FDI in space: Spatial autoregressive relationships in foreign direct investment

Students: Iurii Berezhnoi Egor Cusmaunsa

22/02/2013

Motivation
Since 1980, world wide foreign direct investment (FDI) has grown at a

remarkable rate. According to Markusen (2002), in the latter half of the 1990s FDI ows grew annually by nearly 32% . Thus it drives development of formal economic models of multinational enterprises (MNEs ) and increased empirical investigation of factors driving FDI patterns. The Main empirical determinants of FDI are market size and distance GeneralEquilibrium Model two-country framework models
vertical FDI, MNEs - desire to access cheaper factor inputs abroad
Helpman (1984)
22/02/2013 2

horizontal FDI, MNEs substitute for export flows

Markusen (1984) and Helpman (1984).

Weakness reliance on the two-country (or bilateral) framework

Solutions relax the two-country assumption

FDI decision are multilateral in nature (i.e. not independent across host countries) Solutions: Idea: Export-platform FDI Parent country invest in a particular host country including third markets complex vertical exports of inputs to third markets is processing before being shipped to its nal destination

Authors:

Ekholm et al. (2003), Yeaple Baltagi et al (2003), Bergstrand and Egger (2004)
3

22/02/2013

Research Background
What extent does omission of spatial interactions bias the coefcients on the traditional regressor matrix in empirical FDI studies?
How robust are estimated spatial relationships in FDI patterns across specications and samples?

To what extent can we uncover evidence of various theories of FDI using these techniques and available data?

22/02/2013

Data & Findings

DATA
To explore these issues, we use various samples of US outbound FDI from 1983 through 1998.

Input

get

Results

Findings

We nd that the estimated relationships of traditional determinants of FDI are surprisingly robust to the inclusion of terms to capture spatial interdependence, even though empirical patterns in the data suggest that such interdependence can itself be signicant; analysis also reveals that both the traditional determinants of FDI and the estimated spatial interdependence are sensitive to the sample of countries examined;

22/02/2013

Spatial autoregression (SAR)

Main goal: How MNE motivations may generate important spatial relationships in the data that may not be adequately controlled for using standard econometric techniques on bilateral-country pairs.
22/02/2013 6

Methodology & Data

Where: FDI is an nx1 vector with j equal to FDI from the US(parent country) to host country j; Log-linear form, leads to well behaved residuals Blonigen and Davies (2004) Host Variables - standard gravity-model variables for the host countries (GDP, population, distance between the parent and host countries, and trade/investment friction variables), as well as a measure of skilled-labor endowments.

22/02/2013

Integrated Model

Where: Surrounding-Market Potential variable broadly, where for a country j it is dened as the sum of inverse-distance-weighted GDPs of all other kj countries in the world for which we can obtain GDP data, by year.

*W*FDI is the spatial autoregression term, where W is the spatial lag weighting matrix and is a parameter to be estimated, which will indicate the strength and sign of the spatial relationship in FDI.
W is a block-diagonal matrix of dimension n x n, with each block capturing a single years observations.

22/02/2013

Integrated Model

(, ) denes the functional form of the weights, declining in the distance, , , between any two host countries i and j
Time invariant distance implies that: The shortest distance is equal to 173 km, so that (, ) is adjusted

22/02/2013

Methodology & Data

Table 2 provides a list of the 35 included countries(20 of which are OECD), as well as summary statistics of the variables in our data from 1983 through 1998.
22/02/2013 10

Base results

very strong rejection in the data that host GDP and surrounding-market potential have identical effects on FDI activity;

GDP - positive coefficient, surrounding-market potential negative coefficient; spatial lag term is positive and significant;
5% increase in FDI into a host country lead to 10% increase in the distance-weighted FDI going into surrounding markets. inclusion of country dummies substantially eliminates the statistical and economic significance of the spatial terms
22/02/2013 11

Conclusion
spatial interdependence has been largely ignored by the empirical FDI literature traditional determinants of FDI are surprisingly robust to inclusion of terms to capture spatial interdependence estimates of cross-country determinants of FDI are not very robust to changing the sample of countries we find evidence suggestive of export-platform FDI for most industries within the developed European countries

22/02/2013

12