors either.

If I use the risk-adjusted discount rate, all I get is the rate of return expected by the marginal investors. I wouldn’t want that.

For that purpose, I set my own rate of return. And here’s the catch: I invest in the stock only if the buy price is able to give me that rate of return. If not, I wait. I do believe in market inefficiency. And I am aware that my rate of return should be reasonable.

Would love your thoughts.