Bubble Psychology and Valuations
Courtesy of Pater Tenebrarum of Acting-Man blog
It Can't Be A Bubble!
Articles claiming that the current situation in financial markets does not deserve the epithet “bubble” are a dime a dozen – we come across several every week since at least late 2013. Before continuing, we should point out that there is a big difference between recognition of a bubble and forecasting the timing of its actual bursting. For instance, we were well aware that there was a bubble in the late 1990s, but not only did it still take a good while before it hit its peak (a peak that was then retested in terms of the broader market half a year later), it also expanded considerably further before it did so, and only started collapsing in earnest in late 2000.
It is important to realize in this context that this particular bubble – the one in technology stocks that peaked in early 2000 – is not some sort of “standard measure” for what constitutes a bubble. It was certainly the most extreme stock market bubble in all of history in a major developed market (in terms of valuation expansion in this particular sector) – beating even the Nikkei's famous 1989 blow-out by a huge margin. Again, only if one compares the tech sector's then trailing P/E of more than 300 to the Nikkei's trailing P/E of more than 80 in 1989.
In terms of the broader market's valuation, the bubble peak in 2000 was less than half as spectacular as the Nikkei's, but it was still the top of the greatest valuation expansion ever experienced in the US stock market. We merely want to point out here that it would be wrong to claim that “well, the year 2000 was a bubble, and therefore anything that doesn't look quite as extreme as this one outlier isn't”.
We came across another article of this type recently and want to discuss what we believe the flaws in its arguments are. The article in question is “Bubble paranoia on S&P 500 is a storm in a teacup”, which was posted at Saxo's tradingfloor.com by Mr. Peter Garnry. Note here that we don't want to make an argument about the likely timing of the bubble's bursting or its potential for further expansion (that is a different subject) – we only want to discuss whether a bubble actually exists or not.
Opinions and Bubble Talk
One of the main arguments made in the above article is the following:
“Financial bubbles are characterized by the lack of different opinions, which is clearly not where we are today”
It is explained in the article that this is a reference to a recent increase in people's belief in the existence of a bubble. This is usually reflected in a surge in references to bubble conditions in various financial news media. Mr Garnry (who was 15 years old at the time of the tech mania, and probably remembers the gaudy atmosphere at the time, which was indeed one of a kind) writes in this context:
“The most important point about bubbles is that they exist when no one seems to think they are there. In 2000, very few thought there was a bubble in tech stocks, but with hindsight it is clear that the bubble was massive. Again, very few acknowledged that there was a bubble in US housing in 2006-2007. The fact that three out of five people believe there is a bubble in US equities tells us that we are not in a bubble. When most bears have crawled back into their caves and maybe a few have come out as bulls then we need to worry, but not before.”
As a more grizzled veteran of the markets, we would immediately assert that this just isn't true, based on anecdotal evidence alone. However, we can actually buttress it with data, and have actually already done so late last year (see “Circular Bubble Logic” for details). As a matter of fact, the exact opposite is the case – as a bubble matures and nears its peak, “bubble talk” increases. Anyone who experienced the tech bubble and the housing bubble on the front lines should be aware of this even if the studies proving it didn't exist – but they do exist.
In addition to the Google Trend data we have shown in our earlier article, Bob Prechter has e.g. documented the prevalence of bubble references in newspapers as historical bubbles have progressed. Clearly, as a bubble climbs toward a frenetic peak, talk about bubble conditions begins to literally explode along with prices. Just because Alan Greenspan, Ben Bernanke and 95% of mainstream economists wouldn't recognize a bubble if it bit them in the behind doesn't mean that there aren't a great many other people who do in fact recognize them.
One can also safely ignore the fact that Wall Street sell side analysts will rarely identify bubble conditions in real time publicly. As we know from the bubble post mortems after the peaks of 1929, 2000 and 2007, many of them do in fact acknowledge privately when such conditions exist. We should also note that given the experiences of the past two major asset price collapses as well as the change in the composition of market participants (not least due to the growth of the hedge fund industry), Wall Street research has overall become a great deal more nuanced. In the end though, the business of Wall Street is the selling of securities, and that creates a certain bias in its published assessments of bubble conditions.
Needless to say, there are probably thousands of articles in the Wayback Machine's cache which discuss the bubble conditions in both 1999/2000 and 2006/2007. If bubble talk is on the rise, it is definitely not proof that there is no bubble. Almost all authors writing about this subject do not realize the irony of the fact that there is at the same time always a flood of articles attempting to explain why a bubble is not a bubble, a flood to which they themselves contribute!
If there is one way to fairly objectively ascertain whether the danger posed by a bubble is taken seriously by investors, then it is by looking at measures of investor fear and complacency as well as various positioning data. One of these measures can be seen below:
The St. Louis Fed's financial stress index – click to enlarge.
One of the reasons why the current bubble is not marked by the same outward excitement that was seen in 2000, is precisely that the participants are no longer the same. The current bubble is mainly driven by professionals – institutional investors of every kind – with retail investors only passively participating by e.g. buying mutual fund shares and ETFs. However, these professionals are clearly almost completely oblivious to risk at present – which can inter alia be seen in the historically unprecedented expansion in junk debt issuance over the past two years.
The Monetary Component
One point on which we agree with Mr. Garnry is when he notes that one's analysis of the market must differentiate between the pure fiat money system that has been established in 1971 and the time period preceding it. The fact that money supply and credit expansion have gone into overdrive ever since has altered a number of “tried and true” yardsticks that served as good rules of thumb back when at least still a gold exchange standard was still in place.
For example, prior to the asset bubble becoming greatly extended for the first time in the mid to late 1990s, the market's overall dividend yield never fell much below 3%. This has changed, and the author is on the right trail when he blames the monetary system, or rather, when he points out that different monetary and institutional dispensations do make a difference to market analysis.
We disagree with a subsidiary assumption though, namely that therefore, anything that happened before 1971 is basically irrelevant. This is surely not correct. For instance, during the “roaring 20s”, the true US money supply increased by about 65% (see Rothbard's “America's Great Depression”, which contains a very precise calculation of the era's money supply growth), which is a far cry from today's monetary expansions, but was certainly extreme at the time. There are no fixed quantitative relations between the size of such an expansion and its effects, such as those on the prices of assets or on other goods. At the time, the bulk of the price effects was concentrated first in real estate and later in securities – just as is the case today. Surely we cannot say that everything that happened before 1971 is entirely irrelevant.
However, Mr. Garnry is correct when he states that every slice of economic and financial history is different, and that these differences often encompass major aspects of the contingent historical setting:
“We do not subscribe to the valuation analysis since 1870 being thrown around on Wall Street. This is because the period includes many regime shifts in inflation, real interest rates, nominal interest rates, economic growth, monetary policy, the nature of businesses and the government's size relative to the economy. All these parameters have changed so much that going too far back distorts the overall conclusion and causes inferences that are wrong.”
Sure enough, predictions can not possibly be based solely on empirical data, as there can be no empirical causal constants in human action (thus, every historical boom and bust sequence looks different, even though there are many parallels between them as well). This is where economic theory comes in. While it does not allow us to make precise “predictions” either, it can at least tell us what is and what isn't logically possible. From this we can, in concert with historical understanding, deduce what is likely. The reason why many historical bubbles evince numerous parallels in spite of their differences is precisely that the same economic laws are in operation every time.
Oddly though, Mr. Garnry doesn't discuss the money and credit supply any further. Surely though, it is a major component of the whole “bubble debate”. In fact, we happen to believe that it is the most important component.
A chart of money TMS-2 without memorandum items (which add approx. another $50 billion to the total). Click to enlarge.
The money supply has approximately doubled between 1990 and 2000 and risen more than three and a half times since then. This is the reason why we have a bubble. All other arguments become mere ancillary arguments, including those on valuations.
The Valuation Argument
This brings us the second major argument, which concerns valuations. Mr. Garnry writes:
“Bubbles are normally characterized by the data point (valuations, house prices etc.) being close to two times standard deviations away from the mean. Valuation on S&P 500 is not even close to that scenario. The current 12-month forward P/E ratio is 15.7x, which is a bit above the average of 15.5x since 1990 and the 12-month expected dividend yield is 2.1 percent. The forward P/E ratio translates into an earnings yield of around 6.4 percent. This yield does not smell of a bubble when you consider that the 10-year Treasury yield is at 2.5 percent and corporate bond yields are at historical lows.”
So here we are evidently back to empirical arguments. First of all, at the 2007 peak, market valuations by that very same criterion didn't warn that a bubble was about to burst either. The market's “forward P/E” was also well below the “two standard deviation band” at the time. And yet, the market collapsed by 58% after the bubble had ended. The reason was that the bubble in asset prices was accompanied and caused by a huge credit expansion, in this case specifically in mortgage debt. The current asset bubble is accompanied by a credit bubble as well – only this time, it is concentrated in low grade corporate debt and government debt (there are also smaller credit bubbles on the side, such as that in student debt and sub-prime auto debt for example).
Why is this important? Because it ultimately vitiates the entire valuation argument. Let us think back to 2007 again. What happened after the bubble burst? Corporate earnings collapsed into a heap and turned into the most massive corporate losses of the entire post WW2 era. Obviously, analysts' “forward estimates” completely missed this turning point (generally, these estimates are almost always over-optimistic and tend to be continually revised lower – the only time when they tend to be too pessimistic is near major bear market bottoms).
Current low government and corporate bond yields – which are a direct result of central bank manipulation of interest rates and the money supply – are not a reason to deem current valuations not excessive – quite the contrary. The losses following the 2007 peak were only booked in 2008 and 2009, but they were actually made long before that time. When market interest rates are distorted, economic calculation is falsified, hence the accounting profits booked during the boom period are actually to a large part fictitious – they effectively tend to mask capital consumption. It is easy to see why this must be so: the money supply expansion and artificially lowered market interest rates must lead to capital malinvestment, which by its very nature is destined to destroy wealth.
However, this is never immediately apparent, since it takes time for long term investments to turn out to have been misguided. Amid monetary inflation, businessmen inter alia reckon with depreciation rates that refer to the price structure that existed before the inflationary policy was set into motion. They therefore report a part of the funds that are actually required to maintain their capital as profits. It matters not in what form these funds are then distributed – whether as higher wages, higher dividends or as share buybacks. A large portion of them is paid out of the substance of companies, and as a general rule of thumb, we can say that the bigger the money supply inflation, the more distorted economic calculation will tend to be. Consequently, the errors will tend to be commensurately larger as well (again, no fixed quantitative relationships can be ascertained in this context, thus it is a “rule of thumb”).
This is what is meant by the saying that we are “eating our seed corn”, or are “heating the house by burning the furniture”. Mr. Garnry asked us what we meant by capital consumption and whether it could be measured. The answer is that it cannot be measured while it occurs. However, there will come a point in time when measurements will be taken.
After the housing boom, that time arrived in 2008. Usually, the growth rate of the money supply falls below a threshold that later turns out to have been required to keep the distorted capital structure aloft. Once that happens, the most marginal malinvestments and bubble activities begin to experience difficulties and are liquidated (e.g. the first sub-prime lenders folded in February of 2007 already). From there, the realization that the fanciful reckonings of the boom were erroneous begins to spread. Once the bust arrives with full force, many of the profits made during the boom disappear, as corporate accounts are reconstituted and begin to reflect the true underlying economic conditions – that is when “measurement” takes place.
Year-on-year growth rate of money TMS-2 – click to enlarge.
Where the threshold will be this time is unknowable, but our guess is that it is higher than the last two times, as the underlying real economy seems far weaker (the red parallel lines indicate the area which we guess might prove important). Note that in the mid 1970s, the bust threshold was a growth rate of around 7%.
Conclusion:
If one wants to identify bubbles, one must perforce study monetary conditions. The comparison of historical data on valuations and other ancillary factors can only take one so far. The problem is that in times of strongly inflationary policy, the economy's price structure becomes thoroughly distorted, and that therefore a great many “data” can no longer be regarded as reliable. An added complication is that we e.g. cannot know in advance if the effects of the inflationary policy on prices will broaden out or not. Should “inflation expectations” (expectations regarding future CPI rates of change) rise markedly in the future, this would have a major impact on valuations, which would then begin to contract rather than continue to expand.
However, a bubble can easily burst even if this doesn't happen. Ultimately the question is whether brisk money supply growth will be maintained and whether the economy's real pool of funding is still large enough to allow for additional diversions of scarce resources into bubble activities. Most of the time, it's the eventual slowdown of money supply growth that brings a bubble to its knees.
