Real but Unequal Representation in Welfare State Reform

• Author: Armen Hakhverdian,Wouter Schakel,Brian Burgoon
• Published date: 01 March 2020
• Date: 01 March 2020
• DOI: http://doi.org/10.1177/0032329219897984
• Subject Matter: Articles

https://doi.org/10.1177/0032329219897984

Politics & Society

2020, Vol. 48(1) 131 –163

© The Author(s) 2020

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/0032329219897984

journals.sagepub.com/home/pas

Article

Real but Unequal

Representation in

Welfare State Reform

Wouter Schakel

Leiden University and University of Amsterdam

Brian Burgoon

Armen Hakhverdian

University of Amsterdam

Abstract

Scholars have long debated whether welfare policymaking in industrialized democracies

is responsive to citizen preferences and whether such policymaking is more

responsive to rich than to poor citizens. Debate has been hampered, however,

by difficulties in matching data on attitudes toward particular policies to data on

changes in the generosity of actual policies. This article uses better, more targeted

measures of policy change that allow more valid exploration of responsiveness for

a significant range of democracies. It does so by linking multicountry and multiwave

survey data on attitudes toward health, pension, and unemployment policies and data

on actual policy generosity, not just spending, in these domains. The analysis reveals

that attitudes correlate strongly with subsequent changes in welfare generosity in the

three policy areas and that such responsiveness is much stronger for richer than for

poorer citizens. Representation is likely real but also vastly unequal in the welfare

politics of industrialized democracies.

Keywords

economic inequality, representation, welfare politics

Corresponding Author:

Wouter Schakel, Leiden University, Pieter de la Court, Wassenaarseweg 52, Leiden, 2333 AK, The

Netherlands.

Email: w.schakel@fsw.leidenuniv.nl

897984PASXXX10.1177/0032329219897984Politics & SocietySchakel et al.

research-article2020

132 Politics & Society 48(1)

The quality of democratic representation in policymaking, including in welfare state

policy, remains a source of enduring and fundamental controversy in developed

democracies. A key debate involves the extent to which attitudes of citizens on issues

of welfare state reform actually constrain or impel policymakers to alter existing

social policies in line with citizens’ wants. Some studies find solid links between

opinion and policy;1 others have found links either to be unclear, because data limita-

tions constrain causal inference about such links, or to be weak, because real politics

block any simple aggregation of citizen preferences into policy change.2 A second,

related debate concerns inequality in political representation. Recent studies, particu-

larly of the United States, find that the average voter and certainly poorer citizens

have much less influence on subsequent policy changes than wealthier citizens do.3

The principal controversy, here, is whether such unequal democracy applies to key

features of the welfare state and extends to settings outside the United States, where

traditions and more inclusive institutions of governance might yield more equitable

political representation.

Both debates about representation in welfare policymaking remain unresolved, in

part because of serious empirical constraints. Among the most significant has been the

difficulty of matching valid and systematic measures of citizen attitudes toward par-

ticular social policy reforms to measures of downstream change in social policy regu-

lations sufficient to support meaningful causal inferences about representation. The

most careful existing studies have linked attitudinal patterns of particular groups to

change or incidence of policy in subsequent years.4 But even those studies work with

very rough measures of policy change, such as changes in government spending. Such

measures are likely to suffer from what is known as the “dependent variable problem,”

where spending metrics do not fully coincide with regulatory and legislative changes

in social policies about which citizens actually express preferences. Large shifts in

spending per capita can and do swing free of actual policy changes—for instance,

when entitlement spending rises with higher take-up during economic downturns.

In this article we partially overcome such empirical challenges by matching the

attitudes regarding social policy of (various) segments of a population to better esti-

mates of change in the actual generosity of such policy. The former we accomplish by

focusing on several waves of the multicountry International Social Science Program

(ISSP) and its questions about the provision of social policy with respect to unemploy-

ment, pensions, and healthcare. The data provide leverage to measure support for these

features of the welfare state validly and reliably, across respondents’ household

incomes, for a considerable cross-section of countries and period of time. Equally

important, we match these data to measures of changes in the programmatic generosity

of the same policy areas, relying on generosity measures from the Comparative

Welfare Entitlements Database (CWED).5

On the basis of such empirical improvements, our principal expectations are two-

fold. First, we expect that attitudes, on average, are associated with measures of sub-

sequent policy generosity, even if such attitudes are less associated with subsequent

spending. Second, we expect that wealthier citizens influence actual social policy

development more strongly than do poorer citizens, on the grounds that wealthier

Schakel et al. 133

citizens find both formal and informal footholds for pressuring policy change. Our

study supports these hypotheses in a range of models, specifications, and estimators

linking individual attitudes to subsequent changes in generosity related to unemploy-

ment, pensions, and healthcare. We provide evidence that citizen preferences are not

expressed in vain but tend to show up in subsequent shifts in the policy provisions of

welfare states. However, not all citizens are equal in political expression. The poorest

tend to have no statistic ally significant influence on subsequent policy changes,

whereas the wealthiest do. The political economy of welfare state reform in industrial-

ized democracies can thus best be characterized as manifesting real but unequal repre-

sentation for its citizens.

(Unequal) Representation in Welfare State Policymaking?

Long-standing debate among policymakers, activists, and academics of all disciplin-

ary and methodological stripes concerns the degree and character of political represen-

tation of broad publics and citizen attitudes in the actual development of the welfare

state. The debate involves two related controversies.

One controversy questions the extent to which policymaking is responsive to the

political demands of citizens.6 In developed democracies, some modicum of represen-

tation should be broadly present, but in welfare state policymaking and elsewhere that

presence is far from obvious. On the one hand, plenty of research supports the intuition

that public opinion tends to translate into policy change. James Stimson, Michael

Mackuen, and Robert Erikson, for instance, highlight two electoral mechanisms link-

ing citizen preferences and policy outcomes: electoral turnover, an indirect pathway

flowing from citizen preferences to election outcomes and from election outcomes to

policy; and political actors anticipating electoral defeat and adapting to public prefer-

ences so as to avoid it.7 Beyond the electoral arena, organized interests such as unions,

employer associations, nongovernmental organizations, and civic associations can

serve as channels by which public attitudes influence policy outcomes.8 These mecha-

nisms can give life to the role of the median voter in spatial theories of voting9 or to

various strains of pluralism in policymaking.10 Empirically, many scholars have found

evidence of such representation. For many polities and policy areas, studies have

unearthed (causal or associational) links between public opinion and aggregate policy

outputs in democracies.11

On the other hand, much scholarship questions these findings. Some scholars

have judged citizen attitudes to be so diffuse and vague, or vaguely known, as to be

very blunt instruments in policymaking.12 Important critiques of pluralism, further,

have articulated why popular positions are blocked, where politics is the province

of organized special-interest groups, lobbies, and elites in areas including social

policymaking.13 A related view emphasizes the relative autonomy of the state, where

the prerogatives of state actors and institutions eclipse the pressure of citizen-voters

for welfare state policies.14 Consistent with these critiques is empirical skepticism

that views any association between citizen attitudes and aggregate policy outcomes

as (spurious) correlation rather than causation.15