Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/235830 
Year of Publication: 
2021
Citation: 
[Journal:] REGION [ISSN:] 2409-5370 [Volume:] 8 [Issue:] 1 [Publisher:] European Regional Science Association (ERSA) [Place:] Louvain-la-Neuve [Year:] 2021 [Pages:] 153-180
Publisher: 
European Regional Science Association (ERSA), Louvain-la-Neuve
Abstract: 
Many questions in urban and regional economics can be characterized as including both a spatial and a time dimension. However, often one of these dimensions is neglected in empirical work. This paper highlights the danger of methodological inertia, investigating the effect of neglecting the spatial or the time dimension when in fact both are important. A tale of two research teams, one living in a purely dynamic and the other in a purely spatial world of thinking, sets the scene. Because the researcher teams' choices to omit a dimension change the assumed optimal estimation strategies, the issue is more difficult to analyze than a typical omitted variables problem. First, the bias of omitting a relevant dimension is approximated analytically. Second, Monte Carlo simulations show that the neglected dimension projects onto the other, with potentially disastrous results. Interestingly, dynamic models are bound to overestimate autoregressive behavior whenever the spatial dimension is important. The same holds true for the opposite case. An application using the well-known, openly available cigarette demand data supports these findings.
Subjects: 
Spatial dynamic panel data
Monte Carlo simulation
Spatial interaction
Dynamic model
Omitted variable bias
JEL: 
C13
C23
R10
Persistent Identifier of the first edition: 
Creative Commons License: 
cc-by-nc Logo
Document Type: 
Article

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.