Investment, Speculation, Valuation, and Tinker Bell
March 18, 2013
by John Hussman
of Hussman Funds
Stocks remain in a mature – but not obviously complete – bull market, with the S&P 500 a stone’s throw away from its 2007 record high of 1565. In the midst of it all, investors face a confusing mix of arguments – some that stocks are significantly undervalued, some that they are dangerously elevated; some that sentiment is overly defensive, some that it is overly euphoric.
The most important questions investors should be asking are these: what do they know that can be demonstrated to be true; and what do they believe that can be demonstrated to be untrue. It is best to make these distinctions deliberately, lest the financial markets clarify these distinctions for investors later, against investors’ will, and at great cost. That’s not to say that understanding these distinctions will ensure investors the correct short-term stance in the market. More likely, though, it will help investors to distinguish good investment opportunities from poor ones. Even in the event the present bull market advance has a long way to go, it will also help to distinguish good speculative opportunities from poor ones.
Let’s begin with an observation. Not a prediction, but merely an observation. Last week, Investors Intelligence reported that the percentage of bearish investment advisors has declined to just 18.8%. The last time bearish sentiment was below 20%, at a 4-year market high and a Shiller P/E above 18 (S&P 500 divided by the 10-year average of inflation-adjusted earnings – the present multiple is 23) was for two weeks in May 2007 with the S&P 500 about 1525. The next instance before that was two weeks in August 1987 (bearish sentiment never dipped much below 27 approaching the 2000 peak except for a reading of 22.6 in April 1998, just before the Asian crisis). The next instance before that was for 3 weeks of a 5-week span in December 1972 and January 1973, which was immediately followed by a 50% market plunge. There are a handful of observations in March-May 1972 at a slightly lower level that were punctuated by a modest market decline before the final advance to the late-1972 peak (that lag is enough to discourage any near-term conclusions in the present instance). The instance before that was in February 1966, which was promptly followed by a bear market decline over the following year. You get the picture.
Again, these are observations. Frankly, I’m about as exhausted with such observations as you are. The problem is that particularly since March 2012, our estimates of prospective market returns (on a blended horizon of 2-weeks to 18 months) have moved from the worst 2% of historical observations, to the worst 1% of observations, to what is now such rarified air that only a handful of atoms remain. Still, markets are driven by beliefs, and so long as those beliefs are not challenged by anything to derail them – even if they ultimately prove incorrect – the markets may very well create their own reality for a while.
Speculation and Investment
We generally distinguish speculation from investment. For us, investment is buying a security at a price that is associated with a reasonable expectation of acceptable long-term returns. Speculation is buying a security in the expectation that its price will advance, with less regard for long-term prospects. Both benefit from solid valuation methods and reliable measures of market action, but investment weighs valuation more strongly, while speculation weighs market action more strongly. The problem for active market participants is to distinguish between investment and speculation in the first place, and then to identify good and bad opportunities to accept such positions.
The two adaptations that we’ve made to our approach in the recent market cycle relate to both investment and speculation. The introduction of ensemble methods would have made the greatest difference between 2009 and early-2010. While I correctly anticipated the credit crisis (see Critical Point for a reminder), the unwise response of policymakers – defend the bondholders, avoid debt restructuring, change accounting rules, extend, and pretend – virtually ensured years of economic headwinds, and led me to insist on making our approach robust to even Depression-era outcomes. During the Depression, the market dropped by half just to raise expected 10-year returns up to 10%, yet the market followed by losing two-thirds of its value from there. Even good trend-following models were subject to far more severe whipsaws and deep drawdowns than were observed in post-war data. Effectively solving that “two data sets” problem took some time. The second adaptation about a year ago deals with “speculative merit,” and increases the reliance on trend-following measures - even when our blended return/risk estimates are negative – but only in the absence of an overvalued, overbought, overbullish syndrome of conditions. The fact is that this syndrome of overextended conditions has persisted since March of last year (about 1400 on the S&P 500), so the impact of that adaptation will only be observed over time.
Barring some abrupt shock, I don’t expect that the market will “clear” its present overvalued, overbought, overbullish condition and simultaneously wipe out positive trend-following measures in one fell swoop. So it seems likely that we will observe a set of conditions in the reasonably near future where the absence of strenuously elevated conditions will be joined with still-positive trend-following measures. That would likely be a reasonable point to take a speculative exposure at controlled risk. In a richly valued market, that sort of risk control is most appropriately established using call options having a strike price situated at about the point where various trend-following measures would turn negative – what is known in finance as a “contingent position” because the position creates its own exit if the market deteriorates further without an interim recovery - and particularly if it deteriorates abruptly.
If this bull market has a long future ahead of it – which I strongly doubt, but which we do have to allow – there will likely be several appropriate points to establish a speculative position at controlled risk (speculative, because my view is that the long-term investment merit of stocks is quite weak here). The late-1990’s bubble provides an instructive example. Extremely overvalued, overbought, overbullish conditions emerged in the May-August 1998 time frame, but those conditions were then cleared – coupled with a shift to favorable trend-following measures by November 1998. That identified a reasonable opportunity to establish a speculative position at controlled risk, even in a market that was historically overvalued and on course to advance further. The next emergence of overvalued, overbought, overbullish conditions was in July 1999, about the 1400 level on the S&P 500, at which point the market began a series of 10% swings (e.g. the end of July 1999 and the second-half of January 2000) that gave even speculators some whiplash as the market approached its final peak in 2000.
From a speculative standpoint, then, what we presently observe is a rarified syndrome of overvalued, overbought, overbullish, rising-yield conditions that has typically resulted in profound market losses within about 2 years, but that doesn’t entirely rule out further speculative gains. In periods such as the late-1990’s where the market was overvalued and getting progressively more so, the best speculative opportunities usually appeared at points where overbought conditions were cleared but positive trend-following measures persisted or re-emerged. The other instances, where overextended conditions failed to periodically clear along the way, were ones that we associate with more abrupt market plunges like 1987.
From an investment standpoint, it’s important to recognize that virtually every assertion you hear that “stocks are reasonably valued” is an assertion that rests on the use of a single year of earnings as a proxy for the entire long-term stream of future corporate profitability. This is usually based on Wall Street analyst estimates of year-ahead “forward operating earnings.” The difficulty here is that current profit margins are 70% above the long-term norm. As evidenced by the entire span of available historical data, the elevation of profit margins is directly related – not only in overall level, but also in their point-to-point change over time – to the sum of government and household saving. See the analysis in Two Myths and a Legend to understand the strength of this relationship. The deficits of one sector must be the surplus of another.
What’s amazing to me is how aggressive and apoplectic some people have become when confronted with this data. Why get all upset over a fact? Government transfers have shot higher in recent years, and currently represent 21.5% of total consumption spending. If wage income is at the lowest share of GDP in history and people are still consuming, is it so difficult to believe that deficits in the household and government sector are driving the surplus in the corporate sector? Consider nearly 70 years of GDP history. The level of household and government savings as a share of GDP is strongly and inversely correlated with the level of corporate profits as a share of GDP (particularly after 2-4 quarters). The change in household and government savings as a share of GDP is strongly and inversely correlated with the change in corporate profits as a share of GDP (particularly after 2-4 quarters). Most important, the level of corporate profits as a share of GDP is strongly and inversely correlated with the growth in corporate profits over the following 3-4 year period.
Half of my day is typically spent in research, including testing the countless propositions about markets, valuation, trend-following, economic data, monetary policy, and profit margins that are spouted out by Wall Street observers. We’ve found no convincing data that international activity explains elevated profit margins or ensures their durability. Productivity growth does not explain them. There is a clear, logical, economically sound, and empirically provable reason why profit margins are elevated here. That evidence is only uncomfortable because it implies very uncomfortable conclusions – namely that corporate profits are likely to weaken significantly in the coming years even if the economy expands.
As I noted two weeks ago, I do believe that some of the more difficult challenges for the U.S. economy could be avoided through the combination of efforts to create incentives for real investment and R&D, to restructure underwater mortgage debt, and to increase capital requirements at too-big-to-fail institutions (ideally in the form of mandatorily convertible debt). But none of these appear to be on the horizon.
So objections aren’t enough. Those objections have to be supported in the data. Not just an argument or a theory, but data – decades of it. Warren Buffett was correct about profits as a share of GDP in November 1999 (and it would be useful if he remembered today how correct he was): “In my opinion, you have to be wildly optimistic to believe that corporate profits as a percent of GDP can, for any sustained period, hold much above 6%… Maybe you’d like to argue a different case. Fair enough. But give me your assumptions. The Tinker Bell approach – clap if you believe – just won’t cut it.”
The only way to adequately gauge investment merit here is to have a valid and historically reliable approach for estimating prospective future market returns. What is most uncomfortable about the present market environment is that even some people whom we respect are tossing out comments about market valuation here that are provably wrong, or at least require one to dispense with the entirety of historical evidence if their optimistic views are to be correct.
Again, the Tinker Bell approach won’t cut it. Before you accept someone’s view about market valuation, examine the data – decades of it. Ignore clever-sounding valuation arguments that don’t have a strong, consistent, and demonstrated relationship with subsequent market returns.
Wall Street is placing a pathological over-reliance on a single year of forward operating earnings as a complete summary of future corporate prospects, without any adjustment for the level of profit margins. In my view, this will be the “hook” that ultimately yanks the hope of long-term financial security from countless investors who were pulverized by the 2000-2002 and 2007-2009 declines – and believe that now is somehow the right time to get back in because Alan Greenspan said so on CNBC. As one observer put it, “This is like getting a lecture on seamanship from the captain of the Titanic.”
To put some structure on this, let’s bring together a number of valuation models that I’ve constructed and presented in these weekly comments over time – all are straightforward to calculate. In practice, we use a broader range of fundamentals and growth considerations, but the combination of these models has a strong record in estimating subsequent total returns in stocks. You can see the various references for an explanation of each, but the basic approach is to estimate likely capital gains plus average dividend income over a 10-year period.
I should note that in each of these models, we’re assuming a long-term growth rate for cyclically-adjusted earnings, revenues, dividends, nominal GDP and so forth of about 6.3% annually. Historically, that has entailed about 3.2% real growth and just over 3% inflation. Last week, Jeremy Grantham provided a strong argument that long-term real economic growth is likely to average less than 2% going forward, largely because population growth is about 1% lower than it has been historically, because much of what we count as “growth” reflects increasing dead-weight costs of resource extraction, and because there is no evidence to expect a compensating increase in productivity growth (in fact, real gross domestic investment has been falling as a share of GDP, which tends to push long-term productivity growth lower).
So I’m not sure it’s even reasonable to assume 6.3% nominal GDP growth going forward, unless most of it is inflation (in which case the prospective nominal return may be the same, but the prospective real return would be lower). Still, the estimates of prospective 10-year S&P 500 returns below stick with a 6.3% long-term economic growth assumption, because I want to emphasize that prospective market returns are dismal even on optimistic economic assumptions.
Hussman's Shorthand Estimates of S&P 500 Expected 10-Year Nominal Annual Total Returns
1) Forward Earnings Model – see Valuing the S&P 500 Using forward Operating Earnings:
(1+g)(12.7 / FOPE)^(1/10) - 1 + Dividend_yield*(FOPE / 12.7 + 1) / 2
g = 1.063 x (0.072 / (FOE/S&P 500 Revenues))^(1/10) - 1
FOE: forward operating earnings. FOPE is the forward operating P/E.
Underlying economic growth is modeled at 6.3% (all of the constants should be revisited over time)
2) Shorthand Shiller Model – see The Siren’s Song of the Unfinished Half Cycle:
1.063 * (15 / ShillerPE)^(1/10) – 1 + Dividend_yield (decimal)
3) Dividend Model – see Estimating the Long-Term Return on Stocks (1998):
1.063 * (Dividend_yield / .037)^(1/10) - 1 + (Dividend_yield + .037)/2
4) Shorthand Market Value / GDP Model – new! fun! because I care:
1.063 * (0.65 / (MV / GDP_Nominal))^(1/10) - 1 + Dividend_yield (decimal)
Market value of equities of nonfarm nonfinancial companies from Z.1 Flow of Funds data
The average correlation between these estimates and subsequent 10-year S&P 500 total returns is 84%. Presently, the average estimate of prospective S&P 500 nominal total returns is just 3.6% annually for the coming decade.
I want one thing to be very, very clear. Today is not 2009, it is not 2003, it is not 1982, and it is not 1950. While the period since the late-1990’s has enjoyed a couple of bright spots coming off of the 2003 and 2009 market lows, the overall period has produced dismal total returns for buy-and-hold investors. It should not be surprising why this is so, and it should not be surprising why this will continue to be so for long-term investors until valuations normalize.
On Friday, Alan Greenspan appeared on CNBC with the competing claim that “The basic way of looking at the degree of exuberance or non-exuberance is to take a look at what we call the ‘equity premium.’ As you know, it’s an extent of a measure of whether stocks are overvalued or undervalued. And right now, by historical calculation, we are significantly undervalued.”
The “equity premium” is often referred to as the “Fed Model” – calculated as the difference between the earnings yield on stocks (forward operating earnings divided by the S&P 500 Index), and the 10-year Treasury bond yield. The “equity premium” is also sometimes estimated by setting the stock return component to something like the dividend yield + 6.3% and then subtracting the 10-year Treasury yield. I’ve included both versions in the chart below.
The problem with the Fed Model is that it presumes a one-to-one relationship between those two yields, an assumption which is an artifact of the disinflationary period from 1982 to 1998. Indeed, prior to 1970, earnings yields and bond yields had a strong inverse relationship (even in post-war data). The same is actually true of dividend yields. What investors may not realize is that the correlation between interest rates and earnings yields (as well as dividend yields) has also been negative since late-1990’s.
While “operating earnings” are not even defined under Generally Accepted Accounting Principles, and “forward operating earnings” only surfaced as a creature of Wall Street in the early 1980’s, it’s simple enough to impute historical values of forward operating earnings because they are almost completely explained by observable earnings and employment data – see Long-term Evidence on the Fed Model and Forward Operating Earnings. If you review that weekly comment, it should be clear that the relationship between the Fed Model and subsequent market returns is surprisingly weak.
Notice that the date of that Fed Model piece was August 2007. That alone is instructive, because the arguments that people are making today that “stocks are cheap on the basis of forward operating earnings” are the same ones they were making in 2007. It speaks to an incredible failure of Wall Street to learn from its mistakes (or at least the mistakes that it foists onto unsuspecting investors), and the likelihood of making nearly the same mistake for a third time in just over a decade.
The chart below presents the two versions of the “equity premium” along with the annual total return of the S&P 500 over the following decade. To say the least, the relationship is less informative than the other models we’ve considered above. Importantly, the relationship is nearly as bad even if these “equity premiums” are compared with the difference between the realized 10-year S&P 500 total return and the 10-year Treasury yield (to get a true “excess” return).
By the way, the correlation of “Fed Model” valuations with actual subsequent 10-year S&P 500 total returns is only 47% in the post-war period, compared with 84% for the other models presented above. In case one wishes to discard the record before 1980 from the analysis, it’s worth noting that since 1980, the correlation of the FedModel with subsequent S&P 500 total returns has been just 27%, compared with an average correlation of 90% for the other models since 1980. Ditto, by the way for the relationship of these models with the difference between realized S&P 500 total returns and realized 10-year Treasury returns.
Still, maybe the Fed Model is better at explaining shorter-term market returns. Maybe, but no. It turns out that the correlation of the Fed Model with subsequent one-year S&P 500 total returns is only 23% - regardless of whether one looks at the period since 1948 (which requires imputed forward earnings since 1980), or the period since 1980 itself. All of the other models have better records. Two-year returns? Nope. 20% correlation for the Fed Model, versus an average correlation of 50% for the others.
So when you hear arguments that the “equity risk premium” is wonderful, or that “stocks are cheap on the basis of forward operating earnings,” understand that you are being fed a very thin gruel. Most likely, investors face a prospective 10-year total return on the S&P 500 of about 3.6% annually, with the prospect of extremely high volatility along the way, while the 10-year Treasury yield comes in at 2.0%. Given that stocks have dramatically greater volatility than 10-year Treasury bonds, an apples-to-apples comparison is probably inappropriate. Suffice it to say that per unit of risk, the present equity premium is almost undoubtedly negative.
Meanwhile, it’s worth observing that all of our valuation models have performed quite nicely over time, including the past decade, and including 2009. To clarify why I remained defensive, I’m going to say this again. The 2009-early 2010 period reflected a one-time stress-testing response because we were forced to contemplate Depression-era outcomes. We addressed that, which I expect to be very evident over the coming market cycle, but there’s no time-machine available to recapture the interval that was required to solve that “two data sets” problem. On the other hand, my recent defensiveness – particularly since March 2012 (near 1400 on the S&P 500) – is a defensiveness that is likely to be repeated in any future mature bull market advance where conditions become as strenuously overvalued, overbought, and overbullish as they are today. It has gone on much longer than we would like, but we aren’t about to abandon our long-term discipline over it.
The argument here is not that stocks will decline immediately. I don’t anticipate a long continuation of the recent bull market advance, but that possibility also can’t be ignored. For long-term investors, valuations are already at levels that imply 10-year total returns of only about 3.6% annually for the S&P 500, and the likelihood of that return being achieved without wide market swings along the way is close to zero. If this advance continues, I would expect to see reasonable opportunities for speculative exposure at contained risk, but it is a bad idea to chase market risk here, under conditions that have generally produced dismal outcomes throughout history.
No market condition is permanent, and even the late-1990’s advance included periods where favorable trend-following measures were not joined by hostile syndromes of overbought, overbullish conditions. If you believe that stocks will continue to advance in the months and years ahead, with no intervening bear market decline, those instances are the main points where “don’t fight the trend” might outweigh negative return/risk considerations more generally. For longer-term investors, consider the prospective return you can expect to achieve over time if you are buying, and that you can expect to forego if you are selling. Compare this with your tolerance for volatility, missed short-term returns, and deep interim losses. All of this is what I know and believe. It’s fine to believe something else. But please – insist on supporting evidence and long-term data. The Tinker Bell approach just won’t cut it.
The foregoing comments represent the general investment analysis and economic views of the Advisor, and are provided solely for the purpose of information, instruction and discourse. Only comments in the Fund Notes section relate specifically to the Hussman Funds and the investment positions of the Funds.
As of last week, market conditions remained characterized by a syndrome of overvalued, overbought, overbullish, rising-yield conditions that have historically been hostile for stocks, on average. Investors Intelligence reports that the percentage of bearish advisors has declined to just 18.8%. There are only a few historical points when bearish sentiment below 20% was joined by a 4-year high in the S&P 500 and a Shiller P/E over 18 (S&P 500 divided by the 10-year average of inflation-adjusted earnings – the present level is 23). They are 2007, 1987, 1972 and 1966 – all prior to significant bear market declines, though the market drifted a few percent higher over a 6-month period in the 1972 instance.
This instance may be different, and we have to allow for the possibility of continued market gains. Even in that event, however, I believe that a better opportunity to accept a speculative position at controlled risk would be at a point where the absence of overextended conditions (overvalued, overbought, overbullish) is coupled with the presence of favorable trend-following measures. One example of risk-control is to use call options, placing the strike price at about the level where a breakdown in trend-following measures would be expected to occur, which provides a natural “contingent exit” if prices deteriorate further, particularly if they deteriorate abruptly.
In general, we try to align our investment stance with the market return/risk profile that we estimate at each point in time. The selective willingness to act on trend-following measures even when our market return/risk estimates are negative (but only in the absence of hostile syndromes of market conditions) is one of the two adaptations we’ve introduced to our investment approach in the recent market cycle. Once valuations are rich and our broad return/risk estimates are negative, our willingness to accept market risk generally requires a window with two exits – one below, at the point where the trend-following measures deteriorate, and one above, at the point where overvalued, overbought, overbullish conditions emerge.
We have not had the opportunity to use that window over the past year, but that could change if the recent bull market advance continues for a longer period, mainly because uncorrected diagonal price ramps rarely persist. On our measures, market conditions deteriorated to the most negative 2% of historical data back in March 2012 (near 1400 on the S&P 500), and have become progressively more extreme, to the point where we can count similar instances on one hand. At the same time, the perception of investors seems to be that quantitative easing is sufficient to obviate every other consideration, and it’s possible that this perception will continue to create its own reality for a while. In my view, our best response is to maintain our discipline, accept limited-risk speculative opportunities if and when they emerge, and to rely neither on a severe market decline nor on the absence of one.
Notable item over the weekend - a European bailout deal for banks in Cyprus now includes a haircut provision. But not for bank bondholders. Of course not for bank bondholders. No - it provides for a haircut on depositors that is being called a "stability levy" amounting to 9.9% on deposits over 100,000 euros, and 6.75% below that level, exchanging their deposits for shares of stock in those teetering banks. So insured bank deposits are now effectively subordinate to uninsured European bank debt. It will be interesting to see how that works out. Alan Greenspan suggested on Friday that there has been a "removal of tail risk" from the global financial system. I doubt it, but we'll take the data as it arrives.
Presently, Strategic Growth Fund remains fully hedged, with a staggered-strike position that raises the strike prices of the Fund’s index put options somewhat closer to market levels, representing about 1% of assets looking well into springtime. Meanwhile, day-to-day differences between the performance of the stocks owned by the Fund and the indices it uses to hedge are also a source of fluctuation in Fund value. Strategic Dividend remains hedged at about 50% of the value of its stock holdings. Strategic International remains fully hedged. Strategic Total Return continues to carry a duration of about 3 years in Treasury securities (meaning a 100 basis point move in interest rates would be expected to impact Fund value by about 3% on the basis of bond price fluctuations), with about 10% of assets in precious metals shares, and about 5% of assets in utility shares.
Prospectuses for the Hussman Strategic Growth Fund, the Hussman Strategic Total Return Fund, the Hussman Strategic International Fund, and the Hussman Strategic Dividend Value Fund, as well as Fund reports and other information, are available by clicking "The Funds" menu button from any page of this website.
Estimates of prospective return and risk for equities, bonds, and other financial markets are forward-looking statements based the analysis and reasonable beliefs of Hussman Strategic Advisors. They are not a guarantee of future performance, and are not indicative of the prospective returns of any of the Hussman Funds. Actual returns may differ substantially from the estimates provided. Estimates of prospective long-term returns for the S&P 500 reflect our standard valuation methodology, focusing on the relationship between current market prices and earnings, dividends and other fundamentals, adjusted for variability over the economic cycle (see for example The Likely Range of Market Returns in the Coming Decade and Valuing the S&P 500 Using Forward Operating Earnings ).
© Hussman Funds