John P. Hussman, Ph.D.
In recent weeks, I’ve emphasized that our estimate of prospective market return/risk in stocks has slipped into the most negative 0.5% of historical data (reflecting a range of horizons from 2 weeks to 18 months). Last week that estimate actually deteriorated, but I am reluctant to make comments on such a small sample, as the only more negative estimate in post-Depression history was on September 16, 2000. Even in the conditions that match the worst 2% of our return/risk estimates (which is the part of the tail we have been in since late-February), the market has lost an average of 20-25% just in the following 6-month period. As much as I try to maintain equanimity – focusing on the average outcome of a particular set of market conditions rather than the specific instance at hand – it is very difficult to do so at present.
The green bands in the chart below depict all of the points since 1980 in the neighborhood of present conditions – having a nearly similar prospective return/risk profile, coupled with a particularly hostile « exhaustion syndrome » that has been a hallmark of the worst market outcomes in recent decades. The blue line shows the S&P 500 Index. As I noted in Goat Rodeo, « what this combination picks up is an already fragile set of market internals that has enjoyed an ‘exhaustion rally’ that both exceeds earnings growth and is met with overbullish sentiment. »
I usually show longer-term charts, but there are no green bands prior to 1987. Before that point, valuations were never been as extended as they are today – on the basis of normalized earnings – except in the quarters leading up to the 1929 crash. Exhaustion syndromes prior to 1987, while still very hostile to stocks, didn’t occur in valuation conditions as rich as we have today. It’s worth noting that there is a very narrow band in 2006 that was followed by a decline of only a few percent, but even the seemingly benign instances in 1998 and early 2000 represented losses exceeding 10%. I suspect we’re at risk of something far more significant. Importantly, the drivers of our market risk estimates are largely independent of our measures of recession risk. This may provide some insight into why my concerns have become so strident in recent weeks.
Present market risks involve a confluence of factors. First, valuations remain unusually rich. Though prospective returns are better than at the 2000 and 2007 peaks, valuations remain more elevated than at any point prior to the late-1990’s bubble, save for the period before the 1929 plunge. Notably, valuations only seem « reasonable » on the basis of « forward operating earnings » if one ignores the fact that profit margins are 50-70% above historical norms, and are dependent on unsustainably large fiscal deficits and depressed household saving in order for that to continue (see Too Little to Lock In).
Second, market internals have deteriorated, with an uncomfortably familiar « two-tier » profile developing between a handful of speculative momentum stocks and the broader market. Coupled with an active new issues calendar, near-panic levels of selling by corporate insiders, heavy beta exposure among mutual funds and institutional managers, record-low mutual fund cash levels, and advisory bearishness at just 20.5% (a level last seen before the steep 2011 decline), there appears to be a lopsided exposure to risk among speculators, and a divestment among issuers.
Third, as I’ve frequently noted, the best way to extract meaningful signals and reduce noise in volatile data is to draw those signals from the joint behavior of several indicators. While these methods range from the simple to the complex, we’ve frequently presented variously defined sets of indicators (see An Angry Army of Aunt Minnies) that capture some particular investment environment (such as an overvalued, overbought, overbullish market where favorable drivers have begun to drop away). In recent weeks, we’ve seen several of these hostile syndromes emerge, as I’ve detailed in prior weekly comments.
Finally, though our overall assessment of market return/risk prospects is largely independent of our macroeconomic concerns, the joint deterioration in the growth of real personal income, real personal consumption, real final sales, and employment, coupled with our inference of leading economic pressures from « unobserved components » methods, creates not only the concern but the expectation that the U.S. economy is entering a recession – not a quarter or two from today, but most likely at present. Indeed, Europe already appears to be in a broadening recession, which the U.K. has now joined, and the confluence of economic weakness and already strained government debt conditions in Europe is likely to produce disruptive outcomes in the coming quarters.
Addicted to Hopium
Again last week, some comments from economist Martin Feldstein were worth noting, particularly in contrast to the « hopium » sprinkled around by so many Wall Street analysts.
« The big problem for Spain is not its trade and current account deficit – that’s relatively small. The big problem there is how they’re going to finance both their fiscal deficit and the rollover of the debt that’s coming due. Those two together will be some 20% of GDP of Spain this year. So that’s not something that’s going to be easy to do when the banks are under tremendous pressure, and when foreign bond-buyers of Spanish sovereign debt are walking away.
« We’ve now had QE2 and after that we had Operation Twist, and both of these have helped to lower long-term interest rates and boost the stock market a bit. But they’re not doing anything to help the real economy. They’re not doing anything for economic activity. Housing remains in a slump. Businesses are not investing. Individuals have been reluctant to spend. It helps the financial markets, and that’s why they are cheering for it, but I don’t think it’s doing anything for real economic activity.
« The state of the economy is quite poor. There was a little burst of enthusiasm in the beginning of the year, but that’s come off as job growth fell very sharply in the most recent month, as real incomes are falling, real wages are falling, house prices continue to fall, so it’s not a very pretty picture. There was talk of 2 1/2% and 3% GDP growth this year, but I think we will be lucky if we get as high as 2%. Last year, we only had 1 1/2%. If you look at the first quarter of this year, while the GDP number was 2.2%, almost all of it – nearly three-quarters of it – was automobile purchases; catching up for last year’s shortages of automobiles. So it doesn’t suggest that there is strong consumer spending on ordinary goods and services, and certainly there isn’t on construction and on business investment spending. »
These observations might be dismissed as simply one economist’s opinion, if it weren’t for the fact that Feldstein is the former president of the National Bureau of Economic Research (NBER) and remains a member of its Business Cycle Dating Committee, which is the official arbiter of U.S. recessions. Importantly, the Committee does not define a recession simply as two consecutive quarters of negative GDP growth. Instead, it looks more broadly for a « significant decline in economic activity » that is spread across the economy. In defining economic activity, the Committee « examines and compares the behavior of various measures of broad activity: real GDP measured on the product and income sides, economy-wide employment, and real income. The Committee also may consider indicators that do not cover the entire economy, such as real sales and the Federal Reserve’s index of industrial production (IP)… a well-defined peak or trough in real sales or IP might help to determine the overall peak or trough dates, particularly if the economy-wide indicators are in conflict or do not have well-defined peaks or troughs. »
Again, we’re already seeing considerable weakness in year-over-year growth of real income, real final sales, and real personal consumption. The fact that industrial production posted a decline in March is also notable, though there’s no clearly-defined peak as yet. From April’s tepid employment report and other data, my impression is that the « gradual but accelerating » weakness we’ve expected is proceeding on course, though it is unlikely that the economy entered a recession in April. From what we infer from a variety of methods, particularly « unobserved components » models of the economy, I suspect that the economy will slip into recession in May or June (yes, this month or next). If that is the case, I would also expect that we’ll see a fairly rapid and abrupt acceleration in new unemployment claims in the weeks ahead (watch for upward revisions in prior data as well). Something to watch closely.
I recognize that our Recession Warning Composite, which we relied on to identify the oncoming recessions in 2000 (see our October 2000 Economic Perspectives note) and again in 2007 (see Expecting A Recession), has gone « quiet. » This is largely because the S&P 500 has advanced in recent months, while the other components remain just a hair from their respective triggers. As I noted last week, however, this composite tends to produce a signal only after the market is down 10-15%. Though it’s usually the case, as it was in 2000 and 2007, that recession risk is still widely dismissed at those points, and the market continues to decline substantially, the Recession Warning Composite is not intended to gauge stock market risk. Of course, in the past two years, the economy has twice come to the brink of recession, only to be pulled back slightly from the cliff by enormous monetary interventions. We can’t rule that out in this instance, but it’s notable that the real economic effects of these interventions have become progressively smaller even though the interventions have remained massive. The pattern of diminishing returns isn’t promising.
As a side-note, we don’t fear another round of QE, if that’s what emerges. With regard to our investment strategy, I discussed our basic roadmap in the event of QE several weeks ago: « If the Fed indeed steps in with an additional round of QE, a few distinctions may be helpful. First, regardless of Fed actions, and even in the past few years, the market has invariably suffered significant losses following the emergence of the ‘overvalued, overbought, overbullish, rising-yields’ syndrome that we presently observe. In contrast, the main window where it has not paid to ‘fight the Fed,’ so to speak, has been the period coming off of oversold lows… Strategically, then, we concede that there is some latitude to ease back on defensiveness between the point where QE induces an early improvement in market internals and an upturn in various trend-following indicators (coming off of a previously oversold condition), and the point where an ‘overvalued, overbought, overbullish, rising-yields’ syndrome is established. But once that syndrome is established, it is unwise to ignore it, and a defensive stance becomes essential (as we saw separately in 2010 and 2011, not to mention at most major market tops over history). »
Meanwhile last week, Lakshman Achuthan of the Economic Cycle Research Institute reaffirmed his recession expectations in an interview on Bloomberg. Achuthan expects a recession to begin by mid-year, which is consistent with what we’re inferring from the data:
« The data coming in is confirming that the slowdown is going ahead in lockstep fashion. The only holdout the last time we talked was job growth, it hadn’t rolled over yet. Year-over-year jobs growth peaked in February, and now it’s fallen in a way that – in the last 60 years – during a slowdown, if we see the year-over-year jobs growth rate fall as it has now, that’s ended in a recession. Year-over-year jobs growth – the size of the decline we’ve seen in that jobs growth, in the context of a slowdown, which we already see in GDP and broad sales and income – that is consistent with a recession over the last 60 years. And personal income is a real weak spot.
« [The NBER is] doing the official definition of recession – there is no other definition at the end of the day. You typically see negative GDP, and I think we’re going to get it this time around. But let’s be clear. In the last 6 recessions, it took over 6 months until you saw your first negative GDP print after the recession started. So here we are saying that there is a recession that’s going to start by mid-year, but I don’t think that you’re going to get hit over the head with – it’s going to be obvious to everybody – until after the recession has begun. So it won’t be until the end of the year, I think, until people start to figure out – huh, something happened. Same thing the last couple of recessions. »
« It took more than a year to learn that GDP actually shrank by 1.3% during the first quarter of the 2001 recession. But back then it was initially reported as having grown at 2.0%. That’s not very different from the latest reading for GDP growth in the first quarter of 2012 – 2.2%. In August 2008, just before the Lehman collapse, GDP was reported to have risen in the first and second quarters with the latter revised up sharply, triggering over a 200-point rally in the Dow that day. Today we know that GDP actually shrank in the first quarter while the second has been revised down by two full percentage points. »
For investors, I should note that by the time people said « Huh, something happened » 6 months into recessions in 2001 and 2008, the market had already lost about 20-25% in both cases, was entering free fall, and the best opportunity to reduce risk was long gone.
While there’s no doubt that Feldstein, Achuthan, and I all differ in the specifics of our approach to economic analysis, it should be clear that we all look at many of the same measures too. Why? Because they are the measures that effectively define recession. Also, notice Achuthan’s phrase « in the context of a slowdown. » This is important, because he recognizes that you don’t infer recession risk just from one indicator, but rather from the overall syndrome of evidence (which is why words like « jointly » and « uniformity » appear so frequently in our analysis, and why we try to use multiple sensors whenever noise-reduction or signal-extraction is required).
Suffice it to say that our recession concerns remain intact, as do our separate concerns about extreme stock market risk.
How to Create a Banking Debacle
Friday’s revelation by JP Morgan of an unexpected $2 billion trading loss simply reinforces our concerns about the leverage and risk exposure of the global banking system. Though the loss wasn’t particularly large relative to JPM’s total equity, the stock took a hit on broader concerns about risk management and the possibility that the problem was not isolated. We don’t draw a great deal of information from this particular loss. JP Morgan is simply a reminder that the risks taken by bloated institutions both in the U.S. and in Europe are far more leveraged and far less transparent than is consistent with a well-functioning banking system.
When I was a professor at the University of Michigan, I used to tell my MBA students that the best way to avoid seeing the word « debacle » next to their name in the Wall Street Journal was to avoid situations involving a « leveraged mismatch with a lack of transparency. » Those features are at the heart of nearly every major financial crisis: a) leverage – taking positions with large amounts of borrowed money, which magnifies the impact of errors; b) mismatch – long and short positions in risky securities that don’t closely offset, and; c) lack of transparency – organizational structures or accounting rules that prevent positions from being monitored and marked-to-market. A fourth feature, d) a sudden absence of liquidity – is often the event that triggers a debacle, but the root of the problem is almost always a highly leveraged, mismatched, opaque position that was there in the first place.
This combination is how Nick Leeson single-handedly brought down Barings bank. It’s why AIG and Lehman failed, and why Bear Stearns needed to get an illegal bailout from the Federal Reserve (which created the Maiden Lane shell companies to buy its bad assets). Even seemingly precise risk calculations, when based on hugely leveraged, mismatched positions, are not enough to avoid implosion, as the geniuses at Long Term Capital Management found out. Without the transparency of mark-to-market accounting and regular position audits, there’s no telling what the assets are.
Institutions often argue that their blowups can be traced to a « sudden lack of liquidity, » as if they are blameless and only need the Fed to provide a constant backstop, or to change accounting rules to eliminate mark-to-market requirements. To some extent, one could argue that MF Global would have been able to continue quite a while longer had it not faced a sudden lack of liquidity, or that European banks will be just fine as long as nobody moves their deposits, but this misses the point. Highly leveraged, mismatched, opaque positions, and the regulatory and supervisory culture that allows them – not a lack of liquidity – are what ultimately produce financial crises.
On the subject of regulation, the proper way to run a bank stress test is not to ask « Gee, what do you think would happen if we go into a recession? » Nearly every institution will answer, « Don’t worry, we’ve got it covered. » No. A stress test that requires that high a level of subjectivity isn’t a stress test at all. Instead, you ask how sensitive each piece of the portfolio is to fluctuations in its respective market, you ask what happens if you reduce the assumed correlation of « cross-market » hedges (hedges where the risk in one market is assumed to be offset by a position in a completely different market), you ask what happens if you increase the correlation of assets that are assumed to be uncorrelated (which can reveal actual risks well beyond what is assumed), you shock the position by several standard deviations (assuming particularly large shocks if valuations are well outside of historical norms), you estimate what those losses would be, and then compare that to the amount of regulatory capital that the institution has, on a mark-to-market basis. The stress testing exercise run by the Federal Reserve was a travesty in that regard – right up there with believing the promises of 5-year olds to control themselves as they scramble past each other through the gates of Candyland.
All of this is a very good argument in support of a strong Volcker Rule – separating banking (where depositors are protected by the government) from investment activities (which presently gets the government umbrella in case things go wrong, and leads to reckless behavior as a result). It is an even stronger argument against massive enterprises of the size of JP Morgan, Citigroup and Bank of America. Government-protected banking should almost by definition be a conservative business. Investment banks should be unprotected, but still subject to the same resolution authority that is applied to insolvent banks. Investors can take risk with them on that basis. What should never happen is for protected deposits and Fed-provided liquidity to be used as a back-door way to get cheap capital and subsidize speculative trading under government protection, as JP Morgan just showed is too great a temptation to resist.
Economist Simon Johnson asked a good question on Friday: « If this is what it’s like at a relatively benign moment, what is it like in the future? » In my view, the first glimpse of the future is likely to come out of Europe. In the coming quarters, we are likely to see a wave of European banks going into receivership – either being nationalized or restructured. Last week, Spain took a 45% stake in its third-largest bank, Bankia, converting debt into equity. As Ireland discovered, providing too much support to the banking system can destroy the fiscal position of the government itself, so it’s unlikely that we’ll see much more of this sort of equity-ownership. Spanish regulators were apparently reluctant to simply take the bank over, which would likely have triggered contagion. But in the coming quarters, this sort of « equity-injection » is likely to be replaced by outright receivership of banks.
This actually isn’t a bad thing. As we’ve often noted, the way that Sweden durably solved its own banking crisis in the early 1990’s was to take receivership of a large portion of the banking system, wipe out shareholders, write down bad assets, protect depositors, give partial recovery to bank bondholders, and then recapitalize banks by issuing the restructured, solvent entities back into private ownership. Europe is finding out that default itself can’t be avoided – it’s merely a question of whether you default on your promises to citizens, or whether you default on your promises to bondholders. The recent election upsets in Europe simply show that citizens have had it with defaults on social contracts in order to make good on bond contracts. Bank nationalization on a large scale (but probably not before the imposition of capital controls) seems to be where Europe is headed.
It’s important to understand that I have no intent of encouraging investors – even buy-and-hold investors – to deviate from their investment disciplines, or from thoughtfully structured portfolios. Many investors are comfortable maintaining their exposure to market fluctuations through the complete bull-bear cycle. As long as these investors are committed to that discipline – recognizing the size and regularity of periodic losses we’ve observed particularly over the past 12 years – my views should not affect their investment strategy. My main objective is always the same – that shareholders see the things that I am looking at so they understand what we are doing and why.
That said, there are also many investors – particularly those close to retirement – who have seen their financial security devastated by retaining investment positions that they were actually very unprepared to carry through a major decline, and then selling at panic lows. I would strongly encourage those investors to review their exposure to risk, and their tolerance for loss, and to reduce any « speculative » positions that they don’t actually intend to retain through the completion of the present market cycle.
Near the 2000 market highs, I offered this advice – If you’re holding a speculative position that you want to reduce, recognize that unless you sell at the exact top – which is unlikely – you’re going to have some level of regret. If you sell part of the position and the price advances, you’ll regret having sold any. If you sell part of the position and the price declines, you’ll regret not having sold it all. Once you accept that there will be some regret in any event, the key is to choose an acceptable level. The « 40% rule » is sometimes helpful – immediately execute 40% of the desired transaction, which gets you moving and helps you to be more objective, since you’ve done something but still have the majority of your position. Then you do it again with part of the remaining position, being as opportunistic as possible. Usually, two or three pieces is sufficient. The transaction costs are somewhat higher, but this sort of approach can reduce the risk of being paralyzed in an inappropriate speculative position. I recognize that some investors feel no inclination at all to dispose of speculative positions here, and have a more favorable view of the market’s prospective return/risk tradeoff. That’s fine – my intention isn’t to change anyone’s strategy. Just keep the 40% rule in your pocket in case market conditions deteriorate more than is tolerable.
As noted above, our estimate of the prospective return/risk profile in stocks, based on a range of horizons from 2 weeks to 18 months, could hardly be more negative. In general, similar instances have been followed by market losses of 20-25% or more over a fairly short period. The outcome may not be as negative as we expect in this specific instance, and there is no universal law of physics that requires the market to adhere to these norms, but we have little basis to expect market risk to be appropriately rewarded here. Strategic Growth and Strategic International remain tightly hedged. Strategic Dividend Value is hedged at about 50% of the value of its holdings – its most hedged position. Strategic Total Return continues to carry a duration of about 2.8 years, with just under 14% of assets in precious metals shares, and a few percent of assets in utilities and non-euro foreign currencies.