The effects of double tax treaties for developing countries. A case study of Austria's double tax treaty network.
In the past years, there has been an increasing awareness that governments are losing substantial tax revenues due to "aggressive" tax avoidance schemes. The G20 and the OECD strongly promote the Base Erosion and Profit Shifting initiative (BEPS), which aims at undermining aggressive tax planning structures used by multinational companies. This initiative claims that 'fixing' some individual problems of current international tax rules may solve tax avoidance issues (OECD, 2013).
At the same time, there is a different discussion coming from the perspective of developing countries, calling for a more fundamental 'rethinking' of the international tax system. The question is raised as to how the international tax system generally and Double Tax Treaties (DTTs) in particular impact developing countries. It is being discussed whether developing countries at all benefit from the signature of DTTs under current internationally accepted standards.
These internationally accepted standards, which are embodied in the OECD Model Tax Convention on Income and on Capital (henceforth OECD Model), and to a lesser extent in the UN Model Tax Convention, have a large influence on actual tax treaty practice (Wijnen & de Goede, 2014). As both models favor the residence principle, where tax residents of a country are subject to taxation on their worldwide income, a greater portion of taxation rights are allocated to a residence country (Daurer, 2013). DTTs based on these Conventions thereby shift taxing rights from the source state (capital-importing country) to the residence state (capital-exporting country). Between two economies with largely reciprocal foreign direct investment (FDI) positions, this reallocation of taxing rights is not problematic. When, however, such a treaty is signed between two countries with an asymmetric investment position, the capital-importing country risks to forfeit tax revenues. This is not merely a theoretical discussion, but some developing countries have already renegotiated or terminated specific DTTs that they do not perceive as beneficial for themselves. (1)
This on-going debate motivates the present study, which analyses the Austrian DTT network with developing countries. Austria's 37 DTTs signed with developing countries are based on the OECD Model. We investigate in detail how these DTTs impact developing countries. As previous results from cross-country studies regarding the effects of DTTs for developing countries are inconclusive, we add to the literature by providing a combined legal and economic review of Austria's tax treaty policy and its ramifications for developing countries.
Austria, as a case study, presents how the OECD Model can shape a country's treaty policy and also affect its treaty partners. Given that Austria's DTTs are very similar to the OECD Model, Austria is often thought of as a good representative of OECD standards. On the other hand, Austria has been (in)famous for its rather strict bank secrecy laws and can be viewed as a special case in this respect.
Similar case studies have been provided for the Netherlands and Switzerland, two other small and open economies with large tax treaty networks (Burgi & Mayer-Nandi, 2013; McGauran, 2013). These two studies focus more on the analytical legal part, and are more descriptive, partly basing themselves on anecdotal evidence. The combination of a qualitative legal approach, as well as conceptual and quantitative economic analysis allows us to systematically study the impact of individual tax treaty provisions on developing countries. The three case studies for Switzerland, the Netherlands and Austria highlight that developing countries incur both costs and benefits when signing DTTs. The results of our econometric analysis moreover suggest that developing countries attract additional FDI projects in the years following the signature of a DTT.
The paper proceeds as follows. Section 2 explicates Austria's international tax treaty policy, and particularly highlights the goals of the Austrian policies. In this context, the specific provisions regarding the allocation of taxing rights in Austrian DTTs and their potential effects on developing countries are discussed. Subsequently, Section 3 more generally studies the relationship between DTTs and FDI. After an investigation as to why attracting FDI inflows is considered so important by many developing countries, the existing literature on the relationship between DTTs and FDI is presented. In Section 4, we briefly describe Austrian investment in developing economies and then analyze econometrically to what degree DTTs contribute to encourage Austrian FDI projects in these countries. Section 5 concludes and proposes policy options.
2. AUSTRIA'S INTERNATIONAL TAX TREATY POLICY
2.1 GOALS OF AUSTRIA'S TAX TREATY POLICY
Austria has a large DTT network, which as of July 2014 consists of 86 DTTs, 37 of which are with developing countries. (2) Whereas formerly Austrian DTT negotiators primarily aimed at boosting tax revenues for Austria, increasing the attractiveness of Austria as a business location is now seen as the main function of DTTs (Loukota, Seitz & Toifl, 2004; Lang, 2012). In order to ensure a uniform international tax policy in its DTT network, Austria has established a DTT Model that is very close to the OECD Model. With its DTTs, Austria pursues four goals, namely to: (i) prevent international double taxation, (ii) foster bilateral economic relations, (iii) increase legal certainty, and (iv) prevent international tax avoidance and evasion (Loukota, Seitz & Toifl, 2004).
First, from Austria's perspective, the main purpose of DTTs is to avoid international double taxation (Jirousek, 2013 a). Austria prefers to apply the exemption method as a mechanism to avoid double taxation. (3) Under the exemption method, a "residence country" (i.e., a country where a company or an individual is considered to be a tax resident) is obliged to exclude income arising abroad (the "source country") from the taxable base to determine the tax due. Austria's domestic tax law provides for double taxation relief that is fairly similar to the relief provided under its DTTs (Section 48 of the BAO). The exemption method under Austria's domestic law applies to active income, such as income derived from businesses carried out through a permanent establishment (PE) situated abroad, which is subject to tax of at least 15%. (4)
As is standard with most exemption countries, Austria applies the credit method to passive income, i.e. dividends, interest and royalties. The credit method requires that a residence country first computes tax due on its residents' worldwide income, and subsequently this amount is reduced by taxes previously paid in a source country. However, with no obvious differences between methods to avoid double taxation under Austria's DTTs and its domestic tax law, signing a DTT seems not to be necessary for Austrian tax residents wishing to avoid international double taxation (Loukota, Seitz & Toifl, 2004).
For Austria, a second purpose of DTTs is to foster economic relations. In order to support expansion of its domestic firms, it is crucial from Austria's perspective to negotiate DTTs that reduce source taxation on passive income as much as possible, even below the standards embodied in the OECD Model (see Section 2.2). Usually, a source country is granted the primary (albeit reduced) taxation right, except for royalties under Article 12 OECD Model, and a residence country taxes the remaining amount.
The third goal of Austrian DTTs is to provide legal certainty. DTTs set common rules applicable in both a residence and a source country, and thus provide legal certainty for investors and tax administrations. (5) From the perspective of a residence country, legal certainty is crucial in protecting its residents investing abroad from international tax conflicts, which may give rise to unsolved double taxation. In order to provide increased certainty, Austria tries to ensure that DTT provisions are interpreted in the same way in both the residence and the source country. Austria insists on including a provision in the DTT protocol stating that DTT provisions should be interpreted according to the OECD Commentaries, which are revised periodically (Jirousek, 2013b). 15 of Austria's DTTs with developing countries include such a provision in the protocols; seven of these provisions refer to both the OECD and the UN commentaries. Austria thereby ensures that the latest version of the OECD Commentaries is legally binding and applicable for taxpayers, tax authorities, and even in the courts of signatory countries (Pistone, 2012).
Fourth, preventing international tax avoidance and evasion is a major concern for Austria, as for many other governments. Tax avoidance is not, per se, an illegal way to reduce taxes due, this term usually refers to "unacceptable" taxpayer behavior: although complying with the letter of the law, a taxpayer deliberately acts against the sprit or the intention of the law with the aim to reduce its tax liability. To prevent international tax avoidance, some countries prefer to include anti-avoidance provisions, such as subject-to-tax clauses, in their DTTs. Austria, however, prefers to apply anti-avoidance provisions in its domestic law, and not in its DTTs. Austria's argument is that specific anti-avoidance provisions in DTTs may stimulate creative tax planners to find ways to circumvent them and, therefore, these provisions will make it more difficult for tax authorities to argue that certain applications of the DTT are abusive (Loukota, Seitz & Toifl 2004). However, if requested by the negotiating partner, Austria agrees on inserting subject-to-tax clauses in the DTT (Burgstaller & Schilcher, 2004). Also in some of the DTTs with developing countries such subject-to-tax clauses are inserted, e.g. in the DTTs with Malaysia (Art. 20(1)) and Mexico (Art. 22(4)). It is argued that such specific anti-abuse provisions would be especially important in DTTs with developing countries that may lack effective domestic anti-abuse legislation (Burgi & Meyer-Nandi, 2013).
The exchange of information provisions in DTTs are viewed as useful tools to prevent not only tax avoidance, but also tax evasion, which--in contrast to tax avoidance--is an illegal way of avoiding paying taxes. (6) The exchange of financial information is imperative for a country to effectively enforce taxation of income from foreign sources, for instance with respect to bank accounts in a given country held by tax residents of other countries (McGauran, 2013). Therefore, a mechanism for the exchange of information, which may also be an effective instrument to frighten off potential tax evaders, would be crucial for both developed and developing countries to ensure their tax revenues are collected (Alliance Sud 2012; OECD, 2012).
All Austrian DTTs (except for the one with Luxembourg) provide for the exchange of information concerning tax matters. For a long time, Austria had been rather reluctant to exchange information on a broad basis. Austria had major information exchange clauses mainly with OECD countries. The major clause, in line with OECD standards, obliges signatory countries to exchange relevant information for the application of both DTT provisions and enforcement of domestic laws regarding taxes of every kind (income tax, valued added taxes, etc.) (Lang, 2013). The DTTs with developing countries mostly include minor exchange clauses. A minor clause only allows exchanging information relevant for the application of DTT provisions (i.e. it does not cover exchange of information for enforcement of domestic laws or related to other taxes other than those covered by DTTs, i.e. income tax). Only three of the Austria's DTTs with developing countries are rated by the OECD as meeting the internationally accepted OECD standard on the exchange of information (OECD, n.d.). Apart from concerns about the security and privacy of the exchanged data, also concerns about the Austria's "competitiveness" were drivers of this policy (Loukota, Seitz & Toifl 2004, p. 369). (7)
Since 2009, Austria has, to some extent, been adapting its DTT network in line with the OECD standards. (8) Also, further adjustments have been made in Austria, such as applying its strict bank secrecy rules only to domestic situations. Still, the exchange of information with non-OECD countries occurs via an "on request" basis. This forces the requesting country to provide enough information to clearly identify the person under examination, limiting the power of this provision to only limited and specific cases (Jirousek, 2013b).
Tax Information Exchange Agreements (TIEAs) are alternative bilateral tax agreements enabling the exchange of information concerning tax affairs. Unlike DTTs, TIEAs however do not include provisions concerning the allocation of taxation rights and avoidance of double taxation. In this regard, Austria prefers to negotiate DTTs with an exchange of information clause rather than TIEAs. To date Austria has only six TIEAs; none of them are with developing countries. (9)
Austria has also signed the Multilateral Convention on Mutual Administrative Assistance in Tax Matters. This Convention serves as a legal instrument to address tax avoidance and evasion without a need to sign a bilateral tax agreement like a DTT or a TIEA. Compared with a TIEA, this multilateral agreement is broader in scope, as it provides additional tools to facilitate cooperation between tax administrations, such as joint audits between tax authorities of signatory countries or assistance in recovery of taxes.
In the field of administrative cooperation for the recovery of tax claims, Austria has been slower to adapt international standards. Article 27 of the OECD Model, which allows signatory countries to assist each other in executing tax revenue claims, is included in Austrian DTTs only when requested by the other signatory country, and when Austria assumes that the partner country respects laws concerning confidentiality, and uses such information exclusively for tax matters (Lang, 2012). Austria therefore prefers to include this particular provision in DTTs exclusively with OECD countries. Only three of the DTTs with developing countries provide for the assistance in tax matters (the treaties with Algeria, Mexico, and Turkey).
2.2 THE ALLOCATION OF TAXING RIGHTS IN AUSTRIA'S DTTs AND THE EFFECTS ON DEVELOPING COUNTRIES
In its DTTs, Austria tries to deviate as little as possible from the OECD Model. Austria's view is that a DTT in line with the OECD Model becomes a valuable and attractive instrument for promoting business and bilateral relations (Loukota, Seitz & Toifl, 2004). (10) However, for a developing country (typically a source country), such an agreement would mean shifting some of its taxation rights (acquired by means of its domestic tax legislation) to Austria, as DTTs patterned along the OECD model favor residence-based taxation. From a source country's perspective, this may well be justified if a DTT attracts new investment; thus the loss in tax revenues may be offset. Generally speaking, DTTs provide for different allocation rules with respect to the taxation of active business income, i.e. business profits, and to the taxation of passive income, i.e. royalties, dividends, and interest payments.
With respect to active business income, the DTT business profits provision (Article 7 of the OECD Model) stipulates that when a company resident in one state generates business profits in the other DTT partner country, the profits are only taxable in the first state. However, in the case where there is a PE (i.e. a substantial business presence through a fixed place of business or a dependent agent) in the other signatory country, this other country has the right to tax profits attributable to the PE.
The definition of a PE and the method of computing business profits attributable to the PE are crucial for the allocation of taxation rights between treaty partners. In both regards, Austria seeks to closely follow the OECD Model. Compared to the OECD Model, domestic laws--especially those of developing countries--often provide for a wider definition of what a PE is as well as a broader approach to allocate profits to a PE. Thus, the rather narrow definition of a PE in the OECD Model can lead to a situation where certain activities, which may be regarded as a PE under the domestic law of the source country, are not regarded as a PE according to the definition in the DTT.
As a result, the source country is granted less or no taxation rights. This conflict of interests in clearly reflected in the two Model Conventions. For instance, while the OECD Model establishes that a construction site must exist for 12 months for it to be deemed a PE, the UN Model requires only six months. Austria follows a middle path with 17 of its DTTs with developing countries requiring 12 months, three DTTs nine months, and 17 DTTs 6 months before a construction site is regarded as a PE for tax purposes.
A further delicate issue of DTTs with developing countries is the so-called "Service PE" provision. Such clause provides that a company is deemed to own a PE in a source country if a foreign company renders services through employees and/or other personnel in the source country. This may include, for instance, management fees paid by a subsidiary located in another country to headquarters located in Austria, and may apply to Special Purpose Entities located in Austria that manage subsidiaries belonging to the same business group. This provision, which is not part of the OECD Model but part of the UN Model, is included only in nine of Austria's 37 DTTs with developing countries. (11) In the absence of this Service PE provision in Austrian DTTs, services rendered by Austrian Special Purpose Entities would technically not constitute a PE in other countries. This may open the possibility of shifting taxable profits like management fees from a company located in a developing country to Austria; as management services would not create a PE, they would be treated as business profits to be exclusively taxed in the residence country.
According to the OECD Model the residence country has the right to tax passive income. While royalties are only taxable in the residence state, the source state is granted a limited right to tax dividends and interest payments. The OECD Model suggests a 5% or 15% withholding tax rate for dividends depending on the stake in the company, (12) and 10% rate for interest payments. Although Austria follows the OECD Model in most regards, Austria aims to reduce source taxation of dividends and interest payments even beyond the rates in the OECD Model. Such deviations are in line with Austria's domestic tax law and its goal to promote itself as an attractive business location.
First, with regard to royalties, Austria follows the OECD Model and attempts to negotiate zero taxation in the source state (Loukota, Seitz & Toifl, 2004). Nevertheless, when it comes to DTTs with developing countries, Austria is willing to accept taxation of royalties at the source. All but six of its DTTs with developing countries grant the source state a limited right to tax royalty payments. However, 21 of the analyzed DTTs establish a lower withholding tax rate for royalties than provided in domestic legislation (on average, the rate is reduced by 12 percentage points). Four DTTs provide for the same rate as national legislation, and 12 DTTs include so-called matching credits or tax sparing provisions (see below) and thereby provide for a fixed rate to be credited.
Besides withholding tax rates, the definition of royalties determines the allocation of taxing power between signatory states. Also in this respect, Austria strives to keep the definition of royalties in its DTTs as close as possible to the definition provided in the OECD Model. The practical implication of this policy is that if a DTT allows for the taxation of royalties in a source country, the narrow definition of the term "royalties", which often is narrower than the definition provided under domestic laws of developing countries, implies that some payments do not qualify as royalties but rather as business profits. In this case, the DTT provision regarding business profits patterned after the OECD Model stipulates exclusive taxation rights for the residence country.
Second, regarding dividends, Austria's domestic tax law includes the so called "international participation exemption law", according to which foreign source dividends received by Austrian companies are exempt from taxation. (13) This legislation is derived from the EU Parent Subsidiary Directive, which standardizes tax exemption for internal company dividends within the EU. However, even beyond the intended scope of the EU Directive, Austrian DTTs extends the participation exemption regime to include non-EU countries.
To avoid economic double taxation of company profits, Austria keeps source taxation on dividends as low as possible. The scope of the EU participation exemption requirement is further extended to a minimum shareholder participation of 10%, as stipulated in Austrian DTTs (Loukota, Seitz & Toifl, 2004).
All Austrian DTTs with developing countries allow the taxation of dividends at the source. Most DTTs, however, reduce source taxation compared to national legislation. Withholding tax rates for dividends paid to qualifying companies are in 16 cases below the rates established in national legislation, and in two cases taxation reaches zero percent. (14) On average, in the 37 treaties analyzed, the tax rates for dividends paid to non-qualifying companies and individuals amount to 13.2%, and for dividends paid to qualifying companies to 6.1%. These rates are somewhat below the rates suggested by the OECD Model. From Austria's perspective, this deviation from the OECD Model is a strategy to create an attractive legal environment for international investments. This is especially targeted at foreign companies that use Austria as an investment routing vehicle for further investment in third countries.
Third, interest payments to non-residents are, generally speaking, not subject to taxation under Austrian tax law. (15) Similar to dividend taxation, Austria negotiates its DTTs reducing source taxation on interest to as low as possible. In 17 of the analyzed DTTs, the withholding tax rates are lower than in domestic law, and in some cases taxation can reach zero percent (Loukota, Seitz & Toifl, 2004).
However, if zero percent taxation for passive income is not achieved, Austria tries to insert a so-called "most favored nation clause" in the protocol to the DTTs which ensures that the country partner does not grant third countries lower withholding tax rates (Jirousek, 2013 a; Hofbauer, 2005; Loukota, Seitz & Toifl, 2004). Six of Austria's DTTs with developing countries include most favored nation clauses in their protocols, providing for treatment not less favorable than other third treaty partners are granted under their respective DTTs. These provisions do not only refer to withholding rates (16), but also to the deductibility of royalty payments, (17) and to the method of eliminating double taxation. (18)
If a source country potentially grants lower interest taxation in a DTT with another country, then from Austria's perspective there could be a potential risk of treaty shopping for Austrian and other foreign companies working under their jurisdiction. These firms may be able to use DTTs with other countries to reroute FDI to developing countries, thus harming Austria's intent to establish itself as a business location for routing investment through to developing countries. While "most favored nation clauses" create a significant advantage for Austria as a competitive location for routing investment to third countries, these provisions also increase the complexity of DTTs, especially for developing countries with rapidly expanding DTT networks. In addition, from a developing country's perspective, such "most favored nation clauses" may be critical, as they make it more difficult to benefit from the withholding taxes that have been negotiated, in case lower withholding taxes are agreed on in future treaties with other countries (e.g. Burgi & Meyer-Nandi, 2013).
Finally, thirteen of Austria's DTT with developing countries include so called tax-sparing or matching credit provisions. (19) These clauses are regarded as tools of economic development that foster the flow of FDI to developing countries. A tax sparing provision establishes that taxes that would have been paid in the host country, but are spared from taxation due to special tax incentives are nonetheless credited to the tax debt in the investor's home country. A matching clause provides that a fixed tax rate that is agreed on is deemed to have been paid in the host country, and is thus to be credited against the tax debt in the home country--regardless of the actual tax rate applicable. Beyond the original intention, such clauses can, however, be misused to achieve double-non taxation (EU Commission, 2012). Such provisions are therefore viewed as problematic in the BEPS context (e.g. Burgi & Meyer-Nandi, 2013).
Given all this, it could be conjectured that in its DTTs, Austria: (i) disproportionally allocates taxation rights to the residence country, thus inducing a loss in revenue for developing countries, and (ii) limits a developing country's access to satisfactory equal exchanges of information according to OECD standards. Therefore, developing countries that sign a DTT with Austria can only hope that revenue sacrificed is offset with the attraction of new FDI that a DTT may bring. In the analysis to follow, we investigate from an economic perspective whether Austrian DTTs with developing countries actually trigger an increase in FDI in developing countries.
3. DTTs AND INWARD FDI IN DEVELOPING ECONOMIES
3.1 WHY DEVELOPING ECONOMIES OFTEN HOPE TO ATTRACT FDI BY SIGNING DTTs
For developing countries, attracting FDI inflows is a main--though not the only--motivation to sign DTTs. Although developing countries (typically in the position of a capital-importer and thus of a source country) may forego tax revenues when signing a DTT based on the OECD Model, the rationale behind this is to attract enough direct investment to offset immediate tax revenue losses. Evidently, taxation is only one of many factors determining the location choice of international firms. Yet, it is undoubtedly an important tool that policy makers have at their disposal. Thus, in this section is a focus on one question, namely, whether or not DTTs trigger a boost of FDI in developing countries.
The influx of FDI is often viewed as highly attractive to many countries, as FDI inflows can spur economic growth (OECD, 2002; UNCTAD, 2012). FDI can boost capital accumulation, create job opportunities, increase integration into the international economy and contribute to the formalization of the host economy by extending value chains. As a result, also tax revenues may rise (OECD, 2002). Moreover, affiliates of MNEs can enhance human capital in a host country and generate technological spill-overs to local businesses, such as knowhow regarding various new production or marketing techniques. Productivity of local firms can also increase (Goodspeed, Martinez-Vazquez & Zhang, 2011). FDI can thus be an integral part of a country's strategy to foster economic development (ibid). The theoretical basis for such positive effects is mainly provided by the "capital fundamentalism" approach, as well as the neoclassical and endogenous growth theories (Rostow, 1961; Solow, 1956; Romer, 1990; for an overview see e.g. Todaro & Smith, 2006).
On the other hand, FDI inflows can also trigger considerable downsides. FDI may create economic enclaves that are not connected with the local economy, crowd out domestic investment and/or curtail economic instability. Investments of foreign companies may contribute to environmental pollution and deterioration. MNEs may also be able to circumvent national regulations like those regarding certain labor laws (for an in-depth discussion see e.g. Navaretti & Venables, 2004). The dependence theory moreover emphasizes that FDI influx may contribute to perpetuating the economic and political dependence of developing countries ("the periphery") on developed countries ("the center"). As long as foreign affiliates located in the periphery are constrained to supplying developed countries with natural resources and inexpensive labor, while decision making functions remain in the headquarters in developed countries, the presence of MNEs in the periphery contributes
to sustaining political and economic dependence (Todaro & Smith, 2006).
To what degree potential benefits of FDI materialize largely depends on local, political or institutional factors, as well as on the absorptive capacities of a host economy (Crespo & Fontoura, 2007; OECD, 2002). When a host country has a certain level of technological knowhow, of human capital stock, and when infrastructure, including financial markets, are developed to a certain degree, it is more likely to reap the benefits from FDI inflows (Hermes & Lensink, 2003; Borensztein, De Gregorio & Lee 1998; Crespo & Fontoura, 2007). Generally, middle-income countries are thus found to benefit more from incoming FDI than low-income countries (Blomstrom, Kokko & Zejan, 1994; Narula & Zanfei, 2005).
Clearly, a range of political and economic factors determines a country's attractiveness for FDI. Amongst others, geographical location, political stability, infrastructure, the size of the host market, labor cost, quality of the host country's institutions, and red tape in potential host countries all pla y a role in a MNE's decision where to set up a foreign affiliate (e.g. World Economic Forum, 2013; OECD, 2002). Business surveys and econometric analyses also show that in addition to these determinants, tax factors--including the presence of double tax treaties--impact the location choice of MNEs. From a policy perspective, DTTs are very appealing as an instrument to attract investment, as they can be implemented rather quickly in comparison to changing other factors such as the skill level of workers, which take a long time to show positive results.
3.2 PREVIOUS LITERATURE ON THE EFFECTS OF DTTs ON FDI ACTIVITY
A priori, it is not clear whether and how DTTs impact FDI activity. On the one hand, DTTs may have a positive effect on FDI. The relief from double taxation and the reduction of withholding tax rates on passive income provided for in DTTs may encourage FDI (Barthel, Busse, Krever & Neumayer, 2010; Blonigen, Oldensky & Sly, 2014; Egger, Larch, Pfaffermeyer & Winner, 2006; Lang & Owens, 2013). Moreover, developing countries entering into these agreements signal to the international
community a spirit of openness and willingness to adopt internationally accepted tax standards (Dagan, 2000).
However, studies show that a comprehensive domestic legislation that provides an overall transparent, non-discriminatory and predictable tax environment may be more important for foreign investors than a DTT alone (Pickering, 2013). In fact, a clear relationship between domestic law and DTTs is important for an easier application of DTT provisions. This concerns: (i) terms that are not explicitly defined in DTTs, (ii) procedures that apply the DTT provisions, such as mechanisms to withhold taxes on passive income and methods to avoid double taxation, (iii) procedures to exchange tax information (Nakayama, 2011).
On the other hand, it is argued that DTTs may hamper FDI flows through their prevention of tax avoidance and evasion. Three arguments are generally brought forward. Firstly, DTTs allow for the exchange of information and administrative assistance between tax authorities, thus reducing tax evasion and avoidance (e.g. Barthel et al., 2010). Secondly, DTTs include a provision which establishes that transactions between associated enterprises to be valued according to the arms' length principle, i.e. to value these transactions as if they took place between unrelated parties (Art. 9 OECD Model). This rule is argued to prevent profit shifting based on the strategic underpricing or overpricing of intra-company transactions (e.g. Egger et al., 2006; Blonigen et al., 2014). Thirdly, specific anti-avoidance provisions in DTTs aim to prevent treaty shopping, i.e. the (ab)use of DTTs by persons that actually are not entitled to benefit from the treaty and may thus hamper FDI.
However, the negative effects of these provisions on FDI are put into question by some authors (e.g. Baker, 2014). First, the exchange of information in DTTs is mostly on an on-request basis. Besides, as FDI mostly takes the form of wholly or majority-owned subsidiaries (IMF, 1993), tax authorities of host countries are unlikely to have additional information that residence countries do not have access to (Baker, 2014). Second, with respect to Article 9 OECD Model, this article cannot serve as "an independent legal basis for the adjustment of profits between affiliated companies" (Lang, 2013, p. 146). As such an adjustment can only be based on domestic law, it is unlikely that the treaty provision prevents tax avoidance and thereby impacts FDI. Third, anti-abuse provisions in DTTs aim to prevent persons that are not entitled to the treaty to benefit from it. Thus, by nature, such provisions should not hamper FDI flows between two signatory states (also Baker, 2014).
Taken together, it is not clear which effect we would expect DTTs to have on FDI flows, and it essentially remains an empirical question as to whether or not DTTs help to attract FDI. The economic literature investigating how far DTTs have an impact on FDI has produced mixed results. Some authors find that DTTs promote higher FDI activity (e.g. Barthel et al., 2010; Davies, Norback & Tekin-Koru, 2010; Neumayer, 2007). Other studies find no or negative effects of DTTs on FDI (e.g. Baker, 2014; Egger et al., 2006; Davies et al., 2010; Louie & Rousslang, 2008; Millimet & Kumas, 2008). While some authors like Baker (2014) argue that DTTs simply do not impact FDI decisions, others like Coupe, Orlova and Skiba (2009) or Blonigen et al. (2014) attribute inconclusive findings to the conflicting single provisions in the DTTs.
Blonigen, Oldensky and Sly (2014) try to disentangle the opposing effects of DTTs and find indirect evidence for these countervailing effects, i.e. a negative effect of the exchange of information and a positive effect of lower withholding tax rates. Using microdata on foreign subsidiaries of US MNEs, the authors conclude that it depends on the specific type of firm that is affected by the DTT as to which of these effects prevails.
Regarding the extant literature, two issues should be emphasized. The first relates to the type of investment decision that is analyzed, and the second to the sample of host countries that such analyses cover. First, a firm's international location choice essentially consists of two separate decisions. One, a firm decides whether or not to invest in a specific country, the so-called "extensive margin". Once this decision is made, a firm chooses how much capital to invest in a foreign affiliate, i.e. the firm decides on the "intensive margin of investment". Davies et al. (2010) and Egger and Merlo (2011) are, to our knowledge, the only studies explicitly analyzing the decision at the so-called "extensive margin". Using Swedish and German firm-level data respectively, both studies find that when a DTT is in place between two countries, there is a positive effect on the likelihood of a firm to establish an affiliate in a given host country. Both studies argue that this positive effect may be explained by the tax certainty that DTTs signal.
Second, the samples of most existing studies include both developed and developing countries as potential host countries. Yet, Blonigen and Wang (2005) claim that investment location decisions in developed and developing countries are likely to be determined by very different factors. Thus, the grouping of both types of countries in empirical analysis is considered to be problematic. In this analysis, only developing economies are included as potential host countries.
Existing studies focusing on the effects of DTTs on FDI in non-OECD countries as host countries do not produce clear-cut findings. On the one hand, Coupe et al. (2009) fail to find a consistent impact stemming from DTTs on FDI in transition economies, and Baker (2014) concludes that DTTs do not impact FDI location decisions in developing countries. On the other hand, Neumayer (2007) finds positive effects of DTTs on FDI in middle-income countries (but not in low income countries), and Barthel et al. (2010) find that DTTs encourage FDI in both middle- and low-income countries.
Analyses of individual countries' tax treaty policies can contribute to a better understanding as of how DTTs impact developing countries. There are two similar country case studies: McGauran (2013) analyses the Dutch tax treaty policy and Burgi and Mayer-Nandi (2013) investigate the Swiss DTTs with developing countries. Burgi and Mayer-Nandi (2013) present a qualitative legal analysis of the Swiss tax treaty policy with developing countries, and embed this in the broader context of international development cooperation. McGauran (2013) reviews the literature on the effects of DTTs on FDI and presents published company case studies. Her main contribution lies in the attempt to quantify the tax revenue losses that developing countries incur after signing a DTT with the Netherlands. We decided not to follow this approach as many of the underlying assumptions could be put into question.
Rather, by providing a case study of Austria that analyses the DTTs qualitatively and estimates the effects on FDI econometrically, we aim at adding a piece to the jigsaw puzzle in order to complete the picture of the ramifications of DTTs for developing countries. As the Austrian DTTs are based on the Austrian DTT Model and as they are also interpreted in a very similar way, we expect them to affect signatory states in very similar ways. The Austrian tax treaty network lends itself to a study to pin down the effects of which specific features of DTTs impact FDI flows. In this respect, our case study approach has an advantage over the studies analyzing the effects of DTTs on FDI in a cross-country context with different tax treaty policies. Moreover, our case study differs from the existing case studies that do not use this advantage to analyze the effects on FDI econometrically.
Based on the detailed analysis of the Austrian DTTs in Section 2 and the theoretical discussions above, we do not expect a hampering effect of DTTs, as international tax evasion and avoidance are unlikely to be prevented by Austrian DTTs with developing countries. The elimination of double taxation in DTTs should also not affect FDI significantly, as this elimination is similarly provided for in Austrian domestic law. Rather, the reduction in withholding tax rates may lead to increased FDI. Also the signaling function may play a role. However, while DTTs may lead to increased certainty, only the interplay of DTTs and domestic law can ensure legal certainty for investors. We therefore expect no effect or a positive effect of Austrian DTTs on FDI in developing countries.
4. ECONOMETRIC ANALYSIS OF THE EFFECTS OF DTTs ON AUSTRIAN OFDI IN DEVELOPING COUNTRIES
4.1 AUSTRIA'S OFDI IN DEVELOPING COUNTRIES
Starting from a rather low level, Austrian outward foreign direct investment (OFDI) has soared since 1995, the year of Austria's accession to the EU (see Figure 1). The opening up of the Eastern European markets has accelerated the growth of Austrian OFDI to such a degree that Austria was among the 20 largest foreign investors globally in 2008 (Bellak & Mayer, 2010). Since 2010, Austria's OFDI stocks have been exceeding its inward FDI stocks. Total OFDI stocks reached 167.3 billion Euros and accounted for about 51.9% of the Austrian GDP in 2013 (BMWFW, 2014). Also Austrian OFDI to developing countries has gained in importance, and, in 2011, 15.6% of all Austrian OFDI projects were located in developing countries (OeNB Special Statistical Evaluation).
Austria has also positioned itself as an attractive hub for businesses. About 300 foreign firms have established regional headquarters to serve the Central and Eastern European (CEE) markets and over 1,000 multinational enterprises (MNEs) coordinate their CEE activities from a base in Austria (ABA, n.d.).
Evidently, there are good economic reasons to invest in Austria, or to use Austria as a location for regional headquarters. Yet, Austria's tax policy may arguably also play a role in a company's decision to invest in Austria. Besides favorable features of the domestic tax system for MNEs (such as a generous group taxation regime and favorable taxation of outbound payments of dividends and interest), another essential factor of Austria's attractiveness as a business location is its large DTT network (Loukota, 1998). Of the Austrian OFDI in developing countries, about one third, both in terms of FDI projects and in terms of total capital invested, is attributed to foreign companies that invest in the respective developing countries via an Austrian subsidiary.
Geographically, Austrian OFDI in developing countries is primarily focused in Europe and Asia. In 2011, about 45% of all Austrian OFDI that flowed to developing countries was allocated in Europe (see Table 1). 37% of the Austrian OFDI projects were located in Asia (esp. China, Turkey and India), 12.4% in Latin America (esp. Brazil and Mexico), and 4.4% in Africa (mainly South Africa).
On the country level, in 2011, Austrian firms were active in 50 of the 143 countries that receive official development assistance (ODA). Nevertheless, Austrian FDI activity in developing countries is quite concentrated. 90% of all Austrian OFDI was invested in only 17 ODA-recipient countries. In 2011, the most important investment locations for Austrian firms among developing countries were Serbia, China, Ukraine, and Turkey (see Table 1).
4.2 DATA AND ESTIMATION METHOD
Austrian FDI activity in developing countries has increased significantly since 1990 and also the number of DTTs that Austria has signed with developing countries has risen (see Figure 2). (20) As of December 2013, 37 Austrian DTTs with developing countries were in place. (21)
For the following analysis of Austrian FDI projects in developing economies, the Austrian National Bank has kindly provided the FDI data on special request. From this data, a panel data set that covers 104 potential host countries over the period from 1990 to 2011 has been constructed. (22)
As DTTs may impact international investment at both the extensive and the intensive margin, both effects are studied. First, it is examined whether the existence of a DTT makes it more likely that an Austrian firm invests in a given host country. This effect of DTTs at the extensive margin of investments is analyzed in a logistic regression model. As dependent variable, a dummy variable is used that takes the values one or zero, indicating whether or not Austrian FDI exists in a specific host country.
Second, an analysis whether having a DTT with Austria leads to an increase in the number of Austrian FDI projects in a developing country is conducted. This can be interpreted as the intensive margin. (23) The number of FDI projects in a given country in a given year is represented by the dependent variable, and count data models are used. Due to confidentiality reasons, the amount in EUR of the individual investments is not available for research. Thus, regressions using the actual size of the investment in EUR are not possible.
The explanatory variable of main interest is a dummy variable, indicating whether or not Austria has a DTT in place with a specific partner country. In the economic literature two approaches are found: Some studies use the date when a DTT is signed, others use the date when a treaty becomes effective. The latter has been used here, as this is the date that is most relevant for international investors (also Barthel et al., 2010). As a robustness test, regressions are run with the date of signature, which effectively lead to the same result.
Depending on how the value chain of a company is split geographically, the literature distinguishes between two types of FDI: horizontal and vertical FDI (e.g. Navaretti & Venables, 2004). When a company transfers activities abroad, which are in the "same (horizontal) stage of the production process", this is known as horizontal FDI. Vertical FDI, on the other hand, refers to the international partition of activities along the value chain.
Horizontal FDI is assumed to be more likely in more alike countries. This idea is incorporated in the empirical framework through the similarity index, which indicates how similar a potential host country is to Austria in terms of GDP per capita (similarity). Alternatively, GDP per capita (gdppc) is also used. (24) Trade costs, captured by variables such as distance or trade barriers between two countries, are also seen as a major determinant of FDI. As the regressions include country-fixed effects that account for factors that do not vary over time, the geographical distance between Austria and respective host countries is not included. Rather, the general openness of a country to trade, defined as total exports plus imports divided by GDP, is used as a control variable to represent the general openness of a country (openness). It is expected that a country that is generally more open to international economic activity also attracts more FDI.
As this study is interested in the effects of international tax policy in general, and DTTs in particular, a measure of the corporate income tax rate is included in the analysis (corporate tax (lnj). Ideally, statutory and/or effective corporate tax rates of a host country as well as withholding tax rates on dividends, interest, and royalties that are paid to Austria could also be included in the analysis. All these tax rates potentially play a role in a firm's location decision; however, for a large number of developing countries withholding tax rates are not readily available and are difficult to compile. With this being considered, withholding tax rates have not been included in the empirical analysis. As corporate tax rates for many developing countries are also not available, Egger et al. (2006) and Baker (2014) have been followed, and general government final consumption expenditure as a percentage of GDP is used as a proxy for the corporate tax rate. (25) Higher taxes are expected to make a country less attractive for foreign investors. (26)
As a further control variable, the corruption index, made available by the Heritage Foundation (corruption) has been included. Empirical studies bring about mixed evidence as to whether corruption deters or encourages FDI. Egger and Winner (2005), for instance, find a positive relation between corruption in the host country and FDI. Wei (2000) and Egger and Winner (2006), on the other hand, find that higher levels of corruption deter FDI. Thus, it is not clear which sign to expect for the corruption variable in the regressions.
Additionally, the quality of the infrastructure of potential host countries is used as a control variable (infrastructure). The number of telephone lines per 100 persons serves as a proxy to measure the quality of a country's infrastructure. Countries with a better infrastructure are expected to also be more attractive for Austrian investors. Table 6 in the Annex provides descriptive statistics for the variables, and Table 7 in the Annex gives an overview of the sources of each variable used in our analysis.
Logistic and count data regression models based on maximum likelihood estimators are the methods used for the analysis. The logistic model is a binary response model with the dependent variable being a dummy variable. A prime candidate for count data models is the Poisson specification. However, this model requires equidispersion in the data, i.e. the mean of the dependent variable should be equal to its variance. As the data exhibits overdispersion, a Poisson specification has been rejected in favor of a negative binomial mod el. (27)
For both the logistic and the count data specifications, a fixed effects estimation including time and country dummy variables was implemented. Thereby, time trends and time-invariant country-specific effects (such as geographical distance or cultural and historical ties) are accounted for, which are not captured by our control variables (also Barthel et al., 2010).
A problem with such regressions, which is hard to alleviate, is endogeneity due to reverse causality or omitted variables. In order to mitigate this problem, all explanatory variables have been lagged by one period. In addition, the fixed effects estimation method is used in order to deal with potential endogeneity caused by omitted variables. Thereby, only within-variation in the data is taken into account and variation from across country-pairs is ignored (Blonigen & Davies, 2002).
In addition, by using fixed effects estimation, countries with which DTTs are in place already before the sample period starts (i.e. before 1989) do not impact the estimation results, as there is no within-variation in the DTT variable. As countries are likely to sign DTTs with countries with which they have had close economic ties at an earlier stage (and continue to do so), "older treaties are more likely to be correlated with unobserved variables and therefore [are more likely to be] endogenous" (Barthel et al., 2010, p. 373). Thus, excluding old treaties helps to alleviate the problem of endogeneity (Blonigen & Davies, 2004). Due to these problems, an upward bias for the DTT dummy in the regressions is expected.
4.3 ESTIMATION RESULTS
First, it is investigated whether having a DTT with Austria makes it more likely that a host country receives Austrian FDI. The first two columns in Table 2 show the results of these binary choice models. All regressions include time and country fixed effects and a constant. The sample in Column (1) includes 38 host countries and covers the years 1990-2011. The regression in Column (2), which additionally includes the corruption index of a host country as a control variable, covers fewer countries (30) and a smaller time-span (1996-2011) due to the availability of the corruption data. (28)
The main variable of interest, the dummy variable, stating whether there is a DTT in effect between Austria and a host country (DTT_e), is significant and positive in these logit regressions. This suggests that potential host countries, which have a DTT with Austria in place, are more likely to attract Austrian investment than those that do not. (29)
The control variables, which are all lagged by one year, show the expected signs. Higher taxes in a host country decrease the likelihood that a developing country receives FDI from Austria. (30) The positive and statistically significant coefficient of the similarity variable indicates that the more similar a potential host country is to Austria in terms of GDP per capita, the more likely it is that Austrian firms invest in that country. The quality of the infrastructure in the host country has the expected positive sign, but is only statistically significant in the regression covering a longer time period. Openness to trade of a country also has the expected positive sign, but is statistically not significant in the logit regressions. The corruption index, spanning from 0 to 100, where greater values indicate a lower level of corruption, is positive and statistically significant. This indicates that a lower level of corruption increases the likelihood that Austrian firms invest in a specific host country. The logistic estimation models thus suggest that having a DTT with Austria makes it more likely that a developing country receives Austrian FDI.
Second, it is investigated whether or not DTTs also impact the number of Austrian FDI projects in developing countries. Columns (3) and (4) of Table 2 present the regression results for the negative binomial model. The sample in Column (3) covers the years 1990 to 2011 and includes 104 countries. In Column (4), the sample also includes the corruption index and spans the period 1996-2011, covering 101 countries. (31) The regressions again include time and country fixed effects as well as a constant.
As in the logit regressions, the main variable of interest is whether or not there is a DTT in place. The count data regressions suggest that developing countries that have a DTT in place with Austria are expected to have a 33.7% or 25.2% increase in the number of Austrian FDI projects, depending on the model used (see Columns (3) and (4) respectively). Evaluated at the mean number of FDI projects, this implies that these developing countries are expected to have 0.8 additional FDI projects. This is a sizable effect.
The control variables in the count data models are similar to the ones in the logit regressions. A higher tax rate discourages Austrian investment. The similarity index and the openness of a country have a positive and statistically significant effect on the number of Austrian FDI projects in a country. The coefficient of the quality of the infrastructure variable is again positive and statistically significant in the larger sample (Column 3). The coefficient of the corruption variable indicates that lower levels of corruption cause a country to be more attractive for Austrian investors. The count data models thus suggest that developing countries with DTT attract more Austrian FDI projects than those without a DTT.
To analyze the timing of the effects, we additionally ran negative binomial regressions which included the leads and lags of the DTT dummy for the five years before and after the signature of the DTT. The sample for this analysis includes only the 26 developing countries with which Austria signed a DTT between 1990 and 2011. The regression results are shown in Table 8 in the Annex. It should be noted that the coefficients are not estimated very precisely due to the correlation among the lead and lag coefficients and the problematic inherent in some DTTs being signed in the years close to the beginning/end of the sample period (also Azemar & Dharmapala, 2015). Nevertheless, the analysis provides indicative insights into the temporal structure of the effects of DTTs on the number of FDI projects in signatory countries.
The pattern, which is also illustrated in Figure 3, suggests that the number of FDI projects increases from the year of the signature of a DTT onwards. Also, fewer investment projects seem to be realized in the two years before the DTT is signed. Albeit this effect is not statistically significant, it may be interpreted in a way that firms expect the signature of a DTT and postpone the realization of their investment projects. The picture however gives no indication that the signature of DTTs follows after an increase in the number of investment projects.
4.4 ROBUSTNESS TESTS
In order to check the robustness of the results, a number of alternative specifications were run. The date of signature of a DTT (DTT_s) was trialed instead of the date of effectiveness of a DTT (see Annex, Table 9). In place of the similarity index, data on the GDP per capita of the host countries (gdppc(ln)) was included (see Annex, Table 10). Moreover, the population (pop(ln)) of host countries was used as a proxy for the host country's potential market size (see Annex, Table 11). Against expectations, no positive and significant effect on FDI was found. The size of the host country population does not increase the number of Austrian FDI projects in a statistically significant way. A reason for this may be that relatively small countries like Serbia or Bosnia Herzegovina are among the countries that attract the most Austrian FDI.
For further robustness testing, the sample was restricted in three different ways (see Annex, Tables 12 and 13). First, CEE countries, (32) that historically attract a large part of Austrian OFDI and thus may bias our regressions results, were excluded from the regression (Columns (1) to (4) of Table 12, Annex). Second, B(R)IC countries, (33) which due to their market size and growth rates in the last decades have attracted a lot of FDI regardless of a DTT, were also left out (Column (5) of Table 12, Annex). Third, the sample of host countries includes ten jurisdictions, which are or have been listed as tax havens or offshore centers by the OECD and/or the Bank for International Settlements: Antigua and Barbuda, Belize, Costa Rica, Dominica, Grenada, Lebanon, Liberia, Mauritius, Panama and Uruguay. (34) In the robustness tests, these countries are excluded from the sample of potential host countries, as investing in these countries may arguably be motivated by other factors when compared to investing in "normal" developing countries (Annex, Table 13). All these alternative specifications confirm the results of our baseline regressions, that there is a positive relationship between DTTs and Austrian investment projects in developing countries.
The econometric analysis presented here suggests that DTTs significantly encourage Austrian FDI activity in developing countries. As Austria mainly has DTTs with middle-income countries (35) (except for Tajikistan and Nepal), our results are in line with Neumayer (2007), who also finds that the presence of DTTs triggers increased FDI in middle-income countries. Our results suggest that the number of Austrian investment projects in middle income countries increases by 25.2% to 33.7% when a DTT is in place.
However, these figures should be taken with some caution. First, even though the concerns of endogeneity in the analysis were mitigated, they are arguably not entirely solved. Thus, it is not totally clear, whether DTTs trigger more FDI or whether rather Austria signs DTTs with countries where Austrian firms are active. Also, the increases in FDI and DTTs may both be caused by another independent variable. As an upward bias due to these endogeneity problems can be expected, the actual effect may be smaller than the regression results suggest.
Second, the method used captures short term-effects only. In the long run, the impact of DTTs on FDI may be different. Indeed, the analysis of the timing of the effects on the number of FDI projects suggests a temporary positive effect during the first three years after the treaty becomes effective. Also Weyzig (2013) finds that DTTs only have a temporary positive effect.
It should also be considered that about a third of Austrian FDI projects in developing countries are effectuated by companies that have a parent in a third country. A number of international firms use their Austrian subsidiaries for investing in other countries, including developing countries. In the econometric analysis above, it is impossible to prove or discredit whether firms invest in a developing country via Austria are doing so simply because of a DTT between Austria and a respective developing country. That is, our results may capture some treaty shopping, which could also lead to an overestimation of the effect of DTTs on FDIs originating in Austria (see also Weyzig, 2013). (36)
This study adds to the literature regarding the effects of DTTs for developing countries by providing a case study for Austria. By offering a combined approach of legal and economic perspectives, this paper provides a general picture of the ramifications for developing countries, and shows that signing a DTT with Austria entails both potential benefits and risks for developing countries. This section summarizes the study and then draws some policy recommendations.
The econometric analysis suggests that signing a DTT with Austria encourages Austrian FDI activity in middle-income countries significantly. Due to potential endogeneity concerns, timing effects, and the rather large amount of "indirect" FDI coming from third countries to developing countries via Austria, the econometric approach may overestimate the effect of DTTs on FDI projects. Still, a higher amount of FDI projects indicates that entering into a DTT with Austria may potentially be beneficial for developing countries.
We moreover examine which aspects of a DTT trigger this increase in FDI. The increased legal certainty brought about by a DTT for potential investors makes signatory countries more attractive investment locations. Legal certainty is, however, only achieved through the interplay of a DTT with a comprehensive, stable and transparent domestic tax system. The legal analysis, moreover, suggests that the prevention of double taxation through DTTs is unlikely to be the main impetus for increased FDI activity, as Austria's domestic tax law contains provisions which allow the preventing of double taxation unilaterally.
Alternatively, reduced withholding tax rates on passive income, as compared with the domestic tax rates, may contribute to attracting FDI. Austria's policy goal is to reduce withholding tax rates as much as possible, even below the rates proposed by the OECD. Also, Austria's treaty policy includes proposing a "most favored nation clause", which may lead to further tax rate reductions in the future. Such reduced withholding tax rates, contrarily, are regarded as critical from a BEPS perspective. Withholding taxes in the source country can serve as an important backstop to prevent tax avoidance and evasion, particularly in developing countries with weak tax administrations (Weyzig, 2013; International Monetary Fund, 2011). Additionally, capital-importing countries risk losing tax revenues if the increased FDI inflows are not large enough to make up for lost revenue.
Moreover, DTTs with Austria generally cannot serve as an appropriate instrument to prevent tax avoidance, as Austria's policy is not to include anti-avoidance provisions in DTTs. Also when it comes to mitigating tax evasion, signing a DTT with Austria may not be the best option for developing countries. The exchange of information on request does not always alleviate the problem of tax evasion. For example, wealthy individuals can hide their money in bank accounts in other jurisdictions. Even though this is a problem for both developed and developing countries, developing countries that have weaker tax administrations may arguably suffer more from this form of tax evasion. (37) Information being exchanged only on request requires countries to provide information to identify the taxpayer under examination. Acknowledging that this is a complicated task, it could be argued that through the automatic exchange of information, one could curb the problem of tax evasion to a greater degree.
Also, developing countries should be aware of the alternatives to DTTs. These alternatives include the Multilateral Convention or TIEAs when the prime goal is to receive tax information or assistance in the field of tax collection. For developing countries, these tax agreements may even be more beneficial alternatives for exchanging information, as they do not shift taxation rights to the residence country. Moreover, the multilateral approach of the Mutual Assistance Convention is an advantage in the fight against tax avoidance and evasion, as it also provides, at least in theory, a "level playing field". Similar rules apply to all signatory states and information can be exchanged also with third countries.
Alternatively, developing countries may consider signing a "light treaty". Thuronyi (2010) proposes that it might be beneficial for developing countries to sign limited treaties that only comprise of the most important provisions. He suggests that such a treaty should provide for the exchange of information, administrative cooperation, non-discrimination, and mutual agreement procedures, and should also include a residence tie-breaker rule, i.e. a rule preventing dual residence for tax purposes. While still providing many benefits of DTTs, limited treaties would reduce the administrative burden for the signatory states, both when it comes to negotiating and administering the treaty.
In light of this assessment, the question arises as of how developing countries can reap the benefits of a DTT while mitigating its potential downsides. Foremost, any country should be very clear regarding the objectives it wants to achieve with a DTT and how, i.e. by the inclusion and design of specific provisions to achieve certain goals. Countries might find it useful to conduct DTT impact analyses in order to be able to estimate potential effects.
Such an analysis would be advisable for countries before deciding whether or not to enter into treaty negotiations. Besides, for developing countries that already have a DTT in place with Austria, an impact analysis may be helpful to find out whether the DTT actually serves their interests, or whether terminating or renegotiating and thus adapting certain provisions could be useful.
Also for Austria, the results of this study are relevant. The growing internationalization of the Austrian economy implies that its international tax policy also impacts other countries. Like all member states of the European Union, Austria has subscribed to "policy coherence". Austria commits to consider the goals and principles of its developmental policy in all policy areas that affect developing countries. (38) In the light of this "policy coherence" principle, Austria might, for example, find it appropriate to re-examine how its DTT policy, with regard to withholding tax rates, affects resource mobilization in developing countries.
Finally, there is ample room for further research. Most economic studies, including the present one, assume all tax treaties to be identical. However, even though they may be very similar in structure, each DTT is different. It is surprising that still little is known about how these different types of DTTs impact on FDI. In addition, further empirical evidence on how DTTs affect tax revenues of signatory states is needed. Corporate taxes, as well as withholding taxes, can be a significant source of revenue for developing countries. (39) Also, case studies analyzing the benefits and disadvantages of individual DTTs could be very insightful. However, not least because of the scarcity of available data, conducting such studies may be very challenging.
ZEW Center for European Economic Research
L7, 1, 68161 Mannheim, Germany
Daniel Fuentes *
DIBT (Doctoral Program in International Business Taxation)
Institute for Austrian and International Tax Law
WU Vienna University of Economics and Business
Welthandelsplatz 1, 1020 Vienna, Austria
ABA (n.d.) Headquarters--Standort Osterreich, Accessed July 2014 (http://aba.gv.at/DE/Home/Headquarters/Headquarters Standort+Osterreich.aspx).
Alliance Sud/Erklarung von Bern. (2012) Steuern und Entwicklung. Wie die Steuerflucht die Entwicklung behindert--und was die Schweiz daran andern kann. Bern: Alliance Sud/Zurich: Erklarung von Bern.
Azemar, C. and Dharmapala, D. (2015) "Tax Sparing Agreements, Territorial Tax Reforms, and Foreign Direct Investment", Accessed May 2016 (http://www.sbs.ox.ac.uk/sites/default/files/Business_Taxation/Events/confere nces/symposia/2015/dharmapala-paper.pdf).
Baker, P. (2014) "An Analysis of Double Tax Treaties and their Effect on Foreign Direct Investment", International Journal of the Economics of Business, 21(3):341-377.
Barthel, F., Busse, M., Krever, R., and E. Neumayer (2010) "The relationship between Double Tax Treaties and Foreign Direct Investment", in: M. Lang, P. Pistone, J. Schuch, C. Staringer, A. Storck, and M. Zagler (eds) Tax Treaties: Building Bridges between Law and Economics, pp. 3-18, Amsterdam: IBDF.
Bellak, C. (2001) "The Austrian Investment Development Path", Transnational Corporations, 10(2):107-134.
Bellak, C. and Mayer, S. (2010) "Outward FDI in Austria and its policy context", in: Sauvant et al. (eds) Columbia FDI Profiles, Accessed January 2014 (http://www.vcc.columbia.edu/files/vale/documents/Austria_OFDI_Final.pdf).
Blomstrom, M., Kokko, A., and M. Zejan (1994) "Host country competition, labour skills, and technology transfer by multinationals", Weltwirtschaftliches Archiv, 130(3):521-533.
Blonigen, B., Oldensky, L., and N. Sly (2014) "The Differential Effects of Bilateral Tax Treaties", American Economic Journal Economic Policy, 6(2): 1-18.
Blonigen, B., and Davies, R. (2002) "Do Bilateral Tax Treaties Promote Foreign Direct Investment?", NBER Working Paper No. 8834, Cambridge: National Bureau of Economic Research.
Blonigen, B., and Davies, R. (2004) "The Effects of Bilateral Tax Treaties on U.S. FDI Activity", International Tax and Public Finance, 11(5):601-622.
Blonigen, B. and Wang, M. (2005) "Inappropriate Pooling of Wealthy and Poor Countries in Empirical FDI Studies", in: T. Moran, E. Graham, and M. Blomstrom (eds) Does Foreign Direct Investment Promote Development?, pp. 221-244, Washington: Institute for International Economics/Center for Global Development.
BMWFW (2014). Aufienwirtschaft. Eine osterreichische Erfolgsgeschichte, Accessed July 2016 (http://www.bmwfw.gv.at/Aussenwirtschaft/oesterreichswirtschaftsbeziehung en/DatenundFakten/Documents/Oesterreich_Aussenwirtschaft_7_14.pdf).
Borensztein, E., De Gregorio, J, and J.-W. Lee (1998) "How does Foreign Direct Investment Affect Economic Growth?" NBER Working Paper No. 5057, Cambridge: National Bureau of Economic Research.
Braun, J. and Weichenrieder, A. (2015) "Does Exchange of Information between Tax Authorities Influence Multinationals' Use of Tax Havens?" ZEW Discussion Paper No. 15-015, Zentrum fur Europaische Wirtschaftsforschung.
Braun, J. and Zagler, M. (2014) "An Economic Perspective on Double Tax Treaties with(in) Developing Countries", World Tax Journal, 6(3):242-281.
Bundesministerium fur Finanzen (2014). Liste der osterreichischen Doppelbesteuerungsabkommen auf dem Gebiet der Steuern vom Einkommen und vom Vermogen, Accessed January 2014 (https://www.bmf.gv.at/steuern/intsteuerrecht/oesterreichische-doppelbesteuerungsabkommen.html).
Burgi, E. and Meyer-Nandi, S. (2013) Schweizer Doppelbesteuerungsabkommen: Aktuelle Politik und Entwicklungsrelevanz, World Trade Institute, Accessed July 2016 (http://www.alliancesud.ch/de/ep/finanzen/downloads/DBAStudie%20bei%20WTI.pdf).
Burgstaller, E. and Schilcher, M. (2004) "Subject-to-Tax Clauses in Tax Treaties. Austria", European Taxation, 44(6):266-276.
Coupe, T., Orlova, I., and A. Skiba (2009) "The Effect of Tax and Investment Treaties on Bilateral FDI Flows to Transition Countries", in: K. Sauvant and L. Sachs (eds) The Effect of Treaties on Foreign Direct Investment, pp. 171-224, New York (NY): Oxford University Press.
Crespo, N. and Fontoura, M. (2007) "Determinant Factors of FDI Spillovers--What Do We Really Know?", World Development, 35(3):410-425.
Dagan, T. (2000) "The Tax Treaty Myth", N.Y.U. Journal of International Law and Politics, 32(4):939-996.
Daurer, V. (2013) Tax Treaties and Developing Countries, Alphen aan den Rijn: Wolters Kluwer.
Daurer, V. and Krever, R. (2012) "Choosing between the UN and OECD tax policy models: An African case study", European University Institute Florence Working Papers, Robert Schuman Centre for Advanced Studies. Global Governance Programme 31, 2012/60.
Davies, R., Norback, P., and A. Tekin-Koru (2010) "The effect of tax treaties on multinational firms: new evidence from microdata", IFN Working Paper No. 833, Stockholm: Research Institute of Industrial Economics.
Dressler, D. (2012) "Form Follows Function? Evidence of Tax Savings by Multinational Holding Structures", Discussion Paper No. 12-057. Zentrum fur Europaische Wirtschaftsforschung.
Egger, P. and Merlo, V. (2011) "Statutory Corporate Tax Rates and Double-Taxation Treaties as Determinants of Multinational Firm Activity", FinanzArchiv/Public Finance Analysis, 67(2):145-170.
Egger, P. and Winner, H. (2005) "Evidence of Corruption as an Incentive for Foreign Direct Investment", European Journal of Political Economy, 21(4):932-952.
Egger, P. and Winner, H. (2006) "How Corruption Influences Foreign Direct Investment: A Panel Data Study", Economic Development and Cultural Change, 54(2):459-486.
Egger, P., Larch, M., Pfaffermayer, M., and H. Winner (2006) "The Impact of Endogenous Tax Treaties on Foreign Direct Investment: Theory and Evidence", Canadian Journal of Economics, 39(3):901-931.
EU Commission (2012) "The internal market: Factual examples of double non-taxation cases", Staff working paper, (TAXUD D1 D(2012)). Brussels.
Feenstra, R., Inklaar, R., and M. Timmer (2013) "The Next Generation of the Penn World Table", Accessed January 2014 (www.ggdc.net/pwt).
Feld, L., and Heckemeyer, J. (2011) "FDI and Taxation: A Meta-study", Journal of Economic Surveys, 25(2):233-272.
Goodspeed, T., Martinez-Vazquez, J., and L. Zhang (2011) "Public Policies and FDI Location: Differences between Developing and Developed Countries", Finanzarchiv/Public Finance Analysis, 67(2):171-191.
Hebous, S. (2014) "Money at the Docks of Tax Havens. A Guide", Finanzarchiv/Public Finance Analysis, 70(3):458-485.
Hermes, N. and Lensink, R. (2003) "Foreign Direct Investment, Financial Development and Economic Growth", Journal of Development Studies, 40(1):142-163.
Hofbauer, I. (2005) "Most-Favoured-Nation Clauses in Double Taxation Agreements: A Worldwide Overview", Intertax, 33(10):445-453.
Holler, H. (February 2014) "Von Kiews StraBen ins diskrete Wien", Zeit Online, Accessed March 2014 (http://www.zeit.de/2014/06/ukraineoesterreich-reichtum).
IBFD Tax Research Platform, Glossary, Accessed January 2014 (http://www.ibfd.org/IBFD-Tax-Portal).
IBFD Tax Research Platform, News Report on 17 January 2014, Austria--Ministry of Finance issues draft version of the Tax Law Amendment Act 2014 (www.ibfd.org).
IBFD Research Platform, Country Analyses, Austria, Corporate Taxation. International Monetary Fund (2011) Revenue Mobilization in Developing Countries. Fiscal Affairs Department (http://www.ibfd.org/IBFD-Tax-Portal).
Jirousek, H. (2013a) "Die osterreichische Position beim Abschluss von DBA", in: M. Lang, J. Schuch and C. Staringer (eds) Die Osterreichische DBA Politik. pp. 15-29, Vienna: Linde Verlag.
Jirousek, H. (2013b) "Exchange of Information and Cross-Border Cooperation Between Tax Authorities", Steuer und Wirtschaft International, 23(11):467-480.
Lang, M. (2013) Introduction to the Law of Double Taxation Conventions, Vienna: Linde Verlag.
Lang, M. (2012) "Reflections on Austrian Tax Treaty Policy", Steuer und WirtschaftInternational, 22(3):108-127.
Lang, M. and Owens, J. (2013) "The Role of Tax Treaties in Facilitating Development and Protecting the Tax Base", Daily Tax Report. Arlington: Bloomberg BNA.
Louie, H., and Rousslang, D.J. (2008) "Host country governance, tax treaties, and US direct investment abroad", International Tax and Public Finance, 15(3):256-273.
Loukota, H. (1998) "Stand und Perspektiven der osterreichischen DBA-Verhandlungen", OStZ, 1998(6):127-132.
Loukota, H., Seitz, W., and G. Toifl (2004) "Austria's Tax Treaty Policy", Bulletin for International Taxation, 58(8):364-371.
McGauran, K. (2013) Should the Netherlands sign tax treaties with developing countries? Amsterdam: SOMO.
Millimet, D. and Kumas, A. (2008) Reassessing the Effects of Bilateral Tax Treaties on US FDI Activity, Accessed July 2016 (http://faculty.smu.edu/millimet/pdf/tt.pdf).
Mintz, J. and Weichenrieder, A. (2008) "What Determines the Use of Holding Companies and Ownership Chains?" Working Paper Series No. 803. Oxford: Oxford University Centre for Business Taxation.
Mintz, J. and Weichenrieder, A. (2010) The Indirect Side of Direct Investment: Multinational Company Finance and Taxation. Cambridge, MA: MIT Press.
Nakayama, K. (2011) Tax Policy: Designing and Drafting a Domestic Law to Implement a Tax Treaty, Washington: International Monetary Fund.
Narula, R. and Zanfei, A. (2005) "Globalisation and Innovation: The Role of Multinational Enterprises", in: J. Fagerberg and D. Mowery (eds) The Oxford Handbook of Innovation, pp. 318-345, New York: Oxford University Press.
Navaretti, G. and Venables, A. (2004) Multinational Firms in the World Economy. Princeton: Princeton University Press.
Neumayer, E. and Spess, L. (2005) "Do bilateral investment treaties increase foreign direct investment to developing countries?", World Development, 33(10):1567-1585.
Neumayer, E. (2007) "Do double taxation treaties increase foreign direct investment to developing countries?", Journal of Development Studies, 43(8):1495-1513.
OECD (2002) Foreign Direct Investment for Development. Maximising Benefits, Minimizing Costs. Paris: OECD.
OECD (2010) Model Tax Convention on Income and on Capital, Condensed Version. Paris: OECD.
OECD (2012) Tax and Development: Draft Practical Guide on Exchange of Information for Developing Countries. Paris: OECD.
OECD (2013) Action Plan on Base Erosion and Profit Shifting. Paris: OECD.
OECD (n.d.) Exchange of Tax Information Portal, Accessed January 2016 (http://www.eoi-tax.org/jurisdictions/AT#default).
OECD (2014) Public Discussion Draft--BEPS Action 6: Preventing the Granting of Treaty Benefits in Inappropriate Circumstances, Paris: OECD.
OeNB (2012) Direktinvestitionen 2010. Osterreichische Direktinvestitionen im Ausland und auslandische Direktinvestitionen in Osterreich. Statistiken--Daten undAnalysen. Sonderheft. Vienna: Osterreichische Nationalbank.
OFSE, Oesterreichische Entwicklungspolitik (2012) Die Zukunft der Osterreichischen Entwicklungspolitik. Vienna: Sudwind Verlag, Accessed July 2016 (http://www.oefse.at/fileadmin/content/Downloads/Publikationen/Oepol/OEP OL2012.pdf).
Overesch, M. and Wamser, G. (2009) "Who Cares about Corporate Taxation? Asymmetric Tax Effects on Outbound FDI", The World Economy, 32(12):1657-1684.
Pickering, A. (2013) "Why Negotiate Tax Treaties?" Papers on Selected Topics in Negotiation of Tax Treaties for Developing Countries, No.1-N. New York and Geneva: United Nations.
Pistone, P. (2012) "General Report", in: M. Lang, P. Pistone, J. Schuch and C. Staringer (eds) The Impact of the OECD and UN Model Conventions on Bilateral Tax Treaties, pp. 1-32, Cambridge: Cambridge University Press.
Romer, P. (1990) "Endogenous Technological Change", Journal of Political Economy, 98(5):71-102.
Rostow, W. (1961) The Stages of Economic Growth: A Non-Communist Manifesto, 4th ed., London: Cambridge University Press.
Sauvant, K. (2009) The Effect of Treaties on Foreign Direct Investment: Bilateral Investment Treaties, Double Taxation Treaties, and Investment Flows. Oxford: University Press.
Skjonsberg, M. (2012 August 23) "The Austrian Banking Model, A Well-Kept Secret to Many", Wealth Briefing., Accessed March 2014 (http://www.wealthbriefing.com/html/article.php?id=44574&page=1)
Solow, R. (1956) "A Contribution to the Theory of Economic Growth", Quarterly Journal of Economics, 70(1):65-94.
Steiner, G. (2013) "Neuer OECD-Bericht "Addressing Base Erosion and Profit Shifting"--Ist Osterreich eine "Insel der Seligen" fur GroBkonzerne? (The New OECD Report "Addressing Base Erosion and Profit Shifting"--Is Austria an "Island of the Blessed" for Multinational Enterprises?)", Steuer und WirtschaftInternational, 23(11):150-158.
Thuronyi, V. (2010) "Tax Treaties and developing countries", in: M. Lang, P. Pistone, J. Schuch, C. Staringer, A. Storck, and M. Zagler (eds) Tax Treaties: Building Bridges between Law and Economics, pp. 441-455, Amsterdam: IBDF.
Todaro, M. and Smith, S. (2006) Economic Development, 9th ed., Boston: Pearson.
UN (2011) Model Double Taxation Convention between Developed and Developing Countries. New York and Geneva: United Nations.
UNCTAD (2012) World Investment Report. Towards a New Generation of Investment Policies. New York and Geneva: United Nations.
Wei, S. (2000) "How Taxing is Corruption on International Investors?", The Review of Economics and Statistics, 82(1):1-11.
Weyzig, F. (2012) "Tax Treaty Shopping: Structural determinants of foreign direct investment routed through the Netherlands", International Tax and Public Finance, 20(6):910-937.
Weyzig, F. (2013) Evaluation Issues in Financing for Development. Analysing Effects of Dutch Corporate Tax Policy on Developing Countries. IOB Study. Ministry of Foreign Affairs.
Wijnen, W. and J. de Goede. (2014) "The UN Model in Practice 1997-2013", Bulletin for International Taxation, 2014(3):118-146.
World Bank (2013) New Country Classifications, Accessed January 2014, (http://data.worldbank.org/news/new-country-classifications).
World Economic Forum (2013) Global Competitiveness Report. Geneva: World Economic Forum.
Wooldridge, J. (2010) Econometric Analysis of Cross Section and Panel Data, 2nd ed., Cambridge (MA): MIT Press.
Table 3. List of Effective Austrian DTTs with Developing Countries Albania (2009) Algeria (2007) rmenia (2005) Azerbaijan (2002) Belarus (2003) Belize (2004) Bosnia and Herzegovina (2012) Brazil (1977) China (1993) Cuba (2007) Egypt (1961) Georgia (2007) India (2002) Indonesia (1989) Iran, Islamic Republic of (2005) Kazakhstan (2007) Kyrgyzstan (2004) Macedonia (2008) Malaysia (1988) Mexico (2006) Moldova, Republic of (2006) Mongolia (2005) Morocco (2007) Nepal (2003) Pakistan (1968) Philippines (1983) Serbia, Republic of (2011) South Africa (1998) Tajikistan * (1979) Thailand (1987) Tunisia (1979) Turkey (1974) Turkmenistan * (1979) Ukraine (2000) Uzbekistan (2002) Venezuela (2008) Vietnam (2011) Note: Years in parentheses depict years when DTT becomes applicable; * old DTT with USSR applicable until new DTT signed (Tajikistan: new DTT signed in 2011, applicable as of 2013) Source: Bundesministerium fur Finanzen (2014). Table 4. Countries Included in Binary Choice Models Albania (a) Algeria Antigua and Barbuda (a) Armenia Azerbaijan Belarus Chile Costa Rica Cuba Ecuador Egypt Georgia Ghana Grenada (a) Guatemala Honduras Iran Jordan Kazakhstan Libya Macedonia (a) Malaysia (a) Mauritius Mexico Moldova Morocco Mozambique Namibia (a) Nicaragua Nigeria Pakistan (a) Panama Paraguay Peru (a) Tunisia Turkey Uzbekistan Vietnam Note: Countries marked with an (a) are not included in the corruption index Table 5. Countries Included in Count Data Models Afghanistan (a) Albania Algeria Angola Antigua and Barbuda (a) Argentina Armenia Azerbaijan Bangladesh Belarus Belize Benin Bhutan Bolivia Bosnia and Herzegovina Botswana Brazil Burkina Faso Burundi Cambodia Cameroon Central African Republic Chad Chile China Colombia Congo, Rep. of Costa Rica Cuba Djibouti Dominica (a) Dominican Republic Ecuador Egypt El Salvador Equatorial Guinea Eritrea Ethiopia Gabon Gambia Georgia Ghana Grenada (a) Guatemala Guinea Guinea-Bissau Guyana Haiti Honduras India Indonesia Iran Jamaica Jordan Kazakhstan Kenya Kyrgyz Republic Lebanon Lesotho Liberia Libya Macedonia Malawi Malaysia Mali Mauritania Mauritius Mexico Moldova Mongolia Montenegro Morocco Mozambique Namibia Nepal Nicaragua Niger Nigeria Pakistan Panama Papua New Guinea Paraguay Peru Philippines Rwanda Senegal Serbia Sierra Leone South Africa Sudan Suriname Syrian Arab Republic Tajikistan Thailand Tunisia Turkey Turkmenistan Uganda Ukraine Uzbekistan Venezuela Vietnam Zambia Zimbabwe Note: Countries marked with an (a) are not included in the models including the corruption index. Table 6. Summary Statistics Variable Mean Standard Deviation Minimum Maximum logit 1990-2011 (816 observations) FDI_d 0.44 0.50 0 1 DTT_e 0.24 0.43 0 1 corporate tax (ln) 2.60 0.39 1.58 3.77 similarity 0.09 0.07 0.004 0.42 infrastructure 11.60 9.69 0.12 49.32 openness 0.82 0.38 0.24 2.12 logit 1996-2011 (459 observations) FDI_d 0.47 0.50 0 1 DTT_e 0.27 0.45 0 1 corporate tax (ln) 2.56 0.37 1.61 3.69 similarity 0.08 0.06 0.01 0.31 infrastructure 12.24 8.82 0.32 43.13 openness 0.83 0.32 0.25 1.78 corruption 31.42 15.19 7 79 count data 1990-2011 (2,133 observations) FDI 2.42 9.82 0 126 DTT_e 0.20 0.40 0 1 corporate tax (ln) 2.60 0.45 0.72 4.24 similarity 0.06 0.07 0.004 0.48 infrastructure 7.84 8.84 0.02 49.32 openness 0.76 0.39 0.04 2.89 count data 1996-2011 (1,383 observations) FDI 3.18 11.19 0 126 DTT_e 0.26 0.44 0 1 corporate tax (ln) 2.55 0.42 0.72 3.75 similarity 0.06 0.06 0.004 0.48 infrastructure 8.52 8.37 0.06 43.13 openness 0.77 0.37 0.08 2.12 corruption 28.98 13.87 4 79 Table 7. Data Sources of the Variables Used in the Regression Analysis Variable Explanation Source FDI Number of Austrian Austrian National investments in a Bank (OeNB Special given country in a Statistical given year Evaluation) FDI_d Dummy of whether or Austrian National not there is an Bank (OeNB Special Austrian investment Statistical in a given country Evaluation) in a given year DTT_s Dummy equal to 1 in IBFD and Austrian the year a DTT is Ministry of Finance signed btw Austria and the respective partner country; also 1 in all subsequent years DTT_e Dummy equal to 1 in IBFD and Austrian the year a DTT btw Ministry of Finance Austria and the respective partner country becomes effective; also 1 in all subsequent years similarity "Similarity is an Own calculation; index, defined as based on UN GDP data one minus the ratio of the absolute value of GDP per capita minus GDP per capita in [Austria], relative to the higher of both GDPs per capita" (Overesch & Wamser, 2009: p. 1670). infrastructure Telephone lines (per World Bank, World 100 people) Development Indicators, available at http://data.worldbank.org/ data-catalog/world- development-indicators corruption Index ranging from 0 Heritage Foundation, to 100, where 0 avail. at: means very corrupt http://www.heritage.org/ and 100 very little index/explore?view=by- corrupt region-country-year gdppc (ln) GDP per capita United Nations Openness (Exports + Penn World Table 8.0 imports)/GDP (Feenstra et al., 2013) corporate Host country World Bank, World tax (ln) corporate tax rate; Development Indicators, proxied by log of available at general government http://data.worldbank.org/ final consumption data-catalog/world- expenditure as a development-indicators percentage of GDP Table 8. Timing of the effects of DTTs on the number of FDI projects DTT DTT signature effectiveness (1) (2) corporate tax (ln) -0.50 ** -0.54 ** (0.23) (0.23) Similarity 6.65 *** 6.38 *** (1.36) (1.22) infrastructure 0.02 ** -0.02 ** (0.01) (0.01) openness 1.07 ** 1.01 ** (0.45) (0.43) 5 years prior to DTT 0.02 -0.04 (0.19) (0.15) 4 years prior to DTT -0.2 -0.2 (0.21) (0.14) 3 years prior to DTT 0.05 -0.13 (0.19) (0.14) 2 years prior to DTT -0.15 -0.07 (0.18) (0.12) 1 year prior to DTT -0.25 -0.03 (0.17) (0.12) DTT -0.14 0.03 (0.18) (0.11) 1 year after DTT 0.02 0.15 (0.15) (0.13) 2 years after DTT 0.06 0.21 ** (0.15) (0.10) 3 years after DTT 0.14 0.28 *** (0.14) (0.10) 4 years after DTT 0.26 ** 0.13 (0.12) (0.11) 5 years after DTT 0.27 *** 0.02 (0.06) (0.11) period 1990-2011 1990-2011 observations 540 540 no. of countries 26 26 pseudo-[R.sup.2] 0.50 0.50 log-likelihood -607.12 -603.30 Notes: The dependent variable in columns (1) and (2) is a count variable indicating the number of Austrian FDI projects in a host country; columns denote coefficients; all control variables are lagged by one period and the natural logarithm of the corporate tax rate is taken; all regressions include year and country FE and a constant; robust standard errors (corrected for clustering at country level) are reported in parentheses; stars denote p-values: *** p<0.01; ** p<0.05; * p<0.1. Table 9. Robustness Test 1. Date of signature of DTT logit count data (1) (2) (3) (4) DTT_s 3.29 *** 4.11 *** 0.32 ** 0.26 * (0.84) (1.14) (0.15) (0.13) corporate tax (ln) -1.37 -2.21 ** -0.65 *** -0.52 (0.99) (1.09) (0.22) (0.51) similarity 20.50 *** 21.69 6.07 *** 4.70 ** (7.83) (17.36) (2.01) (2.22) infrastructure 0.08 -0.06 0.01 0.01 (0.09) (0.10) (0.01) (0.02) openness 0.44 -1.07 0.89 ** 0.77 (1.19) (2.45) (0.40) (0.47) corruption 0.05 ** 0.01 ** (0.02) (0.00) constant 14.54 *** 18.21 *** 3.14 *** 3.07 ** (3.85) (4.79) (0.76) (1.35) year FE yes yes yes yes country FE yes yes yes yes period 1990-2011 1996-2011 1990-2011 1996-2011 observations 816 459 2133 1383 no. of countries 30 38 104 101 pseudo-[R.sup.2] 0.47 0.48 0.53 0.55 log-likelihood -296.49 -165.62 -1371.58 -986.64 Notes: The dependent variable in columns (1) and (2) is a binary variable dependent variable indicating whether or not there is Austrian FDI in a host country; dependent variable in columns (3) and (4) is a count variable indicating the number of Austrian FDI projects in a host country; columns denote coefficients; all control variables are lagged by one period and the natural logarithm of the corporate tax rate is taken; robust standard errors (corrected for clustering at country level) are reported in parentheses; stars denote p-values: *** p<0.01; ** p<0.05; * p<0.1. Table 10. Robustness Test 2. GDP per capita logit count data (1) (2) (3) (4) DTT_e 2.96 *** 2.31 ** 0.31 *** 0.24 ** (0.84) (1.08) (0.12) (0.09) corporate tax (ln) -1.24 -1.88 * -0.53 ** -0.38 (0.94) (0.98) (0.24) (0.50) gdppc (ln) 1.27 * 1.87 0.61 *** 0.47 ** (0.73) (1.26) (0.18) (0.23) infrastructure 0.10 0.02 0.00 -0.00 (0.08) (0.11) (0.02) (0.02) openness 0.73 0.21 0.85** 0.68 (1.17) (2.06) (0.42) (0.48) corruption 0.06 *** 0.01 ** (0.02) (0.00) constant 4.02 2.34 -1.29 -0.48 (8.15) (11.75) (1.65) (1.65) year FE yes yes yes yes country FE yes yes yes yes period 1990-2011 1996-2011 1990-2011 1996-2011 observations 816 459 2133 1383 no. of countries 30 38 104 101 pseudo-[R.sup.2] 0.46 0.45 0.53 0.55 log-likelihood -303.94 -174.89 -1374.41 -989.58 Notes: The dependent variable in columns (1) and (2) is a binary variable indicating whether or not there is Austrian FDI in a host country; dependent variable in columns (3) and (4) is a count variable indicating the number of Austrian FDI projects in a host country; columns denote coefficients; all control variables are lagged by one period and the natural logarithm of the corporate tax rate is taken; robust standard errors (corrected for clustering at country level) are reported in parentheses; stars denote p-values: *** p<0.01; ** p<0.05; * p<0.1. Table 11. Robustness Test 3. Inclusion of population as control variable logit count data (1) (2) (3) (4) DTT_e 2.90 *** 1.96 * 0.22 ** 0.18 ** (0.94) (1.18) (0.11) (0.09) corporate tax (ln) -1.26 -1.75 * -0.56 ** -0.28 (0.90) (0.93) (0.25) (0.44) similarity 19.86 *** 30.36 5.61 *** 4.57 ** (7.39) (19.21) (2.12) (1.95) pop (ln) 0.22 -4.29 -2.12 ** -2.20 *** (4.64) (7.81) (0.87) (0.73) infrastructure 0.09 -0.00 0.00 -0.00 (0.09) (0.11) (0.01) (0.02) openness 0.79 0.22 0.85 ** 0.84 * (1.15) (2.23) (0.37) (0.45) corruption 0.05 ** 0.01 ** (0.02) (0.00) constant 9.68 93.13 41.76 *** 42.50 *** (84.71) (145.54) (15.90) (13.61) year FE yes yes yes yes country FE yes yes yes yes period 1990-2011 1996-2011 1990-2011 1996-2011 observations 816 459 2111 1377 no. of countries 30 38 103 100 pseudo-[R.sup.2] 0.46 0.45 0.53 0.55 log-likelihood -302.03 -173.41 -1359.31 -977.76 Notes: The dependent variable in columns (1) and (2) is a binary variable indicating whether or not there is Austrian FDI in a host country; dependent variable in columns (3) and (4) is a count variable indicating the number of Austrian FDI projects in a host country; columns denote coefficients; all control variables are lagged by one period and the natural logarithm of the corporate tax rate is taken; robust standard errors (corrected for clustering at country level) are reported in parentheses; stars denote p-values: *** p<0.01; ** p<0.05; * p<0.1. Table 12. Robustness Test 4. Exclusion of CEECs and the B(R)IC no CEECs logit count data (1) (2) (3) DTT_e 2.91 *** 2.28 * 0.37 *** (0.92) (1.27) (0.14) corporate tax (ln) -1.46 -1.82 * -0.56 ** (0.98) (0.96) (0.22) similarity 19.85 *** 32.51 * 5.73 ** (7.45) (18.59) (2.30) infrastructure 0.09 0.01 0.02 (0.08) (0.11) (0.01) openness 0.59 0.40 0.73 (1.19) (2.12) (0.45) corruption 0.05 *** (0.02) constant 14.46 *** 14.26 *** 2.78 *** (4.21) (4.99) (0.79) year FE yes yes yes country FE yes yes yes period 1990-2011 1996-2011 1990-2011 observations 773 459 2068 no. of countries 36 30 99 pseudo-[R.sup.2] 0.45 0.45 0.53 log-likelihood -289.88 -173.92 -1222.81 no CEECs no B(R)IC count data count data (4) (5) DTT_e 0.26 * 0.24 *** (0.14) (0.07) corporate tax (ln) -0.49 -0.55 *** (0.50) (0.15) similarity 4.50 ** 8.92 *** (2.26) (0.89) infrastructure 0.01 (0.02) openness 0.48 1.25 *** (0.53) (0.17) corruption 0.01 *** (0.00) constant 3.01 ** 2.68 *** (1.37) (0.47) year FE yes yes country FE yes yes period 1996-2011 1990-2011 observations 1345 2093 no. of countries 96 101 pseudo-[R.sup.2] 0.55 0.53 log-likelihood -896.91 -1203.02 Notes: The dependent variable in columns (1) and (2) is a binary variable indicating whether or not there is Austrian FDI in a host country; dependent variable in columns (3) to (5) is a count variable indicating the number of Austrian FDI projects in a host country; columns denote coefficients; all control variables are lagged by one period and the natural logarithm of the corporate tax rate is taken; robust standard errors (corrected for clustering at country level) are reported in parentheses, stars denote p-values: *** p<0.01; ** p<0.05; * p<0.1. Table 13. Robustness Test 5. Exclusion of tax haven countries logit count data (1) (2) (3) (4) DTT_e 3.35 *** 2.15 * 0.34 *** 0.25 *** (1.09) (1.26) (0.12) (0.09) corporate tax (ln) -1.06 -1.86 * -0.54 ** -0.44 (0.92) (1.03) (0.22) (0.45) similarity 26.72 *** 59.84 *** 6.11 *** 4.68 ** (9.59) (20.59) (2.19) (1.97) infrastructure 0.09 0.02 0.01 0.01 (0.12) (0.13) (0.01) (0.02) openness 3 47 *** 2.05 1.17 *** 0.64 (1.31) (2.49) (0.41) (0.49) corruption 0.06 ** 0.01 ** (0.03) (0.00) constant 10.71 ** 10.46 ** 2.64 *** 2.84 ** (4.35) (4.87) (0.72) (1.25) year FE yes yes yes yes country FE yes yes yes yes period 1990-2011 1996-2011 1990-2011 1996-2011 observations 706 415 1950 1304 no. of countries 33 27 95 94 pseudo-[R.sup.2] 0.49 0.49 0.54 0.55 log-likelihood -248.32 -146.86 -1263.21 -947.37 Notes: The dependent variable in columns (1) and (2) is a binary variable indicating whether or not there is Austrian FDI in a host country; dependent variable in columns (3) and (4) is a count variable indicating the number of Austrian FDI projects in a host country; columns denote coefficients; all control variables are lagged by one period and the natural logarithm of the corporate tax rate is taken; robust standard errors (corrected for clustering at country level) are reported in parentheses; stars denote p-values: *** p<0.01; ** p<0.05; * p<0.1.
* The authors wish to thank Christian Bellak, Veronika Daurer, Markus Leibrecht, Martina Neuwirth, Pasquale Pistone, Alexander Rust, Margit Schratzenstaller-Altzinger, Martin Zagler, and the participants of a Round Table at the Institute of International and Austrian Tax Law at the WU Wien for their valuable comments and feedback to earlier versions of this paper. Further, we would like to thank two anonymous referees for their helpful comments and suggestions. Finally, we wish to thank Rene Dell'mour from the Austrian National Bank for providing us with the data on Austrian Foreign Direct Investments. All remaining errors and inaccuracies are, of course, solely ours. The final draft of this document was closed on 31st August 2014. Developments after the referred date are not taken into account. Financial support from the Austrian Science Fund (FWF grant no. W 1235-G16) is gratefully acknowledged. An earlier version was published as "Double Tax Treaties between Austria and Developing Countries. A legal and economic analysis" by the VIDC (Vienna Institute for International Dialogue and Cooperation) (2014).
(1.) For instance, Mongolia has recently terminated its DTT with the Netherlands and in 2008 Argentina terminated its DTT with Austria. Rwanda and South Africa have renegotiated their DTTs with Mauritius, with the new treaties including for instance higher withholding tax rates on passive income.
(2.) For the purpose of this study, we define developing countries as countries that received Official Development Assistance (ODA) in 2012/13. The list of ODA-recipients is taken from OFSE (2012, p. 123).
(3.) See IBFD Tax Research Platform, Country Analysis, International Aspects: "The provisions on unilateral double taxation relief were issued on 17 December 2002 as a Decree of the Minister of Finance (Verordnung des Bundesministers fur Finanzen betreffend die Vermeidung von Doppelbesteuerungen, BGBl II 2002/474) on the basis of the authority given in section 48 of the Federal Fiscal Code (BAO). These provisions are effective for tax years ending in the calendar year 2002 and later tax years. Previously, double taxation relief could be obtained as a concession of the Minister of Finance (section 48 of the BAO)."
(4.) Not all DTTs, however, follow the exemption method. Austrian DTTs that follow the credit method are mostly with countries that used to be seen as tax havens, such as Bahrain, Barbados, or Belize (Lang, 2012).
(5.) Bilateral Investment Treaties (BIT) are another way to achieve legal certainty, however discussion on such treaties is beyond the scope of this paper. The interested reader can for instance refer to Neumayer and Spess (2005) or Sauvant (2009).
(6.) OECD (2012) provides examples of how the exchange of information between tax authorities may help to prevent tax avoidance and evasion.
(7.) According to Lang (2012), another reason for Austria's reluctance to offer major information clauses might be that offering "too much" administrative assistance, i.e., too much information to other tax authorities, may constitute a "competitive disadvantage" for Austria. For the fear that illegal earnings might be reported, there is anecdotal evidence that major orders were shifted from countries with major information clauses to other countries that do not have such a major information clause (Lang, 2012, p. 109).
(8.) To date, more than 20 DTTs have already been revised (Jirousek, 2014).
(9.) Austrian has concluded TIEAs with 6 countries. These are Andorra, Gibraltar, Jersey, Monaco, St. Vincent and the Grenadines and Guernsey (which is not yet in force to date). Further, Austria is negotiating TIEAs with Uruguay and Cayman Islands.
(10.) However, it should be noted that the OECD Model was never intended to be used by countries with asymmetric investment positions, such as it is typically the case between developing and industrialised countries. It was designed for the use by the OECD member states, i.e. countries with similar investment positions (e.g. Daurer & Krever, 2012).
(11.) The DTTs with Albania, Azerbaijan, China, Cuba, Indonesia, Kyrgyzstan, Mexico, Nepal, and Thailand include a "Service PE".
(12.) The OECD Model advocates a 5% withholding tax rate at the source for direct investments in the case where a shareholder holds at least 25% of the capital of the company paying dividends, and a 15% withholding tax rate for portfolio investment where a shareholder holds less than 25% of the capital of the company paying dividends (OECD Model (2010) Article 10 (2) a and b)).
(13.) See IBFD Research Platform, Country Analyses, Austria, Corporate Taxation, International Aspects, 57.
(14.) Only the DTTs with Turkmenistan and Macedonia provide for zero tax rates under certain circumstances.
(15.) See IBFD Research Platform, Country Analyses, Austria, Corporate Taxation, International Aspects, 65; IBFD Research Platform, News, Austria, Ministry of Finance issues draft version of the Tax Law Amendment Act 2014 on January 17, 2014.
(16.) The protocol to the DTT with Vietnam provides for the automatic reduction of withholding tax rates if such lower withholding rates are negotiated between Vietnam and another Member State of the European Union. The protocol to the DTT with Kazakhstan establishes that lower withholding tax rates with respect to Art 11(2) and 12(2) automatically apply if Kazakhstan agrees on such lower rates with other Member States of the Organization for Security and Cooperation in Europe. The protocols to the DTTs with Serbia and Kyrgyzstan provide that the two states will renegotiate withholding tax rates with Austria in the case that they sign a DTT with lower withholding tax rates with another Member State of the European Union.
(17.) According to Paragraph (3) of the protocol to the DTT with Brazil, Austria is automatically granted the same favorable conditions that might be negotiated with another country in this regard.
(18.) Cuba has agreed to renegotiate with Austria on the more favorable treatment given to any third country with respect to Article 24 of their DTT, i.e. the methods for elimination of double taxation.
(19.) The DTT with Malaysia includes a tax sparing provision, and the treaties with Brazil, China, Cuba, Indonesia, Mongolia, Nepal, Pakistan, Philippines, Thailand, Tunisia, Turkey, and Vietnam include a matching tax clause.
(20.) The data on FDI projects should include both subsidiaries and PEs. However, for practical difficulties of collecting the data on PEs, not all Austrian PEs abroad are recorded in the data.
(21.) For a list see Table 3 in the Annex.
(22.) Out of 104 countries, there are some countries that do not publish data available for all years.
(23.) This analysis is a bit of a hybrid, as it can arguably also be interpreted as an extensive-margin decision (Egger and Merlo, 2011).
(24.) Regressions were also run that focus on the dissimilarities between countries which drive vertical FDI. The percentage of persons enrolled in secondary schooling in a potential host country was compared to the Austrian enrolment ratio in secondary schooling (data from the World Development Indicators of the World Bank). It was tested whether a higher difference in secondary enrolment ratios encourages or discourages Austrian FDI in a potential host country. The dissimilarity variable has a negative effect on FDI in the logit regressions, indicating that countries that are more similar to Austria in terms of secondary schooling are more likely to receive Austrian FDI. The count data regressions, on the other hand, suggest that countries that are more dissimilar to Austria receive a larger number of Austrian FDI projects. Also in these regressions, the DTT variable has a positive and significant effect on Austrian FDI activity.
(25.) The regressions were also run using the statutory corporate tax rate of the host countries. For many specifications, the results remain unchanged; however the DTT-variable is not persistently significant. This is probably due to the smaller sample of countries, which excludes notably the CIS-countries, which are important FDI locations for Austria firms, and most African countries. In this smaller sample of countries, the proxy used and the statutory corporate income tax rate exhibit very similar results. Results are not presented here but are available on request; this includes tax data from Mintz and Weichenrieder (2010) and Braun and Weichenrieder (2015).
(26.) For a recent overview of the empirical evidence of the effect of taxation on FDI see Feld and Heckemeyer (2011).
(27.) In some count data regressions, the alpha-likelihood test indicates equidispersion in the data. In these cases, also the Poisson model was estimated. As the results do not change, it was decided for the sake of uniformity to use the negative binomial model in all count data regressions shown here. Due to the large number or zeros in our dependent variable, also a zero-inflated negative binomial model was estimated. The results brought about by the zero-inflated negative binomial model do not differ from the results of the negative binomial model.
(28.) The number of countries is so low because there are many countries with no variation in the FDI variable. See Table 4 in the Annex for the list of host countries included in the binary choice models.
(29.) It would also be desirable to quantify this effect. However, in logit regressions, only the sign, but not the magnitude of the covariates should be interpreted. For many types of logit regressions, marginal effects can be calculated, in order to measure the size of the effects. However, for the fixed-effects model, estimated with the maximum likelihood method that is implemented here, this is not possible (Wooldridge, 2010).
(30.) It was also tested whether the effect of a DTT depends on the level of corporate taxation in the host country, but evidence for this hypothesis was not found.
(31.) See Annex, Table 5, for the list of countries included in these regressions. In this sample, the number of countries is higher than in the logit regressions as the FDI variable in the count data models evidently exhibits more within-variation.
(32.) The CEE countries in the sample are Albania, Bosnia and Herzegovina, Macedonia, Montenegro, and Serbia.
(33.) Brazil, China, and India; Russia is not included in the sample of host countries.
(34.) Costa Rica and Uruguay were in the OECD "List of Jurisdictions That Have Not Committed to the Internationally Agreed Tax Standards". Antigua and Barbuda, Belize, Dominica, Grenada, Liberia, and Panama were included in the OECD "List of Jurisdictions That Have Committed to the Internationally Agreed Tax Standard, But Have Not Yet Substantially Implemented It". Lebanon and Mauritius are in the list of offshore centers published by the Bank of International Settlements (Hebous, 2014). Additionally, also Malaysia and the Philippines were in the OECD "List of Jurisdictions That Have Not Committed to the Internationally Agreed Tax Standards", but these two countries are large countries, which may attract investment also for economic reasons; thus it was decided to leave them in the sample.
(35.) As of July 2013, according to the World Bank (2013), middle-income countries are defined as having a Gross National Income per capita of between 1,036 USD and 12,615 USD.
(36.) There are studies that find empirical evidence that treaty shopping takes place (e.g. Mintz & Weichenrieder, 2008; DreBler, 2012; Weyzig, 2012). We are, however, not aware of any evidence that the Austrian DTT network is used for treaty shopping purposes. Lang (2012) sees a risk that the inclusion of provisions regarding the waiving of the withholding taxation of residents in Austria's tax treaties with countries that used to be considered as tax havens may provide incentives to shift profits out of Austria into third countries via these partner countries, where income is taxed at a very low rate or not at all. Steiner (2013) provides anecdotal evidence that Austria is being used as a conduit country for routing profits generated in a multinational's European affiliates overseas. According to Steiner, this routing via Austria takes place not because of the Austrian DTT network but because overall the Austrian tax system is so attractive for multinationals.
(37.) Anecdotal evidence suggests that such money is also in bank accounts in Austria, that Austria has a very stable political environment and it (used to) offer bank secrecy (e.g. Skjonsberg, 2012; Holler, 2014).
(38.) Austria has embraced the principle of policy coherence in its national law: [section]1 Zi 5 EZA-G "Der Bund berucksichtigt die Ziele und Prinzipien der Entwicklungspolitik bei den von ihm verfolgten Politikbereichen, welche die Entwicklungslander beruhren konnen." (The federation respects goals and principles of development policy in all political areas that may affect developing countries).
(39.) For instance, in 1997, withholding tax revenues made up 3% of GDP in Brazil, while corporate tax revenues excluding withholding taxes accounted for 3.7% (Weyzig, 2013).
Caption: Figure 1: Total amount of Austrian OFDI stocks in million EUR, 1989-2011
Caption: Figure 2: Austrian FDI and DTTs with Developing Countries, 1990-2011
Caption: Figure 3: Estimated Impact of DTTs on the Number of FDI Projects prior to, during and after the signature of DTTs
Caption: Figure 4: Estimated Impact of DTTs on the Number of FDI Projects prior to, during and after the date of effectiveness of DTTs
Table 1: Geographical Distribution of Austrian FDI Project s in 2011: Major Host Countries Region/Country Number of in Country projects percentages Rank Developing countries total 795 100% Europe 364 45.79% Serbia 126 15.85% 1 Ukraine 95 11.95% 3 Bosnia and Herzegovina 60 7.55% 5 Macedonia 28 3.52% 8 Albania 20 2.52% 9 Montenegro 19 2.39% 10 Belarus 11 1.38% 15 Asia 295 37.11% China 115 14.47% 2 Turkey 69 8.68% 4 India 39 4.91% 7 Thailand 15 1.89% 13 Malaysia 14 1.76% 14 Kazakhstan 9 1.13% 16 Africa 35 4.40% South Africa 18 2.23% 11 Tunisia 5 0.63% 17 Algeria 5 0.63% 17 Latin America 101 12.70% Brazil 47 5.91% 6 Mexico 16 2.01% 12 Chile 9 1.13% 16 Colombia 9 1.13% 16 Argentina 9 1.13% 16 Grenada 5 0.63% 17 Data source: OeNB, Special Statistical Evaluation, own calculations. Table 2. Baseline Regression Results logit count data (1) (2) (3) (4) DTT_e 2.88 *** 2.28 *** 0.34 *** 0.25 *** (0.83) (0.88) (0.07) (0.07) corporate tax (ln) -1.26 ** -1.82 ** -0.60 *** -0.43 ** (0.61) (0.87) (0.15) (0.21) similarity 19.85 *** 32.51 ** 5.86 *** 4.61 *** (6.56) (12.99) (0.83) (0.86) infrastructure 0.09 ** 0.01 0.01 *** 0.01 (0.04) (0.07) (0.01) (0.01) openness 0.79 0.40 0.82 *** 0.64 *** (0.70) (1.45) (0.19) (0.23) corruption 0.05 *** 0.01 *** (0.02) (0.00) constant 13.18 14.26 3.06 *** 2.83 *** (505.99) (716.70) (0.49) (0.67) year FE yes yes yes yes country FE yes yes yes yes period 1990-2011 1996-2011 1990-2011 1996-2011 observations 816 459 2133 1383 no. of countries 38 30 104 101 pseudo-[R.sup.2] 0.46 0.45 0.53 0.54 log-likelihood -302.04 -173.92 -1370.48 -985.41 Notes: In Columns (1) and (2) the dependent variable is a binary variable indicating whether or not there is Austrian FDI in a given country; in Columns (3) and (4) the dependent variable is a count variable indicating the number of Austrian FDI projects in a host country. Columns denote coefficients rather than odd ratios. All control variables are lagged by one period and the natural logarithm of the corporate tax rate is taken; robust standard errors (corrected for clustering at country level) are reported in parentheses. Stars denote p-values: *** p<0.01; ** p<0.05; * p<0.1.
|Printer friendly Cite/link Email Feedback|
|Author:||Braun, Julia; Fuentes, Daniel|
|Publication:||Public Finance and Management|
|Article Type:||Case study|
|Date:||Sep 22, 2016|
|Previous Article:||Merger myths: a functional analysis of scale economies in New South Wales local government.|
|Next Article:||Financial planning as an antecedent of fiscal transparency in local government: the mediating effect of civic participation.|