[This is Chapter Five of Murphey’s book The Emerging Crisis of Economic Displacement.]

 

Chapter Five

 

WHY TODAY'S ECONOMIC STATISTICS WON'T BE PARTICULARLY HELPFUL IN OUR INQUIRY

 

            There have been times, such as the summer of 1997, when economic statistics for the United States have told an almost-euphoric story of national well-being: inflation far down, unemployment at the lowest levels in years, increasing productivity, a booming stock market, and the federal budget deficit falling so rapidly that policy-makers have begun to talk about what to do when the budget starts showing a surplus.

            All the while, there have been continuing news reports of corporate mergers, downsizings and restructurings, with abundant layoffs.  In their personal lives many Americans have several acquaintances who are  retiring early, or moving through a series of temporary jobs, or working at employment far below their qualifications, or waiting with bated breath to be ousted in a "reduction in force," or retaining a job only by moving hundreds of miles to a new location, tearing up roots with family and friends that the person had hoped would last a lifetime.  For many people, their own income has hardly increased in years, certainly not to match inflation, and they know that making any progress at all has depended upon a second income brought in by a working spouse.  Often at the same time, however, if they are fortunate, they receive quarterly reports from their retirement stockfund informing them of gains at a dizzying pace, portending (if all remains well) an affluent old age.

            How are these seemingly incongruous elements to be reconciled?

            More by theory, certainly, than by statistics.  It would seem that almost anything can be "reconciled," or not, depending upon ones wishes.  Economic statistics today, in the hands of differing interpreters, can and do give widely varying pictures. 

            It is plausible to speak of "a crisis of statistics."  As we talk in coming chapters about worker displacement and economic polarity, we will see that every subject is lost in a welter of claims and counterclaims.  All points of view cite statistics to confirm their position, so that it is almost impossible to penetrate through to reality on the basis of statistics alone.  In this vein, Business Week in March 1996 spoke of a resulting "cognitive dissonance."[1]  Kenneth Auchincloss in Newsweek, talking about the figures for per capita income, says "the statistical argument is an impenetrable thicket."[2]  In an enlightening article entitled "The Real Truth About the Economy: Are Government Statistics So Much Pulp Fiction?," Business Week in November 1994 said that "the economic statistics that the government issues every week should come with a warning sticker," adding that "the government is pumping out a stream of statistics that are nothing but myths and misinformation."[3] 

            Yet, it is striking that Alan Greenspan, who as chairman of the Federal Reserve Board holds center stage in making economic policy, is willing freely to use standard statistics on such things as unemployment and the Gross Domestic Product as though they are fully meaningful, which they probably are for his purposes, since he's certainly no fool.[4]  He sees the statistics, in their usual form, as conveying significant information.   Just the same, he acknowledges the problem when he says that "the list of shortcomings in U.S. economic data is depressingly long."[5]

            Whatever continuing value the statistics may have for some uses, for the purpose of understanding where the United States' and the world's economies are going in response to global high-tech, the criticisms I just referred to in Business Week have considerable force.  A current statistical category such as "Gross National Product" was adopted after World War II as a tool that was useful to economists at that time.  With changing conditions, it and other measures are becoming increasingly ill-suited.  The Economist suggests that a major reason "for the inadequacy of official statistics may be that the economy is simply becoming unmeasurable.  Conventional methods of measuring the economy are no longer up to the task, and economists cannot agree on how to improve them."[6]   The result is that we are forced to rely more on theoretical insight than we would like to in knowing the current state of things.  We would prefer, whenever we are dealing in the more purely empirical realm of trying simply to "know what's going on," that the "facts do the talking."  The facts should speak loudly, with theory's role being to explain causation and relationships.  But "the facts" are a welter of confusion, and can be chosen to confirm almost anything.

            My chapters on the displacement will cite statistics at length, since that is naturally the form the discussion takes in the literature, but we will need to keep in mind the weakness of the data and eventually rely much more on two things: (a) the agreement that exists among economists of all persuasions about the major trends that are occurring; and (b) the fact that economic theory itself points to those trends and their implications.  

            I offer these thoughts about statistics not as a form of intellectual nihilism seeking to nullify empiricism itself by denigrating its tools, but to inculcate a wholesome wariness about the statistics one confronts at every hand today and a desire to use statistics only with considerable sophistication about what, in any given case, the statistic includes and does not include.  The fact is that a proper use of economic statistics requires great care.  Many of the citations we see of statistics don't inform the reader adequately of the precise methodology used, especially with regard to what's included and what is not, and because of this shouldn't be accepted at face value.  This will become especially clear as we review many of the main statistical categories presently in use and see what those writing on economic problems have had to say about them.

            The commentary that follows is drawn from that literature.  In preparing to write this book, I have read approximately 25,000 pages of recent economic writing and collated notes on hundreds of separate points.  This has made possible a broad overview of the current economic dialogue, including its many disputes on the significance or appropriateness of this or that statistic.  I am not a statistician myself and do not purport to bring expertise of my own to the debate.  My reading has, however, gathered in one place the thoughts of several economists and economic commentators. It will be informative to review what they have to say.

            We will begin with some of the more general observations about economic statistics and proceed to a review of such specific statistical categories as, e.g., "Gross National Product," "unemployment" and the "Consumer Price Index."  Readers interested in the main themes of this book but not in technical detail may prefer to read the general comments, but then just to scan the more specific review of the categories.  

 

General observations in the literature

            .  That the changing economic reality is making many statistical categories inadequate.  The observation is often made that the present and coming reality of goods and services is shifting from underneath the customary statistical categories.  It is with this in mind that Walter Wriston speaks of "the increasing inaccuracy or irrelevance of our standards of economic measurement," and says that "many of the terms we use today to describe the economy no longer reflect reality."[7]  According to Business Week, the problem is largely that "most of the surging information economy -- including software, telecommunications, and entertainment -- is poorly covered by the data."  Its discussion points out that such categories as gross domestic product (GDP), producer price index (PPI), and capacity utilization speak to traditional manufacturing, as was appropriate when they were conceived.  This gives rise to such yawning gaps between the data and current reality as this: "Business investment in equipment, after adjusting for depreciation, is a full 30% bigger than the government statistics say."  This is because most firms' spending on software and telecommunications equipment isn't counted.  Productivity increases are underestimated because services are difficult or impossible to measure, and unemployment data miss a large number of males who have dropped out of the workforce altogether.  As we go through the specific economic measures, we will point to many discrepancies of this sort.

            Of the authors I have read, Alvin Toffler has looked furthest into the future so far as this increasing gap is concerned.  In Future Shock, published as long ago as 1970, he foresaw a day when society will have changed so much that "the very purposes of economic activity" will be different than they now are, with the result that "even the most sophisticated tools of today's economists [will be] helpless."  He asked, as an example, whether Gross National Product would remain a meaningful concept in a country where "no growth" were to replace "growth" as the objective.[8]

            .  That in today's emerging high-tech economy, statistics give little indication of the underlying direction of movement.  Most of the applications and effects of information-based, non-labor-intensive technology have barely begun to catch hold.  They show up in economic data more as the proverbial "small cloud on the horizon" than as already-established phenomena.  A serious discussion of society looking even to the near-term future requires an eye toward emerging developments more than toward data that describe even the very recent past.  The difference in perception that this occasions is part of the statistical debate between those who describe the current market economy in glowing terms and those who see vastly significant issues looming of worker displacement and income/wealth polarity.  For the latter, a look ahead supercharges some of the data with more significance than they would seem to merit at face value.

            .  That there are important reasons to be suspicious of economic statistics.  A century ago, Henry George warned that:

 

                        As for reliance upon statistics, that involves the additional difficulty of knowing whether we have the right statistics.  Though ‘figures cannot lie,' there is in their collection and grouping such liability to oversight and such temptation to bias that they are to be distrusted in matters of controversy until they have been subjected to rigid examination....Under their imposing appearance of exactness may lurk the gravest errors and wildest assumptions.[9]

 

            The innumerable ways that statistics can convey an inaccurate impression comprise a subject of their own.  Odds and ends from my notes help illustrate this: economist Paul Krugman warns, for example, that "it is important to be careful about starting and ending dates."[10]  The varying phases of the trade cycle can make a big difference.  If data are taken from the top of a boom and compared to those later from the bottom of a recession, an unfavorable result will follow; if the peaks and valleys are reversed in arriving at the "before and after," the results will be much more upbeat; the most informative (and honest) finding will come from data taken from comparable points in the trade cycle.

            Some types of "data" are squishy, even though social scientists may use them (with or without caveats).  Surveys, especially if they ask for subjective responses, are especially "soft."  The sample that is taken may in many ways differ from the population the survey seeks to describe.  The authors of one survey about underemployment in Kansas acknowledged, as was proper, that there were "biases" from too-great disparities relating to urban versus rural, men versus women, the ages of the respondents, and levels of education.  Often, many of those surveyed refuse to answer questions, such as to income, and this introduces a significant imponderable.  And when somebody answers a subjective question, such as about whether he believes himself to be "underutilized," the answer itself depends upon personality and subjective factors that obscure the meaning even though the numerical scoring gives an impression of exactitude.[11] 

            In some subjects an element of selectivity is introduced just because researchers have found reasons to study some things more than others (sometimes, no doubt, in response to whatever grants are made available for research).  We are told, for example, that "most of the systematic study of the diffusion of information technology is biased in significant ways.  The majority of studies deal with manufacturing process technologies, and with nothing equivalent for services or office applications...."[12]  The little-studied service area is, of course, the one that looms ever larger as we enter the Information Age.  Similar gaps in the data appear when it comes to comparing something like employment around the world: statistics just aren't available for many countries and certainly for much of the past; and where they are available, the quality differs widely.[13]

            .  That economic statistics are often manipulated for political or ideological purposes.  The point hardly need be made, since it is so obvious, that statistics can easily become the handmaidens of charlatans, ideologues and special pleaders.  Unfortunately, no point along the social-political spectrum can claim that all its partisans are free of this, even though we would like to think our own point of view, being correct, presents its case with an integrity unknown to the others.  The debates of our own time offer many examples of the misuse of statistics.

            Labor Secretary Robert Reich's pessimistic data about American jobs and income is challenged by free-market economist Richard Vedder when he points out "that Reich's data do not include fringe benefits, such as medical insurance, paid vacations, and pension plans."  This criticism makes sense, since data stripped of those elements tell just part of the story.  In the column that speaks of Vedder's criticism, economist Mark Skousen says that it is often better to look at the "quantity, quality, and variety of goods and services" people are consuming than to look at average real wages.[14]

            On the other hand, free-market economists John H. Hinderaker and Scott W. Johnson, in an article much welcomed by free-market supporters defending today's economy from alleged extremes of inequality, use selective data to their own advantage when they note, for purposes of making a comparison with today to show that there is less inequality than there used to be, that "in 1913, the richest five percent of the population earned about thirty percent of aggregate income (excluding capital gains)."[15]  But why exclude capital gains?  A true picture of the "earnings" of well-to-do people can hardly be given while eliminating one of their principal sources of wealth-enhancement.  (We will see in the later section here on income statistics that the capital gains issue is a complicated one, indeed.)  At least Hinderaker and Johnson tell their readers of the exclusion.  But the notification hardly seems enough in a context where the insufficient "earnings" data are used to make a seemingly telling point.  Without the more complete comparison of 1913 with today, we don't know what to think of their argument.  They would like us to believe, though, that they have made their case.  (None of this is to say that their argument is less than honest.  It is almost impossible to make any contention persuasively, especially in writing for a popular audience, while encumbering the material with all of the qualifications that a complex subject requires.)

            Few will be surprised today that the government engages in statistical sleight-of-hand by leaving out important information.  Hank Brown and Richard Lamm, former U. S. Senator and governor respectively of Colorado, make the point that "the government's accounting system does not honestly account for the full yearly deficit or the total debt...The real cost of government services has to include the ‘unfunded liabilities' from services we consumed and promises we have made.  The official debt figures do not reflect the unfunded costs of military pensions, federal Civil Service pensions, unfunded liabilities of Social Security and Medicare...."[16]  Can we imagine someone's filling out a balance sheet as part of an application for a personal loan and not including most of his legally binding liabilities?  He would run considerable risk of being charged with fraud.  

            .  That economic statistics are available to give diametrically opposite impressions about the economy.  In the passage that spoke of "cognitive dissonance," Business Week in 1996 told how some data report "8 million jobs created in four years, the unemployment rate at 5.8%, inflation down to 2.7%, corporate profits on a four-year roll," all very favorable, while at the same time other data say "real wages have been stagnant for most the past two decades" and that "the distribution of income among Americans has become more unequal over the same period."[17]  A good example is provided by David Awbrey's May 1996 column in the Wichita Eagle telling how Kansas had benefited from U.S. trade policies; he cited the thousands of new jobs created by international trade, and how aircraft manufacture and wheat sales were up.  A responding column by Alan Tonelson and Vanessa Fuhrmans of the U.S. Business and Industrial Council Educational Foundation argued that "real median wages in Kansas -- the best measure of living standards -- have dropped by 8 percent for men and 2 percent for women since 1989... [T]he changing composition of jobs in Kansas recently explains why... Kansas' biggest job-creating industries from 1983 to 1993 were in low-paying sectors...."  The debate proceeds from there, with Tonelson-Fuhrmans pointing out how Boeing had recently shifted significant production to China.[18]

            That particular debate is just a microcosm of a similar difference of opinion -- and of statistics -- at all levels and on virtually every point.

 

Specific Statistical Categories

Income Data

            Each statistical category can be useful -- even if a number of things are not counted in arriving at it that seem to pertain -- if the statistic is used by professionals who bring to it a sharp awareness of precisely what they want it to measure.  To the extent, however, that something like "income data" is put forward in the general literature as telling non-specialists about changes in peoples' well-being, the debate over what should and should not be counted is very pertinent.  Recent income data have been criticized as leaving quite a lot out of account.

            .  Failure to consider the effect of taxes on comparative incomes.  In Upward Dreams, Downward Mobility, Frederick Strobel is critical of an income study made by Michael Horrigan and Steven Haugen that found (as summarized by Strobel) that "the middle-class decline in the 1969-1986 period was largely due to increased mobility to the upper class."  The study was based only on before-tax income data, as Horrigan and Haugen themselves pointed out when they wrote that "the ideal data, after-tax income[,] are not available."  Strobel argues that the owners of capital have managed to avoid taxes on income and property, with the incidence of taxation accordingly falling onto others.[19]  This means that the unavailability of after-tax data could reverse the outcome of the study.  Note, however, that from an opposing point of view Edwin Rubenstein argues that the exclusion of taxes from household-income data makes the rich look better off than they are.[20]  Despite their differences, Strobel, Horrigan, Haugen and Rubenstein all apparently agree that before-tax income data give a misleading picture.

            .  Failure adequately to consider capital gains and losses.  Rubenstein says it is "a common complaint" against the Bureau of the Census figures on average household income that "they exclude capital gains and therefore understate income at the top." 

            He counters this in part when he says that "to add the capital gains and not subtract the taxes, as some CBO [Congressional Budget Office] figures do, is indefensible."  To this he adds that it is useless to estimate capital gains from a sample of tax returns, since asset-owners' willingness to sell assets and report capital gains will vary from year to year based on such a thing as a change in the capital gain tax rate.  Further, he argues that it makes no sense to omit capital losses that are above the tax-deductible $3,000, as the CBO has.  Still further, a lot of capital gain isn't real earnings, but the mere increase in nominal value of an asset caused by inflation.[21]  He gives the example of a farmer who buys a farm for $50,000 and forty years later sells it for $260,000.  It looks as though the farmer has made a lot of money, but when inflation is considered he actually isn't any wealthier (and even has taxes to pay on the supposed increase!)[22]  Rubenstein points out, too, that the CBO is inconsistent when it counts some capital gain but doesn't take into consideration "capital gains and other income accruing to the middle-income brackets in the form of pension funds (which amount to some $3 trillion) or home values."[23]  He argues that, in fact, capital gains should be excluded from the income data, since many realizations of capital gains are "one-time or sporadic events" that skew any picture of annual earnings. 

            All this points to the fact that an inclusion of capital gains introduces serious problems.  Just the same, how is it possible to say that income data give an accurate picture without them?  All of Rubenstein's criticisms are well taken, but they leave intact the essential observation that the appreciation of the value of property can be, and is, an important source of earnings for many people (especially the rich, professional people with stockfund holdings for retirement, and homeowners).  Earnings in a modern economy are far from just "payroll checks received."  If what we want to measure are changes in and comparisons of well-being, it is untidy in a variety of ways to include capital gains, but also very deceptive not to.  A point to keep in mind is that most presentations of "income" figures one sees tell the reader nothing, or almost nothing, about how all these issues of inclusion, exclusion and adjustment have been resolved in arriving at the figures.  Accordingly, the true information-content of the data for the average reader should be considered next to nil.

            .  Failure to consider unreported income and non-cash benefits to the poor, the elderly, and others.  The polarization of income will be vastly overstated if major sources of support going to groups with lower or medium earnings aren't counted.  For example, Business Week reports a study that, based on tax rather than survey data, reveals that "Census Bureau surveys of income far understate the amount of pension income going to retired Americans."[24]

            In their 1988 book The Great U-Turn, Bennett Harrison and Barry Bluestone described the polarization of income and wealth in the United States, but in the course of it they rightly pointed to a vast offsetting factor: that even with the Reagan cutbacks in the number of federal transfer programs, "the total level of federal transfer payments to persons has doubled since 1979, while state and local government assistance has expanded by nearly 90 percent," all for an annual total of more than half a trillion dollars in 1987.[25]  Rubenstein cites a study in 1988 showing that the poor manage to spend twice as much as the total of their apparent income.  How is this possible?  Because "many in-kind government transfers (such as Food Stamps) are not counted as ‘money income.'"[26]   Data that don't include transfer payments of all types -- and this should include payments to the middle class and rich as well as to the poor -- just don't convey the whole story.

            The absence of unreported income also skews the results.  Even though his book on the urban poor is sympathetic, William Wilson cites a study of how welfare mothers have supplemented "their monthly AFDC checks" by money from piecework and other jobs, and from boyfriends.  The study says little of this is reported.[27]  Rubenstein tells us that Census surveys show that "a fourth of the cash income from welfare and pensions is unreported."[28]  And of course such income does not show up in income statistics.  Whether it should (even if it could) depends upon the purpose for which we seek the statistics: we will want even criminally-acquired income included if our purpose is to evaluate the actual means that people have; we won't want it included if we desire to assess, say, what comparative incomes would amount to from legitimate sources.  

            .  Incomes don't convey the entire picture if a reason for them is that people are working either more or less.  Frederick Strobel says family income statistics shouldn't be taken at face value when "longer hours worked by more members of the family" have been required to earn the money.[29]

            .  Effect of family size and type on income statistics.  Hinderaker and Johnson point out that the number of adults in the average American family has been declining since the 1970s, due to several factors that include the high divorce rate.  They say this causes an understatement of the increase in real (i.e., adjusted-for-inflation) income, since a household with fewer adults simply isn't comparable in earnings comparisons to one with more adults.  If individual income were used rather than household income, a more favorable picture would emerge.  In addition, they criticize the CBO's lumping of "unrelated individuals" [i.e., single people] into the category of "families."[30]

            Rubenstein observes that household income comparisons are skewed by the fact that there are now many more households headed by women, often not working but without a husband, than there used to be.[31]  Thus, comparisons are an "apples-and-oranges" proposition.        

            Given the nature of all this, it can't be surprising that Harrison and Bluestone make an observation that points the other direction.  Since World War II in the United States, things have changed drastically about how many adults are employed per household.  The norm at one time was that only the husband earned an income; now, many wives bring in earnings.  The result is that "the wages of every single worker in the economy could decline and family income could still rise," which is what Harrison and Bluestone say has happened in many families.[32]          

            .  Effect of weakness of income-survey data.  Karl Zinsmeister in American Enterprise speaks of the unreliability of statistics based on surveys: they "are only as accurate as the answers given to surveyers [sic] who knock on people's doors or call on the phone."[33]  A weak reed, indeed.

            .  Problems with both averages and medians.  Kevin Phillips points out that a serious problem with citing an average for incomes, so far as showing the well-being of most people is concerned, is that the average is skewed upward by large income increases at the top.[34]  Wallace Peterson says this criticism applies especially to averages about the 1980s, "when the income of families in the top tenth of the population grew many times faster than the income of all other families."[35]   A median [the figure at which half the population is above the figure and half below it] is more descriptive than an average in cases where the distribution isn't bell-shaped.  Nevertheless, median-income figures are only as good as the issues of inclusion, exclusion and adjustment we have mentioned (and others that have escaped mention here) allow them to be.[36]

            Much of the difficulty we have noted thus far has had to do with household income comparisons.  If the focus is shifted, though, to per capita comparisons, some of the same problems remain and new problems enter.  Among the former, the problem about averages, if they are used, applies.[37]  Ravi Batra points to a new problem when he says that a woman's work while a full-time homemaker hasn't been counted in the data at all, whereas when she takes an outside job the earnings show up as income.  The result is that per capita income is shown to leap upward, when in fact "the rise does not necessarily reflect rising prosperity."[38]  Wallace Peterson adds that, to know what is happening, it is necessary to look at the rate of increase in a figure like per capita income, and not just at whether there has been an increase.[39]  Kevin Phillips observes that a worker's increased expenses (such as of a woman who is working and incurs child care costs) aren't deducted as offsets to the income.[40]

            .  Figures about "the top 1 percent."  Those who stress the growing economic polarity within the United States often talk about the wealth and earnings of the top one percent.  Rubenstein makes us aware, however, that "the Census Bureau keeps track only of the top 5 percent," so that "nobody knows exactly how much income is needed to be counted among the top 1 percent."[41]  The fact that there are indeterminancies doesn't necessarily invalidate discussions of the top 1%, but a reader needs to know they are there.

            .  Suggestions to improve measures of income.  We have already seen the suggestion that it would be better to take objective measures of well-being -- such as the average number of cars Americans drive, television sets they own, life expectancy, etc. -- than to tip-toe through the minefield of income statistics.  I am reminded, though, of my son, who in his late twenties quite deliberately lives a life with few amenities and low earnings, while enjoying backpacking and playing a musical instrument.  A counting of "objective measures" won't give the picture unless a value is placed on leisure.  How does one do that?  Probably the answer is that no single measure should be relied upon by itself, but rather an array of measures of different kinds.  To this it will be worthwhile to add a recommendation for a fair modicum of intellectual humility, acknowledging that in many ways reality is too complex and transitory to be  measured more than approximately.

            Some authors urge that income averages and medians should be understood largely through the prism of "income mobility."  Thus, Rubenstein about the common practice of describing the population by "income quintiles": "Income mobility is a far better indicator," because non-recurring tragedies or windfalls often put people or families in earning quintiles that are not typical of them.[42]  An individual tends to move from low-income to high-income and then back to low-income again as he moves through the phases of life, with the result that a placement of him in a certain quintile at any given time almost totally misrepresents his economic well-being.

            This suggests that a use of "comparables" might often help, just as it does in real estate appraisal.  How, say, do 25-year-old male college graduates (but with no graduate education) in a given year compare with others just like themselves from twenty years before?  Karl Zinsmeister, accordingly, has argued that it is necessary to "match equivalent households" rather than talk in terms of medians or averages; "you must be sure you are looking at the same kind of family."[43]  This is not to say that "comparables" don't pose problems, especially when values and conditions change over time.

 

National Income Data: Gross National Product (GNP) and Gross Domestic Product (GDP)

            National-income data are of relatively recent origin.  The U.S. Department of Commerce began publishing them in 1942, and Daniel Bell tells us that "the concept of GNP was first broached by Franklin D. Roosevelt in his budget message in 1944."[44]

            These statistical categories involve an adding-together of all monetary transactions.  The advantage is that those are the transactions most easily counted, but Fred Block says anyone is wide of the mark who puts GNP/GDP forward as a way of "trying to track all of the utility that society produces."[45]  In what follows, we will critique national income data as an indicator of well-being, which is the way they are often used.  This use is distinguishable from any more limited purposes to which economists might put them.  We will see that the concept was flawed from the beginning, but it is worth noting that Block says the gap continues to grow "between what the GNP data report and the actual output of the economy."  People are "actually much better off than the GNP data indicate."[46]

            A good place to start is with the oddity that the statistics count only monetary transactions, although there is also some "imputing of income," such as to homeowners for the benefit they receive from their residences.[47]  In the first place, the aggregation of all monetary transactions is inclusive of too much.  Daniel Bell says "it does not discriminate between a genuine addition to welfare and what, in effect, may be a subtraction."  He gives the example of a steel mill.  Its output of steel is counted, but so also is the money that is spent cleaning up a lake if it pollutes one.  Both go into the total because they involve monetary expenditure.[48]  Sir James Goldsmith illustrates the same point by pointing to a hurricane, earthquake or epidemic.[49]  All cause the spending of money, but society will usually be less well off than it was before.  And, Goldsmith says, the value of a product may be counted several times as it passes through middlemen to the ultimate consumer.[50]

            It is at the same time commonplace to point out that there are many value-producing activities that monetary transactions don't include.  Two that are often cited are house-work and home-making.[51] Nobel laureate economist Gary Becker says that "when a family hires someone to care for the children, clean the house, and cook, that work is counted... When a parent does it, it is not."[52]  This is why John Maynard Keynes once observed that "a man who marries his housekeeper diminishes total GNP."[53]  Fred Block points to several other things that are uncounted but that are important to well-being: environmental quality, life expectancy, the value of leisure, nonpecuniary rewards individuals derive from work (or, we might add, from anything else), economic security, and the value of services as distinct from their dollar cost.[54]  There is also an "underground economy" of unreported expenditures, such as in the drug trade and even some activities that are more legitimate in themselves but as to which the parties don't want the transactions known for tax or other purposes.

            Then there are conceptual insufficiencies.  Economist William Allen observes that "from 1979 to 1980, GNP went up almost 9 percent.  That sounds great.  But prices went up close to 10 percent." He warns that "in dealing with GNP, watch out for inflation."  Also, it is well to keep changes in population in mind: "If real output goes up 15 percent over a period while the number of people increases by 20 percent, output per person falls."  Income distribution is also something to consider: "Several oil-producing countries experienced huge increases in GNP during the 1970s, but 95 percent of the residents did not gain."[55]

            Allen further discusses the problem of qualitative change in goods.  This is the phenomenon of "capital savings."  Two computer-purchase transactions involving the expenditure of $2,500 in 1982 and 1997, respectively, would have the same dollar amount so far as being counted in GNP/GDP, but the purchaser received much more for the money in 1997 than in 1982 because of the vast increase in computer power, which made the two computers hardly comparable.  Block says that "if technological advances are making possible significant capital savings, the dollar value of business and government purchases of capital goods will understate the total utility being produced in a given year."[56]

            Ravi Batra mentions the need for reductions.  Tangible durable goods deteriorate constantly (such as when a building or other property depreciates), but this isn't reflected as a reduction in GNP/GDP.  In fact, when money is spent on their replacement or renovation, that adds to GNP.  His point is actually broader than a simple reference to tangibles; he refers to such things as "the homeless millions, urban blight, crumbling roads and bridges, declining test scores, unaffordable homes, desolate factory towns and parking lots."[57]  It is even possible, as Goldsmith notes, for GNP and unemployment to rise simultaneously: "In France over the past twenty years GNP has grown by 80 percent... [while] unemployment has grown from 420,000 to 5.1 million...."[58]

            Wriston raises another conceptual difficulty when he criticizes the selectivity of "imputing" income to homeowners but not doing so to owners of other durable assets such as cars and dishwashers.[59]

            With all these things in mind (and they are just part of what could be mentioned), it is necessary to ask how much real information one receives the next time he reads a report giving national-income statistics. 

 

Unemployment Statistics

            Data for employment and unemployment are frequently reported in the press and are heavily relied upon in economic discussions as indicators of national well-being.  At the annual Federal Reserve symposium at Jackson Hole in 1994, economist John P. Martin said, however, that "there are long-running debates in many countries as to the adequacy of [the] conventional measures of unemployment."[60]

            It is commonly said that the United States opts for flexible wages, and hence for low unemployment, while European countries insist on inflexibly high wages and underwrite them with the welfare state, the result of which is high unemployment and "Eurosclerosis."  So variable is the perception of facts about unemployment, however, that MIT economist Lester Thurow "puts the real unemployment rate [in the United States] at over 14%, or higher than European levels."[61]  Here is the detail Thurow gives in an interview with Technology Review: "The United States nominally has one of the lowest rates of unemployment in the industrialized world because we don't count a lot of people as unemployed."  There are five or six million people, he says, "who say they want to work but who don't meet any of the technical criteria for being unemployed, such as that they visited a potential employer in the previous week.  We've also got about 4 and a half million people who work part time but want to work full time, and we've got 8 million people who are temporary workers.  We have another million people who are self-employed ‘consultants,' many of them professionals affected by downsizing... who would be glad to take a regular job... And then we have, between the ages of 20 and 60, 6 million missing men... they're dropouts."[62]

            Kevin Phillips agrees, pointing to differences in what is included and not included in Europe and the United States.  "In contrast to many European countries, the United States in compiling jobless data excluded persons without employment who had stopped looking for work, while part-time workers who wanted full-time jobs were nevertheless counted as entirely employed."[63]  Aaron Bernstein in Business Week adds that "our current low unemployment rate [in the United States] masks an economy full of workers hungry for better jobs."[64]  It should be noted at the same time that Sir James Goldsmith argues that French statistics have similar problems; they omit, he says, 1.8 million people who should be counted as unemployed.[65]

            Wallace Peterson makes the point that "unemployed" should have added to it the number of people who do have jobs but who are below the "poverty line."   In 1990 this was 14.4 million people.  If they were added to create a truer picture of where the United States is in terms of human well-being, the rate would be 16 percent.[66]  Of course, even though this seems a valid enough point, a political and non-scientific element would be introduced that could greatly affect this statistic: the decision about where the "poverty line" is to be drawn.

            On the other hand, some things are omitted that, if added, would point the other direction, making conditions appear more favorable by boosting employment and lowering unemployment.  John P. Martin mentions "the proportion of the unemployed who are engaged in concealed employment."[67]  William Allen refers to "the bloating of both the labor force and the number of unemployed by those who pretend to want a job but actually will not accept work."[68]  

            Some conceptual matters pertain.  For example, in any given recovery there may be an apparently favorable unemployment rate, but a look at the rate of improvement might show it to be higher or lower than in other recoveries.[69]  Pointing to another factor, Frederick Strobel says that "the economics profession has... by and large accepted uncritically a low unemployment rate as an indicator of economic well-being," to which he adds that "the unemployment rate says very little about the distribution of income between capital and labor, or about the quality of jobs."[70] 

 

The National Debt and the Annual Fiscal Deficit

            We have already seen the statement by former Senator Hank Brown and former governor Richard Lamm to the effect that the national debt figures leave out vast legally-committed-to liabilities.  Similarly, Ross Perot said in 1993 that "the official figure has us $4 trillion in debt, but... the total obligations of the United States government are actually around $15 trillion... We have a total of $15 trillion guaranteed by the United States taxpayers.  These guarantees cover banks, savings and loans, pension funds, home mortgages, and more."[71]

            A related point applies to the annual federal budget deficit (or surplus, if one really emerges).  Michael Boskin, the chairman of the Council of Economic Advisors under President Bush, wrote in 1989 that "a very large amount of government spending occurs off-budget."[72]  He called for revamping the accounting system used by the federal government.  "The fact is that congressmen today are receiving irrelevant and inaccurate information about everything, from the level and growth of federal spending to the nature of costs in the various programs."[73]

            It reinforces the main theme of this discussion to observe that other factors can make the picture look entirely different.  Wallace Peterson tells of the 1985 book on the deficit by Robert Eisner of Northwestern University.  It warned of improper measurement that leads to overstatement.  What is needed is "correcting for inflation, eliminating deposit-insurance transactions from the calculations of the deficit because they don't reflect current income and production, deducting estimated federal outlays for nonmilitary capital, and offsetting the federal deficit by the combined budget surplus of state and local governments."[74]

            Walter Wriston in 1992 criticized what he sees as the aberrational accounting system used in federal government budgeting.  "Since 1969, the U.S. government has used a unified cash-based budget that does not produce results congruent with generally accepted accounting principles."  He points to the failure to include any capital accounts.  Even such a thing as a multi-billion dollar road system is "expensed."  If it were a bookkeeping system used in the private sector, he said, its user might well be taken to court for fraud.[75]

            It is well to remember, as we review these statistical categories, a point we made earlier: that some of the problem comes from the data's being generated and used in a hotly disputed political and ideological context.  Paul Krugman says, for example, that the deficit figure of $123 billion for 1990 "is partly a fraud, thanks to the desperate accounting expedients used to postpone the Gramm-Rudman [deficit reduction] targets; the true deficit was more like $150 billion."[76]  Speaking of the long-term budget projections that feature so prominently in the "budget deals" between the Congress and the President, Rob Norton in Fortune says that they "are almost entirely worthless.  For one thing, they depend crucially on economic forecasts, and economists can't always predict what the economy will do next week, let alone next decade."[77]

 

Balance of Trade Statistics

            Robert Reich describes the almost total bankruptcy of trade statistics as they now stand.  He says they are "notoriously imprecise, subject to wide swings and seemingly inexplicable corrections.  The truth is that these days no one knows exactly, at any given time, whether America's (or any other nation's) international trade is in or out of balance, by how much...."[78]  Wriston agrees: "The current trade accounting system is totally inadequate."  He gives an example in which an American author sends a manuscript to Taiwan, where the book is manufactured and sent to the United States for sale, resulting in royalties to the author.  "So far as the balance-of-trade figures are concerned, Taiwan runs a trade surplus... [even though] the lion's share of the return on this capital is generated in the United States."[79]

            Much international trade occurs among companies and their foreign subsidiaries, contrary to the traditional pattern.  Alfred Balk points to the massive artificialities this introduces: he says that when U.S. corporations sell abroad through their subsidiaries and affiliates there, which is today more and more frequently the way sales are made, "their sales do not count as U.S. exports."  If they were counted, Balk argues, a 1985 trade deficit of $148.5 billion would have been transformed into a trade surplus of $151.4 billion -- a swing of $300 billion!  Cutting the other way, when foreign-owned factories in the United States send goods abroad, those shipments do count as "American exports."  On the import side, if U.S. firms' foreign subsidiaries sell goods in the U.S. market, such sales count as "imports," making up as much as 60 percent of the trade shortfall.[80] 

            A 1994 article in Business Week says government economists believe American exports are undervalued by 10 percent or more.  A major reason is that "computer software, an industry dominated by U.S. producers, barely shows up... The government calculates only the value of instruction manuals and the blank disks... The value of the software itself simply disappears."[81]