denominator in these approaches is to view spatial structures as the outcome of historical
processes, and as conditioning but not fully determining economic behaviour. The explicit
historical nature of evolutionary analysis, however, poses demanding requirements for empirical
research. One needs to collect time-series data of evolving populations, be it from technologies,
sectors, networks, cities or regions, and to apply appropriate methodologies to analyse the data
collected. The contributions by Klepper (forthcoming), Jacob and Los (forthcoming) and
Essletzbichler (forthcoming) are fine examples of the use of econometric techniques applied to
time-series data. However, other methodologies are also available to fruitfully apply evolutionary
economics. For example, case study research, combining written and oral sources, can provide an
understanding of long-term planning processes (Bertolini, forthcoming) and the multi-faceted
process of regional development (Garnsey and Heffernan, forthcoming; Quéré, forthcoming).
Static analysis, although dealing with snapshots of an otherwise evolving process, can also be
approached from an evolutionary perspective, for example, by deriving hypotheses on expected
inequalities in network positions (Giuliani, forthcoming) or rates of technology adoption
(Bonaccorsi et al., forthcoming). Nevertheless, such phenomena could be understood better if
time-series data were available.
Apart from data limitations and methodological challenges ahead, there are still a number of
conceptual weaknesses that hamper the application of evolutionary economics to economic
geography. For example, the concept of routines still needs to be refined (Becker, 2004), and
their role in the development of multi-locational organisations is still quite unclear (Stam, 2003,
2006). For example, the evolutionary theory of the firm has little to say about multinational
organisations, exceptions aside (Kogut and Zander, 1993; Cantwell and Iammarino, 2003).
Another key concept in evolutionary economics is path dependence. Yet, its fruitful application
in economic geography is still surrounded by a number of unsolved issues (Martin and Sunley,
2006). Finally, as Breschi and Lissoni (2001) have argued at length, the concept of knowledge
spillovers is, both conceptually and empirically, still ill-defined. Despite the growing number of
studies on knowledge spillovers, the mechanisms underlying such spillovers are still poorly
understood as well as to what extent these mechanisms are sector and/or region specific.
Furthermore, the importance of knowledge spillovers may be specific for the geographical
distance over which they occur. The more important information flows typically stem from more
distant locations, a geographical principle that might reflect the strength of weak ties
19