Caution to place makers: Greater firm density does not always promote incumbent firm health

William Kerr, Oliver Falck, Christina Günther, Stephan Heblich, 11 February 2013

a

A

A common theme in economic geography is that increasing returns to scale at the local level are essential for explaining the geographical distribution of economic activity. These agglomerative forces are often cited as a rationale for policy intervention to attract new firms to areas. Tight geographic concentration, however, can also raise countervailing costs as firms compete for local inputs. This makes the gains from increased spatial concentration around incumbent firms uncertain. Causal identification in this setting is challenging due to selection effects: economic models typically assume that firms consider the relative costs and benefits of locations and choose the best candidate. This choice process suggests that empirical correlations of changes in local firm concentration and incumbent firm performance or survival are likely to be biased from the true relationship. This endogeneity problem can be overcome with random assignment of locations to entrants. One setting that overcomes the selection bias is where location choice is driven by non-economic factors. Such quasi-experimental variation is rare in regional economics and cannot be generated in controlled experiments (Holmes 2010, Greenstone et al. 2010), but this form of variation is very valuable.

In a forthcoming paper (Falck et al., 2013), we consider this question in an historical setting with quasi-experimental characteristics – the division of Germany after the second world war. By 1949, the three western zones occupied by the UK, France and the US formed the Federal Republic of Germany. The eastern part developed into a satellite state of the Soviet Union and most believed in 1949 that this eastern zone would adopt the Soviet Union’s socialist system. The fear of expropriation (or worse) prompted many firm owners to flee to West Germany. We study this relocation in the context of the machine-tool industry. In total, a fifth of the machine-tool industry present in East Germany migrated during a narrow window of 1949-56. This was a one-time event, as no comparable prior or subsequent migrations occurred within the industry across German regions. This produced a shock representing on average an 8% increase in total industry size for receiving zones.

The relocation of the German machine tool industry

Machine-tool producers are defined as producers of power-driven machines that are used to produce a given work piece by cutting, forming or shaping metal. Based on Who Makes Machinery, a buyer’s guide issued annually since the 1930s, we identify 394 incumbent firms with pre-war experience in the UK or US zones. We are able to follow the fate of these firms after the relocations from East Germany occur. We also identify 33 relocating firms. These firms were often quite strong before the war, and they quickly regained their former strength in their new locations after moving (Buenstorf and Guenther 2011). We can further compare the impact of these relocating firms with new entrants that choose their location more opportunistically and based upon existing conditions (Glaeser and Kerr 2009), again measured through the buyer’s guide on an annual basis.

What factors were important for the destination choices of relocating firms? These choices were mainly driven by non-economic factors that were independent of local industrial conditions. As an example, Figure 1 illustrates the location choice of seven ‘relocators’ from the region of Chemnitz-Ore mountains in eastern Germany. The shading of the left map reflects ‘dialect/cultural similarity’ (Falck et al. 2012), the shading of the right map reflects product similarity within the machine tool industry. The first observation is that distance is a factor – in that the seven firms migrated far into West Germany, more so than what can be considered random. Second, the chosen destination regions tended to have strong cultural similarity with Chemnitz, while there is not a consistent pattern for product similarity.

Figure 1.

Our paper more formally extends this analysis to all relocations and includes many other explanatory factors. We find that the Chemnitz patterns hold generally. Both case studies and empirical models indicate that existing incumbent structures across regions were not a major factor in the location choices of firms fleeing from East Germany. This is not surprising given the lack of information about industrial conditions after World War II and the speed at which these decisions were made. Instead, non-economic factors like greater distance and dialect similarity dominated selection. This provides confidence that these relocations provided arguably exogenous shocks to the local machine-tool industries.

Effect of ‘relocators’ on incumbent survival

We begin our survival estimations with 1949. Figures 2 and 3 present descriptive evidence on the differences between relocating firms and new entrants. Figure 2 divides West German regions into three groups: those regions that experienced no relocations, those that experienced moderate rates and those that experienced high rates. Failure rates are consistently higher for incumbents in regions that experienced high rates of relocation relative to those that experienced moderate rates.

Figure 2.

Figure 3.

Figure 3, on the other hand, shows a very different pattern with new entrants. We consider entrants during the period of relocations that also survived five years, and divide the sample into three equal-sized groups based upon rates of entry relative to incumbent stock. Failure rates are in the opposite pattern from Figure 2: incumbents in regions with the highest entry rates have the lowest failure rates, while incumbents in regions with the lowest entry rates have the highest failure rates.

Formal analysis (using a proportional Cox hazard model with time-varying covariates) finds that relocating firms substantially raised the hazard rate of failure for local incumbents – one additional relocation was associated with a 25% higher likelihood of firm failure compared to the baseline. In contrast, start-up entrants were associated with a very small decline in incumbent hazard rates. These results are robust to a variety of further specification variants.

As a final step, we investigate more closely the input competition mechanism. A total of eight million ‘expellees’ from East Germany and Austria-Hungary came to West Germany late in the second world war. Expellees had little choice in where they were settled, being generally distributed across regions based on the availability of food and housing by the authorities. We group regions into three bins based upon expellee shares. The impact of relocating firms on incumbent survival was particularly strong in regions with lower labour influx due to expellees. Regions with low labour influx experienced the fiercest competition for labour and the effect of relocations on incumbent survival was twice the sample average. The increase in the incumbents’ risk of failure is somewhat smaller in the intermediate group of regions. Most importantly, there was no increase in failure among regions with the largest expellee influx.

Though based on an historical time episode and one specific industry, our results show that heightened firm density can raise costs for incumbent firms in addition to the often-cited agglomeration benefits. This is an important consideration, for example, when policymakers contemplate efforts to promote their local areas by targeted cluster initiatives, bids to attract large firms (Greenstone et al. 2010), and similar. Policy efforts that are neutral in orientation, like physical infrastructure investments or initiatives to improve the generation and dissemination of knowledge, may be more effective alternatives (Falck et al. 2010, Duranton, 2011).

References

Buenstorf, G and C Guenther (2011), “No place like home? Relocation, capabilities, and firm survival in the German machine tool industry after World War II”, Industrial and Corporate Change, 20, 1-28.

Duranton, G (2011), “California dreamin’: the feeble case for cluster policies”, Review of Economic Analysis, 3, 3–45.

Glaeser, E and W Kerr (2009), “Local industrial conditions and entrepreneurship: How much of the spatial distribution can we explain?”, Journal of Economics and Management Strategy, 18, 623-663.

Greenstone M, R Hornbeck and E Moretti (2010), Identifying agglomeration spillovers: Evidence from winners and losers of large plant openings”, Journal of Political Economy, 118, 536-598.

Falck, O, Guenther, C, Heblich, S, Kerr, W (2013), “From Russia with love: The impact of relocated firms on incumbent survival”, Journal of Economic Geography, forthcoming. Permanent link: doi: 10.1093/jeg/lbs035.

Falck, O, Heblich, S, Kipar, S (2010), “Industrial innovation: direct evidence from a cluster-oriented policy”, Regional Science and Urban Economics, 40, 574–582.

Falck, O, Heblich, S, Lameli, A, Südekum, J (2012), “Dialects, cultural identity, and economic exchange”, Journal of Urban Economics, 72, 225–239.

Holmes, T J (2010), "Structuralist, experimentalist, and descriptive approaches to empirical work in regional economics", Journal of Regional Science, 50, 5–22.

 

Topics: Industrial organisation
Tags: agglomeration, clusters, East Germany, Germany

Ifo Professor of Innovation Economics at the University of Munich

Research Associate, Max Planck Institute of Economics

Reader in Economics at the University of Stirling

Assistant Professor, Harvard Business School