A model is a workhorse. Its advantage is that it can provide clear answers to questions that are sufficiently complex to be beyond the reach of unaided intuitive understanding. Models are not alternatives to such understanding; they provide the scaffolding that enables us to grasp what otherwise we might miss. The simplest way of showing how this particular model works is through a
diagram. Diagrams can be clarifying, and this one carries a lot of insight relative to its difficulty. From time to time through the book I will use it to generate new insights, so it is worth a few moments of concentration. All diagrams portray some space: almost everyone
44 THE QUESTIONS AND THE PROCESS
is familiar with the typical newspaper diagram in which time is measured along the bottom (the horizontal axis), and some
newsworthy number such as the unemployment rate is measured
up the side (the vertical axis). The diagram in Figure 2.1 portrays a space in which the rate of migration from Tonga to New Zealand is measured up the side, and the size of the Tongan diaspora in New Zealand—the stock of unabsorbed immigrants and
their descendants already in the host country—is measured along the bottom.
Now depict the first building block: how migration depends upon the diaspora. Of course, migration also depends upon other things, notably the income gap. So for the moment keep the income gap constant so that we can take it off the stage, leaving just the diaspora Diaspora
Diaspora Schedule
Increasing
M’
ration
E
Diaspora
Rate of Mig
Declining
M
0
Size of Diaspora
Figure 2.1 The Opening of New Zealand to Migration from Tonga
WHY MIGRATION ACCELERATES 45
and migration in focus. For example, we might consider a host country such as New Zealand, and a country of origin such as Tonga, and try to picture how the rate of migration from Tonga to New Zealand varies with the size of the Tongalese diaspora in New Zealand. What you will picture is something like the line M-M c in the figure. Even when there is no diaspora, there is some migration, because the income gap induces some people to relocate. But the larger the Tongalese diaspora is, the faster the migration from Tonga. It will be convenient to have a name for this relationship. Genuflecting to economics, I will call it the migration function, but you could equally call it diasporas support migration since that is all it is depicting.
Now I turn to the second building block: the flows into and out of the diaspora. What are the combinations of the diaspora and migration at which the inflow from migration equals the outflow from absorption? Evidently, only if the number of new Tongalese immigrants joining the diaspora is matched by the number of past Tongalese migrants and their offspring who cease to be members of the diaspora will the size of the diaspora stay the same. In turn, only if the diaspora is constant will migration stay constant. While the Tongalese diaspora is growing, migration from Tonga is getting easier and so will be accelerating.
Many combinations of migration and the diaspora keep the diaspora the same size. For example, suppose that each year 2 percent of the Tongalese diaspora leave it. If the Tongalese diaspora in New Zealand is 30,000, then each year 600 places become vacant. So the diaspora will be constant if 600 Tongalese immigrants arrive. This link between the absorption rate and the number of immigrants has a simple implication. The Tongalese diaspora will accumulate until it is fifty times the rate of migration.
The combinations of the diaspora and migration that keep
the diaspora the same yield the diaspora schedule. What does it look
46 THE QUESTIONS AND THE PROCESS
like? One point on it is obvious: if there is neither any diaspora nor any migration the diaspora will stay constant at zero. So one point
on the schedule is the corner of the diagram. 13 To the left of the
schedule the diaspora is too small for the number of vacancies generated by absorption to match new immigration. As a result, the diaspora is growing. To the right of the schedule the diaspora is shrinking.
I show these changes, which economists rather portentously term dynamics, by the arrows.
To take stock, we have a picture showing that migration is helped by the diaspora and that the diaspora is fueled by migration and slimmed by absorption. The last building block shows how the rate of absorption depends on the diaspora. The bigger the diaspora is, the more social interactions its members have with each other, and so the slower their rate of absorption into mainstream society will be. The absorption rate is simply the slope of the schedule. 14 The
slower the absorption, the flatter the schedule will be, so as the diaspora gets larger the schedule gets flatter.
Again, if you are an intuitive genius, then you do not need the model to see how the three different forces play out. But with the model it is straightforward: we can actually predict both where the rate of migration from Tonga to New Zealand will settle down, and the eventual size of the Tongalese diaspora. Of course, our predictions will depend upon our estimate of how Tongalese migration responds to the size of the diaspora and on how the rate of absorption of Tongalese into mainstream society depends upon its size.
The model can be no better than the numbers that go into it. But it tells us how these relationships fit together.
From a glance at the diagram you can instantly see where the
equilibrium will be: where the lines cross each other. At this point the Tongalese migration induced by the diaspora matches the rate of absorption, which keeps the diaspora the same size. For a given
WHY MIGRATION ACCELERATES 47
income gap, the rate of migration remains constant and the Ton-
galese diaspora remains the same size. 15
Not only is it an equilibrium, but the forces of change inexorably bring the society to it. Until migration commences there is no Tongalese diaspora in New Zealand, so migration starts from point M.
As a result, the diaspora grows. But as the diaspora grows, migration becomes easier and so accelerates. Migration and the diaspora fuel each other, marching along the migration function together.
But rising migration and a growing diaspora do not continue
indefinitely. Once migration has accelerated to the point at which it reaches the diaspora schedule, no further change occurs. The diaspora has increased until vacancies from absorption match entry from migration. Migration and the diaspora fuel each other in a burst of mutual acceleration, but the fuel eventually burns out and both stabilize.
My depiction of migration from Tonga to New Zealand is entirely hypothetical: I do not know the actual shape of either the migration function or the diaspora schedule for this pair of countries, and I doubt whether anyone else yet does. In the same spirit of hypothetical analysis I am going to tweak the diagram by taking a pair of countries between which the income gap has been rather wider. We are no longer looking at Tonga and New Zealand in the twenty-first century, but at The Windrush, the boat that brought the first migrants from the Caribbean to Britain in 1948. Once the barriers of the Second World War and the 1930s Depression are over, the incentive to migrate is so powerful that this migration is much larger than that from Tonga to New Zealand. Depicting this in Figure 2.2, the migration function has shifted upward: for any given size of the diaspora there is more migration. The change may sound as if it would be of little consequence, but the result is dramatically different. Whereas previously the migration function and the diaspora
48 THE QUESTIONS AND THE PROCESS
M’
Diaspora Schedule
ration
Rate of Mig
M
0
Size of Diaspora
Figure 2.2 The Opening of Britain to Migration from the Caribbean schedule crossed, now they miss each other. The implication is that there is no equilibrium: migration keeps on accelerating and th
e diaspora keeps on accumulating.
I should emphasize that I have used Tongalese migration to
New Zealand and Caribbean migration to Britain only as stylized examples to illustrate a process. I do not mean to suggest that in actual fact migration from the Caribbean to Britain would not have reached equilibrium. We will never know how unrestricted migration would have played out because in 1968 the British government got sufficiently worried about mounting opposition to accelerating immigration that it imposed restrictions to limit the rate.
WHY MIGRATION ACCELERATES 49
But the real value of a model is not that it can illuminate why something happened, it is that it can be used to predict the effects of hypothetical situations, including changes in policies. This model will be our workhorse when the time comes, in chapters 5 and 12, to analyze migration policies. By using it we will be able to show why reactive policies are liable to be damaging and that better alternatives are available.
So much for the first sense of equilibrium: where the rate of migration stabilizes. The other sense, in which the net flow of people ceases, will only occur once the income gap is eliminated. The system I have sketched is a simple interaction of stocks and flows: the stock of past migrants in the diaspora, and the flow of new migrants.
Simple stock-flow models are common in all sorts of contexts. In the typical stock-flow systems that are crudely analogous to migration, such as water flowing between two tanks with different initial levels, the flow itself gradually closes the gap: one tank fills up, and the other drains down. This would apply in our present context if migration drove incomes down in host countries and raised incomes in countries of origin. The simple economic models used to predict huge income gains from global migration have just this property.
Migrants are the equilibrators: in the absence of impediments to movement, migration continues until incomes are equalized. At this point, migrants themselves may feel a bit like suckers: they have moved for nothing. Those who have remained in countries of origin end up gaining just as much. The indigenous population of host countries loses, but it can comfort itself with the thought that others have gained more than it has lost. As a description of the effects of nineteenth-century migration from Europe to North America, or for that matter from Ernsbach to Bradford, this is not a bad first approximation.16 As the Midwest was opened up, smallholders who
migrated could get larger plots of land than they had farmed back in
50 THE QUESTIONS AND THE PROCESS
Europe. As the Midwest filled up and Europe became less crowded, plot sizes gradually equalized. Eventually, farmer Schmidt in Germany was as well off as farmer Schmidt in Iowa. But as an analysis of migration from a country that has missed out on prosperity and an advanced modern economy, this simple model is worthless.
Modern migration is not a quest for land; it is a quest for efficiency.
As you will see in subsequent chapters, the feedback forces from migration onto income in both host countries and countries of origin are weak and ambiguous. Further, even though migration has accelerated, it is tiny relative to the stocks of labor in both the host countries and the countries of origin. So the feedback mechanism depends upon changes that are small and generate responses that are weak. Migration from poor countries to rich ones is not likely to have much impact on closing the income gap.
Facts and Their Implications
We have arrived at some solidly based facts that have powerful implications. The first fact is that the income gap between poor countries and rich ones is grotesquely wide and the global growth process will leave it wide for several decades. The second is that migration will not significantly narrow this gap because the feedback mechanisms are too weak. The third is that as migration continues, diasporas will continue to accumulate for some decades.
Thus, the income gap will persist, while the facilitator for migration will increase. The implication is that migration from poor countries to rich is set to accelerate. For the foreseeable future, international migration will not reach equilibrium: we have been observing the beginnings of disequilibrium of epic proportions.
The acceleration of migration is apparent from the aggregate
data. Overall, the global stock of immigrants increased from
WHY MIGRATION ACCELERATES 51
92 million in 1960 to 165 million in 2000. But this increase in the total masks the key change in composition. Migration from the rich world to the poor world shrank to just a few millions. Migration within the rich world flatlined: more movement within Europe being offset by less migration from Europe to the New World.
Note that during this period there was a huge increase in both trade and capital flows within the rich world. So much for the inevitability of globalization leading to an increase in migration: within the rich world it didn’t. The stock of migrants who had moved from one developing country to another grew modestly
from around 60 million to 80 million. What took off, from under 20 million to over 60 million, was migration from poor countries to rich ones. Further, the increase accelerated decade by decade.
The largest increase, both absolutely and proportionately was during 1990–2000, at which point the global data currently
stops. It is a reasonable presumption that 2000–2010 continued this acceleration.
As migration accelerated, the high-income societies responded by retightening their immigration controls. Primarily this was because the acceleration in migration coincided with deceleration in the growth of the high-income economies: the Golden Thirty Years came to an end. Unemployment rates, which had dropped to around 2 percent by the time immigration controls were loosened, rose to around 8 percent and stuck there. The rise in unemployment was not caused by immigration, but it eliminated the obvious arguments that had been responsible for opening borders, while introducing an apparently obvious argument for closing them
again. Variations in policy lags and differences in economic cycles between countries led to some countries tightening almost coincident with others liberalizing. The major American liberalization was in 1965; the first British tightening was in 1968. Australia
52 THE QUESTIONS AND THE PROCESS
switched from heavily subsidizing immigration through the 1960s to heavily restricting it by the 1990s.
But just as the initial opening of borders had been based on little more than short-term political opportunism, the subsequent tightening of restrictions was securely based on neither an understanding of the process of migration and its effects nor a thought-through ethical position. Migration policies were furtive and embarrassed.
Astonishingly, as migration policy soared up the rankings of the policy priorities of voters, the mainstream political parties dodged the issue. The stance of the political Left, which by this time was largely pro-migration, appeared to be “downplay the issue, have as much immigration as we can get away with, and claim it is pro-growth.” The stance of the political Right, which by this time was largely anti-immigration, appeared to be “vaguely oppose migration, but do not be explicit for fear of association with racists, and do nothing that would slow growth.” Nature abhors a vacuum, and so do political opportunists. The space left by the mainstream political parties rapidly came to be occupied by a gallery of grotesques: racists, xenophobes, and psychopaths found themselves with an audience of decent, ordinary citizens who were increasingly alarmed by the silence of the mainstream parties. To date, the only thing that has kept extremist parties at bay has been first-past-the-post voting systems. In the United States and Britain, where such voting systems make it hard for third parties to survive, extremist parties have not gained traction. But in virtually all societies with more inclusive voting systems, single-issue anti-immigrant parties now attract a remarkably high share of the vote. Far from forcing sane debate on immigration policy by the mainstream parties, the emergence of extremists has further frightened them away from the issue. Either you regard this outcome as a shocking condemnation of o
rdinary people, or as a shocking condemnation of the mainstream political
WHY MIGRATION ACCELERATES 53
parties: I view it as the latter. It is little short of disastrous that in some European countries around a fifth of the indigenous electorate is wasting its vote on pariah parties because the mainstream parties will not properly debate what these voters regard, rightly or wrongly, as the most important issue facing their country.
So what should an honest discussion of migration policy be
about? First, it should be based on impartially gathered facts such as the big three I set out above. Of course, there are many more, some of which I will cover in subsequent chapters. Based on these facts, there should then be an open discussion of the ethics of immigration restrictions. If all restrictions are a priori ethically illegitimate, migration will build to rates far in excess of those experienced in recent decades. If they are legitimate, they will be confronting greatly increased pressure of demand, and so the principles and mechanics of controls will become far more important.
This page intentionally left blank
PART 2
Host Societies
Welcome or Resentment?
This page intentionally left blank
CHAPTER 3
The Social Consequences
IN THIS PART I AM GOING TO ADDRESS the question of
how future migration might affect the indigenous populations of host societies. The key word in that sentence is “future.” I am not primarily interested in the question “Have the consequences of migration been bad or good?” If pressed for an answer, I would come down on the side of “good,” but it is not the pertinent question. Imagine, for a moment, the improbable: a consensus that the right answer is “bad.” Even in that eventuality no sane person would advocate that migrants and their descendants should be repatri-ated. In modern high-income societies mass expulsions are unthinkable. So although “Have the consequences of migration been bad or good?” is concrete and entirely meaningful, it is as irrelevant as ask-ing, “Should you have been born?” The question I am ultimately going to address is hypothetical: if migration were substantially to increase, how would it affect host populations? As I showed in chapter 2, migration accelerates unless subject to effective controls,
Paul Collier Page 5