Enterprise Value Revisited

Report a problem or error
1

After the introduction of the “canned” EV (see https://www.portfolio123.com/mvnforum/viewthread_thread,5710),

I am interested to revisit the ratio. According to the wikipedia article (Enterprise value - Wikipedia):

The portfolio 123 EV is defined as (quoting Marco):

For a long time I had been using

as a proxy to EV, mainly because the other items were not available at the time. I have found the original formula seems to work better than the new EV factor, even though it is “less correct”. Partly I have noticed the new factor contains a lot more NAs, possibly because the required items are NA, possibly because the of the following problem:

The wikipedia article makes it very clear for each of the item’s listed, it must be at market value. However, when we are listing items to add we typically rely on balance sheet values. Sometimes balance sheet entries can have stark differences between the recorded value (often at cost or at “par”) and the market value. A company may have bought an office building in 1900 for $10,000, and it may still sit on the balance sheet as that despite clearly being worth millions now.

The way I see it, I look at enterprise value from the point of view of an acquiring company - someone interested in buying the whole business. (Buffet said imagine buying the whole business when you buy a stock). The EV is meant to be the “true” cost of the business, not just paying off the equity holders but all other interested parties too. For the debt holders, paying them off at par seems a reasonable approximation. At the very least, if the debt is acquired by the incoming company, the debt will adjust itself to the credit status of the acquirer and knowing the market value pre-acquisition is less important.

The preferred equity is a bit more problematic. This cannot be redeemed at par, it would have to be bought back at market value, and this may be very different from the par value recorded on the balance sheet (presumably this is what PfdEquity is). My proposal is to “infer” the value of the preferred equity by taking the Preferred Dividend, guessing the preferred yield and back-calculating the preferred equity value:

Preferred Equity = Preferred Dividend / Preferred Dividend Yield

Unfortunately it still looks like Preferred Yield is an unknown, so I make a guess, take the 30 year yield (as a proxy to the “zero” rate), and add a risk premium (RP).

Preferred Equity = Preferred Dividend / (30 year yield + RP)

In p123 language I think that is:

= 100*(PfdDivQ / ((close(0,#Bond30Yr)+RP))

What value for the RP? This is harder - I am thinking we could do with knowing the credit rating of the company, or perhaps it does not matter since if we are the the position of an acquirer, the preferred yield would (post acquisition) be based on our own credit rating, so for the purposes of comparison (ranking) it does not matter. I am going to experiment with values from 1% to 5%. Perhaps one could use the “AvgRec”, since this happens to run from 1-5 (best-worst)!

I’d be interested in hearing opinions. As for “non controlling interest”, this seems a bit of a black box and I am having some more trouble tackling this one. I thought the post (https://www.portfolio123.com/mvnforum/viewthread_thread,5447#36772) that contained the formula:

Looks like an interesting start. Perhaps using a similar approach the Minority interest can be backed out. Or perhaps not, opinions welcome!

2

Olikea,

First…cool that you are really digging on this stuff. And posting. This has been where my head’s been at lately also.

Simple answer -
As a general bias, If the ‘difference’ in backtesting is ‘material’ enough given the number of holdings to rule out random variation, then I would use whichever test better as they both ‘make sense.’ If the difference is immaterial given the number of holdings, I would likely use the more ‘correct’ one…because, if nothing else,it’s more likely that most of the market is using that one.

Some reasons it likely makes ‘little difference.’

  1. Starting with this universe - AVgDailyTot(50)>500000 and MktCap>50 and close(0)>2. About 3660 stocks.
  2. Checking what % of the stocks the Preferred Equity and Minority interest piece is more than 10% of estimated EV. It’s under 5%. (NonControlInt(0,TTM)+ PfdEquity(0,TTM))/EV>.10. That’s about 182 stocks. And these stocks are about ‘average’ with the rest of the portfolio in terms of out of sample performance.

So…these additional pieces are likely almost always less than the estimation error and uncertainties in the underlying core formula. So…simpler is probably fine.

Now, for more thinking on the ‘problem.’

Beyond that, I’ve always found the term Enterprise Value misleading. EV is, as commonly used is really better thought of as ‘minimum acquisition cost’ that a single acquirer might have to pay - if there offer attracted no other offers from the market. True market value for many of the underlying terms is often very hard to know - which is why the dispersion between top quartile M&A managers and lower quartile M&A managers can be very large. For many pieces of this formula, it’s dependent on the number and qualities of suitors and varies widely. (See, for example: Vanguard Asset Management | Personal Investing in the UK. Can see 30%/yr or more difference between higher quartile managers and bottom quartile managers).

So…True ‘Enterprise Value’ to an acquirer is more about things like future cash flows plus true market value of all net assets (to the acquirer) minus true value of net liabilities (to the acquirer) minus likely final acquisition cost - all discounted at a firm specific IR rate (based on other investment opp’s and cost of capital). Plus strategic value to the acquirer (both for their own business and for keeping this asset out of the hands of competitors). So…it’s very hard to evaluate for the broad market as a whole with a general formula.

What I learned in selling a business is that these ‘valuations’ can vary wildly. At one point we were worth 2X sales. When a second buyer came in…we were worth much more. When the second buyer dropped out…we became worth only 4X earnings…(about 40% of sales)…and this constantly varied. Another 5 years later, with growth having fallen a lot, we were now worth 6X sales again. Because multiple parties wanted the business. Further, what people wanted varied. For some potential acquirers…all that mattered was our sales stream. They planned to cut out all marketing (10% of sales for us), lay off as many people as possible, lower quality of goods sold and raise margins and draw out as much cash as possible for as long as possible and then sell the shell of what was left. For some acquirers, all that matters was our hi quality and premium brand image. They planned to license us into all areas of products globally. Etc. So…these players valued us very, very differently. And we were worth very different amounts to every suitor based on their core business and what they wanted to do with us.

Some things that always help: 1. Having more suitors. And being in hi demand. What is likely to drive number of suitors way up? First off…a) core IP that’s very hard to recreate and that there is a big rush to own (a timeliness aspect to it). Second…a lower acquisition cost. There are fewer buyers for $50 Billion dollar companies than $200 Million dollar ones.

So…one possible investment strategy might be to look for companies with a likely ‘future enterprise value’ much greater than a current likely ‘market acquisition value.’ With a large number of potential suitors. In a major, meaningful - hi potential market area.

But…how do we figure out both sides of this equation? That’s up to each system builder.

On the debt and preferred equity ‘true value’, it’s going to be very hard to get a detailed bottom-up estimate of what it’s likely to cost as many notes can have a whole host of covenants (variable vs. fixed rates on the debt, special conversion or call terms embedded, etc). There are simple estimates on P123, but they don’t tell the true picture here. These covenants can be very complex and vary widely by company. So…Not sure there’s any simple way to accurately estimate the true cost of this to a likely acquirer from what we have on P123. I have some systems I use. But they involve fairly complex financial statement analysis for internal consistency and longitudinal changes over time.

Random note: There’s some reason to believe that companies with more restrictive debt covenants placed on them by the ‘borrowing market’ are significantly more likely to be timely in reporting ‘true’ financial results in times of distress. See:

So…it would be very fun / helpful to have quantitative signals for this. But likely beyond the scope of P123.

Best,
Tom

3

Oliver,

Preferred equities are definitely not debt securities, I’m afraid it would be very difficult to approximate the market value of your pfd equity without the data. For one thing each pfd equity has different strings attached to it that will substantially impact the way its market value fluctuate. It’s not just the credit rating of the company, what about callable or convertible preferred ? How do you discount these types of options in your formula ?

4

A couple of thoughts:

  1. EV may give best results as a factor only for smaller companies where there is even a possibility they may be acquired. Or where an older company may be broken up and sold in pieces. Does including Apple or GE in a Universe where EV is used make sense? Or does the investment community look at other factors for different size (Mrktcap/EV/revenue) companies?
  2. The role of Goodwill is not clear in the EV formula other than in its Mktcap. It plays a major role in valuation as Tom says but it is very subjective. Goodwill was $100M in Caterpillar’s purchase of ERA but it had to be written off recently. 2000 had many stories of companies paying way too much for intangibles. But it plays little role in the recent the Cerebus/Safeway deal. As a note, Goodwill impairment write downs spiked in 2008 as companies took losses and bundled up their write downs with the kitchen sink. Maybe intangible write downs could be (one) sign of a market bottom?
  3. trying to figure out the true market value of assets is tough because it affects so many things in balance sheets and income statements and companies (even the most conservative) will play games. having said that, using non-database information (like recent suspension of a dividend) can be very illuminating.
5

Bear in mind what Enterprise Value is. It originated as a simple back-of-the-envelope calculation to give a very rough sense of what one would have to pay to acquire a business. Actually back of the envelope may even be a misnomer: I don’t know who, exactly invented EV, but I’m guessing it might have its origins at a three-martini lunch in which an investment banker made notes on a napkin to try to persuade a potential corporate client who fell asleep during a pitch-book presentation that a proposed deal made sense.

It’s very important that one understand this informal origin; especially the fact that as its use spread, there were NO analytically rigorous processes to determine whether and where the ratio made sense. As it turns out, we got lucky and there is a lot the ratio can do for us. For example, I’m starting to migrate from PS to EVS since EVS better facilitates valuation comparisons among companies with different degrees of leverage.

But we also need to recognize its limitations.

Potential disconnect between book numbers and market values is definitely an issue. So, too, is the absence of reliable data-basing of idle assets. However creative people can come up with proxies. Notwithstanding the potential problems with the way Olikea handled preferred (see aurelaurl), I think it’s an interesting idea and consistent with the over-riding back-of-the-envelope personality of EV. You coukld also make a case for creating a custom formula that adds X% to the p123 EV in order to reflect the frequent willingness of acquirers to pay a control premium (one who owns 100% of a business can extract wealth and build in ways not feasible with a publicly-owned company). Etc.

It’s also important to address NA values. Unlike some other outfits, we automatically assign negative EVs and NA value. It’d definitely possible for a company to have negative NA (i.e. it’s especially cash heavy). But analytically, it has no meaning (you can’t expect to contact the Board of a negative EV company and offer to take ownership of all the common if they’ll pay you).

EV is fine to use, but be prepared to do so in a loosey-goosey manner. Don’t expect to always be able to draw on the sort of intellectual rigor you might apply to more traditional ratios.

6

The aim here is not to write a “perfect” formula, as Buffet says “better to be approximately correct than precisely wrong”. In the majority of cases in the past where I have encountered “value traps” it is because of something not considered when doing the initial analysis. If you were buying a house, you would want to know that a person, for example, had a claim to live there.

Similiarly with preferred equity, we cannot just ignore it because it is too difficult or not important.

If you want to invest in 20 holdings, then a crude formula is fine, but if it is more concerntrated, the formula has to be more accurate. You cannot simply select the 5 cheapest (by PE ratio) and expect everything to work out. What is the value of a company is the ultimate question in the stock market, as Tom eludes to, it is a fuzzy concept but that should spur us on to try and figure it out.

7

1

8

Hi Olikea,

I too have looked into variations of EV.

I use my own custom formula for EV. I substitute zeros for the NAs in PfdEquity and NonControlInt. The reason for that is that for most companies PfdEquity and NonControlInt is zero or small enough not to matter when buying baskets of companies. Therefore I assume that unknown (NA) values are meaningless and substitute zero for NAs.

Deducting cash from market cap may make sense even for huge companies that are not being bought out. David Einhorn owns Apple because of all that cash it owns. I don’t think that he thinks that Apple is going to be acquired. Therefore he probably believes that the cash has some value to shareholders; either when paid out as a special dividend, or used to buy something else with decent earnings or whatever.

Regarding adding debt to market cap I tend to agree with you that it may only make sense for companies that are candidates for acquisitions. If a company is too large to be acquired then debt is not a price issue but a Financial strength issue. As a financial strength issue I am more interested in short term debt compared to cash and earnings than long term debt.

In defense of EV, Joel Greenblatt used it and he is among the best in the business. Many members of VIC (which is Greenblatt’s investment club who are some of the best investors in the US) use it too. No measure is infallible; in fact to quote Warren Buffet from his latest AR “As much as Charlie and I talk about intrinsic business value, we cannot tell you precisely what that number is for Berkshire shares (nor, in fact, for any other stock).” If Buffett cannot do it then no one can. But he makes estimates. So should we.

Of course EV is not perfect. I suspect that it works better for US companies that foreign companies because of the relative ease of buyouts in the US. Perhaps that’s why Greenblatt didn’t have success with his International mutual funds.

9

I realise this topic has been revisited several times. However, I am struggling to make sense of the P123 EV, as used in Greenblatt ranking. The formula for $EV is:
MktCap + DbtTotQ - (CashPSQ * ShsOutMR)

But what is DbtTot? P123 appears to use “Short term Debt +Long Term Debt”. However, there are other types of debt. For example, SDRL, Sea Drill (a basket case!), the balance sheet has:

Accounts payable =122m
Short term debt =3,246m
Other Current Liabilities =1,525m
Long Term Debt = 7,062m
Other Liabilities = 319m

P123 has DbtTot= 3,246+7,062 =10,308m
I believe DbtTot is really the sum of all 5 components= 122+10,308+ 1525+319 =12,274m

The discrepancy is nearly $2billion! Next year SeaDrill is projected to earn a bit over $100m. At this rate it would take Seadrill nearly 20 years to simply pay back the discrepancy, so it cannot be considered to be trivial.
Why does P123 use “Short term Debt +Long Term Debt”? Wouldn’t it be better to use LiabTotQ=“Total Liabilities” (from the balance sheet) which for Seadrill is 12,274m.

David

10

David,

Enterprise value is used to measure the cost to an acquirer. There is a reason why acquirers use the regular formula for EV.

BTW, in my backtesting, I got better results using the standard EV.

11

Chipper,
I completely agree that EV is the total cost to an acquirer. Surely this will include all of: Accounts payable, Short term debt, Other Current Liabilities, Long Term Debt and Other Liabilities. The acquirer can’t say to “other current liability holders”- sorry we aren’t going to pay you.
From Greenblatt’s book, I don’t recall him ignoring certain types of debt.
Better performance from the current EV is certainly an important factor, but is it masking a more important issue?
I have recently come across companies (sorry can’t recall which) where DbtTotQ is zero, because all the debt has been stuffed into “Other Current Liabilities” and “Other Liabilities”. Doesn’t appear sensible to me.
Thanks for responding, but I’m not yet convinced DbtTotQ is the correct item to include in $EV.

David

12

I understand your question and don’t have a crystal clear answer. But the facts are that Greenblatt included only interest-bearing debt in his enterprise value calculation. [quote]
Earnings yield was measured by calculating the ratio of pretax operating earnings (EBIT) to enterprise value (market value of equity* + net interest-bearing debt).
*Including preferred equity
[/quote](from The Little Book That Still Beats the Market, by Joel Greenblatt, page 169. Bold added. This book is a good read by the way. I read the book again after designing my own ranking systems and noticed things that I had missed the first time.)

Furthermore, I have read hundreds of individual stock writeups on Greenblatt’s Value Investors Club. These are some of the best hedge fund analysts in the world. They exclude “other debt” from EV. So, too, during stock acquisitions, if anyone mentions the price they paid, it is in terms of EV in the classical sense.

Why?

Non-interest expenses such as accounts payable are part of running a business. As long as the business keeps going, this debt just rolls over. The acquirer will not have to raise the capital to pay off this debt.

Interest bearing debt, on the other hand, is something that an acquirer has to meet head on. Many acquisitions take place with money that is partially borrowed. If the target already has bank debt, it will reduce the amount of additional debt that the acquirer will be able to add.

13

Chipper,

Thanks for your reply. I am trying to understand the benefits of DbtTotQ and LiabTotQ. I agree that Accounts Payable should be excluded from EV. It’s part of running the business. However, what about “Other Current Liabilities”. I understand that Seadrill have new rigs being built, and there are payments to be made as the construction takes place. A new owner would also have to make these payments.

Perhaps one possible reason is that Greenblatt is looking for good companies at a cheap price. They will normally have plenty of cashflow, so perhaps for these companies “other Current Liabilities” can also be considered “part of running the business”. However, Seadrill is likely to have its ordinary shareholders wiped out, and bondholders take control. Hence Seadrill won’t be appearing at the top of any Magic Formula list. Seadrill probably had a rolling program of new builds, so historically it was “part of running the business”. However, in its current circumstances Seadrill is in survival mode, and LiabTotQ is more relevant than DbtTotQ. I don’t believe anybody will be buying Seadrill because its Debt Liabilities are too high.

David

14

EV is a metric that aims at “permanent capital” (or at least capital that’s as permanent as anything can be in today’s ever-changing world). Think of this as the money you’d raise to start a business; contributions-investment from yourself and co-owners (equity), with whom you must share profit, and contributions from permanent (long-term) debt, where you take on a fixed cost but need not share profit and policy voting.

Debt that comes in this way (on day one, or later on as a business increase capital) is a completely different animal from, let’s call it transactional debt. A classic example is money you temporarily raise in order to manufacture your products or buy your inventory. You expect to sell it soon and use the cash that comes in this way to get rid of the debt; until the next product cycle starts all over again. This sort of debt, even though it may be refinanced forever and ever, is not really a permanent part of your capital base. It’s just there to help you balance your cash flows, to tide yourself over between the time you must spend money to make-acquire the stuff you’ll sell and the revenue you get when you sell it. What we’re describing here is “working capital.”

A lot of short-term debt and other liabilities that are not counted in EV are part of this working capital. They don’t stand alone. They are balanced, more or less depending on the nature of the business and the company’s financial management practices, against comparably transactional assets; inventories, receivables, cash and cash equivalents, and “other” transactional type assets.

Greenblatt does make certain iconoclastic adjustments of his own (can’t recall off the top of my head, but it’s in his book) but ultimately, he and everyone else is aiming at the same general thing; a distinction between “permanent” capital and “transactional” stuff.

Actually, it’s critical that you recall and dig deeply into specific companies. Except in the dreams of overly-imaginative click-craving bloggers, fake accounting is very rare probably because of prosecutions in the past and the fact that so many have seen the HBO series “Oz” or Netflix’s “Orange is the New Black” and really don’t want to go to places like that. Typically, if you dig (and that, in some cases, may require studying footnotes and other textual portions of 10-Ks and, perhaps, googling and studying some accounting standards, especially if it’s not a standard-industrial-type company), you’re likely to find that the difference between permanent capital and transactional stuff is there.

15

I’ve been thinking this over a bit.

EV measures the price an acquirer would have to pay. Seadrill’s debt may push the company into bankruptcy. It may eat into earnings for a very long time. But it doesn’t affect the price an acquirer would have to pay. Existing interest bearing debt, however, does affect the amount of capital that an acquirer would have to pay.

EDIT: Compustat lists forty different types of liabilities that could fit into “other liabilities”. Most of them are non-interest bearing debt. Seadrill’s are interest rate swaps which do seem to fit into the spirit of interest bearing debt. In other words, Compustat lumps all types of “other liabilities” together. The good news is that this is a special case. Even better news is that we may be able to filter out these types of situations by filtering out overly indebted stocks.

16

Hi, I’m joining the debate a bit late here but Wikipedia’s definition has the following:

“+ unfunded pension liabilities and other debt-deemed provisions”

Is there any way to account for pension liabilities in Portfolio123? I’ve been looking and couldn’t find anything.

17

This is one of those occasions when we must recall what Wikipedia is or isn’t. That definition is published only because its what was left behind by the last person who edited the article. If I go in and delete it, then nobody will be able to say Wikipedia includes unfunded pension liabilities . . . unless or until somebody else goes in and changes it again.

Regardless of the merits, the item is not typically included. For one thing, it’s not databased. For another, it is an exceptionally soft number based on company pension administrator projections of future returns on plan assets – which almost nobody in the investment community believes are even remotely in any arguable ballpark. The general expectation among those who know is that there’s little if any hope any company with pension obligations would be able to come anywhere close to funding them (and when this is discussed at conferences, there is considerable discomfort as attendees patiently wait for such presentations to end so proceedings can move on to less unmentionable topics). The general sense, albeit typically unspoken-out-loud, is that we, as a whole, will figure it all out eventually.

Politics of pensions aside, if you wanted to somehow factor unfunded pension liability into EV, you’d have an intellectually interesting result but a number that has no relationship to anything done or considered in the investment community.

18

Hi Marc,

Thanks for your response.

I realize that this is not a common item but let me tell you about my situation and perhaps you (and others more knowledgeable about pension than me) can give more color.

I’m looking at a micro-cap (33M mkt cap) company called CSPI. Relevant 10-K here: https://www.sec.gov/Archives/edgar/data/356037/000035603717000005/cspi-10k20160930.htm

My issue is that the company has “Pension and retirement plans” liabilities of 13M. For a market cap of 33M, this is a significant amount. Is this a projected amount?

So I’m trying to see if this should be included in my EV calculation. Is this an “unfunded pension liability” as opposed to a “funded” one?

For whatever it is worth, here’s what the 10-K says: “The funded status of pension and other postretirement benefit plans is recognized on the consolidated balance sheet. … Liabilities for amounts in excess of these funding levels are accrued and reported in the consolidated balance sheets.”

Should this be considered a form of long term debt? How am I supposed to factor this item into my valuation?

I’ve been reading about pension plans just to understand this item but the rabbit hole keeps leading me to other issues.

Thanks for any insight you may offer.

UPDATE: I found this link, which seems to help: http://www.cnbc.com/id/100959648

Best,
Anlam

19

I checked that 10-K as well as several others and good heavens, pension accounting has gotten considerably more complex than it used to be.

This company does seem to have a bigger balance-sheet pension items than usual, and that could be a red flag. Clearly, it can’t be included in debt since it is not a component of capital. It is correctly considered as part of non-current liabilities.

Do I think to redefine your EV question, it Isn’t so much whether the pension item should be included as part of EV (which we can’t because we don’t have it) but whether the total of non-current liabilities should be included. That is a very interesting question.

It would be easy for you to do that using a p123 custom formula. We would hesitate the chance the pre-set item out of deference to current custom (doing so would raise a lot of hackles). But that’s the whole point of custom formulas; we give you the ability to outthink the crowd and to develop, test and use your own unique ideas, and if it can help your portfolios, the last thing in the world you should want is for us to standardize it and thus disseminate your intellectual property to the world.

As to whether this specific idea is a good one, can I ask you to make a not to yourself to email me on or shortly before April 20. I want the reminder because I’m going out Indiana University for a few days and will be visiting with a couple of accounting professors. I’d be curious to hear what they think.

For the short term for you, if ever I find something in financials that looks important which I can’t figure out, that alone is ground for scrapping a company.

20

Thanks Marc, for taking the time to respond to my query and also looking into my issue.

<< whether the total of non-current liabilities should be included. That is a very interesting question. It would be easy for you to do that using a p123 custom formula>>

I will certainly do so - and see if I see any improvements in backtesting.

By the way, yesterday the company’s share price suddenly appreciated %20-%25. (I have no idea why? Hedge fund managers reading Portfolio123 forums? Someone on this site messing with me?)

Well there goes my margin of safety out the window. Case closed/dismissed.

They had a %5 yield, too - so if I could have only figured out the pension bit, I would probably have invested.

It’s no problem to miss winners - I just want to stay clear of losers.