The coefficient on BUYEROUT remains positive and significant in Phoenix and

Tucson. The coefficient on CONDOCONV remains positive and significant in San

Diego. The coefficient on PORTSALE is economically and statistically significant in

Chicago, New York City, Portland and San Francisco. PORTSALE is also positive and

significant in Boston, but negative and significant in Portland and Sacramento. However

these results will not be discussed since they are based on too few observations.

 Based on OLS Model II, the percentage sales price changes corresponding to the

estimated (statistically significant) coefficients for EXREPL, EXRELQ, RELQ REPL,

CONDOCONV, BUYEROUT, and PORTSALE for each market are presented in Table 12.

 Percentage price effects are calculated using 100*g = 100*{exp(x) 1) consistent

with Halvorsen and Palmquist (1980). In this equation g is the relative effect on sale price

(PRICE) of the presence of a condition (e.g. replacement exchange (EXREPL),

relinquishment exchange (EXRELQ), etc); x is EXREPL or EXRELQ, or some other

condition.

 The percentage price effect of the sale being part of a replacement property

exchange ranges from 5.43 percent (in Seattle) to 17.58 percent in Denver, while the

effect of the transaction being part of relinquished property exchange on sales price is

between 4.73 percent in Los Angeles to 20.02 percent in Sacramento. The price premium

associated with a sale being part of both a replacement and relinquished exchange ranges

from 8.81 percent in San Jose to 26.26 percent in Chicago. Although the coefficient on

relinquished property is harder to interpret, because of the issues discussed, it is clear that

replacement exchanges have a positive effect on price. Moreover, in the majority of the

markets (12 out of 15) this price effect is statistically and economically significant.