Foreign Direct Investment in the Banking Sector: Empirical Evidence from Turkey

Abstract

Multinational bank activities have gradually risen in developing countries since the beginning of the globalisation process. Rising foreign bank activities in developing countries have motivated researchers to investigate foreign banks, comprehensively. Turkey is a typical example of a developing country that achieved a tremendous growth rate in foreign bank asset, especially throughout the last decade. The aim of this thesis is to examine two-way linkage; (1) between foreign bank penetration (FBP) and banking variables; (2) between FBP and country risk and (3) between FBP, foreign direct investment (FDI) and foreign portfolio investment (FPI) in Turkey. Therefore, this thesis is constructed by three empirical sections. Moreover the pattern of FDI inflow and outflow in the world and in Turkey has been analysed, chronologically. In addition, the theory of FDI is taken into account and existing FDI theories has been criticised.

In the first empirical work – Chapter 3 - the short run and long run relationship, if it exits, between FBP and determinants of bank performance (namely, domestic bank assets, domestic credit and banking profitability) in Turkey was investigated after controlling DGDP and 2001 financial crisis (DUM2001). The outcome of the Granger causality test indicates that there was unilateral causality which runs from DDB to DFBP . Moreover, I also found feedback causality between DFBP and DCREDIT . By employing impulse response functions, I found that there is positive relationship between DFBP and DCREDIT as I expected. Moreover, the response of DFBP to one standard deviation shock in domestic bank assets is initially statistically significant and positive. The reverse effect is statistically significant and positive. In the final model, the response of DFBP to one standard deviation shock in profitability (PRO) is significant and positive at 3rd quarter. The reverse effect is surprisingly positive but not statistically significant.

Specifically, what has not been also investigated deeply in the empirical literature is the two-way linkage between foreign bank penetration and risk such as political, financial and economic. Thus, in chapter 4, linkage between FBP and country risk (namely, political risk, economic risk and financial risk) was examined in Turkey using quarterly data from 1994Q1 to 2009Q4. My finding indicated that I found one error correction term significant and positive in bivariate vector error correction in model 1 and 2, implying that in the long run, foreign bank penetration has contributed to economic and political risk. Moreover, short run causality based on VAR approach between DFBP and financial risk is investigated but I failed to find any significant causality in the VAR model after controlling DGDP and 2001 financial crisis, even at the 10% level. By analysing impulse response functions, I could not detect any significant relationship between DFBP and host country risk variables in the short run. This is because adding control variables (DGDP and DUM2001) make the relationship between host country risk variables and DFBP statistically insignificant.

Finally, I investigated two-way linkage between FBP, FPI and FDI in Turkey after controlling DGDP and 2001 financial crisis. The finding from the VAR based block exogeneity wald test indicated that changes in DFBP significantly lead to changes in DFDI and there is also unilateral causality which runs from FPI to DFBP. Moreover, using the variance decomposition technique I found that DFDI and FPI have little explanatory power for the evolution of DFBP in Turkey. The contribution of DFBP to the variability of DFDI is more than that of FPI. The contribution of DFDI to FPI variability ranges between 0.000% and 9.122% throughout 12 quarter periods whilst the contribution of DFBP to FPI variability ranges between 0.000% and 7.611%.