Markel Ventures

So the question was raised about MKL management's use of EBITDA to evaluate Markel Ventures (MV).  I wondered about that too, but had totally forgotten that this issue was addressed back in 2010 in the letter to shareholders (thanks to someone pointing it out in the comments section of my previous post).  I figured clarifying this and taking a closer look at MV is worth another post, so think of this as a footnote to the other post about the 2014 annual report.

Buffett and Munger always caution us against people who talk about EBITDA; they say it is total nonsense as ITDA are all real expenses.

Of course, MKL being MKL, they took the time to carefully explain their choice of EBITDA as a metric to measure progress at MV (and I carefully forgot all about it!).

Here is the section from the 2010 letter where they explain:

With the growth of Markel Ventures, it is important to add some new measures when reporting our financial results to you. We will begin to do so this year and in the years to follow by reporting EBITDA, or earnings before interest, taxes and depreciation and amortization, that Markel Ventures has produced for us. In 2010, Markel Ventures EBITDA was $20.4 million as compared to $4.6 million in 2009. For a reconciliation of Markel Ventures EBITDA to net income, see the table on page 130.

While we generally do not like EBITDA as a performance measure, it does provide useful information if you keep in mind several caveats. Here is the way we break it down by its components to make it useful to us. We share this with you so that you can see how we think about it ourselves.

First, we start with the “E,” Earnings. These are the GAAP after-tax earnings of the businesses involved. They are the starting point for the EBITDA calculation and they are calculated in accordance with GAAP. If we had owned these businesses for a long time, rather than through recent acquisitions, we could just stop there.

It is fair to ask then, why are you adding back Interest, Taxes, Depreciation and Amortization? Aren’t they real expenses? The honest answer is both yes and no, and we’ll try to explain why in the paragraphs that follow.

Interest is clearly a real expense. As such, we count it in considering the economics of each of these businesses. Other than the real estate intensive business of ParkLand Ventures, we operate the Markel Ventures businesses with little or no debt. Consequently, the “I” factor of EBITDA is an insignificant difference between GAAP earnings and EBITDA. Whether we adjusted for “I” or not, the answer would be roughly the same under these circumstances.

Taxes are also real expenses. Real taxes though are affected by leverage and the associated deductible interest expense. In order to make effective apples to apples comparisons about the performance of underlying businesses which might have different amounts of debt in their capital structure, we add back the tax expense to make the results comparable.

Depreciation and Amortization get more interesting. Depreciation is the accounting method that tries to capture the sense of how much the capital equipment of a company is wearing out and what it will cost to replace it eventually. Fortunately, the Markel Ventures companies are not capital intensive and do not need massive doses of capital spending to remain competitive. This is an important aspect of what we are looking for when we purchase companies. Normally, we do not want to invest in businesses that require massive capital expenditures. As such, depreciation, like interest, tends to be only a minor factor in the adjustment from GAAP earnings to EBITDA.

Amortization represents the accounting effort to capture the cost of maintaining the intangible assets of a company each year. Given that the Markel Ventures companies have brand power in their markets and produce excellent cash flows, our purchase price reflects that reality and was a bigger number than just the hard asset values of existing working capital and real estate assets. The price we pay in excess of those tangible assets gets assigned to intangible assets and those intangible assets are written off over time in the amortization account.

We add back amortization to earnings as we are looking at the management teams and evaluating these businesses for two major reasons. First, as the CEO’s of these businesses make decisions, amortization of intangible assets doesn’t affect how they interact with their customers, manage their operations, price their products or any other fundamental aspect of running the business. Had we (or someone else) never purchased the business, this amortization would not exist. It is almost a “Lewis Carroll- Through The Looking Glass” type issue. If you look at these businesses from the point of view of Markel’s financial statements, which is what we are doing in this report, the earnings of the companies are penalized by an annual amortization charge that starts on day one of the acquisition and goes away over a number of years.

Second, the other reason we add back amortization is that if the companies are well run, continuing to build the value of their brand and increasing their earnings, the intangible value of these companies should be INCREASING not DECREASING, as the presence of an amortization charge would suggest.

So that's basically the explanation which makes sense.

Let's take a look at some details from the 2014 10-K.


Markel Ventures Balance Sheet
Here is the balance sheet of MV from the 2014 10-K:

	December 31,
(dollars in thousands)	2014				2013
ASSETS
Cash and cash equivalents	$	106,552			$	61,742
Receivables	92,036				78,764
Goodwill	215,967				191,629
Intangible assets	237,070				182,599
Other assets	534,725				421,714
Total Assets	$	1,186,350			$	936,448
LIABILITIES AND EQUITY
Senior long-term debt and other debt (1)	$	359,263			$	217,119
Other liabilities	213,794				146,343
Total Liabilities	573,057				363,462
Redeemable noncontrolling interests	61,048				72,183
Shareholders' equity (2)	553,972				501,370
Noncontrolling interests	(1,727		)		(567		)
Total Equity	552,245				500,803
Total Liabilities and Equity	$	1,186,350			$	936,448

(1)	Senior long-term debt and other debt as of December 31, 2014 and 2013 included $252.9 million and $116.4 million, respectively, of debt due to other subsidiaries of Markel Corporation, which is eliminated in consolidation.

(2)	Shareholders' equity includes $498.6 million and $444.1 million as of December 31, 2014 and 2013, respectively, which represents Markel Corporation's investment in Markel Ventures and is eliminated in consolidation.

Shareholders' equity is $550 million, which is what the segment is carried on the books for.   One of the things I am interested in is if this understates the intrinsic value of the segment.

There are various ways to look at this.   The easiest way to value it is to do what we do for BRK; look at pretax earnings and slap on some sort of multiple on it.  It may be conservative, but I would use 10x pretax.  You can use whatever you feel comfortable with.  We can also look at adjusted, or cash net earnings and also some sort of EV/EBITDA valuation.

Anyway, let's take a look at the income statement and see what we come up with.

	Years ended December 31,
(dollars in thousands)	2014				2013				2012
OPERATING REVENUES
Net investment income	$	4			$	4			$	4
Other revenues	838,121				686,448				489,352
Total Operating Revenues	838,125				686,452				489,356
OPERATING EXPENSES
Amortization of intangible assets	24,283				20,674				18,684
Other expenses	775,219				613,250				432,956
Total Operating Expenses	799,502				633,924				451,640
Operating Income	38,623				52,528				37,716
Interest expense (1)	13,400				11,230				11,269
Income Before Income Taxes	25,223				41,298				26,447
Income tax expense	13,160				14,654				8,109
Net Income	12,063				26,644				18,338
Net income attributable to noncontrolling interests	2,506				2,824				4,863
Net Income to Shareholders	$	9,557			$	23,820			$	13,475

(1)	Interest expense for the years ended December 31, 2014, 2013 and 2012 includes intercompany interest expense of $8.7 million, $6.4 million and $6.4 million, respectively, which is eliminated in consolidation.

And here is the reconciliation for adjusted EBITDA which breaks out the goodwill impairment charge so we can  calculate adjusted pretax earnings and cash net earnings:

	Years Ended December 31,
(dollars in thousands)	2014				2013				2012
Markel Ventures Adjusted EBITDA - Manufacturing	$	71,133			$	64,415			$	44,963
Markel Ventures Adjusted EBITDA - Non-Manufacturing	23,931				19,372				15,398
Markel Ventures Adjusted EBITDA - Total	95,064				83,787				60,361
Goodwill impairment	(13,737		)		—				—
Interest expense (1)	(12,184		)		(9,283		)		(9,782		)
Income tax expense	(12,848		)		(13,988		)		(7,868		)
Depreciation expense	(24,706		)		(19,313		)		(14,205		)
Amortization of intangible assets	(22,032		)		(17,383		)		(15,031		)
Markel Ventures net income to shareholders	9,557				23,820				13,475
Net income from other Markel operations	311,625				257,201				239,910
Net income to shareholders	$	321,182			$	281,021			$	253,385

(1)	Interest expense for the years ended December 31, 2014, 2013 and 2012 includes intercompany interest expense of $8.7 million, $6.4 million and $6.4 million, respectively.

Pretax Earnings
From the above table, we can see that net earnings at MV was $9.6 million in 2014.  We can add back amortization, goodwill impairment and income tax expense to get a pretax earnings figure.  That comes to around $58 million.   Put a 10x multiple on that and we get a valuation of $580 million for MV, slightly more than what it is carried on the books for ($550 million).   In this sense, BPS still closely reflects the value of MKL including MV.

P/E
From the above $58 million pretax earnings, we can deduct income tax expense and get a cash net earnings or adjusted net earnings figure.  That comes to around $45 million.   Put a 14x P/E on that (long running average for U.S. stocks) and we get $630 million.   That's a little more than the $550 million it is carried at.  But $90 million against total MKL shareholders' equity of $7.6 billion is only 1.2%.

EV/EBITDA
I don't know what the right multiple would be for this business, but let's say we use 10x EV/EBITDA.  Adjusted EBITDA is $95 million, so at 10x, it's worth $950 million.  There is cash and debt on the balance sheet so adjusting for that, we get an equity valuation of  $700 million.   So that's still a bit more than the $630 million using the adjusted P/E ratio.  $700 million is $150 million more than it is carried at, but that still comes to around 2% of MKL's total shareholders' equity.

I tend to be more comfortable with adjusted net cash earnings and pretax earnings to value this sort of thing.

Two-Column Valuation for MKL?
So now that we have some of these figures, how about valuing MKL using the two-column approach we use for BRK?

I tend to like the pretax earnings valuation, so let's use $580 million as the value of MV.  There is 14 million shares outstanding so that comes to $41/share.  Investments per share is over $1,300/share, so the two-column method valuation wouldn't make too much sense at this point.   And by the way, there probably should be some deductions against the $1,300/share investments, like long term debt, for a more proper valuation.  Adjusting for that, investments per share net of long term debt is still around $1,200/share.

Conclusion
So just looking at what we have, it doesn't look like there needs to be a big adjustment at this point to MKL's book value because of MV.  More aggressive valuations would, of course, bump up any needed adjustment to MKL BPS.  But at this point, I am comfortable that MKL's BPS captures most of the value at MV.  Over time, the gap will no doubt widen.

I would imagine that over the next few years, the two column approach to value MKL may become increasingly the norm as it makes sense.

Of course, I have my usual reservations about various aspects of that, but as an overall indicator of intrinsic value (and whatever adjustments each investor makes for themselves) it can be pretty useful.

And by the way, while we are on the subject of amortization, I came across a really nice essay that Buffett put at the end of his letter to shareholders back in 1983.

Check it out, from the back of the 1983 BRK letter to shareholders:

Goodwill and its Amortization: The Rules and The Realities

This appendix deals only with economic and accounting Goodwill - not the goodwill of everyday usage. For example, a business may be well liked, even loved, by most of its customers but possess no economic goodwill. (AT&T, before the breakup, was generally well thought of, but possessed not a dime of economic Goodwill.) And, regrettably, a business may be disliked by its customers but possess substantial, and growing, economic Goodwill. So, just for the moment, forget emotions and focus only on economics and accounting.

When a business is purchased, accounting principles require that the purchase price first be assigned to the fair value of the identifiable assets that are acquired. Frequently the sum of the fair values put on the assets (after the deduction of liabilities) is less than the total purchase price of the business. In that case, the difference is assigned to an asset account entitled "excess of cost over equity in net assets acquired". To avoid constant repetition of this mouthful, we will substitute "Goodwill".

Accounting Goodwill arising from businesses purchased before November 1970 has a special standing. Except under rare circumstances, it can remain an asset on the balance sheet as long as the business bought is retained. That means no amortization charges to gradually extinguish that asset need be made against earnings.

The case is different, however, with purchases made from November 1970 on. When these create Goodwill, it must be amortized over not more than 40 years through charges - of equal amount in every year - to the earnings account. Since 40 years is the maximum period allowed, 40 years is what managements (including us) usually elect. This annual charge to earnings is not allowed as a tax deduction and, thus, has an effect on after-tax income that is roughly double that of most other expenses.

That's how accounting Goodwill works. To see how it differs from economic reality, let's look at an example close at hand. We'll round some figures, and greatly oversimplify, to make the example easier to follow. We'll also mention some implications for investors and managers.

Blue Chip Stamps bought See's early in 1972 for $25 million, at which time See's had about $8 million of net tangible assets. (Throughout this discussion, accounts receivable will be classified as tangible assets, a definition proper for business analysis.) This level of tangible assets was adequate to conduct the business without use of debt, except for short periods seasonally. See's was earning about $2 million after tax at the time, and such earnings seemed conservatively representative of future earning power in constant 1972 dollars.

Thus our first lesson: businesses logically are worth far more than net tangible assets when they can be expected to produce earnings on such assets considerably in excess of market rates of return. The capitalized value of this excess return is economic Goodwill.

In 1972 (and now) relatively few businesses could be expected to consistently earn the 25% after tax on net tangible assets that was earned by See's doing it, furthermore, with conservative accounting and no financial leverage. It was not the fair market value of the inventories, receivables or fixed assets that produced the premium rates of return. Rather it was a combination of intangible assets, particularly a pervasive favorable reputation with consumers based upon countless pleasant experiences they have had with both product and personnel.

Such a reputation creates a consumer franchise that allows the value of the product to the purchaser, rather than its production cost, to be the major determinant of selling price. Consumer franchises are a prime source of economic Goodwill. Other sources include governmental franchises not subject to profit regulation, such as television stations, and an enduring position as the low cost producer in an industry.

Let's return to the accounting in the See's example. Blue Chip's purchase of See's at $17 million over net tangible assets required that a Goodwill account of this amount be established as an asset on Blue Chip's books and that $425,000 be charged to income annually for 40 years to amortize that asset. By 1983, after 11 years of such charges, the $17 million had been reduced to about $12.5 million. Berkshire, meanwhile, owned 60% of Blue Chip and, therefore, also 60% of See's. This ownership meant that Berkshire's balance sheet reflected 60% of See's Goodwill, or about $7.5 million.

In 1983 Berkshire acquired the rest of Blue Chip in a merger that required purchase accounting as contrasted to the "pooling" treatment allowed for some mergers. Under purchase accounting, the "fair value" of the shares we gave to (or "paid") Blue Chip holders had to be spread over the net assets acquired from Blue Chip. This "fair value" was measured, as it almost always is when public companies use their shares to make acquisitions, by the market value of the shares given up.

The assets "purchased" consisted of 40% of everything owned by Blue Chip (as noted, Berkshire already owned the other 60%). What Berkshire "paid" was more than the net identifiable assets we received by $51.7 million, and was assigned to two pieces of Goodwill: $28.4 million to See's and $23.3 million to Buffalo Evening News.

After the merger, therefore, Berkshire was left with a Goodwill asset for See's that had two components: the $7.5 million remaining from the 1971 purchase, and $28.4 million newly created by the 40% "purchased" in 1983. Our amortization charge now will be about $1.0 million for the next 28 years, and $.7 million for the following 12 years, 2002 through 2013.

In other words, different purchase dates and prices have given us vastly different asset values and amortization charges for two pieces of the same asset. (We repeat our usual disclaimer: we have no better accounting system to suggest. The problems to be dealt with are mind boggling and require arbitrary rules.)

But what are the economic realities? One reality is that the amortization charges that have been deducted as costs in the earnings statement each year since acquisition of See's were not true economic costs. We know that because See's last year earned $13 million after taxes on about $20 million of net tangible assets - a performance indicating the existence of economic Goodwill far larger than the total original cost of our accounting Goodwill. In other words, while accounting Goodwill regularly decreased from the moment of purchase, economic Goodwill increased in irregular but very substantial fashion.

Another reality is that annual amortization charges in the future will not correspond to economic costs. It is possible, of course, that See's economic Goodwill will disappear. But it won't shrink in even decrements or anything remotely resembling them. What is more likely is that the Goodwill will increase - in current, if not in constant, dollars - because of inflation.

That probability exists because true economic Goodwill tends to rise in nominal value proportionally with inflation. To illustrate how this works, let's contrast a See's kind of business with a more mundane business. When we purchased See's in 1972, it will be recalled, it was earning about $2 million on $8 million of net tangible assets. Let us assume that our hypothetical mundane business then had $2 million of earnings also, but needed $18 million in net tangible assets for normal operations. Earning only 11% on required tangible assets, that mundane business would possess little or no economic Goodwill.

A business like that, therefore, might well have sold for the value of its net tangible assets, or for $18 million. In contrast, we paid $25 million for See's, even though it had no more in earnings and less than half as much in "honest-to-God" assets. Could less really have been more, as our purchase price implied? The answer is "yes" - even if both businesses were expected to have flat unit volume - as long as you anticipated, as we did in 1972, a world of continuous inflation.

To understand why, imagine the effect that a doubling of the price level would subsequently have on the two businesses. Both would need to double their nominal earnings to $4 million to keep themselves even with inflation. This would seem to be no great trick: just sell the same number of units at double earlier prices and, assuming profit margins remain unchanged, profits also must double.

But, crucially, to bring that about, both businesses probably would have to double their nominal investment in net tangible assets, since that is the kind of economic requirement that inflation usually imposes on businesses, both good and bad. A doubling of dollar sales means correspondingly more dollars must be employed immediately in receivables and inventories. Dollars employed in fixed assets will respond more slowly to inflation, but probably just as surely. And all of this inflation-required investment will produce no improvement in rate of return. The motivation for this investment is the survival of the business, not the prosperity of the owner.

Remember, however, that See's had net tangible assets of only $8 million. So it would only have had to commit an additional $8 million to finance the capital needs imposed by inflation. The mundane business, meanwhile, had a burden over twice as large - a need for $18 million of additional capital.

After the dust had settled, the mundane business, now earning $4 million annually, might still be worth the value of its tangible assets, or $36 million. That means its owners would have gained only a dollar of nominal value for every new dollar invested. (This is the same dollar-for-dollar result they would have achieved if they had added money to a savings account.)

See's, however, also earning $4 million, might be worth $50 million if valued (as it logically would be) on the same basis as it was at the time of our purchase. So it would have gained $25 million in nominal value while the owners were putting up only $8 million in additional capital - over $3 of nominal value gained for each $1 invested.

Remember, even so, that the owners of the See's kind of business were forced by inflation to ante up $8 million in additional capital just to stay even in real profits. Any unleveraged business that requires some net tangible assets to operate (and almost all do) is hurt by inflation. Businesses needing little in the way of tangible assets simply are hurt the least.

And that fact, of course, has been hard for many people to grasp. For years the traditional wisdom - long on tradition, short on wisdom - held that inflation protection was best provided by businesses laden with natural resources, plants and machinery, or other tangible assets ("In Goods We Trust"). It doesn't work that way. Asset-heavy businesses generally earn low rates of return - rates that often barely provide enough capital to fund the inflationary needs of the existing business, with nothing left over for real growth, for distribution to owners, or for acquisition of new businesses.

In contrast, a disproportionate number of the great business fortunes built up during the inflationary years arose from ownership of operations that combined intangibles of lasting value with relatively minor requirements for tangible assets. In such cases earnings have bounded upward in nominal dollars, and these dollars have been largely available for the acquisition of additional businesses. This phenomenon has been particularly evident in the communications business. That business has required little in the way of tangible investment - yet its franchises have endured. During inflation, Goodwill is the gift that keeps giving.

But that statement applies, naturally, only to true economic Goodwill. Spurious accounting Goodwill - and there is plenty of it around - is another matter. When an overexcited management purchases a business at a silly price, the same accounting niceties described earlier are observed. Because it can't go anywhere else, the silliness ends up in the Goodwill account. Considering the lack of managerial discipline that created the account, under such circumstances it might better be labeled "No-Will". Whatever the term, the 40-year ritual typically is observed and the adrenalin so capitalized remains on the books as an "asset" just as if the acquisition had been a sensible one.

* * * * *

If you cling to any belief that accounting treatment of Goodwill is the best measure of economic reality, I suggest one final item to ponder.

Assume a company with $20 per share of net worth, all tangible assets. Further assume the company has internally developed some magnificent consumer franchise, or that it was fortunate enough to obtain some important television stations by original FCC grant. Therefore, it earns a great deal on tangible assets, say $5 per share, or 25%.

With such economics, it might sell for $100 per share or more, and it might well also bring that price in a negotiated sale of the entire business.

Assume an investor buys the stock at $100 per share, paying in effect $80 per share for Goodwill (just as would a corporate purchaser buying the whole company). Should the investor impute a $2 per share amortization charge annually ($80 divided by 40 years) to calculate "true" earnings per share? And, if so, should the new "true" earnings of $3 per share cause him to rethink his purchase price?


* * * * *

We believe managers and investors alike should view intangible assets from two perspectives:

(1)  In analysis of operating results - that is, in evaluating the underlying economics of a business unit - amortization charges should be ignored. What a business can be expected to earn on unleveraged net tangible assets, excluding any charges against earnings for amortization of Goodwill, is the best guide to the economic attractiveness of the operation. It is also the best guide to the current value of the operation's economic Goodwill.

(2)  In evaluating the wisdom of business acquisitions, amortization charges should be ignored also. They should be deducted neither from earnings nor from the cost of the business. This means forever viewing purchased Goodwill at its full cost, before any amortization. Furthermore, cost should be defined as including the full intrinsic business value - not just the recorded accounting value - of all consideration given, irrespective of market prices of the securities involved at the time of merger and irrespective of whether pooling treatment was allowed. For example, what we truly paid in the Blue Chip merger for 40% of the Goodwill of See's and the News was considerably more than the $51.7 million entered on our books. This disparity exists because the market value of the Berkshire shares given up in the merger was less than their intrinsic business value, which is the value that defines the true cost to us.

Operations that appear to be winners based upon perspective (1) may pale when viewed from perspective (2). A good business is not always a good purchase - although it's a good place to look for one.


We will try to acquire businesses that have excellent operating economics measured by (1) and that provide reasonable returns measured by (2). Accounting consequences will be totally ignored.

At yearend 1983, net Goodwill on our accounting books totaled $62 million, consisting of the $79 million you see stated on the asset side of our balance sheet, and $17 million of negative Goodwill that is offset against the carrying value of our interest in Mutual Savings and Loan.

We believe net economic Goodwill far exceeds the $62 million accounting number.