"IF ALL ECONOMISTS WERE LAID END TO END, THEY WOULD NOT REACH a conclusion" is a standard witticism about my discipline and one that relates to two made in this paper. First, the witticism itself it simply wrong--a problem that is more characteristic of the field is not excessive disagreement, but its opposite: the propensity of its practitioners to agree too much. One need merely look at the standard economics texts, the subjects they encompass, the tools they utilize to analyze those subjects, and the conclusions they derive to recognize that there are no profound differences in analytic approaches. Of course, economists do disagree, sometimes passionately, primarily because their varied political orientations, which are patently not derived from economic analysis (nor are they claimed to be so derived) do frequently lead them to differ sharply in their recommendations. But that is, evidently, a very different matter from incompatible analysis.
Second, the jest can be interpreted to imply that because economists allegedly so often disagree, at least some of them must frequently be wrong. Of course, I can hardly deny my own errors or those of my colleagues. Indeed, I will offer a number of illustrations in the text that follows. But economists are hardly the only source of economic errors. My discipline is particularly vulnerable to mistaken ideas contributed from the outside.
Unfortunately, unlike fields such as physics, economics is a subject on which even the most ill-informed of individuals are apt to feel themselves qualified to make authoritative pronouncements. After all, everyone participates in the economy in one way or another. Moreover, where widespread misconceptions passionately held are the result, democratic government can be forced to act in accord with them. Politicians, too, frequently just assume that they understand the workings of complex but common economic phenomena, and have been content to proceed to their conclusions with little evidence or analysis.
In this paper, I will deal with a variety of illustrative errors, focusing among others on two that are critical for policy: the notion that a budget deficit constitutes a burden for our grandchildren, and the idea that rising costs of health care and education mean that society will be increasingly incapable of financing them and that cutbacks in both of these vital services are therefore unavoidable.
ECONOMIC ERRORS THAT DAMAGE THE INDIVIDUAL
Sometimes it is the individual committing an economic error who alone bears the cost. An example is the investor who seeks out and follows financial analysts' advice on the purchase and sale of stocks, despite overwhelming statistical evidence demonstrating that, even if such advice were offered without cost, it would generally be valueless or worse. Professional recommendations on stock market purchases and sales have repeatedly been shown to be totally unreliable. Indeed, they must be so because, as the data demonstrate, the behavior of securities prices approximate what statisticians call a "random walk." Random behavior is, by definition, inherently unpredictable even by the best-informed and most intelligent analyst. But stock market analysts' advice is even worse than this for the investor, on two scores. First, it is not costless. The investor is forced to pay for bad information and is thereby put in the position of a bettor in a gambling casino, where the outcomes are not just random but are systematically biased to bring a predictable rake-off to the gambling house. Second, whether or not as a deliberate dereliction of duty, frequent sales and purchases of securities that benefit the stock market analysts' own firms are characteristic of their recommendations. These transactions multiply the investor's total payments to these firms and, incidentally, materially increase the investor's tax bill.
DAMAGE TO THE SOCIETY: THE CASE OF MISTAKEN COUNTERCYCLICAL POLICY
As the next illustration will show, economic errors can burden many beside the individuals who make the mistake, and sometimes even society as a whole. A notable example was the belief that an essential step in extracting an economy from recession or depression is elimination of deficit spending by the government. While there is no longer agreement by economists that expansion of such spending is invariably a sensible step, it is recognized that the simpleminded argument that leads many nonspecialists to conclude that such deficit spending threatens to bankrupt the nation is an exercise in the "fallacy of composition." This fallacy is the presumption that a relationship that is valid for each individual must automatically be valid for the entire group of these persons. One elementary example entails voluntary exchange between two informed and rational individuals and the conclusion that such an exchange must offer some benefit to each of them, or at least no loss to either (otherwise, the prospective trade participant who stood to lose from the transaction would simply refuse to trade). The fallacy of composition enters when this insight about trade between individuals is applied to trade between two countries, where it is neither self-evident nor generally true that exchanges must invariably benefit both countries.
Turning to the issue of deficit spending, the standard view stems from the observation that an individual who is in financial difficulty because of persistent spending beyond his means must somehow succeed in curtailing his overspending now and in the future if he is to avoid exacerbation of his financial troubles. The inference from this observation drawn by analogy for a depression-beleaguered government--whose tax revenues have fallen as a consequence of reduced incomes and whose expenditure has been driven upward by rising obligations--was that, just as in the individual case, fiscal retrenchment was essential. Governments in that position characteristically find themselves plagued by rising debt and the evident, if questionable, conclusion was that material retrenchment was urgent and unavoidable if financial catastrophe for the country was to be avoided.
But, one of the central propositions to emerge in the course of the Keynesian revolution was that this prescription for retrenchment was the precise opposite of what such a situation requires. Rather, an effective governmental weapon--indeed, a critical component of the counter-depression policy that Abba Lerner dubbed "functional finance"--is enhancement of deficit spending, entailing rising public debt, with the shortfall to be made up during the other end of the business fluctuation, when inflation replaces unemployment as the primary threat to the economy.
The way in which the fallacy of composition enters the matter is quite straightforward. Increased spending by an individual (without any offsetting rise in earnings) spells financial peril. For the community of individuals, taken as a group, the situation is, at least in the simplest Keynesian view, usually the reverse of this. The more the government increases spending without a corresponding rise in tax revenues, the better off the community will be economically. This act of magic occurs because the very deficit spending must put purchasing power into the hands of the public, which in turn will serve to raise demand for goods and services. And in a depressed economy, anything that serves to offset lagging demand must be helpful, because it will expand sales, elicit enhanced production, and provide additional jobs. So deficit spending by government is a stimulus of economic activity and a source of added income for the society as a whole. This stimulus effect also helps to cut the government's budget shortfall by automatically adding to total tax revenues as private incomes rise, and by cutting needed government expenditures, such as outlays for support of the unemployed. As Keynes himself pointed out, the apt parable is that of the legendary widow's cruse, which kept refilling itself as its contents were extracted. For, if the argument is valid, it indicates that the more the government overspends, the more net income it can hope to have available in the near future.
This argument, though not universally accepted by economists today, was certainly rejected by many, including President Franklin D. Roosevelt, in the 1930s. It is at least arguable that the resulting efforts to curb government overspending protracted the Great Depression, creating a second economic decline toward the end of the decade, with termination of the Depression left to the onset of the Second World War, which once again imposed substantial deficit spending on the government. If it is true that insufficient government spending exacerbated the effects of the Depression, then it is surely difficult to dispute the conclusion that here was an economic error that caused great and widespread harm, increasing unemployment, reducing incomes, and keeping output and accumulated wealth of the society down well below what it might otherwise have been.
There is an associated popular misunderstanding, which strengthened the determination of the opponents of deficit spending. This is the conclusion that government deficit must constitute a "burden upon our grandchildren." There are, it must be admitted, circumstances in which this could be true, and one must not go so far as to deny the possibility of any detrimental consequences of government debt for future members of the community. But the common and assuredly naive variant of the idea is yet another example of the fallacy of composition. That assets lost by injudicious expenditure during an individual's lifetime can impoverish her heirs is evident. But for a nation, matters are far different. Thus suppose, for example, that a government greatly increases its current expenditure on military equipment, financing it by borrowing, through the issue and sale of additional government bonds. The labor, steel, power, and other inputs that are used to manufacture the armaments immediately become unavailable for civilian use. This is a burden that fails upon the public at once and need not in any way affect future generations whose supply of factors of production need not thereby be diminished. The labor that today is shifted from production of autos to the manufacture of tanks does not reduce the availability of labor to consumers 20 years hence. Reduced resource availability that results from government deficit spending, then, is primarily a burden upon the current generation, not those of the future.
It is not even true that government debt incurred today need entail a financial problem tomorrow, when the debt is to be repaid. But from what source is the repayment to be made? Suppose, for concreteness, that the government bonds that financed the debt are scheduled for redemption 20 years after the deficit spending occurred, and that at that date the government raises taxes by an amount just sufficient to cover the X-dollar debt. Then that is surely a burden for those who must pay the X dollars in taxes, but it is accompanied by a rise of exactly X dollars in the cash that becomes available to the bondholders. If the bonds are not held by foreigners, what will happen at the date of repayment is that the money that financed the purchases will simply have been transferred from one group of citizens to another. Indeed, even that need not take place to any marked extent. If, for example, the government bonds are held by individuals roughly in accord with their incomes, the wealthier the individual, the greater his holdings, then if the tax is also proportioned to income, the repayment process need not incur any significant transfer of purchasing power. The money will be taken from the wealthy, and promptly returned to the same individuals. In the words of Adam Smith, what will have been entailed is simply a transfer of money from the right pocket of the taxpayer to the left.
In short, viewed in terms of its substance, the burden of government expenditure is a burden upon the present, not upon the future. Yet this was apparently not understood by earlier generations of economists and certainly not by the general public. And the error was not just a matter of academic interest. Rather, by preventing the actions that promised a speedy recovery from recession or depression, it had marked and unfortunate consequences for the general welfare.
CAN SOCIETY AFFORD IMPROVED HEALTH CARE AND EDUCATION?
I turn now to a second example, derived from work with which I have been associated. (1) The issue, which has become a matter of persistent and heated political contention throughout the industrialized world, pertains to the cost of a variety of services in which the public sector plays a substantial role and often is their primary provider. The most significant of the services at issue--clearly matters of great importance for the general welfare--are health care and education. The costs of these two services have been rising throughout the industrial world, and have done so persistently, cumulatively, and at rates substantially outstripping the rates of overall inflation in those economies (that is, their cost increases have outpaced the average rate of increase of the entire set of outputs of those economies). This immediately raises two issues. First, how can this disturbing cost performance be explained? Second, does it mean that society is condemned to experience progressive deterioration in the quality and quantities of these services?
Statistics gathered by the US government document the startlingly persistent and dramatic rise in the cost of these personal services. The public is all too painfully aware of the speed at which general medical costs in this country have been increasing. The cost of a hospital stay has been going up even more rapidly. These cost increases have been replicated in other countries, as will be shown presently, and have made health care a prime subject of debate in political contests, not only in the United States but in virtually every industrialized country.
Consider these facts: (2) between 1948 and 2001, the Consumer Price Index (the CPI, a standard official measure of overall price rises in the economy) increased at an average rate of about 3.8 percent per year, whereas the price of physicians' services rose 5.3 percent per year. This difference seems tiny, but compounded over those 53 years it had the effect of increasing the price of a doctor visit more than 125 percent, measured in dollars of constant purchasing power. This history is noteworthy also because of the length of time, about half a century, over which the phenomenon has endured, and the absence of any protracted exceptional periods. During this same time period, the increase in the cost of hospital stays was even more extreme, with the average price per hospital room skyrocketing at an annual rate of almost 8 percent compounded--which amounts to more than an 800 percent increase since 1948, in constant dollars.
The explanation is not obvious. Doctors' earnings barely kept up with the economy's overall inflation rate during this period. Moreover, despite a great variety of government programs with many different approaches (including different forms of price ceilings, inhibition of doctor incomes, and special medical purchase arrangements), virtually every major industrial nation has failed to prevent health-care costs from rising faster than its economy's rate of inflation. The average yearly rate of increase in real (inflation-adjusted) health-care costs per person over the period and in the countries for which the figures are reported confirm that the phenomenon remains nearly universal among industrial nations, despite the variety of countermeasures they have adopted.
The cost of education has a similar record. For example, cost per pupil in the United States has increased an average of 7 percent per year in the past 50 years. And during the approximately 30-year period between 1965 and 1994, US increases in education costs have been lower than education cost increases for four of the six top industrial countries.
What accounts for the ever-increasing hospital and schooling costs? No doubt a number of contributing factors can be cited. But there is a particularly significant reason--one that cannot be avoided by any hospital or school administration no matter how pure its motives and efficient its operations and that must play a primary role in the persistence and universality of the phenomenon. This common influence has been called the cost disease of the personal services.
Continued from page 2.
This cost disease stems from the basic nature of many personal services. Most such services, notably health-care services and schooling, have indispensable handicraft attributes: they require direct personal contact between those who provide the service and those who consume it. Doctors and nurses engage in activities that require direct, person-to-person contact. Moreover, the quality of the service deteriorates if less time is provided to the patients by the medical personnel. In contrast, in other parts of the economy such as manufacturing, no direct personal contact between the consumer and the producer is required. A labor-saving innovation in auto production need not imply a reduction in product quality. As a result, over the years it has proved far easier for technological change to save labor in manufacturing than to save labor in providing services. Labor productivity (output per worker) in US manufacturing and agriculture has increased at an average rate of something like 2 percent a year since World War II, but the productivity of college teaching (crudely measured by number of students taught per teacher) has increased at a rate of only 1 percent per year during that period. And, in elementary and secondary education, labor saving has actually been negative: the average number of pupils per teacher has fallen about 27 percent during this time, with unavoidable consequences for relative costs. Outputs whose labor productivity (that is, output per work hour) increases at a rate below average can, for obvious reasons, be expected to experience a rate of cost increase that is above the average. But the average rise in cost for the economy as a whole is just a synonym for the economy's rate of inflation. Thus, it is virtually tautology to assert that the costs of items whose productivity growth is well below average can confidently be expected to grow at a rate exceeding the economy's rate of inflation. The fault lies in the technology of the processes of education and health care, not in the inefficiency or greed of its suppliers. (3)
Over a period of several decades, this excess in the growth rate of the costs of the handicraft services adds up, making such services enormously expensive relative to other outputs in the economy. So much so that just the two service sectors we have been considering, health care and education (among the thousands of different outputs of the economy) by themselves now constitute more than 20 percent of total US GDP, and are still rising. (4) No wonder these two services and a few others like them are subject to intense political debate through the industrialized countries.
PUTTING PRODUCTIVITY GROWTH DIFFERENCES TO WORK FOR SOCIETY
Finally, we come to the key misunderstanding, in terms of policy, engendered by the failure to understand the nature of the cost disease: the idea that the cost disease will force society (or the government that provides the finances) to retrench and eventually cut back on vital health care and education because of the mistaken belief that their rising cost must make them increasingly unaffordable to society. This belief, it turns out surprisingly, is virtually the reverse of the truth.
In actuality, the very forces that create the cost disease make these services ever more affordable to society. This is so because the source of the problem is that, although productivity is growing in almost every industry, in some industries (particularly personal services) it is growing more slowly than in others. But if output per worker and output per work hour are rising in virtually all industries, then a given quantity of any bundle of outputs requires an ever smaller share of the labor force for its production. What society must do is use part of the cost savings from the industries (like manufacturing or telecommunications) in which productivity is growing at a rapid rate to pay for the personal services (like health care and education) in which productivity is growing at a relatively slower rate. It is simply not true that society cannot afford those costs. On the contrary, rising productivity means that society can afford to consume more of each and every product. It is this observation that led the late Senator Daniel Patrick Moynihan to describe the cost disease analysis as a profoundly optimistic diagnosis.
The danger is that governments, the primary source of financing of these services in most countries, will decide that the cost burden is beyond their capacity to finance, and will decide that cutbacks are their only option. This is already happening in many of the industrialized countries, where an increasing set of medical procedures are denied to patients and cutbacks in financing of universities and their teaching and research activities are all too common. This is unfortunate because, as is shown by the cost disease analysis, their unaffordability is a fiscal illusion, and retrenchment of these arguably vital activities is an unnecessary if understandable response to this illusion. Here, surely, is a case in which misunderstanding can result in totally avoidable damage to the social interest.
CAN PRICE INCREASES EVER SERVE THE PUBLIC INTEREST?
The possibility that, in a wide variety of circumstances, a rise in price may be a substantial benefit to the public is something that people untrained in economics always find extremely difficult to accept. Indeed, the assertion that low prices may not always serve the public interest seems perverse. Yet, the reason can quite readily be explained. If a price, such as the price of crossing a crowded bridge or the price of environmentally damaging gasoline, is set very low, then consumers will be provided an incentive to exacerbate the problems. These market signals will induce them to add to the crowding or to the environmental damage even further. A striking historical illustration brings out this point. In 1834, a University of Dublin professor of economics named Mountifort Longfield lectured about the price system. He offered the following example:
Suppose the crop of potatoes in Ireland was to fall short
in some year one-sixth of the usual consumption. If ]there
were no] increase of price, the whole ... supply of the year
would be exhausted in ten months, and for the remaining
two months a scene of misery and famine beyond description would
ensue.... But when prices [increase] the sufferers [often believe]
that it is not caused by scarcity.... They suppose that there are
provisions enough, but that the distress is caused by the
insatiable rapacity of the possessors ... [and] they have generally
succeeded in obtaining laws against [the price increases] ... which
alone can prevent the provisions from being entirely consumed long
before a new supply can be obtained (Long-field, 1834: 53-6).
It is intriguing that this talk was given some 10 years before the great potato famine, which caused unspeakable misery, death by starvation, and brought emigrants from Ireland to the United States. The story of the actual potato famine in Ireland is, of course, much more complex than Longfield's discussion indicates. Still, the implications of his lecture about the way the price system works are entirely valid.
It is useful to rephrase Longfield's reasoning. The crop failure brought drastic scarcity of potatoes. If society was to minimize the harsh consequences, the potato supply had to be stretched to last until the next crop arrived. This required the country to cut back on the consumption of potatoes during earlier months--which is just what rising prices will do automatically if free-market mechanisms are allowed to work. However, if the price is held artificially low, consumers will use the very scarce resources inefficiently. In this case, the inefficiency shows up in the form of famine and suffering when people deplete this year's crop months before the next one is harvested.
There are many other examples, less dramatic but equally relevant. One telling illustration is the way that landing privileges at crowded airports are often priced. Airports become particularly congested at peak hours, just before 9 a.m. and just after 5 p.m. This is when passengers most often suffer long delays. But many airports continue to charge bargain landing fees throughout the day, even at those crowded hours. That makes it attractive for small corporate jets or other planes carrying only a few passengers to arrive and take off at those hours, worsening the delays. Higher fees for peak-hour landings can discourage such overuse, but they are politically unpopular, and many airports are run by local governments. So we continue to experience late arrivals as a normal feature of air travel.
We know that inappropriately low prices caused nationwide chaos in gasoline distribution after the sudden drop in Iranian oil exports in 1979. In times of war, constraints on prices have even contributed to the surrender of cities under military siege, deterring those who would otherwise have risked smuggling food supplies through enemy lines. Low prices have also discouraged housing construction in cities where rent controls made building a losing proposition. Of course, in some cases it is appropriate to resist price increases--as when unrestrained monopoly would otherwise succeed in gouging the public, and when rising prices fall so heavily on poor people that rationing becomes the more acceptable option. But before tampering with the market mechanism, we must carefully evaluate the potentially serious and even tragic consequences that artificial restrictions on prices can produce, particularly when scarcity threatens or is already damaging the public welfare.
It is not easy to accept the notion that higher prices can serve the public interest better than lower ones. Politicians who voice this view imperil their jobs. Because advocacy of higher prices courts political disaster, the political system often rejects the market solution that automatically raises prices when resources suddenly become scarce. And that only enhances the shortages.
MUST OUTSOURCING TO OTHER NATIONS ALWAYS BENEFIT BOTH AFFECTED COUNTRIES?
I come to my last illustration, this time as a misunderstanding widely current among economists, and one on which the unspecialized general public seems to have arrived at a more defensible conclusion than many of the professionals. Not without reason, economists are usually strongly predisposed to favor free trade, globalization, and market-driven apportionment of industries among nations. But this orientation has led many of them to conclude that when a portion of an economic activity or even an entire industry moves from a high-wage to a low-wage country as a result of an increase of productivity in the latter, both the gainer and the loser of the industry can be expected to benefit. In particular, while some individuals in the country from which the activity has emigrated will evidently be harmed, on this view the country as a whole will normally benefit from the reduced costs of the products whose production has moved abroad, and benefit sufficiently to compensate for the damages and more.
Here, a colleague and I have been driven to disagree with many other economists and have shown that, in what we believe to be a large range of cases, the country that loses the activity can be expected as a result to suffer a decrease, possibly substantial, in its overall per-capita income, that is, in its standard of living, and this damage need not just be a transitory loss (see Gomory and Baumol, 2000 and forthcoming).
Those who believe that macroeconomic policy can effectively limit involuntary unemployment have reason to conclude that loss in the total number of jobs is not an inevitable consequence of globalization, though it does undoubtedly threaten the working positions of at least a few directly affected individuals, for whom the consequences must not be taken lightly. But though we may reject the popular view that globalization is a major threat to employment and an instrument of extensive job loss, we cannot deny that there is reason to be concerned with at least the short-term effects on wages in both developing and developed lands. International competition can influence relative input prices and thereby determine whether machinery will be substituted for labor, for example, or whether skilled labor will be substituted for unskilled. There are, also, more direct implications for wages. Surely, the increased use of computer programmers in India can be expected to reduce the demand for such skills in the United States below what it might otherwise have been.
For the developing countries, economic history suggests that an industrial revolution initially tends to depress real wages and real living standards, thus supporting the concerns of those who fear the consequences of globalization for the world's less prosperous nations. Though the British industrial revolution is usually considered to have taken off about 1760, it was probably not until approximately 1840 that wages began to rise. Data on life expectancy and average height also indicate that the spread of innovation was accompanied by worsening of the economic status of wage earners, perhaps in part as a result of the move from the countryside to crowded, unsanitary slums; the evidence indicates that the US labor force underwent a parallel trajectory. One may surmise that part of the explanation was a rise in the power of employers and an inability of the workers, in the absence of labor organizations, to resist.
The opponents of globalization draw attention to a similar phenomenon in twenty-first-century globalization, with multinational employers subjecting their employees to disturbingly low wages and shocking working conditions, particularly on the criteria widely accepted in the affluent economies (though by no means always adhered to even there). Thus, even if globalization is a very promising influence for the more distant future prospects of the developing countries, there is good reason to fear that in the short run the workers in those lands may gain little and may even lose out in the initial stages of globalization.
It can be argued that all this is transitory and that in the long run the lower-income groups in the developing countries will be better off, as has indeed been true in the developed economies. But the process can easily take decades. We cannot just ignore decades of very substandard earnings that amount to preservation of grinding poverty in a developing country or the permanent structural unemployment in a developed economy that can beset older workers whose skills are made redundant by innovation, and for whom the acquisition of new skills is not a practical option. These are hardships that constitute an extremely painful economic pathology for the affected individuals. At the very least, one can argue that those who stand to benefit from the process should be expected to agree to provide systematic and substantial assistance to the victims, presumably through government channels, and supported liberally by the wealthier communities. If that is not acceptable politically, there is surely little that can be said convincingly in support of a contention that the suffering of the victims will be justified by the promised future benefits to their descendants.
POSSIBLE LONGER-RUN DAMAGES FROM GLOBALIZATION
Though it can be hoped that, in the longer run, globalization will help to reduce (and even eliminate) poverty in the developing countries, as I have stated earlier (and contrary to widely held views), globalization can also permanently damage economic well-being in some of the affected countries, notably those countries that are now in the economic vanguard. This possibility may seem surprising, and even paradoxical. The simple explanation is that competition and mobility of products tend to equalize wages, raising those that are low, but also reducing those that are especially high.
To confirm this conclusion, suppose, for simplicity, that we are dealing with a two-country, two-good world, with one economy developed and at the technological frontier, while the other country, with similar population size and resources, is far behind in terms of productivity, technical competence, and per-capita income, particularly in the production of one of the traded commodities, thus giving the technological laggard a comparative advantage in supply of the other commodity. As any elementary economics textbook will tell you, this represents an opportunity for mutual gains from trade, with each country specializing in supply of the good in which it has a comparative advantage. Assume also, for easier exposition, that there is no further technical progress in the wealthier country, while in the other economy there is catch-up, made possible by globalization of technology, with equipment and ability approaching those in the wealthier country. Before the technical change, both countries could expect gains from trading, that is, they were both better off as a result of exchange than they would have been if each had produced for itself the quantities of the two commodities its inhabitants consumed. But if the technical progress in the laggard country favors the good in which it was most incompetent before, it can catch up in its productivity of that item sufficiently so that the comparative advantage of the other country in supply of that good can disappear. With the laggard country now equally far behind in the productivity of both goods, the gains from trade that both economies formerly enjoyed will also vanish. The poorer country may still gain from the process thanks to the technical progress it has experienced via learning from the industrialized economy. But the wealthier country in the interim will be worse off than it would otherwise have been. It will then be harmed, perhaps markedly and enduringly, by the technical progress in the other country. In short, this scenario shows the possible dangers to wealthier nations that arise from globalization. It does not mean that globalization is inherently undesirable, or that wealthier nations should never make sacrifices to help impoverished societies. But it does mean that we should not proceed under the illusion that we will assuredly profit from the process. (5)
WHAT WE HAVE LEARNED AND WHAT WE NEED TO LEARN
The approach taken here is quite different from most current discussions of globalization. Discussions of the "offshoring" of jobs especially tend to be heated and fragmentary and to focus on the outcome for workers in the industries most affected by globalization. The overall effect on a country is harder to determine and quantify. The approach just described has been used to investigate that overall effect, and has told us that the net consequence for a country of improved productivity abroad is neither always beneficial nor always deleterious but can be either, depending on the circumstances characterized above. To this result there is also something contributed by the more popular discussions that characteristically focus on the possibility that displaced workers will not find new jobs or will only obtain jobs that provide a lower wage than before, because these workers can no longer use the accumulated skills and know-how of a lifetime.
Of course, that is not the end of the story, at least in the long run. For there is a powerful countervailing force that works to produce long-run benefits to all countries affected by globalization. The power of international competition has arguably contributed much of the unparalleled, sustained economic growth and the unrivalled explosion of innovation that the free-market economies have experienced in the past two centuries. In the countries that have participated in this process, the economic benefits are so spectacular that they could hardly have been imagined by our ancestors. Globalization can extend this process to other nations and can strengthen such developments substantially even in the world's leading economies. Although difficult to quantify, this may well be the greatest economic promise of globalization.
ANYONE CAN ERR
If the arguments of this paper are not themselves in error, what I have shown is that the economics profession can, indeed, sometimes show the layperson the error of his or her more common-sense thoughts. But not always. Sometimes the errors and the route toward correction go the other way. This observation is not meant in any way to denigrate the work of my colleagues. After all, it is only through careful analysis that one can discover where it is the specialist who has been wrong and where the often exceedingly fallible common sense of those with no formal training in the field has turned out to be closer to the underlying reality. We have also seen that misunderstanding in the field of economics can have consequences beyond pushing researchers and teachers in misguided directions. Perhaps as much as any discipline, erroneous economic analysis and conclusions can elicit policies severely damaging to the public interest. And, in this, I believe that we economists do have something to answer for. We are all too prone to put more faith in the implications derived from our quite appropriately simplified models, and to draw from those implications policies that really only apply universally in the artificial world of the constructed model. The recommendation to ourselves that seems appropriate here is advocacy of somewhat enhanced modesty when we do offer advice, and more ready willingness to remind our listeners that, though we are offering the best advice we are in a position to provide, they must recognize that the recommended course may yet prove dangerous to the public health.
APPENDIX: POSSIBLE LONGER-RUN DAMAGES FROM GLOBALIZATION: A DETAILED ANALYSIS (6)
The purpose of this appendix is to show more extensively the way in which long-run damage to some countries from globalization and outsourcing may occur. I therefore, turn to a model that permits an analysis of the issue in more universal terms. That issue can be described as the determination of the effect on the welfare of one country when an industry leaves that country and that industry's production is taken over by its trading partner. It will be seen that there are indeed some circumstances in which the transfer is mutually beneficial, but there is also a large range of cases in which the acquisition benefits the recipient economy but damages the other. I will also indicate the circumstances that lead to the one conclusion or the other. Specifically, it will be seen that when the country that is the recipient of the industry is initially sufficiently poor relative to the other country, the industry shift will yield mutual gains, so the pertinent relative income range is called the "zone of mutual gains." But when the two countries' incomes are closer together, the shift will benefit the recipient and harm the loser of the industry, making the pertinent income range a "zone of conflict."
For simplification, it will be assumed that the world consists of two countries trading in n commodities and that each commodity is produced under conditions entailing scale economies (it must be emphasized that the analysis, albeit somewhat more complicated, applies essentially unchanged to cases of technical progress in industries from which scale economies are absent). While scale economies normally complicate analysis in economic theory, here it is a simplification because any assignment in which Country 1 is the exclusive producer of any m of the n traded commodities and Country 2 is the exclusive producer of the remaining n - m items becomes a (locally) stable equilibrium. That is so because if, say, Country 1 is the exclusive producer of good X, then if Country 2 were to attempt small-scale entry into X production the resulting absence of scale economy will prevent it from competing successfully. Hence, since each and every such specialized assignment of the n industries between the two countries is a stable equilibrium, in this model there will be a vast number of locally stable equilibria, a number that increases rapidly as n, the number of traded goods, expands. (7) The scale-economies premise will allow an examination of the range of distributions of industries and the implications of each possible distribution for welfare in the two countries.
For each distribution of industries and its corresponding equilibrium there will be a determinate level of national income, call it [y.sub.1] and [y.sub.2], for countries 1 and 2 respectively. Country 1's share of world income can be readily calculated from the trivial formula,
(1) Z = [y.sub.1]/([y.sub.1] + [y.sub.2]).
For each equilibrium, as a further simplification here, the welfare of country j will be measured as its national income, [y.sub.j]. Because it can be shown that Z, the share of world income received by Country 1, increases monotonically if it adds to the list of the n commodities of which it is the exclusive producer, Z will be used as the indicator of Country 1's share of industries, with 1 - Z obviously representing this magnitude for Country 2. Finally, at any equilibrium total world income in this two-country world can be calculated,
(2) [y.sub.w] = ([y.sub.1] + [y.sub.2]).
Proceeding with the aid of a graph, each possible equilibrium is represented by three dots, one for Country 1, the second for Country 2 and the third for the world. For example, the Country 1 dot in the graph for some particular equilibrium will show on the vertical axis, [y.sub.1], the income that this equilibrium yields to that country, and on the horizontal axis it will show Z, the share of world income that accrues to Country I in that equilibrium. The graph will show all of these equilibrium points as well as the upper boundaries of the region they fill for the world and for each of the countries, respectively. That will also enable us to see how a transfer of industries that increases one country's share at the expense of the other (that is, a change in Z) affects attainable world income and those of the two countries.
We start off with a graph of world income as a function of Z, which, we will argue intuitively, is hill-shaped. This means that attainable world income is not at its highest when either country has succeeded in capturing for itself the bulk of the world's industries (the right-hand or left-hand ends of the graph where the Country 1 share index, Z, is close either to zero or unity). Rather, the attainable world income will be relatively high toward the middle of the graph, where each country has a substantial share of the trading industries. While this has not been found to be subject to what may be deemed "rigorous proof," it will be seen to be highly plausible. But that hill shape is all that it will be necessary to assume in order to derive conclusions and complete the analysis.
Figure 1 is the graph of the upper boundary of world income, with Z shown on the horizontal axis and [y.sub.w] on the vertical axis. The boundary need not be symmetrical, but as shown and already stated, it is assumed to be hill-shaped. There are two primary reasons that make this shape plausible. First, at either the right-hand or left-hand end of the graph, one of the countries contains almost all of the industries. Hence, its labor force will be fragmented among many products, producing them in small quantities and forgoing their scale-economies advantage. Second, the country that has the bulk of the industries at such a point will be producing many products for which it has no "natural" absolute advantage, such as climate or culture, so that some inefficiency must result. Only where the value of Z is far from either zero or unity are both these sources of output loss weakened, and hence the maximum point of the world income frontier will lie somewhere toward the center of the graph, as shown.
[FIGURE 1 OMITTED]
The shape of the upper income frontiers for the two countries from that of the world frontier can be immediately deduced. This follows from (1), and (2) which gives:
(3) [y.sub.1] = [Zy.sub.w] and [y.sub.2] = (1 - Z)[y.sub.w]
The first of these tells us at once that at Z = 0 the upper production frontier for Country 1 will equal zero and that as Z increases toward unity that country's frontier will approach the world frontier asymptotically. The frontier for Country 2 will be a distorted mirror image of that of Country 1, now moving from zero at Z = 1 toward the world frontier as Z approaches zero (see figure 2).
[FIGURE 2 OMITTED]
There is one more step that is necessary to complete the story.
We have:
Proposition: Let [Z.sub.w], [Z.sub.1], and [Z.sub.2], respectively be the value of Z at the maximum point of the world frontier, the Country 1 frontier and the Country 2 frontier. Then, if the world frontier is horizontal at its maximum and the maxima of the frontiers of the two countries are unique and differentiable near Zw, one must have [Z.sub.1] > [Z.sub.w] > [Z.sub.2].
Proof: The derivatives with respect to Z of the two relationships in (3) are
(7) [y.sub.1]' = [y.sub.w] + [Zy.sub.w]' and [y.sub.2]' = - [y.sub.w] + (1 - Z) [y.sub.w]'
and since at the maximum of the world frontier [y.sub.w]' = 0, it follows that at Z = [Z.sub.w]
(8) [y.sub.1]' > 0, [y.sub.2]' < 0
so that the Country 1 frontier rises toward the right of [Z.sub.w], the maximum point of the world frontier, while that of Country 2 rises as one moves toward the left. If the Country 1 and Country 2 frontiers were known to be convex, (8) as they appear to be in the figure, it would immediately follow that [Z.sub.] > [Z.sub.w] > [Z.sub.2].
It follows that the income share of Country 1 at which its own absolute income is maximal must be greater than that at which the absolute income of Country 2 is maximized. In other words, the Country 1 peak, [m.sub.1] of its upper income frontier, must always lie to the right of the Country 2 peak. Two corollaries also follow:
Corollary 1. At [Z.sub.1], the maximizing Z for Country 1, the largest absolute income available to Country 2 will be below (and, very plausibly, substantially below) the Country 2 maximum, [m.sub.2] (figure 2). The relative positions of the two countries will be reversed at [Z.sub.2].
Corollary 2. Between Z = 0 and Z = [Z.sub.2], both country frontiers can be expected to be upward sloping. Between Z = [Z.sub.1] and Z = 1, both frontiers will normally be downward sloping. In the region between [Z.sub.2] and [Z.sub.1], the Country 1 frontier can be expected to have a positive slope while that of Country 2 will be negative.
The two corollaries lead immediately to the pertinent economic interpretation of our result. First, it can be seen from corollary 1 that the point on the frontier of one of the countries that permits maximization of either country's income will condemn the other country to an income below, and plausibly well below, that other country's maximum. In other words, the trade process can inherently entail conflict of interests, with each country able to achieve its maximum only at the expense of the other.
Second, near Z = 1 Country 1 will have acquired too many of the world's industries for its own good. Because both frontiers have negative slopes in this neighborhood of the graph, a leftward move (that is, a reduction in Z), will make them both better off. Thus, if Country 1 is relatively very rich and Country 2 very poor, both can gain by a more equitable sharing of the world's industries. A similar situation holds near Z = 0, only this time with Country 2 having co-opted too large a share of the n industries for its own welfare. Because near Z = 0 or Z = 1 both countries can benefit simultaneously by a reallocation of industries in which some industries move from the richer to the poorer economy, these two outer regions are called the "zones of mutual gains," as noted earlier (see figure 3). In contrast, in the central region between [Z.sub.1] and [Z.sub.2], the slopes of the two frontiers are opposite in sign, so that any move that benefits one of the countries must be detrimental to the other. Any change in Z that brings a higher level of income to one of the countries must reduce that of the other. This zone of conflict is the region in which neither county is exceedingly poor, but one may well be considerably less affluent than the other.
The interpretation should be clear, and relates to the issue of the ubiquity of the long-run benefits of globalization. Thus, far from invariably leaving all economies better off, globalization can lead to increased prosperity in the less-affluent country at the permanent expense of the country that remains, or at least formerly was, in the vanguard. For globalization and the resulting transfer of intellectual property leading to improved products and processes can enable the economy that is somewhat behind, say Country 2, to increase its income share. But if its initial income share, [Z.sub.2], was located inside the zone of conflict, that must plainly be harmful to its trading partner, Country 1, and there is nothing inherent in the behavior of the model to undo the damage. The most direct implication is that US labor unions may well be right in their concerns about globalization and their resistance to policy measures that facilitate it.
NOTES
(1.) See Baumol and Bowen (1966) for the earliest reference.
(2.) The following figures were derived from data provided by the US Department of Labor (2004) and U.S. Department of Education (2004).
Continued from page 6.
(3.) Economists generally agree that in productivity growth calculations it is appropriate to take account of any rise in the quality of the output. But it does not follow that a productivity growth index that takes no account of quality change is irrelevant and invalid. Rather, the proper choice between a quality-adjusted productivity index and one that is unadjusted depends on the purpose for which it is to be used. If the purpose is to study how the change in productivity affects what consumers are getting for their money, then the appropriate index is quality adjusted. On the other hand, if budgeting is the issue, it is the unadjusted figure that, arguably, gives the right answer. For instance, if city X has a 2 million annual attendance at its public schools, and last year the cost per pupil was $9,000, then a 10 percent growth in unadjusted productivity (brought about, presumably, by an increase in average class size) can be expected to bring that cost down by $900, regardless whether the quality of the teaching has or has not increased and resulted in an improvement in student grades on standard tests. This logic also explains why the cost of health-care services continues to rise so quickly despite the extraordinary improvements in medical technology.
(4.) A few economists have recently claimed to have shown that the cost disease has been cured or is "in remission." To see that this is far from the reality, one need just ask anyone supporting a child in college or paying hospital bills. These writers assert that the services as a group have no cost-disease problem because of the falling costs of high-tech services such as computation. My coauthor, William G. Bowen, and I recognized three decades ago, using telecommunications as an example, that such declining-cost services are not subject to the cost disease. Nothing has changed since then, except that computer services have become more important and pulled up the average productivity performance of the services as a whole.
(5.) An analysis confirming more fully the contention that globalization and outsourcing from a wealthier economy to one that is developing can inflict overall reduction of per capita income to the latter, even in the long run, is more technical that the material in the remainder of this article. It has therefore been relegated to an appendix.
(6.) The analysis that follows is based on research by Ralph Gomory and myself (see Gomory and Baumol, forthcoming). The analysis also appeared in our book (Gomory and Baumol, 2000). The conclusions of this section were independently arrived at by Johnson and Stafford (1993) and an analysis by Paul Samuelson along parallel lines is about to appear (Samuelson, 2004).
(7.) Specifically, if one excludes the possibility of any country producing nothing at all, the number of equilibria will be [2.supn] - 1, because for each commodity there will be two possible places in which it can be produced.
(8.) Actually, the Country 1 and Country 2 frontiers, although not convex, can be shown to be quasi-convex, which produces the same result.
REFERENCES
Baumol, William J. The Free-Market Innovation Machine: Analyzing the Growth Miracle of Capitalism. Princeton: Princeton University Press, 2002.
Baumol, William J., and William G. Bowen. Performing Arts: the Economic Dilemma. New York: Twentieth-Century Fund, 1966.
Gomory, Ralph E., and William J. Baumol. Global Trade and Conflicting National Interests. Cambridge: MIT Press, 2000.
Johnson, G. E., and F. P. Stafford. "International Competition and Real Wages." American Economic Review 82 (May 1993): 127-30.
Longfield, Mountiford. Lectures on Political Economy Delivered in Trinity and Michaelmas Terms. Dublin: W. Curry, Jr. and Company, 1834.
Maddison, Angus. Monitoring the World Economy, 1829-1992. Paris: Organization for Economic Cooperation and Development, 1995.
Samuelson, Paul A. "Where Ricardo and Mill Rebut and Confirm Arguments of Mainstream Economists Supporting Globalization." Journal of Economic Perspectives 18:3 (Summer 2004): 135-146.
United States. Department of Education. Digest of Education Statistics. 2004 <http://nces.edu.gov>.
--. Department of Labor. Bureau of Labor Statistics. CPI Detailed Report 2004 <http://stats.bls.gov>.
COPYRIGHT 2005 New School for Social Research
COPYRIGHT 2005 Gale Group
