Betting On a Bubble, Bracing For a Fall

Weekly Market Comment Archive

Investment Research & Insight Page

Betting on a Bubble, Bracing for a Fall John P. Hussman, Ph.D. All rights reserved and actively enforced. Reprint Policy

The financial markets are in a bit of a fight here between technicals and fundamentals. On a technical basis, a variety of widely-followed trendlines, moving average crossings, and resistance areas converge on the 1100 area for the S&P 500. Market internals have also firmed somewhat during the rally in recent weeks, suggesting that investors are eager to re-establish a speculative tone to the market. At the same time, fundamentals are bearing down hard on the market. We continue to observe a clear deterioration in leading indicators of economic activity.

Over the short-term, my impression is that the technicals may hold sway for a bit. The economic data points simply do not come out every day, and to the extent that economic news is not perfectly uniform in its implications, the eagerness of investors to speculate can easily dominate briefly. We established enough contingent call options at lower levels that we've now got about 1% of assets in roughly at-the-money index calls - a modest "anti-hedge" that removes any concern we might have about a frantic short-squeeze if the S&P 500 moves materially above 1100. At the same time, the historical evidence suggests that fundamentals have ultimately trumped technicals when we've observed similar warnings from economic indicators in the past. My impression is that the economic cold water could hit investors very abruptly, so that gains achieved over several weeks may be suddenly erased in a matter of a few days.

My basic concerns are the same here. Investors who will need to fund specific expenses within a short number of years - retirement needs, tuition, health care, home purchases etc - should not be relying on a continued market advance. If your life plans would be significantly derailed by a major market decline, get out. In contrast, if you are pursuing a disciplined, long-term investment strategy, and you know from your own experience of the past decade that you are diversified enough to ride out periodic losses without abandoning that strategy, ignore my views (and those of everyone else) and stick to your discipline.

The evidence from our Recession Warning Composite is already on the table, and would strengthen considerably if the ISM Purchasing Managers Index declines to 54 or less (the ISM services index already dropped to 53.8 last month). Again, no indicator in our composite is decisive on its own, but the combination of factors has always and only been observed during or immediately prior to recessions. The ISM figures and employment numbers will be a focus of investors in early August, but while a 54 on the PMI would be informative from our standpoint, investors may not immediately recognize it as meaningful. Undoubtedly, the most timely indications of recession risk are based on composites of multiple indicators, which is one of the reasons I've focused recent comments on the ECRI Weekly Leading Index (WLI). Last week, the growth rate of that index slipped again, to -10.5%, which we've never observed outside of actual or oncoming recessions, though the ECRI notes that there are two unpublished data points in the early post-war years that weren't associated with recessions.

I should emphasize, contrary to what some analysts have asserted, that this is not simply a "slowdown in the growth rate" of the WLI. To emphasize this point, I've presented three charts below (the ECRI makes this data publicly available at no charge on its website). The first is the level of the ECRI Weekly Leading Index. Not the growth rate - the level.

Note that we are observing a downturn, not simply the slowing of a positive rate of growth of the WLI. The next chart is the growth rate of the WLI, which dropped to -10.5% last week.

Not every downturn in the index is important. The ECRI emphasizes that when interpreting economic data, there are three requirements that have to be satisfied to confidently indicate an oncoming recession - the downturn must be profound, pervasive, and persistent. Profound means a deep decline, which we're clearly observing here. Pervasive means that it must not be driven by simply one or two components (for example, the drop in 1987 was almost exclusively driven by stock prices). ECRI considers a range of factors such as stock prices, housing, employment, money, and confidence measures, among others, but does not articulate the specifics. For our part, we observe not only stock price weakness, but disappointingly high new claims for unemployment, weakening confidence, soft retail sales, easing growth in new factory orders, a flattening yield curve, wider credit spreads, and so forth.This downturn is not limited to stock prices.

Finally, persistence is required. The signal can't last simply for a few weeks. It is here where we require a bit more subtlety than does the ECRI. The reason is that our primary object of interest is the stock market, which is itself a leading - though imperfect - indicator of the economy. The stakeholders of the ECRI are largely interested in economic activity, and can weather a longer period of uncertainty than investors can. I've noted, for example, that the ECRI's admirable recession calls over the past decade have sometimes come only after significant market declines. The March 2001 call, for example, caught the precise beginning of the recession from the standpoint of official recession dating, but the S&P 500 was already down over 25% by that time. This is not at all a criticism of the ECRI, which is an outstanding institution - just an observation that our constituency is different, and timing lags can be very costly.

Since the ECRI data is only one of many indicators we use, we may come to conclusions that ECRI may be reluctant to make for its very different set of stakeholders. Our own Recession Warning Composite is intentionally constructed of several relatively "weak" criteria that are not particularly stringent in themselves (e.g.anything but a very steep yield curve), but carry significant weight when they are all observed together. We're much more interested in the uniformity of evidence than we are in a collapse in any particular component. This is because the most useful information content is almost always in uniformity or divergence. Again, waiting too long can be costly.

Then again, we may not have to wait long. As I've noted before, the Weekly Leading Index growth rate is highly correlated with subsequent changes in the ISM Purchasing Managers Index, with a typical lead of about 13 weeks. It's possible we'll see some of that effect as early as the July report (due out on Monday, August 2), but the August and September reports will be the ones to watch. Also, collapsing readings from the WLI are typically followed by a surge in weekly claims for unemployment, also on a horizon of about 13 weeks. If the current recession concerns are meaningful, we would expect to observe a sharp spike in weekly claims for unemployment well above 500,000 in the near future. The early signs of deterioration may emerge rather quickly.

Again, some analysts have suggested that we are only dealing with a slowdown in the growth rate of the WLI. This is not accurate. We're observing a sharp contraction in the leading index itself. But just for fun, the chart below shows the 13-week change in the growth rate. This would clearly go into the file labeled "Ugly."

Betting on a Bubble

Much of my research last week was spent working with our various measures of valuation. While the extent of implied overvaluation on our best measures does have a range of variation, that range runs between about 25%-40% overvalued. We certainly know of many valuation indicators that suggest that stocks are "cheap" here. Unfortunately, they don't demonstrate any reliability in historical tests. It is almost mind-numbing to observe how many analysts confidently make valuation claims about the market on CNBC, evidently without ever having done any historical research. If you don't require evidence, you can say anything you want.

For our part, the following charts provide a good overview of where stocks are valued at present.

To create the first chart, I used our standard methodology based on normalized earnings. As long-term readers of these comments know, this method generates projected 10-year total returns for the S&P 500 Index. Presently, we estimate that the S&P 500 is priced to deliver a 10-year annual total return of about 6.71% here. So including reinvested dividends, one dollar of value in the S&P 500 10 years from today probably costs investors (1/1.0671^10) = 52.2 cents here. If the S&P 500 was priced to deliver a 10-year total return of 10%, that same dollar of value would cost investors (1/1.10^10) = 38.6 cents here. On that basis, the chart below would show a valuation ratio of 52.2/38.6 = 1.35, implying 35% overvaluation at current prices.

Doing the same exercise historically produces the following picture. Notice that while the 2009 market low did put the S&P 500 slightly below 1.0 for a few weeks, it was clearly not a valuation low anywhere near the great secular buying points of 1950, 1974 and 1982. I continue to view that 2009 low as vulnerable to a further breach. That's not a forecast - simply a recognition of the market's tendency, during periods of economic turmoil, to reach valuation troughs that are a fraction of what we observed in 2009.

In the foregoing chart, I've highlighted the region in excess of 50% overvaluation in green, representing what I view as bubble valuations. I expect that some observers would scoff indignantly at the idea that the S&P 500 valuation multiple was 3.4 times the appropriate level. But observe what this means. If $1 of value a decade later should cost 38.6 cents (at a 10% annual rate of return), a multiple of 3.4 times that would mean that in 2000, investors were paying about (3.4 x .386 = ) $1.31 for $1 of value 10 years later. Do the math. This means that investors could have expected an annual 10-year total return of (1.00/1.31)^(1/10) - 1 = -2.7% annually. In fact, that expectation would have been just about right.

The following chart shows the relationship between the measure above and the actual 10-year total returns achieved by the S&P 500 historically.

Again, right now, investors are looking at an estimated total return of about 6.71% over the next decade, based on our normalized earnings methodology. Now, given the low level of Treasury yields, it's possible that some investors might find that projected level of total returns satisfactory. Again, if you are a disciplined investor, understand the potential returns and the probable risks, and are comfortable that periodic market losses will not shake your discipline, then buy-and-hold may be a perfectly appropriate discipline to pursue.

I suspect that investors will not do significantly better, but if investors in aggregate are comfortable with this projected level of long-term returns, they may not do substantially worse over the next decade either. The problem is that if risk perceptions increase at some point over the next decade, the majority of investors may very well decide that 6.71% is not sufficient, and if they demand higher long-term returns, the current price of stocks will be pressed lower. So even modest long-term return prospects in stocks come with the potential for substantial shorter-term volatility. Valuation isn't a timing tool - but my impression is that investors may not be durably content with the level of long-term returns implied by stock prices here.

One reason for highlighting the period of bubble valuations in green is to make a very emphatic point. If you exclude the bubble valuations of 1995-2007 (as depicted in the chart below), the current valuation of the S&P 500 is near the highest level ever observed in history. To expect valuations to expand from here is to rely on the sustained resumption of bubble valuations that have ultimately been devastating to investors.

"Well, then the Fed would lose money there"

Last week, Ben Bernanke appeared before Congress for his regular Humphrey-Hawkins testimony. For most of that testimony, it fascinated me that every time the Bernanke said that the Fed has taken no losses on its operations, there was absolutely no remark that the reason the Fed has not lost money is that the Treasury, directly (Fannie, Freddie) or indirectly (AIG) has made the liabilities held by the Fed whole.

From that perspective, the critical part of Bernanke's testimony was the following exchange with New Jersey Congressman Scott Garrett of the House Financial Services Committee. Importantly, Bernanke concedes that by placing two-thirds of its balance sheet into the liabilities of insolvent agencies (Fannie Mae and Freddie Mac), now under conservatorship, the Fed is essentially relying on Congress to make these institutions whole at taxpayer expense. The Fed has put the public on the hook to bail out the GSEs.

SCOTT GARRETT: You bought over a trillion dollars of GSE debt, and to that point, under normal circumstances, on the Fed's balance sheet what you have on there are Treasuries, or if you had anything else on there, I assume you would have a repurchase agreement for those securities on your balance sheet. Now of course around two-thirds of that are in GSE debt.

BEN BERNANKE: Correct.

GARRETT: So right now, those are guaranteed - whether they're sovereign debt or not, we don't know - but they're guaranteed by the U.S. government. But they're only guaranteed to when? 2012, right? After that, Congress may in its wisdom make another decision, and at that point in time, you may be holding on your balance sheet - two thirds of your balance sheet - something that is not guaranteed by the Federal government. First of all, you don't have a ... do you have a repurchase agreement on those with anyone? No.

BERNANKE: I don't know what you mean by a repurchase agreement. We own those securities.

GARRETT: You own those securities. Right. So there is no repurchase agreement outside to buy them back. You own them.

BERNANKE: Right.

GARRETT: So after 2012, if they're no longer guaranteed, is it fair to say that you may at that point in time actually engage in fiscal policy, because you basically are creating money at that time? And I know that you'd agree that it would be an unconstitutional role for the Fed to engage in fiscal policy - so where will you be at 2012 if they had to take a haircut on those because they're no longer guaranteed?

BERNANKE: Well, first from the government's perspective, I, uh, such an act would, uh, there would, the Federal Reserve would lose money which the Treasury would gain. There would be no overall change to the position of the U.S. government. Secondly, the Federal Reserve act explicitly gives..

GARRETT: How would we be gaining? How is the Treasury gaining?

BERNANKE: Well, if there's a bad mortgage and the Treasury.. it requires $10 to make it good, if the Treasury refuses to do that then the Fed loses $10, so one way or another the government's going to lose $10. But I would just say two things, one is that I think, uh...

GARRETT: But if you didn't purchase them in the first place, it would just be a total - then what would have occurred? There would not have been the creation of that $10. Now that you've purchased them, and in essence if we don't back them up, then you will have created that additional $10.

BERNANKE: Well, I hope that doesn't happen, because I think it's very important for financial stability and confidence that we, that we guarantee...

GARRETT: Let's play out that hypothetical that it does happen.

BERNANKE: Well, then the Fed would lose money there. But let me just point out that the Federal Reserve Act, that we did not invoke any emergency or unusual powers to buy those agencies. It is explicitly in the Federal Reserve Act that we can buy Treasuries or agency securities and so we did not do anything unusual there.

GARRETT: In what status were they when you bought them? Were they in conservatorship at that point?

BERNANKE: Um, yes.

GARRETT: Is it normal practice for the Fed to buy agency securities when they're in conservatorship? Was that ever done before?

BERNANKE: It's never been in conservatorship before.