Portuguese Direct Investment Abroad and the political orientation of governments.

FACULTY OF ECONOMICS - UNIVERSITY OF PORTO Portuguese Direct Investment Abroad and the political orientation of governments José Daniel Alves Oliveira [email protected] Master in International Business Supervised by Ana Paula Africano de Sousa e Silva Porto, Portugal 2015 i Biographical note José Daniel Alves Oliveira was born in 1975, being resident, by the time of adulthood, in a small town called Cortegaça. The first son of José and Alice soon showed a special interest about the things around him, questioning and experiencing parent’s activities. The curiosity to learn and his dedication for perfection made him the best student in elementary school. After this period, there were some drawbacks in his life, which made him start working at the age of 16. Even though, his persistence and strong will to finish high school, led him to the final average grade of 14. Due to personal difficulties, his college dream had to be postponed. At the age of 19 years old, he joined the multinational Philips in Ovar and then a great life experience began. In this period he had the opportunity to exchange experiences with people and companies from different countries, allowing him to look at the world as something global. After 15 years of strong commitment and although Philips Ovar was considered a successful business example, the effects of the 2008 crisis forced the company to close permanently this subsidiary. With the company closure due to market difficulties in providing work to a person with experience but no college graduation, going to university started to be a high motivation. Daniel joined ISAG in 2009, where he met a group of teachers and students with wide professional knowledge. It was with distinction that he graduated in Business Management in 2012 with Grade 16 During this time, he got a job in the Casa das Lâmpadas as a product manager. Going from a multinational to a family business, gave him the opportunity to come across very dedicated and clever professionals, allowing him to grow his trade skills. Still keeping contacts with multinational companies, he is responsible for drawing up plans for the Portuguese market in the representation of their products. His multinational experience added to his interest in global economy and looking forward academic recognition, were reasons to join the Master's course in Economics and International Management at the Faculty of Economics, University of Porto. Nowadays, being able to finish this stage, one can come to the conclusion that dreams are possible to be fulfilled and achieved with perseverance, honesty and dedication. ii Acknowledgments During the preparation of this work, I had the support and contribution of several people who helped me to finish it. Therefore, I would like to give a word of appreciation and thanks to all those who somehow got involved and were involved in this journey. To my supervisor, Ana Paula Africano, thank you for your time, dedication, insight, suggestions, criticism and experience that allowed me to complete and reach the very end of this work. To my coordinator course, Rosa Forte, who supported me in critical moments, guided me and contributed to the success of this work, many thanks. To Miguel Fonseca for the Dissertation theme suggestion. Thank you Pedro Oliveira and Paulo Guimarães for your availability and contribution. To Silvia Almeida, for her professionalism and cooperation. To Brizida Tomé for her friendship To my family, thank you for all your concern, love and encouragement. To my wife Paula and my daughter Leonor, very special thanks for your understanding, support, warmth, joy, love ... I love you. iii Abstract Globalisation has been integrating markets across the world. The Multinational Enterprises (MNEs) look at the global market and they decide to invest in those countries that present the best conditions to increase their profits. MNEs' Investments have been associated with a stronger social and economic development in the receiving countries, and for that reason governments formulate policies to increase local attractiveness. But besides the advantages of receiving foreign investments, Governments are also keen to help local Enterprises to invest and expand in foreign markets. Giving such support, they look for benefits from profits and knowledge through reverse spillovers and political interests of home country. Outward Foreign Direct Investment is seen as an opportunity, and many Governments aim to increase their interventionism in supporting local companies’ internationalisation. Previous studies have identified several determinants that influence Outward Foreign Direct Investment (FDI) intensity. Economic dimensions, historical and cultural links, and geographic distance between home and host countries are among the most commonly used. Some government policies are also important for MNEs' investment decisions such as tax policy, investment treaties and openness to foreign investment. Besides that, it was found that political stability has a positive impact on investor’s decision, as it reduces risk. However, the existing literature has somehow overlooked the impact that governments' political orientation may have on Outward FDI. The political orientation of governments may interfere with the decisions of Multinationals concerning Outward Foreign Direct Investment, because the right wing parties and the left wing parties have different policies approaches. Apart Schneider and Frey (1985) and to the best of our knowledge there are no other studies about this subject. Focusing on the Portuguese case, the main objective of this Dissertation is to scrutinize if the Outward FDI can be influenced by Government political orientations, in terms of quantity and geography. The study considers a sample of 13 countries for the period between 1996 and 2013, and uses a gravity equation for the Portuguese outward FDI and a panel data analysis. Specific variables are added to test the impact of governments' political orientation. The iv results show that Portuguese Outward FDI is influenced by market dimension, language, political risk, Portuguese Government ideology and also by specific relations with Brazil, Netherlands and Spain. This research provides the first evidence that, in the studied period, right wing governments had a positive impact on Portuguese outward FDI. Yet, in contrast with other studies, political risk has a positive impact on Outward FDI, which may be explained by the particular Portuguese relations with Angola and Brazil. Keywords: Outward Foreign Direct Investment, Political economy, Political orientations, Portugal. v Resumo A globalização tem proporcionado a integração dos mercados a nível mundial. As empresas Multinacionais analisam o mercado global e tomam decisões de investimento em países que apresentam as melhores condições para o aumento dos seus lucros. Os seus investimentos têm sido associados a um desenvolvimento social e económico mais forte nos países de acolhimento e, por essa razão, os governos formulam políticas que visam atrair o investimento. Para além das vantagens de obter investimento estrangeiro, os governos têm também interesse em ajudar as empresas locais a investirem e expandirem em mercados estrangeiros. Facultando esse apoio, obtêm benefícios, dos lucros e do conhecimento através de externalidades (spillovers) internacionais e interesses políticos no país de origem. O Investimento Direto Estrangeiro no Exterior é visto como uma oportunidade, e muitos governos têm como objetivo aumentar o seu intervencionismo no apoio à internacionalização das empresas locais. Estudos anteriores identificaram vários determinantes que influenciam a intensidade do Investimento Direto Estrangeiro no Exterior. Dimensões económicas, laços históricos e culturais, e distância geográfica entre o País emissor e recetor estão entre os mais utilizados. Algumas políticas governamentais também são importantes para as decisões de investimento das Multinacionais, como a política fiscal, os tratados bilaterais para o investimento e a abertura ao investimento estrangeiro. Além disso, verificou-se que a estabilidade política tem um impacto positivo na decisão do investidor, uma vez que reduz o seu risco. No entanto, a literatura existente de alguma forma negligencia o impacto que a orientação política dos governos pode ter sobre Investimento Direto Estrangeiro. A orientação política dos governos podem interferir com as decisões das Multinacionais em matéria de investimento direto no estrangeiro, porque os partidos de direita e os de esquerda têm diferentes abordagens políticas. Para além de Schneider e Frey (1985), e considerando o melhor do nosso conhecimento, cremos que não existam outros estudos sobre o tema. Centrando-se no caso Português, o objetivo principal desta dissertação é analisar se o Investimento Direto Estrangeiro no Exterior pode ser influenciado por orientações políticas do Governo, em termos de volume e geografia. O estudo considera uma amostra de 13 países para o período entre 1996 e 2013, e usa uma equação gravitacional para o Investimento Direto Estrangeiro no Exterior vi Português e uma análise de dados em painel. Variáveis específicas são adicionadas para testar o impacto da orientação política dos governos. Os resultados mostram que o Investimento Direto Estrangeiro no Exterior é influenciado pela dimensão do mercado, língua, risco político, a ideologia do Governo Português e também por relações específicas com o Brasil, Holanda e Espanha. Esta pesquisa fornece a primeira evidência de que, no período estudado, os governos de direita tiveram um impacto positivo sobre a Investimento Direto Estrangeiro no Exterior Português. No entanto, em contraste com outros estudos, o risco político tem um impacto positivo sobre a Investimento Direto Estrangeiro no Exterior, o que pode ser explicado pelas relações portuguesas particulares com Angola e Brasil. Palavras-chave: Investimento Direto Estrangeiro no Exterior, Economia Politica, orientação Politica, Portugal. vii Table Index Table 1 – FDI Outward Determinants Literature Review: A Synthesis ......................... 19 Table 2 – BPM6 and BPM5 methods comparison ......................................................... 26 Table 3 - Portuguese Outward variation between 2000 and 2011 (BPM6 and BPM5 comparison) .................................................................................................................... 28 Table 4 - Portuguese Governments between 1995 and 2014 .......................................... 33 Table 5 – Variables definition, expected sign and data source ....................................... 40 Table 6 - Variables descriptive analysis ......................................................................... 43 Table 7 - Variables correlation ....................................................................................... 48 Table 8 - Econonometric estimation results - Random Effects ...................................... 51 Table 9 – Econometric estimation results - Countries and Years regressions ................ 54 Graph Index Graph 1 - Portuguese FDI (Inward - Outward) BPM6 vs. BPM5 (1996 and 2014) ....... 26 Graph 2 - Portuguese Outward FDI according to BPM6 1996-2014 ............................. 27 Graph 3 - Portuguese Outward FDI by Countries 1996-2013 ........................................ 36 4 SECTION 1 - Literature review and hypotheses This section reviews the literature about FDI and Outward FDI, as well as the relation of Politics to Economy and FDI. Considering the theoretical explanations it will be formulated hypotheses that are also based on tested empirical applications. 1.1. Foreign Direct Investment (FDI) The definition of FDI is referred in the Balance of Payments Manual (IMF, 1993, p. 86) as being a “significant influence that gives the investor an effective voice in management foreign-controlled sector, the primary distinguishing feature is control.”. Furthermore, “direct investor, who is resident in another economy, owns 10 percent or more of the ordinary shares or voting power (for an incorporated enterprise) or the equivalent (for an unincorporated enterprise). Direct investment enterprises comprise those entities that are subsidiaries (a non resident investor owns more than 50 percent), associates (an investor owns 50 percent or less) and branches (wholly or jointly owned unincorporated enterprises) either directly or indirectly owned by the direct investor”. The motivations to firms internationalize are related to the fact that companies have specific advantages which can be exploited in foreign markets. Despite their costs, benefits are greater than the investment. Imperfect markets are one of the reasons for the focus on internationalization, as it allows overcoming barriers of varied nature (Hymer, 1976). The Hymer’s study reinforces the conclusions of Dunning (1973) about the determinants of international production. Internationalization by FDI is according to Dunning (1988) determined via the existence of three types of conditions. The company is expected to have Ownership, Location and Internalization advantages. This is the so called Eclectic Paradigm or OLI paradigm introduced on Dunning’s study in 1977 “Trade, location of economic activity and the MNE: A search for an eclectic approach”. Through the nature or nationality of firms, internationalization occurs by taking advantages of ownership. Sometimes these advantages are related to monopolistic or competitive firms (Dunning, 1988, p. 2). Due to location advantages, manufacturing 5 production tends to be concentrated in regions. This is associated with lower cost of transports, stronger scale economies and share of spending manufactured products, so firms do FDI to gain market advantages (Dunning, 1988, p. 4; Krugman, 1990). Companies look for additional efficiency and increase competitiveness, which could be aimed by internalizing or externalizing activities (Buckley & Casson, 1998, p. 541; Princen, 1997) . Looking at the main proposals of FDI theory, Buckley et al. (2007, p. 500) says “ (1) firms internalise missing or imperfect external markets until the costs of further internalisation outweigh the benefits; and (2) firms choose locations for their constituent activities that minimise the overall costs of their operations”. Based on this, Dunning and Lundan (2008) identify four motivations for internationalization, thus MNEs look for foreign markets based on the following motivations: market seeking, efficiency seeking, resource seeking and strategic asset seeking.  Market seeking FDI Dunning and Lundan (2008) suggest that MNEs seek out for new markets by FDI because e.g. sales potentials corresponding to the size and growth of foreign markets and the possibility to access more markets.  Efficiency seeking FDI FDI it is important to e.g. rationalize structures of supply and distribution, explore gains of economies of scale or risk diversification (Dunning & Lundan, 2008). Concerning risk, Rugman (1979) refers that by diversifying FDI, MNE’s can reduce the impact of market imperfections in relation to e.g. interest rates, wage earnings, share price indexes.  Resource seeking FDI Dunning and Lundan (2008) say that MNEs invest abroad in order to access cheaper raw materials, natural resources, low cost operations, or attainment of low cost of labour.  Asset seeking FDI 6 Additionally Dunning and Lundan (2008) consider as long-term strategies, essentially acquisitions and joint ventures. In this way, MNE can overcome needs of special assets, such as knowledge (R&D), as resources in unfamiliar markets or acquisition of particular information in markets.  Other motivations to FDI Dunning and Lundan (2008) also refer that MNEs may use FDI in order to avoid country of origin restrictive legislation of governments (“escape investments”), or with the purpose of supporting other multinational activities (“support investments”) or even as an indirect result of mergers and acquisitions (passive investments”). 1.2. Political economy view of outward FDI “Political economists have long argued that the interaction between businesses and governments is a complex, dynamic, and interdependent process in which governments create the rules by which businesses must abide, while businesses make efforts to shape governmental policies” (Luo et al., 2010, p. 69). Rugman (1998) referred that the government controls, regulates and judges business, by creating legislation establishing a regulatory environment. Although the main objectives of governments are to increase social welfare (Helpman & Krugman, 1985; King & Levine, 2004; Levine, 1997), yet, government’s interference may create inefficiency and be counterproductive in markets (Edwards & Keen, 1996; Inman & Rubinfeld, 1996), otherwise it can pull over the economy efficiency with the introduction of experience (Charoenrat & Harvie, 2014; Ng & Gujar, 2009). Even though the different positive or negative opinions about political interference, from MNE perspective, home or host country governments may have a special role in its business strategy development. The World Trade Organization (WTO), with its principles of trade liberalisation, had in 2014 the representation of 160 governments. Such large number of countries does not mean there is no influence in the market by the elected representatives. Since the creation of WTO, traditional barriers, such as tariffs, have decreased to the level of 7 3.8% in industrial products1. Nevertheless, according to Rickard and Kono (2014, p. 350) home governments tend to protect local companies, implicitly, as “treaty violations are difficult to observe and prove” which means that “international agreements are largely ineffective in constraining national governments”. In their study, is referred that government procurement in developed countries it amounts to 15% to 20% of GDP, which is interesting to local and foreign companies business. Despite the existence of international procurement agreements, that aim to foresee an open market, this does not have a strong impact on the government’s actions/decisions to pursue the protection of local companies. This approach may spoil the development of the local economy, as it impacts negatively on its progress by reducing imports or FDI. Concerning economic advantages, Lall and Narula (2004, p. 448) postulate that liberalisation towards FDI openness “as a means of acquiring technologies, skills and access to international markets, and of entering dynamic trade and production systems internal to multinational enterprises (MNEs)”. Chakrabarti (2001) says that openness has a strong correlation to the level of FDI and by inherence the development of economy, referring that “policy makers, intent on increasing FDI, to increase participation in international trade”. There are evident advantages having an open market that will impact on the development of economy. Such result derives from receiving FDI into home market, but also by the increase of domestic MNEs knowledge outwards, which means that governments look at FDI as an important issue in their policy. Knoerich (2012) has referred from the point of view of Outward FDI as being positive to home countries progress. In this study about the development of Outward FDI, he says that there are good indicators about the impact that investment done external has into the home economy, via the inflow of profits, but also by the improvement on knowledge and efficiency via reverse spillovers. Although the first perception of Outward FDI is benefits loss, because of capital outflow, the increase means that home economy is in a high level of development. Such idea is largely identified as the 1 e.g from 1995 to 2000 it was cut 40% on tariffs for industrial products (WTO, 2015) 8 Investment Development Path of Dunning’s theory (Dunning, 1981). This theory proposes that companies invest abroad when their home economy is enough developed which relates Outward FDI to a condition of economic development (Dunning & Narula, 1996 : cited by Castro 2000). The importance that economic growth has to the home country concerning social and economic issues it cannot be ignored. It means that governments look at the Outward FDI as a significant subject to be considered in their institutional policies (Steven Globerman & Shapiro, 2002; Luo et al., 2010). The authors have studied the FDI flows (inward and outward) of developing and developed countries, concluding that there is a strong relation between good political governance and economic development. Policies promoting competition, transparency and good institutional governance increase inward FDI, and by inherence promote the appearance and growth of home country MNC, increasing Outward FDI. Concerning the China case, Luo et al. (2010) found that the government has an important role in promoting Outward FDI. The authors refer that the Chinese government have implemented several measures, such as, financial and taxation special policies, risk safeguard mechanisms, information service network and direction guidance of Outward FDI in order to help Chinese MNCs to overcome firm competitive disadvantages in foreign markets. We can conclude that the support a home country government can have to the business development of MNC, should not be ignored. There are mutual interests concerning Outward FDI. From the government point of view, the issue is economic development and from the MNE point of view is profit increase. 1.3. The determinants of Outward FDI: some hypotheses from the theoretical and empirical literature 1.3.1. Market size and Growth rate One of the determinants that it is generally associated as significant to FDI is market size. In the study of the determinants of FDI, Chakrabarti (2001) found in his survey about FDI studies, strong evidence that higher GDP per capita, referred as market size, 9 has a positive relation to FDI. Same conclusions can be found in the study of Blonigen and Piger (2014) about determinants of FDI concerning parent-country GDP. In their study, it was found positive relation of FDI level to the fact of home and host countries have similar levels of GDP. Besides that, host country GDP level has a confirmed effect on the attraction of FDI, considered by the authors that “wealth of the source country is a key determinant of FDI” (Blonigen & Piger, 2014, p. 800). Buckley et al. (2007) by consistent empirical support, show that Chinese Outward direct investment flows depend upon the GDP levels and the rates of GDP growth in host countries. We may conclude that depending on home and host countries GDP indicators performance, firms would do investments where economy expansion is expected to be stronger. This means, that if a home economy has better indicators, it will attract investment of local firms and foreign ones. On the contrary, MNEs would invest outwards if the home economy growth has worse performance than foreign countries. Hypothesis 1a: Host country GDP per capita has a positive effect on Outward FDI Hypothesis 1b: Host country Growth rates has a positive effect on Outward FDI 1.3.2. Natural Resources Big countries have in their FDI objectives, the control of natural resources in order to decrease the dependency on such important assets like Oil – this result was confirmed for Russia (Kalotay & Sulstarova, 2010) and China (Buckley et al., 2007). Concerning Outward FDI of Chinese MNEs, there is a strong evidence in the study of Kolstad and Wiig (2012) that there is a positive relation of investment in countries with institutional problems but with high level of natural resources. These countries are in the world economy those that are considered as the poorest. This paradigm is explained as “The curse of natural resources”, where “resource abundant countries tended to be high-price economies and that, partly as a consequence, these countries tended to miss-out on export-led growth” (Sachs & Warner, 2001). It means that poor countries are likely to be the ones with high levels of natural resources but with lower levels of GDP growth, on the contrary richest countries have lower levels of natural resources but high levels 10 of GDP. In order to control natural resources, big countries do investments in poor countries, by means of outward FDI. Hypothesis 2: Natural resources on host countries have a positive effect in Outward FDI 1.3.3. Taxes and Tariffs The study by Kemsley (1998) which has analysed the effect of taxes rate and foreign tax credit incentives to the local U.S.A MNEs, concluded that MNEs outweigh costs of exporting instead of producing in foreign countries, using relatively more exports instead foreign production to deliver goods to high-tax foreign markets. Depending on the taxes incentives of home or host countries, MNEs define whether to export or produce outwards. It was found strong evidence that U.S.A. MNEs do exports to countries with high-tax instead of foreign production. In terms of after-tax profits, according to Grubert and Mutti (1991) MNEs “have the incentive to increase their after-tax profits by shifting taxable income from affiliates incorporated in high-tax countries to subsidiaries in low-tax countries”. By empirical analysis, it was concluded that MNEs shift their income profits to countries with very low tax, which is a solid reason for MNEs to divert FDI into these countries. Moreover, statutory tax rate have more impact than effective tax rate, meaning MNEs decisions on FDI are further related to the levels of direct tax applied to the performance of MNEs. Further study about taxes influence on investment outwards was developed by the same authors Mutti and Grubert (2004). They analysed the taxes effects of 47 countries on U.S.A. Outward FDI during the 1982, 1989, and 1994 periods. By the analysis of 728 parent firms affiliate location choice, it was concluded that taxes have a strong effect on their investments option. Nevertheless, the effect is less noticed in countries with high level of GDP. Concerning this conclusion, the authors say “taxes are less of a deterrent to location in high income economies, perhaps because they offer better infrastructure, agglomeration benefits, or a uniquely attractive market opportunity. Those high-income countries are better able to ignore competitive pressures to cut their own taxes”. More recently, Barrios, Huizinga, Laeven, and Nicodème (2012) have made a study about the same implications referred by Mutti and Grubert (2004), taking 33 European countries MNEs during the 1999-2003 period. The results point out a strong negative 11 effect of taxes in MNEs decisions, even in case that parent country taxation can be deferred until income is repatriated. Concerning to tariffs, Levy and Nolan (1991) referred that neither uncontrolled tariffs nor restrictions on FDI will increase benefits to the home markets, as protecting home firms could reduce its competitiveness and country welfare. Situations of monopoly, depending on tariffs and restrictions level, equilibrium at the market structures could be possible. This can be achievable when a home market has monopoly companies. Foreign companies having conditions to enter into home market will compete, increasing competitiveness and social welfare. In case of foreign companies be monopoly, entering into the home country market and replacing a local monopoly company, it will enhance benefits and knowledge of home economy. If a foreign company monopoly can substitute a home country monopoly, it is because is more efficient and has higher know-how. Grubert and Mutti (1991) say tariffs have high impact on the decisions of FDI. This is due to the fact that exporting to countries with high tariffs, will increase goods price, what means the products can be ignored by the consumer because of their price. The way to overcome this barrier, and increase sales on markets with high tariffs, is by doing investment in a company at this market by means Outward FDI. About tariffs effects on European Union MNEs investments choice in foreign countries, Cardamone and Scoppola (2015) analysed France, Germany, Netherlands, Italy and United Kingdom Outward FDI made in 24 partner countries during the period of 1995- 2008. Industry levels were disaggregated in order to understand if there were different impact effects on horizontal and vertical FDI. It was concluded that tariffs applied by European Union countries to products that are re-imported, vertical FDI, have a negative impact on Outward FDI. Nevertheless, concerning horizontal FDI, it is clear that tariffs have a positive effect on Outward FDI. Hypothesis 3a: Lower taxes level in host countries than in home country have a positive effect on Outward FDI Hypothesis 3b: Tariffs in host countries have a positive effect on Outward FDI 12 1.3.4. Openness and Incentives Findlay (1978) postulate that FDI increase the pace of technical progress in the host country, being more common having more advantages receiving new technology than lending it. This idea is corroborated by the study of Liu and Wang (2003), having as an evidence that the increasing of R&D level and firm size where the most important factors to enhance productivity in Chinese industries. Barrell and Pain (1999) refer that FDI is directly linked to process growth and higher integration of the European countries and Görg and Greenaway (2004) refer that FDI is a key driver of economic growth and development. Being FDI so important, governments offer incentives to attract this type of investment, motivated by the expectation of spillovers benefits in order to enhance national income from new investments (Görg & Greenaway, 2004). Attractiveness to FDI of the host countries may be influenced by government policies in a wide multiplicity of ways, indirect and implicit or explicit (Steven Globerman & Shapiro, 1999). Also a home country government can be a powerful ally to MNEs, offering diverse institutional tools, such as fiscal incentives and funds, creating conditions for firms’ development. Increasing capabilities, MNEs will be prepared to compete abroad, what means Outward FDI growth. (Luo et al., 2010). Given the above reasons, MNEs and government have mutual interests and it means that flows of FDI also depend on willingness and capacity to attract FDI. Hypothesis 4: Openness and Political incentives have a positive effect on Outward FDI 1.3.5. Geographical and Cultural Distance Buckley and Casson (1976), postulates that FDI by MNEs exists because internalisation benefits are higher than costs and this process would continue until the reverse occurs. According to Uppsala School (Johanson & Vahlne, 1977, 1990; Johanson & Wiedersheim‐Paul, 1975) the involvement that firms arise in the internationalisation process, may be smaller or larger, depending on internal and on external factors. The internal factors mean those the company holds, such as the perception of competitive advantages over its competitors, management strategies and marketing or the erosion of results. The external factors are the progress of its customers or competitors in the 13 international market, government policies or bank financing. Through the above reasons, companies may choose for a gradual process of internationalization, first export, after through licensing schemes and finally with a higher level of commitment by FDI. However, depending on the ability of companies and their degree of involvement, this evolutionary process may not have this progression, and sometimes can be overcome by starting on a higher level (FDI). The influence of physic distance (e.g. cultural and historical) affects the decisions on internationalisation, and also at high level, on FDI.  Geographical distance Geographical distance has impact on the decisions of FDI, as managing an affiliate at a higher distance has greater costs (Drogendijk & Martín Martín, 2014; Grosse & Trevino, 1996; Siegel, Licht, & Schwartz, 2013).  Cultural Distance Cultural distance has a strong negative impact on FDI (Grosse & Trevino, 1996). According to Siegel et al. (2013), differences on egalitarianism affect negatively MNEs on FDI destinations. The same conclusions for Spain are referred by Drogendijk and Martín Martín (2014).  Past colonial Ties In the study by Siegel et al. (2013) about FDI relation to egalitarianism and cultural distance, although with low relevance, there is always a positive effect with several determinants of FDI if exists common languages and colonial heritage. Head and Ries (2008) refer there is a positive relation to FDI when exist past colonial ties taking into consideration the relation between 30 OECD and 32 partner countries. In the particular case of Southern African countries, Mhlanga, Blalock, and Christy (2010) found that colonial ties have an important influence on FDI, which are the examples of English and Portuguese language countries, where their former colonizers have larger volume of investment. Hypothesis 5a: Geographical distance has a negative effect on Outward FDI Hypothesis 5b: Cultural inequality has a negative effect on Outward FDI 20 Regression analysis + Sachs and Warner (2001) Taxes 3a Pooled ordinary least squares (POLS) - Kemsley (1998) Two stage least square (2SLS) - Grubert and Mutti (1991) Pooled ordinary least squares (POLS);0 Random effects (RE) - Mutti and Grubert (2004) Pooled ordinary least squares (POLS) Timed Fixed Effect - Barrios et al. (2012) Tariffs 3b Two stage least square (2SLS) + Grubert and Mutti (1991) Pooled ordinary least squares (POLS); Fixed effects (FE) + Cardamone and Scoppola (2015) Openness and Incentives 4 Regression analysis + Barrell and Pain (1999) Literature review + Görg and Greenaway (2004) Ordinary Least Squares, Regression analysis + Steven Globerman and Shapiro (1999) Empirical analysis + Luo et al. (2010) Geographical distance 5a Case study and empirical analysis (Johanson & Vahlne, 1977); Johanson and Vahlne (1990); (Johanson 21 & Wiedersheim‐Paul, 1975) Partial least squares-based structural equations modelling - Drogendijk and Martín Martín (2014) Ordinary least squares (OLS); Random effects (RE) - Grosse and Trevino (1996) Ordinary least squares (OLS) regression - Siegel et al. (2013) Cultural inequality 5b Case study and empirical analysis (Johanson & Vahlne, 1977); Johanson and Vahlne (1990); (Johanson & Wiedersheim‐Paul, 1975) Ordinary least squares (OLS); Random effects (RE) - Grosse and Trevino (1996) Ordinary least squares (OLS) regression - Siegel et al. (2013) Partial least squares-based structural equations modelling - Drogendijk and Martín Martín (2014) Common language 5c Ordinary least squares (OLS) regression + Siegel et al. (2013) Fixed effects (FE) + Head and Ries (2008) Pooled ordinary least squares (POLS); Fixed effects (FE) + Mhlanga et al. (2010) 22 Exports 6a Cross-section one with a OLS estimation Insignificant Africano and Magalhães (2005) Random effects (RE) - Fonseca et al. (2009b) Imports 6b Cross-section one with a OLS estimation Insignificant Africano and Magalhães (2005) Democracy 7 Pooled ordinary least squares (POLS); Fixed effects (FE) + (Political stability associated to democracy) Roe and Siegel (2011) Pooled ordinary least squares (POLS) + Jensen (2003) Bilateral Investment Treaties 8 Pooled ordinary least squares (POLS) + Zhang et al. (2014) Empirical analysis + Pavlínek (1998) Survey analysis + Vale (2004) Case study and survey analysis Insignificant Fornes and Butt-Philip (2011) Case study and survey analysis Carlos Sosa Varela et al. (2014) 23 Empirical analysis + Jenkins (2012) Pooled ordinary least squares (POLS); Fixed effects (FE) + Desbordes (2010); (Desbordes & Vicard, 2009) Fixed effects (FE) + Neumayer and Spess (2005) Government Ideology 9a Pooled ordinary least squares (POLS) + but Insignificant Schneider and Frey (1985) 9b Random effects (RE) + but Insignificant S. Globerman et al. (2006) Source: The author based on literature review 24 SECTION 2 - Outward FDI in Portugal: descriptive analysis According to Castro (2004, pp. 6, 7) Outward FDI started to have more relevance from the 1990’s in Portugal. In order to understand this development, an analysis on the Portuguese Investment Development Path will be considered, taking as reference Castro’s study. Although the Investment Development Path theory interprets both Inward and Outward FDI to evaluate the Economy Development level, there is a high relation on country progress to high Outward FDI levels. From the moment of this study, a change occurred in the Balance of Payments data, and based on that, a comparison among methods (BPM5 and BPM6) is given to understand Outward FDI differences. Moreover, it is analyzed the Portuguese Outward FDI evolution. Considering the main question of this study, if Governments ideology influence on Outward FDI, it will be described the main FDI Governments programmes during the period between 1996 and 2013. During this period several measures were implemented, nevertheless, not always aiming to increase Outward FDI. Based on the literature review (Section 1) and Portuguese economy characteristics, a regression model is defined. Afterwards, it is described the variables as well as the data source. Before going to the empirical analysis (Section 3), it will be made a descriptive examination as well as a correlation analysis between variables. 2.1. Portuguese Investment Development Path The Investment Development Path of Dunning’s theory, introduced by (Dunning and Narula (1996): cited by Dunning 2001) and subsequently revised by Dunning (2001); Dunning and Narula (1996), it indicates that an investor country with high level of Outward FDI, means that it is more developed. A country with relevant Outward FDI indicates that MNEs reach a high level of ownership advantage what may lead to increase predominance in foreign markets. This theory identifies five stages of development; “Stage 1 is associated with pre industrialization; stage two the development of some location specific advantages namely by government policies; stage three is associated with less spectacular growth rates of inward FDI eventually 25 overtaken by outward FDI; stage 4 when countries turn into net outward investors and finally stage 5 leading developed countries with permanently high stocks of both inward and outward FDI” (Dunning, 2001, pp. 181, 182). Analysing the Portuguese case, according to Castro (2004, p. 80) stage 1 occurred until 1960. Transition to Stage 2, seemed to materialize from middle of 1960’s, nevertheless, it was not completed until the 1980’s. Stage 3, is reached at the middle of 1990’s. However, Úbeda (2000, p. 151) and Dunning and Narula (2003, p. Chapter 7) referred that Portugal had reached the Stage 3 earlier, at the end of 1980’s, since 1995, Portugal has increased Outward FDI exponentially, which seems an entering in Stage 4. On the other hand, a net outward stock of Portuguese FDI conceals deterioration in the country’s attractiveness as a location of foreign investment, which means the Portuguese market became economically less interesting (Castro, 2004, p. 81). According to Dunning (1981, 1986, 2001), to achieve Stage 4 a country needs to have high levels of Inward and Outward FDI simultaneously. Castro (2000, p. 30) says “The investment development path suggests an association between a country’s level of development (proxied by GDP per capita) and its international investment position (net outward FDI stock per capita). The basic hypothesis is that, as the country develops, the conditions facing domestic and foreign companies change. This will have an impact on the flows of inward and outward FDI”. Since the study by Castro, concerning the Portuguese economy, there are some studies about this theme by (Fonseca, Mendonça, & Passos, 2007, 2009a). Although they found relations between economic development level to Inward and Outward FDI intensity, the study had some limitations concerning development stages identification. The above studies about the Investment Development Path were based on data according to the Balance of Payments Manual 5th Edition (BPM5) method which have been released in 1993. In 2009 the IMF has released the Balance Of Payments Manual 6th Edition (BPM6) and since 2014 the European countries have implemented this methodology in their national accounts (European Comission, 2012). Concerning FDI there are significant differences between methodologies. BPM5 measures the FDI on a directional basis, instead BPM6 on gross assets and liabilities source. Besides this difference there is an interpretation concerning fellow enterprises, which is now explicit 26 with BPM6. Although both methodologies do not change the FDI Balance, total Inward and Outward FDI have changes (IMF, 2015, p. 126) Taking into account the Portuguese 2014 FDI data, in Table 2 we can see the differences between new and old methods concerning Assets/Outward and Liabilities/Inward FDI stocks in Million Euros. Table 2 – BPM6 and BPM5 methods comparison BPM6 BPM5 New/Old (%) Direct Investment Assets / Outward FDI 75.239 48.065 157% Direct Investment Liabilities / Inward FDI 116.553 89.379 130% Net 41.314 41.314 100% Source: Own calculations based on Banco de Portugal (2015) Considering the new procedures on national accountancy, some conclusions in previous studies can be questioned nowadays. A graph with the Inward and Outward FDI stocks from 1996 to 2014 using both methods is illustrated bellow. Graph 1 - Portuguese FDI (Inward - Outward) BPM6 vs. BPM5 (1996 and 2014) Source: Own calculations based on Banco de Portugal (2015) 27 2.2. Portuguese Outward FDI Even if the results of Castro and Fonseca studies may have some differences due to the period that they covered and to the Balance Of Payments Manual change, it is patent that Outward FDI level indicates the development of countries, in line with Dunning’s Investment Development Path theory. Simões and Cartaxo (2012) have made a study concerning Outward FDI from Portugal and its policy context. The authors analyzed the Portuguese Outward FDI stocks and flows during the period in between 2000 and 2011. As referred by the authors, Portugal more than tripled between 2000 and 2011 FDI stocks2. In Graph 2 it is visible the evolution of Portuguese Outward FDI. Graph 2 - Portuguese Outward FDI according to BPM6 1996-2014 Source: Own calculations based on Banco de Portugal (2015) 2 Concerning to this consideration, their analysis was based in US Dollar data; which in case of using Euro, due to exchange rates, the increase is not so significant, being from 20.867,60 Million Euro to 50.055,99 Million Euro (see graph 2) 28 Although the difference between BPM5 and BPM6 methods, concerning to stocks variation, there is not a significant change. This means the author’s FDI stock interpretation is adequate with the new method, as it is visible in Table 3. Table 3 - Portuguese Outward variation between 2000 and 2011 (BPM6 and BPM5 comparison) BPM6 BPM5 Direct Investment Assets / Outward FDI (2000) 26.115 20.868 Direct Investment Assets / Outward FDI (2011) 63.749 50.056 Variation % 40,97% 41,69% Source: Own calculations based on Banco de Portugal (2015) Simões and Cartaxo (2012) with reference to Outward FDI flows between 2000 and 2004, say that the average was more or less stable, instead between 2005 and 2009 it was on a declining. In 2007-2008 the deterioration of Outward FDI is associated by the authors to the global financial and economic crisis where Portuguese companies froze international investments. Between 2010 and 2011, according to the authors there was an increase of Outward FDI flows, relating it to the signing of the “Memorandum of Agreement with the International Monetary Fund, the European Central Bank and the European Commission, which enabled the granting of a financial rescue program to Portugal, several Portuguese companies undertook moves to strengthen their positions abroad to improve access to international funding” (Simões & Cartaxo, 2012, p. 2). According to the authors, the fact that Portugal has had several social problems due to the world economic crisis, in 2008, and to the financial rescue program in 2010, the government started to be more focussed on attracting FDI instead of supporting national MNEs on foreign investment. This approach could put at risk the Portuguese internationalization process. Besides that, the negative climate of Portuguese economy, it would impact on MNEs decisions, as most probably they could increase their interest on foreign markets as an escape. In this way, Portugal might lose value-added, what could lead to a decrease interest on domestic investment. 29 2.3. Portuguese Policy for Outward FDI Portugal is a European Union member since 1986. The fundamental principles of the European Union are free movement of people, goods, services and capital within the Union (European Union, 2012). With no restrictions on capital movement, from this period, Portugal has started to encourage local firms to internationalize. Several policies were implemented by the Portuguese Government in order to increase firm competitiveness, being more significant the results on Outward FDI since 1996. From 1996 to 2014, there were seven different Constitutional Governments, with the leadership by one of two major parties – PS, or PSD alone or in coalition with CDS. PS (Partido Socialista) has a centre left-wing political orientation, and PSD (Partido Social Democrata) a centre right-wing (Lobo, 2006). Despite the different political orientations, concerning to Internationalization of Portuguese firms, both parties have always in their agenda legislation regarding to foreign investment. The policies introduced by the Portuguese Governments were related to financial and fiscal support, investment service and network council and direction guides to firm internationalization. 2.3.1 Government programmes for FDI  XIII - XIV Constitutional Governments (PS leadership) In 1997 based on a government Portuguese policy, a fund to support Portuguese enterprises development and foreign market participation was launched 3. This fund (FIEP) was considered an essential support mechanism for internationalization projects. In the next period of Constitutional Government leadership, it was introduced a 3 Resolução de Conselho de Ministros 168/97 de 09 de Outubro de 1997 FIEP- Fundo Fundo para a Internacionalização das Empresas Portuguesas, S. G. P. S., S. A 36 2.4.1 The sample, variables and data sources Portuguese Outward FDI stock is the dependent variable and the data come from Bank of Portugal data base. Desbordes and Vicard (2009) say that FDI stocks are preferred to FDI flows, being less unpredictable, which is important when evaluating yearly data. Despite the differences between the BPM5 and BPM6 methods, the report “Implementing the latest international standards for compiling foreign direct investment statistics” of (OECD, 2014) it refers to studying the nature and motivations of FDI, the Assets basis is not the most appropriated, nevertheless, for macroeconomic analysis it is preferable. With the implementation of BPM6, several changes were made, and at the moment of this study the data for the period between 1996 and 2013 was only available in an Assets basis. Nevertheless, for the analysis it will be considered the main countries that are recipients of Portuguese Outward FDI, ranging from 60,11% to 87,20% of total amount: Angola, Belgium, Brazil, France, Germany, Netherlands, Ireland, Italy, Luxembourg, Spain, Switzerland, United Kingdom and USA. As the study objective is to analyse the spatial orientation of Portuguese FDI outwards, in spite of the missing data, we have a high statistical percentage available. The Direct Investment Assets (Outward FDI) are in Million Euros. Graph 3 - Portuguese Outward FDI by Countries 1996-2013 37 Source: Own calculations based on Banco de Portugal (2015) For the independent variables, I will specify the ones that were used in previous studies of FDI determinants, with special focus on the direction of Portuguese Outward FDI and government’s political orientation relation. One determinant that is largely used in FDI studies, is GDP per capita as market size definition (Blonigen & Piger, 2014; Chakrabarti, 2001). The market size is measured by the GDP per capita, and data came from the World Bank Development database indicator. Markets with larger size are expected to attract FDI, as they indicate more opportunities for MNEs. The fact that MNEs exist, is because they look at foreign markets as an extent of the local one; MNEs are competing globally, where rivalry is very demanding. In order to be successful, MNEs look to improve efficiency by producing and distributing in an economy of scale. Additionally S. Globerman et al. (2006) says that “larger markets may be associated with agglomeration economies that lower costs for all producers in that market”. Considering that, we are in a no boundary market and opportunities are associated to larger markets. I expect that GDP per capita is positively associated with capital outflows. Economies with high GDP growth levels, are referred by S. Globerman et al. (2006) as an opportunity to investors. The rapid expansion can contribute to get fast profits, meaning that MNEs will look at these markets with special attention. According to UNCTAD (2014b) in the World Investment Report of 2014, there are strong indicators of FDI in blocks like Africa that it will be increasing in the following years. The need of new infrastructures, namely in Mozambique, are signs of economic growth in Africa. The historical relation that Portugal has with Africa, namely with Angola and Mozambique, may be reasons to have particular awareness of these markets. Following S. Globerman et al. (2006) interpretation, concerning to Outward FDI, it is expected to have a negative sign if Portuguese GDP growth level is bigger than foreign economies. On the contrary, we would have a positive sign. This conclusion is also referred by Simões and Cartaxo (2012) for the Portuguese market, taking as sample the Outward FDI flows in the period in between 2000 and 2010. The Global Economic Prospects Index will be used for this analysis. 38 Natural resources are important for countries with few such assets, namely on Oil. Sachs and Warner (2001) referred that there is a relation between GDP and natural resource levels, as they called it the curse of natural resources. Countries with low GDP level are paradoxically the ones with higher natural resources. Although it is obvious, as referred previously, that high GDP levels attracts FDI, it is also true that the dependency specifically on oil, makes those countries make investments in these countries. Kolstad and Wiig (2012) say there is a positive relation to FDI in such countries. Considering that Portugal has a high dependency on fuels and lubricants, having at the end of 2013 a negative value of 6.194 Billion Euros in its trade balance (Banco de Portugal, 2014), It is expected that host countries with high level of energy production will have a positive impact on Outward FDI. Energy production index on International Energy Agency production will be the base for this variable analysis. Taxes have a strong effect on Outward FDI decisions (Barrios et al., 2012; Grubert & Mutti, 1991; Herger, Kotsogiannis, & McCorriston, 2014; Kemsley, 1998; Mutti & Grubert, 2004). MNEs evaluate investment decisions taking into consideration various aspects, and one that is standard is tax rate. Concerning to corporate taxes, MNEs measure its effects, pursuing the best strategy to avoid double taxation, find low tax rate countries and evaluate the possibility of repatriating profits to home country. In the case of Portugal, Simões and Cartaxo (2012) say that high shares of Outward FDI is to locations with special tax and financial advantages. In order to measure taxes impact, I will use OECD and national data concerning corporate tax rate. It is expected that Outward FDI has a positive relation to countries with lower taxes rate than Portugal. It is widely mentioned by economists that consumers gain more with a free market. Based on this thought, several measures are being implemented to reduce the intervention of governments, looking forward to getting more advantages to consumers. WTO, with 160 countries represented (Portugal included), has an important role in the world economy concerning the aperture of markets, reducing government interventionism. The cut of tariffs is being gradual over the WTO period existence, and nowadays we have a more free market. This approach leads to free movement of goods, meaning that the effects on import and export will depend on the capacity of MNEs. I will consider that tariffs, exports and imports may be associated with openness of 39 market to FDI, meaning Outward FDI will be positive related to its level. UNCTAD percentage of world FDI openness data base will be used. Although MNEs capability influence the investment level, it is also true that governments play an important role as investment facilitators abroad. By signing bilateral agreements, it creates useful conditions for business relations, giving credibility and decreasing MNEs investment risks. Based on the positive effects mentioned, a dummy concerning BIT’s will be used, expecting a positive sign when BITs agreements exist. The geographical distance between countries may have two kinds of inconvenience, the time necessary to move from different points and time zone. Concerning moving, nowadays it can be easier travelling to different parts of the world, although it can also be expensive. Besides costs, the time wasted on movement is an important factor when MNEs choose the place to do investments. With regard to time zone, the differences can be a limitation to communication. Drogendijk and Martín Martín (2014) say that a larger distance between countries means lower level of Outward FDI. Due to the above arguments, is believed that geographical distance may have a negative impact on Portuguese Outward FDI. In order to calculate distances, I will use the World Distance Calculator available in www.globefeed.com. Cultural distance has influence on Outward FDI decisions, being past colonial ties a good indicator (Head & Ries, 2008; Mhlanga et al., 2010; Siegel et al., 2013). Portugal has a historical past with colonial ties in all continents (e.g. Angola, Brazil, Cape Verde, India and Mozambique). Mhlanga et al. (2010) relates the influence of colonial ties, considering the same language spoken, the variable to measure cultural distance. For this analysis, I will use a dummy concerning Portuguese language spoken in host countries. It is expected to have a positive sign on Portuguese Outward FDI. “Political instability is important to explain variation in financial development around the world” (Roe & Siegel, 2011, p. 307). Instability may indicate possible problems for MNEs, as the control over its assets can be out of hand. This idea it is also corroborated by (Jensen, 2003) saying democratic governments receives 70% of world FDI. In order to evaluate the influence of political instability, I will use as variable the political risk index of Freedom House. Low risk is expected to increase Portuguese Outward FDI. 40 Critical to my analysis, is the effect that political orientation may have on Portuguese Outward FDI. Schneider and Frey (1985) have concluded that political orientation does not have high influence on Outward FDI, instead political stability has. Considering the period of their study and economy basis, the U.S.A., it is relevant to test nowadays this effect. According to (Biresselioglu & Zengin Karaibrahimoglu, 2012; Mudambi & Navarra, 2003) parties with right orientation are more market focussed, meaning I may have a positive relation to Outward FDI when foreign Government is right wing ideology. Besides the ideology in foreign countries, I will check the influence of Portuguese Government political orientation in the Portuguese Outward FDI and also the effect of the same Political orientations between Portugal and host countries. For these variables, I will base my data on countries elections data base. Table 5 – Variables definition, expected sign and data source Variable Definition Expected sign Data Source POFDI (dependent variable International investment position Assets Stock Million Euros, annual, 1996-2014 Bank of Portugal HGDPpc: Host Country GDP per capita Host country GDP per capita constant 2005 US$, annual, 1996- 2013 + World Bank HGDPag: Host Country Market Growth Host country GDP % growth, annual, 1996-2013 + World Bank HenerMtoe: Host Natural Resources endowment Host country M toe of energy production, annual, 1996-2013 + International Energy Agency 41 Htaxperc: Total tax rate Host country total tax rate (% of commercial profits), annual, 1996- 2013 - OECD and country laws HopenFDI: Host country Openness Host country Ratio of inward FDI stock, annual 1996-2013 + UNCTAD Inward FDI BIT: Bilateral Investment treaties Bilateral Investment Treaties (1 Yes; 0 No) + UNCTAD’s BITs report Hgeodist: Geographical distance Geographical distance between Portugal and Host country - World Distance Calculator available in www.globefeed.com HLang: Language Host country same language + Host country data Hprisk: Political risk Host country political risk (Lower 1 to Higher 7) - Political risk index of Freedom House HHgide Government ideology Host government ideologies (Left Wing 0; Right Wing 1) + Country data and international party organizations HHgidec Government ideology Host and home government same ideology (No 0; Yes 1) + Countries data 42 HHgidept Government ideology Portuguese government ideologies (Left Wing 0; Right Wing 1) + Portuguese data Label: + and – appoint to a positive and negative relationship Source: the author 2.4.2 Variables behaviour descriptive analysis For the analysis, 13 countries were chosen and the analysis for the period of 18 years, what means 234 observations. Concerning the variable henermtoe, there were values missing for the 2013 year. Although we have a strong balanced panel data, in order to not reduce the panel data to 221 observations, it was used an interpolation to fulfil these missing values. As it is mentioned by (Meijering, 2002) this method has been used several times and it is consistent The same method was used for the variable hopenfdi to Luxembourg till 200113. In order to understand the variables behaviour in the econometric model, it should be done a statistical descriptive interpretation. In the Table 6 it is shown the mean, minimum, maximum, standard deviations and observation numbers. Looking at the table 6 strong differences between countries are visible, what it is correlated to the fact that we have countries in different levels of development. This fact gives explanation to the standard deviation values. 13 During the analysis I have compared the results with and without interpolation method in order to check consistency 43 Table 6 - Variables descriptive analysis Description Variable Observations Mean Min Max Std. Dev. Portuguese Direct Investment Assets Flows (Million Euros) pofdi 234 2229.764 22.76 27998.24 3920.205 Host country GDP per capita (US Dollar) hgdppc 234 35714.29 1130.046 86127.24 18644.71 Host country GDP growth (%) hgdpag 234 2.726569 -6.370668 22.59305 3.552151 Host country energy production (M Toe) HenerMtoe 234 198.3791 .04 1810 437.3948 Host country total tax rate over commercial profits (%) htaxperc 234 32.48184 12.5 56.8 7.663108 Host country World ratio inward FDI stock (%) hopenfdi 234 4.590824 0 39.25928 6.734325 Bilateral Investment Treaties (1 Yes; 0 No) bit 234 .8418803 0 1 .3656353 Geographical distance Portugal to Host country (Kilometres) hgeodist 234 2694.603 493.96 7280.06 2022.729 Host country same language (1 Yes; 0 No) hlang 234 .1538462 0 1 .3615746 Host country political risk (1 Low to 6 High) hprisk 234 1.487179 1 6 1.356057 Host government ideologies (Left Wing 0; Right Wing 1) hhgide 234 .5769231 0 1 .4951065 Host and home government same ideology (No 0; Yes 1) hhgidec 234 .517094 0 1 .5007789 Portuguese government ideologies (Left Wing 0; Right Wing 1) hhgidept 234 .3333333 0 1 .472415 Source: Own calculations on Stata 44 The dependent variable (pofdi) shows high level of difference, which is related to the fact that Portugal has a strong relation with the Netherlands and Spain, as it is visible in Graph 3. On the contrary, countries like Belgium, Italy and Switzerland have low levels of investment. With a minimum value of 22.76 Million Euros and high level of 27998.24 Million Euros, it leads to a high value of a standard deviation of 3920.205 Million Euros. The mean value is 2229.764 Million Euros. Concerning the independent variable host country GDP per capita (hgdppc), as mentioned before, the fact that in the data we have countries with contrasting development levels, this value is significantly different between minimum and maximum values. For the minimum values we have Angola with 1130.046 US Dollar per capita in 1996 and Brazil nearby with 4286.173 US Dollar per capita in the same year (developing countries). The country forefront these ones, is Spain with almost 5 times more GDP per capita 20853.437 US Dollar in 1996. At a higher level is Luxembourg with 86127.238 US Dollar in the year of 2007. The mean is 35714.29 US Dollar per capita with standard deviation of 18644.71 US Dollar per capita. Considering that Portugal has a GDP per capita of 18214.52 US Dollar in 2013, looking at mean and standard deviation, Portugal is framed in the normal statistical curve at a lower level. In reverse order, GDP percentage growth (hgdpag) has higher levels in the developing countries with maximum value to Angola in 2007 with 22.59% and minimum to Ireland in 2009 with a negative growth of -6,371%. The crisis of 2008, as it is referred by Rose and Spiegel (2011), had a very negative impact to developed countries concerning growth, on the contrary developing countries did not get such strong influence. The variable henermtoe that is related to energy production has a high discrepancy of values. This is due to the fact that there are countries with very low resources, like Luxembourg with 0.04 Million toe that is the minimum value in contrast with the U.S.A. 1810 Million toe. In this variable there is a high discrepancy, as this related with the natural resources endowment, which depends on specific country characteristics. With a mean of 198.3791, in a panel of 13 countries, only three are above this value (the U.S.A., the United Kingdom and Brazil). 45 The minimum value on tax rates (htaxperc) is obtained for Ireland with 12.5; in contrast to the maximum that is 56.8 for Germany. The mean value is 32.48184. If a statistical analysis was made before and after the year of 2000, the mean value would be higher before and lower after this year. About the FDI openness (hopenfdi), we have a mean of 4.604335. The minimum value is 0.0083122 for Angola. This very low value may be related with the high risks on democracy which may give instability. The maximum value is 39.25928 from U.S.A., one of the most open countries in the world concerning receiving FDI. Apart from the U.S.A., the United Kingdom and France and with very low exceptions of the Netherlands and Belgium during some years, all the countries are below the mean. Concerning Bilateral Investment Treaties (bit), the medium is very high (0.841) because Portugal has investment agreements signed with almost all countries in this study. The exceptions are Switzerland and U.S.A., and the case of Angola just in the 1996 year. Portugal is part of the European Union which also explains this statistical value. The countries more distant from Portugal are Angola and Brazil 5752 kilometres and 7280 kilometres correspondingly, being the last one the maximum value found in the hgeodist variable. Only these countries have Portuguese as national language, what justifies the low mean value of hlang variable. The country closer to Portugal is Spain with 494 kilometres distance, being this country the second most important concerning Outward FDI. All European countries in this study are under the hgeodist variable medium value, the other countries are above. In relation to the political risk the medium value is low, what means that Portugal has relations with democratic countries. All the countries have minimum hprisk value, 1, being the exception Angola and Brazil. With concern to Angola the risk is very high, having the maximum value to all the years in analysis (6). With regard to Government ideology, we have three variables into analysis. The variable hhgide indicates that Portugal has relations with a higher relevance to right wing ideology (100 vs 134 observations), what is explained by the means of 0.5769231 and the standard deviation of 0.4951065. Concerning Portuguese Governments ideology, during the period into analysis, it was mainly left. The hhgidept variable means 52 Concerning market dimension the result is positive, which corroborates with past analysis concerning FDI (Blonigen & Piger, 2014; Buckley et al., 2007; Chakrabarti, 2001). As mentioned by Blonigen and Piger (2014) high GDP per capita is a factor of attractiveness to FDI, being significant to MNEs investment decisions. In this study, holding all other variables constant a 1% increase in the GDP per capita in the host country raises the Portuguese Outward FDI stock by 3,94%. On contrary, GDP growth has a negative relation with Outward FDI, which is different from the previous studies mentioned above. Yet this effect is very small, a 1 point % increase in the GDP growth rate decreases Outward FDI stock by 0,04%. Energy production has a positive impact, meaning that Portugal invests in countries with larger amounts of energy production. Due to the fact that Portugal has low energy resources, the dependency where oil is included, ought to influence the location of some MNEs investments. As mentioned by (Buckley et al., 2007; Kalotay & Sulstarova, 2010), these resources are key factors to MNEs investment in foreign countries with high volume of energy production. This study finds that whenever the host country has a 1% energy production increase, the Portuguese FDI stock in that country expands 0,5%, ceteris paribus. Barrios et al. (2012) referred that high taxes on profits have strong negative impact on FDI decisions. Furthermore, Grubert and Mutti (1991) say that MNEs tend to decide to invest in countries with lower taxes, including moving profits to nations with lower tax rates. The Portuguese case is in line with such studies. For an increase in taxes on profits by 1% point it will decrease Outward FDI stock in 0,042%, ceteris paribus. A country with high levels of FDI openness may attract more foreign investment. In the case of Portuguese Outward FDI it is found the opposite. Ceteris paribus, for each percentage unit of additional openness in the host country, Portuguese investment in that country decreases 0,5%. Görg and Greenaway (2004) refer that FDI openness may be related with high level of economic development. In this case most countries in the sample have similar levels of development, except Angola and Brazil that in turn attract relevant amounts of Portuguese FDI. This is possibly the explanation for such odd result. 53 1 kilometer increase on geographic distance between Portugal and host country leads to a decrease of 2,03 Million Euros on Portuguese Outward FDI. This result is in line with Siegel et al. (2013), as they found that this variable has a high and negative impact on investment decisions. Mhlanga et al. (2010) have concluded that Portuguese former colonies have been able to attract large volume of investments from Portugal. The regression analysis results corroborate this conclusion. A host country with Portuguese language means an extra 9,1 Million Euros of Portuguese outward FDI. Political risk, in the case of Portuguese Outward FDI has a positive effect. This is contrary to previous conclusions (Feng, 1997; Jensen, 2003). Political risk 1 point rise, means there is an increase of 1,2 Million Euros holding all other variables constant. As mentioned in Section 2, Angola and Brazil are the countries with higher political risks, which suggest that Portuguese MNEs invest in these countries because other relevant factors such as historical, cultural and political links may discount their political risks. At last, in relation to the Ideology of Governments, a Portuguese right wing Government is associated with a positive impact - 234,275 Euros of Portuguese Outward FDI. This RE model, as mentioned before, is the most adequate to estimate the determinants that explain the Portuguese Outward FDI in the base model. The statistical results are mostly in line with previous studies, namely market size, natural endowments, taxes on profits, geographical distance and cultural similarity. Nevertheless, countries with less openness to foreign investment, low GDP growth and high political risks have been able to attract Portuguese Outward FDI. The Portuguese Outward FDI statistical results are different from (Schneider & Frey, 1985) study. As it was mentioned by the authors, Government Ideology was not relevant to the U.S.A. Outward FDI but a low political risk was. In this study, political risk is a positive cause and right wing Governments at the home country have a positive significant effect, which is the opposite to (Schneider & Frey, 1985) conclusion but in line with the hypotheses. In order to understand the different results of this study to previous ones, the study proceeds to analyse the Countries and years effect on the model. 54 3.1.2 Random effects model – Countries and Years effect One advantage of Random Effects is the possibility that time-invariant variables are included with explanatory power – this is the case of distance and common language variables in the base model. However, it is possible that particular countries and years may have an impact into Portuguese Outward FDI. In order to study the effects associate with particular Countries and Years two regressions were run that include such variables. Table 9 reports the results. The Random Effects regression of the base model has an R2 of 0.6387, the regression with Country variables reveals an improvement to 0.6398 but the regression with both, country and time variables, has an explanatory power of 79%. Considering these results, is relevant to analyze the Country and Year effects as determinants of Portuguese Outward FDI. Table 9, it shows how the dependent variable changes as well Countries and Year effect. Table 9 – Econometric estimation results - Countries and Years regressions Description Variables Countries Countries & Years Coef. Coef. Constant -31.177*** -17.12*** (7.752) (3.489) Host country GDP per capita (US Dollar) Hgdppc 4.118*** 1.720*** (.974) (0. 432) Host country GDP growth (%) Hgdpag -0.039*** -0.015 (0.013) (0. 019) Host country energy production (M Toe) HenerMtoe 0.398 0.325 (0.327) (0. 209) Host country total tax rate over commercial profits (%) Htaxperc -0.041** 0.004 (0. 020) (0. 012) Host country World ratio inward FDI stock (%) Hopenfdi -0.051*** -0. 018 (0. 013) (0. 014) Bilateral Investment Treaties (1 Yes; 0 No) Bit -0.959** -1.113*** (0. 392) (0. 293) Geographical distance Portugal to Host country (Kilometres) Hgeodist 0 0 (omitted) (omitted) Host country same language (1 Yes; 0 No) Hlang 0 0 (omitted) (omitted) Host country political risk (1 Low to 6 Hprisk 1.283*** 1.452*** 55 High) (0. 189) (0. 234) Host government ideologies (Left Wing 0; Right Wing 1) Hhgide 0.373 0.356* (0. 238) (0. 192) Host and home government same ideology (No 0; Yes 1) Hhgidec 0.015 0.018 (0. 113) (0.057) Portuguese government ideologies (Left Wing 0; Right Wing 1) Hhgidpt 0.238*** 0 (0. 079) (omitted) Belgium -7.020** 1.010 (2.911) (1.702) Brazil 1.556** 4.949*** (0.617) (1.009) France -5.710** 2.286 (2.406) (1.644) Germany -5.812** 2.159 (2.463) (1.647) Ireland -6.187* 3.062 (3.520) (1.974) Italy -6.306** 1.579 (2.625) (1.651) Luxembourg -6.792 2.878 (4.817) (2.419) Netherlands -4.749* 3.959** (2.676) (1.750) Spain -2.308 5.077*** (2.398) (1.594) Switzerland -10.571*** -1.079 (3.544) (1.943) USA -8.031*** -0.287 (2.810) (1.837) United Kingdom -6.464*** 2.102 (2.395) (1.752) Year 1997 0. 371* (0. 211) Year 1998 0. 971*** (0. 245) Year 1999 1.054*** (0. 286) Year 2000 1.489*** 56 (0. 282) Year 2001 1.456*** (0. 340) Year 2002 1.347*** (0. 336) Year 2003 1.462*** (0. 315) Year 2004 1.432*** (0. 360) Year 2005 1.623*** (0. 306) Year 2006 1.698*** (0.320) Year 2007 1.697*** (0.373) Year 2008 1.717*** (0.346) Year 2009 2.063*** (0.322) Year 2010 2.288*** (0.286) Year 2011 2.404*** (0.334) Year 2012 2.400*** (0.294) Year 2013 2.390*** (0.318) N 234 234 R2 0.6398 0.7961 SEE or v σ .606 0.474 Notes: *, ** and *** denote, respectively, significance at the 1%, 5% and 10% levels, standard errors are in parentheses. Add the option ‘robust’ to control for heteroskedasticity. Angola and 1996 naturally omitted. Source: Own calculations RE regression on Stata As expected in these models, the variables hgeodist (geographical distance) and hlang (language) are omitted due to collinearity. This is due to the fact that its effects are incorporated into Countries’ characteristics. Additionally, hhgidept (Portuguese Government ideology) is omitted because this variable is captured by Year variables. 57 Comparing the “countries’ regression” with the base model regression the conclusion is that most variables continue to be statistically significant, have the same sign and same value for the coefficients. There are two exceptions: (1) the variable “host country energy production” becomes non-significant, whereas (2) the variable “bilateral investment treaties” becomes statistically significant but with a negative sign. This negative result is contrary to earlier studies, where it is mentioned that bilateral activities are positive (Zhang et al., 2014). Although the example of (Zhang et al., 2014) does not refer to a BIT, but bilateral Government activities, the essence of BIT is there, and in the case of Portugal there are treaties with almost all countries in the analysis (exception to Switzerland, U.S.A. and Angola in 1996). All country-variables are statistically significant except for Spain and Luxembourg, and they also have a negative sign except for Brazil that shows a positive impact. These results need a careful reading because country-variables do compare with Angola which is omitted but works as a benchmark. Thus, a negative sign means that the particular market receives significantly less Portuguese investment when compared to Angola, after considering all other explanatory factors. Likewise a positive sign means that the particular country receives significantly more Portuguese investment. Finally, we highlight that the regression confirms the result of the base model that Portuguese government ideology (right-wing governments) has a positive impact on the outward FDI. Looking at the Country/years’ regression the conclusion is that with few exceptions most explanatory variables become non-significant as their impact is now captured either by country or years variables. The variables host-country GDPpc and hostcountry political risk are significance in all regressions meaning that these effects are robust and not captured by Countries and Years variables. As in the “country regression” bilateral investment treaties is significant with a negative impact. Host country Government Ideology becomes significant and positive in this model. Not considering by now Countries and Year variables, all the others are not significant. Comparing the signs of this model to Random Effects base one, the significant variables mentioned above still have the same signal. In this model, 1% increase in GDP per capita in host country means Portuguese Outward FDI raise by 1,72% maintaining all other variables constant, being according to previous studies as mentioned in the 58 Random Effect base model analysis. With the same criteria, Ceteris Paribus, political risk 1 point increase means a positive impact on Portuguese Outward FDI of 1.45 Million Euros. BIT means a negative impact of 1.11 Million Euros. As mentioned before, the main objective of this model is to analyse the country and year’s effect on Portuguese Outward FDI. Based in the fact there were different Governments during the period under analysis in this study (1996-2013), it is important to check how Governments affect the Outward FDI decisions. As mentioned before, Portuguese Government ideology is absorbed by Years. Based on this and on the Random Effects base model results where Portuguese Government ideology impacts positively, with high levels of confidence we may say that right wing Governments is positive to Portuguese Outward FDI. In relation to countries’ effect, Brazil, Netherlands and Spain, with not less than 95% of confidence, have a positive impact on Portuguese Outward FDI. Although Brazil is a developing country and the most geographic distant considered on this study, the fact that they share the same language, cultural and historical links with Portugal suggest that past colonial ties have decisive influence, like Mhlanga et al. (2010) concluded. Concerning Spain by sharing a common border with Portugal plus sharing a common currency within the within the Euro Zone explains the positive impact found. In relation to Netherlands, there may be specific country characteristics that affect positively the Portuguese capital movements to which tax policy is mostly important. Finally the yearvariables are all statistically significant and positive – in particular from 1997 to 2013 the coefficient value increases on a regular basis and in most years. This is consistent with the trend identified in the Investment Development Path of Portugal studied earlier in this study. 59 Conclusion Nowadays with globalization markets have become increasingly integrated across the world and MNEs take their decisions within this context of greater competition. Due to increasing profits need, efficiency and enhanced competitiveness, MNEs expenditures are weighted. Depending on market advantages, investments in one country instead of another, is a question of opportunity and sustainability. Concerning that, Governments may influence positively, creating conditions for MNEs to invest locally or outwards. It is being largely studied the positive impact that investments have in local economy, as driver to social welfare development and also knowledge improvement. Based on this, there is a particular awareness of Governments establishing conditions to obtain MNEs investments interest, and also helping local MNEs to increase competitiveness in foreign markets, receiving the incursion of earnings and experience development by reverse spillovers. But elected Governments have different ideas due to Party ideologies, and programmes may differ from a right wing Party to a left one. Due to this, it is important to understand if different Government Party ideology may influence MNEs investment decisions, particularly Outward FDI. Despite many studies about FDI determinants, the Governments ideology has not been considered. The determinant that is normally included concerning political issues is democracy as a proxy to political stability. There is one exception the study by (Schneider & Frey, 1985). These authors, besides several other matters, tested Political determinants including Government ideology and also political stability. Taking by reference the USA Outward FDI, they concluded that Government ideology did not have influence, but political stability did. From the period of Schneider and Frey (1985) study several changes have occurred in the world economy, and the results obtained at that time may differ from nowadays. Besides that, the lack of studies concerning this matter is also a good reason to investigate the influence that Government ideology may have on Outward FDI. Thus a regression model was estimated based on a balanced panel data of Portuguese Outward 60 FDI into 13 countries during the period between 1996 and 2013. Besides the most common determinants of outward FDI the model also included three variables related with governments’ political orientation: one for the home country, another for the host country and finally the third one for identical political orientation between home and host countries. Based on the empirical analysis, it was concluded that economic determinants (market dimension and growth), institutional (political risk, FDI openness, taxes on profits and bilateral investment treaties), natural resources, spatial distance and country cultures are important to explain Portuguese Outward FDI as hypothesised. However, GDP growth, FDI openness and political risk had an impact opposite to the expected one. The results achieved show that determinants such as market dimension, language, political risk, Portuguese Government ideology and the relations with Brazil, Netherlands and Spain are significant. In relation to economic determinants, market dimension has a positive high impact. Like Blonigen and Piger (2014) referred, market dimension is the main determinant to analyse FDI. With reference to language, historical links with ex-colonies do impact on Portuguese investment decisions abroad, and are associated with a significant and positive impact of Portuguese idiom. Although Brazil is the most distant country included in this study, it was found a very strong and positive impact on Portuguese MNEs investment decisions in this country, meaning the negative impact of distance and political risk are overcame by the historical links. In the case of Spain, having a common border, it may justify the positive and significant result found in this study. As for Netherlands, there may be specific advantages to MNEs investments, which can be of varied nature, like institutional and/or market advantages. Besides these determinants, Political risk and Portuguese Government Ideology are also important. In contrast with Schneider and Frey (1985), in this study Government ideology has a significant impact but the political risk is positive. Concerning to political risk, the result is contrary to several studies, which may be related to specific characteristics of Portuguese country relations, namely the investment in Angola and Brazil. Although MNEs may not be particularly influenced by Governments, there is in the Portuguese case an important relation to political issues. 61 Concerning the impact of Government ideology on Outward FDI it was found a positive effect associated with right-wing Governments. Although left wing Parties are considered not to be so market oriented as right wing Parties, according to (Vachudova, 2008, p. 391) “the exigencies of the economy are a better predictor of a party’s economic agenda than ideology”. In other words, practical issues may neutralize differences otherwise associated with governments’ political orientation. This is not the case, as far as we know, this is the first study to present evidence that right-wing governments in the home country have a positive impact on Outward FDI. Despite our results, the study presents some limitations. The changes on Balance of Payments Manual from 5th to 6th edition, lead to lack of information concerning Portuguese Outward FDI by country. Although Banco de Portugal provided aggregate data, the information by countries was limited and the study only considered the 13 countries with available data. This influenced the analysis as the panel data was small. For future analysis, it is important to have the most complete information with more countries, and more years, in order to get wider information. Compared to previous studies, the political risk proxy used here was not the best choice. So it is advisable to use a wider and most common source to study this matter, in particular to test if the result is confirmed, or not. 68 UNCTAD. (2015a). International Investment Agreements Retrieved from http://investmentpolicyhub.unctad.org/IIA. UNCTAD. (2015b). Inward and outward foreign direct investment stock - Percentage of total the World Retrieved from http://unctadstat.unctad.org/wds/TableViewer/tableView.aspx. Vachudova, M. A. (2008). Centre-right parties and political outcomes in east central Europe. Party Politics, 14(4), 387-405. Vale, M. (2004). Innovation and Knowledge Driven by a Focal Corporation The Case of the Autoeuropa Supply Chain. European Urban and Regional Studies, 11(2), 124-140. Vernon, R. (1966). International Investment and International Trade in the Product Cycle. The Quarterly Journal of Economics, 80(2), 190-207. doi:10.2307/1880689 Wooldridge, J. M. (2010). Econometric analysis of cross section and panel data: MIT press. World Bank. (2015a). GDP growth (annual %). Retrieved from http://data.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG/countries/1W?displ ay=default World Bank. (2015b). GDP per capita (constant 2005 US$) Retrieved from http://data.worldbank.org/indicator/NY.GDP.PCAP.KD. WTO. (2015). Tariffs: more bindings and closer to zero. Retrieved from https://www.wto.org/english/thewto_e/whatis_e/tif_e/agrm2_e.htm Zhang, J., Jiang, J., & Zhou, C. (2014). Diplomacy and investment–the case of China. International Journal of Emerging Markets, 9(2), 216-235. Secondary Sources Assembleia Nacional de Angola. (1996). Lei n.º 7/96, de 19 de Abril de 1996. Assembleia Nacional de Angola. (1999). Lei n.º 5/99, de 6 de Agosto de 1999. Banco de Portugal. (2014). Fuels and lubricants - External trade of goods Retrieved from http://www.bportugal.pt/EstatisticasWeb/(S(hvtvdam5j5zmb0mwp5hhbwbs))/Se riesCronologicas.aspx. Biblioteca da Presidência da República do Brasil. (2015). Presidente e Ex. Presidentes Retrieved from http://www.biblioteca.presidencia.gov.br/ex-presidentes. http://www.biblioteca.presidencia.gov.br/ex-presidentes Comissão Nacional de Eleições. (2015). Eleição para a Assembleia Legislativa Retrieved from http://eleicoes.cne.pt/sel_eleicoes.cfm?m=vector. http://eleicoes.cne.pt/sel_eleicoes.cfm?m=vector Comissão Permanente da Assembleia do Povo de Angola. (1992). Lei n.º 18/92, de 3 de Julho de 1992 69 Direction des Elections. (2015). Résultats électoraux Retrieved from http://www.ibzdgip.fgov.be/result/fr/main.html. Gobierno de España. (2015). Presidentes desde 1978 Retrieved from http://www.lamoncloa.gob.es/presidente/presidentes/Paginas/index.aspx. Government of United Kingdom. (2015). Present and Past Prime Ministers Retrieved from https://www.gov.uk/government/history/past-prime-ministers. Governo de Angola. (2015). A História Retrieved from http://www.governo.gov.ao/Historia.aspx. Governo de Portugal. (1999). Decreto-Lei nº 401/99 de 14 de Outubro de 1999 Governo de Portugal. (2002). Decreto-Lei nº 225/2002 de 30 de Outubro de 2002 Governo de Portugal. (2003). Resolução do Conselho de Ministros nº 71/2003 de 14 de Maio de 2003. Governo de Portugal. (2006a). Resolução de Conselho de Ministros 25/2006 de 10 de Março de 2006. Governo de Portugal. (2006b). Resolução de Conselho de Ministros 39/2006 de 21 de Abril de 2006 Governo de Portugal. (2007). Decreto-Lei nº 245/2007 de 25 de Junho de 2007 Governo de Portugal. (2009). Resolução de Conselho de Ministros 115/2009 de 15 de Dezembro de 2009 Governo de Portugal. (2011). Resolução do Conselho de Ministros nº 44/2011 de 25 de Outubro de 2011. Governo de Portugal. (2013). Resolução de Conselho de Ministros 33/2013 de 20 de Maio de 2013. Governo de Portugal. (2015). Chefes do Governo desde 1821 Retrieved from http://www.portugal.gov.pt/pt/o-governo/arquivo-historico/chefes-de- governo/chefes-de-estado.aspx. Governo Geral de Angola. (1972). Diploma legislativo n.º 35/72, de 29 de Abril de 1972. Goverrno de Portugal. (1997). Resolução de Conselho de Ministros 168/97 de 09 de Outubro de 1997. Heywood, A. (2012). Political ideologies: An introduction: Palgrave Macmillan. Ignazi, P. (2012). Italy. European Journal of Political Research Political Data Yearbook, 51(1), 159-166. Kiesraad. (2015). Tweede-Kamerverkiezingen Retrieved from http://www.nlverkiezingen.com/. Ministério da Fazenda. (1996). Instrução normativa nº38, de 27 de Junho de 1996. Présidence de la République française. (2015). Les présidents de la République depuis 1848 Retrieved from http://www.elysee.fr/la-presidence/les-presidents-de-la- 70 republique-depuis-1848/. http://www.elysee.fr/la-presidence/les-presidents-de- la-republique-depuis-1848/ Roinn an Taoisigh. (2015). History of Government Ireland Retrieved from http://www.taoiseach.gov.ie/eng/Historical_Information/History_of_Governmen t/. Socialist International. (2015). Membe Parties of the Socialist International Retrieved from http://www.socialistinternational.org/viewArticle.cfm?ArticleID=1927. The federal council of Switzerland. (2015). All federal councillors since 1848 Retrieved from https://www.admin.ch/gov/en/start.html. https://www.admin.ch/gov/en/start.html The International Center Democratic Party. (2015). Parties by Country Retrieved from http://idc-cdi.com/en/parties/. The Press and Information Office of the Federal Government. (2015). Federal Chancellors since 1949 Retrieved from http://www.bundeskanzlerin.de/Webs/BKin/EN/Chancellery/Timeline_Federal_ Chancellors_since_1949/timeline_federal_chancellors_since_1949_node.html. The White House. (2015). The Presidents Retrieved from https://www.whitehouse.gov/1600/Presidents. Thewes, G. (2011). Les gouvernements du Grand-Duché de Luxembourg depuis 1848.