Regional determinants of MNE’s location choice in post-transition economies

Abstract

This article focuses on the impact of agglomeration and labour market factors on the location choice of MNEs in post-transition economies. We compare data from 33 regions in East Germany, the Czech Republic and Poland using a mixed logit model on a sample of 4,343 subsidiaries for the time period between 2000 and 2010. The results show that agglomeration advantages, such as sectoral specialization as well as a region’s economic and technological performance prove to be some of the most important pull factors for FDI in post-transition regions. With respect to access to labour, the results suggest that FDI in post-transition regions is (no longer) only dominated by efficiency seeking behaviour, but also by access to well-qualified labour.

Appendix 1

1.1 Specification of the coefficients of the empirical function

The profit function

$$ \begin{array}{c} \pi_{jk}=(1-t_j)\times\\ \left[\frac{(\sigma-1)^{\sigma-1}}{\sigma^{\sigma}} \left(\left((1+\tau_{j})w_{jk}\right)^{\gamma_1}{r_{j}}^{\gamma_2}{d_{j}}^{\gamma_3} {S_{jk}}^{\delta_1}{T_{j}}^{\delta_2} {H_{j}}^{\delta_3}{E_{j}}^{\delta_4}\right)^{1-\sigma}\sum\limits_{m=1}^{M} \frac{{MA_{m}}^{\sigma-1}}{{\phi_{jm}}^{\sigma-1}}\right], \end{array} $$

can be transformed by taking logs into the following log-linear empirical function with an error term, e _jk :

$$ \begin{array}{c} \ln{\pi_{jk}}=\underbrace{(\sigma-1)\ln{(\sigma-1)}-\sigma\ln{\sigma}}_{\beta_0}+ \underbrace{\ln{(1-t_j)}}_{\approx\beta_1\ln{t_j}}+\underbrace{{\gamma_1}(1-\sigma) \ln{(1+\tau_{j})}}_{\approx\beta_2\ln{\tau_j}}+\\ \underbrace{{\gamma_1}(1-\sigma)}_{\beta_3}\ln{w_{jk}}+ \underbrace{\gamma_2(1-\sigma)}_{\beta_4}\ln{r_{j}}+ \underbrace{\gamma_3(1-\sigma)}_{\beta_5}\ln{d_{j}}+ \underbrace{\delta_1(1-\sigma)}_{\beta_6}\ln{S_{jk}}+ \underbrace{\delta_2(1-\sigma)}_{\beta_7}\ln{T_{j}}+\\ \underbrace{\delta_3(1-\sigma)}_{\beta_8}\ln{H_{j}}+ \underbrace{\delta_4(1-\sigma)}_{\beta_9}\ln{E_{j}}+ \underbrace{(\sigma-1)}_{\beta_{10}}\ln{\left(\sum\limits_{m=1}^{M} \frac{MA_{m}}{\phi_{jm}}\right)}+e_{jk}. \end{array} $$

The definitions of the coefficients above lead to the profit function serving as the foundation for the empirical analysis.

$$ \begin{array}{c} \pi_{jk}=\beta_0+\beta_1\ln{t_j}+\beta_2\ln{\tau_j}+{\beta_3}\ln{w_{jk}}+{\beta_4} \ln{r_{j}}+{\beta_5}\ln{d_{j}}+{\beta_6}\ln{S_{jk}}\\+ {\beta_7}\ln{L_{jk}}+\beta_8\ln{H_{j}}+\beta_9\ln{E_{j}}+{\beta_{10}} \ln{\left(\sum\limits_{m=1}^{M}\frac{MA_{m}}{\phi_{jm}}\right)}+e_{jk}. \end{array} $$

1.2 Appendix 2

See Tables 6, 7, 8

Table 6 The 33 NUTS-2-regions included in the dataset

Table 7 Descriptives of the secondary variables

Table 8 Correlation table of explanatory variables

.