N9 INTERNATIONAL TAX TREATIES AND FDI

thue quoc te

Empirical Analysis of International Tax Treaties and Foreign Direct

Taro Ohno

Economist, Policy Research Institute, Ministry of Finance of Japan

Abstract

The author conducted an empirical analysis as to how the tax treaties Japan executed influenced Japan’s foreign direct investments (covering 13 Asian countries between 1981 and 2003) The result shows that among the tax treaties that Japan executed in the recent 20 years, newly concluded treaties had

a statistically significant long-term positive effects on the investment scale, while tax treaties revised during the same period showed no significant effects on investment scale The author also verified the indirect effects of the tax treaties that Japan executed with another country on Japan’s investments in a third country, and found no statistically significant effects in the cases of both newly concluded treaties and revised treaties (at least in the short-term)

Keyword: International Taxation, International Tax Treaty, Foreign Direct Investment

∗ In writing this article, the author would like to express appreciation to all of the following people for their valuable advice: Mr Masayuki Goto (Ministry of Finance), Mr Motohiro Sato (Hitotsubashi University), Mr Masayoshi Hayashi (Hitotsubashi University), Ms Satoko Maekawa (Kansai University), and Mr Tamon Yamada (Keio University) The author is solely responsible for the content of this article Opinions expressed in this article have nothing to do with any statement issued by the Ministry of Finance or the Policy Research Institute Lastly, this article

is based on a study first reported in the Ohno(2009), ‘Sozei Jouyaku to Kaigai Chokusetsu Toushi no Jisshou Bunseki’,

Financial Review, Vol.94, pp.172-190 (in Japanese)

The rise of international investment activity has raised taxation issues in two respects The first involves the occurrence of international double taxation and the second involves the prevalence of international tax avoidance This article will look into these issues one at a time First, the right of taxation

is a sovereign right and therefore is basically exercised based on the discretion of each nation Therefore, every nation often imposes tax on the domestic source income of foreign corporations and nonresidents, not to mention on the worldwide income of domestic corporations and residents However, this has created a problem For instance, when a multinational corporation gains an income from its production activity at a plant it established in a foreign country, a tax is levied on the corporation first in the income-source country (source taxation) In addition, if the corporation remits the income to its head office in the home country and is levied a tax (residence taxation), it constitutes a double taxation on the same income (occurrence of international double taxation) Since such double taxation may hinder international investment activity, many countries have concluded tax treaties in a bid to solve the problem Second, the internationalization of investment makes it more difficult for tax authorities to capture income For instance, a government may want to levy tax on all income from around the world from the standpoint of fairness, but it is often more difficult to capture foreign source income than to capture domestic source income For this reason, if the government takes no measures, investors may be prompted to avoid tax (prevalence of international tax avoidance) Because this is a problem common to many countries, countries have cooperated with one another to resolve the issue by concluding tax treaties Tax treaties have two objectives, first to eliminate international double taxation and second to prevent international tax avoidance

A tax treaty can influence foreign direct investment in two possible ways From the viewpoint of the first objective, tax treaties contribute to an increase in foreign direct investments because they reduce harm of investments by eliminating international double taxation On the other hand, from the viewpoint

of the second objective, tax treaties may contribute to a reduction in the investment scale, because they discourage international tax avoidance In light of this nature of tax treaties, does the tax treaty actually influence foreign direct investment? If it does, which of the effects above would be relatively greater? The objective of this article is to empirically find the answer to these questions

Here, I would like to explain the features of this article, while referring to existing studies In the last

20 years or so, many studies have been made on the effects of taxation (in particular, the tax rate) on foreign direct investment,1 and in more recent years, the effects of tax treaties on foreign direct investment have come to be examined Blonigen and Davies (2000, 2004) studied the effects of new tax treaties on U.S outbound and inbound foreign direct investments The study found that executing new tax treaties reduced U.S direct foreign investment and concluded that tax treaties contribute largely to prevent international tax avoidance Davies (2003) also examined the effect of the revision of tax treaties

on U.S outbound and inbound foreign direct investments, reaching the conclusion that the revision of tax treaties does not have an impact on foreign direct investment In recent empirical analyses, whether or not

1 Among survey reports concerning these types of studies are those by Gordon and Hines (2002), and de Mooij and Ederveen (2003)

there was a conclusion of tax treaties has come to be added to explanatory variables In his analysis of cross-border M&As, di Giovanni (2005) showed that the conclusion of a tax treaty increases foreign direct investment Stein and Daude (2007), in their analysis of 17 OECD countries’ foreign direct investments in 58 countries, showed that the conclusion of a tax treaty also increases foreign direct investment On the other hand, Louie and Rousslang (2008) showed that the conclusion of a tax treaty does not influence foreign direct investments by U.S corporations Meanwhile, in Japan, although several studies have been made on the relationship between the tax rate and foreign direct investment,2the effects of tax treaties on foreign direct investment have yet to be verified Incidentally, Hines (2001), Hotei (2006), and Azemar et al (2007) focused on Japan’s "tax sparing credit system", which is a provision of tax treaties, and examined its effect on foreign direct investment

Based on the above, the features of this article are set out in order as follows First, this article verifies the effects of the tax treaties that Japan has executed (newly executed or revised) on foreign direct investment by Japan So far, no study has been made to verify the effects of tax treaties executed by Japan When studying the effect of taxation factors on investment, it is also important to take into account factors other than the tax rate For instance, it is important to take into account such factors as avoidance of double taxation, exchanges of information among tax authorities, transfer pricing taxation, and treaty shopping They are measures that should be covered by a tax treaty In this sense, tax treaties form an important element of taxation factors and it is meaningful to verify their effects Second, this article verifies the long-term and short-term effects of both newly executed treaties and revised treaties Existing studies on tax treaties focus only on the long-term effects in the case of newly executed treaties and on short-term effects in the case of revised treaties However, there are no logical reasons to assume that newly executed treaties have only long-term effects, while the revised treaties have only short-term effects, leaving a need for improvement of those studies in the future Third, this article adopted the

"dynamic panel-data estimation" (system GMM estimator) Among already existing studies on tax treaties, none have adopted the dynamic panel-data estimation However, it is important to consider the previous year’s level of foreign direct investment, because it is natural to assume that it significantly influences the following year’s level Also, when verifying the effects of the tax treaties, endogeneity (arising from simultaneity) is a concern, but the dynamic panel estimation makes it possible to deal with this issue as well

The specific target of this article is Japan’s outbound foreign direct investment (in 13 Asian countries

in 1981 through 2003) Tajika, Ohno, and Hotei (2007) looked at the recent trends concerning Japan’s outbound foreign direct investments by using statistical indicators, such as the "Balance of Payments Statistics" of the Ministry of Finance and the "Basic Survey of Overseas Business Activities" of the Ministry of Economy, Trade and Industry They found that Asia accounted for the highest percentage in

2 Among such studies are those by Hidaka and Maeda (1994), Chong (1996), Fukao and Yue (1997), and Maekawa (2005) These studies can be positioned as part of a series of “factor analyses of foreign direct investment” including the tax rate as a factor Among other studies on “factor analysis of foreign direct investment” (not including the tax rate) are those by Tokunaga and Ishii (1995), Urata (1996), Fukao (1996), Fukao and Chong (1996), and Wakasugi (1997)

terms of the amount of income on foreign direct investment returned to Japan from abroad With Japanese corporations’ business activities abroad steadily increasing, the ratio of profits to sales of Japanese corporations operating in Asia is higher than those of their counterparts operating in the U.S., Europe or other regions In addition, in terms of the ratio of retained earnings gained by local subsidiaries, the ratio

of such earnings remaining in Asia is smaller than those remaining in other regions, meaning the amount

of earnings remitted to Japanese head offices from Asia is larger than those from other regions This suggests that Asia has become an important direct investment destination for Japanese corporations In view of the above, this article focuses its analysis on Asian countries

The result shows that among the tax treaties that Japan has executed in the last 20 years, newly concluded treaties had statistically significant long-term positive effects on the scale of investment, while tax treaties revised during the same period showed no significant effects on such scale The author also verified the indirect effects of the tax treaties that Japan executed with other countries on Japan’s investments in third countries, and found no statistically significant effects in the case of both newly concluded treaties and revised treaties (at least in the short-term)

The structure of this article is as follows Section 2 puts in order the concept of the functions of tax treaties and their effects on foreign direct investments, focusing on the contents of the tax treaties that Japan has actually concluded Based on this, Section 3 estimates the effects of the tax treaties that Japan has executed on foreign direct investment by Japan Lastly, in Section 4, the author presents a conclusion based on the results of the analysis

II Tax Treaty

II.1 Objective, Direction and Function of Tax Treaties

A tax treaty is concluded between two countries So far, Japan has concluded 45 tax treaties with 56 countries (as of December 2008).3 Japan’s tax treaty network covers more than 90% of the country’s foreign direct investment (in terms of accumulated value)

Although the contents of tax treaties vary from one tax treaty to another, they have a lot in common,

as most of them are based on the OECD Model Tax Convention Let’s take a look at the objective, direction and function of tax treaties

3 The reason for the number of countries being larger than the number of treaties is that the treaties concluded with the former Soviet Union and Czechoslovakia have been maintained by the countries that came into being after the breakup of the two countries Also, the treaty concluded with the U.K has been inherited by Fiji

Table 1: Japan’s Tax Treaty Network (45 treaties, 56 countries)

Western Europe

(15) Eastern Europe/former Soviet Union (17) Asia (12) Africa/Middle East (5) Oceania (3)

North America/Central and South America (4) Ireland

India Indonesia South Korea Singapore Sri Lanka Thailand China (c) Pakistan Bangladesh Philippines Vietnam Malaysia

Israel Egypt Zambia Turkey South Africa

Australia New Zealand Fiji (d)

U.S

Canada Brazil Mexico

(Source) Ministry of Finance website

(Note 1) As of December 2008

(Note 2) a: The treaty concluded with the former Soviet Union has been inherited

b: The treaty concluded with former Czechoslovakia has been inherited

c: Not applied to Hong Kong and Macao

d: The original treaty with the U.K has been inherited

<Objective of tax treaty>

The objectives of tax treaties are as follows

(1) Elimination of international double taxation

(2) Prevention of international tax avoidance

In order to achieve these objectives, various provisions are provided in tax treaties, which will be described in the <Function of tax treaties> section

<Direction of tax treaty>

The following two points can be cited as the direction of tax treaties

(1) Basically, they aim at allocating taxation rights, mainly residence taxation

(2) However, they allow source taxation in some tax bases, as source taxation has

its own merit

Problems of international double taxation mainly occur when both contracting countries of a tax treaty impose tax on the same income Therefore, if both parties agree to adopt either source taxation or residence taxation, the problem of double taxation can be solved

Then, the question arises as to which taxation—residence or source taxation—the two parties should adopt, and tax treaties mainly offer the above mentioned directions Before explaining the reasons behind the directions, let’s take a look at the good points and bad points of residence taxation and source taxation One of the good points of residence taxation is that it can ensure the fairness of taxation and the efficiency

of investment From the standpoint of fairness, it is desirable for a government to impose a tax on the worldwide income of domestic corporations and residents It is also desirable from the standpoint of efficiency, since it ensures the neutrality (capital export neutrality) of investment by imposing the same tax rate on similar income wherever an investment is made On the other hand, one of the bad points of residence taxation is that it is difficult to capture foreign income From the standpoint of fairness, the government wants to impose a tax on the worldwide income of domestic corporations and residents However, capturing their foreign source income is often more difficult than capturing their domestic source income By contrast, one of the good points of source taxation is that it is easy to capture income and, therefore, taxation cost is relatively low This is because source taxation covers only domestic source income This point is the exact opposite of the bad point of residence taxation However, the bad points of source taxation are that the fairness of taxation and the efficiency of investment are not always ensured Since source taxation does not cover foreign income, the fairness of taxation on a worldwide level cannot

be achieved Moreover, since each country has tax autonomy, the tax rate applied to similar income normally differs from one country to another under source taxation In other words, there is no guarantee that the same tax rate will be applied In this sense, the neutrality of investment (capital export neutrality)

is not ensured and, therefore, source taxation is not desirable also from the standpoint of efficiency This point is the exact opposite of the good point of residence taxation

Therefore, tax treaties are moving in the direction of attaching relative importance to the good points

of residence taxation but also of partially allowing source taxation to cover the bad points of residence taxation.4

4 This direction is in line with the “OECD Model Tax Convention” and is reflected mainly in tax treaties concluded between advanced countries In the case of advanced countries, since their capital exports are relatively large, giving more priority to residence taxation is in the interest of the governments in question In the case of developing countries, however, since their capital imports are often larger than their capital exports, giving more priority to residence taxation is not in their best interests Rather, they want to expand the scope of application of source taxation Therefore, when a tax treaty is signed between an advanced country and a developing country, or between developing countries, the scope of application of source taxation is often wider under such a treaty than under the OECD Model Tax Convention Japan also takes this point into account when it concludes a tax treaty with a developing country, while basing it on the OECD Model Tax Convention

<Functions of tax treaty>

Here, I would like to discuss the functions of tax treaties and their effects on investment A tax treaty has the following five functions:5

(1) Allocation of tax jurisdictions (Positive effect on investment)

(2) Adjustment of double taxation (Positive effect on investment)

(3) Prevention of international tax avoidance (Negative effect on investment)

(4) Resolution of international tax issues (Positive effect on investment)

(5) Economic cooperation for developing countries (Positive effect on investment)

(1) Allocation of tax jurisdictions (Positive effect on investment)

In tax treaties, the allocation of tax jurisdictions on individual income is coordinated between two sovereign nations Specifically, the two sovereign nations decide whether to adopt residence taxation or source taxation depending on the nature of individual income For instance:

• With regard to "international traffic income," both countries adopt residence taxation

• With regard to "real estate income," both countries adopt source taxation

These efforts are expected to contribute to the reduction and prevention of the risk of double taxation for investors and, as a result, have a positive effect on investment activity Incidentally, in the cases of certain incomes, tax jurisdictions are not thoroughly allocated and both residence taxation and source taxation are allowed on an exceptional basis For instance:

• "Business income" basically falls under residence taxation However, when there is a permanent facility, source taxation may be applied within the scope of the income imputed to the facility

• "Investment income (dividends, interest, royalties, etc.)" also basically falls under residence taxation, but source taxation may be applied

The efforts are basically seen as residence taxation but also take the good points of source taxation (ease of capturing income) into account

(2) Adjustment of double taxation (Positive effect on investment)

As described above, in the case of certain incomes, the tax jurisdictions are not thoroughly allocated and both residence taxation and source taxation are in place This may cause the problem of double taxation Therefore, remedial measures are taken, such as setting:

5 The method of classifying the functions is based on Tax Treaty Study Group (1997)

• "double taxation relief" (the "foreign tax credit system" in the case of Japan)

• limiting the tax rates of source taxation (limited rate of taxation or reduced tax rate) on "investment income (dividends, interest, royalties, etc.)."

These efforts are considered to have contributed to the elimination or reduction of double taxation for investors and, therefore, have a positive effect on investment activity However, in the case of the Japanese foreign tax credit system, it should be kept in mind that since the system is applicable even before a treaty is concluded, it does not exert any new effect after the treaty is concluded

(3) Prevention of international tax avoidance (Negative effect on investment)

Tax treaties are also measures to cope with international tax avoidance The following are examples of such measures:

• Measures for transfer pricing (stipulating that transfer pricing is not to be allowed under a "specified affiliated company provision")

• "Exchanges of information" (increasing the capture of foreign source income) between the tax authorities of the two countries involved

• "Mutual assistance in tax recovery" between the tax authorities of the two countries involved (cooperation to recover tax when a person not covered by the treaty is enjoying the benefits of the treaty)

In recent years, other measures have been established to prevent tax avoidance, such as: the

"limitation of benefits provision" (in 2004, following the new Japan-U.S tax treaty), which enables the strict screening of persons eligible for the benefits of the treaty beforehand; the "silent partnership agreement provision" (in 2006, following the new Japan-U.K tax treaty), which stipulates source taxation

on income arising from a silent partnership agreement; and the "anti-conduit provision" and the

"anti-abuse provision" (in 2007, following the new Japan-France tax treaty), both focused on the forms and objectives of transactions to limit treaty benefits on investment income These efforts are expected to contribute to the reduction of tax avoidance and, as a result, have a negative effect on the size of investment (although a question remains as to the quality of investment, in the sense that it is nonproductive activity)

(4) Resolution of international tax issues (Positive effect on investment)

Tax treaties also provide measures to solve problems that were not assumed at the time of the conclusion of the treaties, including:

• mutual agreement procedures for the tax authorities of the two countries involved (government responses in the case of an occurrence of double taxation not stipulated in the treaty)

Incidentally, in the case of double taxation arising from transfer pricing taxation, many treaties have a separate stipulation on coordination through mutual agreement (correlative adjustment provision) They also set a time limit for the correction of transfer pricing and stipulate that the transfer pricing problems of more than 7 years will not be taken up (time limitation on correction) These efforts are expected to contribute to the elimination or reduction of risk uncertainties for investors and, therefore, have a positive effect on investment activity

(5) Economic cooperation for developing countries (Positive effect on investment)

When a tax treaty is concluded between an advanced country and a developing country, measures for economic cooperation for the benefit of the developing country are sometimes added to the treaty Among examples of such measures is:

• The "tax sparing credit system"

A tax sparing credit system has been incorporated in some of the tax treaties that Japan has executed

so far Since the system directly eases investors’ tax burden, it is expected to have a positive effect on investment activity

The above reveals that tax treaties perform more than one function at a time As noted above, the objectives of tax treaties are the elimination of international double taxation and the prevention of international tax avoidance Function (1): allocation of tax jurisdictions, function (2): coordination of double taxation, and function (4): resolution of international tax issues are designed to realize the former objective, while function (3): elimination of international tax avoidance is to realize the latter objective (Function (5): economic cooperation for developing countries can be viewed more or less as an exceptional measure.) Since the former has a positive effect and the latter has a negative effect on investment, the conclusion of a tax treaty can have both positive and negative effects on investment as a result

II.2 Main points of recent treaty revision

It’s not that a full set of the tax treaty functions has been in place since tax treaties began to be concluded in around 1955 The contents of tax treaties have been added, modified and refined in tandem with changes in the economic environment over time (In other words, the contents of newly concluded tax treaties are more refined than those of previous treaties.) Since tax treaties concluded in the past tend

to become obsolete as time passes, leading to double taxation and tax avoidance, it is always necessary to revise such treaties

Here, the author would like to organize the main points addressed in the revision of treaties in recent years (from 1981 to around 2003) The following are the main points of revision in accordance with the classification used in the <Functions of tax treaties> described above

(1) Allocation of tax jurisdictions (positive effect on investment)

• Creation of "real estate income," "capital gain," and "other income" provisions

• Revision of the contents of the "international traffic income" provision

(2) Adjustment of double taxation (positive effect on investment)

• Lowering the limited rate of taxation on "investment income"

(3) Prevention of international tax avoidance (negative effect on investment)

• Establishment of a "mutual assistance" provision (from around 1992)

(4) Resolution of international tax issues (positive effect on investment)

• Establishment of a "correlative adjustment provision" (from around 1999)

• Establishment of a "time limitation for correction (concerning transfer price)" (from around 1999)

(5) Economic cooperation benefiting developing countries (positive effect on investment)

• Establishment of a "tax sparing credit system" (but on a temporary basis since around 1991)

The above shows that in recent revisions, contents have been added or modified in all of the five tax treaty functions Therefore, treaty revision may have both positive and negative effects on investment Lastly, the author would like to confirm the flow of tax treaty conclusion and revision by Japan Table

2 is a list of years in which Japan concluded or revised tax treaties with other countries since 1981 Treaty revision is classified into two types: total revision and partial revision However, there is no clear distinction between the two They are only classified so in response to the announcement by the Ministry

of Finance Basically, the difference seems to be in the number of items In terms of the effect of revision

on investment, there is no particular reason to believe there is a difference between total revision and partial revision The effect of revision depends on the contents of revision

Table 2: Tax Treaty Conclusion/Revision in Japan (year of effect)

(Source) "Tax treaty-related laws; 2008 edition"

(Note 1) As of December 2008

(Note 2) The effect of conclusions and revisions of treaties with the countries underlined are estimated in this article

III Empirical Analysis

III.1 Formulation, variables, sign condition

Before estimating the effect of a tax treaty on foreign direct investment, the author would like to organize the objects to be estimated The object of estimation are Japan’s outbound foreign direct investment Of the tax treaties, those that were newly executed or revised (total revision) by Japan were taken up to estimate their effects The author used panel data for the 23 years from 1981 to 2003 The

analysis covered 13 Asian countries.6 During the estimation period, Japan concluded four tax treaties and revised five (total revision)

In this article, the estimation model used by Blonigen and Davies (2000) was extended to a dynamic model:7

it i j

it j j k

it k k it

y = δ ⋅ −1+ ∑ α ⋅ . + ∑ β ⋅ , + μ +(yit) is a dependent variable, (yit-1) is a lagged dependent variable, (xk,it) is a tax treaty dummy, (zj,it) is

an explanatory variable, (δ), (αk), (βj) are coefficients, (μi) is a fixed effect, (uit) is a disturbance term, (i) is the host country, and (t) is the year

Let’s take a look at the variables and sign conditions.8 The dependent variable is the annual level of Japan’s "direct investment (flow, gross)."

Explanatory variables are as follows One is the "direct investment of the previous year (lagged dependent variable)." The larger the investment in the previous year, the more the investment increases (the sign condition is positive)

Next is the "real GDP of the host country." The larger the economic size of the host country is, the more the investment increases (the sign condition is positive) In this case, however, horizontal direct investment, rather than vertical direct investment, is taken into account.9

Next is the "real GDP of the home country (Japan)." The higher the GDP of the home country is, the more the investment increases (the sign condition is positive)

Next is the "trade cost of the host country." The effect of the trade cost on direct investment depends

on whether tradable goods and investment goods are substitutive or complementary When they are substitutive, the higher the trade (import) cost is, the more the investment increases as an alternative means However, when they are complementary, the higher the trade (import) cost is, the more the investment decreases When tradable goods are production factors or intermediate goods, the higher the trade (import) cost is, the more the investment decreases, as the higher cost impedes material procurement

In the case of vertical direct investment, the higher the trade (export) cost is, the more the investment

8 The system GMM estimator adopted in this article postulates the existence of fixed factors The factors that do not change during the estimation period are all absorbed into fixed factors (language, religion, distance and national land area, etc.) Therefore, the control variables in this article do not include those that may become fixed factors

9 Direct investment can be roughly divided into two different kinds: one is “horizontal direct investment” and the other is “vertical direct investment.” “Horizontal direct investment” is the type of investment that multinational corporations make in order to expand the market for their products Most of the products produced in the host country are sold and consumed in the local markets On the other hand, “vertical direct investment” is the type of investment that multinational corporations make in order to take advantage of low-cost production factors Most of the products produced in the host country are not sold and consumed domestically but are transported to the home country or third countries

decreases, as the higher cost impedes product transportation Therefore, the sign condition can be positive

or negative

Next is the "trade cost of the home country (Japan)." The effect of the trade cost of this country also depends on whether tradable goods and investment goods are substitutive or complementary When they are substitutive, the higher the trade (export) cost is, the more the investment, as an alternative means, increases However, when they are complementary, the higher the trade (export) cost is, the more the investment decreases When tradable goods are production factors or intermediate goods, the higher the trade (import) cost is, the more the investment increases, as the higher cost impedes material procurement

In the case of vertical direct investment, the higher the trade (import) cost is, the more the investment decreases, as the higher cost impedes product transportation Therefore, the sign condition can be positive

Next is the "inflation rate" (Tokunaga and Ishii 1995, Urata 1996, Wakasugi 1997) The higher the inflation rate of the host country is, the more the investment decreases, as higher inflation destabilizes the economic condition of the country (the sign condition is negative)

Nest is the "foreign exchange rate (local currency unit/¥)" (Tokunaga and Ishii 1995, Maekawa 2005) The stronger the Japanese yen is vis-à-vis the currency of the host country (i.e., the higher the indicator is), the more the investment increases, as the stronger yen facilitates material procurement On the other hand, the stronger the Japanese yen is vis-à-vis the currency of the host country (i.e., the higher the indicator is), the more the investment may decrease, as the profits gained in the host country decrease in terms of yen Therefore, the sign condition can be positive or negative

Next is the "level of investment safety" (Chong 1996, Fukao 1996, Fukao and Chong 1996, Fukao and Yue 1997) The higher the level of investment safety of the host country is, the more the investment increases (the sign condition is positive)

Nest is the "time dummy" In this article, the time dummy for each year is incorporated in order to remove annual shocks common to nations

The last is the "tax treaty dummy." As described earlier, the effect of the conclusion of a tax treaty, regardless of whether it is newly executed or revised, on investment can be positive or negative