Executive cash compensation and tax aggressiveness of Chinese firms

Abstract

We examine the influence of corporate compensation policies on firms’ tax aggressiveness in an emerging market where executive compensation is primarily in cash form. Based on a hand-collected dataset of 958 firm-year observations of Chinese listed firms for the 2006–2012 period, we find that firms paying higher executive cash compensation are associated with lower tax aggressiveness. This relationship also holds for the excess cash compensation measures which control for executive shareholding, firm profitability, size, growth opportunity, and board independence. We further document that mutual funds ownership pressure firms paying higher compensation to reduce their tax aggressiveness, suggesting adverse selection by mutual funds on firms exhibiting risky tax avoidance activities. High leverage offsets the negative link between cash compensation and tax aggressiveness, indicating a complementary effect between debt and tax avoidance, and, hence, suggesting that creditor monitoring is weak. These results are robust to the system-GMM estimation, which simultaneously account for the endogeneity of executive compensation, tax aggressiveness, ownership and control, leverage, and corporate governance. Our findings on Chinese firms have important policy implications for developing countries around the world with concentrated ownership structure, weak institutional environment, widespread corruption, ineffective rule of law, and ongoing significant social and political transformation.

Appendices

Appendix A: 000022: Shenzhen CHIWAN WHARF HOLDINGS LIMITED 2012 annual reports

Income tax expenses (in RMB):

Item	2012	2011
Current tax expenses	133,843,163	155,728,890
Deferred tax expenses	(8,718,115)	(7,167,257)
Total	125,125,048	148,561,633

Reconciliation of income tax expenses to the accounting profit is as follows (in RMB):

Item	2012	2011
Accounting profit	740,894,558	816,337,301
Income tax expenses calculated at 25% (the prior year: 24%)	185,223,640	195,920,952
Effect of expenses that are not deductible for tax purposes	3,715,114	2,746,526
Effect of tax-free income	(20,879,728)	(28,374,764)
Effect of unrecognized deductible losses and deductible temporary differences for tax purposes	957,342	1,003,459
Changes in opening balances of deferred tax assets/liabilities due to the adjustment in tax rate		2,290,517
Effect of different tax rates of subsidiaries operating in other jurisdictions	(302,040)	(244,303)
Effect of tax preference policy	(50,664,660)	(30,256,021)
Withholding tax	7,075,380	5,475,267
Income tax expense	125,125,048	148,561,633

1. The tax-effect BTDs can be calculated in two ways. Firstly, employing prima racie income tax expenses minus current tax expenses, in this case, it is calculated as follows: for the 2012, 185,223,640 − 133,843,163 + 8,718,115 = 60,098,592. Secondly, it is the sum of the temporary and permanent differences, in this case, it is the sum of the row 4 to row 10 in reconciliation of income tax expenses, that is − (3,715,114 − 20,879,728 + 957,342 − 302,040 − 50,664,660 + 7,075,380) = 60,098,592

Appendix B1: the book-tax difference (BTD) models

Our model—Eq. (1)		Tang and Firth (2011) and Tang (2015) model—Eq. (2)
Dependent variables	BTDit	Dependent variables	BTDit
OPEXPit	0.000	ΔINVit	0.002
	(0.32)		(1.16)
OPBITit	− 0.091***	ΔREVit	0.000
	(− 4.90)		(0.75)
PBTit	0.171***	NOLit	1.011***
	(8.71)		(14.07)
PBTit−1	− 0.017***	TLUit	0.854***
	(− 3.10)		(8.18)
PBTit−2	− 0.020***	TAX_DIFFit	0.004***
	(2.75)		(2.95)
INVINCit	0.064***
	(3.38)
LOG(ASSETS)it	0.000***
	(4.08)
INTINCit	− 0.165
	(− 1.64)
FSALEit	0.000
	(1.38)
Intercept	0.002*
	(1.74)
Year dummies	Controlled	Year dummies	Controlled
Industry dummies	Controlled	Industry dummies	Controlled
Observations	958	Observations	962
R-square	0.449	R-square	0.401

1. The dependent variable is total BTD calculated using manually collected BTD categories. These BTD models (1) and (2) control for the drivers of mechanical differences in BTDs. In Eq. (1), the independent variables on the right hand are the proxies for BTD drivers listed in Table 1 including current period investment income ${\text{INVINC}}_{\text{it}}$, interest income ${\text{INTINC}}_{\text{it}}$, operating expenses ${\text{OPEXP}}_{\text{it}}$, the percentage of overseas sales $FSALE_{it}$, the log of total assets ${\text{LOG}}\left( {\text{ASSETS}} \right)_{\text{it}}$, operating profit before interest and tax ${\text{OPBIT}}_{\text{it}}$, net profit before tax ${\text{PBT}}_{\text{it}}$, and the net profit before tax in previous periods ${\text{PBT}}_{{{\text{it}} - 1}}$ and ${\text{PBT}}_{{{\text{it}} - 2}}$. In Eq. (2), the independent variables are change in fixed assets investment $\Delta INV_{it}$, change in revenue $\Delta REV_{it}$, the value of operating losses $NOL_{it}$, the value of tax loss utilized $TLU_{it}$, the difference between the consolidated company’s applicable tax rate and the average tax rate in the consolidated group $TAX\_DIFF_{it}$. See Sect. 3.1 for detailed discussions

Appendix B2: the executive compensation model predictions

Model	(1)	(2)	(3)
Dep. var.	LOG(EXEPAY)	LOG(DIRPAY)	LOG(EDSPAY)
EXE.Shareholding	0.270**
	(2.41)
DIR.Shareholding		0.629***
		(3.60)
EDS.Shareholding			0.536***
			(3.49)
LOG(Tobin’Q)	− 0.221***	− 0.281***	− 0.336***
	(− 8.24)	(− 7.74)	(− 11.99)
ROE	0.445***	0.498***	0.368***
	(7.30)	(6.74)	(6.38)
LOGMC	0.286***	0.337***	0.362***
	(12.98)	(11.53)	(15.87)
BOARDIND	0.228	− 0.559***	− 0.123
	(1.46)	(− 2.71)	(− 0.77)
Constant	11.062***	10.819***	11.567***
	(66.81)	(50.22)	(66.48)
Firm fixed effects	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes
Observations	11,420	11,403	11,585
R-squared	0.466	0.361	0.505
# Firms	2392	2386	2393

1. We use all A-share listed non-financial sector firms during the years 2006–2012 to estimate the predicted cash compensation in Eq. (3) of Sect. 3.2. All 3 regressions control for fixed firm and fixed year effects. t-statistics are based on robust standard errors. ***p < 0.01, **p < 0.05, *p < 0.1

Appendix B3: the influence of firm size on the compensation and tax aggressiveness relationship

Model	(1)	(2)	(3)	(4)	(5)	(6)
Dep. var.	TAXAGG	TF_AGG	TAXAGG	TF_AGG	TAXAGG	TF_AGG
LOG(EXEPAY)	− 0.594***	− 0.527***
	(− 6.17)	(− 5.02)
LOG(DIRPAY)			− 0.632***	− 0.575***
			(− 6.31)	(− 5.20)
LOG(EDSPAY)					− 0.710***	− 0.637***
					(− 6.38)	(− 5.23)
LOG(EXEPAY)*LOGMC	0.051***	0.044***
	(4.65)	(3.88)
LOG(DIRPAY)*LOGMC			0.052***	0.045***
			(4.94)	(4.09)
LOG(EDSPAY)*LOGMC					0.059***	0.051***
					(5.15)	(4.26)
LOGMC	− 0.050	− 0.040	− 0.035	− 0.014	− 0.042	− 0.030
	(− 0.72)	(− 0.54)	(− 0.56)	(− 0.21)	(− 0.70)	(− 0.47)
Other firm level controls	Yes	Yes	Yes	Yes	Yes	Yes
Industry effects	Yes	Yes	Yes	Yes	Yes	Yes
Year effects	Yes	Yes	Yes	Yes	Yes	Yes
Observations	865	867	833	835	866	868
R-squared	0.281	0.227	0.280	0.228	0.282	0.228

1. All models are OLS regressions controlling for industry and year fixed-effects. The numbers in parentheses are robust t-statistics for regression coefficients with firm-level clustered standard errors. To reduce the endogeneity problem, all the continuous independent variables are lagged by 1 year. Other control variables follow models in Tables 4 and 5. Other control variables follow models in Tables 4 and 5. See “Appendix C” for variable definitions. ***p < 0.01, **p < 0.05, *p < 0.1

Appendix C: variable definitions

TAXAGG is the measure of the tax aggressiveness, which is the prediction error from our BTD model.

TF_AGG is the measure of tax aggressiveness for Chinese firms following the BTD model specification in Tang and Firth (2011, 2012).

EXEPAY is the top three executives’ cash compensation, which is the total pay of the top three officers, defined as the sum of basic salary and bonus excluding allowance.

DIRPAY is the top three directors’ cash compensation including basic salary and bonus excluding allowance.

EDSPAY is the average per person cash compensation paid to board of directors, supervisors, and executives.

LOG(EXEPAY) is the log of the top three executives’ cash compensation.

LOG(DIRPAY) is the log of the top three directors’ cash compensation.

LOG(EDSPAY) is the log of the average per person cash compensation to directors, supervisors, and executives.

EXCESS LOG(EXEPAY) is the excess cash compensation for top 3 executives calculated as the prediction error of an executive compensation model.

EXCESS LOG(DIRPAY) is the excessive cash compensation for top 3 directors calculated as the prediction error of an director compensation model.

EXCESS LOG(EDSPAY) is the average per person excessive cash compensation for directors, supervisors, and executives calculated as the prediction error of their corresponding per person compensation model.

BOARDIND is the percentage of board members that are independent.

BOARDSIZE is the size of the board as the number of directors.

BOARDMEET is the total number of board meetings in a year.

CEOD is a dummy which equals to 1 if the chair of the board and the CEO are the same person and 0 if they are two persons.

BIG4AUDIT is a dummy which equals to 1 if the firm’s auditor is one of the “Big-4” accounting firms.

AUDITOP is a dummy which equals to 1 if the auditor opinion is standard or 0 if it is non-standard.

LEVERAGE is the market value financial leverage ratio which equals to the book value of debt divided by the total of market capitalization and book value of debt.

HLEV is a dummy which equals to 1 if the leverage ratio is above its median value, or 0 if otherwise.

DACC is the value of discretionary accruals measured as the prediction error when regressing total accruals against change in sales, fixed assets, and industry and year fixed effects.

ROE is the return on equity.

LOGMC is the log of firm market capitalization.

BOOK/PRICE the book-to-price ratio.

LOSS is a dummy which is equal to 1 if the firm’s net income before extraordinary items is negative, or 0 if otherwise.

STASH is the total percentage of shares that are classified as state-shares and state-legal person shares.

GOVCON is a dummy which equals to 1 if the firm controlling shareholder is government or government agency and 0 if it is a private investor.

FUNDSH is the percentage of shares held by mutual funds.

HFUNDSH is a dummy which equals to 1 if the percentage of fund shareholding (FUNDSH) is above its median value, or 0 if otherwise.