However, as the case of British doctors illustrates, the immigration safety valve may, over the long term, weaken the incentive to address the root of the problem, which is training.
The effect of immigration on the training of indigenous workers has not, to my knowledge, been adequately investigated. Recall that in Europe indigenous workers have benefited because immigrants are skilled: unskilled indigenes can work along with skilled immigrants. But while this is directly good news for the unskilled, indirectly it may not be. The training of young workers in skills depends upon firms choosing to invest in training. Because training is costly and workers once trained can leave for other firms, the most profitable strategy for an individual firm is to poach those who are already skilled off other firms. Because poaching is a zero-sum game, industry associations sometimes try to organize a common commitment
THE ECONOMIC CONSEQUENCES 127
to training, policed by peer pressure. All firms in an industry accept the need to do their share of training. However, all outcomes that are dependent upon coordination are potentially fragile. A shock could break the pattern. An influx of trained immigrants could constitute such a shock, destabilizing industry-wide training. With an influx of trained immigrants, hiring already-trained workers temporarily ceases to be a zero-sum game because they do not have to be poached from other firms. Even if training programs collapse, firms may still in the aggregate gain because they now get trained workers without the costs of training. But young indigenes lose because firms are no longer bothering to invest in training them.
Whether this effect is empirically important has not been
researched, but prima facie it may have occurred in Britain. There has been a collapse in firm-based training: most notably, many apprenticeship schemes have been abandoned. The retreat from
training youth was broadly coincident with the rise in immigration: during the peak years 80 percent of new jobs were filled by immigrants, but whether it was cause, coincidence, or consequence is an open question. Either way, once lost, industry-wide apprenticeship schemes are difficult to revive because of the costs of coordination.
What is good for business is not necessarily good for indigenous people. The short-term interest of business is for the open door: it is cheaper to recruit already-skilled migrants than to train indigenous youth, and the pool of talent will be wider when the door is more open. It is in the interest of the indigenous population to force firms that want to benefit from the country’s social model to train its youth and hire its workers. Germany stands as testimony that such a policy need not drive business abroad. But the divergence of interest between business and citizens should make people skeptical of its pronouncements on migration policy. Most weeks I see letters to newspapers signed by some CEO fulminating against
128 HOST SOCIETIES: WELCOME OR RESENTMENT?
restrictions on migration. If they need skilled workers, why don’t they train them? Their portentous pronouncements are but pallid variants of the grandiloquent “What’s good for General Motors is good for the country.”
Does Immigration Induce Emigration?
British policy toward migration is currently defined in terms of the net movement of people: the target is set as immigration minus emigration. For some long-term purposes this is the right way to define policy. If the objective is to preserve the ratio of open space per inhabitant, then net migration should be set to zero or there-abouts, depending upon birthrates. But for other purposes emigration and immigration need to be considered separately. In most high-income countries emigration as such has not been a policy concern. Yet recent evidence suggests that for European countries
emigration inflicts serious losses on the remaining population. 11
Emigrants tend to be more skilled than the average population and are attracted to high-wage, fast-growth countries such as the United States and Australia. Is there any reason to think that immigration accelerates emigration?
Within a standard stylized economic model of migration, points systems that determine eligibility for the right to migrate create a direct link between immigration and emigration. Recall that a standard feature of points systems is to privilege relatives of the diaspora. The past history of global migration consequently gives Europeans much easier access to the United States, Canada, Australia, and New Zealand than the citizens of low-wage countries. To see how this plays out, consider a three-country world: I will give the countries real names, for ease of recall, but they are not meant as actual countries: they are constructs with assumed characteristics.
THE ECONOMIC CONSEQUENCES 129
Countries A (America) and B (Britain) are identical, high-wage economies, while country C (Chad) is a low-wage economy. America permits migration by British citizens, but not by those from Chad.
Britain now adopts an open-door migration policy for the citizens of Chad. The result is that citizens of Chad move to Britain, in the process driving wages slightly down. The small decline in wages is not sufficient to deter continued migration from Chad: the gains to migration remain massive. But now British citizens have an economic incentive to migrate to America. The mechanism that drives emigration from Britain in this simple model is the decline in wages, and this is something that we know does not happen to any significant extent in actual migration. However, that wages do not decline need not imply that living standards are not reduced. For example, as a city becomes more crowded, gains in wages are offset by rising congestion. Over half of the current population of London is immigrant, yet the population of London today is the same as it was in the 1950s when almost the entire population was indigenous. It is not credible that in the absence of migration the population of London would have halved, so the only reasonable interpretation is that immigration has induced the indigenous population to leave London. Where did these people go? Many of them just moved to
the outer suburbs. 12 However, both Britain and the Netherlands are currently experiencing high emigration coincident with high immigration. Whether there is a causal link between the two has not been studied.
A mechanism by which immigration is likely to drive emigration of the indigenous is the cycle of boom and bust. International flows of both capital and labor amplify booms and thereby inadvertently also amplify the busts that follow. In the 1990s capital inflows to East Asia led to the bust of 1998—the East Asian Crisis. Analogously, open-door migration policies amplified the booms of 1997–2007
130 HOST SOCIETIES: WELCOME OR RESENTMENT?
in the United States, Ireland, Britain, and Spain. At the time, politicians such as Gordon Brown claimed that they had abolished the boom-bust cycle. What they had actually done was to intensify it by enabling booms to run for longer: immigration enabled both public and private overspending to continue without triggering the inflation that had previously compelled governments to rein booms in.
The legacy was the superbust of 2008. Immigration did not cause the boom-bust cycle, but just as with international capital flows, it amplified the cycle, thereby deepening the bust. During the bust new hiring collapsed, implying very high unemployment rates for young people entering the labor market. For example, in Spain youth unemployment is currently around 50 percent. There could be no mechanism whereby migrants in employment vacated their
jobs in favor of new indigenous workers. Faced with unemploy-
ment, the indigenous young might reasonably decide to emigrate.
Whether newly unemployed immigrants decide to return to their country of origin depends upon the income gap between it and the host country and the ease of movement. Most of the immigrants to Spain came from Africa, where incomes were far lower, and entry into Spain had often been sufficiently difficult that a decision to return home might prove irreversible. So even being unemployed for some years in Spain might be a better option than leaving. In contrast, most of the immigrants to Ireland during the boom years were from eastern Europe. As a result, the income gap was narrower and migration easier, so th
at during the Irish bust many migrants returned home, easing the labor market adjustment. Nevertheless, by 2011 Ireland was experiencing its fastest rate of indigenous emigration since the nineteenth century. In Portugal, in response to the bust the problem of indigenous youth unemployment became so severe that the government has actively promoted emigration as an official policy. Immigration in the boom years
THE ECONOMIC CONSEQUENCES 131
inadvertently generated emigration of the indigenous in the years of recession.
Even if migration does induce emigration, does it matter? From any individualistic perspective, whether utilitarian or libertarian, such a voluntary relocation of the indigenous population is of no consequence. Indeed, if British citizens receive a capital gain on their houses as a result of immigration and this enables them to move to Spain, everyone has gained. The first best would be to remove all immigration controls, but the second best is to take advantage of national differences, shuffling people around the world according to their legal access to high-wage opportunities. If you feel uncomfortable with that conclusion, it is perhaps because you attach some value to the concept of a nation, beyond seeing it as a vehicle for the provision of individual opportunities. Emigration does not matter, beyond the economic effects noted above, as long as it does not fundamentally alter the composition of the population. But were the immigration-emigration link to become a powerful revolving door that transformed the population, it would surely become a matter of widespread concern. Just as if Chad emptied, there would be a global cultural loss, so too if we imagine that Icelanders moved to Norway as Iceland was repeopled with Chinese.
How, within a viable ethical framework, this might reasonably be seen as a loss I will return to in part 5.
The Economics of Guest Workers
We have now run through a substantial array of economic effects of migration. Both the narratives that immigration drives down wages for indigenous workers and that immigration is economically necessary are false. The truth is that moderate migration has economic effects on the indigenous population that in the short and medium
132 HOST SOCIETIES: WELCOME OR RESENTMENT?
term are marginal, and most probably modestly positive. Any long-term effects are negligible. In contrast, sustained rapid migration would most likely lower the living standards of most of the indigenous population, both through wage effects and through the need to share scarce public capital. So while controls on migration are important to protect living standards, moderate migration is modestly advantageous. If however, like the Japanese, the society wants to remain homogeneous, then the economic costs are sufficiently modest that it can afford to keep the door closed. After all, without any immigration Japan remains one of the richest societies in the world. In other words, the economic evidence suggests that economics should not be a very important criterion for determining immigration policy.
If not economics, then what should be the criteria? Evidently, the more uncertain, potentially adverse consequences for economic well-being are likely to come through the social effects discussed in chapter 3. There is only one way in which virtually all social effects can be avoided, leaving only the economic effects. That is if immigrants are prevented from integrating in any way into the society other than as workers: that is, in the German euphemism, “guest workers.” A genuine guest-worker program delivers the labor markets effect of migration and nothing else.
Some societies, most notably in the Middle East, have chosen to run very substantial guest-worker programs. Since these societies are small and rich, the attractions of such a migration policy to the indigenous population are substantial: they get others to do the work without the composition of the society being changed. Dubai has become a luxury service economy—only 2 percent of its income is now from oil—by this model. An astonishing 95 percent of the resident population of Dubai are immigrants: you might think that no society on earth could tolerate such an influx, but in Dubai
THE ECONOMIC CONSEQUENCES 133
immigrants even in such numbers are unthreatening because they cannot acquire citizenship or even the rights of residency. The stay of guest workers is conditional upon both their employment contract and their behavior. Their wages are unrelated to wage levels set for citizens and simply reflect the prevailing global markets in their level of skill. A visit to Dubai is a stark and unsettling reminder of global inequality precisely because, by design, the business model attracts the world’s extremes of income. The superrich come to stay in the luxury hotels and the superpoor come to work in them.
However, although Dubai exploits the opportunity created by global inequality, it does not cause that inequality. On the contrary, the jobs that Dubai provides help poor people.
In essence, the enthusiasm of economists for migration is
enthusiasm for the guest-worker model. Commonly the espousal of guest-worker programs is implicit, since all the other effects of migration are ignored. But Professor Alan Winters, a distinguished economist who has specialized in migration, has had the intellectual honesty to advocate the guest-worker model explicitly. Specifi-cally, he proposes that all the high-wage countries should encourage the mass temporary immigration of unskilled workers from poor
countries.13 In economic terms it is hard to fault this prescription: it
would indeed generate global economic gains and benefit almost everyone involved. The world of upstairs-downstairs could be re-created: servile maids from the bottom billion could be stuffed into the attics of every middle-class home. But what kills the proposition is its tin-eared detachment from a workable ethics. The closed, autocratic societies of the Gulf States can indeed enforce a ruthless policy of a complete separation of the rights and entitlements of the indigenous population from those of immigrants. Similarly, they can enforce expulsion of immigrants upon completion of fixed-term contracts. But the open, liberal societies of the West could not
134 HOST SOCIETIES: WELCOME OR RESENTMENT?
begin to operate such policies. Once immigrants have arrived in a country, they are extremely difficult to expel: indeed, with the exception of America, “difficult” should read “impossible.” In America the Obama administration has sustained expulsion rates at around 400,000 per year. In contrast, in Europe expulsion is rare, legally protracted, and controversial. Even the original guest workers who came from Turkey to Germany supposedly temporarily in the 1950s turned out to be permanent. Immigrants to high-wage
democracies become not just a part of the labor force, but a part of society. It is best to accept this evident fact and weigh its consequences in the overall balance of benefits and costs to the indigenous population.
CHAPTER 5
Getting Migration Policy Wrong
ON THE LONG MARCH THROUGH THE EFFECTS of migra-
tion on the indigenous populations of host societies, on those left behind in countries of origin, and on migrants themselves, we have reached a convenient resting stage. Having been through the social and economic effects on host populations, it is time for a preliminary assessment and for a preliminary application to migration policy.
Combining the Economic and the Social Effects
A reasonable assessment of the evidence of the previous chapters, stripped of the near-overwhelming desire to see it in the light best suited to whatever are one’s moral prejudices, is that moderate immigration has predominantly favorable economic effects on the indigenous population, and ambiguous social effects. There is a gain from greater cultural variety, offset by the adverse effect of diversity
136 HOST SOCIETIES: WELCOME OR RESENTMENT?
on mutual regard, and the potential weakening of a functional social model by diasporas attached to dysfunctional social models.
Sustained rapid migration would be an entirely different matter: both the economic and the social effects would most probably be adverse for host populations. The fundamental economic forces of the simple models would kick in: wages would be bid down
and
public capital spread more thinly. The social benefits to increased variety are most likely subject to diminishing returns, while the social costs of diversity and dysfunctional social models are likely increasing. To think concretely, consider immigration from a low-income country in which the social model is manifestly highly dysfunctional, namely Somalia. For any host society the first ten thousand Somali immigrants are likely to provide a pleasing gain in cultural variety and little else. But immigration that increases a culturally separate Somali diaspora from one million to two million would bring little additional gain in variety, while weakening mutual regard and giving significant weight to a bad social model.
So some controls are necessary, but their purpose is to prevent migration accelerating rather than to close it down. Since my audience will be split into a pro-migration camp and an anti-migration camp, through this provisional assessment I anticipate that I have already aroused the fundamentalists. Is there, however, any way of bringing these effects together?
The pro-migration camp will, at this stage, respond with the sentiment that it would be outrageous to sacrifice large and solid economic gains together with the pleasures of variety for a few amorphous and disreputable social ripples. Similarly, the anti-migration camp will respond with the sentiment that we should not be prepared to uproot the fabric of our society for the sake of a few ephemeral
GETTING MIGRATION POLICY WRONG 137
dollars. But if the effects are opposing, how might their net effect be determined?
One approach is to determine which effect dominates in the long run. If the costs of migration were to predominate in the short run but the benefits were in the long run, then migration could be reconceptualized as an investment. Restrictions on migration would be short-sighted. But do the effects of migration fit this temporal pattern? In the long run the only effect of migration is that the population is bigger. For sparsely populated countries such as Australia and Canada this would probably be beneficial; for densely populated ones, such as the Netherlands and England, it would probably be detrimental. The clearest economic gains are in the short run. There is an influx of young workers that temporarily reduces the dependency ratio, and the economy can be run on full throttle without inflation, as during 1997–2007. There may be some further gains in the medium run accruing from immigrant
Paul Collier Page 13
