Happiness and the role of social protection: how types of social spending affected individuals’ life satisfaction in OECD countries, 1980–2012

ABSTRACT

The growing literature on individual determinants of subjective well-being has given little attention to political factors. This paper considers the welfare state, and how social expenditure affects individuals’ self-reported life satisfaction. The statistical analysis uses indicators of subjective well-being, reflecting individuals from OECD-countries between 1980 and 2012, with data gathered from the Eurobarometer and the World Values Survey - which are analysed in comparison. The results suggest that social spending should be studied in terms of underlying branches when addressing its implications. The results find social spending to be uncorrelated with levels of subjective well-being when considered in terms of total levels. When considered as types of spending however, a majority of the elements are found to have significant impacts. The findings show mixed results among the two data sets; however, important similarities are found in the way social spending related to health care and poverty are having positive impacts, and spending associated with unemployment and labour market programmes have negative impacts. As the correlations of the underlying elements affect life satisfaction in different directions, total social spending appears to be uncorrelated with subjective well-being, although the true impact depends on which socialpolicies are being promoted through such spending.

KEYWORDS:

• Social spending
• Subjective well-being
• Comparative study
• Multi-level method
• OECD countries

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Oda Nordheim is a PhD-candidate in political science at the Norwegian University of Science and Technology (NTNU). Her research interests are public policy, the welfare state, health policy, social welfare and social inequality. She is currently working on her PhD project: “Social Protection and its Impact on Health: How types of social spending affect self-rated health within OECD-countries”.

Pål E. Martinussen is professor in political science at the Norwegian University of Science and Technology (NTNU) with more than 15 years of experience from health services research. He has numerous publications in international journals on topics as comparative health policy, health reforms, NPM and the medical profession, and hospital management and organisation.

Notes

1 This measure captures quantity and quality of social spending related to pensions, incapacity and unemployment, and is developed by Esping-Andersen (1990).

2 Flavin et al. (2011) controlled for GDP per capita and unemployment rate, while Veenhoven (2000) controlled for national wealth.

3 We also estimated the models with two-way clustered standard errors (country and year), but since this did not alter the results we choose to present the regression models using clustering by country-year.

4 Including: Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden, and the United Kingdom.

5 Most of these countries are however included in the WVS data set.

6 Including Australia, Canada, Chile, Czech Republic, Estonia, Finland, France, Germany, Hungary, Israel, Italy, Japan, South-Korea, Mexico, Netherlands, New Zealand, Norway, Poland, Slovak Republic, Slovenia, Spain, Sweden, Switzerland, Turkey, the United Kingdom, and the United States.

7 Two additional variables which are important predictors of life satisfaction are income and health at the individual level. These are not included in the analyses due to lack of availability in the utilized data sets.

8 Data on economic growth, GDP per capita, inflation and unemployment rate were taken from the OECD statistical database. Missing data are supplemented from the IMF and World Bank databank.

9 The Eurobarometer and WVS are collectively covering all the OECD countries except Iceland.

10 South Korea and Mexico are omitted from this analysis due to missing data.

11 Models with lagged unemployment rates were also tested, however indicating insignificant coefficients which did not prove to drive these results. These were therefore considered unnecessary to include in the paper.