The role of geographical and temporary proximity in MNEs’ location and intra-firm co-location choices

ABSTRACT

This paper investigates the role of geographical and temporary proximity in the location and co-location decisions of manufacturing activities by foreign multinational enterprises (MNEs). Empirical analysis shows that foreign MNEs co-locate their new manufacturing plants with their plants already operating in the same manufacturing activity, while geographical proximity exerts a much weaker role when the latter operates in services activities. This is especially true in the case of knowledge-intensive business services, where the travelling and meeting of professionals allows temporary proximity. Moreover, a spatial econometric extension of the analysis confirms a geographical decay effect for intra-firm co-location with activities located in contiguous provinces.

KEYWORDS:

• multinational enterprise location choices
• intra-firm co-location
• geographical proximity
• temporary proximity
• geographical decay effect

JEL:

• F23
• L2
• R12

View correction statement:

Correction

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

Notes

1. On the similarities and differences between agency theory and transaction cost theory, see Williamson (1996).

2. Accordingly, the economic geography literature has studied the relationship between external agglomeration and distance-based transaction costs (McCann & Shefer, 2003; Wood & Parr, 2005).

3. The analysis focuses on choices relating to the location of new investments, as opposed to the acquisition of already existing activities in the country. Indeed, in the case of acquisitions, location alternatives are restricted to the places in which the potential target firms are already located. In addition, location will be just one of the possible variables that come into play in the selection of target firms, together with other significant firm-specific factors, such as internal resources, technologies, and other tangible and intangible assets (Head et al., 1995).

4. Other studies (e.g., Gallego & Maroto, 2015) include NACE 74 among KIBS. However, they acknowledge that: ‘the inclusion of ‘other business activities’ within the KIBS category leads one to account for a number of business services such as labour recruitment (74.5), investigation activities (74.6), industrial cleaning (74.7) and other miscellaneous business activities (74.8), such as industrial design, which are not identified by literature as KIBS’ (p. 647). Therefore, we prefer to consider a different category (Other) for these activities.

5. However, to provide a robustness check and control for a possible sample selection bias, we estimated the models considering all the 447 manufacturing greenfield investments undertaken by foreign MNEs in Italy in the period 1998–2012. The results are available from the authors upon request.

6. A company can have more than one location decision (event) throughout the period, and we properly take into account this in the econometric analysis with clustered standard errors by MNE.

7. A hierarchy of Nomenclature of Units for Territorial Statistics (NUTS) levels, for each European country, has been established by Eurostat. The current NUTS nomenclature (applicable from 2012) subdivides the economic territory of the EU into 97 regions at NUTS-1 level, 270 regions at NUTS-2 level and 1294 regions at NUTS-3 level. NUTS-3 areas correspond to a population between 150,000 and 800,000 people. For example, Germany is divided into 412 ‘Kreise’, France into 100 ‘Départements’ and Sweden into 21 ‘Län’.

8. Moreover, we estimate the models without those three companies in order to control for possible biases driven by the outlier cases. The findings are similar to the other specifications. The results are available from the authors upon request.

9. A row-normalized weight matrix is scaled by the row's sum, namely, each value in the matrix is divided by the sum of values in its row.