Policy Research Working Paper 9484 The Determinants of Regional Foreign Direct Investment and Its Spatial Dependence Evidence from Tunisia Bechir N. Bouzid Sofiene Toumi Human Resources Policy, Analytics and Knowledge Management November 2020 Policy Research Working Paper 9484 Abstract This paper explores the relationship between key economic to the same regions as well as nearby ones. These agglomer- and institutional attributes of Tunisian governorates and ation forces are relatively strong in Tunisia in the presence their ability to attract foreign direct investment inflows. of vertical foreign direct investment. Further, the results A dynamic generalized method of moments and spatial indicate that a relatively developed market size, an increase autoregressive approaches are used to estimate a model of of regional development areas, as well as robust governance regional foreign direct investment over the recent period. practices and infrastructure are positive determinants of The results provide evidence of regional interdependence regional foreign direct investment inflows. Finally, the of foreign direct investment that appears to be highly clus- paper shows that although some of the determinants exhibit tered along the coastal areas. An increase/decrease of foreign spillover effects on nearby regions, the direct effect on the direct investment inflows to a given region creates an incen- region represents the bulk of the influence over foreign tive/disincentive for other foreign direct investment inflows direct investment inflows. This paper is a product of the Human Resources Policy, Analytics and Knowledge Management. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at bbouzid@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team The Determinants of Regional Foreign Direct Investment and Its Spatial Dependence: Evidence from Tunisia Bechir N. Bouzid Sofiene Toumi World Bank University of Manouba Washington D.C. Tunisia JEL Classification : C33 ; F21; R12. Keywords: FDI determinants; Regional Disparities; GMM; Spatial Dependence. 1. Introduction The liberalization of economies around the world has experienced an accelerating trend over the last three decades. Many emerging and former centrally planned economies have moved toward more liberalization of their production systems and the opening of the economies to external investors. Competition among those countries to attract financial inflows is at the highest level. Authorities are struggling to put in place effective reform policies to provide investors adequate and attractive frameworks to convince them to channel their funds to the local economy. A healthy business environment and investment climate are prerequisites for emerging countries. This can attract multinational firms with high added value that could initiate a strong and sustainable growth dynamic. Today, foreign direct investment is a determining factor for inclusive growth for a number of countries and an accelerating engine for their production systems. Developed countries have largely benefited over the years from FDI inflows and their spillover effects on job creation, improvement of the well-being of the local population, technology transfers, mobilization of external funds and the strengthening of local production and export capacities (Pegkas, 2015). In Tunisia, local authorities consider foreign investment as an important element for the country’s long-term development strategies and place the attraction of multinational firms among the country's economic priorities. Despite efforts and goodwill, it is worth noting that over the year there have been persistent, major disparities in the regional distribution of foreign investments to the country, which led to unbalanced domestic growth between coastal cities attracting the lion’s share of foreign inflows (and experiencing relatively higher economic dynamics) and the interior regions of the country, which continue to crawl behind in the development process. The economic and social emergence of a country like Tunisia remains hostage of this vicious circle created by the regional imbalances and uneven distribution of investments. Using a dynamic GMM and spatial autoregressive approaches, we estimate a model of regional FDI over the recent period that provides an analytical framework that looks at a set of economic, institutional and geographical determinants. The goal of this paper is to identify some of the determinants of the unbalanced regional distribution of foreign direct investments in Tunisia. We aim to assess the key factors in each region and identify the spatial disparities in terms of investment conditions. We start by analyzing the factors influencing the choice of the location of foreign investments. We then look at the distribution of FDI between the Tunisian governorates by highlighting their concentration along certain regions and their role as an incentive for the entry of new investors. The paper is organized as follows. The next section reviews the relevant literature and develops the theoretical framework of our study. Section 3 discusses the current situation of FDI inflows in Tunisia. Section 4 presents the data and empirical methodology. Section 5 discusses findings and analyzes the results. The final section concludes the paper. 2. Determinants of the choice of locations The eclectic paradigm (Dunning, 1979, 1988) remains one of the main theoretical frameworks used to explain the determinants of the firm globalization since it integrates the different theories that consider the firm’s decision to expatriate beyond its homeland borders and links this decision to the stage of production evolution (Vernon, 1966) and to the related costs (Hirsh, 1976; Buckley and Casson, 1981). According to Dunning, the firm chooses a direct investment strategy as a way to penetrate foreign markets if this strategy is found to simultaneously achieve all OLI advantages, namely: the monopoly advantage of the investor (Ownership advantages), the localization abroad (L) and the advantage in internalizing and controlling the market by setting up a subsidiary (I). Why do firms choose to invest in one region and not another? The choice of the location depends on the degree of adequacy of the company's objectives with the attractiveness factors of the host country (Okafor et al, 2015; Mucchielli, 1991). The investor seeks mainly one of the two objectives. On the one hand, the search is for new markets to sell its products in countries with a growing domestic demand as stipulated by the theory of horizontal investments (Chen and Ku, 2000; Markusen and Markus, 1999). On the other hand, in establishing its production process in foreign territories, some investors are looking for a better cost rationalization with a goal to exporting the final outputs outside the country of production as described in the vertical investment framework (Michalet, 1999; Yang, 1999).1 Firms opting for either strategy obey the same investment conditions and are attracted by the same factors. An exception to this is perhaps the size of the domestic market that is relatively neglected in the case of a vertical strategy. Aside from the local market size, the strategy of investments of multinational firms is profoundly influenced by the economic, political and social conditions in the host country (Altamonte, 2007). All attractiveness factors are not however of the same importance for international investors. Michalet (1999) points out the existence of certain fundamental conditions that are found to influence 1 Eckholm et al (2003) highlight another form of direct investment strategy described as an oblique model. According to the authors, multinational firms could consider a host country as a production platform used to export outputs to a group of neighboring countries with sizable domestic markets. investors’ decisions, such as political and economic stability, legislation, infrastructure and human capital. Other conditions are not as important but remain necessary to allow host countries to appear in the short list of foreign investors (geographic and cultural proximity, administrative procedures, corruption, favorable environment for innovation, low production costs, source of raw materials, fiscal and financial incentives, etc.). Differences in location endowments are critical in determining the distribution of foreign direct investments among potential host countries and across regions within the same country. Bailey and Li (2014) studied the determinants for the establishment of US foreign direct investments in 110 host countries between 2006 and 2011. The results show a negative relationship between the flows of these investments and the geographic, cultural and administrative distances. This finding is also confirmed by Blanc-Brude et al (2014). The study found however that the negative impact of geographical distance on FDI inflows is often largely offset by the presence of markets with growing local demands. Human capital also plays a decisive role in the choice of the host territories (Cleeve et al, 2015 and Salike, 2016). Salike (2016) studied the underlying factors of regional distribution of FDI in 31 regions across China. He concludes that foreign investors consider human capital as the most important factor in making their decisions to set up in the region. Investors are attracted by the availability of human resources, in terms of endowments (supply, health and talents of the workforce) as well as growth potential (demographic evolution). He posits that, overall, once countries are able to provide good education for their population, modernized infrastructure and a good legal system with enforcement to protect the investors’ property assets, other things being equal, these countries will be more able to attract foreign investments. Studying Latin American economies over the period 1990-2010, Sanchez-Martin et al. (2014) highlighted the importance of political stability (solidity of the regime, low risk of expropriation) in the foreign investment process. The quality of institutional factors, such as the degree of complexity of administrative procedures, the level of corruption and government tax incentives are all found to exert a positive influence on the attractiveness of foreign firms. On the contrary, the authors show that regional integration agreements do not seem to be decisive in the choice of the locations for FDI. Similarly, robust macroeconomic factors and growth prospects have a positive impact on attracting foreign investments as shown by Boateng et al, (2015) for the Nordic countries with high inflation, unemployment and the interest rate hikes producing significant negative results on FDI inflows. These results show that the consolidation of macroeconomic indicators provides the host country with strong competitive advantage to attract foreign investments. Studying the impact of the exchange rate variation on foreign investment inflows to nine Chinese manufacturing sectors between 1981 and 2002, Xing (2006) shows the role played by the Chinese exchange rate policy in the establishment of foreign firms in the country. The devaluation of the yuan and its attachment to the dollar have improved over the years the country's competitiveness and strengthened its attractiveness to foreign investors. Huyen (2015) conducted a survey of several foreign companies established in provinces in Vietnam to identify the determinants of their FDI decisions. The results show the importance of the quality of infrastructure, access to finance and the availability of raw materials in the process of setting up foreign companies in Vietnam. Bartels et al, (2014) highlighted the importance given by multinationals to the quality of living in the host regions, the existence of a favorable environment for innovation and the emergence of industrial hubs allowing investors to benefit from positive externalities. This finding is however nuanced by certain research which found that FDI attractiveness is a dynamic process that is evolving over time with certain categories of projects requiring specific criteria that may not be necessary in the case of other projects. Thus, certain territories may be considered attractive for a specific sector or production activities and not attractive for others within the same location (Hatem, 2004). 3. FDI inflows to Tunisia: Where do we stand now? The FDI inflows to Tunisia have experienced a decline since the revolution in January 2011. In recent years, a major part of those inflows was mainly channeled by investors already established in the country. In the recent 2020 edition of the “Doing of Business” index of the World Bank, Tunisia has ranked 78th compared to the 88th position in the 2018 edition. The country was able to gain 10 places in two years. However this is still short of the highest ranking the country ever reached in 2012 when it made it to the 46th position (World Bank, 2020). Table 1 presents a summary of the FDI data to the country as released by the Agency for the Promotion of Foreign Investments (FIPA, 2019, 2017, 2016, 2015, 2014). Foreign inflows have substantially declined since 2008. The trend has slightly reversed from 2017 but still about 14% below its level in 2008. It is worth noting that the surge in FDI (excluding energy parts) in 2012 was mainly explained by the privatization and transfer to international investors of the state’s shares in a domestic telephone operator. Table 1. FDI inflows in millions of Tunisian dinars (TND) (2008-2018) Years 2008 2010 2011 2012 2013 2014 2015 2016 2017 2018 FDI 3127 2165 1616 2504 1813 1806 1967 1901 2131 2742 FDI (excl.energy) 1124 848 553 1618 737 913 995 1097 1318 1832 Source : FIPA (2019, 2017, 2016, 2015, 2014) At the regional level, despite the generous financial and fiscal incentives (such as the multiple regional development areas) approved by local authorities, foreign actors continue to be reluctant to invest in the inland regions of the country. Only 13% of all foreign firms established in Tunisia (about 16% of created jobs) are presently operating in those regions (World Bank, 2014a). Table 2 shows the importance of regional disparities in Tunisia in terms of attracting foreign investments to the 24 governorates of the country. As shown below, The North-West governorates (Beja, Jendouba, Le Kef and Siliana) have attracted about 100 foreign companies. In contrast, the coastal cities located along the regions of Grand Tunis, North East and Center East, have totaled 3,197 foreign companies (91.5%). The 13 remaining governorates (in the Center and South West) host, altogether, the other 289 firms. Table 2. Regional distribution of foreign firms in Tunisia in 2018 Regions (governorates) Number of firms Grand Tunis (Tunis/ Ariana/Manouba/ Ben Arous) 1234 (35,5%) North-East (Bizerte/ Nabeul/Zaghouan) 873 (25%) North-West (Beja/ Jendouba/Le Kef/ Siliana) 101 (3%) Center-East (Sousse/ Monastir/Mahdia/ Sfax) 1090 (31%) Center-West (Kairouan/Sidi Bouzid/ Kasserine) 69 (2%) South-East (Gabes/ Medenine/Tataouine) 80 (2,5%) South-West (Gafsa/ Tozeur/Kebili) 39 (1%) Total 3486 (100%) The economic development model adopted by Tunisian authorities since the 1970s was based on the industrial competitiveness and trade openness through manufacturing exports. This policy has resulted in a clustering of offshore companies near the ports in the capital city of Tunis and along coastal regions of the Center and Northeast. This has left the interior regions trailing behind in terms of development of productive activities, job creation and standard of living. Over the years, a growing part of the local population in the interior regions has chosen to migrate to coastal cities seeking decent job opportunities and a better standard of living for their families. Not surprising to find that today nearly 56% of the total population is located within an hour's drive from the country's three largest coastal cities: Tunis, Sfax and Sousse (OECD, 2018). According to the Tunisian Ministry of Regional Development (MDR, 2011), the spatial analysis of FDI distribution shows that most capital inflows are directed toward coastal regions in contrast to the small share channeled toward the interior of the country despite the fiscal and financial incentives offered to foreign firms under the investment incentive code. The concentration of FDI in those regions is also explained by the ripple effect created by the establishment of foreign firms in the coastal regions since the 1970s. This effect has promoted the attraction of new foreign investors who chose to settle in the same cities in order to benefit from the presence of other multinationals already in place. Large FDI inflows to a given region tend to create a spillover effect in the following years with a growing number of foreign actors choosing to invest in the same region or in a nearby city. Several reasons could explain this strategic behavior. First, the concentration of foreign investors in a regional territory becomes an incentive for the attraction of new projects in the same region. Multinational firms, looking to foster their competitiveness while minimizing their operational risks, would feel largely reassured about the local conditions and investment climate in a region or country that has already attracted significant FDI inflows from similar foreign players. Investors seek to benefit from synergies created by a dense network of foreign investors (Mafruza et al, 2019). Further, potential investors are often concerned about the living environment for expatriates. The quality of life in the inland regions and remote areas is a major obstacle not only to attract foreign actors but also to retain local skills of those regions. Several authors have shown that one of the main reasons that motivates foreign investors to gather in the same city is the good ‘atmosphere’ that reigns within the region and the existence of recreational facilities creating a favorable working environment, such as relaxation centers, adequate accommodation, schools offering a good level of education and friendly relationship with the local population (Balasubramanyam et al, 1999). Enhancing the attractiveness of inland regions cannot only be achieved by providing fiscal and financial incentives. It is rather essential to improve the quality of living in the regions, the access to good infrastructure and to basic services like health, education, and transport. A relatively developed transportation system has strengthened the agglomeration of FDI in the coastal regions at the expense of the rest of the country. The latter continues to lack some basic transport infrastructure such as highways, modern railways and domestic airports which could reduce the isolation of these regions and bring them closer to the port infrastructure necessary for export activities. This infrastructure problem is exacerbated by high internal transport costs in the country. The average price of road transport of goods is about 0.22 US $ per ton-kilometer (World Bank, 2014b), two cents less than the average price in the United States. The average price of freight transport by road is much higher in Tunisia than many other developing countries such as India (0.06 US $) and Vietnam (0.14 US $). It is also higher than the average prices in Sub-Saharan African countries, ranging from US $ 0.05 to US $ 0.13 (Teravaninthorn and Raballand, 2009). Table 3 highlights the importance of this ripple effect and shows the extent of the disparities in attracting FDI between the coastal regions where the majority of foreign investors are clustered and the interior regions. For years, the Greater Tunis region ( Tunis, Ben Arous, Ariana, Manouba) has been the most attractive region with more than 55% of FDI inflows, followed by the Northeast region (Zaghouan, Nabeul, Bizerte) with a share of 20% and the central-eastern region (Sousse, Monastir, Sfax, Mahdia) with more than 10% of FDI inflows. The share of all other inland regions combined, does not exceed the 15% of total inflows. More than 75% of yearly FDI inflows are clustered within 60km distance from the capital (FIPA, 2019). Table 3. Regional distribution of FDI inflows (excl. energy): 2014-2018 FDI (Million TND) 2014 2015 2016 2017 2018 Regions Total % Total % Total % Total % Total % Tunis 451,1 49,5 361,5 36,2 424 38,7 364,5 27,6 724,6 39,55 Ben Arous 21,4 2,3 85,6 8,6 142,4 13 309,8 23,5 257,8 14,07 Ariana 47,5 5,2 82 8,2 138,2 12,6 90,9 6,9 60,3 3,29 Manouba 9 1 5,7 0,6 32,8 3 8,9 0,7 19,3 1,05 Total 529 58% 534,8 53,6% 737,4 67,2% 774 58,7% 1062 57,97% Grand-Tunis Zaghouan 42,3 4,4 93,2 9,4 80,5 7,3 148,1 11,2 205,3 11,21 Nabeul 31,5 3,4 75,8 7,6 128,4 11,7 73,7 5,6 173,6 9,47 Bizerte 61,2 6,5 39 4 33,1 3 37,6 2,9 91,9 5,01 Total 135 14,3% 208 21% 242 22% 259,4 19,7% 470,7 25,69% North-East Sousse 33,8 3,7 83,1 8,4 36,1 3,3 70,6 5,4 124,3 6,78 Monastir 26,1 3 19,2 1,9 11,9 1,1 27 2 41,2 2,25 Sfax 5,2 0,6 2,3 0,2 1,4 0,1 9,1 0,7 6,9 0,38 Mahdia 14,8 1,6 23,4 2,3 3,4 0,3 11,2 0,9 6,1 0,33 Total 80 8,9% 128 12,8% 52,8 4,8% 117,9 9% 178,5 9,74% Center-East Gabes 89,1 9,95 16,1 1,6 40,3 3,7 87,1 6,6 52,9 2,89 Médenine 0,3 0,05 46,4 4,7 - - 1,8 0,14 0,3 0,02 Tataouine - - 0,8 0,1 - - - - - - Total 89,4 10% 63,3 6,4% 40,3 3,7% 88,9 6,7% 53,2 2,91% South-East Beja 17,3 1,9 15,6 1,6 15,1 1,4 24,7 1,9 50,7 2,77 Jendouba 43,5 4,8 35,8 3,6 4,1 0,4 12 0,9 4,4 0,24 Siliana 12,4 1,4 4,5 0,4 0,4 0,05 1,3 0,1 2,2 0,12 Kef 0,2 0,05 0,4 0,05 - - 2 0,1 - - Total 73,4 8,1% 56,3 5,7% 19,6 1,8% 40 3% 57,3 3,13% North-West Kairouan 2,4 0,3 2,8 0,3 4,1 0,4 3,1 0,2 6,1 0,33 Kasserine - - - - - - - - 0,1 0,01 Sidi Bouzid 1,9 0,2 1,8 0,2 - - 22 1,7 - - Total 4,3 0,5% 4,6 0,5% 4,1 0,4% 25,1 1,9% 6,2 0,34% Center-West Gafsa 1,1 0,1 - - 0,2 0,05 12,7 1 4 0,22 Tozeur 0,3 0,05 - - 0,4 0,05 - - - - Kebili 0,3 0,05 - - - - - - - - Total 1,9 0,2% 0 0,0% 0,6 0,1% 12,7 1% 4 0,22% South-West Total 913 100% 995 100% 1097 100% 1318 100% 1832 100% Recent annual surveys conducted by the Arab Institute of Business Leaders (IACE, 2019, 2018, 2017, 2016) with companies established in all 24 Tunisian governorates, made it possible to assess the attractiveness of each region by calculating a local business climate index. The objective of the annual study was to assess for each region a set of development indicators such as: living environment, infrastructure, access to information, availability of workforce and the quality of local governance and business regulations. According to surveys, the coastal cities continue to benefit from more satisfactory investment climate than the rest of the country. The capital city of Tunis, followed by the 2 other economic poles Sousse and Sfax are the top 3 ranked regions in terms of business climate index where business climate is considered "very satisfactory" mainly due to the quality of infrastructures, a better living environment, the existence of a dynamic local industrial fabric and the availability of skilled labor force. In contrast, the interior regions (North-West, Center-West and South-West) are ranked at the bottom of this index well below the average, with business climate considered as "not at all satisfactory." Table 4. Regional Ranking based on Business Climate Index Regions Ranking Index Ranking Index Ranking Index Ranking Index 2015 2015 2016 2016 2017 2017 2018 2018 Grand Tunis (Tunis/ Ariana/ 1 4,45 1 2,65 1 3,90 1 4,30 Manouba/ Ben Arous) North East (Bizerte/ Nabeul/ 3 3,70 4 1,96 3 2,95 3 3,40 Zaghouan) North West (Beja/ Jendouba/ 5 2,90 7 1,75 7 2,30 5 2,81 Le Kef/ Siliana) Center East (Sousse/ 2 3,98 2 2,60 2 3,35 2 3,82 Monastir/ Mahdia/ Sfax) Center West (Kairouan/ 7 2,60 5 1,92 5 2,40 6 2,72 Sidi Bouzid/ Kasserine) South East (Gabes/ Medenine/ 4 3,10 3 2,16 4 2,70 4 3,25 Tataouine) South West (Gafsa/ Tozeur/ 6 2,65 6 1,80 6 2,35 7 1,96 Kebili) Average Index 3,34 2,12 2,85 3,18 Source : IACE reports on regional attractiveness, 2016, 2017, 2018, 2019 4. Variables and Methodology The objective of this section is to identify the factors that influence the regional attractiveness of FDI inflows in Tunisia. We will attempt to provide a comprehensive analytical approach that looks at a set of economic, institutional and geographical determinants. The goal of this analysis is two-fold. First, we will ask how each governorate’s own economic and business environment contributes to the attractiveness of this type of export-oriented FDI. Second, we will empirically measure to what extent the inflows of FDI to the country tend to follow an agglomerate approach within certain regions. It is not the purpose nor scope of this paper to unpack the underlying reasons of this approach either from home or host country perspectives. Instead, we focus on investigating the determining factors that have influenced the long- standing geographical distribution of FDI primarily directed toward coastal regions. Using a spatial analysis, we aim to demonstrate empirically how the geographical location of certain regions and their proximity from one another could affect their FDI attractiveness. Similar to past studies, the interest of this type of research is to provide some empirical guidance to local policymakers to help them overcome some of the obstacles that regions are facing in attracting FDI inflows. The empirical analysis presented in the next section will be divided into interrelated parts. We investigate in the first part the significance of some of the determining factors of FDI inflows to various governorates of the country. This analysis is complemented in the second part by a spatial analysis of the distribution of the FDI across various regions and potential agglomeration and spillover effects. 4. 1. Presentation of variables In line with past research, the dependent variable in our model is the FDI which measures the inflows in millions of Tunisian dinars of incoming foreign direct investments to each governorate over the period 2008-2018 (FIPA 2019, 2017, 2016, 2015, 2014). The potential determinants of these inflows are captured by the following economic and institutional factors: - Unemployment rate by governorate: The abundance of human resources in a region is often considered as key incentives for foreign investors who are looking for low production costs and skilled labor force (INS 2019, ITCEQ 2016, ONEQ 2013). - Regional GDP: This variable captures the productive capacity of a region. In the absence of data on regional GDP in Tunisia, we decided to use as a proxy, the regional share of consumption of high and medium voltage electricity (INS 2019, 2016). - Number of private companies per governorate (EPV) : This indicator reflects the density and dynamic of the local industrial network. Foreign actors may choose to invest in a region to leverage new business relationships, to get closer to sources of raw materials and/or to have rapid access to the services and resources necessary for production (INS 2019, 2016, 2013). - Number of regional development zones (ZDR): To attract foreign investors and promote the industrial development of certain regions, the government provides companies that choose to establish their businesses in these zones with specific incentives in the form of tax and financial incentives such as: tax exemption, investment premium, or payment of the employer's social security contribution (FIPA, 2019). - Governance, participatory approach and institutional quality (GOV): This variable is measured by an index derived from qualitative and quantitative surveys carried out by the Arab Institute of Heads of Companies in Tunisia (IACE). Several indicators measuring different aspects of good governance are taken into account, such as: the quality of municipal regulations and administrative services, the quality of bureaucracy, the availability of information on the local economic situation, quality of life and quality of urban development and safety (IACE, 2019, 2018, 2017, 2016). - Proximity to the port (POR): This is a binary variable that indicates the existence or not of a maritime freight terminal within 60 km of the governorate. The proximity to port is a critical determinant for certain FDI decisions particularly for foreign investors with significant share of local production turned to export (INS, 2019). Finally, we include in the model a dummy variable to capture the behavior of foreign investors in response to the 2011 revolution. This variable is given the value 0 for the year preceding 2011 and 1 for the post revolution years. 4.2. Methodology Using regional annual data from 24 governorates over the period spanning from 2008 through 2018, we start by constructing and estimating a model measuring the impact of a set of determining factors for the attractiveness of FDI . Similar to past research that have provided strong empirical evidence for certain persistence of the FDI across regions and an endogeneity in the relationship between investment inflows and certain macroeconomic determinants, we chose to run a dynamic panel model that accounts for the lagged dependent variable in the explanatory variables (Campos and Kinoshita;2003, Walsh and Yu;2010). This approach, Kinoshita and Campos (2002) and Batana (2011) argue, will allow to incorporate the past FDI inflows as a proxy for the agglomeration effect. In this paper, we follow the same approach by using a GMM model. As pointed out by several authors in dynamic panel models where endogeneity of certain variables is suspected, the results of OLS and fixed effects estimations will be biased upwards and downwards, respectively. To correct for this bias (finite sample bias), the specifications of our model below follow the method suggested by Arellano and Bover (1995) and Blundell and Bond (1998) to estimate the model using the system GMM approach. As suggested by Roodman (2006), we will use two lags for the system GMM estimators, and we will apply the Windmeijer (2005) finite sample correction for standard errors. All GMM estimators use robust standard errors. The model takes the following general form; , = 0 + 1,−1 + 2 + , + Where , + are the idiosyncratic disturbances (capturing the unobserved region-specific effect and the observation-specific errors of the regions, respectively) and i,t represent the regions and time periods. FDI lag of one period is used as an independent variable to assess its dynamic effect and to test for the agglomerate effect. To account for the endogeneity between variables of the model, the above equation is estimated using as instruments the lagged values of the right-hand side variables in levels. In the second part of the analysis, using a spatial lag model which incorporates a spatially autoregressive dependent variable, we attempt to unpack the regional interdependencies that we suspect may account for certain spillover effects across regions. Caughlin and Segev (1999) point out several reasons that could provide a theoretical support for this spatial dependence, such as the spillover of beneficial effects of FDI, labor and cost competitiveness and some geographical attributes such as proximity to port, or mountains and other infrastructure platforms. The application of the spatial econometrics, Garretsen and Peeters (2009) argue, can be a very useful tool to improve our understanding of FDI patterns. In line with similar research, the spatial lag is derived by multiplying the dependent variable by a distance matrix2 W that identifies the spatial relationship among regions. This matrix design allows nearby outcome (FDI) to have a spillover effect on the outcome of the region weighted by the distance that separates both regions. We construct the distance matrix based on the two approaches found in the literature (Villaverde and Maza, 2012, Blonigen, Davies and Naughton 2007, Coughlin and Segev 1999 and Villaverde and Maza 2012). The first approach is based on the inverse of the distance between all regions using their respective longitude and latitude coordinates. The second approach uses the contiguity matrix which in its simple form takes a binary value to reflect whether two regions share the same borders. It is equal to 1 in the presence of common borders and 0 otherwise. In this paper we only report the results from the second approach. Our spatial model can be constructed as follows: = 0 + 1 + . . + 2 The matrix W is constructed from shapefiles (maps) we have obtained over the web listing the latitudes and longitude coordinates of all 24 governorates in Tunisia. The term ρ⋅W.FDI is the spatial autoregression term. The matrix W is the spatial lag weighting matrix estimated based on the geographical contiguity between governorates and ρ is a parameter to be estimated measuring the strength and sign of the spatial relationship of FDI inflows. 5. Estimation results Table 5 presents the results of the GMM regression of regional FDI on a set of economic, institutional and geographical determinants. We note that the lagged FDI variable is positive and statistically significant in all specifications of the model confirming the high persistence effect of this variable. This result indicates that past FDI values inflows have a significant influence on current FDI activities. This finding is supported by the work of Campos and Kinoshita (2003) who point out that in presence of agglomeration effects, past FDI inflows will be a good predictor of current FDI even after controlling for other classical factors. According to the same authors, historically, FDI is found to agglomerate more often than financial investments partly because this type of investment is more of a long-term capital investment that is irreversible in the short term. The statistically significant and positive coefficient of regional GDP as a proxy for the living standard and market size of the region lends support to our hypothesis above that regions with high GDP (mainly clustered along the littoral) tend to be more attractive to foreign investors. According to recent government statistics, the size of the markets along the coastal regions is substantially larger than the rest of the country and represents about 65% of the total population of the country, with 3 out of 4 local firms being established in these regions. The proximity from key government institutions that facilitates the access to information and reduces the transaction costs for doing business, the presence of a relatively well-developed infrastructure network and the abundance of relatively well-educated labor force in the coastal regions could also explain this result. This clustering effect of FDI is not surprising. According to Walsh and Yu (2010), historically foreign firms have a tendency to gather together in specific regions for several reasons: (a) leverage production process linkages among projects, (b) benefit from an established international business climate, and (c) take advantage of the economies of scale and positive spillovers from existing investors. Regional incentives appear to have a positive influence on the inflows of FDI to the regions. Fiscal incentives (e.g. tax cuts/exemptions and other workers' legislations advantages) tend to provide attractive business operating conditions to foreign investors particularly in a vertical FDI framework where most production outputs are turned into exports (Blonigen, 2005). These incentives are often offered by local government (and sometimes sought by certain investors) to compensate for the poor investment and institutional climate. This finding is consistent with Boly et al (2019) who studied the effect of fiscal incentives on the attractiveness of FDI to a set of African countries including Tunisia. According to their results, lowering corporate tax for foreign investors increases FDI net inflows not only for the country that is carrying out the reform but also for its neighboring countries. In the case of Tunisia, it is worth noting that while historically major parts of government incentives were granted to foreign and local actors who decided to invest in the interior regions, the regional disparities in terms of attractiveness of FDI remained persistent throughout the past decades. The positive coefficient for unemployment in most specifications suggests a positive effect of the widely available young workforce (with relatively good education levels compared to other countries of the region) in attracting FDI. Similar reasoning is also found in the work of Casi and Resmini (2011) who argue that the degree of human capital development has a favorable impact on FDI inflows particularly in terms of ensuring adequate supply of skilled labor. This coefficient is however not statistically significant in various specifications to enable us to draw robust conclusions on the impact of this variable. One explanation of this result is that today (as opposed to the earlier periods of global liberalization in the 1970s and 1980s) most foreign investors are looking for qualified workers with skills that allow them to work independently and to quickly adjust to continuously changing work requirements and new technology (Campos and Kinoshita, 2002). As a result, it is not surprising to find that the higher unemployment rate in the interior regions relative to the national average (21% compared to 15% in 2018) and the abundance of a cheap but largely illiterate (or with limited schooling) workforce in those regions do not constitute a significant factor for attracting FDI inflows. One refinement for future work to better capture the dynamic of the labor market would be to adjust the unemployment variable for the labor productivity. Unfortunately, at this stage, this indicator is unavailable at the governorate level for the period under study. Similarly, the coefficient capturing the role of networks of local businesses and private enterprises is positive but not statistically significant, confirming the vertical nature of most FDI inflows in Tunisia. Foreign investors are not concerned about having a local network of domestic firms when investing in the regions. Their investments often depend on imported inputs from the country of origin and their outputs are turned into exports overseas. The regional business dynamic, in this vertical model of FDI, does not appear to be a key precondition for investors. This statistically insignificant coefficient of the number of private firms in the region confirms the results of earlier studies that find in the case of vertical FDI, that multinational firms and local industrial networks are very often two separate networks that do not collaborate or compete with each other for local production or local demand (Casi and Resmini, 2011). What is also noteworthy from the results below is the positive and significant coefficient of the Governance Index which is consistent with our hypothesis that FDI tends to be channeled toward regions that have implemented a fair level of good governance, an efficient bureaucracy and accountability practices (Campos and Kinoshita, 2003). These institutional practices are often seen by international firms as prerequisites to invest in any countries or regions. This finding is consistent with Blonigen (2005) who argues that both poor legal protection of assets and poor quality of institutions, necessary for well-functioning markets, tend to increase the cost of doing business and the time lost in dealing with local authorities and thus reduce FDI attractiveness for a given region. Several reasons can be given to explain this result. According to Walsh and Yu (2010), good governance is often associated with economic growth which remains one of the principal factors to attract FDI inflows. Further, poor governance and the resulting economic and political uncertainty motivates corruption and red tape practices that add to the cost of doing business and reduce any incentives to establish business (Campos and Kinoshita, 2003). In Tunisia, since independence in the late 1950s, and despite multiple institutional reforms led by successive governments, the development of interior regions has largely been limited by the lack of quality of public institutions as the main decision-making process and vital government agencies continued to be highly centralized around the capital and in the large coastal cities (ministries, research centers, investment promotion agencies). It is not surprising to find for instance that, as of 2017, the number of research labs in the capital city was about 165 labs and research centers. This number does not exceed 6 centers in 11 governorates of the interior regions altogether. As expected, the presence of maritime ports and other transportation infrastructure, such as highways, that are used to channel FDI-related imports and exports appears to be a significant determinant for regional FDI. This result is in line with other studies that found transportation infrastructure to have positive and statistically significant impact on FDI inflows to the regions (Caughlin and Segev, 1999). The last specification of the model used a lagged term for the regional GDP to assess the dynamic effect of the regional market on the attractiveness of the FDI. The lagged term is significant at 5% with an impact equal to about one-fourth of the contemporaneous effect of the regional GDP found in previous specifications. This result is also supported by Casi and Resmini’s (2011) work who find, in the case of local regions within European countries, that the higher the number of foreign firms and growth rate of regional GDP in the previous period in the region, the larger the number of new foreign firms that a region is able to attract. Overall, most other signs and significance remain similar to previous specifications confirming the individual effects of each determining factor. Table 5 : Results of the GMM estimation System GMM (Two Step Approach) (1) (2) (3) (4) (5) (6) FDIt-1 0.340*** 0.215* 0.269* 0.290** 0.293** 0.332** (3.60) (1.83) (1.80) (2.26) (2.17) (2.21) GDPt-1 0.339** (2.36) GDP 0.519*** 0.300* 0.321** 0.362** 0.294** (7.14) (1.85) (2.33) (2.31) (2.04) Unemployment -0.048** 0.007 0.004 -0.002 0.008 0.008 (-2.04) (0.42) (0.28) (-0.13) (0.44) (0.60) Private Firms 0.373 0.225 0.108 0.217 0.038 (0.97) (0.62) (0.27) (0.63) (010) Governance Index 0.745*** 0.532** 0.615*** 0.632*** 0.694*** (3.01) (2.11) (2.98) (3.05) (2.79) Regional Dev areas 0.261*** 0.189** 0.256*** 0.243** (3.12) (2.56) (3.16) (2.15) Port 0.763** 0.676*** 0.578*** (2.54) (2.80) (0.286) Highways 0.488** (2.12) Dummy Revolution 0.078 0.046 (0.81) (0.63) Observations 210 210 210 210 210 210 Hansen J-test 0.64 0.49 0.41 0.60 0.59 1.00 AR(1) 0.00 0.00 0.01 0.01 0.01 0.01 AR(2) 0.09 0.17 0.12 0.08 0.09 0.1 Notes: Standard errors are in parentheses. All GMM regressions use robust standard errors. For the two‐step GMM, the Windmeijer (2005) finite sample correction for standard errors is employed. *, ** and *** denote significance at the 10% , 5%‐ and 1% level, respectively. The row for the Hansen J‐test reports the p‐values for the null hypothesis of instrument validity. The values reported for AR(1) and AR(2) are the p‐values for first and second order autocorrelated disturbances in the first differences equation. Notice that across various specifications, the Hansen J-test of overidentification restrictions (with P-value > 0.1) and the Arellano-Bond test for autocorrelation (first and second order serial correlation with P-value <0.1 and >0.05, respectively) both confirm the joint validity of our instrument (not correlated with the residuals under the null hypothesis) and the consistency of our estimation. As described in the sections above and shown in the following graph, the spatial distribution of economic and institutional factor endowments in Tunisia is far from being uniform. This uneven distribution has largely influenced the regional FDI to the country which remains mainly clustered within specific geographical areas. This trend appears to persist over time with coastal regions attracting the lion’s share of the FDI in the country at the expense of the inland regions. Next, we test for this agglomeration aspect of FDI inflows to certain regions in the country. We rerun our model using a spatial lag of the dependent variable (FDI) to measure whether the variation of FDI in a given region would have any spillover effect on the FDI attractiveness in the neighboring regions (i.e. spatial effects). Regional Distribution of FDI across 24 Governorates 2008 2013 2018 Table 6 presents the results of this spatial analysis. The sign of the coefficient of spatial term confirms the positive correlation of regional FDI and thus the agglomeration effect of FDI inflows across the regions in the country. An increase of FDI inflows to a given region seems to be highly correlated with a similar trend in neighboring regions. The coefficient of the spatial lag is about 0.31 and is statistically significant at 1%. This result suggests that a 10 percent increase in the FDI into a host region tends to increase the FDI in the nearby region by little more than 3%. This result confirms our initial finding of the lagged variable in the GMM model above with respect to the agglomeration effect. This finding is also supported by the results of Hong, Sun and Li (2008) who pointed out that in countries where regional trade barriers are relatively mild, if any (as in the case of Tunisia), we should expect to see some agglomeration forces in play and thus a positive coefficient of the spatially lagged FDI variable. Barry et al (2001) examined two main reasons that could explain this result: (a) the efficiency agglomeration and (b) the demonstration effect. According to the authors, international firms have a tendency to cluster in the same or nearby regions not only to benefit from positive externalities resulting from knowledge spillovers, informal communication channels and specialized labor as stipulated in the theory of efficiency agglomeration (Campos and Kinoshita, 2003) but also to imitate each other and to rely on previous investors’ judgments and decision-making process. This reasoning is based on the signal sent by investors already established in place as to their confidence and expectations about the country or region’s outlook. This demonstration approach is particularly crucial for international firms in times of crises and uncertainty when information on the region or the country is scarce or does not reflect the long-term perspectives. In such circumstances, new investors have more incentives to follow the signal from previously established ones. Table 6 : Spatial Analysis of FDI inflows (1) (2) (3) Regional GDP 0.254* 0.252* 0.165* (1.85) (1.73) (0.70) Governance Index 0.559*** 0.560*** 0.561*** (4.40) (4.32) (4.33) Unemployment -0.001 -0.001 (-0.03) (-0.02) Private Firms 0.163 (0.47) Spatial Terms W.FDI 0.314*** 0.314*** 0.315*** (3.59) (3.60) (3.24) Observations 240 240 240 Wald test 10.50 10.46 16.50 Direct effects Regional GDP 0.259 * 0.257* 0.166 (1.86) (1.74) (0.47) Governance Index 0.571 *** 0.571*** 0.572*** (4.10) (4.45) (4.46) Unemployment -0.001 -0.001 (-0.03) (-0.0.2) Private Firms 0.167 (0.47) Indirect effects Regional GDP 0.093 0.093 0.229 (1.65) (1.58) (0.70) Governance Index 0.206 *** 0.206*** 0.780*** (2.74) (2.68) (5.03) Unemployment -0.000 -0.001 (-0.03) (-0.02) Private Firms 0.227 (0.70) Notes: Standard errors are in parentheses. *, ** and *** denote significance at the 10%, 5% and 1%, respectively. Binary variables Port and Regional areas are dropped from this fixed effect model. Overall, it is worth noting that the inclusion of the spatial term does not affect our initial results above. Notice that both binary variables Port and Regional Development areas were dropped from this fixed effect model. A general observation of note is that the direct effect appears to be more dominant for all regional FDI determinants. This finding is not surprising in the context of a vertical FDI model. As suggested in similar studies, resource-seeking investors tend to evaluate the attractiveness of destination regions based on low cost and abundance of workforce in the locality rather than in neighboring ones since most or all production output will be channeled outside the country. In such a trade framework, it is expected to find a weak (if any) spillover effect between the economic and institutional dynamic of the region and FDI inflows in the nearby ones. The statistically insignificant coefficient of the indirect effect for the market size and business network in the region provides additional support to the vertical nature of the FDI in Tunisia. Local suppliers of production inputs both in the destination regions and nearby areas are left out of this process and appear neither to benefit nor to influence the inflows of FDI in the country. 6. Conclusion In this paper we explore the relationship between key economic and institutional attributes of Tunisian governorates and their ability to attract FDI inflows. Using a dynamic GMM and spatial autoregressive approaches, we estimate a model of regional FDI over the recent period. Our results provide evidence for the regional interdependence of the FDI that appears to be highly clustered along the coastal areas. An increase/decrease of FDI inflows to a given region creates an incentive/disincentive for other FDI inflows to the same region as well as nearby ones. These agglomeration forces are relatively strong in Tunisia in the presence of vertical FDI. Further, our results indicate that a relatively developed market size, an increase of regional development areas as well as robust governance practices and infrastructure are positive determinants of regional FDI inflows. Finally, we have shown that while some of the determinants do exhibit spillover effects on nearby regions, the direct effect of the region’s own represents the bulk of the influence over the FDI inflows. Our results, though preliminary, have some policy implications. As shown above, it appears that the strengthening of regional attractiveness for the governorates located far from the coasts will necessarily require the elimination of economic and institutional barriers. These obstacles have persisted since the country’s independence in particular with respect to the size of the local markets, the limited purchasing power of the local population, the lack of skills of the workforce and the poor governance and infrastructure compared to coastal cities. The reduction of regional disparities requires the implementation of an inclusive strategy that should not be limited to ad-hoc policy actions that have long been centered around tax incentives offered to foreign investors who decide to set up their business in interior regions. Our study showed that those incentives are at best limited in scope and seem to have a positive impact only if coupled with more drastic measures. This inclusive strategy should begin by strengthening key investment drivers for the regions (e.g., social peace and security, qualified labor, reliable infrastructure and reasonable standards of living). A modern and reliable transportation system (highways, railways, airports) should also be able to support and further enhance the economic development of the regions, leading to innovation, job creation and sustainable growth. Further, the government should heavily invest in supporting the quality of regional governance by creating effective public institutions by easing complex administrative procedures and fighting corruption and red tape. Last but not least, national economic actors must be encouraged by local authorities to invest in the inland regions in order to boost the local industrial network, promote the spatial regrouping of private companies, and stimulate local consumption and sustainable growth. In doing so, they will send a strong signal to the foreign actors who understandably are less inclined to invest in these regions if the local companies themselves flee these territories. All these economic and political reforms are key drivers and important prerequisites to facilitate business operating conditions for foreign investors. The present work can certainly be further refined and complemented in future research. One potential refinement is to consider the sectoral breakdown of the FDI inflows in Tunisia and how it could influence the attractiveness and the spatial interdependence of those inflows. We propose in an upcoming paper to undertake this study. References ALTAMONTE, C. (2007), “Regional economic integration and the location of multinational entreprises ”, Review of World Economics, 143 (2), pp. 277-305. Arellano and Bover (1995), “Another look at instrumental variable estimation of error- components models”, Journal of Econometrics 68, 29-51. BAILEY, N., AND LI, S. (2015), “Cross-national Distance and FDI: The Moderating Role of Host Country Local Demand », Journal of International Management, Volume 21, Issue 4, pp. 267-276 BALASUBRAMANYAM, V. N., M. SALISU, Et D. SAPSFORD. (1999), “Foreign Direct Investment as an Engine of Growth”, Journal of International Trade and Economic Development, n° 8, pp. 27-40 Barry, Frank, Gorg, Holger and Strobl, Eric (2001), “Foreign direct investment, agglomerations and demonstration effects: an empirical investigation”, Kiel Institute for World Economy , Research Paper No. 2001, 25 BARTELS, F., NAPOLITANO, F., TISSI, N. (2014), “FDI in Sub-Saharan Africa: A longitudinal perspective on location-specific factors (2003-2010)”, International Business Review, Volume 23, Issue 3, pp. 516-529 Batana, Yele Maweki (2011), “ Analysis of the determinants of the foreign direct investment flows to the West African economic and monetary union countries:, PEP Economic Department , University of Laval , AERC Research Paper 239. BLANC-BRUDE, F., COOKSON, G., PIESSE, J., STRANGE, R. (2014), “The FDI location decision: Distance and the effects of spatial dependence”, International Business Review, Volume 23, Issue 4, Pages 797-810 Blonigen A. Bruce, Davis Ronald and Naughton Helen (2007), “FDI in Space: Spatial Autroregressive Relationship in Foreign Direct Investment”, European Economic review, Vol 51(5) Blonigen, Bruce (2005), “Review of the empirical literature of FDI determinants”, NBER Working Papers 11299. Blundell and Bond (1998), "Initial conditions and moment restrictions in dynamic panel data models", Journal of Econometrics 87, 115-143. BOATENG, A., HUA, X., NISAR, S., WU, J. (2015), “Examining the determinants of inward FDI: Evidence from Norway”, Economic Modelling, Volume 47, pp. 118-127. Boly, A., Coulibaly, S. and Kere, Eric (2019), “Tax Policy, Foreign Direct Investment and Spillover Effects”, AfDB WP Series - 310. BUCKLEY, P. J., and CASSON. M. (1981), “The optimal timing of foreign direct investment”, Economic Journal (91), pp. 75-87 Campos, Nauro and Kinoshito, Yuko, (2002), “The location determinants of foreign direct investment in transition economies”, https://www.researchgate.net/publication/228710889 Campos, Nauro and Kinoshito, Yuko, (2003), “Why does FDI go where it goes? New Evidence from the transition economies.”, IMF WP/03/228 Casi, laura and Resmini, Laura (2011), “ The spatial distribution of FDI across EuropanEuropean regions: does a country effect exist ?”, FREIT, WP. 294. CHEN, T. J., AND KU, H. Y. (2000), “The effect of Foreign Direct Investment on firm growth: the case of Taiwan’s manufacturer’s”, Japan and the World Economy, Vol. 12, pp. 153-172. CLEEVE, E., YAW, D., ZELEALEM, Y. (2015), “Human Capital and FDI Inflow: An Assessment of the African Case », World Development, Volume 74, pp. 1-14. Coughlin, Cletus and Segev, Eran (1999), “Foreign Direct Investment in China : A spatial Econometric Study”, The Federal Reserve bank of ST. Louis, WP Series 1999 -001. DORNEAN, A., AND OANEA, D. C. (2015), “FDI Territorial Distribution in Romania”, Procedia Economics and Finance, Volume 32, pp. 610-617. DUNING, J. H. (1979), “Explaining changing patterns of international production: in defence of the electic theory”, Oxford Bulletin of Economics and statistics, vol. 41, pp. 269-295. DUNNING, J. H. (1988), “The Electic paradigm of international production: A restatement and some possible extensions”, Journal of international business studies, vol. 19, pp. 1-31. ECKHOLM, K., FORSILD, R., and MARKUSEN, J.R. (2003), “Export-Platform Foreign Direct Investment”, Working Paper, NBER, n° 9517. Eunsuk Hong, Laixiang Sun and Tao Li (2008), “Location of foreign direct investment in China: a spatial dynamic panel data analysis by country of origin”, SOAS University of London , WP 86 FIPA (2017, 2016, 2015, 2014), Rapport sur l’investissement direct international en Tunisie 2016 », Foreign Investment Promotion Agency, Tunisia FIPA (2019), « Rapport sur l’investissement direct international en Tunisie 2018 », Agence de Promotion de l’Investissement Extérieur, Tunisia Juin 2019, pp. 7-14. HATEM, F. (2004), « Investissement international et politiques d’attractivité », Economica HIRSCH, S. (1976), “An International Trade and Investment Theory of the Firm”, Oxford Economic Papers, vol. 28, pp. 258-270 HUYEN, L. (2015), “Determinant of the factors affecting FDI flow to Thanh Hoa Province in Vietnam”, Social and Behavioral Sciences, Volume 172, pp. 26-33 IACE (2019, 2108, 2017, 2016), « Rapport sur l’attractivité régionale, Institut Arabe des Chefs d’Entreprises, Tunisie INS (2016), « Rapport annuel sur les indicateurs d’infrastructure », Institut national de la statistique, Tunisie INS (2016, 2013), « Répertoire national des entreprises », Institut national de la statistique, Tunisie INS (2019), « Indicateurs de l’emploi et du chômage », Institut national de la statistique, Tunisie, Août 2019 INS (2019), « Rapport annuel sur les indicateurs d’infrastructure », Institut national de la statistique, Tunisie INS (2019), « Répertoire national des entreprises », Institut national de la statistique, Tunisie ITCEQ (2016), « Bilan de l’évolution du marché de travail en Tunisie : 2006-2015 », Institut tunisien de la compétitivité et des études quantitatives, notes et analyses de l’ITCEQ, n°44. kam Oi-Yan (2018), “The Trend of Foreign Direct Investment movement: Did Unintended Nation Brand of legal-families play an instrumental role? ”, MAFRUZA. S, VIDUSHI. K, VISHAL M G, SAI PRATYUSH CH, RAJESHRI P. (2019), “Impact of FDI on Indian Economy”, Journal of Critical Review, volume 6, Issue : 6 : 115-119. MARKUSEN, J.R., and MARKUS, K.E. (1999), “Discriminating among Alternative Theories of the Multinational Enterprise”, Working paper NBER, n°7164. Ministère du Développement Régional (2011), « Livre blanc du développement régional. Une nouvelle vision du développement régional », Ministère Tunisien de Développement Régional. MICHALET, C. A. (1999), « La séduction des nations : ou comment attirer les investissements directs étrangers », Economica MUCCHIELLI, J.L.(1991), « Alliance stratégique et firmes multinationales une nouvelle théorie pour de nouvelles formes de multinationalisation », Revue d’économie industrielle, n°55 OCDE (2018), Études économiques de l’OCDE : Tunisie 2018, Éditions OCDE, Paris, pp. 120- 128. OKAFOR, G., PIESSE, J., WEBSTER, A. (2015), “The motives for inward FDI into Sub-Saharan African countries”, Journal of Policy Modeling, Volume 37, Issue 5, pp. 875-890 ONEQ (2013), « Rapport annuel sur le marché du travail en Tunisie », Observatoire national de l’emploi et des qualifications, Novembre 2013 PEGKAS, P. (2015), “The impact of FDI on economic growth in Eurozone countries”, The Journal of Economic Asymmetries, volume 12, Issue 2, pp. 124-132 Roodman, David, (2006), “How to Do Xtabond2: An Introduction to “Difference” and “System” GMM in Stata”, Center for Global Development Working Paper No. 103. SALIKE, N. (2016), « Role of human capital on regional distribution of FDI in China: New evidences », China Economic Review, Volume 37, pp. 66-84. SÁNCHEZ, M., DE ARCE, R., ESCRIBANO, G. (2014), “Do changes in the rules of the game affect FDI flows in Latin America? A look at the macroeconomic, institutional and regional integration determinants of FDI”, European Journal of Political Economy, Volume 34, June 2014, pp. 279- 299. TERAVANINTHORN, S, AND RABALLAND, G. (2009), “Transport Prices and Costs in Africa: A Review of the Main International Corridors”, World Bank, Washington, D.C. VERNON, R. (1966), “International investment and International Trade in the Product Cycle” Quaterly Journal of economics, vol. 80, pp.190-207. Villaverde, Jose and Maza, Rodolfo (2012), “Foreign direct investment in Spain: Regional distribution and determinants”, International Business Review , Vol (21), PP 722 -733. Walsh, James and Yu, Jiangyan, (2010), “Determinants of Foreign Direct Investment : A sectoral and Institutional Approach”, IMF Working Paper 10/87. WORLD BANK (2014 a), « La révolution inachevée: Créer des opportunités, des emplois de qualité et de la richesse pour tous les tunisiens », World Bank, 301-318. WORLD BANK (2014 b), Tunisian Urbanization Review: Reclaiming the Glory of Carthage. World Bank 2014 WORLD BANK (2020), “Doing Business 2020 : comparing Business regulation in 190 Economies ”, Washington D.C : World Bank, pp. 4-12 XING, Y. (2006), “Why is China so attractive for FDI? The role of exchange rates” , China Economic Review, Volume 17, Issue 2, pp. 198-209 YANG, C. H. (1999), “Determinants of Foreign Direct Investment: comparison between expansionary FDI and Defensive FDI”, Taiwan Economic Review, Vol. 27, pp. 215-240.