Union pay premium in China: an individual-level analysis

Abstract

Purpose – The purpose of this paper is to estimate the union-nonunion pay gap impact separately for wages and bonuses as well as total compensation to include both wages and bonuses in China. The way in which the impact varies as control variables are added is illustrated as is how the impact varies by the type of firm ownership. The overall pay gap is also decomposed into a component due to differences in their pay determining characteristics as well as a component due to differences in their returns to those characteristics. These separate components are also calculated throughout the pay distribution. Design/methodology/approach – Using the 2010 China Family Panel Studies Survey, a nationally representative survey in China, the methodology involves different estimation procedures as appropriate for the nature of the data and the dependent variables. First the authors estimate a single equation to determine the union-nonunion pay gap. Then the authors estimate the union impact on the various components of compensation (wages and bonuses). Next the authors decompose the relative contribution of each factor in explaining the wage gap. Finally, quantile regressions are used to examine the union impact across various levels of the pay distribution. Findings – The authors find a gross union-nonunion pay gap (wages + bonuses) of 42 percent, dropping to 12 percent after controlling for the effect of other pay determining factors. The union impact on wages is only 8 percent, but bonuses are about twice as high for union workers. The union impact is essentially zero for (state-owned firms) SOEs and for foreign-owned firms but it is large at 16 percent for private firms and even larger at 22 percent for government agencies. Of the overall pay gap of 42 percent, about three-quarters is attributable to differences in their endowments of pay determining characteristics and about one-quarter to differences in the returns for the same endowments of characteristics. Quantile regressions reveal that the pure or adjusted union wage premium exhibits a u-shaped pattern being highest in the bottom and to a lesser extent the top of the pay distribution. Originality/value – There are a dearth of studies examining the union-nonunion pay gap in China. Of the studies that examine this issue, all of them are at the enterprise level with no studies at the individual level. Taking a nationally representative dataset at the individual level, the authors are able to estimate the union-nonunion pay gap in China. The authors identify the portion of the gap that reflects differences in endowments of pay determining characteristics and the portion that reflects different returns to those characteristics, and the relative contribution of the different variables to those components; and how these components change over the pay distribution. The authors also offer explanations for many of these patterns.

Do local union strategies explain the (unexpected) union pay premium in China?

“…For the non-wage components of compensation, positive effects of unions are also found. The union effect on bonuses was estimated as 12.9% in Gunderson et al (2016), while the effect on social insurance coverage was estimated to be between 7.5 and 12.4% in the study of Lu et al (2010) based on private firm-level data and between 12 and 17.1% in the study of Ge (2014). The effect on the coverage of housing subsidies was estimated at around 5% (Lu et al, 2010;Ge, 2014).…”

“…Union effects on the sum of wage and non-wage compensation were estimated as 8.6% (Ge, 2014 based on firm-level data) and 11.7% (Gunderson et al, 2016 based on individual-level data). The union effects on wages only were estimated as 7.8% (Gunderson et al, 2016) and 12.6% (Yao and Zhong, 2013 based on firm-level data across 12 cities). These are within the range of 8.9-12.4% found in the meta-analysis of 152 estimates from 114 US studies, as documented in Jarrell and Stanley (1990).…”

“…The China Family Panel Studies (CFPS) 2010 is a nationally representative longitudinal survey covering a broad range of themes for multi-disciplinary analysis. As outlined in detail in the study of Gunderson et al (2016), it had a response rate of approximately 81%, representing 95% of the total population, a claim verified by Xu and Xie (2013). It involves 25 provinces, with five excluded because they are generally less populated autonomous regions where ethnic minorities concentrate (Xinjiang, Tibet, Inner Mongolia, Qinghai and Ningxia) and two excluded because they are special administration regions (Hong Kong, Macau).…”

“…The effects that unions in China have on various outcomes are fairly similar to those in many Western countries. These effects include higher compensations and benefits (Anwar and Sun, 2015;Ge, 2014;Gunderson et al, 2016;Yao and Zhong, 2013), increasing productivity (Budd et al, 2014;Ge, 2007;Lu et al, 2010) and decreasing firm profitability (Ge, 2007, albeit mixed in Anwar and Sun, 2015). These similar effects are perhaps somewhat surprising since unions in China are generally regarded as "company unions" controlled by employers and the State and designed to placate workers and serve as a "transmission belt" for the wishes of the Communist Party of China (CPC), to use a phrase coined by Lenin in his 1920 speech to the All-Russia Central Council of Trade Unions (Lenin, 1965).…”

Towards a hybrid model? A systematic review of human resource management research on Chinese state-owned enterprises (1993–2017)

“…Third, among 37 articles related to individual outcomes, only seven (19%) touched on tangible outcomes. This means that existing studies focus extensively on intangible or psychological outcomes while marginalizing tangible or material outcomes like job insecurity (Wong et al, 2005) and financial rewards (Gunderson, Lee, & Wang, 2016). The lacuna of tangible/material individual outcomes is not surprising, given the growing dominance of OB theories that regard employees as psychologized objects and instruments, subject to personal traits and cognitive capabilities (Wood et al, 2011).…”