The impact of tax holidays on earnings management: An empirical study of corporate reporting behavior in a developing economy framework Kenny Z. Lin Department of Accountancy Lingnan University, Hong Kong Abstract: Many developing economies use tax holidays to attract foreign direct investment. This study investigates whether foreign investment enterprises (FIEs) in China alter their corporate reporting behavior in response to a known schedule of increasing tax rates. Foreign investors in different tax-holiday positions are subject to different tax rates in China: normally 0-15% in the tax-holiday period while 30% in the post-holiday period. If managers attempt to maximize firm value by minimizing tax costs, this spread in tax rates between periods surrounding the rate change would provide a substantial incentive to accelerate revenues and defer expenses. Evidence of tax-induced earnings management is examined by focusing on discretionary portion of current accruals most directly related to taxable income. Based on 750 firm-year observations, the results indicate discretionary accruals in the year immediately preceding the tax rate increase are higher, or more income increasing, than those in other years. The evidence, which indicates that firms can shift income between periods to take advantage of the anticipated increases in tax rates, has important implications for tax policymakers. Key Words: Discretionary accruals, Earnings management, Foreign investment enterprises (FIEs), Income shifting, Tax holidays.
32
Embed
An Empirical Study of Tax-Motivated Earnings Management By
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The impact of tax holidays on earnings management: An empirical study of corporate
reporting behavior in a developing economy framework
Kenny Z. Lin
Department of Accountancy
Lingnan University, Hong Kong
Abstract:
Many developing economies use tax holidays to attract foreign direct investment. This study investigates whether foreign investment enterprises (FIEs) in China alter their corporate reporting behavior in response to a known schedule of increasing tax rates. Foreign investors in different tax-holiday positions are subject to different tax rates in China: normally 0-15% in the tax-holiday period while 30% in the post-holiday period. If managers attempt to maximize firm value by minimizing tax costs, this spread in tax rates between periods surrounding the rate change would provide a substantial incentive to accelerate revenues and defer expenses. Evidence of tax-induced earnings management is examined by focusing on discretionary portion of current accruals most directly related to taxable income. Based on 750 firm-year observations, the results indicate discretionary accruals in the year immediately preceding the tax rate increase are higher, or more income increasing, than those in other years. The evidence, which indicates that firms can shift income between periods to take advantage of the anticipated increases in tax rates, has important implications for tax policymakers.
where DAit is the discretionary accruals estimate for observation it, as described earlier in this
section, and YRi2, YR i3, and YRi5 are dummy-coded as 1 if firm i is respectively in the second,
third, and fifth year of the five-year tax-holiday period. Similarly, YRi6 is coded as 1 if firm i is in
the year immediately after the tax-holiday period, and zero otherwise. Since YR i2 and YRi5 are the
years that appear to have an incentive to manage earnings to take the greatest advantage of lower
tax rates, these two variables are expected to be significantly and positively signed (i.e., firms tend
8
to choose income-increasing accruals in these two years). However, YR i3 and YRi6 are expected to
have negative coefficients, because of reversals of the managed accruals in prior years.
Five additional variables are included in the equation to control for the effect of corporate
characteristics on discretionary accounting practices. As industry affiliation, form of investment and
ownership control affect tax noncompliance (Chan and Mo, 2000), three dummy variables, namely,
INDUSTRY (manufacturing vs. others), JV (joint ventures vs. wholly foreign-owned enterprises),
and OWNERSHIP (manager-controlled vs. owner-controlled firms) are included in the model. In
addition, a continuous SIZE variable (which takes the logarithm of the firm’s year-end total assets)
is used to control for the effect of firm size on accrual management (Watts and Zimmerman, 1978).
Finally, as some firms may be less sensitive to the amount of tax paid overseas if the country in
which their parent companies reside operates a tax credit rather than an exemption system, a
dummy variable EXEMPTION is used to control for this confounding effect.
4. Empirical results
4.1 Univariate results
Table 1 presents overall and by-year mean current accruals for equation (1). Mean accruals
are 3.2% of total assets for all firms over the six-year period. On average, the sample firms report
mean accruals of 6.6% of assets in year 2, 5.5% higher than those reported in year 3. Similarly,
mean accruals as a percentage of assets are significantly higher for year 5 than for year 6 (6.9% vs.
1.8%).
9
[Table 1 here]
Table 2 presents the univariate analysis of discretionary accruals estimated over a six-year
period. Mean discretionary accruals, which were 0.0011 two years before the tax rate change (i.e.,
year 1 in Figure 1), increased to 0.0088 in the year before the tax rate change (year 2), and then
decreased to 0.0016 in the year after the rate change (year 3). There is a similar trend from years 4
through 6. Comparisons of mean accruals indicate that firms report discretionary accruals in years 2
and 5 that are significantly higher than those reported in years 3 and 6, respectively to the extent of
0.72% and 0.75% of assets. The pattern shows a trend toward income increasing accruals in years
immediately prior to tax rate increases. Further, while central tendencies of mean accruals for years
2 and 5 are significantly positive, the same measures are not significantly different from zero for
years 1, 3, 4, and 6. Taken at face value, these results suggest that the magnitude of accruals is
related to the level of tax rate in a way that is consistent with the tax-motivated income shifting
behavior.
[Table 2 here]
4.2 Multivariate results
Table 3 reports the results of regressing current accruals on the change in sales and dummy
year and firm variables (Equation (2)). The model assumes the residual is composed of a time
effect, which will be the same for all firms in a particular year, and a firm effect, which will be the
same for each firm over all years (Cahan, 1992). The F-statistic for the model is significant at the
1% level and the adjusted R2 is 8.2%. The variable, SALESit/ASSETSit-1, is significant at the 1%
10
level and the sign is as expected. The YR and FIRM variables as a whole are also significantly
related to current accruals. These results suggest that the variations in accruals can be largely
explained by the change in sales, and time and firm effects. Therefore, the portion of accruals
‘unexplained’ can be assumed to be discretionary accruals.
[Table 3 here]
Table 4 provides the results for equation (3), which formally tests the hypothesis of whether
discretionary accruals are higher in periods prior to the tax rate increase. The model is significant at
the 1% level, which indicates that the model is well specified. All of the correlations among the
independent variables are below 0.653, suggesting that multicollinearity is unlikely to affect the
results. The four YR-dummies measure the incremental ability to explain the remaining cross-
sectional, intertemporal variation in residuals from equation (2). As expected, YRi2 and YRi5 are
significantly and positively signed. A Wald coefficient test (EViews 3 User’s Guide, 1998) indicates
that the coefficients for YRi2 and YRi5 are respectively significantly higher than those for YRi3 and
YRi6. Moreover, the coefficients on YRi2 and YRi5 indicate firms in the year immediately preceding
the tax rate increase respectively report discretionary accruals that are higher than the average of
discretionary accruals reported two years prior to the tax rate change. While Chan and Mo (2000)
find that firms are most compliant in the tax-holiday period, the results of the current study
demonstrate that firms in the tax-holiday period have greater than normal incentives to minimize tax
burden over a longer horizon. The coefficient on SIZEit is (weakly) significantly negative at the 5%
level, suggesting that larger firms may be more sensitive to political costs and thus more likely to
11
use accruals that decrease financial statement income (Watts and Zimmerman, 1978).
[Table 4 here]
5. Conclusions
Tax holidays are by far the most popular type of tax inducement employed by developing
economies. This study investigates whether foreign investors in China’s developing economy
manage accounting earnings to take advantage of lower tax rates available during the tax holiday
period. Evidence of tax-induced earnings management is examined by focusing on discretionary
current accruals that are expected to have a significant effect on taxable income. The study develops
the hypothesis that predicts income-increasing accruals in the year prior to the tax rate increase. The
evidence, which indicates that firms adjust their discretionary accruals in anticipation of the change
in tax rates, is consistent with income shifting behavior. Evidence from this study may be of interest
to tax policymakers in enforcing tax rules designed to prevent abusive tax avoidance. For example,
given that changes in tax rates create incentives for firms to record transactions in one period rather
than another, government tax inspectors should closely scrutinize the exact timing of transactions in
periods surrounding the tax rate revision. The results also have implications for external auditors
who often face conflicting incentives between provision of tax minimization advice on one hand
and detection of material cases of client earnings management on the other.
12
References:
Cahan, S.F., “The Effect of Antitrust Investigation on Discretionary Accruals: A Refine Test of the Political Cost Hypothesis,” The Accounting Review (January 1992), pp.77-96.
Chan, H.K. and P.L. Mo, “Tax Holidays and Tax Noncompliance: An Empirical Study of Corporate Tax Audits in China’s Developing Economy,” The Accounting Review (October 2000), pp. 469-484.
Dechow, P.M., R. Sloan and A. Sweeney, “Detecting Earnings Management,” The Accounting Review (April 1995), pp.193-225.
Guenther, D.A., “Earnings Management in Response to Corporate Tax Rate Changes: Evidence from the 1986 Tax Reform Act,” The Accounting Review (January 1994), pp.230-243.
Harris, D., “The Impact of U.S. Tax Law Revision on Multinational Corporations’ Capital Location and Income-Shifting Decisions,” Journal of Accounting Research (Supplement 1993), pp.111-140.
Jones, J., “Earnings Management during Import Relief Investigations,” Journal of Accounting Research (Autumn 1991), pp.193-228.
Lopez, T.J., P.R. Regier and T., Lee, “Identifying Tax-Induced Earnings Management around TRA 86 as a Function of Prior Tax-Aggressive Behavior,” Journal of the American Taxation Association (Fall 1998), pp.37-56.
Maydew, E., “Tax Induced Earnings Management by Firms with Net Operating Losses,” Journal of Accounting Research (Spring 1997), pp.83-95.
McNichols, M. and G. Wilson, “Evidence of Earnings Management from the Provision for Bad Debts,” Journal of Accounting Research (Supplement 1998.), pp.1-31.
Mills, L.F., “Book-Tax Differences and Internal Revenue Service Adjustments,” Journal of Accounting Research (Autumn 1998), pp.343-356.
Scholes, M.S., G.P. Wilson M.A. Wolfson, “Firms’ Responses to Anticipated Reductions in Tax Rates: The Tax Reform Act of 1986,” Journal of Accounting Research (Supplement 1992.), pp.161-185.
Shackelford, D. and T. Shevlin, “Empirical Tax Research in Accounting,” Journal of Accounting and Economics (September 2001), pp.321-388.
South China Morning Post (SCMP), Beijing Recovers Hong Kong$14b in Clampdown on Tax Dodgers (September 28, 2002).
Tanzil, V. and H. Zee, “Tax Policy for Developing Countries,” Economic Issues (IMF, 2001).
Watts, R.L. and J.L. Zimmerman, “Towards a Positive Theory of the Determination of Accounting Standards,” The Accounting Review (January 1978), pp.112-34.
14
Figure 1
The magnitude of the tax rate change and income shifting across periods
Tax exemption period (0%) Tax reduction period (15%) Post-holiday period (30%)
Year 1 Year 2 Year 3 Year 4 Year 5 Year 6
Accruals increase
Accelerate revenue and/or defer expenses
Income====shifting
Accruals decrease (reversal s )
Accruals increase
Accelerate revenue and/or defer expenses
Income====shifting
Accruals decrease (reversal s )
15
Table 1
Descriptive statistics for sample firms by year (n=750)
Tax exemption period
Tax reductionperiod
Post-holidayperiod
t-tests for significant between-year differences a
Pooled Year 1 Year 2 Year 3 Year 4 Year 5 Year 6
Current accruals / total assets
0.032b
(0.214)c0.013
(0.090)0.066
(0.234)0.011
(0.123)0.020
(0.151)0.069
(0.207)0.018
(0.105)1&2, 1&5, 2&3, 2&6, 3&5, 5&6
Natural log of assets 4.583(0.463)
4.537(0.456)
4.551(0.450)
4.573(0.456)
4.582(0.463)
4.618(0.474)
4.636(0.480)
1&6, 2&6
Accounts receivable / total assets
0.017(0.124)
0.007(0.060)
0.035(0.140)
0.008(0.070)
0.010(0.010)
0.033(0.020)
0.011(0.130)
1&2, 1&5, 2&3, 2&4,2&6, 3&5, 4&5, 5&6
Inventory / totalassets
0.020(0.108)
0.010(0.040)
0.036(0.201)
0.009(0.074)
0.013(0.101)
0.041(0.116)
0.013(0.130)
1&2, 1&4, 2&3,3&5, 4&5, 5&6
Accounts payable /total assets
0.003(0.031)
0.002(0.027)
0.003(0.065)
0.005(0.032)
0.002(0.016)
0.004(0.028)
0.003(0.020)
1&3
Accrued expenses/ total assets
0.002(0.005)
0.002(0.004)
0.002(0.004)
0.001(0.005)
0.001(0.004)
0.001(0.008)
0.003(0.007)
4&6, 5&6
a t-tests of differences in means, significant at the 5% level.b Mean.c Standard deviation
16
Table 2
Mean discretionary accruals for sample firms by year (n=750)
Tax exemption period Tax reduction period Post-holiday t-tests for significant
Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 between-year differences a
Discretionary accruals(two-tailed p-value)
0.0011(0.341)
0.0088(0.048)
0.0016(0.286)
0.0022(0.196)
0.0105(0.001)
0.0030(0.175)
1&2, 1&5, 2&3, 2&4,2&6, 3&5, 4&5, 5&6
Discretionary accruals are computed as the error term from the following regression:
where:CAit = (ARit + INVit) – (APit + AEit)
where:CAit = current accruals for firm i in year t,ARit = the change in accounts receivable for firm i from year t-1 to t,INVit = the change in inventory for firm i from year t-1 to t,APit = the change in accounts payable for firm i from year t-1 to t,AEit = the change in accrued expense for firm i from year t-1 to t.
ASSETSit-1 = total assets for sample firm i in year t-1,SALESit = the change in sales for firm i from year t-1 to t,YRj = 1 for year j (j = 1 to 5), 0 otherwise,FIRMk = 1 for firm k (k = 1 to 124), 0 otherwise,
a t-tests of differences in means, significant at the 5% level.
17
Table 3
OLS regression of current accruals on the change in sales and dummy variables representing each firm and year
Expected Sign Coefficient t-value p-value
1/ASSETS it-1 0.0078 0.531 0.601
SALESit/ASSETS it-1 + 0.0569 3.532 0.000***
F-statistic = 6.55, p = 0.001, adjusted R2 =0.082.
Estimation model:
(See Table 2 for definitions of variables)
The model was estimated by using least square regression for 750 observations related to 125 firms for 6 years. For brevity, coefficients and t-values for the dummy variables are not reported.
*** Significant at the 1% level.
18
Table 4
OLS regression of discretionary accruals on tax-holiday year and control variables (n=750)
Expected sign Coefficient t-value p-value
Intercept 0.0269 0.162 0.750
YR i2 + 0.0089 2.104 0.034**
YR i3 - -0.0012 -1.152 0.148
YR i5 + 0.0113 2.179 0.013**
YR i6 - -0.0003 -0.094 0.821
INDUSTRYi ? 0.0019 0.286 0.570
SIZEit - -0.0038 -1.890 0.053*
JVi - -0.0025 -0.844 0.391
OWNERSHIPi - 0.0013 0.189 0.776
EXEMPTIONi - -0.0021 -0.838 0.412
F-statistic = 7.09, p = 0.001, adjusted R2 = 0.099.
where:DAit = discretionary accruals estimate for firm i in year t,YRi2 = 1, if firm i is in the 2nd year of the five-year tax-holiday period, 0 otherwise,YRi3 = 1, if firm i is in the 3rd year of the five-year tax-holiday period, 0 otherwise,YRi5 = 1, if firm i is in the 5th year of the five-year tax-holiday period, 0 otherwise,YRi6 = 1, if firm i is in the year immediately following the five-year tax-holiday period, 0 otherwise,INDUSTRYi = 1, if firm i is in manufacturing industry, 0 otherwise,SIZEit = natural log of total assets for firm i in year t,JVi = 1, if firm i is a joint venture in year t, 0 otherwise,OWNERSHIPi = 1, if firm i is sourced from Hong Kong or Taiwan in year t, 0 otherwise.EXEMPTIONi = 1, if firm’s parent company uses a tax exemption system, 0 otherwise.
*, ** Significant at the 10% and 5% level, respectively.
19
Endnotes:
1. For example, Harris (1993), Guenther (1994), Lopez et al. (1998), Maydew (1997), and
Scholes et al. (1992) provide evidence that accounting earnings of U.S. publicly held firms
are managed in response to changes in corporate income tax rates. As different social
contexts and business environments may create different incentives and opportunities for
earnings management, corporate reporting behavior in developed countries may not be
similar to that in developing countries.
2. Although both types of firms have incentives to avoid corporate taxation, privately held firms
are more aggressive than publicly traded firms because 1) they are expected to have smaller
financial reporting costs than publicly held firms (Mills, 1998), and 2) their share ownership
is often highly concentrated and the owners are usually managers of the same firm.
3. For example, a firm can shift income across years via accelerating or delaying shipments of
goods to produce a more appropriate sales figure. Firms that accelerate or defer financial
statement income will in many cases be accelerating or deferring taxable income as well
(Guenther, 1994). While firms reporting conforming book income in the same accounting
period are less likely to trigger tax audit adjustments (Mills, 1998), firms shifting income
between periods surrounding the tax rate change could save current taxes.
4. For example, China’s anti-avoidance campaign in 2002 resulted in a recovery of
underreported tax payment of about US$1.82 billion (SCMP, 2002). The effect of tax
avoidance on developing countries’ economy is more pronounced because the governments
face large fiscal deficits and rely heavily on public sector borrowing.
5. When the normal five-year tax holiday expires, technologically advanced or export-oriented
firms can enjoy a 50% tax reduction for another three years.
6. However, firms established in the designated areas such as Special Economic Zones and
Economic and Technological Development Zones pay enterprise income tax at the reduced
rate of 15% in the post-holiday period.
7. However, there are nontax costs associated with accruals management. For example, delaying
deliveries of finished goods may deteriorate customer relations and increase inventory
holding costs. There are also nontax financial reporting costs associated with reporting low
levels of accounting income (Guenther, 1994; Mills, 1998). For example, costs of violating
debt covenant restrictions and costs associated with negative market reaction. Thus, managers
20
who engage in this form of earnings management typically trade off potential tax savings
with nontax costs. As the sample firms in this study are all privately held, accruals
management may have little capital market effect.
8. Because taxable income is not readily available, some researchers (e.g., Harris, 1993;
Guenther, 1994; Maydew, 1997; Scholes et al., 1992) use book income as a surrogate of
taxable income, and argue that book income shifting is equivalent to taxable income shifting.
As there is a close link between financial reporting and tax in China, book income can be used
to infer taxable income.
9. Because accruals cannot be computed without a lagged year, seven consecutive years of data
are needed to compute accruals for six years (i.e., the five tax-holiday years plus one year
after).
10. To facilitate selection of tax audits, most large tax bureaus in China have developed computer
programs to classify FIEs into three audit classes according to the firm’s perceived likelihood
of tax avoidance and evasion. Class “A” firms are “good” taxpayers that are subject to tax
audits every two years. While class “C” firms are taxed based on a deemed profit rate because
they do not keep a full set of books for audit, class “B” firms represent the majority of FIEs
that are subject to annual audit when resources permit.