If the stock market had a living mind, then the primordial force operating within its subconscious would be 1937. Many contemporaries viewed that year’s equities crash (the Dow Jones index dropped 40 percent) and subsequent global depression as bolts from an almost blue sky. The preceding four years of radical monetary-policy experiments under the Roosevelt Administration had pushed speculative temperatures in equity and commodity markets to record high levels. So it’s hardly a surprise that six years of similar experimentation (2009-14) and accompanying asset-market froth have inspired many investors and commentators to start considering the possibility that another 1937 might be looming ahead. Yes, you can do a Google search and find any number of recent false alerts. But a faulty fire alarm doesn’t mean there’s no such thing as a future fire. The real and important question is: How can we design a better alarm system?
“The danger of 1937” can be described in generic terms. Since its very inception, the Federal Reserve has repeatedly experimented with monetary stimulus as a tool to quicken expansion in the aftermath of serious recessions. Some of these stimulus programs have been more radical than others, and some that appeared radical while they were underway now seem somewhat less so in hindsight. But the essential element of all such programs has been the same: aggressive monetary-base expansion, beyond what would occur under a policy principally devoted to monetary stability. The “stimulus” in question always involves some combination of dollar depreciation and asset-price inflation, with ultra-low interest rates first fuelling capital outflows and then a broader chase for alternative sources of yield in trading markets. Asset-price inflation carries an inherent risk of violent reversal due to the uncertain timing of eventual stimulus withdrawal by the Fed, anticipation of which causes the dollar to rebound. That risk is deferred or even “indefinitely postponed” for only so long as a potential or actual near-miracle continues to justify ballooning market prices and compensate for any drag from a stronger dollar (intensified by foreign currency-devaluation policies). In the absence of such an out-of-the ordinary X-factor, the prospect of a 1937-style crash and recession becomes ever more real.
I. BEST-CASE SCENARIOS
Two previous episodes in modern U.S. monetary and business-cycle history represent a reasonably “happy-ending” version of the basic either/or phenomenology at play, here.
The Early-Mid 1920s. America experienced a serious recession in 1920 and 1921. Benjamin Strong’s Federal Reserve Bank responded with a major monetary stimulus, which everyone agrees helped spark the initial stock market boom. In 1923, however, the global situation darkened (the French incursion into the Ruhr; German hyperinflation, the Tokyo earthquake); the U.S. dipped back into a mild recession; and Wall Street started to quiver. It seemed that Fed-induced asset-price inflation was about to give way to asset-price deflation, intensifying the downturn. This time, though, various X-factor miracles did come riding to the rescue: a technological revolution (mass electrification and the automobile) and accompanying surge in productivity and capital spending; Germany’s economic boom following international stabilization of its currency; and global détente in the form of French-German “Locarno Pact” rapprochement. Thus did the Fed enjoy a lucky exit from its early-cycle interventions. (And thereafter did the Fed bungle this happy ending—out of a misplaced focus on stable prices, which were stable or falling, in any case—by attempting to restrain the natural rise in interest rates that accompanies economic good times. The eventual consequence was a global bubble-and-burst of epic proportions. But the Wall Street Crash of 1929 and Germany’s bankruptcy in 1931 is a story for another day!)
The Early-Mid 1990s. In the wake of the late-1980s bust (the stock market crash and S&L, real estate, junk bond, and Enron crises), the Federal Reserve Bank under Alan Greenspan adopted vigorous stimulus policies in an effort to speed recovery from the recession of 1990-2. Again—in combination with a currency offensive launched by the Clinton administration—the Fed’s easy monetary policy produced asset-price inflation: a powerful uptick in the U.S. stock market and a boom in carry-trade activity into Mexican, Italian, and Canadian high-yield sovereign debt. Then, as the Fed moved towards normalization in 1994, asset-price deflation began to emerge (the Mexican debt crisis, a stock market downturn, and a sell-off of high-yield sovereign debt as currency values plunged in Canada and Italy)—and Wall Street started to quiver. But again, there was an X-factor rescue: the internet and telecommunications boom produced a surge in business investment and productivity and the stock market began a powerful new upswing despite the Fed’s monetary policy normalization. (And alas, once more, the Fed proceeded to throw away this good fortune by imposing a regime of artificially low interest rates, thus encouraging investors to seek higher returns in much riskier trading-market instruments, and eventually bringing the cycle to its predictable close with an asset-bubble burst—the most famous example being the Nasdaq crash).
II. THE WORST-CASE SCENARIO
Federal Reserve Bank stimulus programs designed to accelerate recovery from serious economic recessions do not all end the same way, of course. The Fed does not always go on to make a hash of the positive outcome, for one thing—simply because, more often than not, no X-factor “white hat” cavalry charge materializes, and there is no positive outcome to begin with. When the only X-factors on the horizon are threatening ones, in fact, things generally tend to go very badly, indeed. Here, the starkest historical case in point is the then-unprecedented monetary experiment undertaken during the Great Depression by the first Roosevelt Administration beginning in 1933.
The mid-late 1930s. The New Deal monetary experiment involved two major policy decisions. The first was an exit from the gold standard (and sharp devaluation of the dollar). The second—made possible by a wave of foreign capital inflows during 1934-36, which the Treasury Department converted into Fed-financed gold purchases—was an enormous expansion of the monetary base coupled with short-term interest rates pegged at virtually zero. It was unclear even at the time whether this exercise in quantitative easing (though the phrase was not then in use) was “necessary” or particularly helpful. A remarkably strong business-cycle recovery was already underway in 1934-36, primarily and simply because the Depression had pushed the price of goods so low (down by roughly 30 percent) that consumers and businesses were beginning to spend forward in anticipation of a rebound.
What did become clear by the mid-1930s, however, was that New Deal monetary experimentation had a dark side: fantastic asset-price inflation in U.S. equity and commodity markets, still below peak pre-Depression levels, but alarming enough to inspire widespread talk about possible dollar revaluation and Fed-policy normalization. What followed was a period of minor, reactive zigs and zags in Washington. In late 1936 Fed and Treasury officials took several initial steps towards withdrawing monetary stimulus, causing short-term interest rates and long-term bond yields to rise a tiny bit (less than 50 and 40 basis points, respectively) in the early months of 1937. In mid-spring of 1937, in response to a first dip in the equity markets and a flat-lining of general economic indicators, the Fed intervened again, this time to reverse course in the bond market to lower long-term interest rates. But none of it was enough to forestall the inevitable. Speculative temperatures had pushed asset prices to unsustainable highs. And—crucially—all the X-factors were bad: election results and Supreme Court verdicts that spooked Wall Street; a French government unable to stabilize the franc after the demise of the gold bloc; a Japanese Imperial Army engaged in full-scale war with China.
By the second half of August 1937 an inexorable stock market selloff was underway. By the end of March 1938, during a recession that would wipe out 18.2 percent of U.S. GDP before it ended in June, the Dow had lost nearly half its pre-selloff value.
III. 1937 AND TODAY: HOW BEST TO DETECT AND PLAN FOR A POSSIBLE RECURRENCE
On the face of it, the current U.S. business-cycle (dating from the trough of spring 2009) has reached a stage that looks a good deal more like 1937 than 1923 or 1995. The effects of a sustained and quite radical, mid-1930s-style monetary experiment are increasingly obvious. Everyone and his dog is now aware that the Federal Reserve Bank’s super-low interest rates have produced considerable froth across a wide array of markets: high-yield credits, high-end residential real estate, and (most of all) private equity. In certain such markets, this asset-price inflation has already begun to leak significant air; emerging-market currencies, oil credits, and commodities like iron ore have all plummeted (though the Beijing-sponsored Shanghai equity-market bubble has helped to trigger a recent partial rebound by generating optimism about future Chinese economic prospects). The early-stage dollar devaluation in this U.S. business cycle has been fully reversed. A perfectly understandable expectation has taken hold that the Fed will sooner or later decide to normalize its policies, having already “tapered” away from quantitative easing, and now setting a tentative timetable for the return of above-zero interest rates. And there appears to be no imminent X-factor to cushion the blow; quite the contrary, the geo-political skies are now darker than they have been at any point since the end of the Cold War, while U.S. productivity growth continues to languish amidst lacklustre business investment. Put bluntly: the danger of a 1937-style crash and recession is greater than usual—something only a fool would dismiss out of hand.
It’s not something guaranteed to happen, on the other hand. Nor would it be possible to forecast its arrival or severity with meaningful precision, in any event. The key questions, then, are not so much about “if” and “when” and “just bad, or truly cataclysmic?” As suggested at the outset, they key questions, instead, are: What’s the best warning system? Where should investors and policymakers post their sentries? How should those sentries interpret what they see—is it an isolated trashcan fire or a full-scale conflagration? These are all issues of preparation —and continual readjustment in response to varying degrees of risk. And where judgments of risk are concerned, two factors in particular are worthy of special attention and subtle understanding.
The speculative element in Federal Reserve decision-making. Fed policymaking—both for the Bank’s own practitioners and third-party observers—is a fiendishly tricky enterprise. According to conventional folklore, for instance, all would have been well in 1937 had the Fed not prematurely started to normalize monetary policy in late 1936; fears of a steep recession would not have intensified, and investors would not have felt frantic to dump their equity holdings. Maybe. But maybe not. Basic economic laws cannot be permanently revoked, after all. Eventually, whenever overheated commodity and equity markets reach a boiling point (fuelled by quantitative easing’s artificially depressed interest rates, which make safer bonds and bills unattractive), some degree of corrective freeze will set in—and the value of high-risk assets will collapse—unless a near-miracle occurs. It is just as likely that the Fed’s “premature” normalization in 1936 actually eased the pain of Wall Street’s 1937 collapse—by preventing an even steeper pre-crash peak. Today, almost 80 years later, we still can’t say for sure what the “right” monetary-policy move should have been—even after taking account of a whole series of earlier, quite clearly wrong moves.
Fed officials wrestling with these same problems in the moment can hardly be expected to fare much better. A red flag has gone up over anticipated or already evident asset-price inflation. There may be prominent authorities willing to pretend otherwise, but doing nothing at all in the near term, murmuring about normalization at some unspecified point in the future, and assuming that the situation can smoothly cure itself in the meantime—without divine intervention or anybody getting hurt—is not a serious option; it’s a fantasy. In the real world, the Fed’s choice is a stark one: The Bank can start pressing on the brake—normalize now —and risk triggering a market down-turn all by itself. Or the Fed can decline to intervene, hope for the necessary X-factor, and ride things out from the spectator seats. In either case (the necessary X-factor being so rare), an eventual market down-turn is likely to occur. And in either case, the question whether that down-turn might have been less severe had the Fed acted differently will remain effectively unanswerable. What if your father hadn’t run into your mother that day on the subway? There’s no way to know.
Plenty of people do their best to know, anyway, of course. The Fed “does its best.” And outsiders expend an enormous amount of time and energy trying to guess what the Fed will guess is the right course of action, a speculative phenomenon that plays an outsized role in short-term market fluctuations, and one which—for that reason alone—investors cannot afford to ignore. Nevertheless, narrow-focused attention on Federal Reserve Bank decision-making does not make for the most robust or confidence-inspiring fire-alarm system.
Historical and statistical yardsticks. Past experience provides more generally dependable—if necessarily limited—guidance. We know from previous history that price inflation tends to infect different asset markets at different times, but the overarching cycle follows an established and predictable sequence. At mid-phase, the earliest-afflicted markets are already beginning to deflate, even as later-to-the-party markets are still climbing to their highest (and shakiest) ladder rungs. As suggested above, this is where we seem to find ourselves today, with emerging-market currencies, oil credits, and certain commodity markets all past their peaks and in decline—while high-yield credits, high-end real estate, and private equity remain quite frothy. What comes next will be the cycle’s end phase—that, too, we know from previous history—in which a steep slide occurs across nearly every asset market.
History is of relatively little use, however, in assessing when exactly this next end phase will arrive. For that we must largely depend on a reading of “the numbers.” And we must read with discernment. Some conventionally employed statistical measurements are more reliable and meaningful than others. Some conventionally employed statistical measurements aren’t reliable or meaningful at all, in fact. Until very, very recently, for example—on May 6, Janet Yellen finally conceded that U.S. stock valuations are “quite high” and pose “potential dangers” to financial stability—Fed officials have aggressively minimized the risks of equity-market froth by citing the price-to-earnings (P/E) ratio on the S&P 500 (now around 20, still within the upper limit of a normal range). Fed officials have further pointed out that the earnings yield on stocks (the inverse of the P/E ratio) remains about 4.5 percentage points higher than real long-term interest rates (measured by the yield on “TIPS,” 10-year inflation-protected U.S. Treasury bonds), and thus represents a reasonable (if not particularly rich) reward for bearing risk; earnings yields on stocks provided much lower relative rewards during previous periods of obvious market froth like 1929 or 2000.
These numbers are utterly misleading, however—as Ms. Yellen’s space-of-one-day change of mind about their meaning might well suggest. As a matter of historical fact, P/E ratios were also within a normal range on the eve of the 1937 Crash. And as a matter of principle, “normality”—like beauty—lies in the eye of the beholder. If present reported earnings are far above sustainable long-run trends (and inflated by widespread financial engineering), then rational investors should not pay as high a price for equities (relative to earnings) as they would during “normal times.” And today there are indeed grounds for caution of this sort. Underlying U.S. corporate profits relative to national income are edging down from an all-time peak reached in 2013. The corporate sector is massively engaged in so-called “financial arbitrage” (equity buy-backs related to a disguised rise in leverage, “carry trades” in high-yield bonds, and so forth) which produces superficially flattering earnings-per-share performance reports in the near term, but also involves exposure to large and unknown future risks. This is all reminiscent of the Zai-tek boom in Japan during the late 1980s.
Neither should a rational investor be especially reassured by the TIPS part of this equation. Historically low long-term interest rates, the result in large part of Federal Reserve manipulations, mean that the earnings yield spread between stocks and bonds has not been compressed below normal—despite frothy valuations for equities. The statistical evidence on which an investor might otherwise base a judicious risk-reward calculation has thus been distorted by monetary policy experimentation at the Fed. Yes, there are economists who hypothesise about secular stagnation and argue that long-term interest rates will remain low far into the future. But secular stagnation would mean a dismal outlook for corporate earnings growth, hardly a positive sign for equities. So why would a real-world market actor with serious money to lose make a bet on such a hypothetical?
Maybe because he thinks he’s smart enough to stay at the party just a little while longer—and then beat everyone else out the door in the nick of time. Maybe because, once he’s started dancing to the tune, he finds it impossible to resist soothing lyrics like “there is nowhere else to go [but equities] in a zero-rate world” and “1937 is the most overworked analogy in economics” (as a strikingly confident Wall Street Journal editorial announced back in March). Maybe the Fed won’t start a policy normalization any time soon.
But that’s an awful lot of “maybes.” Come what may, it seems highly unlikely that the Ghost of 1937 will make a quiet exit from the scene. Janet Yellen is now hoping to banish the Ghost by talk alone—gently commenting on stretched valuations in order to dilute market froth in a “controlled” manner. Unfortunately, history and economic principle are not on her side.