Sub-project 2b: The impact of ethnic diversity on the preferences for redistribution and the demand for welfare services


Ethnic diversity may impact the demand side of redistributive politics though several mechanisms. The main one is a decline in altruism: where an “in-group bias” leads individuals to support redistribution when it benefits people “like them” (see, e.g., Alesina et al., 1999 Alesina and La Ferrara, 2002, 2005, Vigdor, 2004, Miguel and Gugerty, 2005, Habyarimana et al., 2007, and Dahlberg et al., 2012). Accordingly, if individuals think that ethnic minorities are over-represented among welfare recipients and if they feel less altruistic toward an out-group, then a growth in ethnic diversity will decrease support for policies that benefit the less-well off. Also, an ethnically diverse population might affect the welfare state through a segregated demand for welfare services if ethnic diversity implies heterogeneity of preferences. This might lead to under-provision of local public goods through at least two mechanisms; (i) either by generating unfocused lobbying, implying that no agreement can be reached on which goods should be provided, with the consequence that none (or fewer) are provided; or (ii) by the above mentioned “in-group bias”. A further possible implication of this type of segregation following increased immigration is the increased use of privately provided welfare services (see, e.g., Betts and Fairlie, 2003, and Gerdes, 2010). We will, use individual panel data from UK (British Household Panel Study) and Sweden (register data provided by Statistics Sweden) to investigate the extent to which ethnic diversity matters for preferences in these two countries.

First, we will investigate the impact of a change in local ethnic diversity on a change in the willingness to keep the British National Health Service (NHS) universally accessible. The NHS is the perfect policy issue to test the impact of ethnic diversity on support for redistributive programs. It is redistributive; access is non-discriminatory, and overtly so (people know that immigrants receive free healthcare, even if they are not citizens), and it is organized in a way that means that the usage of the NHS is linked to local conditions (if there are immigrants in your neighbourhood, you see them at your local NHS). The type of social policy we look at also minimizes the potential problem of reversed causality, as it is a policy uniformly provided across the territory and thus cannot be expected to shape patterns of residential mobility (unlike schooling for instance). We argue that a rapidly diversifying neighbourhood often implies economic insecurity or sense of downward mobility and lack of satisfaction. If true then people may increase their support for NHS because they feel financially insecure, leading to the effect in the opposite direction to the “in-group bias” hypothesis. In other words, ethnic diversity might increase support for keeping the rich in paying for the NHS, but decrease support for sharing these resources with others. To our knowledge, no one has looked at this potential impact.

Our approach will link a high quality individual-level panel data covering two decades to Census and Labour Force Survey data on the changes in the ethnic composition of one’s neighbourhood; this design allows us to control for individual-level unobserved variables that might explain both residency in a diverse neighbourhood and welfare attitudes.

Second, using Swedish register data, we will examine how the extensive inflow of non-Western immigrants to Sweden during the last three decades has changed the provision of welfare services. To investigate this, we use a refugee placement program that was in effect in Sweden between 1985 and 1994. Within the program, the refugees could not choose their initial location of residence themselves, with the implication that the ethnic mix changed quite drastically (and exogenously) in many municipalities. Using this shock to the composition of the local population, we can examine both the short- and the long-run effects on the level and the quality of publicly provided services, as well as how the mix between public and private provision has changed.

Another way in which an ethnically heterogeneous population might create a segregated use of welfare services is through segregated neighbourhoods. Specifically, if there exist neighbourhood tipping points – meaning that, as a consequence of the minority share in a neighbourhood reaching a certain point, the growth in the majority population drastically decreases – we would see ethnic segregation by neighbourhoods (see, e.g., Schelling, 1971, and Card et al., 2008). Such tipping points would not only create a segregated use of local public services, but could also affect the level and the quality of the services; the local tax base and consequently the local provision of public goods and services will likely be affected if the tipping point creates segregation also in other dimensions. With access to rich, individual-level register data from Sweden, we are in a good position to examine these types of neighbourhood dynamics as a function of ethnic diversity, and how these dynamic processes affect the welfare services provided.


Professor Matz Dahlberg, Professor of Economics, Uppsala University,

Dr Anja Neundorf, Assistant Professor in Politics, University of Nottingham,

Dr Heléne Lundqvist, Assistant Professor, Stockholm University,

Mr Henrik Andersson, Doctoral student, Uppsala University,


Latest findings, new research

Publications search

Search all research by subject and author


Researchers discuss their findings and what they mean for society


Background and context, methods and data, aims and outputs


Conferences, seminars and workshops

Survey methodology

Specialist research, practice and study

Taking the long view

ISER's annual report


Key research themes and areas of interest