This paper studies the industry-specific relationship between industrial clustering and firm growth. The literature usually considers industrial clusters as positive for the general economic development of regions. In particular, their relationship with the formation rate of new firms and the survival rate of existing firms is well documented. However, the impact of agglomeration effects on growth prospects of firms is less clear. Because of the heterogeneity of industries, different agglomeration mechanisms, or the way how industrial clusters are identified and spatially delimited, the conclusions which are drawn in the empirical literature on the impact agglomerations and clusters on firm growth are partly contradictory. Hence, FRENKEN, CEFIS and STAM (2011) consider this as one of the key questions in economic geography at large. The study at hand analyses and compares the relationship between firm growth and industrial clusters for 23 industries separately. Therefore, for each firm a micro-geographically defined agglomeration measure is calculated, which is free of the modifiable areal unit problem (MAUP), that means it is independent from the chosen regional aggregation level and the shape of the regional boundaries. To assess the impact of related economic as well as knowledge generating activities, the space is modelled by using travel time distances between all actors and a flexible two-parameter log-logistic distance decay function framework based on behavioural assumptions. The resulting model is estimated by using quantile regression techniques to account for the stochastic properties of firm growth rates and to shed light on differences in the relationship between highly growing and declining firms and agglomeration economies. It is found that the firms? growth prospects are not affected or even hampered by the agglomeration of own-industry employment. On the contrary, the impact of proximate scientific publications tends to be positive. In general, better performing firms are less affected by their spatial surrounding, which becomes more influential for declining firms. However, these findings depend on both the kind and the age, i.e. the lifecycle, of an industry. Furthermore, the optimal decay function specifications, which are endogenized in this modelling approach, indicate that the spatial range of agglomeration effects considerably vary both between the industries and variables. Three illustrative cases of industries are discussed in more details.
