RICARDIAN TRADE THEORY THROWS NO LIGHT ON MIGRATION Julian L. Simon1 INTRODUCTION Immigrants are thought to lower income in the country of origin and raise it in the country of destination because they transfer human capital from the former to the latter. The theory offered in this paper implies that this is not so. An example of the use of this concept is found in the work on the brain drain by Herbert C. Grubel and Anthony Scott. They calculate that "On the basis of [T. W. Schultz's estimates of the value of human capital], U. S. per capita income would be raised in the long run by the emigration of any person with less than two years of college education and lowered if the emigrant had more than two years of college" (1977, p. 31)[1] Another type of use of this concept: Larry Neal and Paul Uselding (1972) estimated the historical effect of immigrants on the United States economy on the assumption that natives benefit from the human capital that immigrants bring with them. The implicit idea is that if a native's net contribution to the economy is zero - the social costs of educating the native offsetting the later benefits from that person - then immigrants must make a net contribution because the social costs of educating the person are borne in the country of origin. And this net contribution is implicitly identified with the Ricardian trade-like effect that immigrants are assumed to have. As Gallman puts it, following Neal and Uselding, "Thus one can compute the 'debt' of the United States to other countries, on account of immigration, in terms of the swings in the costs of child-rearing realized by Americans" (Gallman, 1977, p. 28). This idea has been adopted widely, e. g. Lerner and Roy (1984) apply it to an analysis of the immigration of scientists and engineers. In the essay "Economic Implications of Migration into the Federal Republic of Germany 1988-1992" by Arne Geseck, Ullrich Heilemann and Hans Dietrich von Loeffetholz (1994) the authors say that "the immigrants mean an inflow of human capital ... for which Germany hardly had to invest in education or training". (See also Blitz, 1977; , Hassan, 1988).[1] The human-capital analysis reflects the standard view of economists that international trade and international migration are similar phenomena, to be analyzed with the same general theory2. For example, these are the opening lines in the chapter on "The Economic Effects of Immigration" in the 1986 Economic Report of the President. (p. 213): The movement of people between countries links national economies. Like international trade in goods, services, and financial claims, international migration connects domestic and international markets. The free flow of resources in response to market signals promotes efficiency and produces economic gains for both producers and consumers. The migration of labor, both domestically and internationally, represents such a flow of productive resources (italics added). A similar comment: Economists generally like immigration because it permits the movement of an important resource, labor, from lower to higher valued uses. This movement increases wealth, so we are richer as a nation when we permit immigration and poorer when we prevent it (Froeb, 1995, p. 25). Another example: [T]he employemnt of guest workers constitutes an intenational exchange transaction just as the export and import of physical commodities do... The export and import of labor services via guest workers is analogous with the export and import of physical commodities... Since guest workers can be thought of as but another way countries trade with one another, the neo-classical idea of the 'gains from trade can be applied to the guest worker phenomenon (Krauss and Baumol, 1979, p. 37) And this jocular statement by McCracken illustrates the parallel thinking about trade and migration: "Professors would quickly lose their zeal for open trade if universities found that their strained budgets could be relieved by dismissing high- priced domestic professors and importing cheaper (and better?) pedagogy" (McCracken, 1985, italics added). And another: "Increased labor supply due to immigration enhances the welfare of the typical consumer... " (LaLonde and Topel, 1991, p. 167). This idea apparently has not been challenged until now. The basis for the standard view is the Heckscher-Ohlin-Samuelson theory of factor-price equalization that considers movement of workers and movement of capital to be alternative routes to the same production end.[1] This paper suggests, however, that the analogy is not sound. International trade and international migration are theoretically different phenomena rather than similar. Trade theory is necessarily silent about migration, as is confirmed by its small place in texts on that subject. The gains from trade - at least the first-order Ricardian effects, which are the subject here - are quite unlike the gains from migration. Consumers do not obtain the same sort of gains from the international movement of people as they do from international exchange of goods, though there are important gains from immigration to the countries of destination through mechanisms other than trade-like effects. And the "explanation" of movement of people is different from the Ricardian explanation of movement of goods. The reason for the existence of gains from trade is a continuing difference in the markets of the sort that Hecksher- Ohlin theory is supposed to eliminate but in practice does not; in contrast, after an act of migration there is no such continuing difference which is relevant to the analysis. And hence the Neal-Uselding effect does not exist. To put it differently, one cannot construct the sort of simple numerical example showing immediate gains to consumers consequent to the opening-up of migration of the sort Ricardo could show for the opening-up of trade between two nations. Of course there are a variety of second-order effects that follow from both immigration and from opening-up of trade - for example, immigration-induced changes in technology - but these are not the subject here. Physical capital is excluded from the analysis here for several reasons: 1) simplicity and clarity of presentation, 2) comparability with Ricardo's numerical demonstration, 3) the analysis of Yeager, Borts and Stein, and Berry and Soligo (mentioned below) shows that the famous "triangle" net effect on natives in the two-factor model is very small. Before beginning on the analysis, an analogy may show why it is implausible that there should be a trade-like effect. Consider a youth, whose schooling expenditures (and other "costs") have all been private, entering the labor force in his/her country of birth. No one believes that this event has large positive trade-like effects on the community; indeed, the common (but incorrect) Malthusian presumption is that the youth's entry has negative effects through diminishing returns. The entry to the same labor force of an immigrant youth is an identical event economically, so why should the effects now be expected to be positive? One can complicate the story with the foreign-born youth even going to school in the country whose labor force will be entered, or a native youth going to school abroad, with identical theoretical expectations. The label "immigrant" on one youth or another changes no relevant economic facts in the assumed cases, but somehow induces the expectation in economists of Ricardo-like trade effects. I do not argue that it is impossible in principle to make the phenomena of trade and migration logically equivalent; a skilled economic logician can probably find a set of auxiliary propositions that can do so. My argument is that the Ricardian theory is not useful as a practical explanatory tool, which is consistent with the fact that it is not invoked in theoretical discussions of immigration except at the level of vague handwaving, as seen in the introductory quotations given above. This is unlike the case of the Ricardian analysis of trade. The essence of the paper was put succinctly by Leland Yeager (in correspondence, 1 December, 1993): "An analysis assuming no international movement of factors of production cannot itself address the consequences of their movement". The reader who fully accepts that statement need read no further; the rest of the paper elucidates that idea with various rhetorical devices - a numerical example showing that the consequences of immigration are different than the consequences of trade, a set of analogies, and a variety of verbal analyses of the natures of trade and migration. Yeager regards the central proposition as "trivial" theoretically because it does not lead to theorems or testable implications. I consider that an argument against a widely-held erroneous idea that affects policy discussion is not trivial in the larger context of economics. So much paper and ink is devoted to the subject because it proves to be exceedingly difficult to obtain assent to the central proposition. PRELIMINARY STATEMENT OF THE THEORY To focus the discussion, let us consider first a numerical example of the most difficult case - an "occupationally unbalanced" migration wherein the migrants are not in the same proportion as the work force in the country of destination. (The case where immigrants are in the same proportion as the destination's work force is much easier to analyze.) Assume that just a single barber moves between countries ("U.S." and "India") which have only barber and chicken-farming sectors. The analysis proceeds in two stages, the first stage before the destination economy fully adjusts, and the second stage after it reaches its new equilibrium.3 In Transition, Prior to New Equilibrium Assume that initially the "U.S." and "India" each have 400 worker-hours per week available to the economy and that they produce chickens and haircuts as in Table 1. Table 1: Prior to migration U.S. India Total Haircuts 200 200 400 Chickens 600 400 1000 One worker-hour produces 1 haircut in both countries, but produces 3 chickens in the U.S. and 2 chickens in India. One haircut is worth 3 chickens in the U.S. and only 2 in India. Now let one barber migrate, shifting 40 worker-hours per week of barbering as shown in Table 2. Table 2: After migration but before adjustment U.S. India Total Haircuts 240 160 400 Chickens 600 400 1000 In the first stage, the migrant barber forces down U.S. haircut prices sufficiently to create demand for his services, but as yet no barbers have left the trade. The opposite happens in India. Production in the two countries is as shown in Table 2. Total world production is the same as prior to migration (the opposite of what happens following an opening-up of trade), so there is no total gain in consumer welfare. The barber is better off, however, because s/he is paid at a higher price than before -- more than 80 chickens-worth though less than 120 chickens- worth --which explains why the migration takes place. And there must be some gain in consumer welfare for those who had already been living in the country of immigration, because of the increase in what the budget constraint will purchase; this gain is partially offset by the loss of purchasing power of native barbers. No one has shown how to compute a realistic estimate of such a gain, but the amount is unlikely to rival the computed deadweight gains from trade. Another source of gains to natives during a transition period is from employer or transporter rents arising from exploitation of legal power (such as hiring an illegal alien who is in fear of being reported and deported; no one would hire illegals at the same wage as legals because of the business risks involved), imbalance in market knowledge (the new immigrant being unaware of employment opportunities elsewhere, and hence accepting a relatively low wage), or firm-specific capital (such as the owner speaking the language of an immigrant who does not speak English4). These rents - if they are to be considered so, rather than being considered to be entrepreneurial profits - are consistent with the behavior of firms in investing in the transportation and administrative costs in arranging immigration of selected persons abroad. The rents are conceptually similar to the gains from hiring discriminated-against minorities in Becker's model of discrimination economics. But they do not seem similar to gains from trade. The fact that immigrants' education-and-skill-adjusted earnings begin below those of natives but come to equal or pass those of natives is consistent with the proposition that (except where the law serves to maintain them) these rents are transitory, a phenomenon that has no similarity in the case of trade in goods. The most extreme cases are those of slaves and of babies sold for adoption, cases in which the Ricardian framework could apply as well as it does to trade in producer's and consumer goods respectively. But the natures of these exceptions illuminate and support rather than refute the general proposition of the paper, it would seem. After the New Equilibrium is Reached After some time, the shift in prices leads to occupational adjustments. There is no reason to expect the fully-adjusted occupational proportions to be different than before the migration. Hence the structure of production will be as shown in Table 3, and prices should therefore return to what they were prior to the migration. At equilibrium there is an increase of 20 chickens in total world production, but all the increase goes to the migrant barber, in connection with the higher price of haircuts that s/he gives in the U.S., and it constitutes an increase in his/her standard of living. But everyone else's standard of living is as before. So we distinguish three groups -- the natives of the origin country who do not move, the natives of the destination country, and the immigrant. The first two groups do not benefit from the move, while the immigrant does; this lack of consumer benefit is quite unlike the Ricardian trade situation. Table 3: After adjustment to the immigration U.S. India Total Haircuts 220 180 400 Chickens 660 360 1020 To repeat, trade-induced shifts in prices and production benefit consumers in both countries, whereas after equilibrium is reached the shifts due to international migration benefit only the migrant. And after full adjustment, international migration leaves prices as before, whereas trade in goods alters the price structure. (Capital has been omitted from the model, a matter discussed later.) The entire process can be seen as a single stage if migrants come to the destination country in the same proportion as the existing labor force--not wildly different from reality in the U.S. Prices then simply remain unchanged, the total economy expands in proportion to the increase in population due to the migrants, natives' incomes are unchanged, and immigrants obtain a higher standard of living. This discussion does not answer the question of why production of the tradable good is higher in the destination country. There are many possible explanations, just as in trade theory, the simplest in this case being that persons of higher- than-average competence in the origin country become the migrants. But this question is not part of the matter at issue here. Ricardian trade theory refers to those goods which can be sold separately from the services of the persons who own the means of production of these goods - oranges, chickens, autos, computer programs. In contrast, the benefit which the migrant confers upon the person or organization that pays the migrant for her/his services cannot be delivered at a distance. That is, Ricardian trade theory deals by definition with internationally traded and transported goods, while international (economic) migration is a response to the fact that that person's output cannot be directly sold at a distance. It should not be surprising, then, for trade theory to be silent about the immigration that takes place. The distinction between goods and services is not helpful in this context, I believe. What matters is whether the person in question can sell his or her output directly to a buyer in another country or not. If a computer programmer can market programs by mail, and deliver them by modem, the good is a tradable; if not, not. It does not matter whether the person writes programs for an auto plant or for an airline; if the particular auto plant is in the same country and the programmer cannot work by mail for an auto plant in another country, his output is a non-tradable, even though the auto itself is a tradable. A barber cannot work at a distance, whereas a sculptor with an international reputation can; that is why I choose them as examples (and also because they are not complicated by selling their outputs to other than final buyers). Please notice that both marketing and delivery of the goods must be possible across borders in order that the good be a tradable. Invoking the concept of transaction cost as an explanation -- saying that the transaction cost of moving a barber from one country to another to give a shave is high -- does not seem to me to be helpful here. For some products such as consulting the concept may be helpful, as well as for gravel and stone. But it does not serve to distinguish the main cases with which we are dealing here. Standard trade theory does explain why migration need not take place under certain conditions. But it does not explain why the migration that actually take place should occur. To put the point differently, people either migrate internationally or they do not. One class of people who (according to trade theory) do not migrate are those engaged in the production of internationally traded goods; their wages are thereby equalized in all relevant countries (at least the wages are equalized in principle, though by another principle the equality of wages is not testable empirically); they therefore have no motive to move. Another group of people choose not to migrate because their earnings are greater in the country of origin than in the country of destination. But some people do migrate internationally, even some whose skill in their professions is average in the country of origin. Factor-price-equalization theory provides no motive for these people to migrate. And most important, there is no gain to non-moving natives similar to the Ricardian wine-and- cloth increase in total production whose benefit is realized by native consumers. This paper is not a criticism of trade theory in general or of factor-price-equalization theory in particular. Rather, it simply takes notice of the fact that basic Ricardian trade theory does not explain the international migration that takes place, and does not suggest that there are gains to non-migrant consumers. The rest of the paper is simply an elaboration and explanation of the international migration that does occur, and an analysis of its effects, as is the memo which follows the paper. The reader who is satisfied with the foregoing argument may safely stop here. EXPLANATION The logic of trade theory is sufficiently complex, and the definitions of important quantities and ground rules of the theory are sufficiently elusive, so that the argument easily loses its way when one attempts to deal with this issue formally. Therefore I shall elaborate with extremely simple (though realistic) examples of the simplest sort of migration, cases of individuals who work by themselves and for whom physical capital is unimportant. Assume that migration between the U. S. and India (just two countries for convenience) suddenly goes from being totally restricted to totally unrestricted. Assume, too, that trade in goods heretofore had been free. Along with the change in immigration policy goes the abolition of all occupational regulations that influence earnings, e. g. barber licensing laws in the various U. S. states that restrict entry to barbering and raise haircut prices. Still another assumption is that because of "transactions" costs, movement is not completely and instantly free among occupations. Let us consider the explanation of the immigration, and the analysis of its direct effects, in that order. These are the events that would be expected: Persons in India who produce goods that are freely traded internationally -- for example, internationally-known sculptors whose work is sold in galleries in the United States and elsewhere -- will not have economic reason to move, because their incomes would not be increased by moving; those sculptors can just as well produce their sculpture in India as in the U.S., and send it to the U.S. galleries by ship. But persons who presently sell goods or services that are not traded internationally -- for example, barbers in India, and Indian sculptors who presently do not sell their work abroad but believe that with personal access to U. S. galleries and art buyers they will be able to develop a market in the U.S. and worldwide -- will flock to the U.S. because they believe that with only small alteration in their skills they can sell their services for vastly more than at present, as measured by the purchasing power of their incomes. The musician Ravi Shankar would move if he was at the time mainly in the business of giving concerts, but not if he was then mainly in the business of making records (assuming the marketing of the records would be the same no matter where he is living.) Now about the effects: In the case of goods trading, consumers gain because they obtain more goods than when there is no trade. The classic statement is that of Ricardo who in the first sentence of his chapter on "Foreign Trade" wrote that "extension of foreign trade...will very powerfully contribute to increase the mass of commodities, and therefore the sum of enjoyments" (Irwin paperback edition edited by W. Fellner, 1817/1963, p. 67). And in his Portugal-and-England wine-and- cloth example (p. 71), Ricardo shows how the opening of trade enables England to obtain with the labor of 200 men the goods that heretofore required the labor of 220 men, while Portugal obtains with the labor of 160 men what previously required the labor of 170 men. The decrease in labor required for the same basket of goods implies some combination of an increased consumption of these goods, increase in consumption of other market goods, and increase in leisure. We can say, then, that the extension of trade increases the total goods produced and consumed in each country as well as in both countries taken together. This is proven by Ricardo without considering shifts in relative prices, though such shifts would certainly follow. And the system would remain in this equilibrium - unlike the situation after migration as discussed above. But opening immigration to barbers produces no comparable gain to U. S. consumers. There will be a temporary decrease in the price of haircuts. (There will be no price change if immigrants come to the U. S. in the same proportion in all occupations; a similar statement cannot be made about any trade in goods.) During that initial period after the opening of immigration the same world total numbers of haircuts and other goods are produced as before, but there are as many more haircuts in the U. S. as there are fewer in India. That is, opening up trade in goods is immediately beneficial to consumers because it increases total physical production from the same inputs. But migration of barbers does not at first increase the quantity of total physical production, though it increases the value of the production -- all this keeping in mind that we are so far considering only the immediate effects rather than any effects that may occur after an overall readjustment of the systems. The absence of benefit to consumers from the migration is seen when we extend the illustration. Assume that the barber who migrates to the United States accompanies his entire set of clients in India, which happens to be a Maharajah or a group of internationally-known sculptors who decide to migrate for political and cultural reasons alone and not at all for economic reasons. The barber will receive a higher income because he will be paid in the U. S. by the clients at the U.S. haircut rate rather than at the Indian rate. But the total number of haircuts that he will give, and that they will receive, will be the same in the U.S. as had been the case in India. And no one else will be affected in India or the U.S., even by a change in the price of haircuts, which shows that there is no necessary gain to natives in the country of immigration through this channel. Also unlike the case with trade theory, no native importers and consumers gain the benefit (along with exporters and producers), but rather only the migrants. After the initial stage, the two economies gradually adjust toward equilibrium. There is no reason (except capital dilution, which simply hampers the process) that the two systems should not return to the same structures of prices and production as before the migration. Per-person income in both instances therefore is the same as before. Only the immigrant benefits, moving from a lower-income to a higher-income country.5 The sale of an Indian antiquity in the United States is a case falling somewhere between ordinary goods trade and migration. On the one hand, there is no change in the structure of production in the two countries, unlike Ricardian trading; and there also is no overall change in the world's physical wealth, just as with migration. On the other hand, the antiquity itself is sold and moved, and a seller who remains in India earns from the transaction, which increases measured Indian income; in contrast, the barber's output cannot be sold apart from the location of the barber, and (except in the case of slavery, about which more will be said below) no one profits except the migrant, and hence measured Indian income is not increased by the migration (leaving aside remittances, which are not part of the main story.) One similarity between the phenomena of trade and migration is that opening-up of both migration and trade enables resources to be put to work in higher-value uses than before the opening- up. The difference in value arises from the different price structures -- the difference in price ratios among goods -- in the two countries. And this arises from the more efficient production of the other (tradable) good in the destination country. But again, none of this immediate gain to immigration is obtained by natives of the U. S. unless the services that flow from the migrant's human capital cooperate either with native physical or human capital so as to make that native capital more productive; this does not occur in the case of barbering. The barbers obtain approximately the difference in the value of their production in the two countries. Notice that if you take a trip to India you will get a haircut cheaper than in the U.S, but internationally-traded sculpture will be no cheaper. (Of course if you buy sculpture which is not internationally traded, one could say that it is relatively cheap -- but one could also just as well say that it is relatively expensive; given that it is not internationally traded, by definition there are not two prices in the two countries which one can compare to decide whether it is cheap or expensive.) It is crucial to emphasize this distinction between the trading statuses of a haircut and of internationally-traded sculpture. The very essence of the matter at hand is that migration is a phenomenon which is involved with goods that are not traded internationally -- the services of the migrants -- whereas trade theory focuses on internationally-traded goods. This is why trade theory has nothing to say about immigration. And this is the most likely source of misunderstanding in this matter, I believe, which is why I repeat it. One more example may help: Assume (totally contrary to fact, I hope) that you need a set of dentures. If you go to India, you may be able to have a satisfactory set made more cheaply than in the U.S. But if the denture maker migrates to the U.S., you will then pay her/him as much as you would pay a native U.S. denture maker, and the immigrant pockets the difference. If someone finds a way to sell well-fitted dentures by mail-order, however, the price will come to be the same in both countries (in exchange-rate terms), and there will then be no reason for the denture-maker to migrate (assuming no differences in production methods, etc.). 6 So far there has been no discussion of differences in the amounts of human-made physical capital with which barbers and sculptors work in the two countries, or with differences in the natural endowments of the countries. Rather, we would expect barbers to move because of differences in the demand schedules for barbers at the moment of observation. And this differs from the explanation of why oranges or textiles flow from one country to another, the explanation in those cases having more to do with differential supplies of factors of production across countries than with demand. We could reasonably say that before the opening-up of migration the marginal physical product of barbers was the same in both countries but the marginal value products differed because the structures of relative prices differed in the two countries, barriers to migration being an important cause of the differences in structures. In the cases of haircutting and sculpture we could even drop the word "marginal" from the statement. The purpose for choosing such simple examples is that we then do not need to worry about the complication of some people working for others and selling their wares in the form of services used as a factor of production. In the latter case the wage-worker's personal output is not an internationally traded good, though it may affect the output of internationally traded goods to a greater or lesser extent; this complicates the issue and makes it more difficult to understand the nature of the mechanism at work. This completes the restatement of the main argument -- that there are no consumer gains from migration analogous to the consumer gains from trade in goods. This argument has two parts: First, the total amounts of various types of goods produced are not immediately affected by migration (though they may be later on), whereas the comparable magnitudes are changed by opening up trade. And second, the migrants obtain the benefit of the increase in the marginal value product that occurs immediately due to the international migration, and then they benefit from the increase in physical product that occurs later. The next section contains speculation which the reader may forego and upon which the main argument does not depend, plus some additional explanations that may clarify the matter. DISCUSSION 1. Inevitably, some changes in relative prices due to the immigration of a disproportionate number of barbers, say, will persist for some time because of one key part of the argument adduced here -- that some occupations need not migrate because there are no gains to their migration. Inevitably some distributional effect will occur, and it is especially obvious in the case of self-employed immigrants who work alone, e. g. doctors and domestics. The market for medical services in the United States in recent years has undoubtedly been affected by the influx of foreign physicians. And this is not welcome to U. S. physicians; there will be an obvious loss to the previous sellers of that service who were previously sufficiently efficient to reap some producer surplus, as well as the more difficult-to-calculate loss to the previously marginal producers who are forced out of the business. (And the benefits gained by the immigrants are not an offset to the latter's loss, given that the context of our discussion is the welfare of natives.) During the time of adjustment the shift bestows some additional consumer surplus upon all consumers down to the marginal person who would not have bought at the old price. As to the size of such effects and their impacts upon consumers, considered relative to the gains from trade, I do not have an analysis to offer here. More work is called for. After the systems reach equilibrium, however, and to the (considerable) extent that immigrants are distributed evenly throughout the occupations, and assuming that their consumption patterns are similar to those of natives, the wages and prices of goods will be exactly what they would be without immigrants, aside from the effects due to the temporary dilution of the stock of capital. That is, as occupations shift and the stock of capital returns to what it would have been in the absence of immigrants, the situation will be exactly what it would have been without the immigrants. That is, the immigrants would then cause the economy to be larger, in proportion to the amount of immigration, but all other magnitudes would be the same. 2. One may question whether a barber's skills in India are sufficiently similar to those of barbers in the United States to make the analysis meaningful. Indeed, many barbers in India undoubtedly could not make the transition effectively because of their inability to handle other parts of the barber's trade than cutting hair and shaving beards, or because of difficulty in adjusting to different hair styles. But certainly there are many barbers who would have little difficulty making the transition, and it is they who would migrate. The very fact that many persons do choose to migrate, and make successful adjustments, disproves that cultural styles of work are so different that adjustment is impossible. It may be pointed out that operating as a sculptor in India and selling in the U. S. is not identical to operating in the U. S., because of differences in marketing capacities in the two locations. This is undoubtedly true for many sculptors, but it in no way vitiates the main argument. The argument uses the example of a sculptor who can operate equally effectively in both places, and there are some. But even if there were none, the theoretical example would make the point; the fact that there are actual examples is merely frosting on the cake. And the fact that there are only a few examples in no way damages the logic that the example is used to illustrate. 3. The difference between migration and trade in goods may be illuminated by considering some cases of the movement of people when the facts of ownership are different than they ordinarily are in migration. The key, as Ethier notes (1984, p.20) but does not explore, is that "labor movements have the distinction that they involve the migration of the owner of a factor service as well as the migration of the service itself," that is, property rights are different in the case of migration than in the case of international trade in goods. When babies are "sold" for adoption by parents in poor countries to couples in rich countries, the transaction is exactly like a trade in goods. The natural parents produce the baby at a lower cost (leaving the term "cost" vague for now) than can men and women in a rich country, and they reap gains from the sale; the buyers, too, reap gains in the sense that they pay less for the baby than if they purchased from a rich-country "producer." But if a child grows up and moves herself or himself, the parents then cannot sell the child and reap such gains (except through gift remittances), and the employer of the immigrant grown-up child pays as much for the labor as to a native worker, and hence reaps no gains from the "trade." Another example is the smuggling of people across borders by the "coyotes" on the Mexican-U. S. border; the smugglers are collecting some of the gains from trade just as one would if one were trading in goods; the employer does not obtain any trade gains because s/he pays the alien a market- determined wage (unless the employer pays less because of the illegal's status and fear of being apprehended). Still another example is the sale of slaves, which also is like trade in goods; the slave owner pays less in purchase cost plus upkeep than the slave's marginal product, or else there would be no benefit in holding slaves; the difference between the sale of slaves and the movement of free people typifies the difference between the standard theory of trade and a theory of immigration that fits our purposes here. CONCLUSION This article aims to show that the Ricardian theory of trade has nothing to say about migration that takes place, and immigration does not bring the same sort of benefits as does trade. People either migrate, or they do not. Basic Ricardian trade theory explains why one group of people does not migrate-- those who make internationally traded goods and whose wages therefore become (theoretically) equal in both countries. (Not only is this not very likely in practice, but in principle we can never know whether the wages really are equal). Another set of people who do not migrate are those who earnings would be less in the rich country than in the poor country because their productivity is so low--well below the average in the sending country. The people who do migrate are those whose productivity is average or above in the old country, and who could earn more in the new country for that reason and/or because of the difference in price structures in the old and new countries. These are necessarily people whose earnings are not equalized in the two countries. (For example, sculptors who sell their sculpture on an international market and who can live and work anywhere will not migrate.) Trade theory says nothing about this latter group who do migrate. The core of trade theory is Ricardian, the thrust being that consumers are better off if there is trade than if there is no trade. Immigration does not make consumers better off in a Ricardian way. The gains to consumers from trade result from differences in the structures of production before and after trade--a supply- side phenomenon. The gains to migrants result from differences in demand for labor in the two countries, and the supply structures are unchanged (except in size) in the before-migration and after-migration equilibria. Immigration does, however, bring many other sorts of benefits not discussed here, as well as some costs (see Simon, 1989). These are the externalities that Harry Johnson notes are the forces that matter; Johnson's observation is in unusual accord with the theory presented here. The subject of the paper, international migration, is far from identical theoretically to domestic migration. Relative prices differ much more between origin and destination in the former than in the latter case, and wages differ less (and can easily be compared in the same currency), a key matter for the analysis here. And the ratio of domestic migration going in one direction to that going in the opposite direction tends to be much higher than the comparable ratio in international migration to the United States. REFERENCES Albert R. Berry and Ronald Soligo, "Some Welfare Aspects of International Migration", Journal of Political Economy 77:778-34, Sept./Oct. 1969. Blitz, Rudolph C., "A Benefit-Cost Analysis of Foreign Workers in West-Germany, 1957-1973", KYKLOS, 1977 George J. Borts and Jerome L. Stein, Economic Growth in a Free Market (New York: Columbia University Press, 1966). James M. Buchanan and Richard E. Wagner, "An Efficiency Basis for Federal Fiscal Taxation", in Julius Margolis (ed.), The Analysis of Public Output (New York: NBER-Columbia U. P., 1970), 139-162. Gallman, Robert E., "Human Capital in the First 80 Years of the Republic: How Much Did America Owe the Rest of the World?" American Economic Review, May, 1977, Vol 67 No. 1, pp. 27-31. Geseck, Arne, Ullrich Heilemann and Hans Dietrich von Loeffetholz, "Economic Implications of Migration into the Federal Republic of Germany 1988-1992" in Sarah Spencer (ed.), Immigration as an economic asset: the german experience (Stoke-on-Trent: IPPR/Trentham, 1994) Froeb, Luke M., Vanderbilt Magazine, Winter 1995. Herbert G. Grubel and Anthony Scott, The Brain Drain (Waterloo, Ontario: Wilfrid Laurier Press, 1977). Hassan, M. Kahir, "The Immigration of Third World Scientists and Engineers to the United States", Pakistasn Journal of Applied Economics, Vol . VII, 1988 #1, pp. 43- 58. Johnson, Harry G., "In Search of an Analytical Framework", in The Brain Drain, Walter Adams (ed.), (New York: Macmillan, 1968), pp. 69-91.Krauss, Krauss, Melvyn B. and William J. Baumol, "Guest Workers and Income-Transfer Programs Financed by Host Governemnts", Kyklos, 1979, pp. 36-46. LaLonde, Robert J., and Robert H. Topel, "Labor Market Adjustments to Increased Immigration", in Immigration, Trade, and the Labor Market, Edited by John M. Abowd and Richard B. Freeman, (Chicago: University of Chicago Press, 1991), pp. 167-200. Lerner, J., and Rustum Roy, "Numbers, Origins, Economic Value and Quality of Technically Trained Immigrants into the United States", Scientometrics, 6, 1984, pp. 243-259, Neal, Larry and Paul Uselding, "Immigration: A Neglected Source of American Economic Growth, 1790-1917," Oxford Economic Papers, vol. 24, 1972, pp. 66-88. Ricardo, David, The Principles of Political Economy and Taxation (Homewood, Illinois: Richard D. Irwin, Inc., 1963). Julian L. Simon, The Economic Consequences of Immigration into the United States (Oxford and New York: Basil Blackwell, forthcoming). Paul W. McCracken, "Light at the End of the Budget Tunnel", The Wall Street Journal, March 19,. 1985, p. 28. Simon, Julian L., The Economic Consequences of Immigration (Boston: Basil Blackwell, 1989). **FOOTNOTES** [1]: One might wonder whether the schooling being paid for publicly or privately affects the valuation in the sending country; on most asssumptions it does not. But in any case, there certainly is no effect on the country of destination, the focus of attention in this paper. [1]: Nothing said here is intended to suggest that educated immigrants, even more than uneducated immigrants, are not economically beneficial. But the benefits arise from externalities rather than from a Ricardian trade-like effect. [1]: An exception to the standard view was the assertion by Harry Johnson (1968, p. 80) that world or sending-country loss from brain drain must "hinge on a loss of externalities", which implies that gain must also depend upon externalities rather than an "internal" trade-like effect. This paper may be considered to be a proof of Johnson's assertion. page 1 /article7 immtheor June 3, 1996 NOTES 1I am grateful to Stanley Engerman, Milton Friedman, Herbert Grubel, Charles P. Kindleberger, Nathaniel Leff, Wallace Oates, Ivy Papps, Melvin Reder, Albert Rees, Paul Wonnacott, David Colander, and other participants in a Middlebury College conference on immigration for helpful comments on an earlier draft of this material, though it should be noted that some of them strongly disagreed with the argument. The central idea in this paper was presented in brief in Simon (1989, pp. 17-20). 2Such clear-thinking economists as Gary Becker and Herbert Giersch have said this in public comment, adding that they think the central point of the paper is wrong and they can supply counter-examples. (But neither supplied those counter-examples. And at least one equally clear-thinking Nobelist does agree with the paper.) 3The numbers and some of the language in this example are adapted from a letter by Ivy Papps, who kindly granted me permission to do so, though his argument was different than mine. 4Pertinent here is the joke about the Chinese waiter in a Jewish restaurant speaking broken Yiddish. When one of the patrons asked the owner about this phenomenon, the owner responded, "Sssh, he thinks he's learning English". 5Perhaps it will help the reader's intuition to notice that there is an apparent paradox in thinking about migration. One the one hand, additional people (including immigrants) are (wrongly) considered to lower per-capita income in the country of destination and to raise it in the country of origin because of their effect through the stock of physical capital. On the other hand, immigrants are considered to lower income in the country of origin and raise it in the country of destination because they transfer human capital from the former to the latter. That is, the common presumptions that a) natural population growth reduces income (by reducing output per worker), and b) population growth from immigration increases income, directly oppose each other but refer to exactly the same type of event. And indeed, the theory of Borts and Stein (1964) and Berry and Soligo (1969) refers to the two events interchangeably. One might see this paper as merely an application of Borts- Stein and Berry-Soligo. But the connection to their work was not apparent to me until after many drafts, during which time its arguments had been disputed by almost every reader; hence, the thrust of the paper certainly is not obvious. Also, the B-S-B-S argument focuses on the role of capital, whereas this paper abstracts from it. And their arguments focus upon average income whereas this paper focuses on the separate experiences of natives and immigrants. 6 A newspaper story supports the supposition about the dentures: "WORTH A TRIP: CUT-RATE DENTURES LEAD TO THE OZARKS. Folks Hitch Up their RVs and Head for Missouri...The Russells live in Evansville, Why. Twice in the past month, they have made the two-day drive to this little Ozark town so he could have work done at the Mid American Dental cliinc, a big discount-denture operation here. 'Even with two trips, we saved money,' Mrs. Russell says". (Wall Street Journal, Dec 9, 1993, p. 1.) Migration of the dentist, and mail-order dentures, can be seen as substitutes for migration of the patient, but there are barriers against these possibilities. page 2 /article7 immtheor June 3, 1996 MEMO TO REFEREES AND EDITOR Several extensions and replies to referees' objections are discussed in this memo which is not intended for publication. Some of these matters also are discussed in the text, but receive additional treatment here. (1) The title "Immigration and Trade Are Dissimilar Phenomena" might be more evocative than the present title. (2) The paper in no way conflicts with the factor-price- equalization principle (though it is suggested that that theory is in principle untestable). Just as does Ricardian theory, factor-price-equalization theory refers to goods that can be, and are, actually traded. But immigration takes place when the goods at issue cannot be traded at long distance; that is, economic international migration takes place when movement of people, and movement of capital and movement of the produced goods, are not viable alternatives. Trade theory simply does not deal with internationally non-traded goods, and therefore there can be no conflict between it and the argument of this paper. The paper takes barbers as an example because haircuts cannot be sent by ship or plane. In contrast, trade theory deals with goods like haircutting kits, which can be traded. (For reasons given in the paper, the people who make haircutting kits may still have a reason to migrate, because their personal outputs are not internationally tradable in themselves, unlike the case of sculptors whose work has an international market.) (3) The subject of this paper is not the same as "international trade in services," which has aroused recent theoretical interest. There are several differences: (a) Persons performing the services in question do so in countries that are poorer and less-developed than the countries of which they are citizens; an example is U.S. citizen Red Adair and his crew going to the Persian Gulf to put out oil-well fires. The flow of migrants goes mostly in the other direction. (b) The services are likely to be paid for in the currency of the country of origin, or at least priced that way and then paid for according to the exchange rate; this is not true for immigration. Furthermore, the traded services are likely to be paid for at prices similar to those in the home country. This is important because it is differences in relative prices that drive economic migration. (c) The international movement of some services can best be seen as a phenomenon similar to the domestic expansion of firms from one market to another in all of which the relative prices are the same, or in which relative prices are not the driving force, e.g. McDonald's. This can be thought of as pre-equilibrium growth due to new technical or organizational or marketing developments -- more like domestic investment than like international trade (or immigration); the main motivation is not to take advantage of a difference in price structures. In contrast, the movement of a barber is not due to innovation. (d) Indicative of the difference between barbers and McDonalds is that barbers go to work in the new country instead of the old one, whereas McDonald's opens up new operations abroad in addition to the U.S. locations, rather than instead of them. Trade in services does have in common with the subject of this article that Ricardian trade theory cannot be used to calculate the gains in welfare to the two countries and in aggregate. This is because the opening-up of trade does not lead to increased division of labor between the two countries; rather (to a first approximation), the trade in services changes the location where the persons involved perform their services. (4) The examples in the paper were chosen because in them capital is unimportant; this tactic is intended to bring out the central principle. But even if these examples were perverse, proponents of a theory against which it is directed should show why the examples are not relevant, if they wish to maintain their theory. The larger stock of capital per worker in the new country than in the old country works to make immigration relatively attractive to many immigrants, but it has no direct effects upon consumers, and certainly no effects similar to the case of trade. More generally, bringing capital into the analysis lowers the wages of native workers while increasing the returns to native capital-owners by slightly more than the workers' loss, as Berry and Soligo showed (1969), none of which is crucial in the present context. (5) One reader asked why the immigrants with above-average skill are not paid their marginal product in their home countries. Frank (1984) documents that individuals with above- average skills often are not paid their full marginal products, and he suggests a variety of reasons why this is not so. The closest example at hand is academia. Double Nobel-winner John Bardeen, a major cause of the transistorized electronics industry, is paid less than some second-rate professors of finance, and not all that much more than most professors in his own field of physics. This phenomenon probably is less important, however, than relative prices among occupations differing between poor and rich countries. (6) On the assumption that "a barber can at zero cost become a sculptor", Wallace Oates worked out a model with results in a conclusion different than the conclusion here (letter of Oct. 7, 1985). But I think that it would not be realistic to assume that labor is fungible even if there were no natural differences in capacity and taste, because of the sunk cost to investment in learning. Once one has spent the money to get a PhD in economics, for example, it is much cheaper to practice that trade than for someone else who is presently a barber and must yet make an investment in a PhD. And most international immigrants are at the stage of their lives at which they have already made a large part or all of their educational investments, which provides theoretical explanation for the observed fact that people do not easily and frequently make major switches of occupation. (7) It has been suggested that an important reason why wages for a given occupation differ between poor and rich countries is difference in social organization. I very much agree, and this would in no way conflict with my main argument. But this difference affects efficiency in production. And I believe that my argument (if it is to be sustainable) must hold even where the output of the worker is the same in the old and new countries -- the same barber or sculptor, for example. (8) One correspondent wrote that "The shift in international resources is bound to cause a shift in production combinations." Agreed, at least for the first stage prior to new equilibria. But I see no reason why this would necessarily benefit natives of either country; the matter is similar to the shift in relative prices discussed above, as I see it. The correspondent also says that there would be a "more efficient allocation of resources." I'm willing to agree, though "efficient" is hard to define. But the barber might get all of the increase in value. (The quantity of production might stay the same.) (9) The same correspondent wrote that "the people involved would be moving up their supply curve." Why must this be? In cases where the immigrants combine with more and better physical and human capital in the rich receiving country, this would indeed be so. But a barber or physician might do less work and produce less physical output in the new country. (10) Another reader suggested that the argument in this paper might be correct -- that is, there might be no gain to consumers -- in the case of small countries and "inconsequential" immigration. But that in itself shows the difference between trade in goods and immigration, because trade produces gains in all cases. (11) Perhaps the best indirect proof that immigration produces no immediate consumer benefits akin to the benefits from trade is that there is no narrow special-interest pressure group lobbying for immigration on economic grounds. In contrast, consumers are upset at having to pay higher taxes when auto imports are reduced by federal action. And when textile imports are reduced by regulation, the retail clerks union takes out newspaper ads to protest, because of the jobs that may be lost due to the increased price of garments in stores. (12) A by-product of the argument presented in this paper is that it helps distinguish between international migration to the United States, and contemporary internal migration within the United States (as distinct from the rural-to-urban migration of the past), phenomena which have generally not been distinguished in past discussion although they are quite dissimilar. page 3 /article7 immtheor June 3, 1996 A1Empirically, immigration and modern domestic migration are very different in their respective ratios of net to gross migration. The ratio of annual net to gross migration for cities in the U. S. is very large, of the order of 1:10 or 1:20 (see Simon, 1989, p. 16). In contrast, the U.S. ratio of net immigration to the total of emigration and immigration historically is of the order of 3:4. So considered as a group, international immigration is mainly one-way whereas domestic migration is mainly two-way. (These remarks do not pertain to black migration from the rural South in the U. S., , or to migrations from farm to city.) These data fit with the variation in price structures being relatively small among various regions of the United States, compared to the variation in price structures among countries where the ratio of the price of a taxi ride, say, to the price of a VCR is very different in the two places. The phenomenon of internal migration, then, is one of people moving from place to place in order to better match up their specific abilities with the characteristics of specific jobs; the match can be better west or east, north or south, for most persons in perhaps most occupations perhaps most of the time. This can be seen vividly in the both-directions placement of persons newly-graduated from universities. Hence almost all research on internal migration focuses upon the personal characteristics of the migrants, and upon distance, which is symmetrical rather than unidirectional. To put the matter differently, there are broad factors such as differences in average income, as related to the structures of prices and wages, that help us understand international migration. Though the study of individual characteristics would undoubtedly be illuminating, it is probably no accident that individual characteristics have not been the focus of past research on internation migration. In contrast, internal migration seems to be the product of a myriad of relatively small "random" factors related to individual employers and prospective employees rather than to the economic environment generally. To put the matter statistically, a prediction about which way a first-time mover would move based on wage or income or employment differentials would have (say) 51 percent accuracy between two cities, but perhaps 95 percent accuracy between a poor and a rich country. Hence analyses using such an explanatory variable for inter-city moves would not be very useful (would have little statistical power) in understanding inter-country moves. People move for different reasons than goods move. Even if there are only small differences in relative price structures between two places, there may be systematic one-way trade in particular goods between the locations. ("One way" means that the ratio between net movement and gross movement is high.) Chesapeake Bay oysters move to the Mountain states, while Rocky Mountain "oysters" move to the East Coast; there is no reverse movement of those goods. Large differences in the relative price structures of two countries, differences which are closely associated with large differences in the average incomes and levels of development of the two countries, do not constitute sufficient condition for profitable movement of all persons from the poorer to the richer country. Aside from considerations of differences in physical capital between the countries (and such differences are not the key issue in immigration, as discussed in the paper), roughly as many persons would lose by such a move as would gain by it. This is because the average output per person equals the average income per person, and the relative price structure is unfavorable to some just as it is favorable for some. Productivity and marginal products are low on average in poor countries, which is why they are poor. This is why the notion that a country can be characterised as having "cheap labor" is misleading. And the truism of the output-input identity explains why Swedish shipworkers are brought to Polish shipyards to finish difficult jobs and paid many times the Polish daily wage; it also explains why in China domestic television sets cost two or three times as much as the world price of foreign television sets. We may now summarize the differences between the explanations of migration where the relative price structures are similar and where they are different. When the structures are similar, the economic motivation for moving is mainly one-to-one individual matches of job and person. When the structures are different, people will migrate a) because the price structure is relatively more favorable for the person's occupation in the destination country, and/or b) because (i) the person's skill is above the occupational average in the country of origin, and (ii) skill in that occupation in the country of destination is on average higher than in the country of origin (which it is likely to be the case if the country of destination has higher average income than the country of origin). It is important to note that circumstances (a) and (b) are not present for a goodly proportion of any poor country's working population. It should be noted that the subject is contemporary mostly- urban migration, rather than the migration of agriculturists that took place to the New World in the last century. It was not so clear that movement from, say, Germany to the U.S. was movement from a poorer to a richer country. To the extent that the farming of the immigrants was largely for subsistence, the explanation did not depend upon prices but simply upon land prices and productivity. And much the same was true about grain production for the world market. But this case has some special twists that need not be pursued here. (13) The aforegoing leads us to ask: Why are there income differences from place to place? Why are barbers' wages in the U. S. now sufficiently high that Indian barbers believe that they can improve their standard of living by migrating? It is true that immigration has been restricted, and the going wages for barbers in the U. S. are those necessary to compete with opportunities in other industries. But this is not a satisfactory explanation. The question may at first seem trivial. It seems obvious that a person may expect to earn more in a rich country than in a poor country, assuming s/he finds employment in the rich country. But further reflection suggests that the answer is not so obvious. Theory assumes that labor is paid its marginal product. Without additional considerations, and assuming the same amount of cooperating capital in both countries, a given person would be assumed to produce the same marginal physical product in the new country as in the old one, which implies the same earnings in both countries, ceteris paribus. If earnings are indeed different, as much casual evidence suggests is the case, then there must be some additional elements of the situation which we must come to understand. Another candidate for explaining the phenomenon of differential wages for barbers is a difference in the cost of living. But surprisingly, upon examination it is not clear that the cost of a basket of quality- standardized goods differs between rich and poor places, especially (as one would expect) the set of transportable goods, generalizing from evidence among the U. S. states (Love,1982). The better supply of equipment capital with which people work in rich countries than in poor countries is offered by many (e. g. Kindleberger in correspondence; Grubel and Scott, 1977) as a possible explanation. But differences in capital supply within occupations obviously do not explain the differences in earnings of physicians or of journalists in rich and poor countries. One might still argue that the earnings of physicians, journalists, and barbers are affected by the supply of capital in other occupations, which therefore affects the overall structure of wages. But there is no theoretical reason to expect to find different capital-supply situations, to the extent that such a comparison is meaningful. Unlike the situation in the past when a new continent was suddenly opened up, we would not expect to find major differences in the supplies of capital among countries after capital has had opportunity to flow toward opportunities; there would be no theoretical reason to expect to find different amounts of capital cooperating with persons of the same skill in different countries even if one country is poor and the other rich. And there may really not be differences in the supply of capital greater than are warranted by the human capital that is available with which the capital can cooperate. Perhaps the general situation is analogous to the situation in Indian agriculture a couple of decades ago when Schultz concluded that the supposed capital shortage there did not exist. So this explanation is insufficient even if it may contribute in part to the observed effect. The correctness of this view here is not important to the rest of the discussion, however, so we can pass on. Another reason to believe that a difference in the the supply of capital is not a central explanation is a twist in the Leontief paradox: There is migration from rich countries to poor countries (albeit often temporary migration) of experts such as construction engineers; the goods they are selling clearly are labor intensive rather than capital intensive. One plausible explanation of different earnings in different countries for persons with similar skills is that the price of a given product of that skill may be different in the rich and poor countries. This is well-known for goods that cannot easily be traded internationally, e. g. bus travel, and the phenomenon can be seen in the difference in price ratios of the same pairs of goods in different countries (which is also a major cause of the difficulty of comparing standards of living across countries, as is seen in Kravis's work. Furthermore, even in industries which can be compared across nations because they produce international goods, some of the inputs may not be easily traceable to outputs, e. g. the impact upon total production of the nurse on duty at a factory is not easily determined, and hence it is not possible to know if s/he is being paid her/his marginal value product. Additionally, the existence of differences in factor shares over time and across nations seems to confirm that wages are not always set neatly by the value of marginal output. In my view, inter-country differences in structures of prices in goods and in wages, differences which are not erased by international trade in goods, are the most important cause of the differences in earnings for given occupations, which in turn lead to international migration. This recalls the flavor of the "cheap labor" notion, and in fact the present line of thinking points toward what is meaningful in that notion. A country's labor as a whole cannot be cheap or expensive because the income of a country (except under unusual conditions) equals its output, and therefore people are on average paid what they produce. This is seen most clearly in subsistence farming, where the only wage payment that one can imagine is what a person gives to himself/herself in consumption; the notion of cheapness of labor obviously makes no sense in that context. But particular types of labor may be relatively cheap or expensive in an economy, compared to other economies. If, because an individual is tied to a given poor economy, s/he will produce a given amount of his/her services, and therefore of final goods, cheaper (or in some cases, more expensively) than would the same individual in a given rich economy, then the person's labor may sensibly be said to be cheap (or expensive) in the poor economy. And this can indeed sometimes be the case, as suggested by the theorizing above, together with plenty of casual evidence. But nothing so far tells us that barber's wages ought to be relatively high or relatively low, where "relatively" means that the wage is high enough so that the person is better off by remaining in India rather than migrating. We know that since the price structures differ from place to place, there must be some occupations whose pay is relatively high and some whose pay is relatively low, but there is no clearcut basis to predict which will be which. The fact of working in an industry that makes internationally-traded goods should not be crucial, given equalization of wages throughout an economy on the argument given above about barber's wages in the U. S.A2 On the other hand, if we assume that there is some force operating in the direction of factor price equalization between trading countries, then we might expect at least some pressure in that direction. But the best statement might simply be that any two economies are very different from each other in respect to which occupations work relatively productively and which work relatively unproductively, and the difficulty of comparing the purchasing power of a given sum of one currency in the two economies is sufficiently great so that no one can do it with precision; therefore, anything is possible. (This can be seen starkly with some retail price-ratio comparisons. Some years back a man's undershirt cost the same in the U.S. as a pair of underpants, but in Israel the undershirt cost four times as much as underpants, with no obvious physical differences that explain the difference in ratios.) Ethier (1984) systematically examined migration in the light of pure trade theory. His work helps establish the potential contributions of that body of theory, as well as the limits to those contributions. In Ethier's analysis of people as inputs to the production of traded goods, the reason that people move between two countries is that the countries have recently opened up trade and are not yet in equilibrium, and the relative endowments of capital and labor differ between the two countries, or else there is an absolute production cost difference between the countries which causes different wage rates. (The Ricardian model may be seen as a special case of the Heckscher-Ohlin- Samuelson model for the moment.) For the most part, such analysis of people as labor input is the same as the analysis of capital as input. The migration occurs in response to the difference in wage rates, and continues until a wage-rate equilibrium is reached (at wage equality under most sets of assumptions). Migrant agricultural labor in the states bordering Mexico would seem to fit this model, although the same crops might be grown just as well on the Mexican side of the border, and the causes of the location of production are not immediately obvious. There has been, however, considerable time for pairs of countries such as the U. S. and the U. K. to move most of the way toward such an equilibrium between them. One might argue that immigration restrictions prevent an equilibrium from being reached. But prior to 1924 there were few restrictions on immigration into the U. S., and the quota from the U. K. to the U. S. has not been an effective constraint. Yet migration has not equalized wages across the two countries in such fashion as to remove the reasons that cause some persons to migrate. By the 1990s, we should be ready to abandon the practice of talking about "wages" as if they are a homogeneous entity; trying to do so simply introduces confusion. If theory tells us anything conclusive, it is that (apart from differences in the supply of capital and natural endowment, which are increasingly less important with time) average wages are by definition the same across countries. But wages in some occupations may be sufficiently different to induce people to migrate, e. g. physicians' wages in the Philippines and in the U. S. And the wages of particular individuals may be less than their particular skills should command in the country of origin. But the difficulty of establishing the facts in any particular case is very great. Analogously, some years ago it took two researchers six weeks to compare two supermarkets to determine which could be said to be the cheaper (Holdren, 1960); the multiplicity of items, and the differences in them, make such a comparison extremely tricky. It may have been possible to compare "the price of food" century after century, from country to country, hundreds of years ago, but no longer; so it is with wages now. The reader may find it difficult to believe that there are persons in many occupations in poor countries whose earnings would not justify migration. And the fact that so many persons in so many different occupations wish to immigrate to the U. S. seems to substantiate this view (though there seems to be no information on just how many would come if they could). But it is likely that many persons who immigrate do so because they believe themselves to be relatively highly-skilled persons in occupations where the average level of skill is low in their country relative to the United States, and as individuals they can improve themselves even if the average person in that occupation could not. This is consistent with the anecdotes and scraps of statistical evidence indicating that immigrants are self-selected with respect to skill and drive. The unpredictability of the relative price of labor of a given occupation in two countries may be shown by one of the very few occupations where the effective labor rendered by a person is demonstrably much the same in both places: prostitution. And there is cross-migration in this trade, to the extent that the law allows it. According to newspaper stories, prostitutes from the U. S. work in the Philippines and Thailand, and prostitutes from those countries apparently work in the U. S. This case suggests the importance of taste differences, whereas in other cases differences in labor practices are the cause of the wage structures. The cause does not matter here, however, only the fact that there is no known way to deduce theoretically the relative wage situation of a given occupation. (The case of prostitution also illustrates well that the capital with which a person works is not necessarily the key to the wage structure and the cause of a differential that the migrant exploits by migrating.)A3 It is important not to dismiss as a nicety the impossibility of cleanly comparing the prices of labor services across nations. The difficulty is twofold. First, it is practically never possible to be sure that the labor service being priced is the same in the two countries, even if the actual person is the same. Can we determine whether an advertising executive is performing the same service for an agency in Bombay as in Atlanta when the other workers (and in other occupations, the physical capital) that the person in question is cooperating with are so different in the two countries? And second, comparing standards of living at any pair of incomes, one in each of the two currencies, is difficult in considerable part because the market basket differs from person to person, which is compounded by changes in exchange rates from time to time. The comparison is difficult even for the same person; the advertising executive instanced above told me that she could not easily determine whether the salary she was moving back to India to accept was greater or less in purchasing power than the salary she was leaving in the United States. The fact that the labor in an entire organized collectivity -- that is, a firm -- may be more expensive on an efficiency basis in a poor country than in a rich country is proven by the fact that poor countries find it cheaper to import than to produce some goods, even if production capital is a minor matter, as in the case of television set assembly. "[T]he cost of producing a TV set in China runs two to three times the price of an imported one." (Hong [Yeung]), 1986, p. 21) Would we expect that income differences for given occupations and individuals would soon disappear after the opening-up of immigration -- especially if we specify that all barriers to the flow of capital also will be wiped away? I think that there will continue to be income differences for journalists, say, for whom the amount of cooperating capital is not importantly different in the two countries, for a long time after the opening-up. And I think there will continue to be differences for busboys in restaurants. The supervisors of the busboys will be better managers in the U. S. than in Mexico for the foreseeable future, and therefore the output per busboy will be higher. And one of the reasons the supervisors will get more output is the higher wages of the busboys, causation running in both directions. My point is that the stocks of human capital with which people work will continue to differ for a very long time. One cause of the perseveration is that the characteristics of the stocks of human capital are self-perpetuating. Additionally, economic and social and political institutions may differ greatly, e. g. Hong Kong, Taiwan, and China. And as long as those stocks of human capital differ, most journalists and executives in India will not compete with journalists and executives in the U. S. for open jobs on equal terms. One can, of course, rationalize all this by talking about equal pay for equal "effective" labor, but that simply talks away the matter, I believe, rather than helping us understand why there are income differences between countries that motivate people to move. Again, differences in demand in the various countries would seem to be primarily responsible for the movement of migrants. These differences in demand may stem from differences in taste. Or -- and I would think that this is likely to be more important -- the differences stem from differences in production conditions in various industries in the various countries, due to many kinds of historical and cultural factors. Even after a long time, I would expect that a New York-style city desk editor or busboy supervisor would face different demand schedules for his/her labor in New York and Delhi; New York skills cannot be as well exploited in Delhi as in New York. (14) As I was pondering the lack of acceptance of the central idea of this paper by most of those to whom I have presented it, I found myself reading Albert Einstein's Relativity (1916). It is an explanation in simple prose intended for laypeople -- in the words of the dustjacket, "explaining the theory in simple words that anyone with the equivalent of a high school education can understand." (The dustjacket may be correct in principle, but this reader, at least, finds it extraordinarily difficult even after hundreds of hours of reading.) There is a provocative formal similarity between Einstein's problem and the problem under discussion here. Einstein's theory relates events that occur in different places; the same is true of trade and immigration theory. The source of the difficulty in both cases is that relating the separated events seems obvious and easy. In the case at hand, one can correctly say that a person is paid his/her marginal output both in the country of origin and the country of destination. But it is not agreeable to then hear that unless those marginal outputs are related in value terms, they cannot sensibly be compared. The only possible simple comparisons are necessarily inconclusive; the barter value of a haircut in bushels of wheat is not likely to be the same in the two countries, and the barter equivalents of a haircut in wheat and in dental services are likely to be unbalanced in different directions for those two goods. No recourse to a transformation such as the money values figured at the exchange rate can in principle wholly succeed; the inevitable flaw in that process is due to the differences in relative prices and quantities consumed in the various countries, differences which are at the heart of the phenomenon of international migration, it is argued here. The purpose of such work as that of Kravis et al. (1982) is to provide a useful, but necessarily partial, measure of the transformation that would be needed. Perhaps because we are so accustomed to making exchange-rate comparisons for a variety of purposes, and because it is never enjoyable to recognize the inherent impossibility of this or any other measure providing a theoretically complete transformation, one does not feel comfortable with an argument that is based squarely upon this inherent impossibility, as is the case of this paper. It is, however, fatal to the understanding of the present argument to shrug this off and rush on to formal manipulations. Einstein makes a similar point at the very start of his discussion, a point which is the keystone of his thinking and the point from which all of his theory of relativity departs. Lightning has struck the rails on our railway embankment at two places A and B far distant from each other. I make the additional assertion that these two lightning flashes occurred simultaneously. If I ask you whether there is sense in this statement, you will answer my question with a decided "Yes." But if I now approach you with the request to explain to me the sense of the statement more precisely, you find after some consideration that the answer to this question is not so easy as it appears at first sight. After some time perhaps the following answer would occur to you: "The significance of the statement is clear in itself and needs no further explanation; of course it would require some consideration if I were to be commissioned to determine by observations whether in the actual case the two events took place simultaneously or not." I cannot be satisfied with this answer for the following reason. Supposing that as a result of ingenious considerations an able meteorologist were to discover that the lightning must always strike the places A and B simultaneously, then we should be faced with the task of testing whether or not this theoretical result is in accordance with the reality. We encounter the same difficulty with all physical statements in which the conception "simultaneous" plays a part. The concept does not exist for the physicist until he has the possibility of discovering whether or not it is fulfilled in an actual case. We thus require a definition of simultaneity such that this definition supplies us with the method by means of which, in the present case, he can decide by experiment whether or not both the lightning strokes occurred simultaneously. As long as this requirement is not satisfied, I allow myself to be deceived as a physicist (and of course the same applies if I am not a physicist), when I imagine that I am able to attach a meaning to the statement of simultaneity. (I would ask the reader not to proceed farther until he is fully convinced on this point.) (1916, pp. 21-22) The difference between Einstein and those who came before him is that he accepted this difficulty as a problem to be dealt with, rather than either not noticing its existence or closing his eyes to it. Lest it be thought that I am making a laughably inappropriate comparison between the present work and Einstein's (and indeed, I have been reluctant to invoke his name because I recognize how pretentious it may seem), I note several enormous differences: First, I am attempting to use an insight; Einstein discovered it. And the discovery of that insight about the nature of simultaneity is the very heart of his work. Second, Einstein found a transformation (the Lorentz transformation) that solved the problem raised by his insight; this paper only argues that the problem raised is unsolvable in principle. And third, the matter at hand has none of the earthshaking importance of Einstein's work, of course. These remarks are in no way intended to suggest that physics is a good model for economics. Rather, I agree with Hayek that physics is a poor model because of the simplicity of the systems with which physics deals, and the lack of the subjective element in that subject. But it should be noted that the physics that some take as the model for economics is Newtonian rather than modern physics, and perhaps the greatest philosophical difference between the classical and modern is the doing-away with property definitions in such circumstances as Einstein's pathbreaking case. This is the matter which makes Einstein relevant for the present discussion. (15) The use of homely examples and detailed specifications as a mode of explanation is not pleasing to some. Here I seek shelter under the spirit of Einstein as expositor. "The purpose of mechanics is to describe how bodies change their position in space with 'time'. I should load my conscience with grave sins against the sacred spirit of lucidity were I to formulate the aims of mechanics without serious reflection and detailed explanations." (1916, p. 9) Einstein also pleads, "In the interest of clearness, it appeared to me inevitable that I should repeat myself frequently, without paying the slightest attention to the elegance of the presentation. I adhered scrupulously to the precept of that brilliant theoretical physicist, according to whom matters of elegance ought to be left to the tailor and the .pa cobbler." (p.v.) I, too, repeat my argument several times in different ways. page 4 /article7 immtheor June 3, 1996 REFERENCES FOR MEMO See end of paper for references to paper proper Berry, R. Albert, and Ronald Soligo, "Some Welfare Aspects of International Migration," Journal of Political Economy, 77:778-94, Sept.Oct., 1969. Einstein, Albert. Relativity - The Special and General Theory (New York: Crown Publishers, Inc., 1916/61). Ethier, Wilfrid J., "International Trade Theory and International Migration." (xerox, 1984). Frank, Robert H., "Are Workers Paid their Marginal Products?", The American Economic Review, September, 1984. Bob R. Holdren, The Structure of a Retail Market and the Market Behavior of Retail Units (Englewood Cliffs, N. J.: Prentice- Hall, 1960). Yeung, Wai Hong, "China Gets Cold Feet in the Open Market," The Wall Street Journal, May 19, 1986, p. 21. Kravis, Irving B., Alan Heston, and Robert Summers, World Product and Income - International Comparisons of Real Gross Product (Baltimore and London: The Johns Hopkins University Press, 1982). Douglas O. Love, "City Population and Item Prices", Research in Population Economics, Vol. 4, pp. 83-92 (Greenwich, Conn: JAI Press, 1982). Schwartz, Aba, "On Efficiency of Migration," The Journal of Human Resources, VI, 2, Spring 1971. Yeung (see Hong). page 5 /article7 immtheor June 3, 1996 FOOTNOTES FOR MEMO A1International migration also differs from internal migration in magnitude; there may be sufficient domestic migration so that consideration of the use of public goods and other consumption ought to influence policy. Buchanan and Wagner (1970) recommended federal fiscal equalization for its efficiency in keeping natives from moving among cities in search of a better supply of public goods. Whether or not this is a sound argument with respect to domestic migration, it seems much less potent with respect to international migrants, whose aim is more likely to be a change of place in search of conditions where their work can produce output of higher value, that is, a production consideration rather than a consumption consideration. A2Writing this in a hotel in Boston, I am struck by evidence that even being a non-traded good is not a very good predictor. Room service brought me coffee here at just about the same price, and much quicker and with less hassle, than in India in December -- despite the supposedly vastly overvalued U.S. dollar. In India, my order for coffee at 6:00 in the morning was probably the only order that he had all night. And making it up was undoubtedly a major production, judging by the time it took. A3Nothing said here implies that generalizations about relative wages are impossible in principle. It may, for example, be possible to make valid generalizations about the relative pay of classes of skilled versus unskilled labor as a function of average income in a country, though it would certainly be difficult to hold skill constant in inter-country comparisons. But even if it can be established that the relative pay of lower- skilled occupations is less in LDC's than in MDC's, this would not automatically imply a greater incentive for lower-skilled persons to migrate than for higher-skilled persons, because that incentive depends upon the absolute gains to a given person, not the gains relative to other occupations, and the absolute gains may well be larger in higher-skilled occupations even if their relative pay is higher in LDC's. The propensity of persons in different income and skill classes to immigrate to the United States is an interesting empirical question that might be illuminated with data on migration from Puerto Rico to the mainland U.S., and with data on migration of Soviet Jews and North African Jews to Israel and to the United States. page 6 /article7 immtheor June 3, 1996