The Effects of Restrictive Measures on Cross-Border Investment: Evidence from OECD and Emerging Countries

This paper investigates the effects of global and sectoral restrictive measures on crossborder FDI among 49 advanced and emerging countries. We use a gravity model with panel data from 2010 to 2019 and the FDI Restrictiveness Index of the OECD that quantifies the level of restriction in FDI. Our results suggest that global restrictive measures do not significantly affect cross-border FDI in OECD countries, while restrictions in the service sector have negative and significant effects on FDI. Moreover, the overall restrictive measures and those in the service sector negatively impact inward FDI among OECD and big emerging countries. In addition, restrictions in the services sector impede inward FDI in African countries. Interestingly, restrictions in the secondary sector boost FDI between advanced and African countries, with larger effects for inward investments in African countries. The analysis of disaggregated sectoral restrictive measures shows that restrictions in business and other financial services are negatively associated with intra-OECD FDI, while restrictions in the banking sector have a significant positive impact on FDI. We also find that restrictions in the manufacturing sector have restrictive impacts on inwrad FDI in big emerging countries, and those in the mining, quarrying, and oil extraction sector hinder inward FDI in African countries. Reforms to liberalize sectoral restrictions by country have positive effects on FDI, but deregulation of the services sector has beneficial effects on inward FDI in all countries.


I. Introduction
Foreign direct investment (FDI) has become more and more important in economic growth and globalization in the last years. Indeed, FDI can increase productivity and change the comparative advantage of host countries. The establishment of multinational firms, capital accumulation, or delocalization can reduce unemployment, income inequality, poverty, improve technology transfer, and increase product variety through innovation (Hale and Xu, 2016;Dritsaki and Stiakakis, 2014). The empirical literature suggests a positive correlation between FDI and economic growth (Iamsiraroj and Doucouliagos, 2015). However, several factors determine the massive inflow of FDI into a country and its effectiveness in economic growth (Alfaro et al., 2004;Li and Liu, 2005; Batten and Vo, 2009; Desbordes and Vicard, 2009). A strong macroeconomic framework with competitive and effective policies is necessary and contributes to attracting more FDI into a country (tax rates, restrictions on financial transactions, legal framework, economic and political stability, etc.). Indeed, an investment environment that considers the local institutions, regulations and policies in which companies operate stimulates economic growth by encouraging firms to invest. In this case, regulation has an impact on job creation and sustainability (World Bank, 2005). The positive link between FDI and growth is stronger in open economies with an educated workforce and developed financial markets (Bodman and Le, 2013). Moreover, some analyses have showed the positive link between FDI inflows and low economic policy uncertainty (Gulen and Ion, 2015).
Studies showing the relationship between FDI and regulation suggested that the FDI sector is far less liberalized than trade in goods and services (Ghosh et al., 2012). Although regional, bilateral, and multilateral trade and investment agreements have reduced formal barriers, restrictions are still significant in some countries and affect FDI. The regulatory framework plays a key role in FDI flows. Indeed, regulation has a profound and durable impact on a firm's financial choices and is seen as a crucial driver of investment. To encourage FDI, authorities must reduce the costs, minimize the risks associated with investment, and create an appropriate climate for the domestic economy. Regulation must be both optimal and competitive, protecting investors from potential risks, promoting competition among firms across borders, and protecting consumers from higher prices.
In 2018, inward FDI represented about 2% of GDP in the EU and 1.5% of GDP in all OECD countries (see Figure 1). But we find that investments dropped by 3 to 1% of GDP between 2016 and 2018, which is contrary to the acceleration of GDP and trade growth. These trends are more significant in advanced countries than emerging countries (see Figures 2 and 3). In this year, some 55 economies introduced at least 112 measures affecting foreign investment. Two thirds of these measures sought to liberalise, promote and facilitate new investment (falling since 2016). Almost a third of these measures are new restrictions (increasing since 2016) (UNCTAD, 2019). How can we explain the decline in foreign investment? Is it a consequence of restrictive or sub-optimal policies? What is the real impact of restrictive measures on investment? Have restrictions on investment stimulated capital accumulation in the markets?    (Nicoletti et al., 2003;Ghosh et al., 2012). Some studies have underlined the restrictive effects of measures on cross-border M&A in the secondary sector and non-financial services (Gregori and Nardo, 2021). Others have highlighted the negative effects of these measures on cross-border M&A in the services sector (Mistura and Roulet, 2019). Mistura and Roulet (2019) investigate the impacts of restrictions on global inward FDI and cross-border M&A across 60 advanced and emerging countries from 1997 to 2016, and Gregori and Nardo (2021) analyze the effects on EU member countries from 2011 to 2018. Few studies have empirically measured the impact of sectoral restrictive measures on overall inward FDI between advanced and emerging countries.
Our paper contributes to the literature about the effects of restrictive measures on FDI in advanced and emerging countries. However, it is innovative in 3 aspects . First, similar to Mistura and Roulet (2019), we estimate through an augmented gravity model the effects of restrictive measures on cross-border FDI in 49 developed and emerging countries from 2010 to 2019. We use the OECD FDI Regulatory Restrictiveness Index that captures barriers to FDI entry in 22 economic sectors across 69 countries. The index measures institutional factors that could influence FDI, such as restrictions on equity or key personnel for foreign investors, limitations on the establishment of branches, and clauses on profit and capital repatriation. Moreover Mistura and Roulet (2019) investigate the effects of restrictions on global FDI across advanced and emerging countries in 2001 to 2012 and on M&A from 2001-2016. our paper provides a more recent analysis of the effects of restrictions on global FDI among advanced and emerging countries from 2010 to 2019. It considers the latest international guidelines for compiling foreign direct investment (FDI) statistics. 1 Second, Mistura and Roulet (2019); Gregori and Nardo (2021) study the effects of different types of restrictive measures on FDI. Our model examines in more details the impact of global and sectoral restrictive measures on cross-border FDI. We consider FDI restrictions in the primary, secondary, and tertiary sectors, and further investigate the more disaggregated sectoral effects. The last contribution relates to the type of country considered. Contrary to Mistura and Roulet (2019), who analyze the effects of restrictions on FDI between OECD and non-OECD countries, we examine these effects at three levels: first, a study on FDI between advanced countries (intra-OECD), then among OECD countries and emerging countries (BRICS and some Latin American and Asian countries), and finally between OECD countries and middle-income countries (North African countries and South Africa). 2 Indeed, depending on type of economy, the impacts of sectoral measures are different. OECD countries have a more service-oriented economy, emerging countries have a manufacturing and primary sector-based economy and the economy of African countries depends on agriculture and natural resources.
Our results suggest that overall restrictive measures have negative and non-significant effects on cross-border FDI across OECD countries, while restrictions in the service sector have negative and significant effects on FDI. Indeed, an increase of 0.05 points in FDI restrictions in this sector decreases inward FDI by about 17.36%. Moreover, the overall restrictive measures and those in the service sector have negative and significant effects on inward FDI among OECD and emerging countries. An increase of 0.05 points is associated with a decrease in inward FDI respectively by 14.43% for global measures and 31.31% for those in services. In addition, restrictions in the services sector hamper bilateral FDI between OECD and African countries, the effects being more pronounced for African countries. Interestingly, restrictive measures in the secondary sector stimulate FDI between advanced and African countries. The effects are significantly negative for OECD countries, so we can conclude that restrictions in this sector boost inward FDI in African countries. The analysis of disaggregated sectoral restrictive measures shows that restrictions in business and other financial services are negatively associated with intra-OECD FDI and restrictive measures in the banking sector have a positive and significant impact on intra-OECD FDI. A rise of 0.05 points in restrictions in the banking sector is associated with an increase in inward FDI by 28.30% on average. The 1 In 2014, many countries implemented the latest international guidelines for compiling foreign direct investment (FDI) statistics (OECD, 2015). 2 The countries with FDI restrictive measures data available in our sample are North African countries and South Africa. restrictive measures in the manufacturing sector have restrictive impacts on inward FDI in big emerging countries, most pronounced in the BRICS countries, and those in the mining, quarrying, and oil extraction sector are barriers to inward FDI in African countries. Reforms to liberalize sectoral restrictions by country have positive effects on FDI, but deregulation of the services sector has beneficial effects on inward FDI in all countries.
The remainder of this paper is structured as follows. The next section documents recent literature on the effects of FDI regulation on investment. In the second part, we describe our econometric model with data, sources, and the type of regression used. The third section presents and discusses the results. The last section uses the results of this study to perform policy simulation.

II. Literature Review
Empirical studies that examine the impact of restrictive measures on FDI use two main indices. Some authors study the effects of restrictions on FDI using the FDI Regulatory Restrictiveness Index of OECD and others the indexes of capital account restrictions based on IMF's AREAER database. 3 Authors such as Nicoletti et al (2003), based on the original version of the index created by Golub (2003) 4 and the OECD's PMR 5 , explore the effects of FDI restrictions and other policies on foreign direct investment in a panel of 28 OECD countries between 1980 and 2000. The paper uses a gravity model to control bilateral outward FDI flows and a dynamic panel model to explain total multilateral inward FDI stocks. Their results suggest that FDI restrictions could reduce bilateral outward FDI stocks by between 10% and 80% on average, depending on the type of restriction. Inward FDI stocks are also impacted by FDI restrictions, but the results should be treated carefully due to the lack of variability of restrictions across OECD member countries. The analysis of Ghosh et al (2012) similar to the previous one shows the impact of restrictions on inward FDI stocks using panel data ) for 23 OECD countries. Based on the updated of Koyama and Golub (2006) index and a gravity model, they find significant negative effects of restrictions on inward FDI stocks. To determine the short-and long-run effects of the restrictions, they use an autoregressive distributed lag model. Their results show that the short-run elasticity estimated was between 0.06 and -0.14, and the long-run elasticity between -0.64 and -1.49.
The analysis of Ahrend and Goujard (2012) indicates that FDI restrictions may contribute to greater risks of financial crisis. Indeed, higher restrictions in OECD countries, measured by OECD indexes, and anti-competitive product market regulation have contributed to reduced financial stability. That is due to a rise of countries' debt over FDI or capital investment. By contrast, more stringent capital regulations for banks and more openness to foreign bank entry have reduced the vulnerability to financial contagion. Fournier (2015) examines the determinants of foreign direct investment (FDI) from 1998 to 2013, including FDI restrictions. Using gravity models and the recent version of the OECD FDI restrictiveness index, he finds a 3 Database on the exchange rates and trade regimes of all members of the International Monetary Fund (currently 189 countries) and three territories (Aruba, Hong Kong SAR, and Curaao and Sint Maarten -formerly the Netherlands Antilles). 4 The first to construct the aggregate index of FDI restrictions. 5  . It is positively influenced by the size of source country (E i ), as large economies tend to invest more in technological capital. The stock of bilateral FDI is also positively influenced by the size of destination country (Y j ), as large economies can in principle absorb more foreign technology. If the size of the aggregate stock of technological capital in country i is denoted by M i , the ratio Y i M i can be considered as a gross measure of the potential absorptive capacity of country j for FDI-related technological capital from country i. FDI flows are impeded by obstacles or frictions. For FDI, the relative openness of country j to foreign technologies can be represented by ij, which has values from 0 to 1. If w ij = 1, country j is fully open to the entry of technological capital from country i, while in the case of w ij = 0, no technological capital from country i is allowed. All these factors are the main determinants of the bilateral stock of FDI. 7 The general formulation is as follows: Wth E i measures the size of country i as a total expenditure, including expenditures for the development of technological capital; Y j is a measure of the size of host country j. The parameter η is the elasticity of FDI revenue flows with respect to the measure of openness. More openness in country j will lead to more frequent use of the technology stock, which will lead to an increase of FDI revenues. The other elements of equation (1) come from the structural gravity system for trade, in which the FDI determinants are integrated. α groups a set of fixed parameters from the theoretical model. 8 Finally, P i is the inward multilateral resistance term of the gravity trade model. They aggregate the bilateral trade costs of country i with all other countries: With t ji represents the bilateral trade-cost frictions (bilateral distance, having different languages, common border..) that increase bilateral trade cost. Y = ∑ Y j is world production or world GDP, used to normalize the size of destination country (Y j ), and σ is the elasticity of substitution from CES functions used to aggregate multilateral resistance (MR) terms. 9 7 Time indices are omitted in this representation. 8 These include parameters such as the depreciation rate, the utility function discount factor and other parameters that are used in the underlying theoretical model (see Anderson et al., 2016Anderson et al., , 2017. 9 With σ > 1, the elasticity of substitution shows that all countries have a preference for a variety of products and technological capital by origin country. World trade is a fully integrated system, equation (2) also contains the term Π j , which represents the outward multilateral trade resistances of country j. It aggregates the bilateral trade costs of country j with respect to all other countries. The gravity system of the FDI becomes: Equation (3) shows that if trade costs increase in host country j, domestic prices rise and thus reduce the country's real potential to absorb foreign technological capital.
The author has highlighted the gravity estimation of bilateral FDI remains Bénassy-Quéré et al. (2005). The latter study the impacts of FDI determinants on horizontal FDI. In the model, bilateral FDI stocks depend on both economies' GDP, the determinants of supply and demand, and the distance between capital. However, recent theoretical developments have provided other foundations for the application of a gravity model to other With FDI ij,t represents FDI stocks from country i (the reporting or source country) to country j (the partner or host country) in period t (2010-2019)= FDI ij,t GDP de f lator (ij,t) . FDI are calculated by dividing FDI stocks (in U.S. dollars) by the average of the source and host country GDP deflators to remove inflation. FDI ij,t−1 is one-year lagged dependent variable (Egger and Merlo, 2007). 11 Following Mistura and Roulet (2019), all explanatory variables are lagged by one year to reduce potential endogeneity issues. FDI RI s j,t−1 is our interest variable (FDI Regulatory Index), i.e. the restrictive measures on FDI implemented in the destination country j at the time t-1 in the sector s (s= primary, secondary and tertiary sectors). It is measured by the OECD FDI restrictiveness index. t ij,t−1 includes time-invariant bilateral control variables, i.e. bilateral distance, common language, common border, colonial links (Blonigen and Piger, 2014) 12 and time varying variables such as regional trade agreements (RTA), bilateral investment treaties (BIT) and human capital dissimilarity (HCD ij,t ). According to Ethier and Markusen (1996), the difference in factor endowments can affect inward FDI. 13 We include standard gravity variables, specifically the GDP of both the origin country and the destination country. X jt covers destination country specific characteristics such as regulatory quality (Bénassy-Quéré et al., 2007) and FDI determinants in destination country such as trade openess 14 , productivity, labour freedom index 15 , tax burden (Djankov, Ganser, McLiesh, Ramalho, and Shleifer, 2010). 16 α i and φ j represent source-host country fixed effect (dummy variables that control the inward and outward multilateral resistance terms). γ t is a time fixed effect (capturing the global macroeconomic cycle) and ijt is a error term. Standard errors are clustered by country pairs to control for potential heteroskedasticity and to limit the potential effect of persistence over time of FDI stock levels in each pair of countries, see Fournier (2015). Our dependent variable takes a value of zero and an estimation using OLS leads to a bias (zero FDI is associated with high bilateral fixed costs). To avoid biased estimation results, we use the Pseudo-Maximum Likelihood estimator (PPML) suggested by Santos Silva and Tenreyro (2006). The PPML is used in our case in order to deal with the constraints of zero trade between States, and also estimates the non-linear shape of the gravity model in the presence of heteroskedasticity. However, an important assumption of the PPML estimator is equidispersion, which means that the conditional variance of the dependent variable and its conditional mean are equal. PPML estimation can be assessed by solving the following condition: Where p is the country pair, X p is unilateral trade (i.e. exports or imports) between country pairs in non-logarithmic levels and Z p is the complete vector of the gravity equation as defined above.
We include in the estimation fixed effects by origin country to control for unobservable multilateral resistance terms (Olivero and Yotov, 2012). 17

V. Data Description
To analyse the effects of restrictive measures on FDI stocks between OECD and emerging countries, we use panel data across 49 countries from 2010 to 2019. Indeed, we consider OECD 12 They identify as main enabling factors for inward FDI the traditional gravity variables such as cultural distance, difference in labor endowment, and the presence of trade agreements. 13 Golub et al., 2003 define human capital dissimilarity as the difference in absolute value between the human capital index in the destination country and that in the source country HCD ij,t =| (ln(education jt − ln(education it ) |. 14 Chakrabarti (2001) finds that a country's degree of openness to international trade is a relevant determinant of the FDI decision, because most investment projects concern the tradable sector. 15 Nordin et al. (2019). 16 They estimate, using cross-sectional survey data for 85 developed and developing countries, that corporate taxes always have a negative and significant effect on FDI inflows. 17 We also estimated the model using destination country's dummies, but the model presents convergence issues (Santos Silva and Tenreyro, 2011). countries because inward and outward FDI account for a large share of GDP and we include emerging countries due to the high level of restrictions in FDI (see figure 4). Annual data from 2010-2019 to explain the decline in inward FDI since 2016, and to consider the new FDI statistics introduced in 2014. Our dependant variable is aggregate bilateral FDI stock. The data are collected on OECD Foreign Direct Investment Statistics. 18 The data cover a range of advanced and emerging countries in terms of origin and destination. However, we use the latest version set up by the OECD. 19 This database highlights bilateral FDI between OECD member and non-member countries and runs from 2005 to 2019. It also highlights sectoral FDI (primary, manufacturing and service sectors). 20 Missing data (or non-reported, suppressed) and negative FDI are replaced by 0 in our case, because negative values are interpreted as disinvestment and to have a balanced panel (Kox and Rojas, 2019).   19 Benchmark Definition 4th Edition (BMD4). 20 Sector-specific FDI data are not bilateral so the sectoral analysis will be conducted with the sectoral FDI restriction index. In addition, bilateral resistance variables such as the bilateral distance between the two capitals and binary variables (common border, language and colonial links) come from CEPII (Centre d' Etudes Prospectives et d' Informations Internationales) database. Binary variables such as regional trade agreements are obtained from the WTO (Regional Trade Agreements Information System, RTA-IS) and information on the signing and ratification of bilateral investment treaties is taken from on UNCTAD's International Investment Agreements database.
We consider the specific characteristics of the destination country that affect inward FDI such as regulatory quality (data available on Worldwide Governance Indicators). It refers to perceptions of the ability of the government to formulate and implement sound policies and regulations that permit and promote private sector development. The indicator is estimated yearly at the country level, in units of a standard normal distribution, ranging from approximately -2.5 to 2.5.
FDI determinants in destination countries such as tax burden and labour freedom index are collected on The Heritage Foundation database. 21 Tax burden is a composite measure that reflects marginal tax rates on both personal and corporate income and the overall level of taxation (including direct and indirect taxes) imposed by all levels of government as a percentage of gross domestic product (GDP). The labor freedom component is a quantitative measure that considers various aspects of the legal and regulatory framework of a country's labor market, including regulations concerning minimum wages, laws inhibiting layofs, severance requirements, and measurable regulatory restraints on hiring and hours worked, plus the labor force participation rate as an indicative measure of employment opportunities in the labor market. These two indices are graded on a scale of 0 (less flexible) to 100 (more open or flexible). Productivity are collected on OECD database. 22 Human capital index in Penn World Table database of Groningen Growth and Development Centre. 23 Trade openness, GDP of exporting and importing countries are extracted from the World Bank database (World Development Indicators)

VI. FDI Gravity Results and Discussion
This section presents and analyzes the empirical results. We estimate the effects of restrictive measures on FDI stocks between OECD and emerging countries from 2010 to 2019. Tables 2, 3 and 4 present the baseline gravity model results for cross-border FDI using the OECD FDI restrictiveness index. Table 2 presents the results of restrictions on bilateral FDI among advanced countries (intra-OECD) and tables 3 and 4 respectively the results on cross-border FDI among advanced and big emerging countries and between OECD and middle-income 21 https://www.heritage.org/index. 22 It measured by GDP per hour worked (U.S dollars). 23 This index refers to the number of years of schooling and return on education. emerging countries (African). 24 In order to determine whether the effects of FDI restrictions can be expected to differ for developed and developing economies, we include a dummy variable equals 1 if the destination country is an OECD member in year t and an interaction term with the variable of interest capturing their level of FDI restrictiveness (results presented in Tables 3 and 4). We confirmed the validity of our results using a series of robustness tests presented in Tables 5 to 11. We apply the PPML estimator to control for the zero FDI and heteroskedasticity issues, we also include country fixed effects to control for structural multilateral resistance. Therefore, in order to better capture multilateral resistance and produce unbiased estimates, we include country and time fixed effects separately.

i. Baseline results
In terms of variables of main interest -the FDI restrictiveness indices -we find that global FDI restrictions have a negative and non-significant effect on intra-OECD FDI because the sectoral FDI regulation index is larger than the global index (see Figure 5). Moreover, these restrictions represent a barrier to cross-border FDI among OECD countries and the big emerging countries. The introduction of reforms leading to a 0.05 point reduction in level of FDI restrictiveness could increase bilateral FDI inward stocks by around 17.17% on average (column 4 of Table  3). 25 The interaction between our dummy variable and the restrictive measures does not have a significant effect on FDI. The combined negative and significant effect of global restrictive measures and non-significant effect of our interaction term suggests that the negative effects of restrictions on FDI tend to be more accentuated for emerging economies. 26 This result confirms the findings of Mistura and Roulet (2019). Further, global restrictive measures do not affect cross-border FDI between OECD and African countries (Table 4). 27 The sectoral analysis of FDI restrictive measures shows significant negative effects of restrictions in the services sector on intra-OECD FDI. The effect of a 0.05 point reduction in FDI restrictions in services is associated with a 17.36 % increase in global FDI (column 4 of table 2). The reason is that services are the largest sector for inward investment in the OECD and also the most restrictive compared to other sectors, which explains the non-significant results (see Table 1 and figure 5). Indeed, the manufacturing sector, excluding defense and military goods, is the most open sector where countries generally allow foreign investment. In the primary sector, the location-specific and licensing-heavy nature of some of such investments (e.g., extractive industries) may offer relatively few alternatives for foreign investors. The existence of numerous taxes and royalties where economic rents are potentially high, as in the extractive industries, to capture a part of such rents for their nationals limits investment in this sector (Mistura and Roulet, 2019). The analysis on FDI across advanced and large emerging countries suggests significant negative effects of restrictive measures in the services sector. A reduction in service sector restrictions by 0.05 points leads to an increase in global inward FDI by 31.31% on average, and these effects are greater for emerging countries.
Insert Table 1 here. 24 The estimated coefficients from regressions pooling wealthy and poor economies may fail to capture the true relationship between FDI and the explanatory variable of interest (Blonigen and Wang's, 2004). 25 The percentage change in inward FDI from a 0.05 point reduction in the FDI restrictiveness index is calculated as follows: [exp(-0.05*coefficient)-1]*100. 26 The reason is that global and sectoral restrictions, especially in services, are very high in emerging countries. 27 We have a limited sample of African countries, which could explain these results. Moreover, restrictions in the services sector hamper bilateral FDI between OECD and African countries, and as before the effects are more pronounced for African countries (column 8 of table 4). Interestingly, restrictive measures in the secondary sector boost FDI between OECD and African countries. Our interaction term shows significant negative effects of these measures for OECD countries and the combined effects show significant positive effects of restrictions in the manufacturing sector for African countries. The negative effects of restrictive measures in the secondary sector on FDI between OECD countries refers to the existing literature on the effects of restrictions in services on the performance of manufacturing firms (Ariu et al., 2019; Ariu et al., 2020; Amara, 2021). Indeed, restrictions on FDI in services impact the productivity of manufacturing firms (increased costs of sourcing services) as they increasingly use services as an important input of manufacturing value added. The positive effect of restrictions in the manufacturing sector for African countries can be explained by the fact that the index in this sector measures FDI restrictions in the food, chemical, metal, and electronics industries. Contrary to OECD countries where these sectors are developed with more competition, they are still nascent in North and South Africa with lower competition. An increase in restrictions in this sector leads to an increase in economic rents from FDI (Rouzet and Spinelli, 2016). One of our robustness tests incorporates restrictions in the mining and natural resources sector (a developed industry in Africa), and we find negative effects on FDI in Africa. The negative impacts of restrictive measures in FDI are due to high entry costs in the different sectors. In addition to acting as a barrier to entry, this result may also underscore a potential signaling effect of restrictions about the difficulties in doing business as a foreign investor, including outside of the restricted sectors. Services are the sectors that receive the most FDI and thus are the most affected by these restrictions.
Insert Table 2 here.
Insert Table 3 here.
Insert Table 4 here.
In the different specifications, we also find the following effects: the standard gravity model variables such as distance, common language and colonial links have the expected signs and magnitudes and are statistically significant. Indeed, distance has negative and significant effects on cross-border FDI inflows among advanced and emerging countries as in the theory (Bénassy-Quéré et al., 2005; Basile et al., 2008). Binary variables such as common language and colonial links have positive and more significant effects on FDI between OECD and emerging countries. Moreover, the common border has negative and significant impacts on FDI inflows, which is contrary to the theory. This negative and significant sign is due to the substitutability between trade and FDI, especially between countries that share the same border (Gregori and Nardo, 2021 Alfieri, 2020). Regional trade agreements (RTA) have significant negative effects on FDI in advanced countries, while there are significant positive impacts on cross-border FDI among advanced and emerging countries. The negative effect of trade agreements suggests that being a partner in a trade agreement tends to discourage FDI between origin and destination countries, because a trade agreement facilitates access to the destination country's market through other channels, such as exports or greenfield investments (Mistura and Roulet, 2019). Bilateral investment treaties (BIT) have negative and significant effects on cross-border FDI, more meaningful effects on FDI across OECD and emerging countries. The BIT has different characteristics than the RTA, it protects the investor against risks in the market receiving the FDI. It therefore establishes transparency on the host country (Bergstrand and Egger, 2013). These agreements among advanced countries and emerging or developing countries (North-South investment) have much greater effects than North-North agreements (Kox and Rojas, 2019). BITs affect negatively FDI inflows when the political risk in a country is high whereas the opposite occurs when the risk level is low (Tobin and Rose-Ackerman, 2005). 28 If we consider the determinants and explanatory factors of FDI, the regulatory quality variable is positively related to cross-border FDI, highlighting the proactive role that public administration can exert to stimulate foreign investment (Gregori and Nardo, 2021). Trade openness attracts investment because of a complementarity between FDI and trade (Belke and Clemens, 2018). The dynamism of the destination economy has positive and significant impacts on FDI. These results confirm that the size of the destination country's market boosts cross-border investments by creating additional market shares, economies of scale, or reducing production costs for foreign investors (Eicher et al., 2012). Likewise, the dynamism of investing country is positively related to cross-border flows (Gregori and Nardo, 2021). The productivity measured by the GDP per hour employed is positively associated with FDI. Like the GDP of the destination country, productivity in the destination economy creates economies of scale and reduces production costs, which attracts cross-border FDI. The difference in relative human capital endowment is significant with a negative impact 29 Table 5 and show more significant effects in the period 2015 to 2019 and confirm our results found above; however, during this period, global restrictions have negative and significant effects on intra-OECD FDI and on FDI between OECD countries and large emerging economies. Restrictive measures in the post-crisis period had a greater impact on FDI between OECD countries and emerging markets.
Insert Table 5 here.
The second test examines the effects of restrictive measures on FDI between advanced and BRICS countries 32 considered as large emerging countries and between OECD and emerging countries excluding BRICS. This test shows whether the effects are different among emerging countries. The results reported in table 6 show that restrictions have negative and more significant impacts on BRICS than other emerging countries and confirm that restrictions in the service sector have a significantly negative impact on cross-border FDI.
Insert Table 6 here.
OECD inward and outward FDI in the services sector is the largest in comparison to the other two sectors (accounts for almost 60% of total FDI, see table 1) and considering table 7, FDI in the financial sector is the biggest of the cross-border FDI in services. The third test looks at the effects of restrictions in the disaggregated financial services sector on intra-OECD FDI. Table 8 presents the results and shows that restrictions in business and other financial services are negatively associated with cross-border FDI (negative and significant results). However, FDI restrictions in the insurance sector have no significant effect and those in the 30 Their study argues that the growth-effect of FDI is possibly influenced by the flexibility of the labor market in the host country. 31 Gregori and Nardo (2021) observed a stagnation of inward M&As in EU countries during 2011-14. 32 Brazil, Russia, India, China and South Africa.
banking sector have a positive and significant impact on intra-OECD FDI, explained by the profit margin of FDI in this sector as a result of restrictions (Rouzet and Spinelli, 2016). 33 A rise of 0.05 points in restrictions in the banking sector is associated with an increase in inward FDI by 28.30% on average.
Insert Table 8 here.
The two tests below estimate the effects of restrictive measures on FDI between OECD and emerging countries by type of economy. The big emerging countries have economies more oriented to the manufacturing and agricultural sectors than the North and South African countries where the main economic activity is agriculture and natural resources (table 9). In the first part, we study the effects of restrictions in the agricultural and manufacturing sectors on FDI between OECD countries and large emerging economies (results reported in table 10) and the second the effects of restrictions in the agricultural and natural resources sectors (mining, oil and gas, etc.) on FDI among OECD and African countries (table 11).
Insert Table 10 here.
Insert Table 11 here.
The results suggest that FDI restrictions in the manufacturing sector have significant negative effects on FDI between advanced and emerging countries, the effect being more pronounced if the destination country is an emerging country. In addition, restrictive measures in the mining, quarrying, and oil extraction sector have significant negative effects on FDI between advanced and African countries, with higher effects for African countries.
The last test examines the effects of restrictive measures on FDI across advanced and all emerging economies including Africa. North and South African countries are considered upper-middle-income countries according to the World Bank classification. The results presented in table 12 show significant negative impacts on FDI restrictions in the services sector. In addition, measures in the secondary sector are positively associated with inward FDI (significant but low magnitude). The positive impact is mainly due to restrictions in the secondary sector of African countries, as highlighted in our baseline results.
Insert Table 12 here.
These results underscore the implementation by the governments of attractive sectoral regulation of FDI. The effects of FDI vary according to the type of economy, but unanimously, liberalization of the services sector has more beneficial effects on inward FDI in all countries in order to boost the performance of manufacturing industries. The big emerging countries in addition to the services sector should deregulate the manufacturing sector which plays a vital role in economic activity, and for African countries, reforms to liberalize the natural resources sector would further boost inward investment. Governments also need to regulate FDI taking into account the restrictions of other countries, because regulations will affect FDI differently depending on the destination country.

VII. Conclusion
This paper, which investigates the effects of restrictive measures on FDI, contributes to the literature on the impact of restrictions on FDI, but differs from recent studies because it examines the sectoral effects of FDI restrictions on cross-border FDI between advanced and emerging countries.
Using a gravity model we examined the effects of global and sectoral restrictive measures on FDI among OECD countries, between advanced and large emerging countries and finally between advanced and African countries from 2010 to 2019. Our results suggested that global restrictive measures have non-significant negative effects on cross-border FDI between OECD countries, while restrictions on FDI in the services sector are negatively associated with inward FDI. In addition, both overall and service sector restrictions have negative and significant effects on inward FDI among OECD and emerging market countries. In addition, restrictions in the services sector hamper bilateral FDI between OECD and African countries, the effects being more pronounced for African countries. Interestingly, restrictive measures in the manufacturing sector stimulate FDI between advanced and African countries. The effects are significantly negative for OECD countries, so we can conclude that restrictions in this sector stimulate inward FDI in African countries. The results suggested that restrictions in business and other financial services are negatively associated with cross-border FDI. However, restrictions on FDI in the insurance sector have no significant effect and those in the banking sector have a positive and significant impact on intra-OECD FDI. The last two tests showed that restrictive measures in the manufacturing sector have restrictive impacts on inward FDI in emerging economies, particularly in the BRICS, and that restrictions in the mining, quarrying, and oil extraction sector are an obstacle to inward FDI in African countries.
We could improve our study by considering financial FDI such as M&A. However, there are some important limitations, mainly related to the data. First, FDI restrictions have some limitations, notably that they are time-invariant for some sectors. Second, it would be really interesting to also examine the effects of these restrictive measures on domestic investment in future research. This paper highlighted the detrimental impact of restrictive measures on cross-border FDI. It showed the negative impacts of sectoral restrictions depending on the type of economy receiving the FDI. From this study, we conclude that the drop of inward FDI in OECD countries since 2016 is due to a rise of restrictive and protectionist policies in order to protect local firms. It is also the result of Donald Trump's tax cuts since 2017. This measure led to repatriation of profits into United States. This decrease is probably due to the trade war between China and United States, which has a considerable effect on production and investment in global value chains.