Abstract
Bid-rent theory suggests that the dependent variable of a population density function is unobserved (or population equals zero) if land values are relatively low in residential use. If the errors of density and bid-rent function errors are correlated, bid-rent theory implies that OLS population density estimates are subject to selection bias. Combined land use and density estimates for suburban Chicago confirm these predictions. The results provide further evidence that Chicago cannot be modelled adequately as other than a polycentric city.
Original language | English (US) |
---|---|
Pages (from-to) | 1119-1130 |
Number of pages | 12 |
Journal | Urban Studies |
Volume | 35 |
Issue number | 7 |
DOIs | |
State | Published - Jun 1998 |
Externally published | Yes |
ASJC Scopus subject areas
- Environmental Science (miscellaneous)
- Urban Studies