Public Sector Performance in Economic and Social Policy: An Evaluation of 35 Countries

Economic and social policy can reduce unemployment and poverty while also improving life expectancy and happiness, provided that the appropriate policies are adopted. This article explores a recent study of public sector performance in various policy areas such as education, health and housing in 35 countries, including all European Union member states. It considers findings on social security, employment, income and wealth. Countries that succeed in these areas often rely on active labour market policies and target their social spending.

Since 2004, the Dutch Ministry of the Interior and Kingdom Relations has commissioned international comparative studies of public sector performance resulting in reports in 2004, 2012, 2015 and, most recently since 2022.

The latest study has been managed by the European Institute of Public Administration (EIPA) and runs in two waves until 2026. It covers 35 countries (all EU member states, the UK, Switzerland, Norway, Iceland, the USA, Canada, Australia and New Zealand) and nine policy areas: education, health, housing, economy and infrastructure, social safety, environmental protection and climate change, sports, and public administration.¹ This article presents the findings of the chapter on social security, employment, income and wealth (Dauderstädt, 2024).

Structure, indicators and methods used

The analysis of all policy areas follows an identical structure. The inputs and activities are government policies, the outputs are their immediate results, and the outcomes are the desired economic and social developments, which, in turn, influence three final goals, namely life expectancy, satisfaction with life (happiness) and trust in government. These goals are important in themselves but also from the viewpoint of governments in democracies that need the support of voters. All elements are quantitatively measured using appropriate indicators.

As very different countries are compared (for instance, the USA and Estonia), appropriate means that are not absolute values but growth rates and relative shares must be used. In our case, data for 28 indicators were collected covering about 14 years following 2007.² The policy area was divided in two major fields: the first is employment, income and wealth, and the second is social security. The outputs and outcomes in the first field largely depend on markets where policies have limited influence, whereas in the second field, public policies mostly determine the results. Table 1 gives an overview of the 28 indicators analysed.³

The way in which inputs led to outputs, outputs led to outcomes and how these influenced the three final goals was investigated through correlation analysis. In this article, only some correlations are presented and discussed. Correlations cannot establish causality but point out probable connections. After the correlation analysis, we compare the overall performance of all 35 countries and contrast our findings with other international studies. To identify best practices, a deeper analysis of the countries that perform best in preventing or reducing unemployment and poverty is added.

The drivers of life expectancy, happiness and trust

The three ultimate goals are correlated with central outcomes such as the growth and distribution of income, and unemployment. Starting with life expectancy, Figure 1 shows that higher economic growth tends to improve life expectancy. This correlation is hardly surprising, but some outliers are notable: the US is the only country with declining life expectancy despite decent growth. The top performers are the Baltic countries that have enjoyed strong growth since 2007 – albeit from a very low level – and have added significantly more life years than other post-communist countries with similar growth rates.

While higher incomes are generally associated with longer life expectancy, large income inequalities within a country tend to reduce overall life expectancy. The correlation delivers a trend line indicating that, when inequality rises by ten Gini index points, life expectancy is likely to decline by 1.7 years. This finding matches similar earlier assessments (De Vogli et al., 2005).

Even more pronounced is the effect of poverty (see Figure 2). Both life expectancy and poverty are represented by their averages over the period under consideration. A poverty rate that is one percentage point higher lowers the life expectancy by about four months.

Considering happiness (or life satisfaction), the picture is different. Happiness and GDP growth are negatively correlated. This counterintuitive result matches the more general findings of the happiness research (Easterlin Par-adox) that show that above a certain level of income per capita (about €30,000 (PPP)), an even higher income does not increase happiness or may even reduce it (Rustichini & Preto, 2014). However, if we look at how life satisfaction evolves when the gross national income per capita (measured at PPP) grows faster than the average of our country sample, a positive correlation emerges. This correlation is mainly driven by the countries of Central and Eastern Europe (CEE) that combine higher growth with clear rises of happiness,⁴ albeit both starting at low levels. A group of slow growing rich countries confirm the sceptical findings of Easterlin (1974) and Rustichini and Preto (2014). An interesting outlier is Ireland, whose outstanding growth was accompanied by declining happiness.

Inequality and poverty are both negatively correlated with happiness. A ten-point rise in the Gini index is, on average, accompanied by a decline in happiness by almost four points. The trend line indicates that, on average, a rise in inequality by ten Gini index points reduces the happiness score by almost 0.7 points. This result matches with the findings of Wilkinson and Pickett (2010), who posit that more equal societies are happier. The effect is even stronger regarding the poverty rate when we compare average levels. The happiest countries are in Scandinavia while the CEE countries with their high poverty rates are still quite unhappy despite big improvements.

Two other correlations are noteworthy: although unemployment is a major problem for societies and individuals, work as such is not an unmitigated benefit, but rather often a necessity to earn an income. This ambiguous role of labour is reflected in the relationship between work and happiness. On the one hand, unemployment is correlated negatively with happiness as one would expect. But, on the other hand, the number of hours worked per person is correlated negatively, too. People who work fewer hours in their job are happier. On average, ten additional percentage points of unemployment (equalling the difference between Finland and Spain) lower the happiness score by one point, equivalent to 400 more hours worked per year per person (equalling the difference between Austria and Greece).

The picture is less clear when examining trust in government. The correlation between growth and the change in trust is relatively weak. The (negative) correlation with inequality and poverty is stronger. A ten-point rise in the Gini index (higher inequality) leads to a loss of trust by about 8.5 points. A similar picture emerges, when we compare the changes of inequality and trust. When inequality increases (higher Gini), trust in governments tends to decline, with a rise in the Gini index by one point leading to a decline of trust by 1.9 points. Both findings are supported by research, e.g. from Acemoglu and Robinson (2006), Solt (2008) or Krieckhaus et al. (2014). The negative correlation between poverty and trust in government is similarly strong.

The conclusion from these correlation exercises is clear: to get better results regarding the three ultimate goals, governments should try to raise the national income and distribute it more equally. The next section looks at the policies that might help achieve inclusive growth.

How well do policies work?

Monetary and fiscal policies represented by indicators such as central bank policy rates and budget deficits are only weakly correlated with growth and employment as they are mostly used to counteract a recession. Thus, it is hardly surprising that no strong correlation between budget deficits and GDP growth (total over the whole period) can be observed. The trend line indicates that, on average, one additional percentage point of deficit spending increases growth, but by 0.05 percentage points. However, most of the 35 governments in our sample used these policies in the 2009 global financial crisis and coronavirus pandemic.

Regarding tax policies, the study showed that there were only small changes during our reference period. A correlation that compared the top income tax rate with the growth rate of the share of wealth owned by the richest 10% between 2007 and 2021 indicated that if one in-creases the top income tax rate by ten percentage points, the growth of wealth of the richest 10% declines by one percentage point. A similar correlation exercise with the change in the Gini index (distribution of income) showed a much weaker correlation. Interest rates affected the overall growth of wealth as the value of assets increased when interest rates declined as they did dramatically in the period of observation.

Given the fact that most outputs and outcomes regarding employment, income and wealth largely depend on (global) markets, the limited effect of national policies is hardly surprising. In the area of social security, where public policies are dominant, one should expect more pronounced effects. Our analysis delivers sobering results.

The prominent input here is social spending (as a percentage of GDP). Comparing social spending and the poverty rate (both average 2007-2021) shows that more social spending is likely to reduce poverty, but only to some extent (see Figure 3). Increasing the share of social spending by ten percentage points of GDP lowers, on average, the poverty rate by 5.5 percentage points.

The relatively weak impact of social spending on poverty is probably due to the fact that most social protection systems try to maintain former income levels rather than equalise incomes. The levels of most pensions, unemployment benefits or sickness benefits are linked to former incomes, usually wages, thus “protecting” income disparities. Some countries target social benefits better by means-testing them. But this increases the administrative costs which make up, on average, about 2.5% of social spending. However, a study by Stefan (2015) shows that high administrative costs are positively correlated with the degree of poverty reduction.

Governments redistribute market income through taxes and social transfers, thus transforming the income distribution of market income into disposable income. The difference between the two Gini coefficients (of market and disposable income) indicates the redistributive effort of governments. Correlating that effort with the poverty rate, the indicator delivers a trend line where reducing the inequality by 0.1 (Gini coefficient difference) lowers the poverty rate by four percentage points. The impact of redistribution is limited because a lot of taxation is regressive, in particular VAT and “sin taxes” on alcohol or tabaco. Social security contributions are also not progressive (contrary to most income tax regimes), but are flat rates, often capped at certain income thresholds. The most generous welfare states (Scandinavia) rely more on regressive taxes than more frugal ones like the US.⁵

How did public policies affect the three ultimate goals of life expectancy, happiness and trust in government? To assess their impact, we constructed a government policy score that is a composite of the share of income tax of total tax revenue, as well as the top income tax rate, the minimum wage, the strictness of employment protection legislation and the share of social spending in GDP. To produce a consistent indicator, we divided the value of the indicator for each country by the average for all countries. These normalised values for all five indicators are summed up and divided by five. Due to the described normalisation method, the values of the (government) policy score range between 0.7 and 1.3 (average = 1), with the CEE countries showing very low scores. We correlated both average levels and the change over the period 2007-21 of the policy score and the three ultimate goals. The correlation is positive but weaker for the changes. The relationship is strongest regarding life expectancy which increases by 1.4 years for each 0.1 increase in the score (Figure 4).

Turning to life satisfaction/happiness, the correlation becomes weaker, though still positive. On average, a rise in the policy score by 0.1 increases the happiness score by almost 0.3. The weakest, but still positive correlation can be observed between the policy score and the trust in government. A rise in the score of 0.1 points increase the trust value by 0.28.

Our correlation analysis provides a picture about the likelihood of the success of certain policies that confirms a general positive effect of the economic and social policies considered. But most scatterplots display a large dispersion of country cases, often with remarkable outliers. Thus, we look first at the overall performance of the 35 countries in our sample and, second, at the policies adopted by the best performing countries.

Comparing countries’ performances

To compare the countries‘ performances, we create three composite scores, one economic, one representing our three final goals (life expectancy, happiness and trust) and one “total score” combining the previous two. The economic score also consists of three components: average GDP per capita growth, changes of unemployment and inequality (Gini). All values are normalised in the same way as for the policy score (see above). Table 2 shows the results. To check our findings, we compare them with two other indices, the Social Progress Index (SPI) and the Human Development Index (HDI). The SPI is based on a much bigger set of indicators resulting from a huge effort of data gathering. The respective column in Table 2 shows the changes of this index between 2014 and 2022 (except Luxembourg and Malta). The results are often close to our findings, but the Mediterranean countries score much better as the SPI does not include our economic indicators but a much broader set of indicators for basic human needs, wellbeing and opportunities (rights). The only indicator common to the SPI and our set is life expectancy. The HDI is composed of three indicators (income, life expectancy and education), two of them (except education) being elements of our score, too. Thus, the HDI scores tend to confirm our findings, albeit with some exceptions: Canada, Italy, Luxembourg, Norway and Spain score clearly better; Germany, Lithuania and Slovakia score worse. The differences probably result from different scores regarding education.

As shown in Table 2, the clear top performers are the three Baltic countries, Poland, Bulgaria and Romania (all with values above two). The losers are Greece, Spain, the US (mainly due to declining life expectancy), Belgium and the UK (mainly because of lacking trust and happiness). Regarding social progress, the Mediterranean countries performed better while the Anglo-Saxon countries show relatively low scores.

The big differences between the two EU peripheries (CEE and Mediterranean) result from the impact of two crises: first the financial crisis and the subsequent sovereign debt panic, and the coronavirus pandemic – both of which affected the southern part of the EU much more than the eastern part (Dauderstädt 2021a, 2022). Drawing lessons regarding economic policy is easier for the southern periphery than for the eastern one. During the sovereign debt crisis, the EU should have used the ECB as a lender of last resort in a timely and generous way and avoided austerity policies. Eastern European countries benefitted from a low starting point after the collapse of communism and large inflows of aid and investment following EU accession.

These findings, reflecting developments since 2007, conceal the fact that the actual socio-economic conditions in Western Europe and Scandinavia is usually better than in CEE countries. If we use the current values of our five indicators, a more familiar pattern emerges with the Nordic and most Western European countries coming out on top. The outliers among the rich countries are the USA and the UK, where the actual picture confirms the poor longer-term results presented in Table 2. Thus, to identify best practices, we must compare countries with respect to the level as well as the change during our period of analysis.

Fighting unemployment and poverty: Lessons from successful countries

Which countries performed best in fighting unemployment and poverty?⁶ Table 3 shows the six countries we chose to analyse more in depth, to identify best practices. The selection includes countries with very low average levels and those with strong declines in unemployment and poverty rates. The first four countries (Czechia, Iceland, Norway and the Netherlands) combined very low levels of both indicators; the other two (Germany and Poland) were included because of strong declines. Iceland and the Netherlands achieve low poverty with little social spending (see Figure 3).

For each country, one must consider first the general economic and social conditions that affect the development of unemployment and poverty such as, for instance, demography or external economic circumstances. Only after considering these factors, the role of government policies and institutional arrangements can be reliably identified.

Tables 4 and 5 give an overview of the economic and social policies of the selected countries. In summary, with regards to unemployment, all analysed countries used monetary and fiscal policies to stabilise and strengthen demand and employment during recessions. Labour market policies accompanied these macroeconomic measures. During the pandemic, many countries subsidised furloughs, thus stabilising employment throughout the crisis. Many of the successful countries used employment protection legislation, minimum wages and strong unions, which ensure decent wages and reduce market income inequality. The possible exception is Germany, which reduced high unemployment after 2003 through labour market reforms that weakened social protection. But other factors contributed to the decline in unemployment, too: unit labour costs were controlled by strong unions and collective agreements (Dustmann et al., 2014). When Germany introduced a statutory minimum wage in 2015 for the first time, unemployment did not increase.

The rate of poverty depends on the distribution of disposable income which results from the distribution of market income and its redistribution through taxes and social spending. Market incomes stem from capital (profits, rents, etc.) or labour (wages). A more widespread distribution of wealth reduces inequality when, for instance, more people own their homes or have savings for retirement, as is the case in the Dutch pension system. Minimum wages and strong unions can reduce the number of working poor. But redistribution remains crucial to poverty reduction in most cases, albeit with less effect than desired (see Figure 3). One possible cause is that tax policies have relatively little effect. In the best case, the poor pay less or no income taxes. But most government revenue comes from VAT and social security contributions. Both sources are not progressive and have flat rates that favour the rich, who do not consume as much, save more and often pay social insurance premiums only up to a certain income threshold and not at all on capital income (profits, rents). That leaves social expenditure and income support. Social protection often aims at maintaining previous market income levels rather than avoiding poverty and is not means-tested. The successful countries target their social spending better or distribute fixed amounts of benefits that are relatively more valuable to poor households.

• 1 The respective chapters are published on the EIPA website and accompanied by an interactive dashboard presentation of the large set of indicators and data. To get a more comprehensive picture, please see https://docs.eipa.eu/benchmark-study/public-sector-performanceprogramme, particularly the chapters on social security systems (Knol et al., 2024) and economy, infrastructure and science, technology and innovation (Diaz & Clinton, 2024) that are directly relevant for the issues discussed here.
• 2 Data availability that ended in 2021 or even earlier often prevented a deeper analysis of the effects of the coronavirus pandemic, the Ukraine war and the subsequent energy crisis.
• 3 The complete data tables of all 28 indicators covering all 35 countries and the years after 2007 (as far as available) can be found in Dauderstädt (2024).
• 4 https://worldhappiness.report/
• 5 See also Lindert (2021).
• 6 This section draws on the second phase of the EIPA study (Dauderstädt, 2025).

References

Acemoglu, D., & Robinson, J. (2006). Economic Origins of Dictatorship and Democracy. Cambridge University Press.

Bittorf, M. (2017). KfW Research Economics in Brief Labour market in the Czech Republic sets standards. KfW Research Economics in Brief.

Dauderstädt, M. (2021a). Cohesive Growth in Europe: A Tale of Two Peripheries. Intereconomics, 56(2), 120–126.

Dauderstädt, M. (2021b). Wirtschaftsprogramme gegen die Pandemie-Krise – Deutschland im internationalen Vergleich. Wirtschaftsdienst, 101(5), 362–368.

Dauderstädt, M. (2022). We are not (at) all in the same boat: Covid-19 winners and losers. In B. Vanhercke, & S. Spasova (eds), Social policy in the European Union: state of play 2021 Re-emerging social ambitions as the EU recovers from the pandemic.

Dauderstädt, M. (2024). Social Security, Employment, Income and Wealth. EIPA. Public Sector Performance Programme 2022-2025. An International Benchmarking Study. Sub-Study 2023, 176–271.

Dauderstädt, M. (2025). Social Security, Employment, Income and Wealth. EIPA.

De Vogli, R., Mistry, R., Gnesotto, R., & Cornia, G. A. (2005). Has the relation between income inequality and life expectancy disappeared? Evidence from Italy and top industrialised countries. Journal of Epidemiology and Community Health, 59(2).

Diaz-Fuentes, D., & Clifton, J. (2024). Economy, infrastructure and science, technology and innovation. In Public Sector Performance Programme 2022-2025. An International Benchmarking Study. Sub-Study 2023, 176–271. EIPA.

Dustmann, C., Fitzenberger, B., Schönberg, U., & Spitz-Oener, A. (2014). From Sick Man of Europe to Economic Superstar: Germany’s Resurgent Economy. CDP 06/14.

Easterlin, R. D. (1974). Does Economic Growth Improve the Human Lot? Some Empirical Evidence. In P. A. David, & M. W. Reder (Eds.), Nations and households in economic growth: essays in honor of Moses Abramovitz (pp. 89–125). Academic Press.

European Commission Directorate-General for Employment, Social Affairs and Inclusion. (2023). The future of social protection and of the welfare state in the EU. Publications Office of the European Union.

Eydal, G. B., & Gislason, I. V. (2014). Family Policies: The Case of Iceland. In M. Robila (ed.), Handbook of Family Policies Across the Globe. Springer Science+Business Media.

Gielen, A. C., & Schils, T. (2024). Non-Standard Employment Patterns in the Netherlands. IZA Policy Paper, No. 77.

Green, M., Harmacek, J., & Krylová, P. (2019). Social Progress Index. Executive Summary.

Green, M., Harmacek, J., Krylová, P., & Htitich, M. (2022). Social Progress Index. Executive Summary.

Klinker I., & ter Weel, B. (2024). Wages and Employment in the Netherlands, 2017-2023. IZA DP, No. 17049.

Knol, J., van Berkel, K., Schoenmaker, F., van Vuuren, D. (2024). International best practices in social security systems. EIPA Public Sector Performance Programme 2022-2025. International benchmarking study. Sub-study 2024.

Krieckhaus, J., Son, B., Bellinger, N., & Wells, J. (2014). Economic Inequality and Democratic Support. The Journal of Politics, 76, 139–151.

Lewandowski, P., & Magda, I. (2023). The labour market in Poland, 2000−2021. IZA World of Labour, 2023, 426.

Lindert, P. (2021). Making social spending work. Cambridge University Press.

Martin, J. P. (2015). Activation and active labour market policies in OECD countries: stylised facts and evidence on their effectiveness. IZA Journal of Labour Policy, 4(4).

Nielsen, O. A. (2020). The labour market in Norway, 2000–2018. IZA World of Labour, 2020, 424v2.

OECD. (2018a). OECD Economic Surveys: Germany 2018. OECD Publishing.

OECD. (2018b). OECD Economic Surveys: Netherlands 2018. OECD Publishing.

OECD. (2019). OECD Economic Surveys: Iceland 2019. OECD Publishing.

OECD. (2020). OECD Economic Surveys: Poland 2020. OECD Publishing.

OECD. (2021). OECD Economic Surveys: Netherlands 2021. OECD Publishing.

OECD. (2023a). OECD Economic Surveys: Czech Republic 2023. OECD Publishing.

OECD. (2023b). OECD Economic Surveys: Poland 2023. OECD Publishing.

OECD. (2023c). OECD Economic Surveys: Iceland 2023. OECD Publishing.

OECD. (2023d). OECD Economic Surveys: Germany 2023. OECD Publishing.

OECD. (2023e). OECD Economic Surveys: Netherlands 2023. OECD Publishing.

OECD. (2023f). Pensions at a Glance 2023: OECD and G20 Indicators. OECD Publishing.

OECD. (2024a). Taxing wages.

OECD. (2024b). OECD Economic Surveys: Norway 2024. OECD Publishing.

Ólafsdóttir, K. (2020). The labour market in Iceland, 2000-2018. IZA World of Labour, ISSN 2054-9571. Institute of Labour Economics.

Pavlovaite, I. (2018). Social and Employment Policies in the Czech Republic. European Parliament.

Porter, M., & Stern, S. (2014). Social Progress Index. Executive Summary.

Ranaldi, M. (2025, April). Global Distributions of Capital and Labor Incomes. Capitalization of the Global Middle Class. World Development, 188, 106849.

Rustichini, A., & Preto, E. (2014). GDP and life satisfaction: New evidence.

Solt, F. (2008). Economic Inequality and Democratic Political Engagement. American Journal of Political Science, 52(1), 48–60.

Stefan, G. M. (2015). A brief analysis of the administration costs of national social protection systems in EU member states. Procedia Economics and Finance, 30, 780–789.

Vecerník, J. (2001). Labour market Flexibility and Employment Security, Czech Republic. ILO Employment Paper, 2001/27.

Vecerník, J. (2007). The Czech labour market: Historical, structural and policy perspectives. Prague Economic Papers, 3.

Wilkinson, R., & Pickett, K. (2010). The Spirit Level. Penguin.