Do Labor Market Flexibility and Efficiency Increase Government Social Protection Expenditures?

ABSTRACT

This article examines the relationship between labour market flexibility and efficiency and social protection expenditures. Building on the theories of government expenditures and the empirical literature on the determinants of social protection, the article examines empirically the relationship between flexible labour market and social protection expenditures. The empirical model explains social protection mainly in terms of the demand and supply of social protection, modernization theory and political institutions, namely democracy. Labour market flexibility and efficiency indicators include hiring and firing policies, labour-employer cooperation, wage determination flexibility, redundancy costs, linking pay to productivity, reliance on professional management, and the ratio of women in the labour force. We use a sample of 44 developed and emerging economies for the period 2007–2014 and adopt instrumental variables and two-step generalized method of moments estimation methodologies to account for endogeneity. Estimates show that linking pay to productivity, wage flexibility, and redundancy costs reduce social protection expenditures, while the ease of hiring and firing increases them. The article draws attention to the positive impact that flexible and efficient labour markets and organization-level human-resource policies have on reducing social protection expenditures and increasing fiscal space.

KEYWORDS:

• Social protection
• labour market flexibility
• performance pay
• wage flexibility
• hiring and firing

JEL:

• H53
• J28
• J38
• J58

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

² We use the labour efficiency pillar data of the World Economic Forum’s Global Competitiveness Report in this article. The labour efficiency pillar uses a) cooperation in labour-employer relations, b) flexibility of wage determination, c) hiring and firing practices, d) redundancy costs, and e) effect of taxation on incentives to work to measure flexibility. To measure efficiency in using talent, the report uses f) linking pay to productivity, g) reliance on professional management, h) capacity to retain and attract talent, and i) female participation in labour force. We do not use all the labour efficiency pillar indicators.

³ Our review of the literature finds that most explanations are based on demand-side theories comprising public choice (the median-voter model and theory of the bureaucracy) and Wagner’s law.

⁴ Our analysis shows that age dependency and population growth rates are positively and negatively associated with the percentage of social protection expenditures to GDP, respectively.

⁵ The theory of bureaucrats may also be considered a supply factor of public goods and government expenditures.

⁶ In his empirical model, he accounts for GDP per capita, old-age dependency ratio, and the unemployment rate.

⁷ She explains public social expenditure (percentage of GDP) in terms of unemployment, GDP growth, agriculture employment, government expenditure, democracy indicator, the extent to which the executive branch and legislature is left, trade openness, and global crisis dummies.

⁸ Kellermann (2007) reviews the role of power resources theory in income inequality. His literature review shows that the strength of labour unions and left political parties are associated with larger welfare states, generous public pensions, and lower inequality and poverty (Huber and Stephens 1993; Hicks and Swank 1984; Bradley et al. 2003; Moller et al. 2003).

⁹ The demand-side determinants of social protection they account for in the empirical model include per capita income, industrialization, population growth, globalization factors, income inequality, and the demand for income redistribution.

¹⁰ The growth rate of national expenditures on defence, public order and safety, economic affairs, and housing and community amenities decreases significantly relative to social protection in election years. Castro and Martins (2016) used social protection as the baseline category in regression analysis.

¹¹ On the supply side, they include in the empirical model the tax burden, direct tax revenues, and the budget deficit to capture the fiscal illusion and the size of bureaucracy.

¹² Sáenz, Sabaté, and Gadea (2013) find that trade openness increases public spending needed to offset the worsening income distribution resulting from increased openness.

¹³ We use STATA 16 in estimation and Eviews to conduct Granger-causality tests.

¹⁴ The under-identification test examines whether the excluded instruments are relevant and correlated with the endogenous regressors. The null hypothesis is the model is under-identified. A rejection of the null hypothesis indicates the model is identified.

¹⁵ The joint null hypothesis for the over-identification test is the validity of instruments, i.e. uncorrelated with the error term, and that the excluded instruments are correctly excluded from the estimated equation.

¹⁶ The number of work weeks paid are in parentheses.

¹⁷ Selection of fixed as opposed to random effects is based on the Hausman specification test results.

¹⁸ The p-value for Hansen test statistic is 0.088.

¹⁹ Agnello and Sousa (2014), Ball et al. (2013) and Mulas-Granados (2005) point out that cuts in social protection expenditures increase income inequality. This suggests a possible reverse causality between SP and INEQUALITY. Granger-causality test result does not reject the null hypothesis that SP does not cause INEQUALITY.

²⁰ We obtain positive and negative coefficients when undertaking OLS and FE regression of SP on INCOME.

²¹ Since we take the first difference of dependency (D.Population), we can think in terms of changes demographic composition.