Measurement Validity in Comparative Welfare State Research: The Case of Measuring Welfare State Generosity

AuthorSimon Grundt Straubinger,Jon Kvist,Anders Freundt
Publication Date01 December 2013
Date01 December 2013
Jon Kvist, Simon Grundt Straubinger and Anders Freundt*
Examining the generosity of welfare states and individual benefi t schemes is a
classical task in comparative welfare state studies. Th

ree types of welfare states
can be discerned based, in part, on their level of benefi t generosity. Although
signifi cant advances have been made in the development of measures of welfare state
generosity, this progress has not been without its challenges and limitations. In this
article, the authors examine two sets of limitations related to measurement validity
in comparative welfare state research: securing content validation and ensuring
comparability across time and place. Th

rough the use of illustrative examples to
compare the situation of the unemployed in fi ve European countries across several
income levels and two family types, we demonstrate that, by profi ling and stacking
public benefi ts using the OECD Tax-Benefi t micro-simulation model, we are able
to carry out a more informed analysis of the redistributive strategies of the welfare

Keywords: income packaging; micro-simulation; profi ling; stacking; unemployment
How do we measure the generosity of welfare states? How do we avoid comparing
apples and pears? Getting the right answer to these two questions is important in
Jon Kvist, University of Southern Denmark; address: Campusvej 55 5230 Odense M;
phone: +45 51718561; e-mail:; Simon Grundt Straubinger, University of Southern
Denmark; address: Kragemosen 42 5250 Odense SV Denmark; phone: +45 22564154; e-mail: Sistr08@; Anders Freundt, University of Southern Denmark; address: Rådmandsgade 40c, 5,
lejl. 133 2200 Copenhagen N Denmark; phone: +45 51683932; e-mail:
European Journal of Social Security, Volume 15 (2013), No. 4

Jon Kvist, Simon Grundt Straubinger and Anders Freundt
comparative welfare state studies, not least in studies of their causes and consequences,
and for policymakers interested in designing benefi t systems that provide adequate
benefi ts without creating unacceptable work and saving disincentives. Th
e purpose of
this article is to identify the limitations of conventional measures used in comparative
welfare state research on benefi t generosity and to propose ways of overcoming these
e starting point of much comparative welfare state research is the idea of three
worlds of welfare capitalism put forward and advanced by Gøsta Esping-Andersen
(1990; 1999). Th
e three types of welfare states are characterised in terms of the Liberal,
the Conservative and the Social Democratic welfare regimes that are represented
by the Anglo-Saxon, the Continental and the Scandinavian countries respectively.
When investigating how these types of welfare states came about, or what their
consequences are, researchers have oft en looked at the level of generosity, i.e. how
much recipients actually receive in benefi ts. Access to the benefi ts and obligations
for recipients are other parameters that may be included in the evaluation of social
security systems, but that is beyond the scope of this paper (see Kvist 2007: 475 et seq.).
In short, liberal welfare regimes provide meagre benefi ts on the basis of need as a last
resort when markets and families have failed. Conservative welfare regimes provide
generous benefi ts to labour market insiders and meagre benefi ts to outsiders. Social
Democratic welfare regimes provide benefi ts on the basis of citizenship. Low-income
groups receive generous benefi ts, while middle-income and high-income groups
receive relatively less generous benefi ts.
In other words, the generosity of the social security system is not equal for
every citizen but varies for diff erent socio-economic groups in the diff erent welfare
regimes. We can illustrate this heterogeneous impact of the welfare state by referring
to four diff erent strategies of equality. Th
e ‘Robin Hood strategy’ takes from the rich
and gives to the poor, while more is given to rich than the poor in the ‘St. Matthew
strategy’. Th
e ‘quid pro quo strategy’ or ‘something for something’ approach implies
that one receives according to what one pays in. Finally, the ‘egalitarian strategy’
gives equally to everybody, which translates into more generous benefi ts for the poor
than for the rich (Korpi and Palme 2004). Figure 1 shows how these four strategies of
equality translate into four distinct profi les, measured in terms of the generosity of
benefi ts compared with previous earnings. Note that the four profi les are ideal types
in a Weberian sense; no country is likely to be an exact copy of any one of the ideal

Measurement Validity in Comparative Welfare State Research:
e Case of Measuring Welfare State Generosity
Figure 1. Four strategies of equality measured by the generosity of social benefi ts as a
percentage of previous earnings

Coupling the redistributive strategies with our welfare regimes we would expect the
Liberal welfare model to be associated with an egalitarian strategy, the Conservative
model the quid pro quo strategy, and the Social Democratic a combination of the
Robin Hood strategy for the poor and the quid pro quo strategy for middle and high
income earners.
We use this theoretical framework as our starting point to examine whether
conventional measures on benefi t generosity actually refl ect these hypothetical
strategies. We argue that this is not the case. In the following analysis we fi nd that many
of the conventional measures, including some of the measures in the Social Citizenship
Indicator Programme (SCIP) and the Comparative Welfare Entitlements Dataset
(CWED), are limited with regard to the dual challenge of securing content validation
and ensuring comparability. Securing content validation entails understanding
how we can adequately capture the meaning of theoretical concepts and analytical
constructs. Ensuring comparability entails creating measures that are comparable in
diff erent contexts, i.e. in diff erent countries or at diff erent points in time (Adcock and
Collier 2001). To overcome such limitations, we use profi ling and stacking analysis.
is means simultaneously accessing the relevant tax-benefi t situations for diff erent
relevant groups in the population. With an illustrative application of unemployment
insurance in fi ve European countries we suggest how profi ling and stacking analysis
can address the dual challenge of content validation and establishing comparability
over time and space.
European Journal of Social Security, Volume 15 (2013), No. 4

Jon Kvist, Simon Grundt Straubinger and Anders Freundt
e article is set out as follows. In the fi rst section, we set out in more detail
the conventional measures used in comparative welfare state research i.e. social
expenditure, SCIP and CWED. Th
e second section presents our methodological
refl ections on profi ling, stacking and the OECD tax-benefi t micro-simulation model.
In the third section, on content validation, we attempt to illustrate the limitations of
the single case indicator, the country group indicator and the composite indicator
by comparing the tax-benefi t position of diff erent socio-economic groups using the
example of unemployment insurance. In the fourth section, on comparability, we
demonstrate that the conventional measures may not always have the same meaning
over time and across countries, and how equivalence may be established by making
use of context specifi c domains of observation and adjusted common indicators.
Finally, we off er some concluding remarks on the challenges of measurement validity
on welfare benefi t generosity in future comparative welfare state research.
In broad terms, there are two types of measures on benefi t generosity, namely social
expenditure data and institutional data. Social expenditure data is based on the
monetary input into benefi ts, whereas institutional data looks at the level of individual
benefi t schemes. We argue that other types of measures of generosity like the extent
of poverty or inequality refer to the consequences or outcomes of benefi ts rather than
express benefi t generosity as such.
e most frequently used measure of welfare state generosity is undoubtedly social
expenditure as a proportion of GDP. Th
e oft en implicit assumption is that generous
welfare states spend more than non-generous welfare states. Since the 1970s public
social expenditure as a proportion of GDP has been by far the most frequently used
measure in comparative welfare state research, with empirical data easily available
from various sources. Arguably, this measure has been valuable in that it has increased
our understanding of diff erences between countries and among the correlates of
variations in social expenditures. Th
ere are, however, major shortcomings of public
social expenditure data when used in comparative welfare state research as it not only
fails to capture all that is relevant to generosity but it also captures elements that are
not relevant to generosity and the welfare state.
e fi rst problem is that public social expenditure...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT