How much is enough?

For too long we have allowed xenophobes to set the terms of the immigration debate. We do need controls over who comes, but better ones
September 18, 2013

The queue for visas outside the British Embassy in Sofia, 2006, before Bulgaria joined the European Union in 2007. (© Rex/Ray Tang)

British immigration policy clearly needs overhauling. Desperate not to give succour to xenophobes and racists, social scientists have strained every muscle to show that immigration is good for everyone. Inadvertently, this has allowed the terms of the immigration debate to be set by the xenophobes, and for the question to be asked: “Is immigration good or bad?” This is the wrong question. We should be asking not whether immigration is good or bad, but how much immigration is best. And while some immigration is better than none, there are solid reasons for thinking that beyond a certain rate it can be excessive.

Effective controls on immigration, therefore, are neither an anachronistic vestige of nationalism and racism, nor merely the obsession of paranoid xenophobes: they are going to become increasingly necessary in all the societies in which diasporas have accumulated. In the absence of effective policies, immigration tends to accelerate. The reason it does so is straightforward but is little understood and has only recently been decisively established by research. The single most powerful influence on the rate of immigration is the size of diasporas (meaning those immigrants and their descendants who have chosen to retain strong links with their country of origin). They are crucial for the rate of immigration, especially from countries which are poor and distant, because migrating is expensive. Most citizens of poor countries simply cannot afford the costs and risks. This is why immigrants tend to be from middle-income groups rather than the poorest. Having a relative in the country of destination dramatically lowers the costs and risks. As immigration fuels the diaspora, and the enlarged diaspora fuels immigration, it accelerates.

If migration accelerates it is liable to rise beyond the range at which the benefits of further migration exceed their costs. Britain, like other high-income societies, has only had six decades of immigration and so diasporas have grown gradually from negligible beginnings. For much of this time they were small enough to keep entry rates modest. This phase is over.

The process of immigration fuelling diasporas and diasporas fuelling immigration can either spiral explosively until the country of origin has been depopulated, such has occurred in Northern Cyprus, or else it will eventually settle at some rate of equilibrium, as may happen with Polish immigration. This depends upon how rapidly immigrants are absorbed into their host society and what happens to the income gap between the country of origin and the host. If immigrants swiftly integrate and lose all connections with their country of origin, migration will not accelerate significantly. For example, nearly a century ago my grandfather was an immigrant from a then-impoverished German village. But having since lost any connection with Germany, I would be able to offer no help today to an immigrant from Ernsbach. In any case, in a reversal of fortunes, incomes in Ernsbach have risen well above those in Bradford, which was the richest city in Europe when my grandfather arrived there. Immigration flows are most likely to accelerate without limit when both the culture gap and the income gap are wide. A wide culture gap slows the pace at which immigrants lose connections with their home societies so that diasporas keep growing; a wide income gap sustains the economic incentive to migrate.

Does it matter if immigration keeps accelerating? Why, if some immigration is good, is not more immigration better? Part of the answer follows from elementary economics, but mostly it derives from the complexities of social organisation.

Elementary economics provides two clear predictions about the effects of immigration on the host populations that receive them. These turn out to be too simplistic, but they are not entirely misleading. Economic wellbeing derives partly from private income and partly from government services. Where income is concerned, immigration would be expected to reduce wages and increase the returns on capital. As a result, indigenous workers would be worse off and indigenous wealth owners better off. In the case of government-provided services, the existing stock of public capital—schools, hospitals, roads—would be shared among more people and so per capita provision would deteriorate.

Poorer people receive more of their income from work and less from wealth, and more of their overall wellbeing from government-provided services. Hence, the prediction from elementary economics is that immigration benefits those indigenous people who are wealthy, but makes poorer indigenous people worse off. In parody, this already over-simplified analysis amounts to the assessment that the middle classes benefit from cleaners and nannies, but that the working classes lose from competition with workers willing to accept lower pay, and competition with immigrant families using social services.

At modest rates of immigration these predictions turn out to be largely wrong because there are offsetting effects at work. But at sufficiently high rates they would most likely be correct. A credible recent study has found that while at the bottom of the wage spectrum immigration to this country has indeed reduced wages, along most of the spectrum it has increased them. The research also found that the increases were larger than the reduction: most indigenous workers gained from immigration. The researchers speculate that the fluidity introduced by immigrant workers improved the efficiency of the labour market—immigrants concentrated in the expanding service economy of southeast England. This helped entrepreneurs to increase productivity and so pay higher wages. Consequently, the most likely effects of past immigration on wages are that most indigenous workers ended up gaining, while the poorest ended up losing. However, both these effects are tiny relative to the fuss that has been made about them.

In Britain, housing is the single most important asset, so here the effect of migration is potentially significant. Migrants increase the pressure on the housing stock. A recent estimate is that house prices are around 10 per cent higher due to immigration. Since the housing stock is disproportionately owned by older and richer people, the appreciation in house prices has entailed a large regressive transfer from lower income groups. And since immigration has been geographically highly concentrated, it will have affected regions very differently. That 10 per cent rise in national house prices due to immigration masks a negligible effect in much of the country and very large increases in London, the southeast and a few other pockets of high immigration. Paradoxically, by severely widening the north-south divide in house prices, this has made it more difficult to move to the southeast from other parts of the country. Immigration has increased the ability of firms in growth areas to recruit workers, but it has reduced the ability of indigenous workers to move to these new jobs.

A further effect is that immigrants who arrive poor compete with the indigenous poor for social housing. While the effects on the wages of low-income indigenous workers are tiny, competition for social housing has been much more substantial: immigrants who are poor tend to concentrate in a few poor neighbourhoods. A continued acceleration in immigration could potentially seriously reduce the access of the indigenous poor to social housing.

Many immigrants strive to succeed through education, and at sufficient scale this can become a problem. Among the least successful part of the indigenous population, the success of immigrants can demoralise rather than inspire. In Britain, a perennial problem has been the low aspirations of children from the working class-. Faced by decades of frustrated hopes, the indigenous underclass has settled into fatalism, seeking to avoid disappointment by not trying. Being overtaken by immigrants can further deepen this sense of the inevitability of failure. Even those children of immigrants who do not speak English at home now outperform the children of the bottom half of the indigenous working class. The problems faced by the children of immigrants—language and discrimination—are real, of course, but they can be addressed by sufficiently active policy. Yet doing so can tend to crowd out the more nebulous and apparently intractable problem of low aspirations among sections of the indigenous population.

Even towards the top of the spectrum of achievement, the success of immigrants can cause problems. We know that East Asian “tiger mothers” drive their children to attain outstanding accomplishments, with the result that in a society into which East Asians have immigrated the cream of selective educational places will be taken by this particular group. One consequence of this is that fewer children of the indigenous population will achieve the “glittering prizes”.


There are other supposed effects of immigration that are worth considering. Do we, for example, need immigrants because they are exceptionally innovative? It is often pointed out that in the USA immigrants and their children account for a disproportionate number of patented inventions. In short, the argument runs, immigrants tend to be exceptional. However, the American experience may be due more to the exceptional nature of America, as a magnet for innovative entrepreneurs, than to the exceptional nature of immigrants globally. In the long term, immigrants are absorbed into the society and so cease to be exceptional.

What about the contention that we need immigrants because we are ageing as a society? Even a continuous influx of young immigrants gives a society only a one-off fiscal windfall, whereas increased life expectancy is a continuing process. And a one-off windfall cannot be used to finance ever-rising obligations for pensions. Further, this argument presupposes that immigrants reduce the ratio of dependents to workers: being young, they are in the workforce and so balance the expanding retired indigenous population. But working immigrants also have both children and parents. By 1997, the desire among immigrants to the UK from low-income countries to bring in dependent relatives was so considerable that only 12 per cent of migrants were coming for work.

Do we need immigrants to fill skill shortages? From time to time, particular skill niches may not be filled by the indigenous population and are most readily met by selective immigration. In the 1970s, for instance, Britain found itself short of nurses and recruited them from the Commonwealth. No society can anticipate all its needs for skills, but the immigration safety valve may, over the long term, weaken the incentive to address the root of the problem of skills shortages, which is training. Firms may gain because they now get trained workers without the costs of training. But young workers born in this country lose because employers are no longer bothering to invest in training them. Hardly a week goes by without some CEO or other fulminating in a letter to a newspaper against restrictions on immigration. Such letters remind me of the old saying, “What’s good for General Motors is good for America.”

Whereas the elementary economic effects of migration are modest, the complex social effects of the diversity generated by diasporas are potentially considerable. Diversity enriches economies by bringing fresh perspectives for problem solving, and the variety it brings enhances the pleasures of life. But diversity also brings risks. Enoch Powell conjured up entirely bogus fears about diversity, envisioning “rivers of blood.” This was nonsense. In Britain, as in all high-income societies, people of diverse cultures have learned how to coexist peacefully. Diversity is fully consistent with a society of mutual respect.

But what diversity does tend to undermine is mutual regard. That is what supports the cooperation and generosity necessary for an equal society. The public goods which modern egalitarian societies provide for their citizens depend upon myriad complex cooperative games sustained by social conventions. Cooperation and generosity would not collapse however much diversity in Britain were to increase, but it would be complacent to discount the substantial evidence that rising diversity at some point threatens them. The legitimate concerns about rising diversity are about not what it has done, but what further increases might do.

There is a trade-off to be made, therefore, between the costs and benefits of further diversity. The benefits of variety are probably subject to diminishing returns, as with any other form of variety. In contrast, the costs of moderate diversity are likely to be negligible; but beyond a certain level greater diversity will jeopardise cooperation games and undermine the willingness to redistribute income. So the costs of diversity are likely to rise at an increasing rate. At some point, the incremental costs of diversity are likely to exceed the incremental gains from variety. So the right way of posing the diversity question is not whether it is good or bad—pitting the xenophobe against the progressive—but how much is best. Unfortunately, social research is currently nowhere near the level of sophistication needed to estimate at what point diversity would become seriously costly—from which one may conclude that the concerns are scaremongering. Or one may see it as grounds for caution.


What is embarrassing about current immigration controls in this country is not their existence but their inept design. That we are reduced to putting posters on the side of lorries encapsulates the cumulative legacy of ineptitude. We have an arbitrary target for immigration which is net of emigration but gross of temporary immigrants. We have no comprehensive policy on the composition of immigration, and indeed we cannot even properly monitor it. Nor do we have a measure of diversity, let alone an objective for it.

Good controls would derive both a ceiling and an assessment of the appropriate composition of immigration from empirically-grounded analysis. Since ever-rising diversity is the fundamental concern, capping it should be the objective of an immigration ceiling which would set the gross inflow of permanent immigrants equal to the rate at which diasporas are absorbed into the general population. Unfortunately, current policy on the ceiling, which is specified net rather than gross, fails to distinguish between permanent and temporary immigrants and is set without regard to the pace at which immigrants integrate—something we have not got round to measuring. Our current ceiling, even if enforced, is so mis-specified as to leave ever-rising diversity unchecked, while being incompatible with the growth of a core 21st-century export industry—tertiary education.

Research by Frédéric Docquier and his colleagues, the leading team investigating migration, suggests that the composition of immigration is likely to be more important than its scale. Skilled and employable immigrants are beneficial; dependents of the diaspora are not. A points system can screen for whether migrants are educated, but not for whether they are employable. This can only be determined by employers. Germany and New Zealand operate a double hurdle system: a threshold of education points, plus a job offer. If the diaspora have unrestricted rights to bring in dependents, then this will pre-empt the places set by the ceiling. Docquier finds that diasporas are the most important influence on immigration—they “increase migration flows and lower their average education level.” Other criteria, such as skills and employability, would become irrelevant. A reasonable way of limiting the entry of relatives might be to set an annual quota and run a lottery for places. Lotteries are long-established in America, and also New Zealand.

As a society we have to learn how to discuss the details of migration controls without descending into paroxysms of embarrassment and anger. Because of their history, the immigrant societies of Australia, Canada and the USA have long been able to do so. As a consequence, they all have policies that are more coherent and sophisticated than our own. Until we do the same, our policies will continue to be the playthings of the tabloids.

More on the immigration debate in this month's Prospect:

Mine host: An exclusive poll by Peter Kellner shows that attitudes towards immigrants have hardened considerably over the past eight years

Nation shopping? Immigration ought to be understood in the context of global economic injustice, says Rowan Williams, as he reviews Paul Collier's new book, Exodus: Immigration and Multiculturalism in the 21st Century