Conventional wisdom is that centralised wage bargaining usually reduces earning and income inequalities. In a survey of multidisciplinary studies on the determinants of income distributions, Atkinson and Brandolini (2006) list sociological and political science studies including Wallerstein (1999), Rueda and Pontusson (2000), and Mahler (2004) that conclude that centralised wage setting, and co-ordination in bargaining, reduce income inequality.
But for labour markets in which there are large differences in productivity across regions, and between urban and rural areas, binding national wage agreements may increase the dispersion of earnings and incomes. Centralised bargaining compresses nominal earning dispersion, but this tends to increase real earning dispersion after accounting for differences in the cost of living across regions. Nominally compressed wage structures aligned to market-clearing wages in high-productivity areas also increase income inequality because they generate large misallocations, notably creating more unemployment in low-productivity areas.
Italy and Germany
As we discuss in a recent paper (Boeri et al. 2019), we can learn about the relationship between centralised bargaining and income distribution by comparing wage-setting institutions and their effects on regional labour markets in Italy and Germany. Following reunification in Germany, the countries have similar geographical differences in productivity. In Italy, the north is more productive than the south. In Germany, the west is more productive than the east. The Italian north-south productivity gap is remarkably similar to the German west-east gap. In 2014, the difference in mean value added between the northern Italian and southern Italian firms was 19.0%; the corresponding difference between western and eastern German firms was 19.9%.
The two countries have adopted different models of wage bargaining. Italy sets wages based on nationwide contracts that allow for limited local wage adjustments. This means that, in each sector, firms in high-productivity and low-productivity areas face largely the same wage schedule. Germany has a more flexible system that allows for local bargaining. Due to concerns about lower productivity in the east, since 1996 Germany has used 'opening clauses' that allow firms to negotiate locally with unions, potentially undercutting nationwide agreements (Schnabel 1998).
As a result of these different bargaining structures in the two countries, nominal wage dispersion is lower in Italy than in Germany. In Italy, there is much stronger wage equalisation across provinces. After controlling for worker characteristics, we find that the 90-10 percentile difference in mean wages across provinces is 10.3% in Italy, and 42.9% in Germany. The mean wage difference between north and south in Italy is 4.2%, while the mean west-east difference in Germany is 28.2%.
German firms are significantly more able to adjust nominal wages to local productivity than Italian firms. We find that wages in Germany respond to differences in local value added four times as much as in Italy (the elasticity of log wages with respect to log value added is 0.19 in Italy, and 0.73 in Germany).
Different spatial equilibria
In Italy, where nominal wages cannot fully adjust, provinces with low productivity have higher non-employment rates. This is because firms in provinces in which productivity is low need to pay wages above the local market-clearing level.
In Germany, where wages can adjust more to local productivity, unemployment is much less responsive to local differences in value added (the elasticity of unemployment to value added in Italy is almost six times larger than in Germany).
At the same time, low-productivity provinces in both Italy and Germany have lower housing prices, inasmuch as workers move from low-productivity to high-productivity areas. This generates a positive relationship between local productivity and housing prices, driving cost-of-living differences across areas.
In Italy, where nominal wages are compressed across provinces, the purchasing power of wages (nominal wages deflated by the local cost of living) is lower in northern provinces than in southern provinces, since the south has low housing costs but similar nominal wages. By contrast, in Germany, we do not see that real wages in the west are lower than the east, since nominal wages are spatially more flexible. These striking differences in the way real wages react to differences in productivity in the two countries are illustrated in Figure 1.
Figure 1 Differences in real wage response to variation in value added, Germany and Italy.
Source: Boeri et al. (2019).
National wage contracts have therefore created a spatial equilibrium in Italy in which workers queue for jobs in the south. If they find a job, they are better off than their colleagues in the north in terms of real wages, but while queuing they are not employed. This results in a misallocation of resources, with high unemployment and a high dispersion of incomes.
We have tried to quantify the aggregate costs stemming from spatial misallocation in Italy in terms of forgone aggregate earnings and employment. Estimates from a counterfactual scenario in which Italy adopted a system similar to Germany – with the Italian relationships between wages and value added and between non-employment and value added the same as those observed in Germany – indicate that average wages in southern provinces would decrease by 5.9% (or 53 cents an hour), while southern employment would increase by 12.85 percentage points. On net, aggregate earnings in southern provinces would increase on average by 16.6%, or 114 euros a month.
Nationwide, we estimate that aggregate employment would increase by 5.77 percentage points and aggregate earnings by 7.45%. This amounts to around 600 euros per year for each working-age adult. The north-south gap in income per capita would decrease from 28% to 11%.
Our findings are relevant to other countries too. The Italian and German systems are not unique. Broadly speaking, France, Belgium, Portugal, Finland, Iceland, and Slovenia are similar to the Italian model, while Austria, Denmark, the Netherlands, Norway, and Sweden are closer to the German model (OECD 2017 and 2018). Greece, Portugal, and Spain have recently moved from a bargaining system similar to Italy’s to 'controlled decentralisation', similar to the German system. France has long debated a more decentralised bargaining system. This reform was initially part of the labour market reforms proposed by President Macron in 2017, though it was subsequently dropped due to strong union opposition. Our study suggests that, contrary to common belief, this would have increased GDP, employment and wages, while reducing income inequalities.
Atkinson A B and A Brandolini (2006), “The panel of Country Approach to Explain Income Inequality: An Interdisciplinary Research Agenda”, Chapter 4 in S L Morgan, D B Grusky, and G S Fields (eds), Mobility and Inequality. Frontiers of Research in Sociology and Economics, Princeton University Press: 400-448.
Boeri, T, A Ichino, E Moretti and J Posch (2019), "Wage Equalization and Regional Misallocation: Evidence from Italian and German Provinces", CEPR discussion paper 13545.
Mahler V (2004), “Economic Globalization, Domestic Politics and Income Inequality in the Developed Countries: A Cross-National Study”, Comparative Political Studies 37: 1025-1053.
OECD (2017), Employment Outlook 2017.
OECD (2018), The role of collective bargaining systems for good labour market performance, OECD Employment Outlook 2018.
Rueda D and J Pontusson (2000), "Wage Inequality and Varieties of Capitalism”, World Politics 52: 350-383.
Schnabel, C (1998), The reform of collective bargaining in Germany: corporatist stability vs. firm flexibility, in in R Hoffmann, O Jacobi, B Keller and M Weiss (eds), The German Model of Industrial Relations between Adaptation and Erosion, Hans-BöcklerStiftung: 83–95.
Wallerstein M (1999), “Wage-Setting Institutions and Pay Inequality in Advanced Industrial Societies”, American Journal of Political Science 43: 649-680.