Governance and foreign direct investment: a panel gravity model approach

Abstract

This paper investigates the link between governance and foreign direct investment in the case of 14 transition countries by using a panel gravity model approach in two alternative ways. First, the level of governance in the target country is studied. Second, the absolute difference in the governance level between the source and target country is investigated. In both cases the results suggest that the lack of good governance does not deter, in fact it encourages, foreign direct investment.

Keywords:

• governance
• corruption
• FDI
• transition
• gravity model

JEL Classifications:

• D73
• F23
• P20
• C23

Notes

*significant at the 10% level

**significant at the 5% level

***significant at the 1% level.

*significant at the 10% level

**significant at the 5% level

***significant at the 1% level.

*significant at the 10% level

**significant at the 5% level

***significant at the 1% level.

*significant at the 10% level.

**significant at the 5% level.

***significant at the 1% level.

1. Other factors include market size, proximity to source market, wages, skilled labour force, physical infrastructure, trade openness, favourable tax treatment, availability of natural resources. The agglomeration effect is also important.

2. These are two former Soviet Union countries (Russia and Ukraine), four Balkan countries (Bulgaria, Croatia, Romania and Serbia) and eight Central European countries (Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia and Slovenia). The selection of countries is determined by the availability of data.

3. Foreign investors adopt alternative strategies to minimize the risk in countries with poor governance. As Smarzynska and Wei (2000) argue, investors often prefer to have a joint-venture to wholly-owned subsidiary when corruption makes dealing with government officials less transparent and more costly. Wei (2000) also suggests that foreign investors prefer financial investment over FDI as the current international financial architecture provides more protection for bank lenders than for direct investors.

4. The Business and Advisory Committee to the OECD, November 2002 (in Arndt and Oman 2006, 19).

5. Moreover, MNCs do not only strategically adjust to the local environment in order to get business contracts but also actively try to influence the local environment to achieve favourable conditions. They do not passively respond to the local market conditions but adopt proactive policies in their pursuit to control markets. This involves lobbying as well as other illegal activities, such as corrupting policy makers and influencing the legal system.

6. While it is beyond the aims of this paper to investigate the details of companies that have been involved in corruption, here are a few well known examples: Siemens (see Sheenagh 2008 and Associated Free Press 2007 for details), Aon Limited (see FSA 2009 for details), Johnson & Johnson (see Seelye 2007 for details), Halliburton (see Driver 2009 and Leigh 2005 for details), ABB Vetco Gray (see Department of Justice 2004 and Economic Times 2007 for details).

7. According to the PRS Group International Country Risk Guide data, law and order, democratic accountability and corruption are scaled from 0 to 6. Bureaucratic quality is scaled from 0 to 4. High values indicate better governance. See Figure 3.

8. See for example Eaton and Tamura (1994), Wei (1997, 2000), Stein and Daude (2001), Habib and Zurawicki (2002), Bevan and Estrin (2004), Bevan, Estrin and Meyer (2004), Bos and Van de Laar (2004), Gopinath and Echeverria (2004), Guerin and Manzocchi (2006), Guerin (2006).

9. Vienna Institute for International Economic Studies does not provide bilateral FDI flow data.

10. The GDPs of the countries can be replaced with their per capita GDPs and populations but per capita GDP interacts with the other independent variables including the governance variables.

11. See the PRS website for further details and related formulas for calculating risk for all the governance variables.

12. The lower limit of the time period is 1990, which is considered to be the beginning of the transition period and widely used in the majority of the studies concerning the transition countries. The upper limit of the time period is determined by the availability of FDI data. The data are rarely available for the entire period for all the countries. Therefore, we use unbalanced panel data.

13. The sample contains all the transition economies of the Balkan and Central Europe, but only two former Soviet countries, Russia and Ukraine. The dataset used in this paper (Vienna Institute) does not provide data for the rest of the former Soviet countries. However, the included countries correspond to the 95% of the overall average FDI stock in the transition countries during the studied period.

14. Despite these limitations we have estimated the same equation by using the Fixed Effect and Pooled Least Squares methods. Although we do not report it here, we confirm that the results are consistent with the above method for the governance variables.

15. High values indicate better governance.

16. Indeed, by using FDI flows from seven high income countries to 89 countries, Habib and Zurawicki (2002) found a negative correlation between the two variables. Similarly, Cuervo-Cazurra (2006) tested the same hypotheses by using data on bilateral FDI inflows from 183 sources to 106 target economies and concluded that investors from high corruption countries are not deterred by a high level of corruption as they have experienced corruption at home.

17. An obvious way to assess this claim would be to estimate similar regressions for the non-transition countries and compare the results, but due to data limitations this could not be done.

18. By focusing on the different types of corruption, Cuervo-Cazurra also suggested that although both ‘arbitrary’ and ‘pervasive’ corruption have a negative impact on FDI, pervasive (arbitrary) corruption has a larger (smaller) negative impact on FDI in transition countries than in other countries.

19. The same exercise for 175 countries produces similar results. Although the R-bar-square is low (0.053), the coefficient remains positive, which suggests that high income countries receive more FDI.