Does withholding tax on interest limit international profit-shifting by FDI?

Research background: Poland is a significant recipient of intercompany loans as a part of foreign direct investment (FDI) debt instruments reported in the Balance of Payments. Most of them come from the developed West European countries — Netherlands, Luxembourg, France, Germany, and Belgium. Igan et al. (2020) confirm debt-based FDI inflows to emerging markets had a higher impact on the industries’ growth in the pre-crisis period 1998–2007 than after (till 2010). Purpose of the article: We aim to identify withholding tax (WHT) impact on intercorporate loans inflow to Poland and analyse the relationship between trade credit and intercompany loans to assess the importance of the profit-shifting role of FDI after 2010. Methods: To reflect the impact of withholding tax and trade credit on inter-company loans (inward debt-based FDI) in 2011–2017 to Poland, we use Arellano-Bond and random effects panel model estimators. The estimated specification is derived from the knowledge-capital model and includes two types of capital: human and physical. Findings & value added: We show that WHT on interests reduces profit-shifting by multinational companies’ intercompany lending to Poland. But intercompany loans are positively related to foreign trade credit. Unlike in the case of total FDI inward to Poland (Cieślik, 2019), we identified that vertically integrated multinational enterprises are more likely to provide loans to Polish firms. This study is the first to confirm that withholding tax of interests reduces international profit-shifting by FDI and to provide evidence on the relation-ship between foreign trade credit and intercompany loans provided by multi-national companies. Equilibrium. Quarterly Journal of Economics and Economic Policy, 16(1), 11–44


Introduction
This paper aims to identify withholding tax impact on intercorporate loan inflow to Poland and analyse relationships between trade credit and intercompany loans to assess the significance of the profit-shifting role of foreign direct investment (FDI).
FDI is an integral part of cross-border capital flows and forms a part of the financial account in the country's balance of payments. FDIs are an instrument of the multinational enterprise' (MNE's) global activity based on equity or debt investments. Igan et al. (2020) analysed these data from 1998 to 2010, and we started our analysis in 2011. Igan et al. (2020) showed that debt-based FDI financing played a significant role in industries' growth in emerging markets, including Poland. Both intercompany loans and trade credit are components of debt-based FDI instruments. According to the National Bank of Poland data, Polish enterprises use these sources to finance their activity.
But along with trading, MNE found the scheme to shift the profits from affiliates in countries with higher taxes through intercompany loans, in which interests are deducted from taxable income (and allow to benefit from interest tax shield). This problem is widely announced in the academic and non-academic literature. Governments are looking for instruments to limit such unfair practices. Several initiatives at the EU level were made: Base Erosion and Profit Shifting by OECD (2015), Anti-Tax Avoidance Directive by European Commission, and an allowance for corporate equity (ACE) rule in European countries (European Committee, 2018). However, the problem of profit-shifting is still actual and unsolved (Kudła, 2018;Shimanski, 2018;Polish Economic Institute, 2020), as it is well seen in Poland's case (Polish Economic Institute, 2020).
Among different economic factors that influence the FDI inflow into the host country, a significant role is played by comparative advantages described in trade theories (Gudowski & Piasecki, 2020) and tax competition (Kudła et al., 2015). After 1989, a financial capital movement liberalisation led to implementation of the withholding tax (WHT) rate for foreign investments related to profit-shifting (royalties, dividends, and interests). This instrument's effectiveness is broadly discussed in the literature, but it gives a background for a new research question about its influence on the profitshifting function of FDI. Based on this, our paper aims to analyse the impact of WHT on profit-shifting by FDI. Poland is the chosen setting.
We analyse the determinants of intercompany loans as a component of debt inward FDI in 2011-2017 to Poland using Arellano-Bond and random effects panel model estimators. The estimated specification is derived from the knowledge-capital (KC) model. It includes two types of capital: human and physical, analysed by Cieślik (2019) for total FDI inflow to Poland from 15 EU members till 2015. We focus on the impact of WHT and trade credit on intercompany loans as debt instruments of FDI. We use data from the Central Bank of Poland (Narodowy Bank Polski, NBP) about FDI debt instruments (in position, not flows) which allow us to separate intercompany loans from trade credit and other debts. Therefore, our paper uses unique data and is the first to shed light on the role of intercompany loans in Poland. We use data on debt-based FDI inflow to Poland (available on the NBP's website) for the robustness check.
The remainder of this paper is structured as follows. Section 2 provides a literature review on foreign direct investments focusing on intercompany loans and trade credit, profit shifting, and WHT taxation. Next, we describe the data and research design in Part 3, show the empirical analysis results with interpretation in Section 4, and discuss them based on the previous selected studies in Part 5. The article is completed with conclusions.

FDI in the trade theory
There are two alternative dimensions of multinational activity that combine ownership and location advantages with technology and country characteristics in the new trade theory: horizontal and vertical. The former is directing to overcome distance and use lower costs of foreign markets access, while, according to the latter, -companies integrate vertically to decrease production costs. In theory, horizontally integrated multinational enterprises (MNEs) dominate when countries are similar in economic size, humane and physical capital factor proportions, and when trade costs are high (Faeth, 2009). Cieślik (2019) showed that the horizontal dimension explains FDI inflow to Poland from 15 EU members.
Analysis of integration processes globally shows that globalisation is still mainly exploiting relatively cheap labour resources in developing countries (Makhavikova, 2018). Intense competition forces companies to enter international markets to minimise their costs, especially in high-tech industries. Human resources play the most critical role in variable expenses rather than natural resources.
Based on the theory developed by Dunning (2000), neoclassical economists proposed several models of transnational companies' behaviour based on the idea of public welfare. Vernon (1966) proposed a "product life cycle model", explaining the increase in companies entering the world market and opening foreign branches. According to this theory, FDI flows begin to emerge when a product enters the "stage of maturity", when both the demand for it in emerging markets and the competition level increases (the monopolistic advantage of a new product disappears). Porter (1998) proposed a model of internationalisation considering the competitive advantages of countries-recipients of FDI. In line with the imperfect market theory and competitive advantages, Moosa (2002) presented three FDI hypotheses: the differential return rates, the diversification, and the output and market size. Large market size and low entry barriers strongly positively affect FDI inflow to Central Eastern European countries (Neuhaus, 2006).
Public policy and political conditions influence FDI because the government can create a suitable environment for competitiveness or decide to implement trade tariffs and barriers. Moosa (2002) underlined governmentinitiated FDI that can be triggered when a government offers incentives to foreign investors to eliminate a balance of payment deficit. The influence of public policy (sociopolitical factor) on FDI was studied, also focusing on the environmental effect (pollution reduction) (Dang, 2019).
Next, the institutional environment for multinational companies determines taxation and access to capital in the host country. Countries with high economic stability, often reflected by tight monetary and sustainable fiscal policies, are more attractive for FDI. For example, government restrictions have reduced the FDI inflow to Canada, negatively influencing employment and productivity (Hejazi & Trefler, 2019). Camarero et al. (2018) showed that monetary policy positively affects international trade and FDI by creating the currency union inside the EU. On the contrary, Duboz et al. (2019) proved that implementing uniform economic policies across sectors and places among EU countries does not attract FDI in service sectors. Jovanovic and Jovanovic (2018) analysed the influence of the business environment on FDI by using the Doing Business Indicator for 27 ex-socialist countries. They found a weak significance, except for trading across borders.
Trade credit is more frequently used than bank loans (i.e., Białek-Jaworska & Nehrebecka, 2016). According to the Balance of Payments' methodology, trade credit and advances arise when payment for goods or services (other than financial intermediation services measured indirectly or prepayment of insurance services) is not made simultaneously with the change in ownership of a good or provision of a service. If payment is made before the change of ownership, there is an advance (IMF, 2009). In 2011-2018 Polish companies were recipients of trade credit from foreign companies from Germany, Switzerland, France, Belgium, Italy, and the Netherlands (see Fig. 1 that ranks countries in descending order by volume of trade credit liabilities). The share of debt-based FDI financing in total FDI measured by positions (as liabilities at the end of the fiscal year) over 60 % is observed in Honk Kong, Ireland, Turkey, Slovak Rep., Czech Rep., Israel, and Belgium (see Fig. 2). There is redistribution of bank loans via the trade credit channel described in the literature (Guariglia & Mateut, 2006;Taketa & Udell, 2007) in case of supply limitation when banks restrict their credit policy. Polish firms have problems accessing bank loans. As a result, Poland has the lowest nonfinancial enterprises' debt to GDP ratio in the EU, next to Romania (Białek-Jaworska, 2018). It equals approximately 46%.

Profit-shifting via debt-based FDI
Instead of debt in banks, Polish companies use intercompany loans that are a constituent part of FDI debt instruments, particularly from the Netherlands, Luxembourg, France, Germany, and Belgium (see Fig. 1). Fig. 1 ranks countries in descending order by volume of payables due to intercompany loans at the end of the fiscal year. After the 2007-2008 global financial crisis, followed by bank lending constraints, the influence of such sources of alternative financing also increased for European small and medium-sized enterprises (Casey & O'Toole, 2014;Durovic, 2017).
In tax avoidance and tax planning literature, intercompany lending is an instrument for profit-shifting from high-tax to low-tax countries. This phenomenon is discussed in the literature review provided by Fonseca and Juca (2020). Borrowing from affiliates in business groups located in low-tax countries and lending to affiliates in high-tax locations will allow the latter to deduct interest payments from taxable income and save taxes (Mintz & Smart, 2004). It is essential to admit that the difference in tax rates between countries stimulates vertical integration due to transfer pricing for multinational companies (Egger & Seidel, 2013). However, there are different opinions about such a tax-planning strategy. As shown by Buttner and Wamser (2007), the tax-revenue effects for German multinationals were minor. The costs related to adjusting the capital structure for profit shifting purposes were substantial. Damgaard et al. (2019) pointed out that the possibility of tax payment reduction guides FDI, and parts of the investments are channelled to tax havens. Buettner et al. (2018) suggested that multinational companies may use foreign operations' financial structure to shift taxable profits. Intercompany loans may be used to create interest payments, which are deducted in coun-tries with high taxes (Fuest et al., 2011;Dischinger et al., 2014;Hansson & Olofsdotter, 2014;Devereux et al., 2018). In our opinion, in Poland's case, interest capitalisation also results in a debt increase. Multinational entities operating in a developed capital market optimise their tax payments by managing loans inside their business groups, from parent to subsidiary and among subsidiaries (Desai et al., 2004;Bénassy-Quéré et al., 2005;Azémar, 2010;Barrios et al., 2012;Egger et al., 2014;Hebous & Ruf, 2017).
Governments try to prevent debt financing by intercompany loans and limit the possibility for subsidiaries to transfer profit abroad through thincapitalisation rules with an ambivalent effect (positive in curbing tax planning and negative on investment (Merlo et al., 2020)). The introduction of an allowance for corporate equity (ACE) to achieve tax neutrality in some EU countries decreased debt financing. Still, it generated an increase in intra-group lending within multinational companies for tax planning purposes (Hebous & Ruf, 2017). Furthermore, empirical evidence shows that anti-profit-shifting legislation intensifies the adverse tax effects on FDI if imposed by host countries with relatively high tax rates (Buettner et al., 2018).
In Poland, the problem with profit-shifting is also widely discussed. In the report based on national accounts from 2018, Polish Economic Institute (PEI) shows that PLN 17 billion of corporate profit was transferred abroad, causing the corporate income tax (CIT) gap (Polish Economic Institute, 2020). Among the receiving countries were Ireland, the Netherlands, Luxemburg, Switzerland, Cyprus, Belgium, and Malta. But PEI analyses this problem only through the CIT gap's perspective, without taking WHT on interests of intercompany loans into consideration, which limited conclusions. Our analysis sheds light on this issue by focusing on the WHT. Fig. 3 illustrates the basis for the profit-shifting, that is, twelve countries with the highest amounts of Poland's liabilities due to intercompany borrowings: the Netherlands, Luxembourg, France, Germany, Belgium, Spain, Austria, the UK, the United States, Cyprus, Switzerland, and Ireland. Fig. 4 presents the volume of profits transferred abroad through interest payments in the case of those countries. The interest payments transferred in 2018 totalled over PLN 9.5 billion for the analysed countries.

Taxation of FDI
In the late 20th century, the liberalisation process in the international movement of capital in the EU and concerns about tax evasion led to bringing up an idea of a wide WHT on interest payment. According to double taxation conventions, the proposed minimum rate of WHT was 15%, with further diversification between WHT on interests and dividends (Table 1). The idea of tax heavens is broadly discussed in the literature. Previous articles such as Yaniv (1988), Huizinga and Nielsen (2000), and Huizinga et al. (2008) have studied the effect of taxes on international business capital movement. The rights to tax active income at the source are almost universally granted when such income is attributable to a permanent establishment of a foreign enterprise in the source country (Zee, 1998). Comparing two approaches to taxation: a residence-based one and a pure source-based one, Zee (1998) argued that the latter could require net capital importing (usually emerging) countries to give up their rights to tax a significant share of income generated from operations carried out within their national borders. Huizinga et al. (2008) suggested that source-level taxation affects leverage more significantly than the residence-level taxation levied by a multinational's parent country. Therefore they argue that corporate tax rates, rather than non-resident dividend WHT, appear to matter for leverage. However, they overlook WHT on interests paid for loans granted by MNEs, which is our paper's subject. Thus, our study covers this research gap and contributes to the literature.
Zee (1998) argues that capital flow in FDI is less prone to tax arbitrage considerations than portfolio investments because it deals with higher transaction costs, is more sensitive to nontax factors, and faces stricter regulatory requirements. Arena and Roper (2010) explain the influence of WHT rates on dividends and interest. MNEs issue debt through subsidiaries located in countries with high WHT on dividends. At the same time, MNEs issue larger debt in foreign countries with low or non-existing WHT on interests.
In contrast to previous studies, our paper checks whether WHT is a tool to reduce the profit-shifting function of FDI. We contribute to the state of the art of FDI by a separate analysis of trade credit and intercompany loans as components of FDI debt instruments and an extension of the knowledgecapital (KC) model analysed for Poland by Cieślik (2019). The usage of Kaufmann's indices, which measure the institutional environment, adds value to the literature.

Research methodology
We analyse factors that determine intercompany loans as a component of debt inward FDI in 2011-2017 to Poland. We use data received from the Central Bank of Poland (Narodowy Bank Polski, NBP) about FDI debt instruments (in position, not flows), which allow us to separate intercom-pany loans from trade credit and other debts. For the robustness check, we use data on debt-based FDI inflow to Poland in 2010-2018, available on the website of NBP. Previously, debt-based FDI data (a component of the Balance of Payments) was used in the research of Igan et al. (2020) to study the debt-based FDI inflow impact on the industries' growth in the emerging markets (including Poland) in 1998-2010. Table 2 presents the definitions of variables and data sources describing the material selection for our study. Polish liabilities due to intercompany loans are used as a dependent variable. Intercompany loans represent direct investment debt positions between affiliated enterprises (IMF, 2014). They are identified separately from other debt instruments for debt analysis because this lending has different implications for risk and vulnerability compared with debt between unrelated parties. In general, intercompany lending has three components: debt liabilities of parents to their affiliates, debt liabilities of affiliates to their parents, and debt liabilities between related affiliates. Based on the Balance of Payments' methodology, trade credit and advances rise when payment for goods or services (other than financial intermediation services indirectly measured and prepayment of insurance services) is not made simultaneously with the change in ownership of a good or provision of a service. If payment is made before the change of ownership, there is an advance (IMF, 2009).
We focus on the impact of WHT (on interests of foreign loans) and trade credit on intercompany loans as debt instruments of FDI. We suppose that:

H2: Foreign trade credit is positively related to intercompany loans provided by multinational companies to Poland.
The estimated specification is derived from the knowledge-capital (KC) model and includes two types of capital: human and physical. In this part of the model, we follow Cieślik (2019), but we use intercompany loans as a component of FDI debt instruments (in positions), contrary to FDI inflow to Poland (used in his paper). In the KC model, the explanatory variables are the similarity in economic size and relative physical and human capital factor endowments between Poland and the country-lenders. We measure the similarity in relative economic size (sdij) by a bilateral style of the Helpman's size dispersion index. This Helpman's index is calculated using data on output-side real GDP at chained purchasing power parity (PPP) rates and expressed in constant 2011 US dollars for Poland and the particular country that provides loans: where the variable is an i country's share of countries: lender's and borrower's GDPs calculated as described above.
The relative physical capital factor endowment between Poland and the country-lender (ln_kdiff) is proxied by the logarithm of physical capital (infrastructure) per worker difference between country-lender and countryborrower. The physical capital is calculated using the national capital stocks expressed in PPPs in constant 2011 USD and the number of workers employed. At the same time, we use the logarithm of the differences in human capital endowments calculated using the human capital indexes for the source and host countries to calculate relative human capital factor endowments between Poland and each country-lender (ln_hdiff). The human capital indexes are based on the average years of schooling and return to education.
These variables allow us to distinguish between horizontal and vertical theoretical models. By assessing the statistical significance and signs of the estimated coefficients on these variables, we identify which investment motive explains the cross-country pattern of intercompany loans as a part of debt-based FDI liabilities better. A positive coefficient at the similarity in pair countries' economic size supports horizontal integration and KC models (importance of the market access motive). In comparison, its insignificance confirms the vertical integration model. Negative coefficients at the differences in relative physical and human capital factor endowments favour the horizontal integration model. In contrast, the pure vertical and KC models assume a significant positive relationship (when the efficiencyseeking motive is of higher importance than the market access motive). We use the Penn World Table 9.1 (Freenstra et al., 2019) to measure the differences in relative factor endowments between Poland and countrieslenders using both human and physical capital per worker. We use the summation of Poland's and the home (lender) country's GDPs (ln_sum) as a control variable. For this variable, we expect that the coefficient at the combined market size of investment partners is positive, similar to theoretical predictions for total FDI. To control trade costs, we use geographical distance and expect a negative effect, as in Cieślik (2019) ) variables control institutional environment and governance. Financial freedom, trade freedom indicators, market capitalisation, and total FDI of home countries are control variables that catch differences in macroeconomic development. Because we focus on WHT on interests, we also control WHT on dividends on equity received from country-lender, amount of interests (potentially used for profit-shifting), dividends, and bonds as another component of FDI debt instruments. Table 2 describes in detail all test and control variables. Table 3 presents descriptive statistics for FDI data in positions (panel A) and inflows (panel B), while Table 4 provides their correlation matrix. Because of possible multicollinearity, we consider highly correlated control variables in separate models.
We estimate the relationships between intercompany loans and explanatory variables described above using a random-effects estimator for panel data. This approach allows us to include variables constant in time.
For robustness check, we first control for Kaufmann's governance indicators measured not only by estimates but also by ranks. Second, we use GMM Arellano-Bond dynamic panel-data estimator to focus on the impact of withholding tax and trade credit on intercompany loans. This approach uses instruments for the differenced equation but eliminates data constant in time. However, the generalised method of moment (GMM) approach unifies the estimators and eliminates the disadvantages of reduced sample size (Arellano & Bond, 1991). The dynamic panel-data approach allows a better understanding of the adjustment dynamics, especially when current behaviour depends on past behaviour. Third, for the robustness check, we control the explained variable's measure using data on debt instruments of FDI inflow to Poland instead of positions on unpaid liabilities due to intercompany loans. For this purpose, we use data publicly available on the website of the Central Bank of Poland. Table 5 presents results obtained from the estimations of models using random-effects estimator with WHT on dividends, bonds, amounts of interests, and dividends to foreign investors (incl. parent companies) used as control (2) variables. Our study's findings in Table 4 show a negative correlation between WHT on interests and intercompany loans borrowed by Polish firms, which is in line with H1. This outcome confirms that WHT on interests reduces profit-shifting by intercompany lending. Positive coefficients at the trade credit variable follow our expectations and support hypothesis H2. We show that intercompany lending increases with the combined market size in both lending and borrowing countries (ln_sum). The results indicate that the market access matters based on positive significant coefficients at the similarity in the economic size of pair-countries sdi variable. However, the efficiency-seeking motive is more important than the market access motive because of positive coefficients at the differences in relative factor endowments (ln_kdiff, ln_hdiff). Thus, our findings support the pure vertical and KC models in intercompany lending as a component of FDI. The estimated significant coefficients at the distance variable in columns (2) and (3) are negative. This finding supports the model of vertically integrated multinational firms to be more likely to provide intercompany loans. Regarding control variables, parameters aside bond and wht_div variables are insignificant, while intercompany loans increase with both amounts of interests and dividends paid. Table 6 provides results of model estimations where we control for governance and institutional environment. Considering Kaufmann's Governance Indicators (estimates) allows us to identify that countries with lower WHT on dividends grant higher loans (models presented in columns (3)-(6) that control for governance efficiency, regulatory quality, the rule of law, as well as control for corruption). Conversely, coefficients at the voice and accountability and political stability, and absence of violence variables appear insignificant when we use these governance indicators' estimates. For robustness checks, we use an alternative measure of Kaufmann's Governance Indicators in the form of ranks instead of estimates. Table 7 presents the outcomes of these models, which are stronger; all parameters at the Kaufmann's Governance Indicators measured by ranks are significantly and positively related to intercompany loans. These findings allow us to show that countries with higher governance and better institutional environment provide higher intercompany loans to Poland. Results of these models delivered in Tables 6 and 7 are in consistent with results presented in Table 5, which strongly confirms hypothesis H1 for the WHT on interests impact on reducing profit-shifting with intercompany loans. However, a positive relationship between trade credit and intercompany loans is weaker (insignificant in models presented in columns (4)-(6) in Table 6 and columns (3)-(6) in Table 7). Table 8 provides the results of model estimations with the use of the GMM Arellano-Bond dynamic panel-data estimator. This approach uses instruments for a differentiated equation. Our findings confirm the negative impact of withholding tax on intercompany loans, in line with the H1 hypothesis. The positive relationship between trade credit and intercompany loans (H2) is confirmed by the model (5) when we control market capitalisation. In the other models, such relation is visible for the lagged trade credit. When we control for financial freedom and trade freedom, the relationship between trade credit and borrowings from foreign MNEs is the least significant (models (2) and (3) in Table 8). Countries with higher financial freedom and trade freedom provide fewer loans to Poland, similar to host countries that benefit more from total FDI inflow. Conversely, MNEs from countries with higher market capitalisation and higher governance (measured by an arithmetic average of the Kaufmann's Worldwide Governance Indicators' ranks: regulator, voice, political, govefec, ruleoflaw and corruption) provide more loans to Polish firms.

Results
Negative coefficients at the differences in relative human capital factor endowments (ln_hdiff) favor the horizontal integration model of intercompany lending as a component of FDI. In contrast, a significant positive relationship between the differences in relative physical capital factor endowments (ln_kdiff) and the dependent variable supports the vertical integration between lenders and borrowers from Poland (Table 8). We can explain this by the different work-intensive and capital-intensive industries' debt financing practices. Besides, parameters at the combined market's size of both lender and borrower countries (ln_sum) seem insignificant. We overlook the sdi variable because its coefficient is insignificant as well.
Next, for the robustness check, we control for the alternative measure of the explained variable. For this purpose, we use data on debt-based FDI inflow to Poland instead of positions on unpaid liabilities due to intercompany loans. We provide results of these models' estimations with controls for Kaufmann's Governance Indicators, measured by both estimates and ranks, in Table 9. Despite the usage of data on debt-based FDI inflow to Poland, our results presented in Table 9 still confirm the significant negative relationships between WHT on interests and debt-based FDI inflow to Poland. These outcomes support hypothesis H1 and shed light on reducing profit-shifting with FDI by WHT on interests. Moreover, all Kaufmann's Governance Indicators, measured by both estimates and ranks, are significantly and positively related to debt-based FDI inflow to Poland.
The research problem is original not only in the case of Poland but also on the European level. Our results are a kind of novelty and contribute to the literature on the profit-shifting role of debt-based FDI instruments and FDI taxation. Our research also adds value to the discussion on the effects of debt-based FDI inflow to emerging markets opened by Igan et al. (2020).

Discussion
Contrary to the outcomes of the Cieślik (2019) study for the total FDI inward to Poland until 2015, we identify that vertically integrated MNEs are more likely to provide loans to Polish firms. Using the dynamic panel-data approach proposed by Arellano and Bond (1991), we find evidence to confirm the difference between work-intensive and capital-intensive industries in shaping intercompany debt relationships or choosing a lender from abroad. The horizontal integration motive of intercompany lending is more familiar among work-intensive industries, that is, service industries. In the case of physical capital, our results support the vertical integration of intercompany lending as a component of FDI. When we use a less statistically robust standard panel-data model, findings support the vertical integration motive.
Conversely, Cieślik (2019) pointed out the horizontal integration motives as the primary reason for total FDI inward to Poland. Unlike Jovanovic and Jovanovic (2018), we find that a better institutional environment of home countries has a significant positive impact on the debt-based FDI inflow to Poland as a host country. However, home countries with less financial freedom, less trade freedom, and lower FDI provide more intercompany loans to Poland. These results are achieved with the use of the Arellano-Bond estimator.
Our results support statements of Fuest et al. (2011), Dischinger et al. (2014, Hansson and Olofsdotter (2014), and Devereux et al. (2018) that MNEs optimise their tax payment by managing intercompany loans. However, we prove that governments can limit these practices by increasing WHT on interest income.
Simultaneously, we extend the discussion on the profit-shifting causing income tax gap in Poland conducted in a report of the Polish Economic Institute (2020) by adding the WHT on interest impact on reducing this tax avoidance practice. Furthermore, analysis of the unpaid liabilities due to foreign borrowings at the end of the fiscal year identifies a different list of top country-lenders -the Netherlands, Luxembourg, France, Germany, and Belgium compared to Polish Economic Institute (2020). Furthermore, we support the conclusion of Arena and Roper (2010) that MNEs issue more debt through subsidiaries located in countries that have lower with-holding taxes on interest income and higher withholding taxes on dividends with the state where the parent company resides.

Conclusions
The causes and role of FDI in the development of emerging markets, including the Polish economy, have been studied extensively in the literature. The previous empirical analyses show that in the investment decisions, MNEs are motivated by different advantages (market access motive, efficiency-seeking by vertical or horizontal integration of business groups) of home and host countries. However, FDI can also be used to profit-shifting and tax avoidance studied in depth from a different perspective by Kałdoński (2016). Therefore, in our paper, we focus on withholding tax to check if WHT reduces the profit-shifting role of debt-based FDI, and in particular intercorporate loans, inflow to Poland.
Our results can raise policymakers' attention as intercompany loans are used as a tool for profit-shifting between firms located in high taxes and low taxes countries. We show that to limit profit-shifting, the government can effectively use WHT on interests, not only on dividends. Our findings confirm that WHT on interests reduces profit-shifting by intercompany loans provided to Poland by MNEs.
Besides WHT, our paper analyses trade credit (as a debt-based FDI instrument) impact on intercompany loans inflow to Poland in 2011-2017. The estimated specification is derived from the knowledge-capital model and includes two types of capital: human and physical. Using the Arellano-Bond next to random effects panel model estimators, we find that foreign intercompany loans are positively related to and even caused by trade credits abroad. Unlike in the case of total FDI inward to Poland (Cieślik, 2019), we find that vertically integrated MNEs are more likely to provide loans to Polish firms. A better institutional environment of home countries has a significant positive impact on the debt-based FDI inflow to Poland as a host country. MNEs from home countries with higher market capitalisation and higher governance (measured by Kaufmann's Worldwide Governance Indicators) provide more loans to Polish firms. However, MNEs from home countries with more financial freedom, greater trade freedom, and higher FDI beneficiaries provide fewer intercompany loans to Poland.
Our paper contributes to the discussion on the profit-shifting causing income tax gap in Poland conducted by Polish Economic Institute (2020). Policymakers and governments can use the results to limit profit-shifting practices via intercompany lending by increasing WHT on interest income.
The research limitations deal with the availability of the Penn World data only until 2017. Future studies can focus on equity FDI and reinvestments of MNE's profits in Poland. The comparative analysis of the impact of WHT on dividends and interest income on the capital inflows to Poland and other emerging markets is still an open question. The study of multinational firms' investment decisions from selected industries comparing debtbased FDI to equity FDI inflow is also a possible future research direction. Kudła    Helpman's size dispersion index that is calculated using data on output-side real GDP at chained purchasing power parity (PPP) rates and expressed in constant 2011 US dollars for Poland and particular country that provides loans = 1 − − is country i's share of a group of country-lender's and Poland's GDPs calculated using data on output-side real GDP at chained purchasing power parity (PPP) rates and expressed in constant 2011 US dollars is country j's share of a group of country-lender's and Poland's GDPs calculated using data on output-side real GDP at chained purchasing power parity (PPP) rates and expressed in constant 2011 US dollars ln_kdiff the logarithm of capital per worker difference that is calculated using the national capital stocks expressed in PPPs in constant 2011 USD and the number of workers employed. ln_hdiff the logarithm of the differences in human capital endowments calculated using the human capital indexes for the source and host countries that are based on the average years of schooling and return to education ln_distance A proxy for trade costs is a logarithm of geographical distance of each parent country's (lender's) capital city from Warsaw that is measured "as the crow flies" distance between the capitals of the country-lender and the capital city of Poland  voice and accountability captures perceptions of the extent to which a country's citizens can participate in selecting their government, as well as freedom of expression, freedom of association, and a free media, measured by estimate (voicee) or rank (voicer) political political stability and absence of violence/terrorism measure perceptions of the likelihood of political instability and/or politically motivated violence, including terrorism, measured by estimate (politicale) or rank (politicalr). govefec government effectiveness captures perceptions of the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government's commitment to such policies, measured by estimate (govefece) or rank (govefecr). ruleoflaw the rule of law captures perceptions of the extent to which agents have confidence in and abide by the rules of society, and in particular the quality of contract enforcement, property rights, the police, and the courts, as well as the likelihood of crime and violence, measured by estimate (ruleoflawe) or rank (ruleoflawr). corruption control of corruption captures perceptions of the extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as "capture" of the state by elites and private interests, measured by estimate (corruptione) or rank (corruptionr).     Standard errors are given below the coefficients. Significance levels are *** p < 0.01, ** p < 0.05, and * p < 0.1. Standard errors are given below the coefficients. Significance levels are *** p < 0.01, ** p < 0.05, * p < 0.1, # p < 0.15, and ## p < 0.2. Table 9. Note: Standard errors are given below the coefficients. Significance levels are *** p < 0.01, ** p < 0.05, and * p < 0.1.