Semper Augustus Investments Group annual letter to clients for the year ended December 31, 2015; title, “Party Like It’s Nineteen Ninety-Nine.”
Bad Breadth, Intrinsic Value, And A Deep Dive Into Berkshire Hathaway
I was dreamin’ when I wrote this, forgive me if it goes astray;
But when I woke up this mornin’, could’ve sworn it was judgment day.
The sky was all purple, there were people runnin’ everywhere;
Tryin’ to run from the destruction, you know I didn’t even care.
Say two thousand zero zero party over, oops, out of time;
So tonight I’m gonna party like its nineteen ninety-nine - Prince
2015 went out with shades of 1999. The Federal Reserve raised interest rates, Luke Skywalker owned the box office, and a narrowing group of popular and increasingly overvalued stocks led the market to new highs, while the majority of stocks, including Berkshire Hathaway, lost altitude.
This year’s letter begins with an overview of the unfolding deterioration of breadth in the stock market. It compares the current bifurcation to the period leading up to the 2000 bubble, then discusses our intrinsic value model, a tool we created in March 2000 that demonstrates the degree to which the S&P 500 is nearly as overvalued today as it was then. We ruminate about a group of four highflying tech names known fondly as the FANG’s (Facebook, Amazon, Netflix and Google). We sense they will be licking major wounds going forward. Finally we examine Berkshire Hathaway, our largest holding, in detail so deep that it would put both Warren Buffett and Charlie Munger to sleep, permanently. We address last year’s 12.5% decline in the stock and what it means for expected returns going forward. The report describes multiple approaches we use to value the business. It is a lengthy analysis; so long in fact, that the competition’s drumming pink bunny can’t keep going and going. The bottom line: Berkshire is increasingly durable and the stock is considerably undervalued. We expect to earn 8% to 12% per year depending on multiple expansion for at least the next decade and a half holding Berkshire, and the shares may earn double the return produced by the S&P 500. Berkshire is trading at 70% of intrinsic value, giving it 45% upside from today, just to get to fair value.
Semper Augustus - Breadth Leading Up To The 2000 Bubble Peak, And Today…
Yogi Berra’s best malapropisms and expressions lacking logic included:
- The future ain’t what it used to be.
- No one goes there nowadays, it’s too crowded.
- You can observe a lot just by watching.
- It ain’t over till it’s over.
- When you come to a fork in the road, take it.
- It’s like déjà vu all over again.
Each of these “Yogiisms” fits today’s investment climate. The last is extremely apt here at the outset of 2016. The run-up to the bubble that burst in early 2000 involved such excesses and extremes that we really believed we’d never again see anything like it again. Well, history repeats, or at least rhymes, and in the words of Yogi, “It’s like déjà vu all over again.” Valuations and market behavior are frighteningly similar to what we saw in the late 1990’s, even more extreme in some cases. The number of stocks moving down has exceeded those moving ahead since May. EVERY major stock market decline since 1929 was preceded by a breakdown in breadth leading up to each market peak. While we saw such divergence in late 2007 through the fall of 2008, the episode leading up to 2000’s market peak is worth revisiting.
A Look Back at 1999
The stock market bubble peaked on March 10, 2000, with the NASDAQ closing at 5,048. The S&P 500 peaked two weeks later at 1,553, trading for about 40 times normalized earnings. The NASDAQ, dominated by high-flying tech and Internet shares, traded at an unbelievable 242 times earnings. We have fond memories of March 10, 2000, because from that day forward, for the first two years of the market collapse, which took the S&P down 50% and the NASDAQ by more than 80%, our stocks essentially went straight up. Under the surface, the market was breaking down long before March 10.
The late 1990’s were brutal for value-oriented investors. We launched Semper Augustus in late 1998 and managed to generate decent gains on the stocks in our portfolios, keeping pace with the S&P in 1999. But we were so far behind the 84% posted by the NASDAQ that the pressure to own and chase the tech names that were making everybody else “richer” was enormous. The first 70 days of 2000 were brutal. The markets, led by the insanely overvalued tech names, marched higher on a daily basis, while nearly every stock we owned was in decline. Investors liquidated real businesses with real value for the names creating arguably the greatest bubbles in capital markets history. Janus’ stock funds, all loaded with the same inflated tech names, were getting half of all of the money flowing into the entire mutual fund complex during the six months leading up to the March 10 peak. When the “markets” blew up, Janus blew up.
It took enormous effort to keep clients from chasing the tech bubble. We had cash on hand and as the stocks of real businesses got cheaper and cheaper, we put money to work at increasingly favorable prices. Several of our buys from that period we still own today. But it took every bit of energy we had to keep our clients on board and confident in the quality of businesses we were buying and in our favorable expectations, regardless of what the S&P or the NASDAQ did.
We penned a client letter in July 1999 titled, Large Cap Stocks Still Overvalued; Some Bargains in Small-Caps and Mid-Caps. The letter discussed a breakdown in breadth, the bifurcation in the market as a small handful of names drove the markets higher while the average stock lagged behind, many declining in price.
The blue chips, the GE’s, Coca-Cola’s, Wal-Mart’s and Exxon’s - the “new nifty fifty,” peaked at over 40 times earnings during 1998. Tech heavy indices continued to defy logic, narrowing through the remainder of 1999 and into early 2000. Tech, media and Internet stocks led the way toward the end of the advance. Everything unrelated to tech or telecom was in decline while tech sprinted ahead.
Our stocks managed to keep pace with the S&P 500’s 21% gain during 1999, but the NASDAQ was up an incredible 84%. We had lots of cash on hand, which muted our portfolio gains even more. As breadth deteriorated, we were able to put money to work. Toward the end of the bubble, in early March of 2000, nearly everything we owned fell. On the morning of March 10, the day of the peak, I was sitting at my desk literally shaking as Microsoft, Sun Microsystems, Lucent, Cisco and Pets.com screamed upward as our portfolio names bled red. To maintain sanity I wrote a piece for Alan Abelson at Barron’s which pointed out the market cap of the NASDAQ was set to pass that of the NYSE, yet the businesses on the NASDAQ generated less than 10% of the profits of those on the big board. The P/E ratio of the NASDAQ was 242 while that of the S&P was over 40. We had a portfolio trading for less than 15 times earnings, and it was falling fast on that day.
Then, it ended. Everything changed. The NASDAQ went red; my screen went green. We went on to post positive gains during the bear market that ensued, which took the S&P down by 50% and the NASDAQ by 80%. Late in the 2000-2002 rout, the Federal Reserve intervened, talking about “helicopter drops” of money and lowering Fed Funds to 3%. This sounds absurdly high by today’s standards. Thanks, Ben. Short-term rates stayed at 3% for several years until the real estate bubble fully developed and the S&P, at least, by 2007 had recovered all of its losses. Then, in 2007, a breakdown in breadth began again, setting the stage for the 2008-2009 crisis. The movie began anew…
Fast Forward to 2015
We could well have written our July 1999 letter today. Star Wars is back, dominating the box office, and the Federal Reserve just raised interest rates by ¼ percent. As it was then, stock gains in 2015 were uneven with much of the market in decline, a decline in breadth that began in May 2015. The average S&P 500 stock was down almost 4% last year, but the median stock finished 2015 down more than 22%. Half of the stocks in the S&P 500 finished 2015 down more than 20% from their 52-week highs. The Russell 2000 was off 5%, the Value Line Geometric was down 11% and the Dow Transports shed almost 18%. Uglier still, more than 70% of Russell 2000 and 68% of NASDAQ stocks finished the year more than 20% below their yearly highs. Profit margins, which peaked at an all-time record in the third quarter of 2014 at 10.1% (8.7% estimated for the fourth quarter 2015) for the companies making up the S&P 500, weren’t evenly distributed. As an extreme example, Apple contributed 22% of overall S&P 500’s margin expansion since 2009.
Financial academics define risk as volatility. That may be fine for theory, but for those of us who live in the real world, we define risk as a permanent loss of capital. The likelihood of risk using our definition is always highest at the point where the general perception of risk is lowest. The years 1929, 1937, 1966, 1972, 2000 and 2007 all marked moments when the perception of risk troughed, and the stock market peaked. The ensuing bear markets saw stock indices plummet anywhere from 50% to 89%. As markets plunge, the perception of risk rises, with fear peaking as markets hit lows in 1932, 1942, 1970, 1974, 1982, 2002 and 2009. Fear, and the perception of risk, peak at a market low, not at market highs. Only after suffering devastating losses do people believe stocks are “risky”. The reality is: we think 2015 will be added to the list above. But, which list?
The stock market is expensive. We said that last year, and the year before that and the year before that. Just because the stock market is expensive doesn’t mean prices can’t go up. Markets peak, however, following a period of time where the breadth of the market erodes, that is, where more stocks are declining than advancing. Yet, a narrowing group of popular leading stocks continues to make new highs. Breadth was positive and virtually straight up from the 2009 market low through May 2015.
These “bad breadth” episodes preceded every single market peak listed above. They also have lasted for anywhere from several months to two years. The three most extreme cases of divergence occurred in the years leading up to the 1929, 1972 and 2000 tops. For the holder of broadly diversified funds and indices, the deterioration in breadth is sending a warning shot over the bow. Coupled with extremely expensive valuations, risk is high.
Intrinsic Value In 2000 And 2016
By the end of March 2000, despite a massive reversal that began on the 10th of that month, we still had no objective evidence that the bubble had popped. We needed a tool to help clients understand that we owned a portfolio of undervalued quality businesses. We also wanted to demonstrate how overvalued the market was. Thus, we developed our intrinsic value report (March 31, 2000) and have regularly updated since then. The report highlights each holding, what we paid for it, its current price, and how we valued its intrinsic worth. The report contains our estimate of normalized earnings, and the P/E, earnings yield and dividend yield for each holding and for the portfolio as a whole. The earnings yield is essentially our base-case expectation for prospective annual returns over a long period of time. Each yardstick is measured on cost basis, current market value and on intrinsic value.
On top of expecting to reap the earnings yield, we further expect to earn any accretion of a discount to fair value that exists over some period of time (the market is very efficient over the very long haul). As stocks move up to fair value and are sold, we expect to earn an additional premium to the extent we are able to find additional undervalued businesses and repeat the process. Of course, you have to subtract for those mistakes we will invariably make, where any combination of lower business quality than envisioned, or developing poor economics, or simply erring on our appraisal of value or of lower actual growth than expected unfolds. Our expectation of prospective returns is:
Per the first March 2000 report, we owned a portfolio of stocks trading at 15.7 times normalized earnings, which gave us an earnings yield of 6.4%. Next, our holdings were trading at 83 cents on the dollar of fair value, which mathematically implied we should earn about 20% over some period of time as the valuation gap closed. An immediate closing would be big but the gap seems to take on average about five years to close. At ten years the accretion would yield us a little less than an additional 2% annually, on top of our 6.4% earnings yield. Further add a little premium for repeatability, subtract a little for mistakes, and we thus expected to earn about 8.5% to 9.5% per year over a long number of years, beginning at March 31, 2000. How accurate was the report? Our stocks indeed produced our expected return over more than 15 years, which lends integrity to the report, its assumptions and expectations.
So what did the intrinsic value report say about expectations for the market? On March 31, 2000, the S&P closed at 1499. Our estimation of intrinsic value was 590. We calculated $37.05 in normalized earnings, scrubbed for accounting aggressiveness, where trailing operating earnings were $48.26. Trailing reported dividends of $16.79 equated to a scant 1.1% dividend yield, a fraction of the return produced historically by dividends over the years. In fact, that was the lowest dividend yield on record. Using our normalized earnings number, the index thus was trading at a nosebleed 40.4 times earnings, which equated to a puny 2.5% earnings yield. Even using trailing operating earnings without our accounting adjustments the P/E was 31, giving the market an earnings yield of only 3.2%.
Our adjusted 2.5% earnings yield was thus the base-case expectation for annual market gains for a long period of time. Know that the investment world expected far, far more. Stocks produced a total return of more than 17% per year since the bear market low of 1982, when P/E’s were in single digits (7 times) and profit margins were extremely depressed (3% after-tax). Virtually nobody in the year 2000, save the small universe of investors grounded in price, believed stocks would only produce returns in the low single digits, or even losses, for far more than the next decade.
Our appraisal of fair value for the S&P 500 at March 31, 2000 assigned a 15.9 multiple to our normalized $37.05 in earnings, suggesting fair value at 590, demanding a not small 61% decline from 1499, or, as mentioned, requiring a whole bunch of years for underlying fair value to catch up with the inflated price.
We all know how the S&P fared subsequently. The index declined by 50% for the balance of 2000 through the low in late 2002. It then recovered by 2007, back to its former 2000 high, ballooned upward alongside the housing market with the Federal Reserve’s hot air. By late 2007, valuations were again very high, though not quite fundamentally as extreme as in 2000 (same index price but with 40% growth in GDP). The market then rolled over again during the “Great Recession”, seeing almost 70% of its price shaved away, bottoming at the demonic 666 level in early 2009. Enter Sandman; rather exit Greenspan and Enter Bernanke. With helicopter drops of monetary intervention, rounds of QE’s accompanied by matching monetary policy abroad, we saw the S&P blow through its former highs in 2013 and touch 2134 this past April.
Putting it all together, the S&P produced a total return of only 4.0% per year from March 31, 2000 through the end of 2015, barely more than our normalized 2.5% earnings yield. The market spent much of that time underwater. Only with a push from central banking’s quantitative easing did the market eclipse its 2000 and 2007 highs. Even with an ephemeral annual gain of only 4.0% per year, the market remains far from cheap. In fact, we still peg fair value below the 2000 and 2007 highs. The bear market that began more than 16 years ago has more work to do.
Using our intrinsic value report today, the index closed out 2015 at 2044. Trailing operating GAAP earnings (excluding write-offs and write-downs) will likely come in around $103 – down $10 or almost 10% year over year. Using this number, the market ended 2015 at 20 times earnings, which gives you an earnings yield of 5%. Not bad in a world of low interest rates. But we didn’t use GAAP earnings in 2000 and we don’t today.
The quality of earnings is again incredibly poor today. We normalize free cash profits at $70 for the S&P, 85% higher than our calculation in 2000. Nominal GDP grew by about 80% in 16 years. In addition to a poor quality of earnings, we also allow for some profit margin contraction over time from levels we think are too high. If our normalized number is more free cash correct, then the market is really trading for 29.2 times (2044/70), which equates to an earnings yield of 3.4%. That gets our long-range expectations for the market down to about what its done for the past 16 years.
A reasonable long-term multiple on profits is 15 times (6.7% earnings yield). Even if you don’t use our more conservative adjusted number, using GAAP earnings of $103, fair value for the index would be 1545, demanding a nearly 24% drop to fair value, or, as before, simply not making money for a whole bunch of years until fair value catches price.
Incidentally, an index price of 1545 gets you precisely back to the 2000 and 2007 highs. Using our more conservative $70 in normalized free cash earnings, fair value drops all the way to 1050, nearly half of today’s 2044 level.
The market forecast using our intrinsic value methodology is thus ugly. It’s not quite as grim as 2000 but still pretty bad. So how do we size up against the market now?
From our current intrinsic value report, our prospective advantage versus the index is nearly as great as it was in March 2000. Our stocks are trading for 12.1 times our calculation of normalized earnings, which gives us an earnings yield of 8.2%. Compared to the S&P 500’s GAAP earnings yield of 5.1%, or our more conservatively estimated 3.4%, our annual advantage is between 3.1% and 4.8%. We believe we have a huge long-term prospective advantage versus the market. Our report in March 2000 suggested that then, and it suggests it today. The report now has nearly 16 years of history behind it so it’s not merely a hypothetical. Again, the market is pretty darn efficient over the long haul, and those that appreciate that can benefit.
Of course, our stock portfolio didn’t get to 12.1 times earnings all of a sudden. Sixty percent of our long equity holdings were down last year. We owned 31 long equity positions during the year. Of those, 18 were down. Of the 13 that were up, one was sold entirely for a nice gain and would have been down had we kept it; and one was half sold and would have also been down otherwise. Our biggest position, Berkshire Hathaway, was down more than 12% last year, though it grew its profits and underlying value considerably. Our second largest holding, Exxon Mobil, was also down more than 12%, falling far less than the price of oil, but down nonetheless. Our third largest holding, Mercury General, was off almost 14% last year. The list goes on. This is most definitely the definition of declining breadth. Despite the price declines, on average our businesses grew their profits (even allowing for the notable exception of those who extract things from Mother Earth), which means the P/E multiple necessarily declined and our prospective earnings yield necessarily grew. The same thing happened in 1999 and early 2000, and also toward the end of 2007 and into 2008.
Combining asset prices in decline with underlying fundamental growth equals the expansion of value. How investors react to the unfolding of value dictates either long-term success or permanent failure. If you don’t expect to need all of your money for consumption for many years, then such short-term changes in price, particularly extant correspondent changes in value, shouldn’t matter. To the extent price deviates markedly from value, opportunity presents itself, whether as a buyer or as a seller. Understanding the difference between price and value is perhaps the most critical aspect of investing. Was the decline in Berkshire last year a loss? Was it giving back some of the gain from much lower cost bases? Does the decline indicate deteriorating permanent fundamentals inside the company and a reduction in long-term expected profits and gains? Is the decline in the shares really an opportunity to acquire shares, either by us or by management on our behalf at attractive prices? Does the decline increase our expected long-term return or lower it? Depending on how you answer those questions dictates long-term success or failure. It also necessarily dictates how well you sleep at night worrying about things beyond your control.
We spent the latter part of last year with our stocks in decline. It felt like late 1999 and early 2000. We watched a narrowing band of names; many of the tech variety, as well as a handful of branded consumer names, march ahead. Share prices in companies like Colgate, Clorox, Proctor & Gamble and Costco marched higher to very expensive levels, but none more so than a new era of tech favorites.