Endogenous local labour markets, regional aggregation and agglomeration economies

Agglomeration Economies and the Urban Wage Premium in Australia

“…Importantly, it can be seen that the urban wage premium estimates are remarkably stable based on density effects estimated in the range of 1,800 to 300 unique spatial units. The scale effects of the modifiable areal unit problem in the context of Australia appear relatively small when compared to the international literature (Briant et al, 2010;Burger et al, 2010;Meekes and Hassink, 2019), where it is generally found that estimates of agglomeration externalities on wages are higher when using fewer and larger spatial units.…”

“…Table 2 shows that the empirical evidence of positive density effects on hourly wages is mixed, as the size of the urban wage premium estimate as well as the statistical significance strongly depends on the spatial structure and the model specification. Compared to the literature (see, for example, D'Costa and Overman (2014) and Meekes and Hassink (2019)), it is surprising that the density effect on wages based on SA2s is higher after including individual fixed effects, as it controls for the endogenous sorting of more able workers to denser areas. By contrast, the estimates of the urban wage premium based on LGAs is indeed smaller for specifications that include individual fixed effects.…”

“…For other countries, many studies have analysed various research questions related to agglomeration externalities over the last two decades. A key topic has been the heterogeneity in agglomeration benefits among subgroups of the population, for example based on gender, educational attainment, skill, occupation and economic sector (Adamson, Clark, and Partridge, 2004;Di Addario and Patacchini, 2008;Meekes and Hassink, 2019). Other research has analysed to what extent agglomeration benefits attenuate with the distance from the core of an urban area (Rosenthal and Strange, 2008).…”

“…For the Netherlands, Burger, Van Oort, and Van der Knaap (2010) found that scale effects matter for agglomeration estimates, using spatial units based on three pre-defined administrative classifications (municipality, district and region). More recent work by Meekes and Hassink (2019) is the first to estimate the urban wage premium for a continuum of regional aggregations. Using data for the Netherlands from 2006 to 2014, the authors repeatedly estimate the urban wage premium for different sets of aggregation, ranging between 398 disaggregated unique spatial units (municipalities) and 13 highly aggregated unique spatial units.…”

“…This paper applied the same approach used by Meekes and Hassink (2019) to estimate the urban wage premium for a continuum of regional aggregations. A key limitation of Meekes and Hassink, however, is to apply the approach to the Netherlands, which arguably has only very few local labour markets, given its land area size of 33,671 km 2 .…”