Immigration, Multiculturalism,

and the Welfare State

Will Kymlicka and Keith Banting

I
n most countries today, the general public does not support large-scale per-
manent immigration, and policies designed to exclude, marginalize, or as-
similate immigrants reflect this popular attitude. Some would-be
immigrants are unable to gain legal admission and are thus excluded entirely
from a country. Others are allowed to enter only as temporary workers or
asylum seekers, without any secure right to settle permanently in the country
or to take up citizenship, and are so condemned to a marginalized status. Even
in countries where (some) immigrants have the right to become permanent resi-
dents and citizens, they may be expected to hide their ethnicity and assimilate
into the mainstream society; they are not allowed to seek any public recognition
or accommodation of their distinct identity and culture. Such policies reflect a
widespread perception of immigration as a threat and a burden, which must
therefore be contained and minimized.
There have also been experiments with a more inclusive and accommodating
approach to immigration, however. In some countries, public policies support
large-scale immigration, provide newcomers with the rights needed to partici-
pate in and integrate into the larger society, and endorse a more ‘‘multicultural’’
conception of citizenship that seeks to accommodate rather than suppress immi-
grant ethnic identities. Such policies reflect a perception that—properly man-
aged—immigration can be a benefit and a resource to the country rather than a
threat to it.
Of course, these policies and the perceptions on which they are based are ideal
types. Most citizens have a more mixed and ambivalent view, and are torn be-
tween fears of ‘‘the other’’ and an impulse toward tolerance. And this ambiva-
lence is reflected in the trajectory of public policies in many countries. Hesitant
moves toward greater openness to and accommodation of immigrants alternate
with periods of backlash and retrenchment. Exaggerated claims about the tri-
umph of new models of postnational multiculturalism alternate with equally
exaggerated claims about the resurgence of nativist populism and the death of
multiculturalism. In reality, there has been no clear victor in these ongoing

281
debates about immigration, and public attitudes change significantly in response
to exposure to new ideas and important local and international events.
Prominent among these new events, of course, are the September 11 attacks
and the related actions of violent anti-Western Islamist groups and movements.
These events have cast a deep pall over immigration debates in many countries,
particularly where Muslims form a majority or significant minority of immi-
grants. In this essay, however, we want to look at a different issue that has re-
cently emerged in debates surrounding immigration—namely, the impact of
1
increasing ethnic and racial diversity on the welfare state. This is a rather diffuse
concern that takes various forms, as we will see below. But the general idea is that
a viable welfare state, which commits substantial resources to health care, income
transfers, and social services, depends on achieving and maintaining a high level
of solidarity among citizens, and that this in turn rests on feelings of common-
2
ality among citizens. If this idea is right, then there is a potential trade-off
between a more open and accommodating approach to immigration, on the one
hand, and the maintenance of a robust welfare state, on the other.
While this issue has not raised the same level of public anxiety as issues of
security and terrorism, it has become influential in academic debates and is
beginning to shape debates among policy-makers as well. Indeed, the belief
that such a trade-off exists is now taken for granted in many circles and treated
as if it were a well-established fact, creating a new basis for opposition to at-
tempts to adopt more inclusive and accommodating policies regarding the ad-
mission of immigrants, the rights of noncitizens, and multiculturalism.
In reality, however, there is surprisingly little evidence about the impact of in-
creasing ethnic diversity on the welfare state, and claims that there is an inevit-
able trade-off between them may be premature. In this essay, we will examine
the purported trade-off between diversity and the welfare state, survey the avail-
able evidence, and discuss some preliminary results of our own research on the

1 There are various ways of defining and measuring the impact of immigration on ethnic and racial diversity.

Some studies focus on the percentage of the population that is foreign-born; others focus on the percentage of
people whose mother tongue is a foreign language; others focus on phenotypic differences (the percentage of
‘‘visible minorities’’ or ‘‘racial minorities’’). There is no single term that fully captures this range of possible
understandings and definitions of diversity. For the rest of this essay, we will use ‘‘ethnic diversity’’ or ‘‘ethnic
heterogeneity’’ as an umbrella term, but this should be understood as shorthand for ‘‘ethnic, linguistic, and
racial diversity.’’ We will discuss the use of more specific indicators for measuring ethnic diversity below.
2 In our discussion, the ‘‘welfare state’’ and the related concepts of ‘‘social programs’’ and ‘‘social spending’’ are

all taken to encompass health care, income security programs, and social services, but not education. In this, we
follow the conventional approach adopted by the OECD in its compilation of data on what it calls public social
expenditures.

282 Will Kymlicka and Keith Banting

question. We will conclude with some brief reflections on the implications of
this issue for debates about the legal status of noncitizens.

CLARIFYING THE TRADE-OFF BETWEEN DIVERSITY AND THE

WELFARE STATE

Why would increasing ethnic diversity due to immigration pose a threat to the
welfare state? There are actually two concerns here:

1. That ethnic diversity as such makes it more difficult to sustain expansive

social programs and to achieve substantial redistribution toward the poor
through taxes and transfers. This is said to be true because it is difficult to
generate feelings of national solidarity and trust across ethnic lines. We can
formulate this as the hypothesis that the greater the size of ethnic minor-
ities as a percentage of the population, the harder it is to sustain a robust
welfare state. We will call this the heterogeneity/redistribution trade-off.
2. That the ‘‘multiculturalism’’ policies adopted to recognize or accom-
modate ethnic groups tend to further undermine national solidarity and
trust. Such policies (described in detail below) include multicultural edu-
cation, legal exemptions, and funding of ethnic groups. We can formulate
this as the hypothesis that the more a country embraces the multicultural
politics of (ethnic) recognition, the harder it is to sustain the politics of
(economic) redistribution. We will call this the recognition/redistribution
trade-off.
The first hypothesis argues that the very presence of sizeable ethnic diversity
erodes the welfare state, regardless of what sorts of policies governments adopt
to manage that diversity. The second hypothesis argues that a common way in
which Western governments today attempt to manage diversity—namely, by
attempting to accommodate it through multiculturalism policies, rather than
ignoring or suppressing it—aggravates the problem.
If either of these hypotheses is true, we face a serious and growing problem,
because there is no reason to expect either that ethnic minorities will diminish as
a percentage of the overall population in most Western countries, or that these
groups will abandon their claims for multicultural accommodations. On the
contrary, there is every reason to expect that immigration rates in the Western
democracies will continue to grow, partly because of concerns about the declin-
ing birth rate and aging population, and partly because there are limits on the

immigration, multiculturalism, and the welfare state 283

state’s ability to stop would-be migrants from entering the country. Similarly,
there is every reason to expect that minorities, whether they are historically
rooted or newer migrants, will continue to press demands for recognition, which
3
grow out of deep forces in contemporary societies.
So if there is a tendency for either ethnic heterogeneity or multiculturalism to
erode the welfare state, the problem is likely to get worse. If either of these
hypotheses were true, the very idea of a ‘‘multicultural welfare state’’—a welfare
state that respects and accommodates diversity—would be destined to failure.
4
This worry has been labeled as the ‘‘progressive’s dilemma.’’ Social democrats, it
is said, are faced with a tragic trade-off between sustaining their traditional agenda
of economic redistribution and embracing immigration and multiculturalism.
The belief in such a trade-off is creating a major political realignment on these
issues. In the past, most resistance to immigration and multiculturalism came
from the Right, which viewed them as a threat to cherished national traditions
and values. Today, however, opposition to immigration and multiculturalism is
5
emerging within the Left as a perceived threat to the welfare state.
But are the hypotheses true? Why have so many people come to believe that
ethnic heterogeneity threatens the welfare state, and that multiculturalism poli-
cies exacerbate the problem? We will start with the heterogeneity debate, and
then examine multiculturalism.

The Case for a Heterogeneity/Redistribution Trade-Off

In one sense, the idea that ethnic heterogeneity can weaken the pursuit of a
robust welfare state is an old one. Karl Marx argued that racial divisions within
the working class in the United States would undermine its capacity to demand
6
progressive reforms, and this has been a recurring theme in American politics.
Yet, until very recently, no one has attempted to systematically test the impact
of heterogeneity on welfare-state levels.

3 For a discussion of these forces, including the postwar human rights revolution, the desecuritization of state-

minority relations, and democratization, see Will Kymlicka, ‘‘Culturally Responsive Policies’’ (background
paper prepared for the 2004 United Nations Human Development Report, June 15, 2004); available at hdr.
undp.org/docs/publications/background_papers/2004/HDR2004_Will_Kymlicka.pdf.
4 See David Goodhart, ‘‘Too Diverse?’’ Prospect 95 (February 2004), pp. 30–37; and Nick Pearce, ‘‘Diversity

versus Solidarity: A New Progressive Dilemma,’’ Renewal: A Journal of Labour Politics 12, no. 3 (2004).
5 For an overview of the debates within European social democratic parties on these issues, see René Cuperus,

Karl Duffek, and Johannes Kandel, eds., The Challenge of Diversity: European Social Democracy Facing Migration,
Integration and Multiculturalism (Innsbruck: StudienVerlag, 2003).
6 For a discussion of the history of this argument, see Seymour Martin Lipset and Gary Marks, It Didn’t Happen

Here: Why Socialism Failed in the United States (New York: W. W. Norton, 2000).

284 Will Kymlicka and Keith Banting

 Attention to ethnic heterogeneity as an explanatory variable has recently
emerged, initially in two discrete geographical contexts. First, development econ-
omists, including some associated with the World Bank, pointed to ethnic and
tribal diversity in attempting to explain the poor economic and social perform-
ance of countries in sub-Saharan Africa. The focus here was initially on the
impact of ethnic heterogeneity on economic growth, but subsequent research
has extended the focus to include the negative impact of heterogeneity on the
7
provision of public goods, such as public education.
The second context concerns the United States, where ethnic and racial heter-
ogeneity has been invoked to explain differences in social expenditures across
8
cities and states within the country. These studies consistently show, for exam-
ple, that the higher the proportion of African Americans within a state, the more
restrictive state-level welfare programs such as Medicaid are. More recently,
Alberto Alesina and Edward Glaeser have extended this analysis to the cross-
national level, arguing that differences in racial diversity are a significant part
of the explanation of why the United States did not develop a European-style
9
welfare state.

7 See William Easterly and Ross Levine, ‘‘Africa’s Growth Tragedy: Policies and Ethnic Divisions,’’ Quarterly

Journal of Economics 112 (1997), pp. 1203–50; William Easterly, ‘‘Can Institutions Resolve Ethnic Conflict?’’ Eco-
nomic Development and Cultural Change 49, no. 4 (2001), pp. 687–706; William Easterly, The Elusive Quest for
Growth: Economists’ Adventures and Misadventures in the Tropics (Cambridge: MIT Press, 2001); Daniel Nettle,
‘‘Linguistic Fragmentation and the Wealth of Nations,’’ Economic Development and Cultural Change 49 (2000),
pp. 335–48; and Quentin Grafton, Stephen Knowles, and Dorian Owen, ‘‘Social Divergence and Productivity:
Making a Connection,’’ in Andrew Sharpe, France St-Hilaire, and Keith Banting, eds., The Review of Economic
Performance and Social Progress: Towards a Social Understanding of Productivity (Montreal: Institute for Re-
search in Public Policy, 2002).
8 Alberto Alesina and Edward Glaeser, Fighting Poverty in the US and Europe: A World of Difference (Oxford:

Oxford University Press, 2004); Rodney Hero, Faces of Inequality: Social Diversity in American Politics (New
York: Oxford University Press, 1998); Rodney Hero and Caroline Tolbert, ‘‘A Racial/Ethnic Diversity Inter-
pretation of Politics and Policy in the States of the U.S.,’’ American Journal of Political Science 40 (1996), pp. 851–
71; Rodney Hero and Rob Preuhs, ‘‘Multiculturalism and Welfare Policies in the US States: A State-level
Comparative Analysis,’’ in Keith Banting and Will Kymlicka, eds., Multiculturalism and the Welfare State: Re-
cognition and Redistribution in Contemporary Democracies (Oxford: Oxford University Press, forthcoming);
Martin Johnson, ‘‘The Impact of Social Diversity and Racial Attitudes on Social Welfare Policy,’’ State Politics
and Policy Quarterly 1, no. 1 (2001), pp. 27–49; Joe Soss, Sanford Schram, and Richard Fording, eds., Race and
the Politics of Welfare Reform (Ann Arbor, Mich.: University of Michigan Press, 2003); Martin Johnson,
‘‘Racial Context, Public Attitudes, and Welfare Effort in the American States,’’ in Soss, Schram, and Fording,
eds., Race and Politics, pp. 151–70; Joe Soss, Sanford Schram, Thomas Vartanian, and Erin O’Brien, ‘‘Setting the
Terms of Relief: Explaining State Policy Choices in the Devolution Revolution,’’ American Journal of Political
Science 45, no. 2 (2001), pp. 378–95; Matthew C. Fellowes and Gretchen Rowe, ‘‘Politics and the New American
Welfare State,’’ American Journal of Political Science 48 (April 2004), pp. 362–73; and Erzo Luttmer, ‘‘Group
Loyalty and the Taste for Redistribution,’’ Journal of Political Economy 109, no. 3 (2001), pp. 500–28.
9 Alesina and Glaeser, Fighting Poverty.

immigration, multiculturalism, and the welfare state 285

 Based on these studies from Africa and the United States, several scholars have
concluded that there is a universal tendency for people to resist interethnic redis-
10
tribution. Explanations differ as to why this tendency exists. Some invoke an
11
inherent genetic disposition to favor co-ethnics; others see it as a predictable
12 13
by-product of social capital deficits or electoral dynamics. But whatever the
explanation, there is an increasing tendency to assume that ethnic heterogeneity
erodes support for redistributive policies. Although the main evidence for this
assumption comes from two specific contexts—namely, sub-Saharan Africa and
the United States—it is increasingly treated as a universal tendency. The strongly
racialized dimension of U.S. welfare politics is no longer seen as an anomaly—a
pernicious legacy of the peculiar American history of slavery and segregation—
but rather as a normal, even inevitable, reaction to the simple fact of ethnic
heterogeneity. Indeed, the United States has come to represent the leading in-
ternational example of the proposition that heterogeneity as such erodes redis-
tribution. Its story has emerged as a sort of ‘‘master narrative,’’ the quintessential
model of the (inherently weak) heterogeneous welfare state.
Based on this assumption, scholars have drawn rather dire predictions about
the future of the welfare state across the Western democracies. If it is the mere
presence of ethnic minorities that has weakened the welfare state in the United
States, then increasing immigration threatens to do the same in Europe. As early
as 1986, Gary Freeman predicted that immigration would lead to ‘‘the American-
14
ization of European welfare politics,’’ and more recent studies have reiterated
15
this prediction. In their comparison of U.S. and European welfare states,
Alesina and Glaeser conclude with ‘‘a caution about current directions in
European politics. . . . As Europe has become more diverse, Europeans have
increasingly been susceptible to exactly the same form of racist, anti-welfare

10 It is important here to distinguish episodic ‘‘humanitarian’’ charity in response to disasters from ongoing

institutionally compelled redistribution. The debate concerns resistance to the latter.

11 Frank Kemp Salter, ed., Welfare, Ethnicity, and Altruism: New Findings and Evolutionary Theory (London:

Frank Cass, 2004).

12 Robert Putnam, ‘‘Who Bonds? Who Bridges? Findings from the Social Capital Benchmark Survey’’ (pre-

sentation to the annual meeting of the American Political Science Association, Chicago, September 2004).
13 Herbert Kitschelt, The Radical Right in Western Europe: A Comparative Analysis (Ann Arbor, Mich.: Uni-

versity of Michigan Press, 1995).

14 Gary Freeman, ‘‘Migration and the Political Economy of the Welfare State,’’ Annals of the American Academy

of Political and Social Science 485 (1986), p. 62.

15 Nathan Glazer, ‘‘The American Welfare State: Exceptional No Longer?’’ in Henry Cavanna, ed., Challenges to

the Welfare State: Internal and External Dynamics for Change (Cheltenham, U.K.: Edward Elgar, 1998); and
Goodhart, ‘‘Too Diverse?’’

286 Will Kymlicka and Keith Banting

demagoguery that worked so well in the United States. We shall see whether the
16
generous welfare state can really survive in a heterogeneous society.’’
This, then, is the basic case for a heterogeneity/redistribution trade-off. The
case begins with strong empirical evidence from two specific contexts—Africa
and the United States—where negative correlations between ethnic heterogeneity
and public spending have been found in several studies. This evidence is then
combined with some speculations about the underlying evolutionary or socio-
logical mechanisms at work, and the conclusion is drawn that there is a general
tendency for heterogeneity to erode redistribution.
This basic case has some initial plausibility, but it is not airtight. The empirical
evidence is drawn from two contexts that are arguably atypical and quite un-
related to contemporary issues of immigration. In the sub-Saharan context, the
artificiality of state boundaries, combined with the weakness of state institutions
at the time of independence, meant that states had no usable traditions or in-
stitutional capacity for dealing with diversity. In the American context, racial
animosity had been sedimented by centuries of slavery and segregation, the
maintenance of which depended on state-sponsored ideologies and practices that
dehumanized blacks. One could argue that neither of these contexts provides
a reliable basis for predicting the impact of, say, increasing Turkish immigra-
tion on the German welfare state or increasing Filipino immigration on the
Canadian welfare state. Where minorities are newcomers rather than historically
enslaved groups, and where state institutions are strong rather than weak, the
impact of increasing heterogeneity may be quite different. We therefore need
more research before drawing any firm conclusions about the heterogeneity/
17
redistribution trade-off.

The Case for a Recognition/Redistribution Trade-Off

Much of the literature to date has focused on the link between levels of ethnic
heterogeneity and the size or performance of the welfare state. But for most
practical purposes, the level of ethnic heterogeneity is a given—a country simply

16Alesina and Glaeser, Fighting Poverty, pp. 180–81.

17We should also be cautious about the data used in these studies to measure correlations of heterogeneity and
redistribution. For example, the classification of ethnic groups in the Index of Ethno-linguistic Fractionalization
used in the Alesina and Glaeser study is not consistent, but varies significantly across countries. For example,
the U.K. data reflect racial differences: e.g., White 93.7 percent; Indian 1.8 percent; Black 1.4 percent. In Canada,
however, the data represent an amalgam of linguistic and national origins: e.g., French 22.8 percent; Other
Canadian 43.5 percent; British 20.8 percent; German 3.4 percent.

immigration, multiculturalism, and the welfare state 287

finds itself with certain ethnic minorities. The size of ethnic minority popula-
tions can be affected at the margins by increasing or decreasing immigration, but
this typically only has a significant effect over the long term.
For most policy-makers, therefore, the crucial issue is not ‘‘what level of ethnic
heterogeneity is desirable,’’ but rather ‘‘how should we respond to the ethnic
heterogeneity that already exists in our society?’’ In the past, ethnic diversity was
often seen as a threat to political stability, and hence as something to be discour-
aged by public policies. As we have seen, immigrants have been subject to a
range of policies intended to exclude, marginalize, or assimilate them. Compara-
ble policies of exclusion and assimilation have also been adopted toward nonim-
migrant minorities, such as indigenous peoples. Today, however, many Western
democracies have abandoned these earlier policies and have shifted toward a
more accommodating approach to diversity. This is reflected, for example, in the
widespread adoption of multiculturalism policies for immigrant groups, the
acceptance of territorial autonomy and language rights for national minorities,
and the recognition of land claims and self-government rights for indigenous
peoples. We will refer to all such policies as ‘‘multiculturalism policies,’’ or MCPs.
This trend has generated the second main worry that we wish to discuss—
namely, the fear that there is a trade-off between multiculturalism policies and
the welfare state. MCPs publicly recognize and institutionalize heterogeneity; if
we assume that heterogeneity weakens the welfare state, surely policies that
heighten the public visibility and political salience of diversity will only com-
pound its negative impact. Critics worry, therefore, that there is a conflict be-
tween the politics of recognition and the politics of redistribution. These critics
generally acknowledge that defenders of MCPs do not intend to weaken the
welfare state. On the contrary, most defenders of MCPs are also strong de-
fenders of the welfare state, and view both as flowing from the same underlying
principled commitment to remedying injustice and reducing inequality. The
conflict between MCPs and the welfare state, therefore, is not so much a matter
of competing ideals or principles, but of unintended sociological dynamics.
MCPs, critics worry, erode the interpersonal trust, social solidarity, and political
coalitions that sustain the welfare state.
Why have so many critics assumed that there is a recognition/redistribution
trade-off? In the case of the heterogeneity/redistribution trade-off, the basic case
rests on a few empirical studies from Africa and America, which are interpreted
as reflections of a universal trend or tendency. In the case of the recognition/

288 Will Kymlicka and Keith Banting

redistribution trade-off, by contrast, there are no empirical studies that have re-
vealed a negative correlation between the adoption of MCPs and the welfare
state. No one has even attempted such an empirical test of the recognition/
redistribution hypothesis.
Instead, the argument for a recognition/redistribution trade-off is entirely
speculative and conjectural. Critics have speculated about a range of mechanisms
by which the adoption of MCPs could inadvertently erode the welfare state. We
have elsewhere summarized these as the ‘‘crowding-out’’ effect (that is, time and
energy spent on MCPs leaves less time and energy for redistribution struggles),
the ‘‘corrosion’’ effect (MCP policies that heighten the salience of ethnic differ-
ence reduce feelings of commonality and solidarity), and the ‘‘misdiagnosis’’ ef-
fect (the ideology of multiculturalism encourages people to explain inequalities
as the result of cultural misunderstandings, rather than the result of class struc-
18
tures or racism).
This, then, is the basic case for the existence of a recognition/redistribution
19
trade-off. It has some initial plausibility, and even some defenders of MCPs
have acknowledged its force. Anne Phillips, for example, who defended a multi-
culturalist conception of democracy in her 1995 book, has subsequently said,
‘‘I cannot avoid troubled thoughts about the way developments I otherwise sup-
port have contributed (however inadvertently) to a declining interest in
20
economic equality.’’

18 In identifying these complaints, we have drawn in particular on the writings of a set of critics whose works
have become widely cited in the literature. See Brian Barry, Culture and Equality: An Egalitarian Critique of
Multiculturalism (Cambridge: Polity Press, 2001); Todd Gitlin, The Twilight of Common Dreams: Why America
Is Wracked by Culture Wars (New York: Metropolitan Books, 1995); Richard Rorty, Achieving Our Country:
Leftist Thought in Twentieth-Century America (Cambridge: Harvard University Press, 1999); Richard Rorty, ‘‘Is
‘Cultural Recognition’ a Useful Concept for Leftist Politics?’’ Critical Horizons 1 (2000), pp. 7–20; Alan Wolfe
and Jyette Klausen, ‘‘Identity Politics and the Welfare State,’’ Social Philosophy and Policy 14, no. 2 (1997),
pp. 213–55; and Alan Wolfe and Jyette Klausen, ‘‘Other Peoples,’’ Prospect, December 2000, pp. 28–33. When
referring to ‘‘the critics,’’ we have these authors in mind, as well as the many commentators who have en-
dorsed their arguments. For a detailed discussion of these three effects, see Keith Banting and Will Kymlicka,
‘‘Do Multiculturalism Policies Erode the Welfare State?’’ in Philippe van Parijs, ed., Cultural Diversity versus
Economic Solidarity (Brussels: Editions De Boeck Université, 2004), pp. 227–84.
19 There are other accounts in the literature. Nancy Fraser, for example, suggests that recognition and redis-

tribution conflict because the latter is ‘‘de-differentiating’’ (i.e., aims at reducing differences between groups, by
creating greater similarity in life conditions), whereas the former is ‘‘differentiating’’ (i.e., affirms group boun-
daries). See Nancy Fraser, ‘‘Social Justice in the Age of Identity Politics: Redistribution, Recognition and Partic-
ipation,’’ in The Tanner Lectures on Human Values 19 (1998), pp. 1–67; and Nancy Fraser, ‘‘Rethinking Recognition,’’
New Left Review, (May 2000), pp. 107–120. But it’s not clear why reducing differences in economic circumstances
should conflict with recognition of differences in cultural identities. If we ask why the former conflicts with the
latter, Fraser’s answer would probably end up invoking one or more of the three mechanisms we have listed above.
20 Anne Phillips, Which Equalities Matter? (Cambridge: Polity Press, 1999).

immigration, multiculturalism, and the welfare state 289

 As we mentioned earlier, however, there is as yet no empirical evidence for the
existence of such a trade-off. It remains wholly speculative. Defenders of MCPs
have responded that, far from eroding solidarity, multiculturalism policies can
actually strengthen it by reducing prejudice and mistrust, and thereby help build
21
more stable pan-ethnic coalitions for social justice generally. Here, again, more
research is required before we make any definitive judgments about the recog-
nition/redistribution trade-off.

TESTING THE TRADE-OFFS: SOME PRELIMINARY RESULTS

Any attempt to systematically confirm or refute the two hypotheses would re-
quire a massive and long-term undertaking, in part because the existing data are
insufficient to test the claims. We do not have reliable cross-national data over
time on many of the crucial variables, such as levels of different types of ethnic
heterogeneity, or levels of different types of MCPs. Even the data on welfare
spending is not as consistent as one would like. Nonetheless, there are some
more preliminary tests that can be conducted regarding the two hypotheses. The
following section describes some recent findings from studies that we have
conducted in collaboration with several colleagues.

Testing the Heterogeneity/Redistribution Trade-Off

As we noted earlier, those who believe that there is a heterogeneity/redistribution
trade-off invoke it both to explain the historical weakness of the American
welfare state and to predict the future of European welfare states as immigration
increases. Skeptics respond that the racialization of American welfare politics is
an idiosyncratic product of the history of American race relations and need not
be a harbinger of the impact of immigration on the welfare state.
Is there a way to test this dispute? In a recent study, Stuart Soroka, Keith
Banting, and Richard Johnston analyzed the relationship between immigration

21 For versions of these responses, see Richard Caputo, ‘‘Multiculturalism and Social Justice in the United States:

An Attempt to Reconcile the Irreconcilable within a Pragmatic Liberal Framework,’’ Race, Gender and Class 7,
no. 4 (2001), pp. 161–82; James Tully, ‘‘Struggles over Recognition and Distribution,’’ Constellations 7 (2000);
Bhikhu Parekh, ‘‘Redistribution or Recognition? A Misguided Debate,’’ in Stephen May, Tariq Modood, and
Judith Squires, eds., Ethnicity, Nationalism, and Minority Rights (Cambridge: Cambridge University Press,
2004), pp. 199–213; and Christopher Zurn, ‘‘Group Balkanization or Societal Homogenization: Is There a
Dilemma between Recognition and Distribution Struggles?’’ Public Affairs Quarterly 18, no. 2 (2004), pp. 159–86.
As with the critiques of MCPs, these responses are entirely speculative. For a case study of how the struggle for
MCPs has reinvigorated left-wing struggles for redistribution, see Donna Lee Van Cott, ‘‘Multiculturalism
versus Neoliberalism in Latin America,’’ in Banting and Kymlicka, eds., Multiculturalism and the Welfare State.

290 Will Kymlicka and Keith Banting

and the change in the level of social spending across OECD countries from 1970
22
to 1998. In this study, immigration is measured using United Nations data on
‘‘migrant stock,’’ the proportion of the population born outside the country. To
analyze the role of migrant stock in the evolution of the welfare state, the study
begins by adopting the most sophisticated existing models of the factors asso-
23
ciated with variation in social spending. These models incorporate a range of
factors that have been shown to be important determinants of social spending,
such as the proportion of the population over age sixty-four, the percentage of
women in the labor force, and the strength of the political left in government.
These models, however, have not typically included ethno-racial heterogeneity
as a variable. So this study adds ‘‘migrant stock’’ to the model to see what effect,
if any, it has on social spending.
If a heterogeneity/redistribution trade-off existed, we would expect that those
countries with higher levels of migrant stock would have exhibited either a de-
crease in social spending or at least slower rates of growth compared to countries
with lower levels of immigration. The study revealed, however, that there is no
relationship between the proportion of the population born outside the country
and growth in social spending over the last three decades of the twentieth cen-
tury, controlling for other factors associated with social spending. There was
simply no evidence that countries with large foreign-born populations had more
trouble sustaining and developing their social programs over these three decades
than countries with small immigrant communities.
24
This finding has been confirmed in a more recent study we conducted and
25
in a separate analysis conducted by Peter Taylor-Gooby. He too shows that

22 Stuart Soroka, Keith Banting, and Richard Johnston, ‘‘Immigration and Redistribution in the Global Era,’’ in
Pranab Bardham, Samuel Bowles, and Michael Wallerstein, eds., Globalization and Egalitarian Redistribution
(Princeton: Princeton University Press, 2006).
23 In particular, the models developed in Duane Swank, Global Capital, Political Institutions, and Policy Change

in Developed Welfare States (Cambridge: Cambridge University Press, 2002); and Evelyne Huber and John
Stephens, Development and Crisis of the Welfare State: Parties and Policies in Global Markets (Chicago: Univer-
sity of Chicago Press, 2001).
24 See Keith Banting, Richard Johnston, Will Kymlicka, and Stuart Soroka, ‘‘Do Multiculturalism Policies Erode

the Welfare State? An Empirical Analysis,’’ in Banting and Kymlicka, eds., Multiculturalism and the Welfare
State. This more recent study also examined the impact of two other (nonimmigrant) forms of ethno-racial
diversity: namely, indigenous peoples (e.g., American Indians, Maori, Sami) and national minorities (e.g.,
Scots, Catalans, Flemish, Quebecois). Here again, in both cases, there is no correlation between the size of the
minority and change in welfare spending over the past thirty years. Countries with larger indigenous
populations or national minorities had no more difficulty sustaining their welfare spending than countries
with smaller such groups.
25 See Peter Taylor-Gooby, ‘‘Is the Future American? Or, Can Left Politics Preserve European Welfare States

from Erosion through Growing ‘Racial’ Diversity?’’ Journal of Social Policy 34, no. 4 (2005).

immigration, multiculturalism, and the welfare state 291

once the other factors that affect social spending are controlled for, the size of
ethno-linguistic and racial groups has no statistically significant effect on social
spending within Western democracies (apart from the United States). He con-
cludes that there is, as yet, no evidence that immigration will have the same
effect on European welfare states that race has historically had on the American
welfare state.
One might argue that this simply reflects a time lag: perhaps the corroding
effects of heterogeneity have not yet appeared. Welfare programs are large,
complex, slow-moving entities that do not change overnight. For this reason,
it is useful to look not only at welfare spending levels but also at public support
for the welfare state. As Markus Crepaz notes, we can view public attitudes
as the ‘‘canary in the mine.’’ If heterogeneity is going to have an eroding impact
on the politics of redistribution, this will likely show up first as a drop in public
support for the welfare state, before it shows up in actual changes in spending
levels. Crepaz’s research shows, however, that there is no clear tendency for in-
creasing immigration to erode support for social spending in the Western
democracies. After a comprehensive analysis of cross-national data on public atti-
tudes, Crepaz warns against extrapolating from the American experience. The
challenge posed by immigration to European welfare states is unfolding now that
these welfare systems have reached maturity and are deeply embedded in public
expectations. In contrast, in the American experience, racial heterogeneity ham-
pered movement toward a more comprehensive welfare state from the outset.
As a result, he concludes, ‘‘there is little evidence that immigration-induced
26
diversity will lead to an ‘Americanization’ of the European welfare state.’’
In short, we have found little support for the idea that the size of the immi-
grant population exercises a systematic or inevitable downward pressure on the
welfare state. The correlations regarding African Americans in the United States,
or ‘‘tribal’’ groups in sub-Saharan Africa, cannot be taken as universal tendencies
that apply to immigrant groups in other countries.
These studies, however, do reveal one note of caution relating to the case of
immigrants: the pace of change does seem to matter. When the analysis examines
the relationship between growth in the foreign-born population and change in

26 See Markus Crepaz, ‘‘‘If You Are My Brother, I May Give You a Dime!’ Public Opinion on Multiculturalism,

Trust and the Welfare State,’’ in Banting and Kymlicka, eds., Multiculturalism and the Welfare State; and Markus
Crepaz, Trust Without Borders: Immigration, the Welfare State and Identity in Modern Societies (Ann Arbor,
Mich.: University of Michigan Press, forthcoming).

292 Will Kymlicka and Keith Banting

social spending as a proportion of GDP between 1970 and 1998, an interesting re-
sult emerges: countries with large increases in the proportion of their population
born outside the country tend to have smaller increases in social spending. This
relationship remained statistically significant in multivariate analysis. Social
spending as a proportion of GDP rose in every country in the sample during this
period, including in countries with substantial growth in migrant stock. But the
growth was smaller in countries that saw a significant increase in the portion of
the population born outside the country, other things held constant.
While there is still much work to be done in this field, these preliminary find-
ings are suggestive. There is no evidence here that countries with large immi-
grant populations have greater difficulty in sustaining and enhancing their
historic welfare commitments. But large changes do seem to matter. It is the
pace of social change rather than the fact of ethnic diversity per se that stands
out here as politically unsettling.

Testing the Recognition/Redistribution Trade-Off

It seems, then, the sheer fact of immigration-induced ethnic diversity does not,
by itself, have a depressive effect on the welfare state. And perhaps that should
not be surprising. The effects of ethnic diversity are likely to depend on many
factors, including the public policies that countries adopt in response to that
diversity. Perhaps ethnic heterogeneity only erodes the welfare state when the
government mismanages diversity and makes it a source of social conflict and
political division, thereby corroding trust and solidarity.
But which kinds of state policies toward immigrants corrode trust and solidar-
ity? This is the crux of the debate over the recognition/redistribution trade-off. As
we have seen, some commentators argue that multiculturalism policies that give
recognition to immigrant ethnicity are likely to corrode solidarity, since they en-
courage people to focus on their ethnic differences and separate group identities
rather than on the things they share with other members of society. Defenders of
MCPs respond that multiculturalism can strengthen solidarity by building mutual
respect and understanding (for example, by highlighting the positive contribu-
tions that different groups have made in the school curricula, or by creating
public spaces where minorities can explain their views and traditions).
Is there any way to test this dispute? In order to test the recognition/
redistribution trade-off hypothesis, we would need to find a database that
measures the level of multiculturalism policies in different countries around

immigration, multiculturalism, and the welfare state 293

the world. We could then plug ‘‘extent of MCPs’’ into models of welfare state
spending, to see whether ethnocultural recognition erodes economic redistribu-
tion. Unfortunately, there are no such databases. To our knowledge, no one has
attempted to systematically measure the extent to which different countries have
adopted multiculturalism policies.
In order to test the hypothesis, therefore, we have constructed such an index.
This involved several steps. We first developed a representative list of multicul-
turalism policies; we then checked to see which countries had adopted which
policies, and on that basis categorized countries as stronger or weaker in their
level of MCPs. Finally, we asked whether countries that have adopted a wider
range of multicultural policies over the last two decades have, in fact, experi-
enced a weakening or even just slower growth in their welfare states compared
to countries that have resisted such policies.
To begin, then, we selected the following eight policies as the most common
or emblematic of a multicultural approach to immigrant integration:

1. Constitutional, legislative, or parliamentary affirmation of multi-

culturalism;
2. the adoption of multiculturalism in school curricula;
3. the inclusion of ethnic representation and/or sensitivity in the mandate of
public media or media licensing;
4. exemptions from dress codes, Sunday-closing legislation, and so forth
(either by statute or by court cases);
5. allowance of dual citizenship;
6. the funding of ethnic group organizations to support cultural activities;
7. the funding of bilingual education or mother tongue instruction;
8. affirmative action for disadvantaged immigrant groups.

The first three policies celebrate multiculturalism; the middle two reduce legal
constraints on diversity; and the final three represent forms of active support for
immigrant communities and individuals.
We then examined twenty-one established democracies to see which of these
eight policies had been adopted. A country that had adopted six or more of these
policies was classified as ‘‘strong’’ in its commitment to MCPs; a country that
had adopted two or less of these policies was classified as ‘‘weak’’; countries fall-
ing in between were categorized as ‘‘modest.’’ The resulting groupings of coun-
tries are reported in Table 1.

294 Will Kymlicka and Keith Banting

 Table 1. The Strength of Immigrant Multiculturalism Policies in Western Democracies

STRONG Australia, Canada

MODEST Belgium, Netherlands, New Zealand, Sweden, United Kingdom, United States
WEAK Austria, Denmark, Finland, France, Germany, Greece, Ireland, Italy,
Japan, Norway, Portugal, Spain, Switzerland

The second step was to examine how the three groups of countries fared in
terms of change in the strength of their welfare state between 1980 and the end
of the 1990s. Is it true that countries that adopted strong multiculturalism poli-
cies for immigrants had more difficulty in maintaining and enhancing their level
of social spending and the redistributive impact of their tax and transfers than
countries that resisted such approaches over the last two decades of the twentieth
century? Table 2 provides a first cut at the issues. In short, we find no evidence of
a systematic tendency for immigrant multiculturalism policies to weaken the
welfare state. Countries that adopted such programs did not experience erosion
of their welfare states or even slower growth in social spending than countries
that resisted such programs. Indeed, on our two measures—change in social
spending as a proportion of GDP and change in the redistributive impact of
taxes and transfers between 1980 and 2000—the countries with the strongest im-
migrant multiculturalism policies did better than the other groups, providing a
hint that perhaps multiculturalism policies may actually ease possible tensions
27
between heterogeneity and redistribution.
As part of this study, we conducted a similar analysis for indigenous peoples
and national minorities. In each case, we identified a list of representative MCPs
applicable to such groups, categorized countries as strong, modest, or weak in
their level of MCPs, and tested whether strong-MCP countries experienced more
difficulty sustaining their welfare states compared to other countries. Here again,
as with immigrants, we found no evidence that engaging in a strong politics of
recognition regarding indigenous peoples and national minorities entails a
28
trade-off with a politics of redistribution.
As with the heterogeneity hypothesis, critics might argue that the hypothesized
29
negative effects of MCPs have simply not had time to show up in spending levels.

27 More recently, this conclusion has been tested in multivariate analysis, with the same result. See Banting,

Johnston, Kymlicka, and Soroka, ‘‘Do Multiculturalism Policies Erode the Welfare State?’’
28 Banting and Kymlicka, ‘‘Do Multiculturalism Policies Erode the Welfare State?’’
29 See van Parijs, ed., Cultural Diversity.

immigration, multiculturalism, and the welfare state 295

 Table 2. Multiculturalism Policies and Change in Social Spending and Redistribution

Social Spending Redistribution

Multiculturalism Policies Average % Change Average % Change

Strong 42.8 11.8

Medium 3.8 29.2
Weak 18.3 10.6

Notes
1. Change in social spending represents change in ‘‘public social expenditures’’ on health care, income trans-
fers, and social services between 1980 and 2000. Based on data in OECD Social Expenditure Database.
2. Redistribution is measured as the difference between the Gini coefficient for the distribution of market
income and the Gini coefficient for the distribution of final income, which incorporates the effects of taxes
and transfers. Change in redistribution therefore represents change in the difference between the distribution
of market and final income resulting from taxes and transfers between the early 1980s and 2000 or near years.
Based on data provided by the Luxembourg Income Study.
Source: For details of the calculations, see Keith Banting and Will Kymlicka, ‘‘Do Multiculturalism Policies
Erode the Welfare State?’’ in Philippe van Parijs, ed., Cultural Diversity versus Economic Solidarity (Brussels:
Editions De Boeck Université, 2004), Appendix 1.

To address this worry, we need to see whether there has been a drop in public
support for the welfare state, even if it has not yet affected actual changes in
spending levels. This question has been examined in a recent study by Crepaz.
Drawing on public opinion data from a variety of surveys across the Western de-
mocracies, Crepaz asks whether states with higher levels of MCPs have seen an
erosion in public support for redistribution in comparison with countries with
lower levels of MCPs. Here again, the results are encouraging: he finds no evidence
30
that adopting MCPs erodes trust, solidarity, or support for redistribution.
In short, we have found no support for the claim that there is an inherent or
systematic trade-off between policies of ethnocultural recognition and economic
redistribution. It is possible, even likely, that there are more localized circumstan-
ces where particular forms of recognition erode particular forms of redistribution.
But given the overall results, it is equally likely that there are other circumstances
where the politics of recognition enhances redistribution. It would obviously be
of great interest to try to identify these more localized cases of either mutual
31
interference or mutual support between recognition and redistribution.

30 Crepaz, ‘‘‘If You Are My Brother, I May Give You a Dime!’’’ in Banting and Kymlicka, eds., Multiculturalism

and the Welfare State.

31 For some preliminary exploration of this topic, see the case studies in Banting and Kymlicka, eds., Multi-

culturalism and the Welfare State.

296 Will Kymlicka and Keith Banting

But the evidence to date provides no support for the bold claim that ‘‘a politics
32
of multiculturalism undermines a politics of redistribution.’’

NATIONHOOD, MULTICULTURALISM, AND THE RIGHTS OF

NONCITIZENS

These results are preliminary and the hypotheses clearly require further testing,
including research on the possible causal mechanisms that connect redistribution
to recognition. But assuming that the basic findings hold up, do they have impli-
cations for debates over the legal status of noncitizens? Our studies do not di-
rectly address specific issues regarding the rights of noncitizens, including their
right to stay in a country, to access social benefits and employment opportuni-
ties, to exercise civil and political rights, to naturalize, and so on. Western de-
mocracies vary in how they approach these matters, and we have not attempted
to directly evaluate their merits.
Our goal, rather, has been to address one of the objections that is commonly
raised to the claims of immigrants generally, at all stages of the migration and in-
tegration process, from decisions about who gets admitted, to the legal status of
noncitizens, to the terms of naturalization and expectations of citizenship.
Underlying all of these debates is the fear that the presence of large numbers of
newcomers and/or their claims to accommodation and recognition are inher-
ently corrosive of solidarity and economic redistribution. Fears of this kind have
been invoked to defend restrictive admissions policies, to justify limiting the
rights of noncitizens, and to defend assimilationist models of naturalization and
citizenship.
If our findings are valid, this objection can be answered. Acknowledging the le-
gitimate presence of immigrants and enabling them to participate in society with-
out having to hide or relinquish their ethnic identity seems to pose no general
threat to the welfare state. If so, this should help reduce one important source of
public opposition to a broad range of liberal immigration proposals, whether it
concerns adopting a more open admissions policy, or providing greater legal pro-
tections to noncitizens, or adopting models of multicultural citizenship.
While these findings should reduce opposition to the claims of noncitizens, it is
important to emphasize that they do not necessarily entail support for removing
all distinctions in rights between citizens and noncitizens or for diminishing the

32 Barry, Culture and Equality, p. 8.

immigration, multiculturalism, and the welfare state 297

importance of national citizenship. On the contrary, it may well be that one of the
factors that has enabled some countries to successfully manage issues of ethnic di-
versity and multiculturalism is precisely their commitment to preserving a strong
sense of national citizenship, with a distinctive set of rights and responsibilities.
To see this, consider the main argument that is given for the recognition/
redistribution trade-off hypothesis, which we have called the corroding effect.
In its usual form, the argument goes like this:

d Multiculturalism policies emphasize diversity;

d
emphasizing diversity undermines the sense of common national identity;
d
feelings of national solidarity are necessary for a robust welfare state.

We can imagine two very different responses to this argument. One response,
advocated by some postnational or cosmopolitan theorists, is to accept that
MCPs erode national identity, and hence jeopardize the national welfare state as
currently understood, but to argue that we need to separate out redistribution
from ideas of nationhood.
Redistribution, on this postnational model, should be seen as an obligation owed
to human beings as such, based on their universal personhood and human rights
rather than their status as ‘‘co-nationals,’’ or their shared membership in some na-
33
tional collectivity. Acquiring the legal nationality of a country could continue to
have some vestigial symbolic significance—some immigrants might want to natu-
ralize as a way of officially declaring their loyalty to their country of residence—but
it should not have any bearing on one’s substantive rights. On this view, we cannot
truly accommodate heterogeneity and multiculturalism if we continue to pursue
the nation-building goal of turning immigrants into ‘‘national citizens.’’
We do not share this postnational view. One difficulty with this suggestion—
apart from the fact that it has no hope of being adopted in the foreseeable fu-
ture—is that connecting redistribution to personhood rather than nationhood is
more likely to lead to a leveling down than a leveling up. Contemporary welfare
states typically operate on a model of ‘‘egalitarianism among ourselves, humani-
tarianism for others.’’ All human beings should have their basic human needs met,
for example through foreign aid for the poor abroad or through emergency medi-
cal care for visitors and tourists in our country. But among co-nationals, we aim
for more than this, including equality of economic opportunity, equal capacity to

33 See Yasemin Soysal, The Limits of Citizenship (Chicago: University of Chicago Press, 1994).

298 Will Kymlicka and Keith Banting

participate in political life, and some sense of a minimally adequate standard of
34
living needed to avoid social exclusion. If we disconnect redistribution from na-
tionhood, and assume that visitors and co-nationals should be treated the same
way, the result may not be to extend egalitarianism to nonnationals but rather to
reduce intra-national redistribution to mere humanitarianism.
Since we believe that national welfare states are vital instruments for protect-
ing the vulnerable and ameliorating disadvantage and inequality, our response to
the corroding effect argument is different. We accept that national solidarity is
important to sustaining a welfare state that goes beyond mere humanitarianism,
but we dispute the claim that MCPs necessarily corrode solidarity. There are sev-
eral ways in which MCPs can be designed or supplemented to protect or even
enhance national solidarity. As we will see, however, some of these strategies de-
pend on preserving the legal significance of national citizenship and the political
salience of naturalization, and as such are inconsistent with postnational theories
that seek to eliminate these distinctions.
There are at least three ways in which MCPs may promote greater social solid-
arity. First, MCPs can help to combat the stereotypes and stigmatizations that
currently erode feelings of solidarity across ethnic and racial lines. Critics of
MCPs assume that by heightening the public salience of ethnicity, they necessa-
rily reinforce perceptions of ‘‘difference’’ and ‘‘otherness,’’ creating and reinforc-
ing boundaries between people. But the reality is that such boundaries are
already intensely marked in contemporary societies and are often accompanied
by deeply rooted feelings of prejudice, fear, indifference, and contempt, deriving
from a long history of racism and colonialism in the West. The power of such
boundaries and their corrosive effects on solidarity are no less powerful for being
‘‘private’’ and informal.
Under these conditions, the adoption of MCPs can be seen as a ‘‘destigmatiza-
tion’’ strategy. MCPs may heighten the public salience of ethnicity, but they do so
precisely in order to contest the inherited stigmas that create feelings of distrust
and antipathy between people. MCPs, on this view, focus on our differences, but
do so in a way that allows us to see each other as equally worthy of respect. We
would argue that MCPs often have this reconstructive and destigmatizing goal in

34 Versions of this argument have been articulated in recent works by David Miller (Citizenship and National

Identity [Cambridge: Polity Press, 2000]), John Rawls (The Law of Peoples [Cambridge: Harvard University
Press, 1999]), and Thomas Nagel (‘‘The Problem of Global Justice’’, Philosophy
Public Affairs 33, no. 2 [2005],
pp. 113–47), to name a few.

immigration, multiculturalism, and the welfare state 299

contemporary Western democracies. The emergence of MCPs in the West in the
last forty years is directly linked to the human rights revolution, and to the post-
war commitment to the equality of races and peoples. MCPs have been demanded
by historically marginalized groups as a way of challenging inherited racial and
ethnic hierarchies, and the sort of public recognition they provide is often de-
signed in light of the sort of informal stigmatizations that persist. That is to say,
MCPs do not necessarily protect those aspects of a group’s culture that are seen by
group members as most essential or valuable. Rather, they often recognize those
aspects of a group’s culture and history that have been most dishonored by the
larger society. (This explains the tendency for politics of recognition to merge into
politics of redress and demands for the acknowledgment of historic injustice.)
Viewed this way, the adoption of MCPs can be seen as an indirect tool for
building national solidarity by reducing the inherited antipathies that currently
35
prevent people from feeling solidarity across ethnic and racial lines. Such des-
tigmatizing strategies, by themselves, however, do not explain how a sense of
common national identity arises. They may help citizens view immigrants as
worthy and respectable persons, and as such entitled to humanitarian concern,
but they do not yet explain why these others should be seen as ‘‘one of us’’ for the
purposes of communally bounded redistribution.
To address this problem, the destigmatizing components of MCPs are often
supplemented with more explicit ‘‘nation-building’’ policies. Such policies include,
for example, language training programs, citizenship education in the schools, the
shared celebration of national heroes and holidays, citizenship ceremonies for im-
migrants, and so on. According to David Miller, it is the presence of these nation-
building policies that prevents MCPs from having a corrosive effect. He argues
that ‘‘radical’’ defenders of multiculturalism promote MCPs without simultane-
ously promoting nation building, and the inevitable result is to erode national sol-
idarity and national redistribution. But a more ‘‘moderate’’ multiculturalism
combines MCPs with nation-building policies, so that the potentially corrosive ef-
36
fects of the former are offset by the bonding effects of the latter.

35 Of course, whether MCPs actually achieve this intended goal is another question. We can imagine scenarios
in which the adoption of MCPs would backfire, and reinforce the perception of minority groups as ‘‘needy,’’
‘‘undeserving,’’ ‘‘ungrateful,’’ and ‘‘dependent.’’ Some critics seem to assume that such unintended re-stigmatizing
effects are inevitable. See Fraser, ‘‘Social Justice in the Age of Identity Politics.’’ But the evidence mentioned in the
previous section suggests that MCPs have not systematically had this effect.
36 See David Miller, On Nationality (Oxford: Oxford University Press, 1995), and Citizenship and National

Identity.

300 Will Kymlicka and Keith Banting

 On Miller’s view, these nation-building policies are separate from, but necessa-
rily supplemental to, MCPs. We would go further, however, and argue that MCPs
can themselves serve as a tool of nation building, by providing a focal point for
national pride. The decision to adopt MCPs can be seen as a collective national
project, as something that ‘‘we’’ as a nation chose to do, a national project that
we can take pride in. It is a way of showing that we are a ‘‘modern,’’ ‘‘progres-
sive,’’ ‘‘cosmopolitan,’’ and ‘‘tolerant’’ nation that has transcended outdated ideas
of national homogeneity. The adoption of MCPs has clearly played this role in
many democratizing countries, serving as a symbolic marker of the break with
the narrow and insular nationalism enforced by previous authoritarian regimes.
Indeed, multiculturalism can come to be seen as one of a nation’s defining and
distinguishing characteristics. Multiculturalism arguably plays this role in Canada.
Affirming multiculturalism has become part of what it means to ‘‘be Canadian,’’
37
and, more specifically, part of what it means to be a ‘‘good Canadian.’’ It is ‘‘un-
Canadian’’ to oppose multiculturalism, a betrayal of the national code. Opposing
multiculturalism in Canada is not only or even primarily an attack on particular
minorities, it is an attack on the symbol and substance of Canadian nationhood.
In short, MCPs are likely to sustain, rather than erode, national solidarity
when (a) they contain a destigmatizing dimension, (b) they are supplemented
with nation-building policies, and (c) they are integrated into a national narra-
tive that is a source of collective pride. When these three conditions are in place,
the effects of heterogeneity and recognition are unlikely to be corrosive.
It is important to note, however, that this package is a very ‘‘nation-centered’’
one. It extends rights and recognition to immigrants, and hence reflects a dis-
tinctly multicultural form of nationalism, one in which the national identity has
been modified (and ‘‘thinned’’) to be more inclusive of ethnic heterogeneity. But
equally it is a very nation-centered form of multiculturalism, premised on ac-
cepting larger national frameworks and narratives.
We can see a clear example in the Canadian context. Immigrants can legiti-
mately seek to benefit from Canada’s multiculturalism policies, but they are ex-
pected to do so qua Canadian. It is as Canadians that they can claim
multicultural rights and privileges, as part of the exercise of Canadian citizenship,
and as part of the playing out of a national narrative. As such, immigrants are

37 Will Kymlicka, ‘‘Being Canadian,’’ Government and Opposition 38, no. 3 (2003), pp. 357–85.

immigration, multiculturalism, and the welfare state 301

expected to accept the nation-building policies that accompany multiculturalism
(with its expectation of learning an official language, naturalization, citizenship
education, and so on), and to internalize the national narrative in which multi-
culturalism is portrayed as a distinctive and worthy collective national project.
For immigrants to demand multicultural accommodations while resisting these
nation-building components would be seen as insulting and abusive.
Indeed, in those few cases where minorities do not follow the script, and in-
stead assert multiculturalism as an inherent right divorced from being Canadian,
38
there is often a strong backlash from the larger society. Perhaps for that reason,
most immigrant groups in Canada carefully play along with the script, and insist
that they have no greater desire than to be good Canadians, and that the enacting
of their ethnic identities in public space is their way of proudly and loyally ‘‘being
Canadian.’’
Under these conditions, it is perhaps not surprising that multiculturalism does
39
not corrode national solidarity. But notice that this happy compatibility of di-
versity and solidarity depends on preserving the category of national citizenship
as a reference point for debates on the management of ethnocultural diversity. It
is the willingness of immigrants to make a commitment to becoming Canadian
citizens that sustains public commitment to multiculturalism, and if there were
no tangible distinction between citizens and noncitizens, there would be no
basis for this reciprocal commitment. Becoming a Canadian citizen involves a

38 There was a strong backlash in Canada when an Aboriginal leader (Matthew Coon-Come) asserted defiantly
that he ‘‘wasn’t a Canadian,’’ yet nonetheless expected recognition of various rights from the Canadian state.
The same sort of resentment arises when anglophone Canadians are asked to make accommodations by Quebe-
cois nationalists who, it is widely believed, do not really want to be part of the country.
39 This capacity for multiculturalism to be captured by the nation-state, and to be deployed as a tool of nation

building, has been bemoaned by some postmodernist critics. On their view, the emancipatory potential for
multiculturalism depended precisely on the possibility that it would not be contained with national narratives
or nationalist ideologies, and that it would push us toward a new postnational order that abandoned any op-
pressive fantasies of national cohesion. Instead, multiculturalism has been used to buttress and relegitimize the
nation-state, and invoked as a tool for ‘‘normalizing’’ and ‘‘disciplining’’ minorities as ‘‘national citizens.’’ See
Richard Day, Multiculturalism and the History of Canadian Diversity (Toronto: University of Toronto Press,
2000). This postmodernist critique of multiculturalism stands in stark contrast with the social-democratic cri-
tique represented by Miller and Barry. The social-democratic left criticizes MCPs for eroding national solidarity;
the postmodernist left criticizes MCPs for presupposing and relegitimating nationhood. From our perspective,
postmodernists are right that multiculturalism has often become a tool of nation building, and a tool for nor-
malizing immigrants as national citizens. We would argue, however, that this is legitimate and indeed desirable,
so long as (a) the conception of national citizenship respects the legitimate minority and cultural rights of all
groups, (b) the means used to promote this national identity are morally permissible, and (c) the resulting sense
of national solidarity is used to advance legitimate public goals, including redistribution. For one vision of such
a multicultural nationalism, see Will Kymlicka, Politics in the Vernacular: Nationalism, Multiculturalism, and
Citizenship (Oxford: Oxford University Press, 2001).

302 Will Kymlicka and Keith Banting

significant amount of time, effort, and money, which immigrants would be un-
willing to expend if citizenship carried no tangible benefits. And if immigrants
did not make an observable effort to become citizens, native-born Canadians
would not continue to support multiculturalism.
Of course, this does not mean that there should be gratuitous or arbitrary dif-
ferences between the rights of citizens and noncitizens. Many rights are indeed
owed to people simply on the basis of their personhood, regardless of their legal
nationality: for instance, respect of basic civil liberties, and provision of health
care and education. And other rights are owed simply on the basis of contribu-
tion: if immigrants pay into unemployment insurance, for example, they should
be able to access those benefits. And yet other rights should be based on length
of residency, regardless of citizenship: long-term residents should probably have
guarantees against deportation, for example, even if they were unable or unwill-
40
ing to take out citizenship.
The importance of national solidarity, however, suggests that the state has a
legitimate interest in preserving the category of national citizenship as the locus
of certain tangible rights, such as national voting rights, and responsibilities,
such as jury duty and military service. It also suggests that the status of nonciti-
zens can legitimately be defined in such a way as to encourage and facilitate their
41
transition to full citizenship.
The implications of our argument for the rights of noncitizens, therefore, are
neither direct nor simple. On the one hand, our findings cast doubt on pessimis-
tic claims about the conflict between diversity and solidarity, and this should in-
directly lend support to the struggles of noncitizens, making greater space for an
open-minded consideration of their claims on their own merits. On the other
hand, however, our findings do not cast doubt on the importance of national
solidarity, and our own best guess is that those countries that have done well in
combining diversity and solidarity have done so by linking immigration and

40 There have been several shocking cases where Canada deported men convicted of crimes to Jamaica, even
though they had been brought to Canada from Jamaica as young children.
41 In many cases, the best way to encourage noncitizens to become citizens may be to extend certain proto-cit-

izenship rights to them. For example, extending local voting rights for noncitizens may encourage them to take
a greater interest in the political system generally, and ultimately to take a greater interest in national citizen-
ship. But if so, this would reflect a judgment about how to encourage national citizenship, and not a postna-
tional desire to transcend or decenter national citizenship. And if the evidence showed instead that immigrants
with local voting rights were less likely to take out national citizenship, then it would be legitimate to withhold
local voting rights as an incentive to naturalize.

immigration, multiculturalism, and the welfare state 303

multiculturalism to the promotion of national citizenship. If so, the reserving of
certain rights and responsibilities for national citizens should not be dismissed as
the relic of a nationalist ideology that has been rendered obsolete by diversity
and global migration. Rather, it may be a crucial factor that enables states to deal
with these new challenges.

304 Will Kymlicka and Keith Banting