Flexibility in HRM and foreign direct investment: Do international investors self-select?

Abstract

Only a few studies so far have tried to assess the role of HRM-flexibility in cross-national location decisions. None of these, however, investigates a potential heterogeneity across multinationals in terms of their specific needs concerning different types of flexibility in HRM. In this paper, I theoretically and empirically analyse the impact of HRM-flexibility on the type of investment being attracted. The theoretical implications are based on a real options framework and are derived in the context of a British–German comparison. They are confronted with data from an original survey on US-multinationals with manufacturing subsidiaries in Germany and the UK. The analysis reveals that in UK subsidiaries, investments in R&D are significantly lower, and technological uncertainty is significantly felt less strongly than in their German counterparts. Both results are compatible with the implications derived from the real options framework.

Keywords:

• HR-flexibility
• cross-national location decisions
• multinationals
• heterogeneity
• real options
• labour regulations

Acknowledgement

The author wishes to thank the German Science Foundation DFG for financial support in collecting the data.

Notes

1. While there is literature on real options theory and FDI (see, e.g., the recent collection edited by Rugman and Li 2005), the fact that the value of real options may indeed vary from location to location and may be influenced by the regulatory environment or by different types of flexibility is (to the best of my knowledge) only taken into account by Moretto and Valbonesi (1999) and by Sureth (2002). These two studies, however, are not concerned with labour regulations or HRM-flexibility.

2. To be able to focus on the effect of HRM-flexibility on firm location decisions, I assume that international investors otherwise are indifferent between investing in the two country locations, with the implicit assumption that, e.g., higher labour costs in Germany may indeed be compensated by a higher labour productivity (see Oulton 1994, p. 58; Ross, Bamber and Whitehouse 1998, p. 342) and/or a higher quality of products (see Jarvis and Prais, 1997). This assumption is, e.g., supported by O'Mahoney's (1995, pp. 87f.) data measuring the ratio of real unit labour costs in Germany to real unit labour costs in the UK between 1989 and 1993. According to his research, this ratio lies between 0.92 and 1.17, i.e. is close to one.

3. According to a recent survey among 361 UK firms by Michie and Sheehan-Quinn (2004), 71% of participating firms used formal induction programmes for new employees (arguably basically on an on-the-job basis), whereas only 26% of firms stated that ‘at least some non-managerial employees received formal off-the-job training in the past 12 months’. Apparently, then, UK firms seek to make up for the deficiencies of its VET system (the so-called ‘failure of training’) by basically investing in firm- and job-specific skills which, however, are less susceptible to functional flexibility: a ‘redistribution of tasks amongst non-managerial employees’ by, e.g., job rotation or by decreased job demarcations is undertaken by a comparatively low fraction of 32% of participating firms.

4. Similarly, Görg (2005, p. 521) argues that in industries confronted with high risk of failure, exit costs will be particularly relevant when it comes to entry decisions (for a related argument see Haaland, Wooton and Faggio 2003, p. 15).

5. The information was eventually conveyed by executives from the highest hierarchical levels of the parent firms (CEO, Chairman, President, Vice President, CFO or COO). With regard to their experience with the subsidiaries, not only close involvement, line responsibility, direct reporting relationship and frequent visits were reported, but some respondents were even involved in the establishment of the subsidiary (or its acquisition) or held prominent functions within the subsidiary, e.g. head of the board, etc.

6. Moreover, as the scale is designed to measure product demand volatility in all of its facets, it must by necessity display a certain degree of heterogeneity. As Streiner (2003, p. 102) emphasizes, Cronbach's alpha does not only measure the homogeneity of the items, but also the homogeneity of what is being assessed.

7. Even when controlling for the buffering of excess demand by backlogging customer orders or coping with excess capacity by building up inventories, no significant effect of product demand volatility on the location of the subsidiary could be detected.

8. A closer look at the data reveals the following. A cluster analysis based on the items concerning the use of flexible technologies shows that the way in which flexibility of technology is achieved in German subsidiaries varies systematically from that in UK subsidiaries. While, in German subsidiaries, flexibility of technology takes shape in form of modular product design (item 2.2), in the UK, it is rather achieved by rapid changes of the product mix and switches among production lines (items 2.1 and 2.3). The additive item used in the Logit-analysis might then simply overstate the extent to which the technology in UK subsidiaries is flexible. When, instead of a simple additive item, a dummy variable ‘affiliation to the cluster of firms using flexible technologies’ is used as independent variable in the Logit-analysis, the sign of the corresponding coefficient turns insignificant, hinting at a lack of robustness of this particular result.

9. Including available control variables like size or industry in the Logit-analysis undertaken here, however, would use up degrees of freedom very rapidly.