## Sea-Jin Chang

Print publication date: 2013

Print ISBN-13: 9780199687077

Published to Oxford Scholarship Online: January 2014

DOI: 10.1093/acprof:oso/9780199687077.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 25 February 2017

# (p.215) Appendix 10 Income Shifting Practices in China

Source:
Multinational Firms in China
Publisher:
Oxford University Press

This appendix summarizes the major findings of my ongoing work with research colleagues that examines the degree of income shifting among multinational firms (Chang, Chung, and Moon 2012). If income shifting is occurring among multinational firms operating in China, foreign subsidiaries with parents from countries where corporate income tax rates are lower than that of China will have motivation to shift income to their parent firms in order to minimize global corporate tax liability (e.g., Grubert and Mutti 1991; Hines and Rice 1994).

In order to test this income shifting hypothesis, we first need to determine the nationality of multinational parents. The Survey of Foreign-Invested Enterprises conducted by the Chinese National Bureau of Statistics (NBS) in 2001 allows us to identify the nationality of multinational parents who together owned 150, 435 foreign subsidiaries in 2001. We then match this information with the Annual Industrial Survey Database (2001–7) from NBS in order to acquire each firm’s financial information. Matching the 2001 Survey of Foreign-Invested Enterprises and Annual Industrial Survey Database generates a list of 51, 118 foreign subsidiaries. We further restricted our sample to firms whose parents are from OECD countries as of 2001 or from Hong Kong, Macao, Taiwan, or Singapore (HMTS), as well as to wholly owned subsidiaries. We focus on OECD countries since we can obtain reliable data on corporate tax rates. We also include subsidiaries from HMTS because they represent the lion’s share of FDI inflows to China, though they are not members of the (p.216) OECD. Of the 42, 388 firms from OECD or HMTS countries, 47.4% are from Hong Kong, 17.9% from Taiwan, 10.9% from Japan, 7.3% from the US, 6.9% from Korea, 2.2% from the UK, 1.6% from Germany, 1.5% from Macao, 0.8% from Australia, and 0.6% from France, the Netherlands, and Canada, respectively. These top 12 countries represent 98.08% of our sample firms. We also focus on wholly owned subsidiaries because here, parent firms command full control over wholly owned subsidiaries in deciding whether or not to income shift and adjust transfer prices. Of the 42, 388 foreign subsidiaries of multinational parents from OECD and HMTS countries, 25, 177 firms are wholly owned subsidiaries. We then build panel data for these 25, 177 (104, 105 firm-year observations). Our panel data for 2001–7 coincides with the aftermath of China’s initiation into the World Trade Organization in 2001, when China officially removed various trade and investment barriers like joint venture requirements and restrictions on official repatriation of profits. Thus, we observe income shifting practices of foreign multinationals in China when there are no official barriers to profit repatriation.

We use operating profit as our measure of firm performance, defined as profit before deducting tax and financial expenses (in thousand RMB). We focus on operating profit as multinational firms’ various uses of transfer prices, for the purpose of income shifting, will directly affect profits from their operations (e.g., Jacob 1996; Newberry and Dhaliwal 2001; Clausing 2003; Desai, Foley, and Hines 2004, 2006; Eden, Juarez Valdez, and Li 2005; Bernard, Jensen, and Schott, 2006). Home tax rate is the corporate income tax rate of the foreign parent’s home country, measured by summing the national and average local corporate income tax rates, following the OECD’s reporting standard as collected by the OECD Tax Database.9

Corporate tax information for HMTS countries is provided by KPMG (2007). We include the tax rates of our sample firms’ countries of origin in Appendix Table 10.1. During our sample period, 2001–7, China’s corporate income tax rate remained stable at 33%. During our sample period, there remained some tax benefits in various Special Economic Zones (SEZs). For instance, manufacturers in designated SEZs could (p.217)

Appendix Table 10.1. Combined corporate income tax rates by foreign firms’ home countries

OECD MEMBERSHIP

COUNTRY

2001

2002

2003

2004

2005

2006

2007

Yes

Australia

30

30

30

30

30

30

30

Yes

Austria

34

34

34

34

25

25

25

Yes

Belgium

40.2

40.2

33.99

33.99

33.99

33.99

33.99

Yes

40.51

38.05

35.95

34.42

34.36

34.07

34.09

Yes

Czech Rep.

31

31

31

28

26

24

24

Yes

Denmark

30

30

30

30

28

28

25

Yes

Finland

29

29

29

29

26

26

26

Yes

France

36.43

35.43

35.43

35.43

34.95

34.43

34.43

Yes

Germany

38.9

38.9

40.2

38.9

38.9

38.9

38.9

Yes

Greece

37.5

35

35

35

32

29

25

Yes

Hungary

18

18

18

16

16

17.33

20

Yes

Iceland

30

18

18

18

18

18

18

Yes

Ireland

20

16

12.5

12.5

12.5

12.5

12.5

Yes

Italy

36

36

34

33

33

33

33

Yes

Japan

40.9

40.9

40.9

39.54

39.54

39.54

39.54

Yes

Luxembourg

37.5

30.38

30.38

30.38

30.38

29.63

29.63

Yes

Mexico

35

35

34

33

30

29

28

Yes

Netherlands

35

34.5

34.5

34.5

31.5

29.6

25.5

Yes

New Zealand

33

33

33

33

33

33

33

Yes

Norway

28

28

28

28

28

28

28

Yes

Poland

28

28

27

19

19

19

19

Yes

Portugal

35.2

33

33

27.5

27.5

27.5

26.5

Yes

S. Korea

30.8

29.7

29.7

29.7

27.5

27.5

27.5

Yes

Slovakia

29

25

25

19

19

19

19

Yes

Spain

35

35

35

35

35

35

32.5

Yes

Sweden

28

28

28

28

28

28

28

Yes

Switzerland

24.7

24.4

24.1

24.1

21.3

21.32

21.32

Yes

Turkey

33

33

30

33

30

20

20

Yes

UK

30

30

30

30

30

30

30

Yes

US

39.26

39.3

39.33

39.31

39.28

39.3

39.26

No

Hong Kong

16

16

16

17.5

17.5

17.5

17.5

No

Macao

15

15

15

15

12

12

12

No

Singapore

25.5

24.5

22

22

20

20

20

No

Taiwan

25

25

25

25

25

25

25

No

China

33

33

33

33

33

33

33

Note: OECD data is from the OECD Tax Database (Section C, Table II. 1, accessible at www.oecd.org/ctp/taxdatabase), and non-OECD data is from KPMG (2007) Hong Kong Tax Competitiveness Series: Corporate Tax Rates.

(p.218) enjoy 15% tax rates. Foreign firms located in various development zones and cities could also enjoy 24% tax rates, depending upon the location of the project. We thus need to control for these various economic zones with preferential tax rates. The dichotomous variable Economic zones with preferential tax denotes such locations.

In order to find evidence of income shifting at the subsidiary level, we follow the methodology developed by Bertrand, Mehta, and Mullainathan (2002) to trace the propagation of profit shock through a multinational firm. Consider, for example, a multinational firm with headquarters in a home country and a subsidiary in a host country. Suppose that the corporate income tax rate is lower in the home country than in the host country. Further, suppose that the subsidiary experiences a profit shock such as a sudden and unexpected increase in demand that would cause its profits to rise by RMB 1 in the absence of income shifting. Because the multinational firm has incentive to minimize corporate taxes, some increased profits in the foreign subsidiary may be shifted out of the country. As a result, the actual profits of the subsidiary will rise by less than RMB 1. If the home country has a higher corporate income tax rate than the host country, the situation would be reversed.

In order to calculate the predicted value of the operating profit of a subsidiary after experiencing a profit shock from an industry and a region where the subsidiary belongs, we first calculate the asset-weighted average profitability by dividing the aggregate operating profit of all local firms and foreign subsidiaries in the same industry and province by the aggregate assets of these firms in a given year. For each subsidiary i at time t, predicted profit is calculated by multiplying the asset-weighted industry-region-level average profitability for each industry-region pair JR with the total assets of each subsidiary using the following equation:

(1)
$Display mathematics$
(p.219)

Next, we use a subsidiary fixed-effect model to estimate a subsidiary’s operating profit as a function of predicted profit using the following specification:

(2)
$Display mathematics$

Since a fixed-effect model is used, both performance and predicted performance are measured as deviation from subsidiary-level averages, allowing us to interpret the deviation in predicted operating profit as a response to profit shock. Thus, the coefficient β captures the response of subsidiary performance with respect to market-level shocks in profitability.

(3)
$Display mathematics$

In equation (3), we add the interaction between predicted profit and home country tax rates. We also add the coefficient γ to capture the differential response of the performance variable with respect to market-level shocks in profitability according to those subsidiary-level characteristics. Additionally, we add the interaction between predicted profit and economic zones with special tax rates to control for any tax benefits for firms in those locations.

Appendix Table 10.2 reports the main results. Model (1) is the baseline model. The coefficient for predicted profit is 1.310, which means foreign firms increase their operating profit by RMB 1.310 for each RMB 1 profit shock in the market. Multinational firms may receive more benefit from unexpected profit shocks than local firms because foreign multinationals possess stronger technology or brands. In models (2) and (3), we enter the interaction term between predicted profitability and home country tax as a continuous variable and as a dummy variable. Model (2) shows (p.220)

Appendix Table 10.2 The Extent of Income Shifting According to Tax Rates among Wholly-Owned Subsidiaries

(1)

(2)

(3)

(4)

(5)

SAMPLE

ALL

ALL

ALL

ECONOMIC ZONESWITH SPECIAL TAX RATES

NON-ECONOMICZONES

Predicted profit

1.310***

0.668***

0.861***

0.143

0.843***

(0.038)

(0.195)

(0.187)

(0.382)

(0.196)

Predicted profit * home tax rate

0.017**

0.011*

0.041**

0.012*

(0.005)

(0.005)

(0.016)

(0.005)

Predicted profit * economic zones

-0.745+

with preferential tax

(0.416)

Predicted profit * home tax rate*economic

0.031+

zones with preferential tax

(0.016)

Log(assets)

-1,673.290***

-272.370

-936.175

-1,137.832

-512.492

(391.825)

(494.241)

(646.907)

(938.245)

(603.454)

Year fixed effects

Included

Included

Included

Included

Included

Firm fixed effects

Included

Included

Included

Included

Included

No. of observations

104,105

104,105

104,105

69,578

34,527

0.941

0.942

0.945

0.868

0.983

Note: Standard errors in parentheses;

+ 10%,

(*) 5%,

(**) 1%,

(***) 0.1% significance level, respectively.

(p.221) that wholly owned subsidiaries whose parents are headquartered in a country with higher tax rates report higher profits than expected from these market-level shocks. Other variables being equal, a firm whose parent company comes from a country with a 15% corporate income tax rate will report a RMB 0.923 (= 0.668+0.017*15) increase in operating profit when faced with a RMB 1 profit shock, whereas a firm whose parent comes from a country with 40% corporate income tax rate will report a RMB 1.348 (= 0.668+0.017*40) increase in operating profit in response to the same RMB 1 profit shock.

Model (3) uses the dichotomous variable economic zones with special tax rates and creates a two-way interaction term with predicted profit and a three-way interaction term with predicted profit and economic zones with preferential tax. The three-way interaction term is positively signed and significant, suggesting that firms in various economic zones with special tax rates respond positively to an external shock. Since the tax rates in these various economic zones are lower than other locations, foreign firms located in these economic zones report higher profits than those located elsewhere.

In models (4) and (5), we split the sample into firms located in economic zones with special tax rates and firms located elsewhere. Results show that the coefficient for predicted profit is insignificant for firms located in economic zones, suggesting that foreign firms in economic zones do not respond to industry-wide profitability. Rather, their reported profits depend entirely on the home tax rates. In other words, foreign firms whose home tax rates are high will report higher profits, while those whose home tax rates are low will report lower profits, even though both enjoy the same special tax rates in the same economic zones. Profits of foreign firms located in places other than economic zones also respond positively to their home country tax rate.

These results support our conjecture that the profits of foreign subsidiaries depend on home country corporate tax rate, regardless of location in China. Even foreign firms in economic zones with lower tax rates demonstrate similar patterns. This finding confirms extensive income (p.222) shifting among foreign subsidiaries in China. Thus, the reported low profitability of foreign firms in China does not necessarily mean that they are unprofitable, rather, they just report less profit to the tax authorities in China.

## Notes:

(9) . This database is accessible at www.oecd.org/ctp/taxdatabase. See Section C, Table II.1.