Abstract

     This paper analyses US payroll share and components across states for the period 1977–۲۰۱۷٫ Findings include that spatial clustering in payroll shares decreased until the year 2000. States are clustered in low and high productivity groups. We relate this phenomenon to dualism. High labour productivity states featured high payroll shares early on, but now feature low payroll shares. We label this phenomenon decoupling of labour productivity from real wages. A Divisia Decomposition documents that Rust Belt states dominated the decline in US payroll share in the 1980s, whereas more recently large (coastal) states dominated. While we do not explain (potentially different) mechanisms, decoupling is apparent throughout.

Introduction

     The labour share of income is equivalent to the ratio of real wage and labour productivity. Its precipitous decline has led to a growing literature investigating underlying causes, not least since it illuminates interactions between labour incomes and technology. This paper provides novel evidence on the sources of the decline with a focus on spatial dimensions. We decompose the aggregate US payroll share – which excludes income from self-employment from compensation – by states for the period 1977–۲۰۱۷ and analyse its components from several angles. These state payroll shares are driven by the sectoral composition of each state’s economy, and therefore related to geographical factors and agglomeration effects. The former include endowments – a bay facilitating a port and transportation hub; the Bakken Formation facilitating an extractive boom – but the focus of the latter has been on the returns to density. These ideas build on Smith (1776), Marshall (1890) and Krugman (1991, 2011) and have been categorized as related to (1) reduced transportation costs, and improved flows of (2) labour and (3) ideas. More recently, such agglomeration economies have been critically examined in light of the ubiquity of information technology (Ellison et al., 2010; Giuliano et al., 2019; Glaeser & Gottlieb, 2009). The evidence suggests that agglomeration still matters in manufacturing, albeit less today than earlier; and that certain service activities crucially benefit from agglomeration due to (2) and (3). Throughout our analysis, we distinguish between (1) progressive and stagnant sectors and (2) states with (predominantly) progressive or stagnant sectors. In his seminal paper, Baumol (1967, p. 415) defines ‘technologically progressive activities in which innovations, capital accumulation, and economies of large scale all make for a cumulative rise in output per man hour’ (emphasis added).

Conclusion

     This paper investigates the decline of the US labour share by states. We conduct three exercises. First, a spatial Gini index illuminates regional patterns in payroll share changes. A key result is that spatial dependence of payroll shares decreased throughout the 1980s and 1990s. Further, spatial dependence is highest for manufacturing employment shares. The decline in spatial dependence occurs simultaneously with deindustrialization in (spatially clustered) Rust Belt states. Future research could investigate whether these results are robust to further disaggregation.

     Second, uni- and bivariate mixture models indicate that the distribution of labour productivity – but not that of real wages – across states has become increasingly dislocated. Our results support the hypothesis that labour productivity in US states differs in a statistically significant manner. Particularly bivariate mixture model results show that during the last two periods, states with higher labour productivity feature relatively low payroll shares, and vice versa. In the earlier periods, states with high labour productivity featured high payroll shares. This emerging dualism does not feature strong spatial clustering at the level of the state. Nevertheless, results appear consistent with recent literature on agglomeration economies (Ellison et al., 2010; Giuliano et al., 2019; Glaeser & Gottlieb, 2009).