SEVEN
UNCERTAINTY UNDERMINES INVESTMENT
The first recovery out of recession I ever tracked, as a neophyte economist, was in 1949. I had recently joined the National Industrial Conference Board and every researcher in the shop was focused on an economic phenomenon no one had experienced for years. The American economy had, of course, faltered in 1945 as military spending fell sharply. But output of consumer goods surged from late 1945 through 1946 in a drive to fill inventories of such goods after years of shortages and rationing. The associated acceleration in inventory accumulation could not be sustained and production decelerated in late 1948 and 1949. The inventory recession, however, was short-lived. Marriages had hit a high of 2.3 million in 1946, sparking the beginning of the baby-boom generation a year later. Newly formed households extended the wartime boom in home ownership. In 1950, single-family housing starts rose to a then-unprecedented level of 1.2 million units. The sharp resurgence in building construction in general was a major force for recovery. Although I couldn’t know it at the time, that cyclical pattern of construction was to dominate the American business cycle for more than a half century thereafter.
In fact, private construction has been a major contributor to every recovery out of recession since 1949—except that of 2009. Both nonresidential and, especially, residential construction had important roles in the previous ten recoveries of postwar America. But construction as a share of GDP, after falling sharply in the wake of the economic collapse of 2008, has failed to fully recover since the recession officially ended in the second quarter of 2009 (Exhibit 7.1).1 A deep-seated reluctance of business and households to invest in projects with a life expectancy, or durability, of more than twenty years (predominantly buildings) explains virtually all of the weakness in business activity and rise in the unemployment rate following the Lehman Brothers default in September 2008 (see Box 7.1).
BOX 7.1: THE MATURITY OF GDP
To measure the degree of aversion to investing in long-lived assets, I constructed a series of the average maturity or durability of personal consumption expenditures and private fixed investment (combined, they compose almost nine tenths of total GDP). Software, according to the Bureau of Economic Analysis, lasts 3.5 years, nonresidential structures 38 years, and homes approximately 75 years. Where there are no official estimates, I fill in the missing numbers—mainly for short-lived services—haircuts one month, for example. Not unexpectedly, the resulting series parallels the share of construction in GDP closely. Other than construction, most components of GDP have recovered more or less as would have been expected in a “normal” recovery.
The business community’s willingness to invest in fixed assets, as I noted in Chapter 4, is best captured by the proportion of liquid cash flow that nonfinancial corporate businesses choose to commit to difficult-to-liquidate equipment and structures—the “cap-ex ratio.”2 It is a useful measure of the degree of business confidence about the future. It doesn’t rely on what people say but on what they do. In 2009, that ratio had fallen to the lowest peacetime annual level since 1938 (Exhibit 7.2).3 More economically relevant is what market-driven forces “explain” the cap-ex ratio. As can be seen from Exhibit 7.3, the operating rate of nonfarm business is, not surprisingly, an important explanatory variable of the measured variance. Less obvious is the very significant impact of the federal deficit, adjusted for business cycle variation, on capital investment. This displacement (crowding out) by federal deficits is covered more generally in Chapter 9. Similarly, the cyclically adjusted interest rate spread between the U.S. Treasury’s thirty-year and five-year obligations, currently the widest in history, reflects the degree of heightened uncertainty beyond five years and explains why long-lived assets have been so heavily discounted and scaled back in recent years.
Individuals displayed the same aversion to investing in long-term assets. Their disengagement was reflected in the sharp fall in the value of their purchases of illiquid homes4 as a ratio of household gross savings5 (the equivalent of business cash flow). In 2010, that ratio fell to a quarter-century low (and barely above the postwar low), reflecting the diversion of gross savings from household illiquid fixed investments to the paying down of mortgage and consumer debt and the accumulation of liquid assets. Fearful of their future finances, households curtailed committing a large part of their gross savings to an asset which might be difficult to sell, except at a loss. Signs of a recovery have been mounting as the overhang of vacant homes has been drawn down. Homes prices have recovered a third of their losses. But homeownership as share of total occupied dwelling units through the second quarter of 2013 failed to show signs of recovery.
In short, the sharp increase in the aversion to long-term asset investment risk had induced an equivalent withdrawal from long-term commitments by both households and businesses. As the global economy imploded, both groups turned on a dime and retrenched, abruptly switching from long-lived to short-term asset investment with an emphasis on cash accumulation and debt repayment.
In the boom years that preceded the crash of 2008, business executives were drawn to the higher, though riskier, rates of return on long-term investments, and households (prior to 2006) to the prospect of capital gains through homeownership.6, 7 Price bids above the latest public offered price on a home were common, as anxious buyers, goaded by herd instinct, wanted to make sure they obtained the home of their choice. Homeownership rose by more than 1.3 million units annually between 2001 and 2004,8 becoming the principal driver of the boom.
Cash flow and capital gains are the dominant factors in determining overall business capital investment. The use of cash flow is not costless, because it can be employed either to fund capital investment, to build up cash balances, or to pay down debt. The cap-ex ratio is thus significantly affected by business confidence overall, in addition to the intrinsic value of individual projects. It is high in periods of prosperity (indeed it is a major cause), and low as business slumps. The cap-ex ratio in peacetime has not exceeded 1.29 (excluding 1974) or fallen below 0.67 since 1938 (Exhibit 7.2). These data suggest that the degree of leveraging and deleveraging of investment is relatively tightly constrained, an issue I addressed in Chapter 4.
COST SAVING OUT FRONT
The significant rise in nonfinancial corporate domestic profitability from the spring of 2009 through the end of 2010 appears to have resulted almost wholly from gains in output per hour, most likely resulting from investments in cost-saving facilities.9 Throughout the prolonged expansion in economic activity between 1983 and 2006,10 capital investment was predominantly for risky, high leverage capacity expansion, and was expected to yield quite high rates of return—and it did. Cost-saving investments during those years, though moderately profitable, were markedly less risk laden than capacity-expanding outlays. In discussing the issue with business executives during those years of euphoria, I sensed that many were reluctant to engage in cost-saving investments in part because they viewed such investments as diverting cash flow from more profitable capacity-expanding investments. Furthermore, cost-saving investments that entailed discharging people, an unsavory activity, are largely avoided in times of boom and euphoria.
Following the crash of 2008, capacity-expanding capital investments declined sharply. But the backlog of potential cost-saving investments, forgone during the protracted economic expansion, offered significant opportunities for investment. Reluctance to shed workers vanished. The payoff was a major increase in profit margins and profits through the fourth quarter of 2010. (See Box 7.2.)
BOX 7.2: COST-SAVING INVESTMENTS ARE LESS RISKY
In contrast to an illiquid investment made to foster expansion, a cost-saving replacement of an existing facility is almost always a safer, more predictable wager, and thus requires a much lower rate of return to be authorized. The reason is that the potential increase in demand from a new market, a source of considerable uncertainty, does not enter the replacement facility calculation as it does for capacity-expanding i
nvestments. For cost-saving investments, analysts can take the level of future expected demand as a given, and evaluate only the alternatives to the existing facilities for achieving that output. The variance of returns to cost-saving investments therefore tends to be considerably less than the variance of returns to riskier expansion projects. Acceptable rates of return on cost-saving projects thus tend to be far lower than those required to green-light an expansion.
FULL RECOVERY THWARTED
Investment in longer-lived assets (structures) was shunned. They are far more sensitive to uncertainty than investment in shorter-lived assets (equipment and software).11 Thus, the greater the life expectancy of the investment, the higher the rate of return required to justify the outlay. That propensity has never been more pronounced, in my experience, than in the period of stunted recovery in the immediate aftermath of the crisis of 2008.
I have found the yield spread between the U.S. Treasury’s thirty-year bond and five-year note to be the most useful measure of uncertainty beyond five years, and hence beyond most effects of the business cycle. It captures the degree of uncertainty associated with expectations of inflation, taxation, climate change, future technologies, and the innumerable only partially identified events that could alter investment plans over the very long run.12
The shunning of homeownership, long-term commercial lease rental commitments, and new factories resulted in the dramatic greater than 40 percent decline in new construction, in real terms, from its cyclical peak reached in 2006 to its trough in the first quarter of 2011. Home price increases over the past year are encouraging signs that the depressing overhang of vacant single-family homes is largely behind us. Nonetheless, single-family housing starts in June 2013 were still only a third of their 2006 peak.
THE WEDGE
The collapse in construction during the years 2008 to 2011 opened up a wide chasm in growth rates between total real GDP and real GDP, excluding private structures. The share of private fixed investment going to structures declined by more than a third from 1955 to 2011. The huge shortfall in real GDP caused by this decline (Exhibit 7.6) accounted for the equivalent of a more than 2-percentage-point rise in the unemployment rate13 by the first quarter of 2011.14 It has since receded in part. The implied gap represented by the ratio of employment to the civilian noninstitutional population is far more worrisome.
The effect has been to create a two-tier economy: one comprising more than 90 percent of the economy, producing goods and services with a duration of less than twenty years, and since 2009 operating at a modestly respectable rate of potential; and another comprising the remainder of economic activity—that encompassing components with a duration of more than twenty years (almost all construction)—operating at barely more than half of potential. The distortion, as I noted earlier, resulted in sharply lower expected profit returns on business structures and imputed returns on homeownership (mostly expected home price inflation) because they were discounted at the ever increasing rates associated with longer-lived assets. Those elevated discount rates sharply suppressed the level of such investment. A significant part of real GDP, in effect, has been put on the shelf and removed from prospective output. The CBO’s real output gap, a measure of economic slack, is 6 percent (Exhibit 7.7). The shortfall in effective demand is also suppressing credit demand (see Chapter 12). Presumably, so long as the current degree of uncertainty persists, the disproportionate loss of long-lived asset investment is likely to frustrate full recovery.15
AVERSION TO INVESTMENT
That business had become markedly averse to investment in fixed long-term assets appears indisputable. The critical question is why? Although most in the business community attribute the massive rise in their fear and uncertainty to the collapse of economic activity,16 many judge its continuance since the recovery took hold in early 2009 largely to be the result of widespread government activism in its all-embracing attempt to accelerate the path of economic recovery and regulate finance. The evidence tends to largely support the latter judgments.
POLICY DISAGREEMENTS
In these extraordinarily turbulent times, it is not surprising that important disagreements among policy makers and economists have emerged on the issue of the size of government and the extent of policy activism. Almost all agree that activist government policy was necessary in the immediate aftermath of the Lehman bankruptcy, when many critical overnight markets ceased to function, arguably a once-in-a-century event. This necessitated special policy treatment because it was a unique market breakdown that did not occur in pre-2008 financial crises where prices crashed but market structures remained intact.17 The tendency for market prices to seek equilibrium cannot occur if market structure is dysfunctional. The powerful forces of free market adjustments fail. Under such circumstances, the substitution of sovereign credit for private credit is essential to quickly restore market structure. The U.S. Treasury’s equity support of banks through the Troubled Asset Relief Program (TARP) and the Federal Reserve’s support of the commercial paper market and money market mutual funds, for example, were critical in arresting the free fall. All financial markets, however, were functioning again by early 2009.
But the government activism as represented by the massive 2009 $831 billion federal program of fiscal stimulus18 (the American Recovery and Reinvestment Act, or ARRA),19 in addition to housing and motor vehicle subsidies, and innumerable regulatory interventions, continues to be the subject of wide debate. The evidence of this once-in-a-lifetime event is such that policy makers and economists can harbor different, seemingly reasonable expositions of the forces that govern modern economies. At root these differences reach down to the fundamentals of how economies—free market based or otherwise—operate. If common agreement on how capitalist markets function existed, I presume the differences among economic policy makers would have long since been resolved. Regrettably, that is not the case.
THE DEEP DIVIDE
Most difficult to bridge are the differences among forecasters, or economists more generally, with respect to the broad conceptual framework of models of how all the elements of a market economy interact and how the economic world works. But many, if not most, of the key interrelationships are not directly observable—such as in a simple case of the shape of supply and demand curves that create prices and interest rates. We never observe the internal market process directly because it exists only as abstract mathematical constructs of economists. Perhaps the most important divergence between “liberal” and “conservative” economic policy makers is the extent to which each view market-driven economies as competitive, flexible, and hence self-correcting. Programmed government expenditures and taxes (fiscal policy) or central bank interest rates (monetary policy) exhibit unobservable channels of causation through which these policies affect economic outcomes. In short, all forecasts of economic developments ultimately rest on how each analyst fits the ex post facts he or she observes in the marketplace into an all-encompassing, unobservable ex ante construct of the way an economy works toward its observable ex post data.20
A rare partial insight into how financial markets under extreme stress work behind the scenes became accessible as the 2008 financial crisis unfolded. A half century ago, economists had envisioned a risk distribution of market outcomes represented by a bell curve (that is, a normal distribution) whose outcomes were determined wholly by chance, such as coin tossing. Later in the postwar years, it became apparent that there were more fat-tail events—that is, extreme outcomes previously perceived as low probability—than would be expected from a normal distribution of expected outcomes. On the morning of October 19, 1987, for example, prior to the opening of the New York Stock Exchange, the probability of a more than 20 percent decline in stock prices that day, eclipsing any previous one-day collapse, was at least ten thousand to one, and more likely a million to one. It happened. And during the lead up to the crisis of 2008, many events occurred that could not easily be explained by any stretch of conventional wisdom.
Economists, chastened by many such back-to-back highly improbable economic outcomes, tilted toward assuming more fat-tail distributions. But as the years 2007 to 2009 traced out the outcomes of the crisis, we came to understand that the “extreme” tails of probability outcomes such as 2008 were downright obese—“highly improbable” catastrophic market breakdowns had begun to happen too often. Examination of the data has altered previous views of the way financial risk operates—mine in particular.
It is dismaying when the real world runs counter to a cherished paradigm or, in the words of Thomas Huxley, the famous nineteenth-century biologist, commits “the slaying of a beautiful hypothesis by an ugly fact.”21 Physical scientists go through the same routine as economists and other social scientists, but the world with which they deal is far more stable and predictable than the economic world.22 The vagaries of human nature do not affect the physical world. The laws of physics, for example, once identified, rarely have to be revised. Yet there are still famous differences among physicists. Albert Einstein never felt comfortable with the probabilistic nature of quantum mechanics. As he famously said, “God doesn’t play dice with the universe.” He sought certainty in the world, not probability.
ENTER POLITICS
As I’ve noted, at the very center of the postcrisis political confrontations is the debate over whether free markets are ultimately self-correcting, or as a large number, perhaps a majority, of economists and policy makers currently believe, the shortfalls of faulty human nature‒driven markets require significant regulatory direction, periodic fiscal stimulus, and a vast safety net for those who fall through the cracks of a largely self-regulated market system.
The Map and the Territory Page 13