Multiple large shareholders and cost stickiness: evidence from China

AbstractUsing a sample of Chinese listed firms from 2001 to 2017, this study investigates the impact of multiple large shareholders (MLS) on cost stickiness from the agency costs perspective. We find a positive association between MLS and cost stickiness after controlling for various determinants of cost stickiness. The results of additional analyses suggest that coordination costs among large shareholders make it challenging to monitor managers, and stronger protection of minority shareholders helps to eliminate the effect of MLS on cost stickiness. This paper extends our understanding of the ‘dark side’ of MLS and complements existing research investigating the determinants of cost stickiness from the ownership structure perspective.Keywords: multiple large shareholdersownership structurecost stickinessJEL classification: M41G32 Disclosure statementNo potential conflict of interest was reported by the author(s).Notes1 Since selling, general, and administrative (SG&A) costs account for a main portion, most research on cost management focuses on SG&A costs. In this study, ‘cost stickiness’ and ‘SG&A cost stickiness’ are used interchangeably.2 Our study focuses on the agency-driven incentives to build empires or enjoy a quiet life inducing cost stickiness. However, agency-driven incentives to meet earnings target reduce cost stickiness (Kama and Weiss, Citation2013). Therefore, we have controlled managerial incentives to meet earnings target throughout all regressions. We thank an anonymous reviewer for the insightful comment about this issue.3 Following Jiang et al. (Citation2018), we alternatively define the large shareholder as one who owns at least 5% ownership and our results hold for using this alternative measure.4 In China, shareholders with more than 10% ownership have the right to require an interim shareholders’ meeting or a board meeting besides the regular meetings and they could put forward or vote for a proposal in these meetings.5 Obvious errors include negative total assets, negative sales, negative SG&A costs, and negative number of employees.6 To mitigate the potential influence of China’s new accounting standards adopted in 2007, we have conducted a robustness test using an alternative sample from 2007 to 2017. Our results remain robust. We thank an anonymous reviewer for the insightful suggestion about this issue.7 Following Kama and Weiss (Citation2013), the interpretations of the coefficient β4 is presented below. In model (1), the slope for the sales increase in MLS firms is β1 + α1 + α2AI + α3 EI + α4SuccessiveDecrease + α5GDP + α6FCF + α7Target. In addition, the slope for sales decrease in MLS firms is β1 + β3 + β4 + δ1AI + δ2EI + δ3SuccessiveDecrease + δ4GDP + δ5FCF + δ6Target + α1 + α2AI + α3 EI + α4SuccessiveDecrease + α5GDP + α6FCF + α7Target. Thus, cost stickiness in MLS firms is the difference between the two slopes, which equals β3 + β4 + δ1AI + δ2EI + δ3SuccessiveDecrease + δ4GDP + δ5FCF + δ6Target. Similarly, cost stickiness in non-MLS firms is the difference between the two slopes, which equals β3 + δ1AI + δ2EI + δ3SuccessiveDecrease + δ4GDP + δ5FCF + δ6Target. Taken together, the cost stickiness difference between MLS firms and non-MLS firms is β4, which measures the effect of MLS on cost stickiness.8 Following Hartlieb et al. (Citation2020), we alternatively define the variable SuccessiveDecrease equals one if sales decrease in two consecutive years, and zero otherwise and our results hold for using this alternative measure.9 Specifically, we divide large shareholders in China into four types, i.e., financial institution shareholders, state shareholders, individual shareholders, and non-state companies (Cronqvist and Fahlenbrach, Citation2009; Schwartz-Ziv and Volkova, Citation2021). The reason is that Cronqvist and Fahlenbrach (Citation2009) and Schwartz-Ziv and Volkova (Citation2021) divide large shareholders into four categories, i.e., financial institution, hedge fund, individuals and non-state corporations. However, there are few hedge funds holding more than 10% shareholdings in the Chinese firms. The possible reason is that hedge fund in China has a short history (Li et al., Citation2011). Moreover, since state ownership is significantly different from other types in China, we add state shareholder as another type of large shareholders.10 There might be a potential concern that industry-year average might not perfectly satisfy the excluding restriction since industry average might be correlated with industry characteristics that drive both ownership structure and cost pattern in a firm. Thus, the results of the Heckman two-step method in this paper should be interpreted with caution.11 Following Banker et al. (Citation2011), we use the predicted value of (SG&A)i,t estimated from the following model (F1) as the instrument variable. (SG&A)i,t = a+b (SG&A)2 +ui,t (F1). We then include the instrument variable in the model (F2) to obtain the measures of value creation of SG&A costs. (OI/TA)i,t = β0+β1 (1/TA)i,t-1+∑k = nk = 0β2,k(SG&A/TA)i,t-k+ei,t (F2), where OI is operating income before depreciation; SG&A is SG&A expenses; TA is total assets. We estimate the coefficients on each lag of (SG&A/TA) for each industry and control for year-fixed effects. We require the coefficients to be positive and significant and we select the best model by Akaike Information Criterion (AIC) and the Schwartz Bayesian Criterion (SBC). Finally, these coefficients are discounted at a discount rate of 10% and the sum of these discounted values is the value creation of SG&A costs (FV).