Abstract: Achieving policy goals often requires different policies for different places, but the assignment of places to policies is often arbitrary, political, or based on anecdotal evidence. We argue that there are simple analytical techniques to improve policy by allocating places into corresponding ‘policy regimes’ in a more objective manner. We show how to implement this approach using a threshold model and relate the policy design to the underlying concept of agglomeration economies. Policies are implicitly based on an underlying hypothesis that adjusting specific factors will generate the desired outcome. The threshold approach modifies the underlying theory to allow for stepwise regimes, rather than a continuous function. These regimes determine bands of similar regions, or thresholds define when key variables have the greatest rate of rapid change in slope. Policy-makers can then assign places to policy regimes either according to bands in which similar places would require similar policy settings, or to target places just below thresholds to achieve greater impact by shifting places between thresholds. Bands and thresholds are determined by the data, rather than by anecdotal evidence, arbitrary assignment, bureaucratic experience, or political aims. We use the example of agglomeration economies in Australian cities to demonstrate this suggested approach.
Published: Bond-Smith, S., & Leishman, C. (2024). Threshold regressions for more objective urban and regional policies. Cities, 149. https://doi.org/10.1016/j.cities.2024.104925