What an Analysis of Variation in the ‘Partner Pay-Gap' Tells Us about the Persistence of Income and Gendered Inequalities within Couples: A Comparative Analysis Using Quantile Fixed-Effects Regression.

Income inequality has risen substantially since the mid-1980s in Europe and America (Fredriksen, 2012) with many countries accused of doing too little to either prevent lower earners’ exposure to poverty and too little to abate the very high earnings of those at the top of the widening earnings distribution (Atkinson, 2013). Inequality in households’ access to work, employment inequality, has also risen with a polarizing trend in work-rich and work-poor households found in many countries (Gregg, Hanson and Wadsworth, 2002). Here marital homogamy is seen to reinforce earning and employment inequalities: couples with weak human capital are found at the bottom of the distribution and are unable to access (stable and well paid) jobs which perpetuates poverty risk and keeps entire households’ ‘work-poor’, whilst those with considerable human (and other) capital resources are concentrated at the top of the distribution with both partners in stable employment with high earnings, which serves to widen employment and earnings inequalities. Several authors report increases in marital homogamy (Blossfeld and Buchholz, 2009; Schwartz and Mare 2005), with the relationship between educational homogamy and increased income inequality confirmed using US data (Park et al. 2009). The literature on gender inequalities has also turned its attention to within household dynamics (Dieckhoff et al. 2015), with the earnings’ gap between men and women, the ‘partner pay-gap’, examined as a mechanism behind persisting gender inequalities. While theorists of marital homogamy often assume that similarity between couples assumes equality in their relationships, research on the partner pay-gap suggests otherwise: there is a remarkable traditionalism in couples’ economic outcomes, with men in the majority of instances out-earning their female partners by quite high levels (Bertrand et al. 2015; Dieckhoff et al. 2015). Moreover, these gaps in earned income have been found to have negative effects on women’s economics outcomes (Dieckhoff et al. 2015) which in turn expose women and children to poverty risk in instances of relationship dissolution. While there is a considerable literature on the sources and consequences of income and employment inequalities across households, the literature which examines dynamics within households, a dynamic which is presented as causal of income and employment inequalities, is scant. Our paper adds to this literature in its examination of variation in the partner gap in pay and employment across the income distribution. Our paper reveals; the extent to which partners in low income households are more or less equally contributing to household income compared to wealthy households. We also examine variation in the work-life reconciliation strategies of low and high income households in three different institutional and cultural contexts.

Data and Method. Our paper looks at variation in the partner pay-gap (that is the difference between husbands’ and wives’ labour income) by quantiles and over time using six waves of the German Socio-Economic Panel (SOEP) and the United Kingdom Household Longitudinal Study (UKHLS) which spans the period 2008-2014. The datasets are regarded to be first-class, nationally representative longitudinal panel surveys (Wagner et al., 2007) and contain many cross-nationally comparable variables of interest. We examine women and men in partnerships (either married or in a cohabiting union) aged 20-59 years. We drop those with less than two years of full-information on key covariates, given our application of quantile fixed-effects models (Powell, 2014; Lamarche, 2010) which require at least two data points for correct estimations. The paper examines three different institutional and cultural contexts: The United Kingdom, Eastern and Western Germany. Our three-way comparison allows for an assessment of variance in the structuring effect of national institutions and national cultures on gendered employment strategies. Although Germany and the United Kingdom are similar in their residual institutional support for working-motherhood in terms of childcare provision, they differ markedly in parental leave policies and income support for parents with income support for parents shown to affect within-couple earnings inequalities (Dotti Sani, 2015). Moreover, the two countries have very different labour market institutions deemed relevant in shaping female labour market behaviour. The German market is typically regarded as rigid or co-ordinated whereas the UK is often classified as flexible and/or liberal (Hall and Soskice, 2001). Our ability to measure the impact of culture on gendered behaviour is further enhanced by our separate analysis of Eastern and Western Germany. Although Eastern and Western Germany share the same institutional structure, the partitioned country has been reunified since 1990, the historical legacy of state socialism in the German Democratic Republic (GDR) continues to structure cultural norms pertaining to working motherhood in a manner quite different to West-Germany (Rosenfeld et al., 2004).

Research Strategy and Preliminary Findings. We find an increasing tendency for men to out-earn their partners in each country case over-time, a finding found in earlier research using data from a different time period (Dieckhoff et al., 2015). Looking at the distribution of relative household contributions, we find a big drop-off in the proportion of couples where she out-earns her partner in each country case, a finding which replicates those of researchers using American data (Bertrand et al. 2015). Our finding of within-couple earnings inequality varies significantly by quantiles of household income. Here we find that the partner pay-gap identified at the mean is driven primarily by those in the upper quantiles. At the top of the earnings distribution we find the partner pay-gap for the highest quantile to be double that of the second highest quantile. While among the lowest earners we find a very small gender pay-gap, however this is driven by the low pay of lower skilled workers as well as lower earning men’s disproportionate exposure to unemployment. So we find gender equality for lower quantiles to be the result of disadvantageous market conditions rather than due to gender egalitarian ideals. There are also interesting differences by country case. While the UK and Western Germany are similar in the structure of the partner pay-gap by household quantiles, Eastern Germany is very different; with a comparatively smaller partner pay-gap and smaller, and in some cases declining, differences by income quintile.