Maybe This Time is Different
By Andrew Redleaf
August 14, 2012
This Time Is Different, the catchy title of the popular book by economists Carmen Reinhart and Kenneth Rogoff, has also become a catchphrase summing up the world-weary wisdom of our time. Reinhart and Rogoff in recounting eight hundred years of "nancial follies and investment bubbles gleefully point out that in every case experts offered plausible arguments for dispensing with traditional rules of valuation, i.e., “this time it’s different.” Surely the book owes its popularity to our having endured, in less than a decade, two of the biggest valuation boondoggles in history. In the aftermath it no doubt makes people feel all worldly wise and cynical-smart to arch their eyebrows and sarcastically reply, “right, this time it’s different” to every suggestion that something fundamental might really be changing.
In short, mean reversion is hip.
And that is tickling my contrarian funny bone.
Don’t get me wrong. Mean reversion is one of my very favorite things in the whole world. Betting against the infinite extension of a happy trend is a terriffic way to make money. Even better in my experience is to bet against the infinite extension of an unhappy trend. Markets often take years to grasp that the next New Jerusalem is but another mirage. They usually catch on sooner to the news that the world has yet again failed to come to an end.
We ourselves have been heard to chuckle cynically at theories for why, say, housing prices would never fall or high yield markets would never recover. Regular readers of these letters may recall how much fun we’ve had over the years mocking the increasingly fatuous arguments people make for increasingly improbable valuation trends right before the trends reverse.
Still what are we to think when mean reversion itself becomes a fad, the latest way to be in the know without really knowing anything? Are dour fads any less faddish simply for being dour? Should mass skepticism be exempt from our own skepticism about mass opinion?
In particular we have been wondering about the growing chorus proclaiming that the elevated corporate profit margins and earnings we have seen since roughly the beginning of this century must revert to the average levels seen for most of the last century. Jeremy Grantham and his colleagues at GMO have been vocal on this subject for years now, the latest sally coming from GMO analyst James Montier in a piece called “What Goes Up Must Come Down!” Montier offers a complicated macroeconomic argument to the effect that anomalously high profit margins have been driven by anomalously high government spending and must revert when our government is forced to change its spendthrift ways.
I thought most of the Montier piece nonsense not because it was bad macroeconomics but because it was typical macroeconomics. Macroeconomics arguments are based always on a “circular flow” account of the economy. The circular flow describes the normal roundabout of making and trading, getting and spending, and all the other clearly “economic” activities of a society governed by ordinary economic motives and disciplined by generally recognizable price signals. The circular flow describes not “the society” but “the economy” which consists of “markets” tending toward equilibrium. Circular flow arguments are thus almost entirely useless for analyzing fundamental economic change because as Schumpeter showed every important “development” in the economy, every change of note happens outside the circle, indeed outside the economy in the conventional sense of that term. If anything really is different “this time” no abstraction of the circular flow will tell us about it.
Still, as silly as I found the Montier piece his core point—that we should expect mean reversion—generally holds. A man can write thousands of words of nonsense and yet end his treatise with the simple phrase “expect reversion to the mean” and still prove a sage. In any event “profits must revert to the mean” seems now to echo from every organ of the "nancial press, from countless blogs, from one dour gray-haired countenance after another glaring at us from CNBC. For all I know it is written on the subway walls and tenement halls.
And I just don’t quite believe it.
Could profits revert to the mean? Sure. Likely will? OK. But must? Why? Because nothing ever changes? Because no one ever invents the steam engine or the assembly line or the mass production of steel or the microchip? Because social conditions never alter?
Oh, sure, we’re going to see some reversion. We’ve got to see some reversion for the same reason the New York Times has at some point declared every recovery in my professional lifetime to be the first “jobless recovery” ever in history. Most recoveries start out “jobless” because all recessions begin to end in the same way. Productivity rises ahead of wages or hours worked. The immediate result is a recovery in profit margins even as revenues stay depressed. Eventually revenues recover, but gun-shy firms postpone hiring, further increasing margins and cash reserves until management finally gains the confidence to hire and increase wages. That pushes margins on sales back down, though returns on equity may continue to rise for some time.
As this picture suggests, the big variable on the other side of the equation from profts is usually wages. It is wages, traditionally, that drive profits back to the mean. And for all the talk we’ve heard lately about growing income inequality, over the long run the shares of national income paid out to labor and capital have been remarkably stable.
For many decades the labor share of U.S. income fluctuated around two-thirds. Over the past decade the labor share of nonfarm business income has taken a hit, falling from 65% around 1980 to near 57% now. The capital share has risen, though more modestly. But there have been short-term variations before. With the long-term trend so stable the blip over the last decade can be read just as much as an indicator that we are due for some mean reversion as an argument that we are seeing fundamental change. That said, it’s interesting to think a bit about why fundamental change might be happening. I can think of two reasons, both more or less obvious, plus a third that might ultimately benefit labor and yet still be good news for earnings.
The first is that we have seen a remarkable increase in wage flexibility, overthrowing some basic economic assumptions that have reigned at least since Keynes penned The General Theory. The motive for The General Theory was Keynes’s observation that social developments, particularly the rise of unions but other complexities as well, were invalidating the classical assumption that high levels of involuntary unemployment would always be rapidly ameliorated by a decline in the cost of wages, making it affordable for employers to hire. Keynes saw that wages in a modern economy had become too “sticky” to allow for rapid downward adjustment. With the nominal price of labor so much less elastic than in the economy of a century before Keynesian policy was to force down the real price of labor by inflation.
Whatever one thinks of the record of the Keynesians after Keynes, his insight on wage stickiness was a powerful one and seemed to remain essentially true in the world in which I grew up. That world is gone. The high watermark of U.S. private sector unions came four years before I was born, in 1953 when some 35.7% of the U.S. workforce, and by far the majority of the industrial workforce, was unionized. Even in 1980 just before the last big recession some 20% of private sector workers were unionized. In 2011 that figure was 7%. In 1980 GM employed well over half a million UAW members in the U.S., more than one-half of 1% of all U.S. workers. By 1994 that number had shrunk to about a quarter of a million. Today that number is about 64,000, less than 0.05% of total U.S. employment
Manufacturing jobs are lost in two ways: they can migrate elsewhere, or they can be lost locally to automation. Migration driven by international labor cost arbitrage has been a public agony in the U.S. at least since we began to lose domestic auto sales to Japan. The migration of manufacturing jobs overseas often hits unionized and well-paid industries especially hard. One result is to decrease labor’s share of income because industries where labor has the best deal are hit hard. The other result is to increase elasticity of manufacturing wages in the U.S. This is not only because some of the least elastic jobs no longer exist, thus shifting the total proportion of jobs toward more elasticity, but because manufacturing jobs don’t only migrate overseas; they migrate to right-to-work states in the U.S. Forty years ago not a whole lot of work, or at least factory work, got done in most right-to-work states, nearly all of which were farm states and most of which were in the South. But U.S. Bureau of Economic Analysis data show that between 1990 and 2010, employment in right-to-work states grew 26% compared to 8% for all other states. Manufacturing employment in right-to-work states grew 84% vs. 19% for all other states. Meanwhile, though wages in right-to-work states rose, because the rise was from a low base relative wage costs moved very little. Manufacturing wages
per worker grew 109% in right-to-work states vs. 96.1%.
After decades of assuming wage stickiness and conceding, “Oh, Keynes was at least right about that,” to look around at the U.S. private economy today is a bit like waking from a dream. Suppose the question had no history. Suppose no one had ever before asked, “Are wages made sticky by institutional arrangements such that it is implausible to expect rapid wage adjustments in a recession?” If someone asked that question for the first time today, would the most common answer be “obviously” or “let me get back to you”?
Increased elasticity in the price of labor would certainly help explain one of the curious and hopeful phenomena of the current recovery. In every previous business cycle of my professional life the recession devastated manufacturing employment and the ensuing recovery did little to bring it back. This time, to coin a phrase, “it’s different.” Manufacturing employment has been one of the few relatively bright spots in the recovery, at least in terms of jobs and hours worked. True, most of the new manufacturing jobs, especially in right-to-work states, don’t pay as well as the ones that have disappeared. That is exactly what one would expect in a reversion from a Keynesian to a neo-classical world. In a neo-classical world labor avoids unemployment by putting itself on sale. An odd example of this, because it involves public sector employment, is the denouement in Wisconsin. One reason Scott Walker ultimately triumphed is that even the modicum of wage flexibility gained by removing benefits from collective bargaining allowed school districts and municipalities to forestall or reduce layoffs while so many states with stronger public service unions faced massive job cuts.
The increase in wage elasticity cuts both ways. If we really are back in a neo-classical world then as earnings increase so will wages. Productivity will fall and eventually margins on sales will follow. But if the question before the house is the durability of elevated corporate earnings, on balance a return to a neo-classical world is likely a positive. Toward the peak of the business cycle corporate earnings do continue to rise even as margins on sales fall. Even in a frictionless world of perfect information and no mistakes employers are right to add shifts and increase wages even as margins fall until the last marginal worker is actually detracting from returns on equity. Booms are not a negative for shareholders. The risk to shareholders in Keynes’s world is the difficulty of getting wages down as management (operating in the real world with scarce information and faulty judgment) over-expands and earnings fall. Make it easier to cut wages and earnings should suffer less and recover faster in a downturn.
The second reason to think that the shift of income from labor to capital might continue is far more basic and fundamental: the entire history of economic development is the history of shifting income from labor to capital.
In a primal economy with no money, no trade, and thus no buying or selling or owning of either land or capital goods the share of income claimed by labor is always 100%. There are no capitalist hunter-gatherers. Economic development is precisely the process of adding capital, including intellectual capital, to the system to increase productivity and giving capital a claim on income. The reciprocal of a hunter-gatherer society is an economy in which essentially every manufacturing task is automated and industrial employment and payments to industrial labor cease. In that state 100% of revenue goes to capital and materials, with most materials being similarly produced by automated systems. In principle the same is true even of most services though it is hard still to imagine 100% of the service economy being automated. Industrial automation, however, is already so fantastically advanced that we can easily imagine an industrial sector many times more productive than today’s yet employing not 10%, perhaps not 1%, as many factory workers. Heck, if the special double-secret-euthanasia-for-entrepreneurs provision of Obamacare is repealed we might even live to see it.
So the natural progression of history, occasionally interrupted by Pestilence, War, Famine, and Death, is toward capital capturing 100% of at least manufacturing income. A reasonable question is how this occurs—as a steady progression, or in spurts? If it happens at least partly in spurts then presumably those spurts will occur in response to developments impinging on the economy from outside. And developments there certainly have been.
Ever since Japan emerged as the first major threat to America’s post WW II total domination of world manufacturing the U.S. has been doing the five stages of grief thing. We’ve gone through anger—“ungrateful jerks”; denial, “we’re still the best”; bargaining, “if Washington would just keep out imports for a few years we promise to . . . ,” and depression, “those jobs are gone and they ain’t comin’ back,” or whatever your favorite Springsteen post-industrial lament is. After the first four, though, come acceptance and making the best of things. In the four decades since the oil embargo first brought the Japanese car threat to the forefront, the percentage of the cost of an American-built car that comes from labor costs has fallen by more than half and is now well under 10%. We found it surprisingly difficult to get similar statistics for U.S. manufacturing generally but the trend is clear and overwhelming. As incomes and thus domestic demand in the new industrial behemoths of Asia, Latin America, and soon even Africa rise, and as the labor/unit cost of U.S. manufacturing goods becomes a trivial consideration, the international labor arbitrage that has been so agonizing for the U.S. is about to be reduced to an episode for the history books. A rebound in U.S. manufacturing may not bring much of an absolute rebound in manufacturing employment, though it may staunch the bleeding. But it will certainly bring rising sales and quite likely greater pricing power and higher sales margins to U.S. manufacturing firms. Isn’t that likely, along with the energy boom, to give some support to current elevated U.S. earnings?
Still as the eminent Mr. Grantham has pointed out even a very large increase in the share of income going to capital should not, in theory, permanently increase returns to capital. For even if capital no longer has to compete as strenuously against labor for income, capital still must compete with itself. If the rewards to capital invested in U.S. industry suddenly double from 10% to 20%, economic theory says that more capital will rush in from elsewhere bidding the price down again. If that’s not true, Mr. Grantham reminds us, capitalism does not work.
This is a powerful argument that under most circumstances only a fool would wantonly defy. And yet it is not quite foolproof. Let the fool offer three objections:
First, even if the return to capital is inevitably bid down this is not quite the same as return on equity being bid down. If the new factory consists of fewer men and better machines that is another way of saying it consists of more collateral and requires less in the way of unsecured capital, including equity. Earnings are thus effectively more leveraged without a corresponding increase in economic risk, facilitating a new higher equilibrium ROE.
Second, “eventually” has a lot of stretch in it. There is a big di!erence between a temporary anomaly in earnings levels that cannot be justi"ed in economic terms and a rise in earnings off a major shift in the fundamentals of the economy. Both may be temporary but one temporary is likely to be a lot longer than the other.
Finally, are we actually to believe that the average level of ROE has been divinely established and unchangeable since the beginning of time and regardless of the nature of the economy? If we had the data would it really turn out to be the case that ROE in Augustan Rome was the same as ROE during the dark centuries of barbarism and again unchanged by flowering of the Middle Ages; that it was unaffected by Columbus’s ventures on the Ocean Blue or the launching of the steamship onto those same oceans? Are global capital markets so efficient that ROE in Russia over the next fifty years will be indistinguishable from ROE in the United States, or that China’s will match both?
Nine times out of ten, nay ninety-nine out of a hundred, bet on mean reversion; bet on things being the same yesterday, today, and tomorrow; bet that the man with a vision is having visions, that the self-proclaimed prophet is a knave and his followers fools.
All too true. And, yet, sometimes it is different.
—Andrew J. Redleaf
July 30, 2012
(c) Whitebox Advisors