The Trouble with Economic Statistics
[The Independent Review, (1998)]
Economists have been grousing a good deal lately about the deteriorating quality of basic economic statistics — official data on prices, incomes, employment, productivity, and poverty, among other things — and about the lack of government funding to remedy the problem. On its face, the complaint seems reasonable and practical.
But I wonder. Having used official economic statistics from time to time for some 35 years, I would miss them if they were to disappear. Yet, however put out I might be as an economic analyst, I suspect that the world would be a happier place had these figures never been created. Certainly the statistics are often inaccurate or otherwise flawed and hence misleading. An even more serious consideration, however, is that the official statistics help to provide rationales for pernicious policy making.
Poorly Defined, Imprecise, and Invidious
Because they are ill-defined conceptually, many official economic statistics fail to capture what they purport to measure. Figures on "poverty," for instance, are notorious in this regard. Is poverty an absolute or a relative condition? If the latter, what is the proper standard of comparison?
Obviously, the living conditions of many Americans below the "poverty line" must seem affluent to billions of submerged denizens of the Third World. Apart from international comparisons, many Americans now classified as poor would have seemed well-to-do in the eyes of, say, their grandparents. Above a certain absolute income, "poverty" becomes less a definite condition than a staging area from which armies of redistributionists launch their attacks on higher-income people.
Aside from the conceptual questions, the mere measurement of personal income as currently defined poses nearly insurmountable difficulties. For example, among those in "poverty," illegal (hence unreported) incomes loom large — earnings from drug dealing, prostitution, gambling enterprises, and everyday theft. If the poor have only the income they report to the Internal Revenue Service or the Bureau of the Census, how do they come by the automobiles, televisions, jewelry, and other visible adornments of their homes and persons?
Of course, the poor are scarcely the only class concealing real income, whether honestly or illicitly acquired. The wealthy support an entire stratum of professional attendants — lawyers, accountants, financial gurus — whose sole mission in economic affairs is to remove income from the gaze of the tax collector. Small-business people notoriously accept payments "under the table," and hosts of carpenters, painters, electricians, plumbers, and gardeners, not to mention the nannies, earn income that is wholly or partly unreported.
In one of the most important and unjustly neglected economics books of the past 50 years, Oskar Morgenstern warned, "We must carefully distinguish between what we think we know and what we really do and can know."1
Yet all too often economists avert their eyes, plowing blithely ahead with exquisitely sophisticated econometric analyses of virtually meaningless or inaccurately measured variables. As Michael J. Boskin attests,
Both the economics and statistics professions have become more theoretical and spend less time on the practical issues of sampling, data collection, quality of data, and providing professional rewards in terms of standing in the profession for those who show great skill in finding, developing, or improving data.2
One hesitates, then, to blame laypersons for reacting to the drumbeat of media reports of a widening distribution of income during recent decades in the United States. Is this "growing inequality" not a fact? Who really knows? But whether in some purely arithmetic sense it is or not, it would never have been made the basis for public-policy proposals to "correct" the situation if statisticians had not constructed "the distribution of income" in the first place.
It is hard to imagine another statistical artifact better calculated to feed the fires of envy and political rapacity. Such information is unnecessary for the conduct of a just government but well-nigh indispensable for the operation of a predatory one.3
Nourishing the Mercantilists
In an uncertain world, one thing is sure: every month, without fail, the press will prominently report the latest official figures for the US international "trade deficit." Even in a relatively intelligent article,4 the graph is labeled "That Pesky Trade Deficit." Of course, the very term "deficit" has a negative connotation, suggesting a shortfall of some, presumably regrettable, kind. Clearly the journalists, along with the proverbial man in the street, regard the trade deficit as a bad thing.
Often they highlight some ominous bilateral trade imbalance, especially the perennial trade deficit of the United States with Japan. No doubt it dampens one's spirits to be told repeatedly that the nation is being "flooded" by imports, that it is "awash" in cheap foreign goods.
Yet anyone who stops to consider how someone might keep track of all the goods and services being exchanged across America's borders must develop some fundamental doubts. Upon being informed that the trade deficit is X, one might ask: how do they know? Of course, the Customs Service generates mountains of data on international trade, but surely many transactions escape the agency's surveillance — for instance, the sizable commerce in illegal drugs, estimated at $400 billion per year worldwide, in which the United States looms large as a net importer.5 Nor are illicit drugs the only products covertly imported or exported.
Morgenstern aptly warned against the unreliability of the international trade data when he wrote,
Any one who has ever sat through meetings (as the author has) in which final balance of payment figures for most invisible items were put together, can only marvel at the naïveté with which these products of fantasy, policy, and imagination, combined with figures diligently arrived at, are gravely used in subsequent publications…. Writers on all phases of foreign trade will have to assume the burden of proof that the figures on commodity movements are good enough to warrant the manipulation and the reasoning to which they are customarily subject.6
Morgenstern was writing decades ago, but the deficiencies to which he called attention have persisted. According to Boskin's recent assessment, "the trade statistics have serious flaws" and "it is becoming more and more difficult to measure trade accurately."7
Actually, the compilers of the international-trade statistics confess their inability to identify all the relevant transactions or to measure correctly the ones they do identify. Because every exchange has two sides, the overall balance of payments must necessarily balance. But in practice it never does, and the Commerce Department reconciles the two sides of the account by inserting a fudge factor called "statistical discrepancy" (formerly "errors and unrecorded transactions").
In 1996, for example, this amounted to minus $46.9 billion, equivalent to 32 percent of that year's deficit on current account. (The current account includes investment income and transfers as well as sales of goods and services.)8
The statistical discrepancy varies widely from year to year. For instance, in 1992 it was negative $43.6 billion, in 1993 positive $5.6 billion. In view of the violent fluctuations of this fudge factor, how much confidence can one place in the "fact" that between 1992 and 1993 the deficit on current account increased by $34.4 billion?9 Perhaps, despite all the hand-wringing occasioned by the increase of the measured current-account deficit in 1993, the true deficit did not increase at all. And considering the inaccuracy of even the annual data, the monthly reports featured in the press deserve no credence whatsoever.
If balance-of-trade data merely served as one more excuse for econometricians to waste their time, the data would be relatively innocuous. Unfortunately, by virtue of their routine and widespread dissemination by the news media, these figures play an important role in the politics of rent-seeking. As Paul Heyne has written, allegations of a trade deficit
provide political arguments that can be used by people who want protection from foreign competitors or subsidies for their efforts to sell abroad. For the existence of a trade deficit implies that the ratio of imports to exports must eventually decline, since no deficit can continue forever. So we might as well get on with it now: Fund the Export-Import Bank, restrict imports from nations that interfere with our exports, slap penalties on foreign firms that are "dumping" in our markets, and face up in general to the fact that free trade is good trade only if it is fair trade.10
Anyone who pays attention to the news will recognize the refrain. As Heyne observes,
The declaration of a trade deficit amounts in practice to a kind of declaration of martial law. What is most dangerous about such a declaration is that it gives government officials a license to subordinate the rule of law and respect for established rights to considerations of political advantage.11
Far better for both justice and economic prosperity if the international trade statistics had never been collected. They have been and continue to be major means to the thoroughly mischievous ends of pandering politicians and their rent-seeking supporters.
Defining Government Spending as Productive
William Petty (1623–1687), an Englishman who practiced "the art of political arithmetic," has been called "the first econometrician" and identified as "the author of the first known national income estimates."12 It would have been a boon to honest humanity had "political arithmetic" stuck as the name for economic statistics of the sort now assembled in the official national income and product accounts. This designation would have alerted one and all to the political purposes lurking beneath the construction of such figures.
Consider, for instance, how the statisticians arrive at the amount of gross domestic product: add the values at market prices of all newly produced, domestic, final goods and services purchased by consumers, investors, and governments, then throw in net exports.
So accustomed are we to this setup that no one pauses to ask, why make government purchases a separate category? And if we include government services, how can we value them at market prices, inasmuch as they are generally provided without charge and financed by taxation?
Nowadays not even many economists know that prior to World War II the inclusion of a government category in the national income and product accounts was a hotly debated issue. Now, however, as Ellen O'Brien has written, "it is rare that someone suggests that the current treatment of the government product in national income is flawed and it is nearly inconceivable that it would be suggested that a government product doesn't belong at all.13
As O'Brien notes, "the treatment of the government sector put in place in 1947 (which has remained standard practice in the US since that date) was initiated by estimators in order to assess the impact of the tremendous increase in war expenditures on the economy" during World War II. The "theoretical debates from the pre-war period continued through 1947 and were never fully resolved."14
Simon Kuznets, the principal architect of the early national income accounts in the United States, played David in this David-and-Goliath struggle, but ultimately he could not prevail against the vastly superior resources of the US Department of Commerce, and he retreated from the field of battle.15
In reviewing this dispute, O'Brien puts her finger on a critically important but scarcely appreciated consideration:
It seems as if the government bureaucrats were determined to emphasize the importance of government's role in the economy by enlarging the share of government expenditure in the national income. While this official change was motivated by the great increase in government expenditures caused by the war, the underlying explanation must relate to the Commerce estimators' philosophy of the proper role of government in the economy.16
It is easy to see how postwar Keynesian doctrine and policy making meshed with a system of national accounts in which all government purchases of goods and services — the services of government employees being valued at whatever the employees happen to be paid — count equally with private purchases of only final goods and services in the market, where consumers and investors demonstrate their valuations by spending their own money.
Among the many repercussions of adopting the official national income and product accounts was the perpetuation of the myth of "wartime prosperity" during World War II, along with the crackpot theories of "military Keynesianism" that the myth fostered. Eliminate the government component, which was almost entirely devoted to purchasing munitions and paying the personnel of the bloated, mainly conscripted armed forces, and you are left with national product data indicative of wartime recession.17
Later, especially in the 1950s, the official accounts gave rise to a distorted picture of the business cycle.18
For more than half a century the official national income and product accounts have served as the map used by policy warriors to plan their assaults on Fine-Tune Mountain. (For this indictment, the annual reports of the Council of Economic Advisers contain incontestable evidence, as do successive generations of macroeconomics textbooks, written as though intended to serve as cookbooks for policy makers.) If the macroeconomic warriors failed to capture the hill, perhaps the blame belongs to their tactics or weaponry as well as to their faulty map. But lacking that map, they might have been more reluctant to sally forth, and hence the American economy might have been spared the ravages of these pretentious idiot savants.
That economists have passively accepted economic statistics designed and constructed by government bureaucrats ranks among the more shameful aspects of their professional conduct in the 20th century. One wonders, who was using whom?
In the post–World War II era, this intermingling of an ever more intrusive government and an economics profession dedicated to instructing the intruders achieved solid institutionalization. It is now the status quo, with every prospect of remaining so. But the economic statistics joining the two sides of this symbiosis are often ill-defined, inaccurate, and productive of mischief when used in policy making.
Morgenstern considered it
necessary that worthless statistics be completely and mercilessly rejected on the ground that it is usually better to say nothing than to give wrong information which — quite apart from its practical, political abuse — in turn misleads hosts of later investigators who are not always able to check the quality of the data processed by earlier investigators.19
The advice remains sound, however little it has been or will be heeded. Economics would progress faster if economists asked more hard questions before admitting official data into their analyses.
On a wider front, where the interests of the general public come into play, economic statistics might be put to an even more stringent test. Suppose we were to pose seriously the question asked rhetorically by journalist Peter 20Passell in his discussion of international trade statistics "The transactions are voluntary and generally take place between consenting adults," Passell observed. "So why is it anyone's business but theirs?" Why indeed?
A just government, one that confines itself to protecting the citizens' rights to life, liberty, and property, has no need for figures on the distribution of personal income; no need for data on international trade and finance; no need for national income and product accounts. None of these statistics can assist in the defense of the citizens' just rights. Such figures belong to "political arithmetic." They are raw materials for rent-seekers and government officials who would make government an engine of predation and a destroyer of just rights.
- 1. On the Accuracy of Economic Observations, 2d ed. [Princeton: Princeton University Press, 1963], vii.
- 2. "Some Thoughts on Improving Economic Statistics to Make Them More Relevant in the Information Age," prepared for the Joint Economic Committee, Office of the Vice Chairman, United States Congress, October 1997, 5.
- 3. Here I stand by my previous statement, "Is More Economic Equality Better?" Intercollegiate Review 16 (Spring/Summer 1981): pp. 99–102.
- 4. Peter Passell, "The Fear Is Gone, Not the Danger," New York Times, March 1, 1998.
- 5. Mark J. Porubcansky, "U.N.: Drug Dealing Is 8% of All Trade," Seattle Times, June 26, 1997.
- 6. On the Accuracy, p. 180.
- 7. "Some Thoughts," p. 9.
- 8. For the data, see US Council of Economic Advisers, Annual Report for 1998 [Washington, D.C.: US Government Printing Office, 1998], p. 399.
- 9. Ibid., pp. 398–99.
- 10. "Do Trade Deficits Matter?" Cato Journal 3 (Winter 1983–84): p. 711.
- 11. Ibid., p. 715.
- 12. Phyllis Deane, "Petty, William," in International Encyclopedia of the Social Sciences, vol. 12, edited by David L. Sills (New York: Macmillan and The Free Press, 1968), p. 67.
- 13. How the 'G' Got into the GNP," in Perspectives on the History of Economic Thought, vol. 10, Method, Competition, Conflict and Measurement in the Twentieth Century, edited by Karen I. Vaughn (Aldershot, Eng.: Elgar, 1994), p. 247.
- 14. Ibid., 242.
- 15. See, however, the rearguard action he mounted in his Capital in the American Economy (Princeton: Princeton University Press, 1961), pp. 465–84.
- 16. O'Brien, p. 252.
- 17. Robert Higgs, "Wartime Prosperity? A Reassessment of the U.S. Economy in the 1940s," Journal of Economic History 52 (March 1992): pp. 41–60.
- 18. Robert Higgs, "The Cold War Economy: Opportunity Costs, Ideology, and the Politics of Crisis," Explorations in Economic History 31 (July 1994): pp. 283–312, esp. pp. 297–98.
- 19. On the Accuracy, pp. 55. For a similar warning, see Boskin, "Some Thoughts," p. 20.
- 20. Passell, "The Fear Is Gone."