Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Rethinking Macro Measurement

Rethinking Macro Measurement VOLUME 32 NUMBER 4 FALL 2020 Journal of AP PLIED COR PORATE FINANC E IN THIS ISSUE: Public James Sweeney, Chief Economist, Credit Suisse 17 The First Modern Financial Crises: The South Sea and Mississippi Finance and Bubbles in Historical Perspective Robert F. Bruner, University of Virginia; and Scott C. Miller, Yale University Central Banks Convergence and Reversion: China’s Banking System at 70 Carl Walter Apocalypse Averted: The COVID-Caused Liquidity Trap, Dodd-Frank, and the Fed Craig Pirrong, University of Houston The Poverty of Monetarism Patrick Bolton, Columbia University Columbia Business School Roundtable on Broken Models of Public Finance Panelists: Jared Bernstein, Center for Budget and Policy Priorities; and Paul Kazarian, Japonica Partners. Moderated by Shiva Rajgopal, Columbia University 82 Columbia Business School Roundtable on The Fed’s Response to the Global Financial Crisis—and Now the Pandemic Panelists: Frederic Mishkin and Patricia Mosser, Columbia University. Moderated by Kate Davidson, The Wall Street Journal The Euro @ 20: How Economic and Financial “Asymmetries” Marred the Promise of the Single Currency George Alogoskoufis, Athens University of Economics and Business; and Laurent Jacque, Tufts University The Benefits of Buying Distressed Assets Jean-Marie Meier, University of Texas at Dallas; and Henri Servaes, London Business School, CEPR, and ECGI 117 Using ESG to Enhance Fixed-Income Returns: The Case of Inherent Group Nikhil Mirchandani and Chelsea Rossetti, Inherent Group The Economic (Not Literary) Offenses of Michael Lewis: The Case of The Big Short and the Global Financial Crisis Don Chew, Journal of Applied Corporate Finance Rethinking Macro Measurement by James Sweeney, Chief Economist, Credit Suisse new empiricism is sweeping through macroeconomics. It is based on an old under- A standing—that some widely followed and near universally accepted economic indicators do not do a good job of capturing the concepts with which most of the general public associates them. It is also based on new methods of gathering data. The COVID-19 pandemic has contributed to increased aware - I will show that in many ways the national accounts ness of these developments. Some macroeconomic indicators data are valuable, despite their limitations. Still, there is have proved to be too slow or imprecise to explain what was much room for new measurement approaches that can be happening in real time. Meanwhile, a range of new indica - employed alongside while even making use of the existing tors, often higher frequency and based on “bigger” data, have numbers. With the new possibilities created by big data provided genuinely useful guidance. and rising computational power, the social importance of National accounts data that have long been widely used GDP and related measures has probably already peaked, and to represent concepts such as real income, inflation rates, new measures are reshaping economic thinking. A greater and productivity are now being questioned more than ever emphasis on “administrative data” and the use of “quasi- before. The appropriate scope, inherent biases, and potential experimental evidence” is now clear in the business world misuse of the national accounts have long been controv-er and in the empirical academic literature. sial.1 In recent years, these debates have gone mainstream, Although we cannot precisely predict how a macro revolu - as widely perceived changes in the rates of technological, tion will play out, and the extent to which it might lead to economic, and social change have not always matched changes to orthodox theory, understanding the problems with what’s in the numbers. traditional data helps us stay open-minded about whether One response has been a proliferation of altern - a popular views about recent economic performance might in tive measures. These include new indicators of everyday the future appear to have been misguided. phenomena, including foot traffic and web traffic data, tax return data, app data, and a wide range of company- Mismeasurement specific data. There is work on disease-based price indexes, Let’s start with a simple question. Could widely accepted facts happiness measures, “billion price” aggregates, and activity on recent economic performance be wrong on account of bad trackers sourced from satellites. There are also attempts to data? If we compare U.S. performance in the period from “correct” the traditional measures with ad hoc adjustments 1995 to 2007 with the period from 2008-2019, we arrive at or by switching emphasis toward less well-known variants the following propositions: of the main measures. 1. N ominal GDP growth has slowed (from 5.4% p.a. to I will focus here on mismeasurement and misuse of the 3.4%) national accounts and suggest some ways to make better 2. Working age population growth has slowed (from decisions once these are taken into consideration. Mismea - 1.3% to 0.5%) surement refers to whether a statistic is an accurate measure 3. Nominal GDP per capita and per worker have slowed of what it purports to measure; misuse refers to whether a (from 4.4% to 2.7%) concept associated with a statistic is appropriately invoked. 4. e Th share of income going to labor has continued a Some concepts are just not easily quantie fi d. secular decline (from 55.1% to 53.1%) 5. Core ina fl tion has slowed slightly (from 1.8% to 1.6%) 1 Diane Coyle (2014), GDP: A Brief But Affectionate History. Princeton, NJ: 6. Real GDP growth has slowed (from 3.2% to 1.7%) Princeton University Press. 8 Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 7. Labor productivity (real GDP per hour) has slowed themselves for living in their own homes. Estimation relies (from 2.1% to 1.1%) on many assumptions, such as whether local rents are a good 8. Business investment has fallen (from 6.3% to 3%) proxy for what owner-occupied housing would rent for if 9. Total factor productivity has slowed (from 1.4% to put on the market. 0.5%) Shortcomings aside, nominal GDP data serve many In short, the U.S. economy appears to be sue ff ring from purposes well because their source components are usually both a demographic and a productivity slowdown. Ina fl tion collected as nominal figures before being adjusted and aggre - has fallen and investment has been lackluster. However, not all gated. Data are collected in three categories: expenditure, of these claims are equally supported. This is not just because income, and value added (profit plus labor cost), with expen- of differences in the measured magnitudes, but also in the diture figures usually arriving first. These three perspectives reliability of the measures themselves. should sum to roughly the same total, so national accoun - The first three propositions are based on nominal and tants can cross-check their estimates, even if they never have demographic data, which most of my macro colleagues and I complete data from any of the perspectives. Large revisions consider to be highly accurate. Controversies about nominal to GDP often follow when tax data become available. The GDP measurement are mostly about scope: what should be usual stability of the ratio of collected taxes to nominal GDP counted? This has been debated for most of the past century, provides some evidence that nominal GDP measures are but the main conclusion is settled: GDP measures market reasonably accurate. transactions without making (or purporting to make) any attempt to capture all valuable activity. GDP thus ignores home production and free goods. And for this reason, swings “ in female and child participation in the market economy have Problems arise when we use market GDP data to significantly impacted GDP over time, as activities shifted stand in for broad social concepts like productivity between paid and unpaid work. For example, as one economic historian has pointed out, and value added. “Cooking must have been the most widespread, and generally time-consuming, expenditure of labor throughout history.” 2 But cooking, retrieving water from a well, and washing clothes outside have never been counted in GDP—unless they were part of a market transaction; that is, outside help was hired, Our fourth proposition is about the labor income share. It or the eating was done in restaurants. comes from the income accounts and can be supported with In sum, many valuable but “free” goods and services are surveys and tax return data. After massive recent public atten - ignored because they do not involve market transactions. And tion to wealth and income inequality, it has become widely in the modern economy, the importance and value of free accepted that, during the past three decades, the pretax income search engines and social media are self-evident, as they are distribution has moved sharply in favor of very high earners for many other freely provided goods and services, such as in the United States and several other countries. And as crowd-sourced trac ffi apps, which create value by getting us inequality within labor income has worsened, labor’s overall to home and work faster. share has fallen—developments that have long been ree fl cted You can’t fault an accounting system for failing to measure in media and political debates. what it does not claim to measure. Nevertheless, problems Propositions 1-4 are well enough established to be called arise when we try to use market GDP data to stand in for “facts,” and though measurement issues exist, they do not broad social concepts like productivity and value added, a undermine the use of these data for many purposes. What subject we turn to later. distinguishes them from propositions 5-9 is that they are all There are a few specific measurement difficulties for nominal concepts. For calculations that rely on inflation adjust - subcomponents of nominal GDP that are worth noting ments, our confidence in their accuracy should be reduced to here. Take “imputed rents,” which make up about 8% of reflect both measurement and conceptual challenges. Real GDP U.S. GDP. This is what homeowners are assumed to pay growth equals nominal growth minus inflation. To the extent inflation measures are unreliable, real growth measures are too. And since ina fl tion measurement is the heart of our topic, we 2 Jurgen Osterhammel (2014), The Transformation of the World: A Global History turn to it first, and give it extra attention. of the Nineteenth Century. Princeton, NJ: Princeton University Press. Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 9 Inflation Bias weights, price time series for each (often changing) good, Many financial market participants talk as if inflation data and sometimes quality adjustment measures. Each of these are precise. But consider this excerpt from the transcript of involves considerable dic ffi ulty and numerous assumptions. an FOMC meeting on July 2-3, 1996: Economics students learn about two types of inflation indexes, which use different weighting formulas: Laspeyres CHAIRMAN GREENSPAN: Is long-term price stability an and Paasche Indexes. The U.S. deflator for personal consump - appropriate goal of the Federal Reserve System? tion expenditures, known as the “PCE inflation index,” uses the Fisher Ideal method, which is a geometric average of the MS. YELLEN: Mr. Chairman, will you define “price stabil - two. The U.S. consumer price index (CPI) uses a version ity” for me? of the Laspeyres method with fixed weights periodically adjusted after spending surveys. e Th CPI is never signic fi antly CHAIRMAN GREENSPAN: Price stability is that state in which revised, but the PCE inflation index is revised often, and expected changes in the general price level do not ee ff ctively often substantially. alter business or household decisions. Different vintages of the core PCE deflator show that revisions can push measured inflation as much as 0.5% away MS. YELLEN: Could you please put a number on that? from the original estimate.4 These revisions do not affect the CPI, even though that is the source of much of the PCE CHAIRMAN GREENSPAN: I would say the number is zero, if deflator’s data. Not revising the CPI prevents some annoy - ina fl tion is properly measured. ances because it is used to index government benet fi s and inflation-linked bonds. The bond market reacts strongly to MS. YELLEN: Improperly measured, I believe that heading both indexes, and Fed officials have said that they focus on toward 2 percent inflation would be a good idea, and that the PCE deflator, partly because of its flexible weights and we should do so in a slow fashion, looking at what happens revisions. along the way. My presumption based on the literature is, Measurement uncertainty should cause market partici - as Bob Parry summarized it, that given current inaccurate pants and policymakers to reduce their confidence in measurements, heading toward 2 percent is most likely to be short-term inflation observations. Several years ago, when benec fi ial. core U.S. PCE inflation was less than the market expected, the Dallas Fed argued that “it’s likely that these figures will be CHAIRMAN GREENSPAN: Could we leave it at that? Let’s now revised higher… at the next annual revision in summer 2018.” move on.3 And, indeed, signic fi ant upward revisions did later occur. During the 2020 COVID-19 pandemic, the measured This exchange between sitting and future Fed chairs CPI and PCE ina fl tion both fell sharply, but large changes in occurred during a discussion on whether to use monetary the actual composition of underlying spending raised unc- er policy to drive inflation permanently to 0% following years tainty about the reliability of these real-time measures and the of declining measured ina fl tion. Before the 1970s, peacetime possibility of future volatility and revisions. In my view, it is inflation had probably been near zero for most of U.S. history. clear that the pandemic has resulted in short-term downward One reason the Fed, along with a number of other central pressure on ina fl tion, especially driven by falling commodity banks, chose 2% is that it would allow more room to cut prices. But the imprecision of the inflation data is even greater interest rates when needed. Another was that research had than usual, and the many product-specific supply and demand suggested that inflation data were upwardly biased. “One shocks that have occurred have likely led to much “noisier” can argue that we have roughly a 1 percent bias in the CPI,” component level data. Yellen said in the same meeting. This exchange, and the studies Obtaining time series of price data for specific goods as that emerged from the 1990s Boskin Commission, show that inputs to the inflation index sounds like an easy task, especially Greenspan, Yellen, and others were well aware of the impr- eci for something like a pound of bananas. But exactly when sion of ina fl tion estimates. should a new good be included? VCRs were first sold for more Ina fl tion indexes are much harder to estimate than price than $1,000 in the late 1970s, but they did not appear in the changes through time of a particular good. An index requires 4 Incidentally, recent large revisions to PCE have tended to come from financial 3 https://www.federalreserve.gov/fomc/minutes/19960702.htm. Pages 50-51. services prices. 10 Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 CPI until 1987, when prices were down to a few hundred quality adjustments—have created biases and uncertainties dollars and most households already owned them. 5 The prior in inflation measures. We have focused so far on consumption price declines never affected measured inflation because VCRs goods, but GDP inflation includes the prices of investment were not in the index when their greatest dea fl tion occurred. and traded goods, too, leading to more room for error7 . Goods evolve, sometimes into performing entirely Not all biases are in the direction of underestimating different functions. Today’s smartphones are yesterday’s output and overestimating inflation. For example, when cameras, fax machines, televisions, tape recorders, music companies substitute cheaper foreign-sourced inputs for players, handheld video game players, pocket organizers, domestic ones, output is actually overstated. And many alarm clocks, and many more things—some of which are professionals in the financial services industry profess to new. If smartphone prices were compared to the total prices believe that inflation is understated. But this view can, and of all of those things in the past, then phones would have should, be dismissed as the working of local or “availability” be shown to have deflated massively. bias: individuals living in areas with particularly strong income To be sure, phone prices have fallen a lot, and the cost of growth witness many firsthand examples of conspicuous price transmitting telecommunications data—a service, distinct increases—think restaurants, rents, and tuition. But econo - from the hardware—has dropped too, but there are other mists would identify these as changes in “relative prices,” not forms of “deflation” in a smartphone that have not been signs of national ina fl tion. captured. Statistical agencies simply cannot capture it all, but Physical goods constitute about one-third of consumption they are always working on ways of improving their measures. in the developed markets, and the rest are services. Short- But as a result of this mismeasurement, phone “inflation” is term inflation expectation surveys are strongly correlated with almost certainly measured and reported to be higher than it energy prices, the most visible price changes in the economy. could be, and “real expenditure” on phones is understated. In general, when people think about inflation, they think So-called “hedonic” adjustments—those which control of goods prices, even though housing, health care, financial for the change in certain attributes of a good—are one way services, and other services involve much more spending, with of dealing with this issue, but such adjustments are used prices that are often opaque. in only a subset of the goods and services whose quality Health care prices are particularly hard to collect. Health changes. Automobiles are a well-known case where hedonic care spending is 17% of U.S. GDP. This figure comes from adjustments cause the national accounts prices to differ from adding the industry’s reported profits to its workers’ wages. retail prices. Carving nominal health care spending into ina fl tion and real Technology prices can involve hedonic adjustments for growth is a formidable challenge. For Americans, discovering attributes such as memory and processing speed. During the the price of specific health care services is often equally formi - tech boom of the late ’90s, jumps in processing speed amid dable! But the government uses surveys and special accounts high chip and hardware production in the United States to index prices to arrive at the sector’s ina fl tion. prompted hedonic adjustments that, by pushing down the Health care prices have risen faster than overall inflation in prices of such goods, led to significant increases in estimates most years. At first, one might view this as confirming evidence of real output, and hence a reported surge in productivity. of the notorious U.S. inefficiency in this sector—until one As much of the hardware industry left the United States recognizes that health care relative prices tend to rise in other after 2000, U.S. data lost the benefit of this fast-growing, economies too, and in general, the richer the economy, the fast-deflating production, one significant factor in the lower higher the health care spending. This phenomenon is in part a rates of real growth reported in the past decade. This failure reflection of “Baumol’s cost disease.”8 Factory-produced goods to take consistent account of such changes has profoundly get cheaper with technology; but because the principal actors impacted—and I would argue distorted—industrial produ-c in this case, the doctors, are in fixed supply, relatively speaking, tion data by emphasizing some productivity gains while the price of doctors’ services tends to rise relative to those of completely missing others.6 manufactured goods. And the higher the relative productiv - A whole series of statistical challenges, then—those ity in manufacturing, the stronger this effect is likely to be. arising from weighting issues, price series continuity, and 7 There are separate literatures on biases in import and export prices, and invest- ment goods’ prices. 5 http://www.nber.org/chapters/c6072.pdf. Page 388. 8 William Baumol and William Bowen (1966), Performing Arts, The Economic 6 https://www.brookings.edu/wp-content/uploads/2016/06/us-manufacturing-past- Dilemma: A Study of Problems Common to Theater, Opera, Music and Dance (New and-potential-future-baily-bosworth.pdf. York: Twentieth Century Fund). Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 11 The sharp increase in the use of online doctor visits during ling arguments that the statistics for that era were incapable of the COVID-19 pandemic has the potential to bring about accounting for the true deflation once quality adjustment is an interesting short-term reversal in this situation, allowing considered. To see the problem, it helps to consider William some costs to be squeezed out of health care. Nordhaus’s famous analysis of the historical cost of a lumen But the big story in health care ina fl tion has to begin by of light, in which he observed that recognizing the quality improvements that have occurred, and the inability of the data to capture those innovations, that have a comparison of the pure price of light with a traditional taken place over time. In 1900, world average life expectancy light price indicates that traditional price indexes overstate price at birth was near 30; today it is over 70. Infant mortality has growth, and therefore understate output growth, by a factor declined by more than 90%. There has been great progress between 900 and 1600 since the beginning of the nineteenth recently against chronic diseases, and in the past century, most century. This finding suggests that the “true” growth of real wages infectious diseases have been conquered—notwithstanding and real output may have been signic fi antly understated during the recent appearance of one particularly infectious one. the period since the Industrial Revolution.11 Cold numbers on life expectancy and infant mortality are meant to stand in for millions of smiles and memories that In Nordhaus’s view, the keepers of the national income could not have occurred before the 20th century. The topic is accounts (and macroeconomists) must find a way to measure too overwhelming to summarize with simple stories, but to the price of light, not of light bulbs. In other words, to grasp the potential impact of proper “quality adjustment” for understand the value of the goods and services created by an health care data, try imagining yourself as the patient awaiting economy, we would ideally like to know the demand curve a 19th-century surgery, pre-anesthesia, pre-germ theory, with for specific quantities of what consumers are seeking, such a surgeon standing over you with saw and scissors. as lumens of light. This would make possible true quality- Of course, quality adjustments are not much applied to adjusted measurement. health care prices, although the U.S. Bureau of Economic Another example: two centuries ago, only the rich or Analysis is experimenting with disease-based cost indexes that powerful could demand music. They had to pay a string try to measure prices in terms of the cost of conquering or quartet or some such. Eventually, music was put on a record controlling certain health problems. and made cheaper. Now, for a small fee, listeners can hear any The economics literature has plausibly argued that song they want, on numerous devices, wherever, whenever, “true” inflation, after the elimination of all major biases, has for a low fee. Thus, the cost of “music services” has fallen far been consistently and significantly lower than what’s been more than the product price indexes for record players, CD measured and reported. In 1996, the Boskin Commission players, or iPods is likely to suggest. concluded that the CPI has been systematically overestimated Robert Gordon famously dismisses the growth benefits by 1.1%. And a recent paper by U.S. officials from the Bureau of recent technology by arguing that information technol - of Labor Statistics and the Bureau of Economic Analysis ogy innovations pertain to a subset of consumer spending, argued that, in the prior 16 years, “the reduction in measured while the innovations in the first half of the 20th century real GDP growth from biases in both personal consumption were relevant for the lion’s share of purchases. We agree that expenditures and private fixed investment would be about “quality adjustment bias” was more widespread in the prewar -0.4 percent.” 9 period than today. Nevertheless, we think today’s bias is worse But where the BEA and BLS authors suggest that this than that which prevailed in the ’80s and ’90s because of the downward bias has not varied much since 2000, I find it evolving digital economy; however, much such advances pale unlikely that this bias has been constant over longer periods. in comparison to those that took place over a century ago. Robert Gordon, in his much-cited book called The Rise and Fall of American Growth,10 argued that the cluster of new Real GDP technologies that came to market in the early part of the 20th If we begin by assuming that nominal GDP is accurate, then century led to a very large upward inflation bias. Research on to the extent (PCE) inflation is biased up, real growth is automobiles, lighting, and other goods have provided compel - systematically understated. If the upward inflation bias has only gotten worse in recent years, then the slowdown in real 9 https://www.aeaweb.org/articles?id=10.1257/jep.31.2.187. 10 Robert Gordon (2016). The Rise and Fall of American Growth. Princeton, NJ: Princeton University Press. 11 http://www.nber.org/chapters/c6064.pdf. 12 Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 growth in recent years has been less pronounced than what has 2009, higher growth in hours worked as the labor market been reported, and seems to be almost universally accepted. recovered from the recession. Downward biases in real GDP However, the abrupt change in trend nominal growth imply downward biases in labor productivity. For given real in most developed countries around 2008 does not suggest GDP growth, high growth in hours depresses labor produc - a gradual change in measurement bias. Developed market tivity. On a short-term basis, productivity data can be hugely growth shifted down in both real and nominal terms. Tech volatile, with large jumps and drops. Productivity trends are optimists argue that the recent raft of under-measured, Inter - best assessed using long-term averages; but even then, the net-related innovations are sufficient to fill the gap in missing basic inflation biases we mentioned above affect the data. growth. But recent research has shown that technology-related The Cleveland Fed recently drew attention to a “systematic biases are probably insufficient to account for the reduction tendency to understate growth in real time, suggesting that in productivity trends after 2005. 12 In other words, the gap the average rate of the past six years will likely be revised up between post-2005 productivity and extrapolated pre-2005 in future.”13 trends is larger than what technology can account for across Theoretical explanations for weak recent labor productiv - the rich economies. ity include lower rates of business investment, which limits Nevertheless, firm belief in a downward shift in real what workers can produce; income distribution changes and growth involves both highly questionable assumptions and a social problems; and the slowdown in the growth rate of touching degree of trust in the statistical agencies. Growth has average human capital, measured as the growth of average slowed, but that’s “nominal growth.” Real growth has prob-a years of schooling (which has stopped trending up in the bly slowed, too, but almost certainly by less than the data U.S.). 14 suggest. Whereas conventional reported macro data indicate Whereas labor productivity measures control for the a decline of some 1.5% in real growth since 2007, the results growth of labor input, total factor productivity measures of the Boskin Commission suggest that real GDP growth control for both labor and capital. As we have moved down has been understated by at least 1% a year since the 1990s. our list of nine propositions, measurement uncertainty has New technologies might have worsened this bias. And more risen. The culmination of this movement toward increasingly conceptual critiques such as Nordhaus’s suggest that the “cost unreliable measures is reached with the case of total factor of living” has been falling significantly all along, especially productivity, which practical people find hard to take seriously when we consider the services provided by many new goods for a host of reasons. that have become available only in recent years. For starters, how do we measure the capital stock? Do we include intangible capital—things like code, patents, and Productivity unique business plans—and, if so, how do we measure it? And Macroeconomic productivity is usually measured in one of for ordinary physical capital, how do we measure the relative two ways: as real GDP or business sector output per hour contributions of residential structures (houses), public inf- ra worked (known as “labor productivity”) or residual real structure, non-residential structures (businesses of all types), GDP (“total factor productivity”) once the contributions equipment, and software? from changes in labor and capital inputs are removed. Such “Net” measures subtract depreciation of the capital stock. measures often bear little relation to what business leaders But depreciation measurement is as fraught with uncertainty think of as productivity, which tends to be some measure of as inflation measurement. The way things like natural disas - revenue per worker, the spread of specific technologies, or the ters are captured has changed sharply through the history of obsolescence rate of business models arising from technolo-gi GDP, a fact that surely matters during the current pandemic. cal change and, more generally, the entire process of creative e Th attempt to identify economic well-being with a national destruction. In fact, my own experience is that many CEOs income account concept should surely give more weight to net laugh openly at the idea that productivity has slowed down than to gross figures. But this is almost never done. in recent years. Still, the fact that total factor productivity measures But even so, the slowdown in the growth of real GDP per have fallen somewhat in the U.S. and Europe in the past ten hour worked (labor productivity) on both sides of the Atlantic reflects slower overall real GDP growth and, in the U.S. since 13 https://www.clevelandfed.org/~/media/content/newsroom%20and%20events/ publications/economic%20commentary/2016/ec%20201616/ec%20201616%20pdf. pdf?la=en. 12 http://faculty.chicagobooth.edu/chad.syverson/research/productivityslowdown. 14 Claudia Goldin and Lawrence Katz (2010), The Race between Education and pdf. Technology. Cambridge, MA: Harvard University Press. Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 13 years suggests that demographics and weak investment do To consider whether living standards have grown steadily, not fully explain all of the slowdown in growth. Economists we have to ask what living standards are. Consider this quota - as early as the 1960s began to observe a tendency of general tion from a recent history of the 19th century: economic growth to exceed what the increases in labor and In 1900 or 1914 most people around the world were engaged capital inputs seemed to imply. And this was considered an in agriculture. They worked on and with the soil. They mainly embarrassment to the profession: the mystery of growth could toiled in the open air, where they were dependent on the elements. be explained only as “a residual.” Nobel Laureate Robert That an ever-increasing share of all work came to be performed Solow is credited with naming this residual, which he called indoors was a great novelty of the nineteenth century. For someone “productivity.”15 newly arrived from the country, the first impression of a factory In conclusion, nominal GDP data have limited measur- e must have been of a workhouse.16 ment problems, inflation, real growth data have bigger problems, and productivity measures have larger problems This reminds us of the gap between GDP statistics and still. None of these data should be disregarded, but an living standards in a world in the process of being transformed. awareness of their relative precision is critically important, It helps to think about how people were living rather than especially given the kinds of business and policy decisions about what they were producing. Manufacturing employment that are influenced by the narratives constructed around ascended in the late 19th century, stabilized in the early 20th them, by the media and politicians as well as CEOs and century, and declined during the past 50 years. This hump- policymakers. like pattern is similar in other early industrializing economies. One final observation on measurement: our remarks have Surely the inexorable decline of farm jobs, the rise and fall focused on the United States. These issues are almost certainly of manufacturing employment, and the increase in services more challenging for many other countries. employment tell us more about how lives were being trans - formed than GDP data. The move from outdoor to indoor Misuse of the Data employment is described by economists as a “discontinu - Now let’s assume our official measures are perfect—completely ity” in a person’s living standards, and these discontinuities reliable quantifications, in a statistical sense, of what they happened frequently in the 19th century. Measured farm or purport to represent. The next question we have to ask is, do factory output per worker cannot capture the change when a such measures really provide useful indicators of the concepts worker in, say, Northern England, whose forefathers picked that tend to be associated with them? For example, does nomi - crops in a cold dark field, became permanently engaged in nal GDP really tell us much about productivity? And can we sheltered indoor employment. really justify the near universal tendency to use statistics such e Th cluster of technologies emerging in the 19th century as GDP, CPI/PCE inflation, or real GDP growth per hour transformed lives in other ways. e Th re were railroads, lights, to stand in for concepts such as living standards, the cost of telegraphs, and telephones. In the r fi st half of the 20th century, living, worker ec ffi iency, technology growth, the production there was widespread electricity, antibiotics, air condition - frontier, consumer surplus, and welfare? ing, and automobiles. When we consider the actual lifestyles, The answer to each of these questions I offer here is yes, health care, transportation, and entertainment that were but only with the sternest of caveats. available 100 or 150 years ago, the idea that productivity was According to Angus Maddison’s widely used estimates, the growing only modestly more than it grows now seems laug- h growth rate of U.S. real per capita GDP, when smoothed over able. And today’s pandemic-induced changes in Internet use ten-year periods, has ranged between zero and 3% per annum for both shopping and work may well represent a new discon - over the past two centuries. This performance of course far tinuity. We must respect the differences between concepts and exceeds that in all previous history—the r fi st modern econo- statistics by resisting the temptation to cona fl te productivity mies, the U.K. and U.S. did not achieve 1% growth until the and real GDP measures with likely unmeasurable notions of start of the 19th century. And if we translate these percent - consumer surplus, value added, utility, and progress. ages into how long it takes for real GDP to double, the range One can certainly argue that recent technologies have not becomes 25 years to inn fi ity. been as transformative for the scope of market transactions as the innovations of the late 19th century or the early 20th century. But artificial intelligence, robotics, genomics and a 15 http://www.nber.org/chapters/c8352.pdf. 16 Ibid. 14 Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 host of other developments are changing the environment for utilization rates. Big data are making it possible to improve r fi ms and consumers. measurements of all sorts of things. GDP’s conceptual problems will not go away because In What Data Should We Trust? of better data, better models, and better thinking, but a The attention and deference accorded today’s national accounts range of alternative measures may gradually replace official data are well in excess of what they are able to measure. Today’s GDP, inflation, and productivity measures when various productivity measures cannot truly capture the advance of an concepts are discussed. During the 2020 pandemic, data economy’s technology or its productive capacity. Real GDP on broad trends in individual’s temperatures, foot traffic cannot measure consumer surplus or utility. And inflation to various types of establishments, web traffic, road traffic, cannot measure the change of the cost of living over time, if and the exact timing of stimulus and other benet c fi hecks only because the meaning of “living” changes continuously. to individuals all proved to be essential to understan -d An idea sometimes used in philosophy and psychol - ing how economic activity was behaving in real time. The ogy, though not economics, is that of an “affordance.”17 In actual tracking of COVID transmission data, hospitaliz-a economics, an ao ff rdance might be thought of as something tion, mortality, and other correlates also proved useful in that enables a person the ability to accomplish some type of assessing economic prospects. In the second quarter of 2020, work, but that also ae ff cts how they live by, say, changing the an extreme drop in GDP occurred in many countries, and environment in which it occurs. The invention of fire was in the third quarter there was a rebound whose strength an affordance that enabled cooking, which in turn dramati- surprised many analysts. But infections rebounded from cally improved human life. In past millenia, humans, like a brief decline in the third quarter too, and expectations other primates, likely spent much of their time masticating for a strong GDP bounce hardly told the story of what was raw meat. After generations of cooking food, humans began happening. High frequency measures of foot traffic and to lose the ability to digest some things they used to eat. Fire other things suggested that growth was facing limits locally and cooking began as a luxury, but they created change, new where new outbreaks were underway. liabilities, a requirement to find fuel, a need to “afford the ao ff rdance.” The smartphone is a modern-day affordance. For many “ people, it has already become hard to imagine life without one. If the most prestigious data reinforce the wrong Expensive phones and data eat up household budgets, and message, then the data itself can be a cause of genuine low-income households struggle, work more, and incur debt to pay for their phone and data. Although this new affordance mischief, including misguided public policy as well as led to additional measured economic activity, comparing distorted private decision-making. “living standards” before and after smartphones (or fire) is not as straightforward as GDP data suggest. It’s hard enough to measure the relative happiness or well- being of one person compared to another. The task becomes e Th broad subjective impression of the economy’s success even less well-den fi ed once you acknowledge how dic ffi ult it according to households, voters, and policymakers can also is to measure the happiness or well-being of the same person be useful information. Even when wonderful technological over time. Does this mean the data are useless? Of course gains occur in a time of peace, consumer opportunities, and not. Nominal GDP is a useful starting point for measuring improving health care, if people believe the economy is not economic activity. making them better o,ff then their decisions are likely to be Some economists—and lots of non-economists—scoff at inu fl enced by that mindset. (And the real possibility that the macroeconomic indicators that markets and policymak - some improvements in income distribution have actually ers focus on now. What matters are things that are genuinely been taking place in recent years could easily get lost amid observable and relevant to the decisions made by households, the widespread messaging and general impression of “no firms, and policymakers: nominal income, the profit and wage progress.”) shares, the income distribution across households, and labor If the most prestigious data reinforce the wrong message, then the data itself can be a cause of genuine mischief, inclu - d ing misguided public policy as well as distorted private decision-making. Take, for instance, the claim that “real 17 The term is attributed to psychologist James J. Gibson. Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 15 median income has fallen for decades.” Would the median highs. To be sure, the gains have been especially conc- en household really prefer the opportunities available to them in trated in large technology companies, including some whose 1980 to what they can have now with the supposedly same business models relied heavily on intangible capital. But the real income? relationship between GDP and equity gains is loose in the Meanwhile, guided by central bank behavior, some inv- es best of times. And the wedge between the value creation tors take ideas like potential growth quite literally, forgetting being projected in today’s stock prices, and the stagnation the error bands that even policymakers warn people to place that appears to be evident in national accounts indicators, around their estimates of real GDP. Statistics and their histo -r provides yet another alert to the differences between concepts ical distributions can lead to hubristic forecasts derived from and statistics that this essay has covered. past data in a world still clearly in the midst of transforming. e Th re are other measures of the net worth of the private Janet Yellen knows the problems of inflation measurement, sector besides the equity market. If measured at market value, but 2% inflation has been a useful anchor for market expecta - these may reflect the capital gains that have accrued to the tions of monetary policy. In one sense, the cost of living might owners of real estate and business equity. The strong increase not really be rising by close to 2% year after year, but at least in private sector net worth in the U.S. in recent years is at odds this inflation measure is anchoring monetary policy, much as with slow growth in GDP, and is another hint that something the price of gold alone once did. is o i ff n its measurement. If we insist on measures of real data, one alternative We should think about the past, the present, and the to GDP is industrial production (IP), whose measurement future differently. We don’t know much about the future predates GDP. IP measures physical goods that can usually and never have. As for the present, we have big income and be stored and traded. These need financing, and IP data are wealth disparities, global imbalances, amazing new techn-ol highly correlated with financial market measures. Histori- ogies, slower-than-usual nominal growth, high corporate cal approaches to national income have sometimes focused prot fi s relative to nominal GDP and labor income, unusual on only physical or agricultural goods. This is wrong in monetary policy, changing demographics, debt/budget many ways, but short-term swings in the physical good s concerns, political economy tensions, and a broad sense of sector are nevertheless useful for investors who want to discomfort reported and no doubt experienced by many track the business cycle over short periods. For example, my households. Condensing all that into a single productivity Credit Suisse colleagues and I make extensive use of indu - s or GDP statistic does not begin to do justice to the complex trial production and related goods sector data when trying reality of the U.S. economy.   to understand short-term financial market and economic developments. Other alternatives include various attempts at “happ - i James Sweeney is Credit Suisse’s Chief Economist and Regional CIO ness indices,” ad hoc constructed “quality of life” indices for the Americas. He manages a 30-person economics department spread that impose a fixed definition of what a quality of life is, and worldwide. His written topics focus on the global and U.S. economies, direct measures of innovation. In many OECD economies, particularly on global industrial production and the workings of the global the growth of patents and the general spread of published financial system. His work appears regularly in major international finan - information have shown no signs of a slump. Intuitively, the cial media and James travels extensively to meet clients and policymakers. evolution of the Internet has allowed for a faster dissemination James is a member of Credit Suisse’s Zurich-based investment committee, of information, increased the possibilities for collaborative which determines investments for some of the firm’s more than $800bn research, and generally made formerly arcane knowledge in assets under management, the Americas Operating Committee of the widely available. New technologies, including innovations Global Markets Division, and the Americas Pension Investment Committee. in machine learning, robotics, and biology, are obvious drivers of profound innovation. The recent development of Previously, James held roles including Head of Fixed Income Research and the CRISPR gene editing technology has the potential to Chief Fixed Income Global Strategist. He joined Credit Suisse in 2000, profoundly shape our future, but its signic fi ance cannot (and having previously worked as a political speechwriter. He holds an MSc in will not) be captured with patent data or GDP. economics from the London School of Economics and a BS, magna cum Of course, one forward-looking measure of future income laude, from Florida State University. James is a member of the Council is the stock market, which did very well in the years between on Foreign Relations, the Economic Club of New York, and the National 2008 and the 2020 pandemic. And after plunging in March Committee on United States – China Relations. 2020, both the S&P 500 and Nasdaq have set new record 16 Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 ADVISORY BOARD EDITORIAL Yakov Amihud Carl Ferenbach Donald Lessard Clifford Smith, Jr. Editor-in-Chief New York University High Meadows Foundation Massachusetts Institute of University of Rochester Donald H. Chew, Jr. Technology Mary Barth Kenneth French Charles Smithson Associate Editor Stanford University Dartmouth College John McConnell Rutter Associates John L. McCormack Purdue University Amar Bhidé Martin Fridson Laura Starks Design and Production Robert Merton Tufts University Lehmann, Livian, Fridson University of Texas at Austin Mary McBride Massachusetts Institute of Advisors LLC Erik Stern Michael Bradley Assistant Editor Technology Stern Value Management Duke University Stuart L. Gillan Michael E. Chew Gregory V. Milano University of Georgia G. Bennett Stewart Fortuna Advisors LLC Richard Brealey Institutional Shareholder London Business School Richard Greco Stewart Myers Services Filangieri Capital Partners Michael Brennan Massachusetts Institute of René Stulz University of California, Trevor Harris Technology The Ohio State University Los Angeles Columbia University Robert Parrino Sheridan Titman Robert Bruner Glenn Hubbard University of Texas at Austin University of Texas at Austin University of Virginia Columbia University Richard Ruback Alex Triantis Charles Calomiris Michael Jensen Harvard Business School University of Maryland Columbia University Harvard University G. William Schwert Laura D’Andrea Tyson Christopher Culp Steven Kaplan University of Rochester University of California, Johns Hopkins Institute for University of Chicago Alan Shapiro Berkeley Applied Economics University of Southern David Larcker Ross Watts California Howard Davies Stanford University Massachusetts Institute Institut d’Études Politiques Martin Leibowitz Betty Simkins of Technology de Paris Morgan Stanley Oklahoma State University Jerold Zimmerman Robert Eccles University of Rochester Harvard Business School Journal of Applied Corporate Finance (ISSN 1078-1196 [print], ISSN 1745-6622 Statement on Research4Life [online]) is published quarterly per year by Wiley Subscription Services, Inc., a Wiley Wiley is a founding member of the UN-backed HINARI, AGORA, and OARE initiatives. Company, 111 River St., Hoboken, NJ 07030-5774 USA. They are now collectively known as Research4Life, making online scientific content available free or at nominal cost to researchers in developing countries. Please visit Postmaster: Send all address changes to JOURNAL OF APPLIED CORPORATE FI- Wiley’s Content Access – Corporate Citizenship site: http://www.wiley.com/WileyCDA/ NANCE, John Wiley & Sons Inc., c/o The Sheridan Press, PO Box 465, Hanover, PA Section/id-390082.html 17331 USA. Journal of Applied Corporate Finance accepts articles for Open Access publication. Information for Subscribers Please visit https://authorservices.wiley.com/author-resources/Journal-Authors/open- Journal of Applied Corporate Finance is published quarterly per year. Institutional sub- access/onlineopen.html for further information about OnlineOpen. scription prices for 2020 are: Print & Online: US$844 (US), US$1007 (Rest of World), €656, (Europe), £516 Wiley’s Corporate Citizenship initiative seeks to address the environmental, social, (UK). Commercial subscription prices for 2020 are: Print & Online: US$1123 (US), economic, and ethical challenges faced in our business and which are important to US$1339 (Rest of World), €872 (Europe), £686 (UK). Individual subscription prices our diverse stakeholder groups. Since launching the initiative, we have focused on for 2020 are: Print & Online: US$137 (US), $137 (Rest of World), €115 (Europe), sharing our content with those in need, enhancing community philanthropy, reducing £79 (UK). Student subscription prices for 2020 are: Print & Online: US$49 (US), our carbon impact, creating global guidelines and best practices for paper use, estab- $49 (Rest of World), €41 (Europe), £28 (UK). Prices are exclusive of tax. Asia-Pacific lishing a vendor code of ethics, and engaging our colleagues and other stakeholders GST, Canadian GST/HST and European VAT will be applied at the appropriate rates. in our efforts. For more information on current tax rates, please go to https://onlinelibrary.wiley.com/ library-info/products/price-lists/payment. The price includes online access to the cur- Follow our progress at www.wiley.com/go/citizenship. rent and all online back files to January 1, 2016, where available. For other pricing options, including access information and terms and conditions, please visit https://on- Abstracting and Indexing Services linelibrary.wiley.com/library-info/products/price-lists. Terms of use can be found here: The Journal is indexed by Accounting and Tax Index, Emerald Management https://onlinelibrary.wiley.com/terms-and-conditions. Reviews (Online Edition), Environmental Science and Pollution Management, Risk Abstracts (Online Edition), and Banking Information Index. Delivery Terms and Legal Title Where the subscription price includes print issues and delivery is to the recipient’s Disclaimer address, delivery terms are Delivered at Place (DAP); the recipient is responsible for The Publisher, Cantillon and Mann, its affiliates, and Editors cannot be held respon - paying any import duty or taxes. Title to all issues transfers FOB our shipping point, sible for errors or any consequences arising from the use of information contained in freight prepaid. We will endeavour to fulfil claims for missing or damaged copies within this journal; the views and opinions expressed do not necessarily reflect those of the six months of publication, within our reasonable discretion and subject to availability. Publisher, Cantillon and Mann, its affiliates, and Editors, neither does the publication of advertisements constitute any endorsement by the Publisher, Cantillon and Mann, its Journal Customer Services: For ordering information, claims and any inquiry concern- affiliates, and Editors of the products advertised. ing your journal subscription please go to https://hub.wiley.com/community/support/ onlinelibrary or contact your nearest office. Copyright and Copying Americas: Email: cs-journals@wiley.com; Tel: +1 781 388 8598 or Copyright © 2020 Cantillon and Mann. All rights reserved. No part of this publication +1 800 835 6770 (toll free in the USA and Canada). may be reproduced, stored or transmitted in any form or by any means without the Europe, Middle East and Africa: Email: cs-journals@wiley.com; prior permission in writing from the copyright holder. Authorization to photocopy items Tel: +44 (0) 1865 778315. for internal and personal use is granted by the copyright holder for libraries and other Asia Pacific : Email: cs-journals@wiley.com; Tel: +65 6511 8000. users registered with their local Reproduction Rights Organization (RRO), e.g., Copy- Japan: For Japanese speaking support, Email: cs-japan@wiley.com right Clearance Center (CCC), 222 Rosewood Drive, Danvers, MA 01923, USA (www. Visit our Online Customer Help at https://hub.wiley.com/community/support/onlineli- copyright.com), provided the appropriate fee is paid directly to the RRO. This consent brary does not extend to other kinds of copying such as copying for general distribution, for advertising or promotional purposes, for republication, for creating new collective Production Editor: Shalini Chawla (email: jacf@wiley.com). works or for resale. Permissions for such reuse can be obtained using the RightsLink “Request Permissions” link on Wiley Online Library. Special requests should be ad- Back Issues: Single issues from current and recent volumes are available at the dressed to: permissions@wiley.com. current single issue price from cs-journals@wiley.com. Earlier issues may be obtained from Periodicals Service Company, 351 Fairview Avenue – Ste 300, Hudson, NY 12534, USA. Tel: +1 518 537 4700, Fax: +1 518 537 5899, Email: psc@periodicals.com View this journal online at wileyonlinelibrary.com/journal/jacf. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Applied Corporate Finance Wiley

Rethinking Macro Measurement

Journal of Applied Corporate Finance , Volume 32 (4) – Dec 1, 2020

Loading next page...
 
/lp/wiley/rethinking-macro-measurement-7hhRFyJ0Tc

References (0)

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Wiley
Copyright
Copyright © 2020 Cantillon and Mann
ISSN
1078-1196
eISSN
1745-6622
DOI
10.1111/jacf.12428
Publisher site
See Article on Publisher Site

Abstract

VOLUME 32 NUMBER 4 FALL 2020 Journal of AP PLIED COR PORATE FINANC E IN THIS ISSUE: Public James Sweeney, Chief Economist, Credit Suisse 17 The First Modern Financial Crises: The South Sea and Mississippi Finance and Bubbles in Historical Perspective Robert F. Bruner, University of Virginia; and Scott C. Miller, Yale University Central Banks Convergence and Reversion: China’s Banking System at 70 Carl Walter Apocalypse Averted: The COVID-Caused Liquidity Trap, Dodd-Frank, and the Fed Craig Pirrong, University of Houston The Poverty of Monetarism Patrick Bolton, Columbia University Columbia Business School Roundtable on Broken Models of Public Finance Panelists: Jared Bernstein, Center for Budget and Policy Priorities; and Paul Kazarian, Japonica Partners. Moderated by Shiva Rajgopal, Columbia University 82 Columbia Business School Roundtable on The Fed’s Response to the Global Financial Crisis—and Now the Pandemic Panelists: Frederic Mishkin and Patricia Mosser, Columbia University. Moderated by Kate Davidson, The Wall Street Journal The Euro @ 20: How Economic and Financial “Asymmetries” Marred the Promise of the Single Currency George Alogoskoufis, Athens University of Economics and Business; and Laurent Jacque, Tufts University The Benefits of Buying Distressed Assets Jean-Marie Meier, University of Texas at Dallas; and Henri Servaes, London Business School, CEPR, and ECGI 117 Using ESG to Enhance Fixed-Income Returns: The Case of Inherent Group Nikhil Mirchandani and Chelsea Rossetti, Inherent Group The Economic (Not Literary) Offenses of Michael Lewis: The Case of The Big Short and the Global Financial Crisis Don Chew, Journal of Applied Corporate Finance Rethinking Macro Measurement by James Sweeney, Chief Economist, Credit Suisse new empiricism is sweeping through macroeconomics. It is based on an old under- A standing—that some widely followed and near universally accepted economic indicators do not do a good job of capturing the concepts with which most of the general public associates them. It is also based on new methods of gathering data. The COVID-19 pandemic has contributed to increased aware - I will show that in many ways the national accounts ness of these developments. Some macroeconomic indicators data are valuable, despite their limitations. Still, there is have proved to be too slow or imprecise to explain what was much room for new measurement approaches that can be happening in real time. Meanwhile, a range of new indica - employed alongside while even making use of the existing tors, often higher frequency and based on “bigger” data, have numbers. With the new possibilities created by big data provided genuinely useful guidance. and rising computational power, the social importance of National accounts data that have long been widely used GDP and related measures has probably already peaked, and to represent concepts such as real income, inflation rates, new measures are reshaping economic thinking. A greater and productivity are now being questioned more than ever emphasis on “administrative data” and the use of “quasi- before. The appropriate scope, inherent biases, and potential experimental evidence” is now clear in the business world misuse of the national accounts have long been controv-er and in the empirical academic literature. sial.1 In recent years, these debates have gone mainstream, Although we cannot precisely predict how a macro revolu - as widely perceived changes in the rates of technological, tion will play out, and the extent to which it might lead to economic, and social change have not always matched changes to orthodox theory, understanding the problems with what’s in the numbers. traditional data helps us stay open-minded about whether One response has been a proliferation of altern - a popular views about recent economic performance might in tive measures. These include new indicators of everyday the future appear to have been misguided. phenomena, including foot traffic and web traffic data, tax return data, app data, and a wide range of company- Mismeasurement specific data. There is work on disease-based price indexes, Let’s start with a simple question. Could widely accepted facts happiness measures, “billion price” aggregates, and activity on recent economic performance be wrong on account of bad trackers sourced from satellites. There are also attempts to data? If we compare U.S. performance in the period from “correct” the traditional measures with ad hoc adjustments 1995 to 2007 with the period from 2008-2019, we arrive at or by switching emphasis toward less well-known variants the following propositions: of the main measures. 1. N ominal GDP growth has slowed (from 5.4% p.a. to I will focus here on mismeasurement and misuse of the 3.4%) national accounts and suggest some ways to make better 2. Working age population growth has slowed (from decisions once these are taken into consideration. Mismea - 1.3% to 0.5%) surement refers to whether a statistic is an accurate measure 3. Nominal GDP per capita and per worker have slowed of what it purports to measure; misuse refers to whether a (from 4.4% to 2.7%) concept associated with a statistic is appropriately invoked. 4. e Th share of income going to labor has continued a Some concepts are just not easily quantie fi d. secular decline (from 55.1% to 53.1%) 5. Core ina fl tion has slowed slightly (from 1.8% to 1.6%) 1 Diane Coyle (2014), GDP: A Brief But Affectionate History. Princeton, NJ: 6. Real GDP growth has slowed (from 3.2% to 1.7%) Princeton University Press. 8 Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 7. Labor productivity (real GDP per hour) has slowed themselves for living in their own homes. Estimation relies (from 2.1% to 1.1%) on many assumptions, such as whether local rents are a good 8. Business investment has fallen (from 6.3% to 3%) proxy for what owner-occupied housing would rent for if 9. Total factor productivity has slowed (from 1.4% to put on the market. 0.5%) Shortcomings aside, nominal GDP data serve many In short, the U.S. economy appears to be sue ff ring from purposes well because their source components are usually both a demographic and a productivity slowdown. Ina fl tion collected as nominal figures before being adjusted and aggre - has fallen and investment has been lackluster. However, not all gated. Data are collected in three categories: expenditure, of these claims are equally supported. This is not just because income, and value added (profit plus labor cost), with expen- of differences in the measured magnitudes, but also in the diture figures usually arriving first. These three perspectives reliability of the measures themselves. should sum to roughly the same total, so national accoun - The first three propositions are based on nominal and tants can cross-check their estimates, even if they never have demographic data, which most of my macro colleagues and I complete data from any of the perspectives. Large revisions consider to be highly accurate. Controversies about nominal to GDP often follow when tax data become available. The GDP measurement are mostly about scope: what should be usual stability of the ratio of collected taxes to nominal GDP counted? This has been debated for most of the past century, provides some evidence that nominal GDP measures are but the main conclusion is settled: GDP measures market reasonably accurate. transactions without making (or purporting to make) any attempt to capture all valuable activity. GDP thus ignores home production and free goods. And for this reason, swings “ in female and child participation in the market economy have Problems arise when we use market GDP data to significantly impacted GDP over time, as activities shifted stand in for broad social concepts like productivity between paid and unpaid work. For example, as one economic historian has pointed out, and value added. “Cooking must have been the most widespread, and generally time-consuming, expenditure of labor throughout history.” 2 But cooking, retrieving water from a well, and washing clothes outside have never been counted in GDP—unless they were part of a market transaction; that is, outside help was hired, Our fourth proposition is about the labor income share. It or the eating was done in restaurants. comes from the income accounts and can be supported with In sum, many valuable but “free” goods and services are surveys and tax return data. After massive recent public atten - ignored because they do not involve market transactions. And tion to wealth and income inequality, it has become widely in the modern economy, the importance and value of free accepted that, during the past three decades, the pretax income search engines and social media are self-evident, as they are distribution has moved sharply in favor of very high earners for many other freely provided goods and services, such as in the United States and several other countries. And as crowd-sourced trac ffi apps, which create value by getting us inequality within labor income has worsened, labor’s overall to home and work faster. share has fallen—developments that have long been ree fl cted You can’t fault an accounting system for failing to measure in media and political debates. what it does not claim to measure. Nevertheless, problems Propositions 1-4 are well enough established to be called arise when we try to use market GDP data to stand in for “facts,” and though measurement issues exist, they do not broad social concepts like productivity and value added, a undermine the use of these data for many purposes. What subject we turn to later. distinguishes them from propositions 5-9 is that they are all There are a few specific measurement difficulties for nominal concepts. For calculations that rely on inflation adjust - subcomponents of nominal GDP that are worth noting ments, our confidence in their accuracy should be reduced to here. Take “imputed rents,” which make up about 8% of reflect both measurement and conceptual challenges. Real GDP U.S. GDP. This is what homeowners are assumed to pay growth equals nominal growth minus inflation. To the extent inflation measures are unreliable, real growth measures are too. And since ina fl tion measurement is the heart of our topic, we 2 Jurgen Osterhammel (2014), The Transformation of the World: A Global History turn to it first, and give it extra attention. of the Nineteenth Century. Princeton, NJ: Princeton University Press. Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 9 Inflation Bias weights, price time series for each (often changing) good, Many financial market participants talk as if inflation data and sometimes quality adjustment measures. Each of these are precise. But consider this excerpt from the transcript of involves considerable dic ffi ulty and numerous assumptions. an FOMC meeting on July 2-3, 1996: Economics students learn about two types of inflation indexes, which use different weighting formulas: Laspeyres CHAIRMAN GREENSPAN: Is long-term price stability an and Paasche Indexes. The U.S. deflator for personal consump - appropriate goal of the Federal Reserve System? tion expenditures, known as the “PCE inflation index,” uses the Fisher Ideal method, which is a geometric average of the MS. YELLEN: Mr. Chairman, will you define “price stabil - two. The U.S. consumer price index (CPI) uses a version ity” for me? of the Laspeyres method with fixed weights periodically adjusted after spending surveys. e Th CPI is never signic fi antly CHAIRMAN GREENSPAN: Price stability is that state in which revised, but the PCE inflation index is revised often, and expected changes in the general price level do not ee ff ctively often substantially. alter business or household decisions. Different vintages of the core PCE deflator show that revisions can push measured inflation as much as 0.5% away MS. YELLEN: Could you please put a number on that? from the original estimate.4 These revisions do not affect the CPI, even though that is the source of much of the PCE CHAIRMAN GREENSPAN: I would say the number is zero, if deflator’s data. Not revising the CPI prevents some annoy - ina fl tion is properly measured. ances because it is used to index government benet fi s and inflation-linked bonds. The bond market reacts strongly to MS. YELLEN: Improperly measured, I believe that heading both indexes, and Fed officials have said that they focus on toward 2 percent inflation would be a good idea, and that the PCE deflator, partly because of its flexible weights and we should do so in a slow fashion, looking at what happens revisions. along the way. My presumption based on the literature is, Measurement uncertainty should cause market partici - as Bob Parry summarized it, that given current inaccurate pants and policymakers to reduce their confidence in measurements, heading toward 2 percent is most likely to be short-term inflation observations. Several years ago, when benec fi ial. core U.S. PCE inflation was less than the market expected, the Dallas Fed argued that “it’s likely that these figures will be CHAIRMAN GREENSPAN: Could we leave it at that? Let’s now revised higher… at the next annual revision in summer 2018.” move on.3 And, indeed, signic fi ant upward revisions did later occur. During the 2020 COVID-19 pandemic, the measured This exchange between sitting and future Fed chairs CPI and PCE ina fl tion both fell sharply, but large changes in occurred during a discussion on whether to use monetary the actual composition of underlying spending raised unc- er policy to drive inflation permanently to 0% following years tainty about the reliability of these real-time measures and the of declining measured ina fl tion. Before the 1970s, peacetime possibility of future volatility and revisions. In my view, it is inflation had probably been near zero for most of U.S. history. clear that the pandemic has resulted in short-term downward One reason the Fed, along with a number of other central pressure on ina fl tion, especially driven by falling commodity banks, chose 2% is that it would allow more room to cut prices. But the imprecision of the inflation data is even greater interest rates when needed. Another was that research had than usual, and the many product-specific supply and demand suggested that inflation data were upwardly biased. “One shocks that have occurred have likely led to much “noisier” can argue that we have roughly a 1 percent bias in the CPI,” component level data. Yellen said in the same meeting. This exchange, and the studies Obtaining time series of price data for specific goods as that emerged from the 1990s Boskin Commission, show that inputs to the inflation index sounds like an easy task, especially Greenspan, Yellen, and others were well aware of the impr- eci for something like a pound of bananas. But exactly when sion of ina fl tion estimates. should a new good be included? VCRs were first sold for more Ina fl tion indexes are much harder to estimate than price than $1,000 in the late 1970s, but they did not appear in the changes through time of a particular good. An index requires 4 Incidentally, recent large revisions to PCE have tended to come from financial 3 https://www.federalreserve.gov/fomc/minutes/19960702.htm. Pages 50-51. services prices. 10 Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 CPI until 1987, when prices were down to a few hundred quality adjustments—have created biases and uncertainties dollars and most households already owned them. 5 The prior in inflation measures. We have focused so far on consumption price declines never affected measured inflation because VCRs goods, but GDP inflation includes the prices of investment were not in the index when their greatest dea fl tion occurred. and traded goods, too, leading to more room for error7 . Goods evolve, sometimes into performing entirely Not all biases are in the direction of underestimating different functions. Today’s smartphones are yesterday’s output and overestimating inflation. For example, when cameras, fax machines, televisions, tape recorders, music companies substitute cheaper foreign-sourced inputs for players, handheld video game players, pocket organizers, domestic ones, output is actually overstated. And many alarm clocks, and many more things—some of which are professionals in the financial services industry profess to new. If smartphone prices were compared to the total prices believe that inflation is understated. But this view can, and of all of those things in the past, then phones would have should, be dismissed as the working of local or “availability” be shown to have deflated massively. bias: individuals living in areas with particularly strong income To be sure, phone prices have fallen a lot, and the cost of growth witness many firsthand examples of conspicuous price transmitting telecommunications data—a service, distinct increases—think restaurants, rents, and tuition. But econo - from the hardware—has dropped too, but there are other mists would identify these as changes in “relative prices,” not forms of “deflation” in a smartphone that have not been signs of national ina fl tion. captured. Statistical agencies simply cannot capture it all, but Physical goods constitute about one-third of consumption they are always working on ways of improving their measures. in the developed markets, and the rest are services. Short- But as a result of this mismeasurement, phone “inflation” is term inflation expectation surveys are strongly correlated with almost certainly measured and reported to be higher than it energy prices, the most visible price changes in the economy. could be, and “real expenditure” on phones is understated. In general, when people think about inflation, they think So-called “hedonic” adjustments—those which control of goods prices, even though housing, health care, financial for the change in certain attributes of a good—are one way services, and other services involve much more spending, with of dealing with this issue, but such adjustments are used prices that are often opaque. in only a subset of the goods and services whose quality Health care prices are particularly hard to collect. Health changes. Automobiles are a well-known case where hedonic care spending is 17% of U.S. GDP. This figure comes from adjustments cause the national accounts prices to differ from adding the industry’s reported profits to its workers’ wages. retail prices. Carving nominal health care spending into ina fl tion and real Technology prices can involve hedonic adjustments for growth is a formidable challenge. For Americans, discovering attributes such as memory and processing speed. During the the price of specific health care services is often equally formi - tech boom of the late ’90s, jumps in processing speed amid dable! But the government uses surveys and special accounts high chip and hardware production in the United States to index prices to arrive at the sector’s ina fl tion. prompted hedonic adjustments that, by pushing down the Health care prices have risen faster than overall inflation in prices of such goods, led to significant increases in estimates most years. At first, one might view this as confirming evidence of real output, and hence a reported surge in productivity. of the notorious U.S. inefficiency in this sector—until one As much of the hardware industry left the United States recognizes that health care relative prices tend to rise in other after 2000, U.S. data lost the benefit of this fast-growing, economies too, and in general, the richer the economy, the fast-deflating production, one significant factor in the lower higher the health care spending. This phenomenon is in part a rates of real growth reported in the past decade. This failure reflection of “Baumol’s cost disease.”8 Factory-produced goods to take consistent account of such changes has profoundly get cheaper with technology; but because the principal actors impacted—and I would argue distorted—industrial produ-c in this case, the doctors, are in fixed supply, relatively speaking, tion data by emphasizing some productivity gains while the price of doctors’ services tends to rise relative to those of completely missing others.6 manufactured goods. And the higher the relative productiv - A whole series of statistical challenges, then—those ity in manufacturing, the stronger this effect is likely to be. arising from weighting issues, price series continuity, and 7 There are separate literatures on biases in import and export prices, and invest- ment goods’ prices. 5 http://www.nber.org/chapters/c6072.pdf. Page 388. 8 William Baumol and William Bowen (1966), Performing Arts, The Economic 6 https://www.brookings.edu/wp-content/uploads/2016/06/us-manufacturing-past- Dilemma: A Study of Problems Common to Theater, Opera, Music and Dance (New and-potential-future-baily-bosworth.pdf. York: Twentieth Century Fund). Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 11 The sharp increase in the use of online doctor visits during ling arguments that the statistics for that era were incapable of the COVID-19 pandemic has the potential to bring about accounting for the true deflation once quality adjustment is an interesting short-term reversal in this situation, allowing considered. To see the problem, it helps to consider William some costs to be squeezed out of health care. Nordhaus’s famous analysis of the historical cost of a lumen But the big story in health care ina fl tion has to begin by of light, in which he observed that recognizing the quality improvements that have occurred, and the inability of the data to capture those innovations, that have a comparison of the pure price of light with a traditional taken place over time. In 1900, world average life expectancy light price indicates that traditional price indexes overstate price at birth was near 30; today it is over 70. Infant mortality has growth, and therefore understate output growth, by a factor declined by more than 90%. There has been great progress between 900 and 1600 since the beginning of the nineteenth recently against chronic diseases, and in the past century, most century. This finding suggests that the “true” growth of real wages infectious diseases have been conquered—notwithstanding and real output may have been signic fi antly understated during the recent appearance of one particularly infectious one. the period since the Industrial Revolution.11 Cold numbers on life expectancy and infant mortality are meant to stand in for millions of smiles and memories that In Nordhaus’s view, the keepers of the national income could not have occurred before the 20th century. The topic is accounts (and macroeconomists) must find a way to measure too overwhelming to summarize with simple stories, but to the price of light, not of light bulbs. In other words, to grasp the potential impact of proper “quality adjustment” for understand the value of the goods and services created by an health care data, try imagining yourself as the patient awaiting economy, we would ideally like to know the demand curve a 19th-century surgery, pre-anesthesia, pre-germ theory, with for specific quantities of what consumers are seeking, such a surgeon standing over you with saw and scissors. as lumens of light. This would make possible true quality- Of course, quality adjustments are not much applied to adjusted measurement. health care prices, although the U.S. Bureau of Economic Another example: two centuries ago, only the rich or Analysis is experimenting with disease-based cost indexes that powerful could demand music. They had to pay a string try to measure prices in terms of the cost of conquering or quartet or some such. Eventually, music was put on a record controlling certain health problems. and made cheaper. Now, for a small fee, listeners can hear any The economics literature has plausibly argued that song they want, on numerous devices, wherever, whenever, “true” inflation, after the elimination of all major biases, has for a low fee. Thus, the cost of “music services” has fallen far been consistently and significantly lower than what’s been more than the product price indexes for record players, CD measured and reported. In 1996, the Boskin Commission players, or iPods is likely to suggest. concluded that the CPI has been systematically overestimated Robert Gordon famously dismisses the growth benefits by 1.1%. And a recent paper by U.S. officials from the Bureau of recent technology by arguing that information technol - of Labor Statistics and the Bureau of Economic Analysis ogy innovations pertain to a subset of consumer spending, argued that, in the prior 16 years, “the reduction in measured while the innovations in the first half of the 20th century real GDP growth from biases in both personal consumption were relevant for the lion’s share of purchases. We agree that expenditures and private fixed investment would be about “quality adjustment bias” was more widespread in the prewar -0.4 percent.” 9 period than today. Nevertheless, we think today’s bias is worse But where the BEA and BLS authors suggest that this than that which prevailed in the ’80s and ’90s because of the downward bias has not varied much since 2000, I find it evolving digital economy; however, much such advances pale unlikely that this bias has been constant over longer periods. in comparison to those that took place over a century ago. Robert Gordon, in his much-cited book called The Rise and Fall of American Growth,10 argued that the cluster of new Real GDP technologies that came to market in the early part of the 20th If we begin by assuming that nominal GDP is accurate, then century led to a very large upward inflation bias. Research on to the extent (PCE) inflation is biased up, real growth is automobiles, lighting, and other goods have provided compel - systematically understated. If the upward inflation bias has only gotten worse in recent years, then the slowdown in real 9 https://www.aeaweb.org/articles?id=10.1257/jep.31.2.187. 10 Robert Gordon (2016). The Rise and Fall of American Growth. Princeton, NJ: Princeton University Press. 11 http://www.nber.org/chapters/c6064.pdf. 12 Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 growth in recent years has been less pronounced than what has 2009, higher growth in hours worked as the labor market been reported, and seems to be almost universally accepted. recovered from the recession. Downward biases in real GDP However, the abrupt change in trend nominal growth imply downward biases in labor productivity. For given real in most developed countries around 2008 does not suggest GDP growth, high growth in hours depresses labor produc - a gradual change in measurement bias. Developed market tivity. On a short-term basis, productivity data can be hugely growth shifted down in both real and nominal terms. Tech volatile, with large jumps and drops. Productivity trends are optimists argue that the recent raft of under-measured, Inter - best assessed using long-term averages; but even then, the net-related innovations are sufficient to fill the gap in missing basic inflation biases we mentioned above affect the data. growth. But recent research has shown that technology-related The Cleveland Fed recently drew attention to a “systematic biases are probably insufficient to account for the reduction tendency to understate growth in real time, suggesting that in productivity trends after 2005. 12 In other words, the gap the average rate of the past six years will likely be revised up between post-2005 productivity and extrapolated pre-2005 in future.”13 trends is larger than what technology can account for across Theoretical explanations for weak recent labor productiv - the rich economies. ity include lower rates of business investment, which limits Nevertheless, firm belief in a downward shift in real what workers can produce; income distribution changes and growth involves both highly questionable assumptions and a social problems; and the slowdown in the growth rate of touching degree of trust in the statistical agencies. Growth has average human capital, measured as the growth of average slowed, but that’s “nominal growth.” Real growth has prob-a years of schooling (which has stopped trending up in the bly slowed, too, but almost certainly by less than the data U.S.). 14 suggest. Whereas conventional reported macro data indicate Whereas labor productivity measures control for the a decline of some 1.5% in real growth since 2007, the results growth of labor input, total factor productivity measures of the Boskin Commission suggest that real GDP growth control for both labor and capital. As we have moved down has been understated by at least 1% a year since the 1990s. our list of nine propositions, measurement uncertainty has New technologies might have worsened this bias. And more risen. The culmination of this movement toward increasingly conceptual critiques such as Nordhaus’s suggest that the “cost unreliable measures is reached with the case of total factor of living” has been falling significantly all along, especially productivity, which practical people find hard to take seriously when we consider the services provided by many new goods for a host of reasons. that have become available only in recent years. For starters, how do we measure the capital stock? Do we include intangible capital—things like code, patents, and Productivity unique business plans—and, if so, how do we measure it? And Macroeconomic productivity is usually measured in one of for ordinary physical capital, how do we measure the relative two ways: as real GDP or business sector output per hour contributions of residential structures (houses), public inf- ra worked (known as “labor productivity”) or residual real structure, non-residential structures (businesses of all types), GDP (“total factor productivity”) once the contributions equipment, and software? from changes in labor and capital inputs are removed. Such “Net” measures subtract depreciation of the capital stock. measures often bear little relation to what business leaders But depreciation measurement is as fraught with uncertainty think of as productivity, which tends to be some measure of as inflation measurement. The way things like natural disas - revenue per worker, the spread of specific technologies, or the ters are captured has changed sharply through the history of obsolescence rate of business models arising from technolo-gi GDP, a fact that surely matters during the current pandemic. cal change and, more generally, the entire process of creative e Th attempt to identify economic well-being with a national destruction. In fact, my own experience is that many CEOs income account concept should surely give more weight to net laugh openly at the idea that productivity has slowed down than to gross figures. But this is almost never done. in recent years. Still, the fact that total factor productivity measures But even so, the slowdown in the growth of real GDP per have fallen somewhat in the U.S. and Europe in the past ten hour worked (labor productivity) on both sides of the Atlantic reflects slower overall real GDP growth and, in the U.S. since 13 https://www.clevelandfed.org/~/media/content/newsroom%20and%20events/ publications/economic%20commentary/2016/ec%20201616/ec%20201616%20pdf. pdf?la=en. 12 http://faculty.chicagobooth.edu/chad.syverson/research/productivityslowdown. 14 Claudia Goldin and Lawrence Katz (2010), The Race between Education and pdf. Technology. Cambridge, MA: Harvard University Press. Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 13 years suggests that demographics and weak investment do To consider whether living standards have grown steadily, not fully explain all of the slowdown in growth. Economists we have to ask what living standards are. Consider this quota - as early as the 1960s began to observe a tendency of general tion from a recent history of the 19th century: economic growth to exceed what the increases in labor and In 1900 or 1914 most people around the world were engaged capital inputs seemed to imply. And this was considered an in agriculture. They worked on and with the soil. They mainly embarrassment to the profession: the mystery of growth could toiled in the open air, where they were dependent on the elements. be explained only as “a residual.” Nobel Laureate Robert That an ever-increasing share of all work came to be performed Solow is credited with naming this residual, which he called indoors was a great novelty of the nineteenth century. For someone “productivity.”15 newly arrived from the country, the first impression of a factory In conclusion, nominal GDP data have limited measur- e must have been of a workhouse.16 ment problems, inflation, real growth data have bigger problems, and productivity measures have larger problems This reminds us of the gap between GDP statistics and still. None of these data should be disregarded, but an living standards in a world in the process of being transformed. awareness of their relative precision is critically important, It helps to think about how people were living rather than especially given the kinds of business and policy decisions about what they were producing. Manufacturing employment that are influenced by the narratives constructed around ascended in the late 19th century, stabilized in the early 20th them, by the media and politicians as well as CEOs and century, and declined during the past 50 years. This hump- policymakers. like pattern is similar in other early industrializing economies. One final observation on measurement: our remarks have Surely the inexorable decline of farm jobs, the rise and fall focused on the United States. These issues are almost certainly of manufacturing employment, and the increase in services more challenging for many other countries. employment tell us more about how lives were being trans - formed than GDP data. The move from outdoor to indoor Misuse of the Data employment is described by economists as a “discontinu - Now let’s assume our official measures are perfect—completely ity” in a person’s living standards, and these discontinuities reliable quantifications, in a statistical sense, of what they happened frequently in the 19th century. Measured farm or purport to represent. The next question we have to ask is, do factory output per worker cannot capture the change when a such measures really provide useful indicators of the concepts worker in, say, Northern England, whose forefathers picked that tend to be associated with them? For example, does nomi - crops in a cold dark field, became permanently engaged in nal GDP really tell us much about productivity? And can we sheltered indoor employment. really justify the near universal tendency to use statistics such e Th cluster of technologies emerging in the 19th century as GDP, CPI/PCE inflation, or real GDP growth per hour transformed lives in other ways. e Th re were railroads, lights, to stand in for concepts such as living standards, the cost of telegraphs, and telephones. In the r fi st half of the 20th century, living, worker ec ffi iency, technology growth, the production there was widespread electricity, antibiotics, air condition - frontier, consumer surplus, and welfare? ing, and automobiles. When we consider the actual lifestyles, The answer to each of these questions I offer here is yes, health care, transportation, and entertainment that were but only with the sternest of caveats. available 100 or 150 years ago, the idea that productivity was According to Angus Maddison’s widely used estimates, the growing only modestly more than it grows now seems laug- h growth rate of U.S. real per capita GDP, when smoothed over able. And today’s pandemic-induced changes in Internet use ten-year periods, has ranged between zero and 3% per annum for both shopping and work may well represent a new discon - over the past two centuries. This performance of course far tinuity. We must respect the differences between concepts and exceeds that in all previous history—the r fi st modern econo- statistics by resisting the temptation to cona fl te productivity mies, the U.K. and U.S. did not achieve 1% growth until the and real GDP measures with likely unmeasurable notions of start of the 19th century. And if we translate these percent - consumer surplus, value added, utility, and progress. ages into how long it takes for real GDP to double, the range One can certainly argue that recent technologies have not becomes 25 years to inn fi ity. been as transformative for the scope of market transactions as the innovations of the late 19th century or the early 20th century. But artificial intelligence, robotics, genomics and a 15 http://www.nber.org/chapters/c8352.pdf. 16 Ibid. 14 Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 host of other developments are changing the environment for utilization rates. Big data are making it possible to improve r fi ms and consumers. measurements of all sorts of things. GDP’s conceptual problems will not go away because In What Data Should We Trust? of better data, better models, and better thinking, but a The attention and deference accorded today’s national accounts range of alternative measures may gradually replace official data are well in excess of what they are able to measure. Today’s GDP, inflation, and productivity measures when various productivity measures cannot truly capture the advance of an concepts are discussed. During the 2020 pandemic, data economy’s technology or its productive capacity. Real GDP on broad trends in individual’s temperatures, foot traffic cannot measure consumer surplus or utility. And inflation to various types of establishments, web traffic, road traffic, cannot measure the change of the cost of living over time, if and the exact timing of stimulus and other benet c fi hecks only because the meaning of “living” changes continuously. to individuals all proved to be essential to understan -d An idea sometimes used in philosophy and psychol - ing how economic activity was behaving in real time. The ogy, though not economics, is that of an “affordance.”17 In actual tracking of COVID transmission data, hospitaliz-a economics, an ao ff rdance might be thought of as something tion, mortality, and other correlates also proved useful in that enables a person the ability to accomplish some type of assessing economic prospects. In the second quarter of 2020, work, but that also ae ff cts how they live by, say, changing the an extreme drop in GDP occurred in many countries, and environment in which it occurs. The invention of fire was in the third quarter there was a rebound whose strength an affordance that enabled cooking, which in turn dramati- surprised many analysts. But infections rebounded from cally improved human life. In past millenia, humans, like a brief decline in the third quarter too, and expectations other primates, likely spent much of their time masticating for a strong GDP bounce hardly told the story of what was raw meat. After generations of cooking food, humans began happening. High frequency measures of foot traffic and to lose the ability to digest some things they used to eat. Fire other things suggested that growth was facing limits locally and cooking began as a luxury, but they created change, new where new outbreaks were underway. liabilities, a requirement to find fuel, a need to “afford the ao ff rdance.” The smartphone is a modern-day affordance. For many “ people, it has already become hard to imagine life without one. If the most prestigious data reinforce the wrong Expensive phones and data eat up household budgets, and message, then the data itself can be a cause of genuine low-income households struggle, work more, and incur debt to pay for their phone and data. Although this new affordance mischief, including misguided public policy as well as led to additional measured economic activity, comparing distorted private decision-making. “living standards” before and after smartphones (or fire) is not as straightforward as GDP data suggest. It’s hard enough to measure the relative happiness or well- being of one person compared to another. The task becomes e Th broad subjective impression of the economy’s success even less well-den fi ed once you acknowledge how dic ffi ult it according to households, voters, and policymakers can also is to measure the happiness or well-being of the same person be useful information. Even when wonderful technological over time. Does this mean the data are useless? Of course gains occur in a time of peace, consumer opportunities, and not. Nominal GDP is a useful starting point for measuring improving health care, if people believe the economy is not economic activity. making them better o,ff then their decisions are likely to be Some economists—and lots of non-economists—scoff at inu fl enced by that mindset. (And the real possibility that the macroeconomic indicators that markets and policymak - some improvements in income distribution have actually ers focus on now. What matters are things that are genuinely been taking place in recent years could easily get lost amid observable and relevant to the decisions made by households, the widespread messaging and general impression of “no firms, and policymakers: nominal income, the profit and wage progress.”) shares, the income distribution across households, and labor If the most prestigious data reinforce the wrong message, then the data itself can be a cause of genuine mischief, inclu - d ing misguided public policy as well as distorted private decision-making. Take, for instance, the claim that “real 17 The term is attributed to psychologist James J. Gibson. Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 15 median income has fallen for decades.” Would the median highs. To be sure, the gains have been especially conc- en household really prefer the opportunities available to them in trated in large technology companies, including some whose 1980 to what they can have now with the supposedly same business models relied heavily on intangible capital. But the real income? relationship between GDP and equity gains is loose in the Meanwhile, guided by central bank behavior, some inv- es best of times. And the wedge between the value creation tors take ideas like potential growth quite literally, forgetting being projected in today’s stock prices, and the stagnation the error bands that even policymakers warn people to place that appears to be evident in national accounts indicators, around their estimates of real GDP. Statistics and their histo -r provides yet another alert to the differences between concepts ical distributions can lead to hubristic forecasts derived from and statistics that this essay has covered. past data in a world still clearly in the midst of transforming. e Th re are other measures of the net worth of the private Janet Yellen knows the problems of inflation measurement, sector besides the equity market. If measured at market value, but 2% inflation has been a useful anchor for market expecta - these may reflect the capital gains that have accrued to the tions of monetary policy. In one sense, the cost of living might owners of real estate and business equity. The strong increase not really be rising by close to 2% year after year, but at least in private sector net worth in the U.S. in recent years is at odds this inflation measure is anchoring monetary policy, much as with slow growth in GDP, and is another hint that something the price of gold alone once did. is o i ff n its measurement. If we insist on measures of real data, one alternative We should think about the past, the present, and the to GDP is industrial production (IP), whose measurement future differently. We don’t know much about the future predates GDP. IP measures physical goods that can usually and never have. As for the present, we have big income and be stored and traded. These need financing, and IP data are wealth disparities, global imbalances, amazing new techn-ol highly correlated with financial market measures. Histori- ogies, slower-than-usual nominal growth, high corporate cal approaches to national income have sometimes focused prot fi s relative to nominal GDP and labor income, unusual on only physical or agricultural goods. This is wrong in monetary policy, changing demographics, debt/budget many ways, but short-term swings in the physical good s concerns, political economy tensions, and a broad sense of sector are nevertheless useful for investors who want to discomfort reported and no doubt experienced by many track the business cycle over short periods. For example, my households. Condensing all that into a single productivity Credit Suisse colleagues and I make extensive use of indu - s or GDP statistic does not begin to do justice to the complex trial production and related goods sector data when trying reality of the U.S. economy.   to understand short-term financial market and economic developments. Other alternatives include various attempts at “happ - i James Sweeney is Credit Suisse’s Chief Economist and Regional CIO ness indices,” ad hoc constructed “quality of life” indices for the Americas. He manages a 30-person economics department spread that impose a fixed definition of what a quality of life is, and worldwide. His written topics focus on the global and U.S. economies, direct measures of innovation. In many OECD economies, particularly on global industrial production and the workings of the global the growth of patents and the general spread of published financial system. His work appears regularly in major international finan - information have shown no signs of a slump. Intuitively, the cial media and James travels extensively to meet clients and policymakers. evolution of the Internet has allowed for a faster dissemination James is a member of Credit Suisse’s Zurich-based investment committee, of information, increased the possibilities for collaborative which determines investments for some of the firm’s more than $800bn research, and generally made formerly arcane knowledge in assets under management, the Americas Operating Committee of the widely available. New technologies, including innovations Global Markets Division, and the Americas Pension Investment Committee. in machine learning, robotics, and biology, are obvious drivers of profound innovation. The recent development of Previously, James held roles including Head of Fixed Income Research and the CRISPR gene editing technology has the potential to Chief Fixed Income Global Strategist. He joined Credit Suisse in 2000, profoundly shape our future, but its signic fi ance cannot (and having previously worked as a political speechwriter. He holds an MSc in will not) be captured with patent data or GDP. economics from the London School of Economics and a BS, magna cum Of course, one forward-looking measure of future income laude, from Florida State University. James is a member of the Council is the stock market, which did very well in the years between on Foreign Relations, the Economic Club of New York, and the National 2008 and the 2020 pandemic. And after plunging in March Committee on United States – China Relations. 2020, both the S&P 500 and Nasdaq have set new record 16 Journal of Applied Corporate Finance • Volume 32 Number 4 Fall 2020 ADVISORY BOARD EDITORIAL Yakov Amihud Carl Ferenbach Donald Lessard Clifford Smith, Jr. Editor-in-Chief New York University High Meadows Foundation Massachusetts Institute of University of Rochester Donald H. Chew, Jr. Technology Mary Barth Kenneth French Charles Smithson Associate Editor Stanford University Dartmouth College John McConnell Rutter Associates John L. McCormack Purdue University Amar Bhidé Martin Fridson Laura Starks Design and Production Robert Merton Tufts University Lehmann, Livian, Fridson University of Texas at Austin Mary McBride Massachusetts Institute of Advisors LLC Erik Stern Michael Bradley Assistant Editor Technology Stern Value Management Duke University Stuart L. Gillan Michael E. Chew Gregory V. Milano University of Georgia G. Bennett Stewart Fortuna Advisors LLC Richard Brealey Institutional Shareholder London Business School Richard Greco Stewart Myers Services Filangieri Capital Partners Michael Brennan Massachusetts Institute of René Stulz University of California, Trevor Harris Technology The Ohio State University Los Angeles Columbia University Robert Parrino Sheridan Titman Robert Bruner Glenn Hubbard University of Texas at Austin University of Texas at Austin University of Virginia Columbia University Richard Ruback Alex Triantis Charles Calomiris Michael Jensen Harvard Business School University of Maryland Columbia University Harvard University G. William Schwert Laura D’Andrea Tyson Christopher Culp Steven Kaplan University of Rochester University of California, Johns Hopkins Institute for University of Chicago Alan Shapiro Berkeley Applied Economics University of Southern David Larcker Ross Watts California Howard Davies Stanford University Massachusetts Institute Institut d’Études Politiques Martin Leibowitz Betty Simkins of Technology de Paris Morgan Stanley Oklahoma State University Jerold Zimmerman Robert Eccles University of Rochester Harvard Business School Journal of Applied Corporate Finance (ISSN 1078-1196 [print], ISSN 1745-6622 Statement on Research4Life [online]) is published quarterly per year by Wiley Subscription Services, Inc., a Wiley Wiley is a founding member of the UN-backed HINARI, AGORA, and OARE initiatives. Company, 111 River St., Hoboken, NJ 07030-5774 USA. They are now collectively known as Research4Life, making online scientific content available free or at nominal cost to researchers in developing countries. Please visit Postmaster: Send all address changes to JOURNAL OF APPLIED CORPORATE FI- Wiley’s Content Access – Corporate Citizenship site: http://www.wiley.com/WileyCDA/ NANCE, John Wiley & Sons Inc., c/o The Sheridan Press, PO Box 465, Hanover, PA Section/id-390082.html 17331 USA. Journal of Applied Corporate Finance accepts articles for Open Access publication. Information for Subscribers Please visit https://authorservices.wiley.com/author-resources/Journal-Authors/open- Journal of Applied Corporate Finance is published quarterly per year. Institutional sub- access/onlineopen.html for further information about OnlineOpen. scription prices for 2020 are: Print & Online: US$844 (US), US$1007 (Rest of World), €656, (Europe), £516 Wiley’s Corporate Citizenship initiative seeks to address the environmental, social, (UK). Commercial subscription prices for 2020 are: Print & Online: US$1123 (US), economic, and ethical challenges faced in our business and which are important to US$1339 (Rest of World), €872 (Europe), £686 (UK). Individual subscription prices our diverse stakeholder groups. Since launching the initiative, we have focused on for 2020 are: Print & Online: US$137 (US), $137 (Rest of World), €115 (Europe), sharing our content with those in need, enhancing community philanthropy, reducing £79 (UK). Student subscription prices for 2020 are: Print & Online: US$49 (US), our carbon impact, creating global guidelines and best practices for paper use, estab- $49 (Rest of World), €41 (Europe), £28 (UK). Prices are exclusive of tax. Asia-Pacific lishing a vendor code of ethics, and engaging our colleagues and other stakeholders GST, Canadian GST/HST and European VAT will be applied at the appropriate rates. in our efforts. For more information on current tax rates, please go to https://onlinelibrary.wiley.com/ library-info/products/price-lists/payment. The price includes online access to the cur- Follow our progress at www.wiley.com/go/citizenship. rent and all online back files to January 1, 2016, where available. For other pricing options, including access information and terms and conditions, please visit https://on- Abstracting and Indexing Services linelibrary.wiley.com/library-info/products/price-lists. Terms of use can be found here: The Journal is indexed by Accounting and Tax Index, Emerald Management https://onlinelibrary.wiley.com/terms-and-conditions. Reviews (Online Edition), Environmental Science and Pollution Management, Risk Abstracts (Online Edition), and Banking Information Index. Delivery Terms and Legal Title Where the subscription price includes print issues and delivery is to the recipient’s Disclaimer address, delivery terms are Delivered at Place (DAP); the recipient is responsible for The Publisher, Cantillon and Mann, its affiliates, and Editors cannot be held respon - paying any import duty or taxes. Title to all issues transfers FOB our shipping point, sible for errors or any consequences arising from the use of information contained in freight prepaid. We will endeavour to fulfil claims for missing or damaged copies within this journal; the views and opinions expressed do not necessarily reflect those of the six months of publication, within our reasonable discretion and subject to availability. Publisher, Cantillon and Mann, its affiliates, and Editors, neither does the publication of advertisements constitute any endorsement by the Publisher, Cantillon and Mann, its Journal Customer Services: For ordering information, claims and any inquiry concern- affiliates, and Editors of the products advertised. ing your journal subscription please go to https://hub.wiley.com/community/support/ onlinelibrary or contact your nearest office. Copyright and Copying Americas: Email: cs-journals@wiley.com; Tel: +1 781 388 8598 or Copyright © 2020 Cantillon and Mann. All rights reserved. No part of this publication +1 800 835 6770 (toll free in the USA and Canada). may be reproduced, stored or transmitted in any form or by any means without the Europe, Middle East and Africa: Email: cs-journals@wiley.com; prior permission in writing from the copyright holder. Authorization to photocopy items Tel: +44 (0) 1865 778315. for internal and personal use is granted by the copyright holder for libraries and other Asia Pacific : Email: cs-journals@wiley.com; Tel: +65 6511 8000. users registered with their local Reproduction Rights Organization (RRO), e.g., Copy- Japan: For Japanese speaking support, Email: cs-japan@wiley.com right Clearance Center (CCC), 222 Rosewood Drive, Danvers, MA 01923, USA (www. Visit our Online Customer Help at https://hub.wiley.com/community/support/onlineli- copyright.com), provided the appropriate fee is paid directly to the RRO. This consent brary does not extend to other kinds of copying such as copying for general distribution, for advertising or promotional purposes, for republication, for creating new collective Production Editor: Shalini Chawla (email: jacf@wiley.com). works or for resale. Permissions for such reuse can be obtained using the RightsLink “Request Permissions” link on Wiley Online Library. Special requests should be ad- Back Issues: Single issues from current and recent volumes are available at the dressed to: permissions@wiley.com. current single issue price from cs-journals@wiley.com. Earlier issues may be obtained from Periodicals Service Company, 351 Fairview Avenue – Ste 300, Hudson, NY 12534, USA. Tel: +1 518 537 4700, Fax: +1 518 537 5899, Email: psc@periodicals.com View this journal online at wileyonlinelibrary.com/journal/jacf.

Journal

Journal of Applied Corporate FinanceWiley

Published: Dec 1, 2020

There are no references for this article.