In recent reviews we have commented on what we believe to be increasingly obvious risks in the equity markets. One of those appears to be in technology stocks, once again the largest sector of the S&P 500 Index, yet now seeming subject to a new round of disruptive technology (the transition from P.C. to tablet computers).
A broader risk to the market is likely to be profit margins. Of the 50 largest market capitalization companies that outperformed the S&P 500 in 2010, a great many recently reported net profit margins above 20%, and of the others reported margins near or above historical peaks. Currently, the 20 largest companies in the S&P 500 with the highest profit margins, ranging from Amgen (AMGN) and Google (GOOG) (at around 30%), to Cisco (CSCO) and Altria (MO) (near 18%), comprise well over 15% of the market value of the index. If the average profit margin of companies that comprise more than 15% of the S&P 500 well exceeds 20%, how is it possible that the S&P will exhibit its historical post-recession expansion of profit margins and earnings? It's not likely to happen.
Ultimately, this situation poses a great danger to investors. Oddly, an extremely high profit margin is not the most desirable of business characteristics – it is a competition attractor of the first order. It is for this reason that sustained profit margins of 20% or higher are a historical rarity. The risk of increased competition is difficult to quantify, because it doesn't exist at the moment. Yet, it's reasonable to suppose that it will. The maintenance of profit margins at these high levels is, paradoxically, an earnings risk in itself. For example, if a company with a temporary profit margin of 25% ultimately stabilizes at 20%, an extraordinary accomplishment on an absolute basis, it has actually experienced a 20% earnings decline from its peak to its stabilized margin.
A related risk: valuation compression.
Today's more-or-less average P/E ratios are deceptive as valuation tools because of the aforementioned inflated margins. There is, though, an alternate method to assess the considerable valuation risk now inherent in the broad market. The following table displays the price-to-book value ratios of the 10 largest companies in the S&P 500. These range from a high of 8.70x to a low of 1.09x. The S&P 500 price-to-book ratio is 3.68x. It makes no allowance for intangibles. One might note, then, that 3 of the top 10 companies, Procter & Gamble (PG), IBM (IBM) and AT&T (T) have no tangible equity whatsoever; 60% of General Electric's (GE) book value consists of intangibles. Many might find that fact surprising. The market's price-to-tangible book value ratio is inordinately high.
As a frame of reference, Berkshire Hathaway (BRK.A)(BRK.B) trades at about 1.3x book value, so the price-to-book value ratio of almost every significant company in the S&P is enormously higher than that of Berkshire Hathaway. This does not represent merely a difference in pricing; it is a backdoor method to assess investors' earnings expectations for the stock market. This method can determine if those expectations are unrealistically high because the price-to-book value ratio is, to some degree, a forecast of return-on-equity, or ROE.
By way of explanation, shareholders' equity simply represents a company's net cash, property, equipment and other assets, including intangible assets such as trademarks, after deducting all liabilities. It is through the use of these net assets that a company earns its revenue and income. Some businesses require less in the way of net assets to generate a given level of earnings than others – perhaps less in the way of factories and more in the way of phones. As a generic example, if one personally owned a company with $1 million in shareholders' equity, let us say a manufacturer of a commoditized product such as plastic eating utensils, and if this business were to earn $50,000 by the end of the year, one's return on those net assets would be 5%; that would be your ROE. If that were the best that could be hoped for, perhaps you would prefer to close the business, considering all the effort and risk involved, and invest the $1 million in a municipal bond fund earning 7%.
If that were an insufficient trade-off, perhaps you would allocate that capital to the purchase of a better business, say one that manufactures surgical equipment. This new business also has $1 million of net assets, but in an average year produces a 20% ROE, or $200,000. If you could actually purchase that surgical equipment business for $1 million, or at a price-to-book value of 1.0x, you would be doing very well. However, it is unlikely that the owner would be willing to part with such a fine business at such a low price. Would you pay $2 million for it, or 2x book value? If you did, you would still earn $200,000, but on $2 million; this would be a 10% return on your investment, which is better than you were doing.
But you don't have $2 million available to invest, only the $1 million. So, perhaps you are willing to purchase a 50% interest in the business for $1 million, still paying a price of 2x book value ($1 million for $500,000 of shareholders' equity), and earning $100,000 (your $500,000 share of book value x 20% ROE) on the $1 million purchase price, and still be twice as well off as before. Would you pay 3x book value? The answer will be a choice each investor must make individually, but it is clear that one's willingness to pay a higher multiple of book value is a function of how profitable those net assets are expected to be. The market is effectively saying that most companies in the S&P will have considerably higher ROE than Berkshire Hathaway. Perhaps this is an implied rather than a conscious forecast, but it happens to be astonishing. It's also worth noting, in light of the S&P price-to-book value ratio of 3.68x, that the Japanese market traded at its historical high in 1989 at 5x book value.
An interesting question to ask is if the price-to-book value ratio is 2.85x that of Berkshire Hathaway, then will the average S&P company have 2.85x the ROE of Berkshire Hathaway?
There is another way to look at it. Standard & Poor's itself estimates the S&P 500 earnings for 2011 at more or less $96 on a unit value basis, and the index level is now about 1,325, suggesting that the index trades at 13.8x earnings (1,325 ÷ 96 = 13.8). This is within what would be considered a historically average range of valuation – neither obviously discounted nor expensive. Yet, quite a different picture emerges if those estimated earnings are viewed from an ROE perspective. The book value of the S&P 500 is about $360. Accordingly, to earn $96 on $360 of book value would require a ROE of 26.6% for the average company, which is an extraordinary number. Remarkably, the S&P produced about a 21% ROE in 2010.
Now it's time for another frame of reference. The highest ROE ever sustained for the longest time period by any company is, of course, that of Berkshire Hathaway at about 20% per annum. If the earnings estimates are to be believed, they would imply that the average S&P company is earning a return on its equity capital that is more or less consistent with the historical rate of return of Berkshire Hathaway, which is the greatest return there ever was.
The valuation risk quantified?
If one expects a 10% rate of return on the S&P over time, and if the time period in question is five years, and if, at the end of five years, the average price-to-book value ratio is close to 2.5x, which is a more normal level, then the current price of the S&P implies less than a 2% rate of return on the investment for the next five years. Thus, a necessary precondition for a successful 10% rate of return is that there should be no meaningful shrinkage in the current price-to-book value ratios.
There are, incidentally, more reasons why the S&P 500 companies will not sustain above-average profitability than reasons they might succeed in sustaining them. Stocks in general, and large-capitalization stocks in particular, have benefited from a number of factors over the past 50 years, which are not likely to be sustained or repeated in the future. The first is a declining tax rate following the World War II recuperation period. From 1952 to 1963 the corporate income tax rate was 52%. Since then, with the exception of 1968 and 1969, which included a 10% surcharge for the Vietnam War, the tax rate has been in fairly regular and significant decline. Given the current 35% rate, the decline has been 17% points. What would the historical S&P 500 return on equity have been in the absence of this benefit? Certainly, this phenomenon will not likely be repeated, and it might be partially reversed.
A second factor is the vast increase in government spending during and since the Second World War. In one way or another that spending has benefited the large companies. Those benefits could be in the form of transfer payments to individuals who use that money to buy products and services. Alternatively, it may be from money that is spent by the government itself when it buys products and services. In the current era of federal and municipal fiscal challenges, this phenomenon is likewise unlikely to be repeated.
The third factor is that increasing profit margins comprised a great part of the rate of return experience for the S&P companies over the last 50 years. A large part of that experience was enabled by the expansion and dominance of the largest U.S. companies, first domestically, then internationally. The international expansion was largely a function of the export of U.S. brands, from Coca-Cola (KO) to Disney (DIS) and Deere (DE), in a global economy in which geo-political constraints limited the free movement of a great deal of monetary as well as human capital. In today's global and digitally enabled economy, financial and labor capital are much more fluid, and the accompanying ability to source services and production globally is imposing competitive pricing and margin pressures on U.S. companies, not enabling margin expansion.
Since the beginning of the financial crisis of 2008/2009, there has been a substantial, even accelerating movement towards indexation and passive strategies. The desire is to avoid security-specific risk (any real estate stock can go to zero, but it's not likely that all of the stocks in a real estate exchange-traded fund ("ETF") will go to zero) and to develop tools and tactics to limit portfolio volatility.
The ability to transact on a momentary basis is a related factor, one that cannot be accomplished with a mutual fund (who wants to wait for 4 o'clock?), but can with an ETF. As recently as two years ago, in March 2009, the U.S. ETF industry had $485 billion of assets under management. That figure now exceeds $1.0 trillion, which is a 2.2-fold increase.
Of course, stock valuations have risen since then, but the ETF expansion may be compared with the experience of equity mutual funds, for which AUM rose by approximately 1.2x during the period, to a current figure of almost $6 trillion during the same period. Therefore, ETFs are now about 17% the size of the equity mutual funds, up from 10% two years ago. Moreover, during the first three months of this year, the flow of funds into equity ETFs increased by +$17.2 billion compared to +$30.5 billion into equity mutual funds. Given the much smaller base, that's over an annualized 8% increase for equity ETFs versus 2% for equity mutual funds.
This movement to indexation is impacting the valuation parameters of individual stocks. How so? Indexation by its nature favors the more liquid companies and disfavors the less liquid. This preference can be explained by viewing indexation as an asset-gathering business as opposed to an investment management business. Its object is not to outperform, per se, other indices or managers, but to gather the maximum volume of assets. However, the quantity of assets that any index can gather is self-limited by its least liquid member. An index is simply a fixed rule system by which a set of securities is selected for inclusion or exclusion. Therefore, when constructing one, the rules tend to be set so that companies with less share liquidity will be excluded.
Typically, indexes are float-adjusted to avoid being so limited. Companies with substantial inside ownership and hence, low float, are likely to be excluded. As well, companies that do not otherwise qualify for the rule-set, such as those that operate in multiple industries or in unconventional industries, or which manifest other idiosyncrasies would tend to be excluded as well. Thus, many companies, for reasons entirely apart from investment merit, will not participate in the flow-of-funds buying pressure of the expanding ETF industry.
Value, however, is usually found in the obscure companies that are less liquid and lack broad-based analytical coverage. This year, companies of that type are significantly underperforming the larger, more liquid companies. There is some empirical evidence of this phenomenon: A number of deservedly well-regarded mutual funds, which have superb long-term track records, each managed by value-oriented managers who pursue independent philosophies, are all underperforming the market this year. That's a very unusual set of circumstances, particularly given that these funds have almost no holdings in common.
Berkshire Hathaway is a not unrelated example. It might be considered a fund of sorts, because it is a collection of many assets, and it is certainly as well-diversified as any fund. It is also underperforming the S&P. Consider the tremendous value that Warren Buffett has amassed for Berkshire Hathaway through the financial crisis. At the end of 2007, the company had a stated book value of $78,008 per share and a share price of $141,600. Berkshire concluded 2010 with a stated book value of $95,453 per share, roughly 21% higher than at the onset of the crisis, yet a share price of $120,450, a decline of approximately 15%. Thus, the price-to-book ratio for the company contracted from 1.81x to 1.26x.
The valuation-altering impact of the ETF phenomenon works in the inverse as well. Not all the companies that qualify for inclusion in an ETF will necessarily be such wonderful companies. To illustrate the point, let's create a theoretical construct. Let's assume there exists an ETF with 100 members, which generates an attractive rate of return. The members in the ETF are extraordinarily good companies with one exception. That exception is an extraordinarily bad company, and comprises a 1% position in the ETF. This hypothetical ETF would attract a tremendous amount of assets and, as a consequence of its structure, shares of the one very bad company would be purchased as well. The bad company's stock would then trade at an unusually high valuation relative to its prospects simply because of its weight in the ETF.
The odds of that happening are actually fairly good; it's all a question of size and scale. The more ETFs available to investors in the market, the more possible combinations of portfolio securities will exist, and the probability of a dynamic like the one described in this example becomes greater. As purchasers, the ETFs are effectively indifferent to the price, because their models are all computer-driven. This situation has important implications.
One implication is that there must now be shares of companies that, despite poor or deteriorating businesses or business models, are supported simply because they are constituents of major equity indices. ETFs keyed to those indices are recipients of growing inflows of funds. This can go on for some time. Ultimately, though, investors who look to share-price performance as a precursor of business performance might one day find themselves painfully re-educated in the verity that whichever security or sector, technique or instrument of the day, becomes avidly popular becomes mispriced.
The other major implication is quite constructive for the so-called stock picker. Just as the rise of the ETF industry is distorting upwards the clearing prices of shares that are included in the underlying indices, so too is the ETF phenomenon distorting downward the prices of the non-standard company relative to its fundamental merits. This is not bad – it's good. Although the earnings or book value multiples of such companies might be too low, the underlying businesses are not injured. The high-ROE company that was compounding book value at a remunerative rate before the impact of the ETF phenomenon will continue to compound as before. The difference is that the shares can be now be purchased at a discounted valuation, which redounds to the benefit of the stock pickers – those who select for the valuation and risk/reward merits of each security as an idiosyncratic case study, without reference to formalistic constraints such as market capitalization, conformity with an index or other non-investment-based rule systems.
In particular, owner-operator companies of the type we have focused on increasingly this past year, yet which are much more often non-conformist in their structure or style of operation, have lost much of the constituency they justly enjoyed before the financial crisis. Accordingly, many of them are on sale – some on fire sale. Meanwhile, they continue to operate as they always have, making opportunistic use of discounted assets that litter the landscape, building their shareholders' equity. This should be a boon to our portfolios over the long term.
In various strategies, among the very largest positions, there are shares of companies that trade at a discount or even deep discount to book value or liquidation value (Henderson Land (HKG), Icahn Enterprises (IEP), Howard Hughes Corp. (HHC)), Global Logistic Properties) at P/Es of 10x or lower relative to current or normalized earnings, which is to say a 10% after-tax earnings yield in a 4% Treasury market (Jarden (JAH), AutoNation (AN)) at discounts to net asset value, in such cases that the company owns significant value in other publicly-traded companies (Loews (L), Jardine Strategic, Genting Berhad) at a small fraction of the valuation of comparable companies (Virtus Investment Partners (VRTS)) or merely at much too low a valuation multiple for the high profitability and quality of the business (Berkshire Hathaway (BRK.A)(BRK.B), Greenlight Re (GLRE)). This is not an all-inclusive list. Its purpose is merely to suggest the wealth of opportunity for building very salutary, likely idiosyncratic, returns in the coming periods. A casual review of these companies will show them to be, in their respective fashion, superior businesses. And they're on sale.
One interesting measure: As a group, these companies also tend to share a relative absence of Wall Street coverage. They are not selected for this measure – it is a symptom, not a causative factor. For simplicity, using the eight domestically traded U.S. companies mentioned above, the following table provides the number of Wall Street analysts who provide their year-forward earnings estimates to the various financial news distributors. Three of those companies have only one or no analyst providing estimates; one has as many as 12; the average is 3.75. As a comparison, the largest eight companies in the S&P 500 have over 26 analysts posting their earnings estimates, 51 for Apple. Granting that these are the spotlight companies, the S&P constituents numbered from 151 through 158 average more than 18 analysts each.
Since our founding in 1994, we have prided ourselves on investments in areas that are less covered and not easily understood. We believe today's investment landscape, driven by increasing commitment to indexation and ever-increasing complexity and choice, is impacting valuations and providing excellent opportunities for our long-term investors.