R&D public expenditure, knowledge spillovers and agglomeration: Comparative statics and dynamics

Since the introduction of the influential core-periphery (CP) model by Krugman (1991),New EconomicGeography (NEG) models have provided a natural framework for non-linear dynamic analysis.¹ Moreover, as shown by the comprehensive picture of policy implications of the NEG paradigm provided by Baldwin et al. (2003), a well established finding is that policy changes have non-linear effects on industrial location, in general. Why economic activities tend to cluster in space is an old question. Often referring to Marshall's famous classification, regional and urban economists analysed the agglomerative effects of better access to public goods in central locations, of knowledge spillovers between firms and of labour market pooling (see for a recent survey Duranton and Puga 2004). Instead, the NEG focuses on the trade costs, increasing returns at the firm level and factor mobility, and determines endogenously the spatial distribution of (monopolistically competitive) firms by the interplay of agglomeration and dispersion forces (see Venables 2008).