The Map and the Territory

by Alan Greenspan

  THE MAGIC 3 PERCENT?

  With the exception of the immediate postwar years,12 output-per-hour growth in most advanced economies appears to have been subject to the 3 percent growth ceiling.13 But why couldn’t the current level of technology and productivity have been achieved in, say, 1960, rather than a half century later? The answer appears to be that we human beings are apparently not smart enough to have produced such a leap.

  I suspect that the relatively stable rate of growth of productivity that we observe from 1870 to 1970, despite the persistent degree of market churning, reflected a combination of the long-term unchanging inbred rate of time preference, which sets the time horizon for our innovative initiatives, and our inbred propensity toward optimism and competitiveness. They, I assume, are capped by the upside limit of human capability to create and apply knowledge over the long run. Certainly there is nothing to demonstrate a major difference during the past two millennia in the degree of intelligence of, for example, Euclid, Newton, and Einstein, the icons of outer-edge human intelligence of their respective eras.14 Technology may accumulate, but given the apparent ceiling to intelligence, the pace of knowledge accumulation, of necessity, is limited.

  MULTIFACTOR PRODUCTIVITY

  The widely accepted paradigm for evaluating and projecting productivity is called multifactor productivity (MFP).15 Simply put, MFP is a change in output that cannot be accounted for by changes in the combined inputs of labor and capital. It is presumed that an increase in MFP is owed mainly to improvements in technical efficiency and to the new ideas that lie behind those improvements. We use the term “technology,” or, broadly speaking, applied “innovation.” MFP will also reflect increased timeliness in the transportation of people and goods; advances in communication capabilities; improved efficiencies in the use of energy, materials, and services; economies of scale; better ways to organize production; and finally, the increasingly important use of robots that sustain the level of production but significantly decrease the number of human hours required to produce the output. It would be useful if we could quantify these cost reductions in the official estimates of MFP. (Not to be readily dismissed as a factor is measurement error. It can sometimes be disturbingly large.) Applied innovation is a major component of output per hour, as can be seen in the equation in endnote 19 of this chapter.

  My former colleagues at the Federal Reserve, in one of their many related analyses, documented the prominent role of both the production of information technologies and their increasing use in the acceleration of labor productivity between 1995 and 2006.16, 17

  THE METRICS

  Multifactor analysis can be employed to forecast business output and output per hour given projections of hours, an adjustment to labor quality (the educational attainment of the workforce, for example), capital services,18 and finally the extent of applied innovation (MFP).19 Capital expenditures, their level and mix, and applied innovation statistically prove to be the most important determinants of labor productivity growth.

  The share of GDP devoted to capital investment peaked in the late 1970s at around 25 percent and has since trended steadily downward to a low of 18 percent in 2009. The major reason, as I note in Chapter 9, is the dramatic decline in the rate of domestic savings, the prime source of funding of domestic capital expenditures (along with borrowed savings from abroad). The trade-off between social benefits (mainly Social Security, Medicare, and Medicaid) and savings, as I demonstrate in Chapter 9, is almost a dollar-for-dollar substitution of consumption20 for gross domestic savings. That detracts from the funding of capital investments, and hence from output per hour.

  The diversion of the flow of savings of households from capital investment to consumption depressed the rate of growth of the stock of productive capital, a key component of productivity growth. As I note in Chapter 9, the pace of growth of output per hour has slowed since the mid-70s, but remained at a still respectable 2 percent per year until 2010, when it dipped to a less than 1 percent annual rate.

  MFP grew at an annual rate of 2.1 percent between 1947 and 1965. That pace almost surely reflected the buildup of a large backlog of applications of technological insights that could not be applied during the suppressed environment of the 1930s, and, of course, during World War II. Apparently having run through the significant backlog of new ideas, the process significantly slowed its pace of rise from 1965 to 1995. But for the next decade it regained momentum as the Internet’s role in information technologies during the dot-com boom broadened, and the pace of new insights and applications accelerated at close to the early post–World War II level. The crisis of 2008, of course, put a hold on many (but by no means all) innovative applications.

  FORECASTING INNOVATION

  Forecasting innovation is particularly challenging largely because innovation (an idea or application of an idea), by definition, is something which nobody has previously discovered. In that sense, innovation is presumably unforecastable. It is very easy in retrospect to identify the emergence of telegraphy as a major innovation in the nineteenth century. But would it even have been imaginable to anticipate it before the identification of electricity? Certainly before the electrical properties of silicon were discovered, no one would have imagined its monumental implications for the future.21

  I suspect that the only forecasting assistance historical data can offer is to spell out the ranges of growth in innovation, say, over the past half century. We can generalize that the necessary environmental conditions for the blossoming of innovation are the enforcement of property rights and other political and economic conditions that have proved conducive to innovation. Propensities to innovate can be thwarted by culture, religion, or state repression.

  EBBS AND FLOWS OF PRODUCTIVITY AT HOME AND ABROAD

  For much of the near century and a half that U.S. output per hour has been more or less reliably measured, shortfalls in productivity growth have been followed by catch-ups that brought its level back to its long-term trend. But the unprecedented slowdown in savings starting in 1965 eventually funded a persistent below-average growth of capital investment and hence of nonfarm output per hour between 1973 and 1995, with estimates of growth closer to a 1.4 percent annual rate than to the average annual rate between 1870 and 1970 of 2.2 percent.

  Throughout much of the 1970s and 1980s, Japan was heralded as the economy that would soon displace the United States as the world’s productive powerhouse. Herman Kahn’s highly influential 1971 treatise The Emerging Japanese Superstate: Challenge and Response22 was close to subsequent conventional wisdom. That perception was reinforced by the relatively lackluster performance of productivity in the U.S. economy and the apparent dynamism and productiveness of Japan. Lying ahead was the dot-com dynamism of the United States and the prolonged post–1989 stagnation of Japan.

  Now there is an emerging belief that China will soon be displacing the United States not only in the level of GDP but eventually even in per capita GDP as well. That conclusion is by no means self-evident. China’s spectacular growth in recent years rests on technology largely borrowed from the rest of the world. Little to date, apparently, has been homegrown. China’s businesses have yet to demonstrate a capacity for innovation on a par with those in many advanced economies, most notably those of the United States. In a November 2011 study, Thomson Reuters identified what it saw as the one hundred most innovative global companies. No Chinese company was on the list; America had forty.23 As I note in Chapter 10, it is not surprising that an authoritarian state that discourages political and other nonconformity will not sanction a climate that fosters ideas wholly out of the mainstream. Yet such ideas are the defining characteristic of innovation. It is precisely those who tread beyond the bounds of conventional behavior who innovate. Yes, growth in Chinese output per employee is reported at a stunning 9.5 percent annual rate between 1990 and 2011. But much, if not most, of Chinese innovative technology originated in developed-country firms. China is already running into rising real wage costs th
at have significantly narrowed their competitive advantage in global manufacturing. Lurking in the background is a rapidly growing robotics industry that substitutes low-“wage” robots24 for people. Even China will have trouble competing with them. As Japan found out, forecasting innovation is precarious.

  IT TAKES TIME

  Most productivity increases are incremental. It takes time for new ideas to filter through the process of trial and error and meet the test of the marketplace before adding to the level of a nation’s capacity to produce.25

  Forecasting productivity from the pace of new inventions and the level of patents has always proved challenging (see Box 8.1). Only if an economic infrastructure is ready to capitalize on an innovation can that innovation percolate up through the economy to add to productivity. According to economic historian Paul David of Stanford University, writing in 1989, it often takes a number of decades for major innovations to be applied in a manner that increases output per hour. As I noted in The Age of Turbulence, following Thomas Edison’s spectacular illumination of lower Manhattan in 1882, it took some four decades for even half of the nation’s factories to be electrified. Electric power did not fully exhibit its superiority over steam power until a whole generation of multistory factories was displaced after World War I. David explains vividly what caused the delay. The best factory buildings of the day were poorly designed to take advantage of the new technology. They ran on so-called group drives, elaborate arrangements of pulleys and shafts that transferred power from a central source—a steam engine or water turbine—to machines throughout the plant. To avoid power losses and breakdowns, the lengths of the shared drive shafts had to be limited. This was best achieved when factories rose vertically, with one or more shafts per floor, each driving a group of machines.26

  Simply substituting large electric motors to power the existing drive shafts, even when feasible, did not improve productivity very much. Factory owners realized that electricity’s revolutionary potential would require far more dramatic change: Power delivered by wire made central power sources, group drives, and the very buildings that housed them obsolete. Because electricity opened the way to equipping each production machine with its own small, efficient motor, sprawling single-story plants came into vogue. In them, machinery could readily be arranged and rearranged for greatest efficiency and materials could be moved about with greater ease. But abandoning city factories and moving to the wider spaces of the countryside was a slow, capital-intensive process. That was why, David explains, electrifying America’s factories took dozens of years. But eventually millions of acres of one-story plants embedding electric, motor-driven power dotted America’s midwestern industrial belt, and growth in output per hour finally began to accelerate.

  In more microexamples of absorption delay, color fashions in everyday clothing did not become widespread until vat dying became commonplace in the 1930s, and, of course, truck transportation did not take on its full potential until a highway system was created to facilitate its commerce after World War II. Moreover, highways fostered the development of suburbs that, in turn, improved the productivity of land use. The value of land almost always rises as its use shifts from sparsely settled rural environments to today’s population concentrations in urban areas. Denser populations tend to further refine the division of labor and, accordingly, boost incomes and increase the value of the land on which people live.

  Thus, much of today’s level of productivity rests on insights that were spawned decades ago but could not be capitalized upon until other insights facilitated their integration. A vivid example of backlogged innovation emerging as increased output per hour occurred following the sharp contraction in economic activity that resulted from the debacle of 2008, as I explain in Chapter 7. The surge in output per hour from the first quarter of 2009 through the first quarter of 2010 reflected the dramatic crisis-driven efforts to cut costs and concentrate on cost-saving capital investment, a backlog of which had been built up during the boom years when capital investment was primarily dedicated to expanding markets rather than reducing costs (see Chapter 7).

  BOX 8.1: PATENTS

  What I find puzzling is that we gain so little insight into productivity forecasting from data on patents that the federal government has tabulated since 1790. Patents unquestionably cover much of the innovation that in the end is reflected in gains in productivity. Accordingly, I would have expected that they would be a useful indicator of forthcoming productivity gains. The result of the analysis that I exhibit in Statistical Appendix 8.1 is disappointing: Patent issuance parallels but does not, as might have been expected, statistically lead new additions to the level of productivity.

  ENTER FINANCE

  The speed of absorption of innovations often depends importantly on the efficiency of financial markets in signaling which of a whole array of prospective investments will add to productivity and hence profitability.

  As I noted in Chapter 5, the purpose of finance is to direct the scarce savings of a society, including depreciation plus borrowed savings from abroad,27 if any, to our most potentially productive intellectual and physical investments. (In Chapter 10 I address the determinants of savings and the culture that breeds it.) When competition is pervasive and allowed to function,28 those investments with the highest rates of prospective return and least variance of expected profit also promise to contribute the most to growth in output per hour. It is the gap between the level of output per hour embodied in new capital investment, presumably employing up-to-date technologies, and that of obsolescent, low-productivity facilities (that are gradually being retired) that over time engenders net gains in average output per hour, and hence increases in real per capita income and standards of living.

  HOW IS IT DONE?

  But how do profit-seeking savers or their surrogates, such as hedge funds or banks, know where to place new savings? They are guided by the signals of the marketplace: stock prices, asset prices more generally, interest rates, exchange rates, and the whole panoply of information that pours out of research departments of financial firms, governments, and academia. The market signals, in turn, are driven by the chronic imbalances endemic to financial markets. These imbalances arise because some prospective investments are being inadequately financed relative to competing investments with similar risk profiles, and hence the latter yield abnormally (uncompetitive) high profits in the marketplace. Rising stock prices, reflecting a shortage of invested capital relative to profitable performance, attract new capital inputs until demand to acquire the stock of a company is sated as the company becomes adequately funded. “Overvalued” and hence overcapitalized companies, on the other hand, will yield low prospective rates of profit until the excess capital is withdrawn and presumably reinvested in more promising ventures.29

  These investments address the initial capital misallocation and, because of the associated expanding or diminishing returns, continually adjust the investment’s prospective rate of return on equity to competitive levels. This complex process is constantly churning as new information on supply, demand, and price emerges out of the ongoing competitive trading of markets and the arbitraging of all prospective rates of return (adjusted for risk) toward equality. That equality, however, is never fully achieved because new forces invariably intercede before the adjustment is complete.

  The equilibria that markets are seeking and generally closely approximate are those governed by human propensities best described as an admixture of neoclassical and behavioral economics. While over the long run, market prices do seem to converge on values defined by humans’ rational long-term self-interests, as described by the neoclassical school, it is decidedly not necessarily so in the short run.30

  Indeed, as the episode of the dot-com bubble made clear, on occasion, euphoria combined with herd behavior can result in financial markets that for a time become divorced from realistic evaluations of future corporate earnings prospects. During this period of dot-com euphoria, funds flowed to virtually any
activity with a “dot-com” affixed to its name, whether or not the firm had a reasonable and realistic business plan and any significant prospect for making a profit in the foreseeable future.

  The dot-com productivity acceleration, as my colleagues at the Federal Reserve Board demonstrated,31 was primarily driven by major advances in information-processing equipment and software. First, the industry itself exhibited outsized gains in output per hour. But of even greater importance, the rapid application and absorption of such technologies had a major effect on virtually every niche and corner of American business and on households as well. Fed economists Stacey Tevlin and Karl Whelan32 have demonstrated the implications for productivity measurement of the rapid declines in prices of information-processing equipment and software, especially computers. After a long period of relative price stability, prices fell at a 1.5 percent annual rate between 1980 and 1985, accelerating to a 3.9 percent rate decrease between 1985 and 2010, then slowing to a less than 1 percent annual rate of decline during the last two years (Exhibit 8.4). If profit margins in the production of information-processing equipment and software were essentially stable, as they apparently were, unit costs would, of necessity, be falling pari passu. Adjusting for wage rate changes, that implies an extraordinary rise in productivity. The dot-com boom was very largely pure innovation and would be expected to enhance MFP, as it did. Price declines for information-processing equipment and software have correlated remarkably well with MFP over the last three decades (Exhibit 8.4).

  Innovation, as a driver of productivity, has always been a key component of rising output per hour, though not always as dramatically as was evident in the few years of the dot-com boom. Manufacturing technologies that reduce the rate of scrappage and rejects and hence reduce materials input, for example, are an insufficiently heralded source of lowered unit cost and rising output per hour. Prior to the 1950s, for example, steel mill products’ yield from ingot averaged less than 75 percent. Ingots had to be cropped before entering the rolling mills. But with the advent of continuous casting, “home scrap” (that is, internally generated scrap) has fallen dramatically.