CHAPTER 8 THE EFFECTS ON TECHNOLOGY, PRODUCTIVITY, AND NATIVE HUMAN CAPITAL By far the most important determinant of society's standard of living is the productivity of the persons who make the goods and services. Productivity depends, of course, upon education, upon the amount of capital with which people cooperate, and upon the efficiency with which that capital is used. But productivity also depends upon the types of techniques, including the organizational practices, that are in use at any given time. And the types of techniques that are in use are heavily influenced by the state of technical knowledge ("technology"). Knowledge stems from human minds. Minds matter economically as much as, or more than, mouths or hands. In the long run, the most important economic effect of immigrants is their contribution to our stock of useful knowledge. And this contribution is large enough in the long run to dominate all the other benefits and costs of immigration. Immigrants affect productivity and technology partly in their special role as immigrants by stimulating both natives and immigrants to invent new ideas that are some combination of the transported ideas and the ideas that are already present in the country of immigration. They also carry ideas and practices from one society to another, thereby inducing natives to adopt the transported ideas. The special role of immigrants as idea- transporters has been important throughout history. In earlier times, the movement of people was crucial in transmitting ideas from one part of the world to another. As Childe tells us about ancient Greece, after it had learned some tolerance: The foreigners in each city brought with them and celebrated there their own native cults. With these religions and their officiants spread new brands of magic and philosophy--a motley horde of quacks, astrologers, alchemists, and oraclemongers--competing with traditional beliefs and legitimate sciences. (l942, p. 254) In the Middle Ages, migrants carried new techniques such as advances in weaving looms from one part of Europe to another, thereby fructifying the countries where they settled. An American example is "the introduction of rice culture in South Carolina. . . . it was accelerated, if not determined, by the skills of the slaves brought into Carolina, many of whom had grown rice in Africa. Apparently the masters didn't know the first thing about rice culture." (Gallman ?, 1977, p. 29). And though modern communications greatly speed the transfer process, the movement of people is still very important in transmitting knowledge. For example, I learn much about Japanese and Indian ways of doing things from informal conversations with colleagues from those countries. Such is the stuff of advance in civilization and economy. One of the nicer minor benefits of immigration is that it increases cultural variety--for example, Chinese and French restaurants. But the benefits of variety go beyond consumer and esthetic pleasures and spill over into economics. Variety is a key ingredient of invention. A minor illustration: as a result of eating in a Chinese or French restaurant and comparing the food with what he or she cooks, a homemaker or commercial chef or food processing executive may get new ideas for recipes combining the foreign and native cuisines. This minor example has often been writ large in major cultural advances produced by, or from contact with, immigrants. Because scientific and technical inventions occur inside the mind, and the process is not visible on the outside, it is difficult to find convincing examples of this process at work. But the process is quite obvious in artistic and religious meldings of many traditions. Foreign graduate students raise the productivity of American univrsity scientists as well as directly contributing their own new ideas and discoveries. Herbert Simon (along with Jean-Jacques Servan-Schreiber, 1987) recently wrote about the situation in his field of computer science, as well as engineering, the two most popular fields for foreign students. "As of 1986, more than 300,000 foreign students were enrolled in American universities, 60 percent of them in technical fields". They account for between 52 percent and 68 percent of the graduate students in the various engineering programs, and 40 percent in computer science. In direct fiscal terms, these foreign graduate students are subsidized by American taxpayers to the tune of more than half their tuition. But, Servan-Schreiber and H. Simon say, "Foreign students give the United Sates as much as they get. They are paying for their long years of study with the most precious and expensive commodity, the one the United States most needs today: more knowledge, new knowledge, provided by their labor. By working in American laboratories for three to seven years of post-graduate study, thousands of young experts are by themselves the most efficient 'subsidy' to scientific progress and economic development." And as to whether the foreigners are displacing Americans, they quote Karl Willenbrock, director of the American Society for Engineering Education: "We don't have too many foreign students, we have too few Americans. We are not attracting enough of our students into graduate schools." A second way in which immigrants affect productivity is in their roles as "regular people" who participate just like natives in the production process and then buy goods with their incomes. As consumers they increase the total volume of goods produced, which increases productivity through "learning by doing". As workers they increase productivity through their contributions to the development of improved production processes and better products. Both of these mechanisms will be discussed at length below because of their great importance in a complete assessment of the overall economic effect of immigrants, even though the evidence concerning these mechanisms comes from observations on the society as a whole rather than from observations of immigrants alone. Please keep in mind that this chapter (drawn mainly from Simon, 1981) discusses partial effects of immigrants upon natives' standard of living, effects through changes in technology and productivity. This chapter does not take up the overall effect of more immigrants or more people generally; that topic is tackled in Chapter 10, after this and the previous chapters have laid the groundwork for that overall assessment. EFFECTS THROUGH INCREASED MARKET SIZE Immigrants constitute additional consumers who increase the size of the markets for the goods which they consume. Economists since William Petty and Adam Smith have understood the importance of market size in influencing productivity. As market size increases, greater efficiency results from the division of labor and other well-known economies of scale. In more recent years, economists have also noted the influence of the size of the market--the total output and income in a market--on the decision to invest. As Nurkse put it: "The inducement to invest is limited by the size of the market...the level of productivity depends--not entirely by any means, but largely--on the use of capital in pro- duction. But the use of capital is inhibited, to start with, by the small size of the market." (Quoted by Agarwala and Singh, l963). Increased size of market also implies an increase in research development and in the adoption of new technology, with a consequent improvement in technique in practice. The modernization in technique occurs partly through the simplest route: New physical capital embodies newer technology. But additionally, as Schmookler (1962a; 1962b; 1966) and Simon and Sullivan (1985) have shown, increased investment in an industry induces a higher level of invention in the industry, as indicated by patents and technical book publications. Yet many economists--at least in this century--have refused to draw the most obvious conclusion from this line of reasoning: Additional people lead to faster economic growth by increasing the size of the market, and hence boosting productivity and investment. That is, it is obvious that an increase in the labor force is overwhelmingly important in increasing total output and hence the size of the market. But nevertheless, very few writers have drawn the conclusion that in this way more people have a positive effect on overall economic well-being. Sometimes it is said that additional population growth no longer is beneficial through this mechanism because the market is _a_l_r_e_a_d_y large enough. But no evidence in support of that qualification has been given. Thus, this qualification may be regarded simply as an unsupported opinion. Let us now consider the matter in more sober detail. The Theory The greater efficiency of larger-scale production stems from (l) the ability to use larger and more efficient machinery, (2) the greater division of labor in situations where the market is larger, (3) knowledge creation and technological change, and (4) improved transportation and communication. We shall take up each of these factors briefly, then in more detail. Please keep in mind as we proceed that there is no easy and neat distinction between increases in productivity due to increased knowledge, and increases in productivity due to economies of scale; they are interdependent, and both are accelerated by population growth and therefore by immigration. (1) A bigger population implies a bigger market, all else equal. A bigger market promotes bigger manufacturing plants that are likely to be more efficient than smaller ones, as well as longer production runs and hence lower setup costs per unit of output. (2) A larger market also makes possible a greater division of labor and hence an increase in the skill with which goods and services are made. Adam Smith emphasized the importance of the division of labor and used the example of pinmaking; his predecessor Petty made the same point when talking of the advantages of a large city like London over a small city, and he used a more vivid example than did Smith: . . . the Gain which is made by Manufactures, will be greater, as the Manufacture it self is greater and better . . . each Manufacture will be divided into as many parts as possible, whereby the Work of each Artisan will be simple and easie; As for Example. In the making of a Watch, If one Man shall make the Wheels, another the Spring, another shall Engrave the Dial-plate, and another shall make the Cases, then the Watch will be better and cheaper, than if the whole Work be put upon any one Man. And we also see that in Towns, and in the Streets of a great Town, where all the inhabitants are almost of one Trade, the Commodity peculiar to those places is made better and cheaper than elsewhere . . . (Petty, 1682/2 1899, p. 473) Specialization can also occur with respect to machinery. If the market for its goods is small, a firm will buy multipurpose machines that can be used in the production of several kinds of products. If its market is larger, the firm can afford to buy more efficient specialized machines for each operation. Larger markets also support a wider variety of services. If population is too small, there may be too few people to constitute a profitable market for a given product or service. In such a case there will be no seller, and people who need the product or service will suffer by not being able to obtain it. (3) Economies of scale also stem from learning. The more television sets or bridges or airplanes that a group of people produces, the more chances they have to improve their skills with "learning by doing"--a very important factor in increasing productivity, as many studies show. The bigger the population, the more of everything that is produced, which promotes learning by doing. (4) As discussed in Chapter 7, a bigger population makes profitable many major social investments that would not otherwise be profitable--for example, railroads, irrigation systems, and ports. The amount of such construction often depends upon the population density per given land area. For example, if an Australian farmer were to clear a piece of land very far from the nearest neighboring farm, he/she might have no way to ship her/his produce to market and may have difficulty in obtaining labor and supplies. But when more farms are established nearby, roads will be built that will link him/her with markets and supplies. Such reasoning lay behind Australia's desire for more immigrants and a larger population; this was also the case for the American West during the 19th century. Public services such as fire protection are other social activities that can often be carried on at lower cost per person when the population is larger. There may also be diseconomies of increased scale, however, such as congestion. As the number of sellers and activity in, say, a city's wholesale fruit-and-vegetable market increases, transacting one's business may become more difficult because of crowding and confusion. Each additional person imposes some costs on other people by decreasing the space in which the other person can move around, and by inflicting her/his pollution (soot, noise) on other people. Therefore, the more people there are, the less space each person has and the more pollution each suffers from, all else equal. These effects would be felt both in a decreased ease and joy of living and in higher prices due to the higher costs of production caused by congestion. This sort of diseconomy is very much like the concept of diminishing returns from a given acre of land that is at the heart of Malthusian reasoning. Ultimately congestion must occur as long as there is some factor of production that remains fixed in size, be it land for the farmer or market area for the wholesaler. But if that factor can be increased rather than remaining fixed--by building a bigger market or by bringing new land into cultivation--then the diseconomies of scale, especially congestion, can be avoided or reduced. An example of increasing congestion as population and economic activity increase is found in this news story: NEW YORK -- Lower Manhattan, the financial heart of the country, is suffering from massive hardening of its arteries. The arteries are the subway tunnels and streets that lie at the base of the commercial center, which one planning organization describes as "the most densely populated square mile in the Western world." Though lower Manhattan probably has more transportation services that any site in the country, much of the time the decayed network is tied into a strangled knot of disruption and delay. . . The potential for total blockage is apparent in the cramped, twisted innards of many lower Manhattan subway stations, compared by one planner to a "rat maze." At the Whitehall Street-South Ferry station, for example, arriving passengers often take several minutes to clear the platform as they wait to mount four-foot-wide stairs to the street. . . Another problem is lower Manhattan's narrow, tangled streets. Laid out by Dutch colonists some 300 years ago, the roads are overburdened with cars, trucks, buses and pedestrians, who are squeezed off even narrower sidewalks. Passage is difficult; parking is almost impossible. One result: Wall Street Mail Pickup Service, a private mail carrier, recently added a second man to its trucks so that one employee can drive around the block while the other makes pickups. (The Wall Street Journal, October 15, l984, p. 35.) But such congestion is not a permanent condition. The news story goes on to quote New York's city planners about the situation: "There's nothing wrong with very high density as long as you take care of the transportation design." The subway system in lower Manhattan was built many years ago for a lower level of economic activity. When New York eventually deals with this situation, people on average will probably wind up better off than if the congestion had not for a while increased to the point of discomfort. THE STATISTICAL EVIDENCE Let's begin our examination of the statistical evidence with an estimate of the overall effects of population size on productivity in less-developed countries (LDCs). Chenery compared the manufacturing sectors in a variety of countries and found that, all else being equal, if one country is twice as populous as another, output per worker is 20 percent larger. This is a very large positive effect of population size no matter how you look at it. Now let us move from the national level down to the industry level, and let us shift from LDCs to more-developed countries (MDCs) because most of the available information pertains to MDCs. In every industry, there is some minimum size of factory that must be attained in order to reach a reasonable operating efficiency. But though this is the sort of economy of scale that has been most studied in the past (because of its industrial applications), it is not the economy of scale that is most relevant to population questions. More relevant are studies of industries as wholes. As mentioned above, it is an important and well-established phenomenon that the faster an industry grows, the faster its efficiency increases--even compared with the same industry in other countries. The most recent and complete analysis is shown in Figure 8-l. There we see comparisons of the productivity of U. S. industries in l950 and l963, and of U. K. industries in l963, with U. K. industries in l950--and also comparisons of U. S. industries in l963 with those of Canada in the same year. The larger the industry relative to the U. K. or Canada base, the higher its productivity. This effect is very large: Productivity goes up roughly with the square root of output. That is, if you quadruple the size of an industry, you may expect to double the output per worker and per unit of capital employed.1 ---------- Figure 8-l ---------- The effect Chenery observed in economies as wholes, together with the effects seen in individual industries, constitutes strong evidence that a larger and faster-growing population produces a greater rate of increase in economic efficiency. The phenomenon called "learning by doing" is surely a key factor in the improvement of productivity in particular industries and in the economy as a whole. The idea is a simple one: The more units produced in a plant or an industry, the more efficiently they are produced, as people learn and develop better methods. Industrial engineers have understood learning by doing for many decades, but economists first grasped its importance for the production of airplanes in World War II, when it was referred to as the "80 percent curve": A doubling in the cumulative production of a particular airplane led to a 20 percent reduction in labor per plane. That is, if the first airplane required l,000 units of labor, the second would require 80 percent of l,000 or 800 units, the fourth would require 80 percent of 800, or 640 units, and so on, though after some time the rate of learning probably slows up. Similar "progress ratios" have been found for lathes, machine tools, textile machines, and ships. The economic importance of learning by doing is very great. The effect of learning by doing can also be seen in the progressive reduction in prices of new consumer devices in the years following their introduction to the market. The examples of room air conditioners and color television sets are shown in figure 8-2. ---------- Figure 8-2 ---------- The studies discussed above automatically subtract any costs of congestion from the positive effects of scale. But it should be interesting to many readers to know how large the congestion costs are by themselves. If there really are important congestion problems in bigger cities, for example, one would expect them to be reflected in the cost-of-living data for cities of different sizes. But no strong relationship between size of city and cost of living is apparent. More detailed statistical studies of this evidence reveal that at most there is a tiny effect: The largest estimate is a 1 percent increase in the cost of living for each additional million people, for people living on a high budget; other estimates range downward to no effect at all. A study by Love (l976) of the relationship of city size to the prices of over 200 individual goods and services found that although more prices go up with increasing city size than go down, for almost every good or service, workers are more productive in the larger cities after the higher wage in bigger cities is allowed for. And the higher incomes in larger cities more than make up for the higher prices, so that the overall purchasing power of a person's labor is greater in the bigger cities (Love and Simon, no date). This suggests that the disadvantages of congestion are less than the positive effects of greater population (including better communications and more competition) upon the standard of living in larger cities. And in a general study of the costs of manufacturing production, Sveikauskas (l975) found an economically-important advantage in efficiency in larger cities. There is also evidence that less capital is needed to produce a given amount of output in larger cities (summarized by Alonso, l975). And the cost of capital is lower in larger communities, as measured by bank rates (Riefler, 1930; Stevens, l978). Another important element is the greater density of communications and transportation networks that accompanies denser population. This may be seen casually in the larger number of radio and television stations in larger cities. And Segal (l976) found that SMSA's with populations of two million or more have 8% higher productivity than smaller SMSA's, because "economies exist in transport and communications in the very largest cities..." EFFECTS FROM MORE CREATORS It is a simple fact that the source of improvements in productivity is the human mind, and a human mind is seldom found apart from a human body. Because improvements--their invention and their adoption--come from people, it seems reasonable to assume that the amount of improvement depends on the number of people available to use their minds. This is an old idea, going back at least as far as Petty in l682. As for the Arts of Delight and Ornament, they are best promoted by the greatest number of emulators. And it is more likely that one ingenious curious man may rather be found among 4 million than 400 persons. . . . And for the progagation and improvement of useful learning, the same may be said concerning it as above-said concerning . . . the Arts of Delight and Ornaments. . . . (1682/1899, p. 473) More recently, this benefit of population size has been urged upon us by Kuznets (1960). In contrast, many of the doomsday writers completely omit from consideration the possibility that, all else equal, more people imply more knowledge and greater productivity. For example: "It is difficult to see how any further growth in world population could enhance the quality of human existence. On the other hand, one can readily envisage ways in which further population increments will diminish our well-being." (Brown, 1974, p. 149) It cannot be emphasized too strongly that "technological advance" does not mean only "science," and scientific geniuses are just one part of the knowledge process. Many technological advances come from people who are neither well educated nor well paid: the dispatcher who develops a slightly better way of deploying the taxis in his ten-taxi fleet; the shipper who discovers that garbage cans make excellent cheap containers; the supermarket manager who finds a way to display more merchandise in a given space; the supermarket clerk who finds a quicker way to stamp prices on cans; the market researcher in the supermarket chain who experiments and finds more efficient and cheaper means of advertising the store's prices and sale items, and so on. The scarcity of additional producers of knowledge, and of their potential contribution to resources and the economy, also is manifest. Nobel prize winner Bethe tells us that the future cost and availability of nuclear power--and hence the cost and availability of energy generally--would be a rosier prospect if the population of scientific workers were larger. Talking specifically about nuclear fusion, Bethe said, "Money is not the limiting factor. . . . Progress is limited rather by the availability of highly trained workers." (1976, p. 2) Still another important element is the greater propensity to produce new ideas that accompanies living in larger cities (Higgs, l97l; Kelley, l972), and the greater propensity for new ideas and trends to diffuse and be adopted in larger cities (Simon and Golembo, l967). Just _w_h_y people are more likely to create and pick up new ideas where population is denser is still the subject of sociological speculation, but the evidence for the phenomenon is rather solid. Students of organizational behavior also tell us that, all else being equal, the larger an organization's resources in numbers of people and amounts of money, the more innovations it will come up with. "If any one group of variables may be said to stand out among all others as empirically determined correlates of innovation, it is the group of inter-related factors indicating size, wealth, or the availability of resources." A variety of investigators "all conclude that organizational size and wealth are among the strongest predictors of innovation in the sense of readiness to adopt new patterns of behavior." (Mohr, 1962, p. 112) Even a casual inspection of the historical record confirms this connection between population size and growth of knowledge. There have been many more discoveries and a faster rate of productivity growth in the past century than in previous centuries, when there were fewer people alive. True, l0,000 years ago there wasn't much knowledge upon which to build new ideas. But, seen differently, it should have been all the easier l0,000 years ago to find important improvements, because so much still lay undiscovered. Progress surely was agonizingly slow in prehistoric times. For example, whereas now we develop new materials (metals and plastics) almost every day, it took centuries between the discovery and use of, say, copper and iron. It makes sense that if there had been a larger population in earlier times, the pace of increase in technological practice would have been faster. Population growth spurs the adoption of existing technology as well as the invention of new technology. This has been well documented in agriculture (Boserup, 1965) where people turn to successively more "advanced" but more laborious methods of getting food as population density increases--methods that were previously known but that were not used because they were not needed earlier. This scheme well describes the passage from hunting and gathering--which we now know requires extraordinarily few hours of work a week to provide a full diet--to migratory slash-and-burn agriculture, and thence to settled long-fallow agriculture, to short-fallow agriculture, and eventually to the use of fertilizer, irrigation, and multiple cropping. Though each stage initially requires more labor than the previous one, the endpoint is a more efficient and productive system that requires much less labor. This phenomenon also throws light on why the advance of civilization is not a "race" between technology and population advancing independently of each other. Contrary to the Malthusian view, there is no immediate linkage between each food- increasing invention and increased production of food. Some inventions--the "invention-pull" type, such as a better calendar- -may be adopted as soon as they are proven successful, because they will increase production with no more labor (or will enable less labor to produce the same amount of food). But other inventions--the "population push" type, such as settled agriculture or irrigated multicropping--require more labor, and hence will not be adopted until demand from additional population warrants the adoption. (Simon, 1977, Chapter 8; Simon, 1978) The Malthusian invention-pull innovation is indeed in a sort of race between population and technology. But the adoption of the population-push inventions is not in a race at all; rather, it is the sort of process discussed at length in the chapters on natural resources. If a larger labor force causes a faster rate of productivity increase, one would expect to find that productivity has advanced faster and faster as population has grown. Ancient Greece and Rome have been offered as counter-examples to this line of reasoning. Therefore I plotted the numbers of great discoveries, as recorded by historians of science who have made such lists, against population size in various centuries. Figure 8-3 shows that population growth or size, or both, were associated with an increase in scientific activity, and population decline with a decrease. (Of course other factors come to bear, too.) ---------- Figure 8-3 __________ As for the contemporary scene and better data, Solow concludes that the yearly rate of increase of productivity doubled, from l percent to 2 percent, between the periods l909-29 and l929-49 periods (1957, p. 320); and the populations and and labor forces of the U. S. and of the developed world were larger in the latter period than in the earlier period. Fellner found the following rates of increase in productivity (using two methods of calculation): l900-29--l.8 (or l.5) percent; l929-48- -2.3 (or 2.0) percent; l948-66--2.8 percent (1970, pp. 11,12). These results are consistent with the assumption that productivity indeed increases faster when population is larger-- though of course other factors could explain part of the acceleration. Here an important caution is needed: Because of the economic inter-relatedness of all modern countries, we should think about the population and productivity growth of the developed world--or indeed of the world as a whole--rather than think about any particular country. One country can, to some extent, ride on the coattails of the developed world as a whole, but this is less likely than is often thought, because local research and development are needed to adapt international knowledge to local conditions. For example, high-yielding seeds cannot simply be imported and planted successfully without extensive adaptation to the local sunlight angle, temperature, water and soil conditions, and so on. So, though our data refers to individual countries, or to cross-sections of countries, the unit to which our discussion applies best is the developed world as a whole. But is it certain that the recent acceleration of productivity would not have occurred, even if population had been smaller? The connections between numbers of scientists, inventors and ideas, and the adoption and use of new discoveries, are difficult to delineate clearly. But the links needed to confirm this connection seem very obvious and strong. For example, the data show clearly that the bigger the population of a country, the greater the number of scientists and the larger the amount of scientific knowledge produced. More specifically, scientific output is proportional to population size in countries at the same level of income (Love and Pashute, 1978). The U. S., for example, is much larger than Sweden, and it produces much more scientific knowledge. Sweden benefits from the larger U. S. population because it "imports" much more knowledge from the U. S. than the U. S. imports from Sweden; this can be seen in the references used in Swedish and U. S. scientific writings, and in the number of patented processes licensed from each other. Then why aren't populous China and India the most advanced countries of all, given that they have the largest populations? The obvious answer is that China and India do not produce as much new knowledge as the U. S. or even the U. S. S. R., because China and India are relatively poor, and hence they are able to educate relatively fewer people. Yet it is instructive that despite its poverty, India has one of the largest scientific communities in the world, just because it has such a large population. Put differently, would you bet on Sweden or Holland, against Great Britain and the U. S. S. R., to produce the great discoveries that will make nuclear fusion practical? (I have omitted the U. S. from this bet because of its higher per person income than Britain or the U. S. S. R.). I am saying that, in the long run, the most important economic impact of population size and growth is the effect of additional people upon the stock of useful knowledge employed in the production of goods and services. And this positive effect is large enough (in the long run) to dominate all the negative effects of population growth. This is a strong statement, but the evidence for it seems strong, as I shall now try to show. (For a more detailed presentation of the evidence, see chapters 4 and 6 of Simon, l977). Let's begin with a question: Why is the standard of living so much higher in the U.S. or Sweden than in India or Mali? And why is the standard of living so much higher in the U.S. or Sweden now than it was two hundred years ago? The proximate cause is that the average worker in the U.S. or Sweden now produces x times as much goods and services per day as does the average worker in India or Mali, or as did the average worker in the U.S. or Sweden two hundred years ago, where x is the ratio of the standard of living now in the U.S. or Sweden, relative to that in India or Mali now or relative to the U.S. or Sweden then. Though the first answer is almost definitional, it points us to the important next question: Just _w_h_y does the average worker in Sweden now produce so much more per day than does the average worker in Mali, or than did the average worker in Sweden two hundred years ago? Part of the answer is that the average worker in Sweden today has available to him or her a much larger supply of capital equipment to work with--more buildings, tools, and transportation equipment. But that is only a minor factor; as proof, notice how fast West Germany and Japan were able to regain a high standard of living even after much of their capital was destroyed in World War II. The all-important difference between the U.S. or Sweden now, and those countries two hundred years ago or India now, is that there is a much greater stock of technological know-how available now, and people are educated to learn and use that knowledge. The knowledge and the schooling are intertwined; in India now, unlike the U.S. two hundred years ago, the knowledge is available in books in the library, but without the schooling the knowledge cannot be adapted to local needs and then put to work. The stock of industrial capital is also intertwined with the stock of knowledge and with education; the value of much of our stock of capital, such as computers and jet airplanes, consists largely of the new knowledge that is built into them. And without educated workers, these chunks of capital cannot be operated and hence would be worthless. The importance of the technical knowledge factor has clearly emerged in two famous studies, one by Solow (l957) and the other by Denison (l976). Using different methods, they calculated the extent to which the growth of physical capital and the labor force could account for economic growth in the U.S. and Europe. Both found that even after capital and labor are allowed for, much of economic growth cannot reasonably be explained by any factor other than an improvement in the level of technological practice (including improved organizational methods). Economies of scale due to larger factory size do not appear to be very important in this context, though in larger and faster-growing industries the level of technology improves more rapidly than in smaller and slower-growing economies. This improvement in productivity with technological practice did not come for free, of course; much of it was "bought" with investments in research- and-development (R&D). But that does not diminish the importance to us of the gains in technological knowledge. How do immigration, population size, and population growth come into the picture? To repeat, the source of improvements in productivity is the human mind, and a human mind is seldom found unaccompanied by a human body. And because improvements--their invention and their adoption--come from people, it seems reasonable to assume that the amount of improvement depends on the number of people available to use their minds. The reader may wonder whether a person need live in the United States for the United States to get the benefit of the person's impact on productivity. The answer differs somewhat depending on the person's origin--that is, whether the person is from a more-developed or a less- developed country. The answer may also depend on the person's education and occupation, but the effect of the former is clearer and probably much more important. Recall, please, that a person may influence technical progress through both his/her demand for goods and her/his supply of knowledge. Let us consider demand and supply separately, looking first at the more problematic case, that of the person who already lives in a more-developed country such as Sweden or Japan, and then examining the case of a person who moves from a poor country to a rich country. It is indeed true that there is international trade; a Swede's demand for goods may be satisfied by imports from the United States. It is also true, and more relevant, that only a small proportion of U. S. goods are sold abroad. It is more likely that an increment of U. S.-made autos or newspapers or smoke detectors will be sold if a given person chooses to reside this year in the United States rather than in Sweden. This should be enough evidence to make the point. An even stronger argument comes from a more general view of trade. If a Swede migrates to the United States and still imports a Swedish auto, Sweden's imports (directly or indirectly) from the United States will rise by the amount of other goods equal in trade value to the auto. Total production in the United States, therefore, will rise by the amount of the immigrant's output and income, which will cause learning-by-doing and other demand-induced productivity-increasing mechanisms. We must also consider, however, whether the flow of technology among developed countries is so free that it does not matter in which country the technical progress is first made. By now there seems to be consensus among students of the subject that it does matter. A first reason is that there is a time lag of, say, a minimum of three years. Second, much technical progress is a matter of local adaptation, such as new agricultural varieties and techniques that depend on particular soil and climatic conditions; this is why even individual states within the United States can get a high return on research and development in agriculture. (Griliches, 1958; Evenson, 1968) If a person goes from a poor country, where little new technology is being created, to a rich country where much technology is being created, this argument is obviously even stronger. In this case the United States benefits not merely by the person contributing to technology that will be differentially helpful to the United States but also by the absolute increment of technology that the person creates. The more technically advanced (relative to the state of the art) is the industry in which a person works, the greater the opportunity for that person to advance the state of the art. It is not a contradiction to this line of thought that the rate of economic growth per capita in the post-World War II period has been proportionately as high or higher in the poorer countries as in the United States. The poorer countries can take advantage of the technological progress in the richer countries much more than the reverse can occur. A reduction in the quality or amount of education that children receive is another possible negative effect of population growth on immigration upon the growth of knowledge. Human capital as well as physical capital is crucial in the productivity of an economy. And people might not wish to provide (or authorities might not demand) enough additional tax revenues to maintain an equivalent level of schooling in the face of population growth. If so, a larger population, with its larger proportion of children, might lead to less education on the average, and thus to less potential for individuals to increase the stock of knowledge, than a smaller population. But this reduction is doubtful. The conventional theory underlying this idea is straightforward Malthus: A fixed educational budget of money and resources divided among more students implies less education per student. But we know from casual evidence that people and institutions often respond to population growth by altering the apparently fixed conditions. In agricultural countries, for example, having more children causes parents to increase their labor on the land. And in industrial countries, when there are additional profitable opportunities for investment, people will shift some resources from consumption to investment; additional children constitute such an opportunity. Therefore, we must allow for responses contrary to the simple Malthusian pie-sharing theory. There is no way of knowing from theory alone which of the two effects--dilution of resources or increase of work--will dominate. Therefore we must turn to empirical analysis. A comparison of rates of population growth in LDCs with the amounts of education given to children shows that an increase in the birthrate reduces educational expenditures per child, and also secondary enrollment, but does not reduce primary or postsecondary enrollment (Simon and Pilarski, 1976). Perhaps the most important result of that study is that the negative effects are nowhere near as great as the simple Malthusian theory would suggest, and in general the effects do not seem to be large. One may wonder whether the immigrants themselves obtain, through their earnings, all the benefits of the new knowledge that they create. This question can be answered confidently in the negative. A large body of excellent econometric studies have shown that the social rate of return to research and development tends to exceed the internal rate of return to investment at large. And given that firms are willing to pay their employees only for the benefits that the firm can obtain from their output, we can conclude that immigrant knowledge producers are not themselves able to extract all the economic benefit of the knowledge they produce, and hence the remainder is enjoyed by the rest of society. HUMAN CAPITAL EXTERNALITIES A possible drawback to very large-scale immigration--which seems to have been overlooked in past technical discussions of immigration--is a force which we may call "human capital externalities." A person's output depends not only on the person's own skills and the quality of the machines one works with, but also on the quality of the skills of the people one works with. Immigrants from poor countries possess poorer productive skills than do people with the same amount of formal education in richer countries; this is almost definitionally true, and the effect can be seen in the lower incomes of those immigrants in the countries from which they have come than in the incomes they expect in the U.S. (which is exactly why they come). So until they improve their informally learned sills--handling modern communications systems, for example, or getting used to doing things by telephone and computer rather than in person and with pencil and paper--they represent lower-quality human capital for American workers to cooperate with. And this will reduce the productivity (and growth of productivity) of American workers until the immigrants--in perhaps 2 or 5 years--pick up the informal learning (after which they forge ahead of natives). Of course there is a linked positive effect, the beneficial impact of working with someone from a different culture. Immigrants cause new ideas to arise, even when higher-skilled persons are exposed to more "primitive" ways of doing a job. This effect is also likely to run out, however, as the immigrants become more Americanized. And there is no strong reason to believe that this positive effect outweighs the linked negative human-capital effect. The size of the negative human-capital externality effect depends upon the proportion of immigrants that natives work with. If your work companions are one percent new immigrants, the negative effect may be small, and outweighed by the positive effect. But if you work with immigrants 50% of the time, the outcome might be quite different. There is an analogy here to busing. The original notion of busing was that kids from poor backgrounds could gain from the school environment (including the other kids) that better-off kids have, and the better-off kids would not suffer thereby. Maybe this is true for 1% or 10% bused-in kids. But what if it is 90% or 99% of the class? Quite obviously the result for the 90% or 99% would be no different than if they never left their neighborhood schools (aside from better teachers, if it is so), and for the better-off kids representing 10% or 1% of the population, it would be as if they had been bused to the poor neighborhood school. If 1 million Russians or Indians or Chinese move to Champaign-Urbana, Illinois, the schools and workplaces would become like those in the countries of origin. And this is why towns on both sides of the U.S.-Mexican border are somewhere between mainstream U.S. and mainstream Mexican culture and productivity. This line of reasoning just spells out the person- in-the-street's thoughts about being "inundated" by a great many immigrants. But--the average immigrant worker comes to have higher earnings than the average American worker after a few years in the U.S., due in large part to greater youth, and a high level of average education. This suggests that the human capital externality effect comes to be positive in a fairly short time. And it is probably positive on balance, as long as we weigh the future and the present normally. IMMIGRATION, TECHNOLOGY, AND THE COMPETITIVE FUTURE OF THE UNITED STATES As economics usually does, this book addresses the effects of immigration upon individuals -- mainly U. S. residents, and also the immigrants themselves. Also following the usual practice of economics, the book mainly analyses immigration incrementally -- that is, determining the effect of the "next" single immigrant -- rather than asking about the effect of an entire large cohort of immigrants. At a very few points the question has been raised whether a large number of immigrants might have a different effect than a much smaller number of immigrants, or different than effects that are experienced with the current or historical volumes of immigration; for example, the effect on natives of working together with a relatively large number of immigrants was considered in Chapter --. But even in these exceptional cases only the possible special difficulties and not the possible benefits of a large cohort of immigrants were discussed. In connection with consideration of a large cohort of immigrants, some readers might also wish to consider the political implications of the economic effect of immigration -- and necessarily, the implications of a fairly large flow of special immigrants, rather than just the "next" ordinary immigrant. That is, the issue is the political standing of the U. S. relative to other countries. It may reasonably be assumed that a nation's political standing is heavily influenced by its economic situation, both the standard of living of individuals and the total output of the residents taken together. And when we ask this question, we find that immigration represents perhaps the most amazing opportunity for the U. S. that any country has ever had to get ahead of its political rival or rivals -- the safest, cheapest, surest alternative ever available to a country. This statement is not made as a recommendation but rather simply as observation. Nowadays the future of any country, and most especially the future of a major country that is in the vanguard with respect to production and living standards, hangs completely on its advance in knowledge and skill and productivity. This is more true at present than in the past because technology changes so rapidly nowadays. Even a single invention can radically alter a country's economic or military future -- consider, for example, the atom bomb or the computer -- with a speed that no invention could in the past, even the invention of the gun. All that the U. S. need do to sharply increase the rate of advance in its technology and its industrial productivity is relax its barriers against the immigration of skilled creators of knowledge. It could simply give permanent-resident visas to those foreign students who come to the U. S. to study. Many foreign students already find ways to remain under the present rules -- about half among the students of engineering and science (H. Simon and Servan-Schreiber, 1987), a group that those writers argue persuasively is crucial to present progress. But even more foreign graduates would remain if allowed to, and they could push up the rate of progress even further. Going even further, if young people abroad knew that they would be able to remain in the United States after completing their education here, more would choose to come to the U. S. to study. This would provide multiple benefits to the United States. Given assurance that they could remain, these students could afford more realistic tuition rates than are now charged, which would benefit U. S. universities. And these increased rates would enable the universities to expand their programs to serve both foreign and native students better. Most important of all, of course, would be the increased number of highly competent engineers and scientists of all kinds who would be part of the American work force. Going even further, a larger number of students requires a larger number of professors. And a larger number of openings for professors, especially in such fields as engineering and science, would attract from abroad more of the world's best scientists. This would reinforce the process which has brought to the United States so many foreigners who have won Nobel prizes in the U. S. to the advantage as well as honor of the U. S. Just why knowledge workers produce so much more when they are located in the U. S. than in poor countries is exceedingly hard to state clearly. It must be some combination of the reward structure, the resources available to work with, the colleagues (which is a self-reinforcing matter), closeness to the center of scientific activities, and so on. But the fact that the difference is great is indubitable. Where would the physical capital come from for the additional scientists to work with, and to use for housing and schooling their families? Would this not mean a "dilution" of the capital supply in such fashion that there is less capital for natives to work with, and hence a diminution in the productivity of native workers? This bugaboo is discussed in Chapter 7. The conclusion arrived at there, contrary to everyday common sense, is that aside from the shortest- run considerations, physical capital does not pose a constraint. One matter not emphasized there that deserves special mention in the present context is that new investment opportunities are created by immigrant scientists and engineers, and capital must flow into the country from abroad to invest in these opportunities, which works to keep the supply of capital per worker from being diluted. No other country has such opportunity. The Soviet Union must even resort to compulsion to prevent its native scientists from leaving, and there obviously is little demand from foreigners to work there. Europe draws some, but is perhaps hampered by the fact that English is the international language of science, which makes it advantageous to live in an English- speaking country such as the U. S., Canada, Australia, or Great Britain. CONCLUSIONS Immigrants influence productivity both in their special roles as immigrants and in their general roles as additional persons. As immigrants, they bring new and different ideas from their old societies to their new society which may lead to useful improvements. As additional persons they increase productivity both directly as additional ingenious minds, and also indirectly by the impetus that their increased demand and consequent increased production volume gives to productivity by way of learning by doing. 86-84 Prod8 1/16/87 REFERENCES Robert E. Evenson, "The Contribution of Agricultural Research and Extension to Agricultural Production" (Ph.D. diss., University of Chicago, l968), summarized in Yujiro Hayami and Vernon W. Ruttan, Agricultural Development: An International Perspective (Baltimore: Johns Hopkins University Press, l97l). Zvi Griliches, "Research Costs and Social Returns: Hybrid Corn and Related Innovation," Journal of Political Economy, Vol. 66 (October l958), pp. 419-531. 86-84 Prod8 1/16/87 FOOTNOTES 1The U.S.-U.K. comparisons in the same year are relatively free of the potential bias arising from the fact that those industries where world technology grew faster exogenously were also those whose scale of production therefore expanded faster, a bias which afflicts analagous time-series studies within a single country. 86-84 Prod8 12/16/87