although insignificant. The year 2002 marks the beginning of a trend of substantial price
appreciation. This trend is continued throughout 2003-2005, with year dummies
significant in 9, 13 and 12 markets, respectively. One exception to this general trend in
price appreciation after 2002 is notable. Dallas/Forth Worth is the only market in which
all year dummies are negative, and therefore no price appreciation is observed.
Submarket dummy variables are not reported in the regression outputs, but they are
the main instrument used to control for location within the metropolitan area and their
inclusion improves dramatically the fit of the models.
Next, I examine OLS Model II, which adds variables that control for the absolute
location of the property by using a third order expansion of the latitude and longitude
coordinates. The third order expansion rather than a simple linear form is entered in the
equation to effectively draw a price surface based on location. The resulting model
specification has the following form
LNPRICE, = aC + aEXREPL+a2EXARELQ+a3RELQ_ REPL + a4AGE+aAGE2 + a6SQFT
3
+a7SQFT2+c 1.. I DSQFT+ a9LANDSQF72 + aloPARKING+ a, FLOORS+ 2 CONDITION,
z=2
2005 P<43
+al2BUYEROUT+a13SALELEASEBICK+a14PORTSALE+ XYR,+ Z6 k Y )rui
n=2000 s=2
+ (_,1 (Xk, Y) (20)
k=01=0
Table 16 presents the regression results for Model II. Controlling for absolute
location in addition to relative location does not lead to any significant changes in the
estimated coefficients relative to OLS Model I. Moreover, the fitness of the models
improves only modestly with an increase in R-squared ranging from half to one percent.
With respect to the key variables of interest, the results are re-confirmed. The
coefficient on the variable indicating that the sale is part of a replacement exchange,