Minimum wages and non-listed firm’s earnings management: Evidence from a geographic discontinuity design

Highlights

• We examine how and to what extent minimum wage shapes private firms’ earnings management.

• We introduce GIS technique to locate firms in areas that straddle the continuities of county borders to exploit discontinuities of minimum wage at county borders.

• We find that firms significantly manage earnings information upward to respond to increases in minimum wages, particularly for firms with financial constraints.

• Our findings shed light on the effects of labor policy on earnings management at the firm-level.

Abstract

We examine how and to what extent minimum wage shapes non-listed firms’ earnings information production. To identify the causality, we introduce an improved geographic information system technique to locate firms in areas that straddle the continuities of county borders and then exploit discontinuities of minimum wage at county borders. We find that firms significantly manage earnings information upward as response to increases in minimum wages, particularly for firms with financial constraints. Our findings shed light on the effects of labor policy on the information production of firms and provide policy implications to regulators concerned with the allocation efficiency of capital markets.

Introduction

This study evaluates the causal effect of minimum wages on the information production of non-listed firms. Although minimum wage is a crucial policy affecting labor costs, related studies mostly focus on whether such a provision affects employment, household income, inequality, and poverty. Surprisingly, little is known about how and to what extent the minimum wage policy shapes the information production of firms. Given that wage-related costs typically represent more than 70% of economy-wide value added, an evaluation of the causal effects of minimum wages on firm-level decisions should be of particular interest to economists and regulators who are concerned with labor policies and information efficiency.¹

In this study, we focus on one of the most tangible signs of distorted information (Massa et al., 2015), that is, earnings management. Earnings management can be used to “mislead some stakeholders about the underlying economic performance of the company or to influence contractual outcomes that depend on reported accounting numbers.”²

We expect that minimum wage hikes could affect firm’s costs and capital allocation and therefore, their earnings and profits. This assumption implies that a firm may exhibit incentives to manage its earnings due to outside pressure (e.g., raising external capital or avoiding debt covenant violations). The literature well supports the idea that a minimum wage policy substantially increases the labor costs of firms (Grossman, 1983; Gustafson and Kotter, 2021), and the minimum wage policy in China has been considered a major force driving increases in wages and putting labor cost pressure on firms (Geng et al., in press).

However, the impact direction of minimum wage on firm’s earnings management is ambiguous. On the one hand, minimum wage hikes may provide the incentive for a firm to manage earnings upwardly because the increasing labor cost could crowd out capital investment. Anecdotal evidence from media reports argues that increases in minimum wages squeeze small firms.³ Empirical evidence (Dechow et al., 2011, Linck et al., 2013, Dou et al., 2016) shows that the earnings transfer is useful information to banks or other market participants, and firms which want to borrow from external channels upwardly manage earnings to signal a positive outlook. Linck et al. (2013) find that financially constrained firms with valuable projects can use discretionary accruals to credibly signal positive prospects and in turn help the firm obtain bank loans. Thus, to raise external capital to offset the rising costs, firms exhibit incentives to manage earnings upward and deliver positive information after an increase in minimum wage.

On the other hand, minimum wage hikes may lead a firm to manage earnings downward. Labor investment is more flexible and reversible than capital investment (Hall, 2016, Beladi et al., 2020). Therefore, firms affected by wage hikes might cut labor-related expenditures or do not proportionally increase all employees’ wages in response to increasing labor costs. Accordingly, the firm needs a defense to cost-cutting, and managers have incentives to strategically withhold good news and promote bad news to preserve their bargaining power against labor (Bova, 2013, Chung et al., 2016).

Given the lack of a unique prediction about the effect of minimum wage on earnings management, the issue is essentially an empirical one. While previous literature treats minimum wage hikes as plausible exogenous shocks (Aaronson et al., 2012, Aaronson, 2001), we still attempt to construct identification design to address the potential confounding factors in minimum wage endogenously with regional and market conditions.⁴

To overcome this difficulty, we introduce an improved geographic information system (GIS) technique and use policy discontinuities at county borders to identify the variation in minimum wages within each of these county pairs located on opposite sides of a county border (Dube et al., 2010). Given that unobserved heterogeneity in economic activity is likely to be reduced when we consider a contiguous county-pair or a metropolitan area, we come closer to “quasi-experimental” conditions than extant studies using counties’ variation. We thus consider that firms located on opposite sides of county borders may feature variation in minimum wage and constitute counterfactual pairs. Firms in approximate geographical areas demonstrate a sufficiently high level of similarity in labor supply, geographic resource endowment (e.g., geographic and geomorphic conditions), public infrastructure (e.g., airport and railway), market condition, local culture, and other unobservable factors. Therefore, we can observe and examine variations within each of the contiguous county pairs with a minimum wage gradient. We enhance the cross-sectional variation by relying on the large geographical and inter-temporal variations of minimum wage policies in China, where the minimum wage varies across more than 2,800 counties. During our sample period (2004–2013), China implemented more than 19,406 local minimum wage changes, more than 53% of which were greater than 10%.

China, as the study setting, provides an ideal laboratory to examine the effects of minimum wage on a firm’s earnings management. First, China promulgated the minimum wage regulations on January 20, 2004, which expanded the scope of application of the minimum wage in more than 2850 counties in 31 provinces or autonomous regions. Moreover, the governments require adjustment in minimum wage frequency. Given the exogeneity, large intertemporal and geographic variations are present in China’s minimum wage policy, providing an ideal quasi-experimental scenario to address the endogenous issue.

Second, our sample is based on a large dataset of the China Industrial Enterprises Database (CIED), which is a census dataset that comprises the entire universe of Chinese manufacturing firms. Data from CIED allow sufficient observations and geographical information to conduct our identification strategy. Private firms are predominant in nearly all economies, and more than 99% of limited liability companies are not publicly tradable.⁵ Bank financing is usually a major source of finance in private firms (Zou and Adams, 2008), which accordingly induces the earnings management of private firms. By using the unique dataset, we overcome the obstacle of obtaining a large sample of financial data at the firm level.

To conduct the empirical study, following Dube et al., 2010, Geng et al., in press, we introduce GIS techniques to pinpoint the precise location of each firm according to their longitude and latitude. The GIS technique allows us to exploit the discontinuities of minimum wage laws at the same county borders. Firms in approximate geographical areas demonstrate a sufficiently high level of similarity in labor supply, geographic resource endowment (e.g., geographic and geomorphic conditions), public infrastructure (e.g., airport and railway), market condition, local culture, and other unobservable factors.

We construct contiguous county pairs that contain firms located within a short distance (i.e., 25, 50, and 75 km) of the shared border. Our baseline results accounts for the 14.84% increase in earnings management relative to its mean. To check the robustness of our results, we introduce three alternative measures of earnings management, and highly consistent results are found. Our finding is consistent with the hypotheses that minimum wages causally increase the incentives of firms to distort the earnings information for obtaining advantages in external financing. Then, we conduct a number of cross-sectional tests on the bases of the level of average wage, labor capital and IT technology. We find that the incentive of earnings management is stronger in low levels of average wage, labor capital and IT technology firms when they react to minimum wage hikes.

Our paper makes two contributions to the literature. First, we present the causal effects of labor cost on the distortion of firm information, deepening our understanding of the determinants of a firm’s information production and resource misallocation at the firm level. Goldstein and Yang (2019) point out that, following the recent global financial crisis, one main conclusion in the literature is that firms should be required to provide better disclosure and more precise information. Given that non-listed firms constitute the majority of the modern economy and the absence of stock price in aggregating dispersed information, studying information production behaviors of non-listed firms is relevant. Second, we contribute to the emerging literature of labor and finance by evaluating the economic consequences of a prevalent labor policy worldwide (i.e., minimum wage policy). Prior studies mostly investigate whether such provisions affect employment, income, inequality, and poverty. Little is known about how and to what extent minimum wages shapes firm’s earnings information production.

In addition, we also believe that this topic offers clear policy implications to regulators because labor cost, the most important production input, typically represents roughly two-thirds of economy-wide value added (Gan et al., 2016) and information production at the firm level eventually affects resource allocation of the entire economy.