CHAPTER 11 JOB DISPLACEMENT: THEORY OF IMMIGRANTS AND NATIVE UNEMPLOYMENT Jacques Chirac, Prime Minister of France . . . for several years now, and particularly since 1981, large numbers of immigrants have been arriving in France. Given the economic situation, there are too many of them. Their numbers will thus have to decline. Liberation [A Paris daily newspaper]: Can you assert that there is a link between the number of immigrants and the economic situation? Chirac: Naturally. If there were fewer immigrants, there would be less unemployment, fewer tensions in certain towns and neighborhoods, a lower social cost. Liberation: That has never been formally proven. Chirac: It is easy to imagine, nevertheless. This means that we must not accept any new immigration and that we must severly curb illegal immigrants, rigorously apply the laws of the Republic, systematically expel those whose status is irregular . . . (October 30, 1984, taken from Population and Development Review, March, 1985, p. 164. INTRODUCTION "Displacement" of natives by immigrants is the most emotional and politically-influential fear about immigration in the U.S. and elsewhere. As the Committee for the Economic Development of Australia put it, "No issue fires the debate about immigration more than its alleged effects on unemployment in Australia." (Norman and Meikle, 1983). Concern about wages being driven down is a related issue, but wages seem to enter less into public rhetoric than does job displacement. It is commonly asserted that immigrants "take jobs from citizens." Let us consider some typical statements: A candidate for U.S. Senate: Sixty-five Americans lose their jobs for every 100 undocumented workers who are here. (Kent Hance; The Washington Post, May 7, 1984, pp. A1 and A4.) A low-skilled worker: Belle Glade, Fla. (AP). [Headline] "Haitian Refugees Take Away Jobs" [text] "'Man, I can't find no work,' complains 20-year-old Mark Lane, a black American who once packed lettuce for $250 a week. 'Now the Haitians have got all the jobs. They're willing to do anything for $20 a day. Now, all I can do is stand on the corner.' Officials say more than 3,500 Haitians have flooded this city of 18,000 residents -- 8,000 of them seasonal farm laborers -- in the past year. They have...crowded locals out of the hard-labor, low-pay farm jobs...Some observers say the Haitians have unwittingly created a 'slave trade' for farmers."1 *****(Champaign-Urbana News Gazette, Nov. 29, 1980, p. A- 7) Anti-immigration organizations: [I]f 2 million immigrants settle in the United States every year, the unemployment rate could rise to 15.2% by 1990 and 19.1% by 2000. If immigration is within the range of 1.1 to 1.6 million annually, the unemployment rate could range from 12.2% to 13.9% by 1990, and from 12.8% to 16.4% by 2000. Such high levels of immigration will create unemployment rates that could seriously damage our economy and standard of living. (The Data, The Environmental Fund, Nov, 1983, p. 2) If illegal immigrants were not taking so many jobs in America, unemployment could be brought under 4% according to former Secretary of Labor, Ray Marshall -- saving the American taxpayer millions that now go for unemployment compensation and other assistance programs. (Roger Conner, fundraising letter from FAIR, no date). TEF [The Environment Fund] had ample opportunity in September to explain the effects of unchecked immigraton on the U. S. labor force. In an unprecedented 15 sessions on Capitol Hill, TEF ran hourly briefings for congressional staffers on three successive days using our color-graphic computer presention, QUIC DATA. With QUIC DATA projections, TEF was able to demonstrate that unemployment may rise to 14.7% by the year 1990, if present immigration and economic trends Those that participated in these briefings were unanimous in their praise of TEF's comprehensive analysis of the effects of continuing massive immigration on the U. S. unemployment problem. (The Other Side, October, 1983, No. 34, p.1).2 ***** The governor of Colorado: [I]t is folly to import more labor. We are the only country in the world that imports a generation of poor people every year. The average American will earn less in real terms this year than a decade ago. Yet, we accept twice as many immigrants as the rest of the world combined...With 11 million citizens unemployed, this is demographic and economic insanity. (Richard Lamm, 1983) In Germany: About 50% of the West Germans surveyed worried that the East Germans are crowding the job market. (Wall Street Journal, Nov. 20, 1985, p.1). Unions cite job loss when they argue for a policy of fewer immigrants.3 The "official AFL-CIO spokesperson on matters ****concerning immigration and refugee policy" has written as follows: The AFL-CIO opposes any program that would permit importation of foreign labor to undercut U. S. wages and working conditions...abuses of our immigration laws and the lack of essential protections against displacement for American workers...we do not favor the new [higher] levels of legal immigration which are part of the [Select] Commission's recommendations. (Otero, 1982) And during 1982 Congressional discussion of the Simpson- Mazzoli bill Ray Denison, chief lobbyist for the AFL-CIO, disclosed that the labor federation will oppose the bill unless it's assured that the Reagan administration and a majority of the House endorse proposed restrictions on U. S. hiring of foreigners as temporary workers...the AFL-CIO long has been the prime backer of the bill's main element, which would make the hiring of illegal aliens by U. S. employers a federal offense. (The Wall Street Journal, December 2, 1982, p.7) This comment comes from an Australian sociologist: In present conditions migration will not lead to greatly increased spending (and thus job creation) because there is a limited pool of jobs in Australia, and migrants are winning jobs from the pool at the expense of Australian job-seekers (Birrell,1982a)...I certainly think that the 25,000 New Zealanders who settle in Australia each year and come in virtually unchecked ought to be looked at. New Zealand is virtually exporting its unemployment problem to us. (Birrell, 1982b) And the complaint of job displacement is an old one, probably as old as hired labor. John Toland wrote as follows in England in 17l4 (p. 14): The vulgar, I confess, are seldom pleas'd in an country with the coming in of Foreners among em...from their grudging at more persons sharing the same trades or busines with them, which they call taking the bread out of their mouths, p.14. As with popular accounts, the basis of professional discussion of the displacement issue frequently is simply that "everyone knows" displacement must occur. For example, Greenwood, one of the relatively few economists who has specialized in immigration, writes that "we can probably conclude with some assurance that New York, Los Angeles and Chicago [areas with high immigrant concentrations and highly developed industrial economies] experience labor market impacts due to immigration that are absolutely somewhat greater than those experienced elsewhere. The relative impacts are also likely to be fairly sizable compared to those on most other cities." (Select Commission Staff Report, 1981, p. 276). But this "conclusion" is no more than supposition, and hardly a reasonable basis for important national decisions. Instead of supposition we need persuasive theory and/or reliable empirical evidence. The displacement argument is often made with anecdotes about persons who formerly held jobs now being filled by immigrants. These anecdotes are undoubtedly painful. But when reading such anecdotes, it would seem appropriate also to review in one's mind another sort of anecdote, that of the useful and productive immigrants one has known working in a variety of fields. In my own case, a quick survey turns up: C., a Cuban refugee and remarkable mathematician-physicist-computer expert whom I met when he was a 19 year-old student, just finishing a master's degree, who after two weeks as a hired research assistant turned into a full-fledged colleague and co-author. N., my older son's Tae Kwon Do (Korean Karate) teacher who had a remarkable effect upon my son's personality and confidence as he went from being a novice to a black belt; Dr. C., our children's Irish pediatrician from first birth for the next 18 years. I find it hard to imagine that these persons "displaced" native Americans. And all of them became employers making jobs for natives. Two more anecdotes illustrate the same point: Item: GEORGIA HIRES WEST GERMANS FOR SCIENCE AND MATH Atlanta -- Georgia education officials believe they may have found a way to ease the shortage of science and mathematics teachers that plagues high schools in this state and throught the country: important West Germans." West Germany has a surplus of such teachers. And the first ones who have come have been a considerable success. (The New York Times, September 24, 1984, p. A14.) Item: Immigrants from Afghanistan seem to have opened fried-chicken restaurants in every corner of [New York]. Some 10,000 to 15,000 refugees have arrived here since Soviet troops marched into their country in December 1979, and there are already about 110 Afghan fast food shops in the New York area. They bear such popular all-American names as Boston Fried Chicken, Harlem Fried Chicken and Texas Fried Chicken. Afghans have also opened restaurants in New Jersey, Pennsylvania, Washington, D.C., and most recently California, for a total of about 200 chicken places in the country. Although many of the latest Afghan arrivals have entered the fried-chicken business, the man who started it all and some of the major chicken entrepeneurs [sic] from Afghanistan arrived before the Soviet invasion. Taeb Zia, who is known in Afghan circles as the father of Kennedy Fried Chicken, studied engineering in Baku in the Soviet Union. "But I hated Russia and didn't like the factory work," he recalled the other day. He came to the United States in 1972 and got a job with Kansas Fried Chicken through an employment agency. Three years later he had earned enough to buy the franchise. "Then I saw I could do chicken just as well and 15 to 20 percent cheaper with my own recipe and spices," Mr. Zia said. Today he runs six Kennedy Fried Chicken Shops in Manhattan, Queens and Brooklyn, employing mainly Afghan refugees and providing technical assistance to many other aspiring Afghan entrepeneurs (sic). His chain is named after the American President "just because Afghans like him a lot." The Afghan incursion into chicken is a direct result of Mr. Zia's involvement. In a kind of pyramid effect, he hired Afghan refugees who worked in his shops and then went on to open their own. Although many Afghans arrived in the United States with language problems and little experience, they learned from him. "We don't know what to do, and then we see Zia and learn how an Afghan can do business here," said Abdul Mosaver, who owns two successful New York Fried Chicken stores in Queens and one in Brooklyn, employing a total of 50 people, mostly Afghans." (The New York Times, October 7, 1984, p. 47.) It would seem that at least some displacement and some reduction in wages must immediately occur when potential workers in a particular occupation are added to the labor force, whether those potential workers be laborers or physicians, unless the immigration has been expected and allowed for. Even if immigrants' occupations are in the same proportion as the distribution of occupations in the economy-- and they are not too far from that, as Chapter 3 shows--one might expect frictional increases in unemployment for at least some period of time. But expectations of the arrival of immigrants may generate new jobs in advance of the immigrants' arrival that offset the frictional competitive effect. And Harrison's analysis of the imbalance between the demand for goods and the supply of job seekers also works against frictional effects, as discussed below. In the years before World War I when the U.S. did not restrict general immigration, and also in the 1930's, there was a natural check on job displacement. Immigrants already in the U.S. often financed the transportation and settling expenses of their relatives in the "old country," as Isaacs noted then (p. 34). And when times were hard, relatives furnished less such support, hence reducing new entries to the labor force when there already was unemployment. But now that quotas are always filled, this check is absent. This chapter presents some new theory that goes beyond the standard considerations. Empirical estimates of unemployment and wage effects are found in the following chapter. CONCEPTUAL FRAMEWORK The idea underlying fear of job displacement is simple: If the number of jobs is fixed, and immigrants occupy some jobs, there are fewer available jobs for natives. This apparently- obvious idea is found in several of the quotations adduced above, and it is embodied in The Environmental Fund's QUIC DATA model, as seen in Figure 11-1. The model -- billed as "recent studies by TEF" (TEF News, Sep. 21, 1983, p. 1) -- is nothing more than an estimate of the number of jobs that would exist under conditions of zero immigration, followed by an assumption that every immigrant will cause one person to be unemployed. The calculated estimates of the unemployment rate at various immigration levels are simply the total number of persons (natives plus immigrants) divided by the number of jobs previously assumed on the basis of no immigrants. For example, "Figure [11-1] shows that, if 2 million immigrants settle in the United States every year, the unemployment rate could rise to 15.2% by 1990 and 19.1% by 2000." TEF offers as a disclaimer, "Note that these graphs are projections, not predictions. In all likelihood the pressure for high unemployment would trigger measures to prevent joblessness from rising to the levels shown above." (TEF DATA, November, 1983, p. 1, italics in original) In other words, TEF's assertions about the unemployment rates that will prevail at various immigration levels are without meaning, having -- by their own admission -- no connection with any expected reality, but are, rather, explicitly counterfactuals conditioned on the absence of events predicted to occur.4 This is misleading. **** ----------- Figure 11-1 ----------- The opposite over-simplified view of the matter is that the supply of jobs expands instantly and without limit in response to market changes in wages, the way the supply of purchasers of vegetables seems to expand just before closing time in a farmers' market as prices are knocked down by sellers, so that no vegetables go unsold. In the 1930's, people concluded that this model does not describe the labor market; because it was discredited, its core insight was lost. An adequate theory of the unemployment effect of immigration -- or of school leavers or additional women entering the labor force, for that matter -- must be more complex. It must treat both the wage response and the unemployment response, as well as the internal migration response that Manson, Espenshade, and Muller (1985) call to our attention and (reminding us of the work of Eldridge and Thomas, 1964) documents in the case of Hispanic immigration into California. They show that, unlike the western region of the U.S. as a whole, net im-migration to California declined between 1965-1970 and 1975-1980 (p. 17). Furthermore, they found that "the inflow between 1970 and 1983 was weighted toward more-skilled professional and nonprofessional workers, and that the outflow was largely comprised of less-skilled workers" (p. 21), the latter the type of worker the Hispanic immigrants most strongly compete with. Satisfactorily complex theory is not at hand. But two pieces of theory may advance our understanding, and help us have narrower bounds for the effect than the poles of the two above-mentioned over-simplified notions. The first new piece of theory, that of Harrison (1984), ingeniously analyzes how exogenous increases in demand caused by immigrants consumption affect the native and immigrant unemployment rates differently, and can reduce native unemployment even as immigrant unemployment rises. The second, newly presented here (though in brief) looks at the job market competition as a problem in queue theory, and some available parameters suggest upper bounds for the damage done to native unemployment.4.1 ***** Harrison's Theory of the Effect of Exogenous Immigrant Consumption 99 ------------------------- 99 The following formalization of Harrison's analysis was done jointly with Yew-Kwang Ng, and will appear as an article entitled "THE EFFECT OF IMMIGRATION UPON NATIVE UNEMPLOYMENT: A FORMALIZATION OF HARRISON'S ANALYSIS" by Julian L. Simon and Yew-Kwang Ng.------------------------- Harrison (1984) has contributed a quite new idea to the understanding of the effects of immigrants on native unemployment. He notes that an immigrant increases the demand for goods and services immediately upon arrival, and hence increases the demand for labor independently of starting work. If immigrant consumption is assumed (for now only) equal to the average native or family, the demand for labor goes up by one full job, and hence by the proportion 1/L that the immigrant bears to the labor force as a whole. The key point of the analysis is that the immigrant then has a greater than 1/L chance of being unemployed because most of the employed labor force does not change jobs each year. This would decrease native unemployment under conditions of wage fixity and an unchanged overall unemployment rate. Harrison illustrates his argument with numerical examples. Simple formalization seems to bring out his point more clearly, and allows us to better analyse its implications with different sets of parameters. We assumed that the ratio of workers to consumers is the same among natives and immigrants, and other conditions at large remain the same. Let Un = the number of natives unemployed in the absence of the group of immigrants whose effect is being analysed Un' = the number of natives unemployed in the presence of the group of immigrants En = the number of natives employed pre-immigrants En' = the number of natives employed post-immigrants M = the number of immigrants in the cohorts whose effect is being analysed s = the proportion of natives who leave their jobs each year and seek new jobs, i. e. the job turnover rate d = average immigrant's spending for consumption relative to average native spending for consumption, immediately upon arrival, a fraction likely to be unity or less, which may be thought of as the fraction of a new job created by an immigrant. a = the relative likelihoods of an immigrant and a native of being hired for a particular job opening, i. e. an indicator of the relative job-getting success rates. aM = the "effective" number of job-seekers in the cohort immigrants. We begin with the number of natives unemployed following upon the entry of the immigrants whose effect is being analysed ("immigration" hereafter). ( sEn + Un ) (1) U'native = Un + sEn - (sEn + dM) (-------------) (sEn + Un + aM) The first two terms on the r.h.s. of (1) are, respectively, the numbers of natives originally unemployed, and job turnovers. From these are subtracted the sum of the jobs open due to turnover and to the increased demand due to the immigration (sEn + dM), multiplied by the ratio of the number of natives seeking work (sEn + Un) to the sum of the natives seeking work plus the "effective" number of immigrants seeking work (sE + Un + aM). Subtracting the "pre" native unemployment from the "post" native unemployment (a - d) sEnM - dMUn (2) U'n - Un = ------------------- sEn + Un + aM For native immigration to fall, that is, for U` - U to be negative, it is necessary and sufficient that (3) (a - d) sEn - dUn < O , or d(sEn + Un) > asEn which is necessarily satisfied (but not only then) if Un > 0 and d > a. That is, as long as there is any native unemployment it will fall due to immigration if an immigrant's consumption bears a higher proportion to native consumption than an immigrant's propensity to find a job bears to an native's propensity to find a job in the same market. It is interesting to note that the job-turnover does not enter into these conditions; it could even be zero. In fact, the situation is easiest to envisage with a zero turnovear rate; if turnover is zero, in the absence of immigrants there are no jobs for unemployed natives to find. But the coming of immigrants creates some new openings due to increased demand, and hence some unemployed natives can find jobs. Re-arranging we get sEn d (4) ________ < - Un + sEn a which shows that the conditions for native unemployment to fall are quite weak with realistic parameters. For example, if the native unemployment rate and the turnover rate are about the same (say 8%), then d need only be half as large as a for native unemployment to fall; if a is .8, d could be as low as .4. The analysis has not referred to immigrant unemployment. It obviously will be higher than native unemployment at first even if a = 1, i. e. the job-finding propensity is the same for both, because some (most) natives have jobs whereas no immigants do, and more immigrants are hunting for jobs. The relationship of the immigrant and native rates obviously is heavily dependent upon the turnover rate. Estimates can easily be derived with various numerical assumptions. When a higher rate among immigrants than among natives is indeed observed -- as has been the case in Australia recently (see Chapter 4) -- it should not be viewed as a sign of added pressure upon natives, but rather the opposite; it may well be the sign of increased job opportunities and diminished unemployment for natives. Harrison's analysis is not affected by whether the immigrants do or do not receive financial assistance from the government prior to finding employment. But the economic welfare implications are different if the financing comes from private immigrant funds or from government assistance. Harrison models new immigrants like a Marxian "reserve army of the unemployed." This may or may not be an ethically desirable situation. But the immigrants clearly choose this period of unemployment as part of their long-term investment in migration. And native "workers" may be helped thereby. (In contrast, Marx's "reserve army" was for the benefit of the capitalists.) Even if the actual parameters are such that Harrison's mechanism by itself does not produce an on-balance employment advantage for natives, the mechanism must work to substantially reduce the immigrant-caused increase in waiting period for natives seeking jobs that is modeled in the analysis in the next section. Together, the two lines of analysis together suggest that the unemployment burden upon natives is small even if it exists at all. A Queue-Theoretic Analysis of the Effect on Natives of Added Congestion in Job Search Caused by Immigrants The aim of the theory newly offered in this section is to estimate the added time waiting for a job by natives who are unemployed, due to additional competition by immigrants. The analysis begins with two extreme hypotheses about the main forces which work in opposite directions, and then examines other influences which keep these forces from continuing to work to their extremes. The first force is momentary job fixity and the associated queuing effect. To make the process concrete, I suggest that the main causes of delay in finding a job are, first, the time needed to search for job openings and to apply for them, and second, the time spent in waiting for several job candidates to be evaluated and one among them to be selected and hired. To some extent these waiting periods can be concurrent, but they will also be sequential to some extent. The more applicants there are for jobs, the more of a given persons's applications will be unsuccessful, and the larger the total time waiting to be selected. (This phenomenon may be thought of as a congestion effect, though there is no reason to believe that the increase in waiting time is non-linear with the number of competitors for jobs.) If the number of jobs really were fixed and the persons in the queue are homogeneous and wait their turn the way taxis queue up at the airport, a single immigrant added to the queue would impose an infinitely large total added waiting time upon the aggregate of other people now living and to be born in the future. That is, if the person in the queue behind the immigrant has to wait an additional x hours for a job, the second person will also have to wait an additional x hours, as well as the third and nth person. If the number of persons who will eventually come is N, the total added waiting time is N multiplied by x, and if N is infinite, Nx is infinite. For this not to be the case -- that is, for a single immigrant not to impose a very large waiting-time burden upon natives -- the number of jobs must expand as the queue grows larger. Of course this is what does happen. And therefore, we must inquire what causes the number of jobs to increase, and how fast they increase in response to the queue size and to other possible influences. The opposed force that expands jobs also has an infinitely great impact in abstract unmodified form. Imagine an immigrant who by persuasive skill induces a business which would not otherwise have hired anyone to hire the immigrant. This might be because the immigrant has skills that the employer wants but cannot find among natives now looking for work. Or the additional hire might be due to the immigrant offering to work at a wage lower than prevailing, thereby converting the employer's calculation into a positive reckoning for the hiring whereas it was negative at the prevailing wage. (That is, assume that the employer was just at the margin of decision before, and the lower wage tips the balance.) Or it might be that this immigrant is just a very good salesperson and talks the employer into the hiring. So the labor force is thereby increased by one person. But we must also trace the longer-run consequences of this occurrence. After an immigrant finds such an "added" job, and after a period of adjustment to be discussed below, natives who are seeking employment are no worse off than if the immigrant had not taken that job. Almost as soon as the immigrant begins work, s/he spends his/her earnings either on consumption items or on portfolio investment which is then spent on producer goods.5 ***This spending may be seen as complementary to the pre-earnings consumption that drives Harrison's analysis discussed above. We may expect that in a short space of time, these added earnings will produce roughly one full job someplace else in the economy; this increase in employment must take place unless there are economies of scale working to reduce the effect6. And though **this adjustment does not happen instantaneously, it should happen sufficiently quickly so that for many purposes we may safely ignore the time lag. (And we should keep in mind that if the entry of the immigrants into the economy is anticipated by managers, on the basis of normal flows, there need not be any lag at all.) At the bottom of popular thinking about immigration is a very different view of the aggregate production function, that no additional jobs are created when additional persons begin working and spend their earnings (e. g. The Environmental Fund's "model" discussed earlier). The implicit assumption of such a view is that the economy somehow has the capacity to produce more goods without more labor, and therefore there will be no increase in jobs when the demand for goods increases. And we should not be surprised that people hold such a view; I confess to slipping into this sort of thinking myself when I am not careful. The intellectual difficulty is that the process is very indirect and very diffuse; though it is easy to imagine in one's mind's eye an immigrant sitting down before a machine and taking a place that a native might otherwise occupy, one cannot so easily picture in one's imagination the small effect on the labor force in the plant that manufactures the refrigerator that the immigrant buys, at the trucking firms that move the refrigerator along the channels of distribution, in the wholesaler's sales force or front office, in the number of sales clerks in retail stores, and so on. Each of these increases is far less than a single person, and in almost all of these places of work an additional person is not hired in response to the immigrant's earnings. But in one place the additional volume of business will be enough to cause a manager to move from a decision that previously was just marginally not to hire an additional person, to a decision to hire an additional person. And the next immigrant's spending will cause this sort of response someplace else. This process is very hard to visualize, and would be even harder to pin down if one should try to trace the linkage organizationally. Yet on average it must occur. The process may become clearer by specifying more exactly what is meant by an "open job." A "job opening" is not a physical place, but rather a set of marks on a personnel recruiter's docket indicating that a worker is being sought. In contrast, an "existing job" may safely be thought of as a physical work place and an accompanying pay check regularly being issued. Given the state of economic activity at any given moment (as measured, perhaps by GNP), some firms will desire to hire additional people; and these desires are indicated on internal documents and perhaps by advertisements. Now imagine that one person comes unexpectedly into the society, somehow causes there to be an additional "open job," and manages to get hired. When the person begins working, economic activity becomes greater than before the person began working. The signals emanating from this change in economic activity, between the situation before and the situation after the person begins working, induce a change in plans at some other firm elsewhere in the economy, based on the expectations of future sales and profits that have been revised in light of the change. This revision of business expectations results in a new set of marks on one personnel recruiter's papers somewhere in the economy, which constitutes another "open job" that did not exist before this train of events began. This analysis in terms of signals and expectations, rather than in terms of workplaces connected with physical capital, should make clear that the process is not simply a case of an immigrant filling an "existing" job. Rather, when an immigrant manages to induce a new slot to be opened, there then occurs a series of changes in the demand for labor elsewhere in the economy as a result of the immigrant coming to work and collecting a paycheck. Admittedly, this is a rather slippery idea. At first thought it may seem as if the newly-hired immigrant is only producing goods to exchange for the goods that s/he will then purchase and consume, without there being any other changes. But at a minimum, it should be obvious that there must be a change in the total production of some goods following upon the newly-hired immigrant starting to work, and also a change in the mix of goods, because the immigrant produces only one kind of goods. Therefore, if the immigrant does not produce shoes but does wear them, makers of shoes will produce more shoes than before the immigrant goes to work. This change may result in a change in the hiring plans of a shoe manufacturer--or a hat manufacturer, or a violin academy. The only case in which the immigrant would not affect anyone else is if the immigrant becomes a subsistence farmer. Contrast the above sequence of events with the sequence which begins with a native being laid off from an "existing job" which then is not refilled because of lack of demand for the product being produced, continuing with the event of the laid- off native or an immigrant finding what is then an "open job" elsewhere in the economy and going to work there, at which time that open job becomes an "existing job." There will be no change in total business activity in the economy (aside from any change flowing from the change in mix of output and a difference in salary), and no new demand for an additional worker elsewhere in the economy, following upon this sequence of events which is different from the case we are discussing. The key element in the process under discussion is that the change in economic activity which occurs as soon as the immigrant begins working at the newly-created job sends a signal out through the economy which then induces a change in expectations. So, another employee is hired somewhere on account of this immigrant. But there is more to come. On account of the second person's spending, a third person is hired somewhere. And on and on indefinitely. This implies that a newly-created job opening filled by an immigrant leads to an infinite number of other people also being hired as a consequence. Just as we noted earlier that the waiting time for jobs caused to natives by the additional immigrant does not accumulate indefinitely, this process also cannot continue indefinitely, if only because at some point there are no more persons to hire (aside from the "frictional" unemployed); various influences must come into play so that the force of the immigrant's hiring eventually plays itself out.7 Also, the process does not take **place instantaneously; each hire takes time to generate another job. Toland well understood this process back in 1714: We deny not that there will thus be more taylors and shoomakers; but there will also be more suits and shoos made than before. If there were more weavers, watchmakers, and other artificers, we can for this reason export more cloth, watches, and more of all other commodities than formerly: and not only have 'em better made by the emulation of so many workmen, of such different Nations; but likewise have 'em quicker sold off, for being cheapter wrought than thoise of others who come to the same market. (1714/- , p.14). One might wonder whether the same long chain of additional hirings takes place when a native replaces another native on a job, and if not, why not. The distinction between the "new" job which the immigrant (or anyone else) causes to be open, and an "old" job which was formerly filled by one person who leaves or is fired and then is filled by another person, is, of course, artificial. In both cases, employers make calculations about the profitability of making a new hire. But where a person was already working and spending money, it is reasonable to think that the adjustments in the rest of the system consequent on the start of that job have already occurred, and there is therefore no new impetus to additional economic activity when a replacement is hired. (Of course if there is a lengthy hiatus between the first person leaving the job and the second beginning on it, there would be a shrinking of economic activity and then a re-expansion, and the re-expansion would be analogous to the situation of the immigrant described above.) It may be easier to visualize the process if one imagines the reverse process -- the United States after a random half of the working population were suddenly removed. For a while, workers would have larger incomes due to the larger supply of capital to work with, and due to the lack of necessity to build new capital for a while. But after the capital effect peters out after a few years, nothing would be different than if the halving of the work force had not taken place, and incomes would be the same as otherwise (though lower because of productivity effects, as discussed in Chapter 8). There would be no increase in employment. And so it is with immigration except in reverse. We can make concrete the two opposing effects with the device of a group of immigrants. Assume there are five in the cohort -- Able, Baker, Charlie, Dog, and Easy. Able "opens up" a new job the first day, a job that previously did not exist and that Able brought about with a combination of persuasion and a low salary. This sets in motion the sequence that produces jobs for other persons, natives or immigrants, though we do not know how many additional persons or how long it takes for them to get hired, one after the other. The other four immigrants -- Baker to Easy -- do not find work immediately and stand in the queue searching and waiting for jobs, and thereby making it harder for natives to find jobs. Again, just how much extra waiting time they cause for natives is not known, nor do we have any way of estimating it. An illustrative calculation may render the matter more understandable. For simplicity, assume that all hiring takes place discretely at ten-week intervals. At the moment when Baker or Easy get jobs after ten weeks, they cause four persons who would have gotten jobs then to wait an additional ten weeks for jobs. Now consider that Albe generates another new job ten weeks after he begins work -- a period equal to the average length of a spell of unemployment in the United States (see old paper **** ). This removes one person from the queue. And at the twenty-week mark, three workers who would otherwise have gotten jobs now have to wait an extra ten weeks, having been pushed back by the sequence of the four immigrants but diminished by the one new job created. After another ten weeks the sequence begun by Albe generates a second job, and now two workers who would otherwise have gotten jobs then have to wait an extra ten weeks. And at the thirty-week mark there is one worker who has to wait an extra ten weeks. So the combination of Baker to Easy standing in the queue, and Albe starting a sequence of new job creation, results in 4+3+2+1=10 periods of ten weeks, or 100 weeks in total, of addtional unemployment, or 20 weeks per immigrant -- up to this point. The positive effects of the job-creation sequence continue, however, and at some time the continuing reductions in time unemployed will completely offset the 100 weeks in increased unemployment time for natives. This process theoretically would continue forever, or until prevented from further continuation by the economy running out of available workers or by leakages of one sort or another. And of course we must recognize that the bad effects of this sequence occur earlier than the good effects, and discounting would therefore make the overall assessment somewhat more negative than it appears in this scenario. Wage decrease in response to the increased supply of labor in particular categories has, however, been omitted from the story until now, and it exerts a powerful force in mitigation of that increase. Though those persons steeped in the Keynesian analysis of unemployment are predisposed to assume that wages are rigid with respect to downward movement, an increase in immigrants (and in the size of an age cohort, which is quite comparable in this respect) results in significant wage change, evidence which is reviewed in Chapter 12. (With respect to studies outside of the field of immigration, the review is selective rather than comprehensive, but this should be acceptable because the main aim is simply to communicate the main point of this literature rather than to critically assess the body of work.) Perhaps the best estimate for our purposes is that of Morgan and Gardner (198?, pp. 398-399; see next chapter for details), even though it is short-run in nature and limited to agriculture in a few states. They find an elasticity of -3.2 between employment and wages (stated in this fashion because the causation must be mutual and simultaneous). And they find that while 210,000 additional persons were added to the work rolls as the wage declined by 9%, only 51,000 natives failed to find jobs who would otherwise have worked in the industry (and of course some portion of these persons moved on to jobs in other industries rather than remaining unemployed. And when considering a single industry, new jobs are not created by way of the income spent by newly-hired workers, so the unemployment estimate is not biased downwards. The Morgan-Gardner results imply that one immigrant may have to suffer spells of unemployment for each four immigrants who find work immediately. If this is so, the economy-wide job- creation effects following on their spending of their incomes should operate so strongly that the total unemployment caused to natives by the coming of the immigrants should be negligible. It should be remembered that a reduction in wages in a particular sector due to immigration does not mean an overall loss of wage income to natives. Rather, most of what is lost to one sector is likely to be gained in wages by another sector, and not resulting in an increase in capital's share, simply because the wage share of total income is so large. It should be remembered, too, that the relative decline in wages is by no means always in the poorest sectors; in recent years, physicians have probably suffered a greater relative decline in income due to immigration than any other occupation. SUMMARY The two analyses in this chapter -- Harrison's, and the queue theory newly presented here -- are complementary. Harrison focuses upon the exogenous consumption effect, and my analysis focuses upon the extent of additional labor-market "congestion" caused by additional competitors for jobs. Both analyses suggest that the effect upon natives' unemployment is much less than common belief has it, and may even be lessened rather than increased native unemployment due to immigrants. The next chapter describes empirical analyses that estimate the size of the unemployment effect, as well as the related wage effect. 86-85 Displ11 12/11/87 FOOTNOTES 1The importation of sugar workers nicely illustrates how we get ourselves into an unncessary ethical difficulty in one market by attempting to manipulate another market for the gain of a few. The only reason that the sugar industry still exists in the U.S. is the scheme of sugar quotas to protect the U.S. growers at the expense of growers elsewhere; without quotas, there would be no foreign workers whose entry one could oppose on grounds that we are "exploiting" them. On the other hand, abolishing the quotas would mean that those who now come to work in the U.S. would earn less at home, either in sugar or in other industries. It seems to me that objecting to the importation of sugar cane workers is not a useful place to begin a reform campaign -- unless one is only using it as a smokescreen for opposition to guest workers. 2It should be noted that the QUIC DATA "model" is simply a single equation specifying a fixed number of jobs, which therefore produces apparent unemployment by simple arithmetic whenever the labor force is increased. More about this below. 3The usual reaction is a call for restriction of immigration. An interesting example of a response to fear of displacement more subtle than quotas is the set of state restrictions on women's work, which, as Landes ( ) shows, tended to be enacted in states with heavy immigration. 4 It is of interest that TEF implicitly assumes it is government which creates jobs -- "high unemployment would trigger measures," presumably government measures. This language implies lack of understanding that it is markets for goods and labor that lead to jobs, and that historically the process is roughly automatic, though not instantly responsive. From the TEF job-fixity way of thinking inevitably follows the belief that immigrants cause unemployment. 4.1Warren (1983), following on Hughes (1975, reference from Warren), discussed still another possible influence, an increase in job-matching efficiency caused by a larger labor force swelled by immigrants. Warren does not find evidence for this hypothesis in the Australian data, and it does not seem to me of likely importance, because a large increase in the labor force would seem necessary to make a small improvement in matching efficiency. 5The notion that some of the earnings disappear in gold coins under the mattress has little reality these days. If the immigrant remits some of the earnings abroad, the effect is much the same as domestic purchases, because those dollars eventually buy imports from the United States, though remittances sent abroad may slow the adjustment process a bit; see Chapter 2 for more discussion. 6The only fashion in which the economy could provide the additional goods and services bought with the immigrant's earnings without there being the equivalent of an additional person being hired throughout the economy is if there are economies of scale so that the cost in labor necessary for the creation of the incremental goods and services would be less than for the same amount of goods at lower levels of production. And even if there are some economies of scale, it is inconceivable that on average it might take less than, say, 90% of an additional person's labor -- or 80% at the absolute outside -- to provide the additional goods and services. There is nothing in the literature on industrial organization to suggest economies of scale of any such magnitude throughout the economy. A side-benefit of the additional person joining the economy, as compared to not immigrating, is that the new job is likely to be in the most dynamic part of the economy; this is one of the advantages of any labor-force growth. 7This description is related to the concept of the multiplier in regional analysis. But it is not clear whether there is anything worthwhile about that concept. Some may worry that this analysis is faulty because it constitutes an infinite process. The Keynesian less-than-full- employment multiplier also raised such a question until the additional idea of "leakage" was introduced. (This analogy should not be interpreted as an endorsement of Keynesian analysis or policies, of course). 86-85 1/16/87