oo oo oo o oo
oooo oo oo
oo o o o o
oo o o
oo o o
oo ooo o
An unsettled issue in the empirical literature n m deling price indices is whether frequently
s ld pr perties are representative f pr perties that d n t transact r transact nly nce dur-
ing the sample peri d. The expl rati n f this questi n has imp rtant implicati ns f r the
interpretati n and use f repeat-transacti ns price indices. This is the case, because these
indices are, ut f necessity, based n pr perties that transact at least, and ften m re than,
tw times during the sample peri d. T gain insight n this issue, we examine the char-
acteristics and price behavi r f repeatedly transacted pr perties by transacti n frequency.
Using a rich data set fr m f ur different c unties distributed acr ss the U.S., this article
investigates whether mean appreciati n rates and their reliability vary with frequency f
transacti n and hyp thesizes plausible s urces f r p ssible variati ns in these measures.
Frequency f Transacti n and H use Price M deling
Yale University, New Haven, Connecticut
Boston College, Boston, Massachusetts
The University of Pennsylvania, Philadelphia, Pennsylvania
oo o o
oo o ooo o o o
oooo o o o
oo o ooo o
oo o o o oo o oo o
oo oo o oo
J urnal f Real Estate Finance and Ec n mics, 14: 173–187 (1997)
1997 Kluwer Academic Publishers
HENRY O. POLLAKOWSKI
SUSAN M. WACHTER
This article examines the characteristics and price behavi r f repeatedly transacted pr perties. Using data fr m
f ur U.S. c unties, we estimate hed nic price m dels f pr perties gr uped by transacti n frequency, and c mpare
estimated standard deviati ns and estimated appreciati n rates by gr up.
F r each f f ur c unties studied, we ﬁndthat estimated h use price appreciati n is systematically higher am ng
pr perties that transact m re frequently. One p ssible explanati n f r this result is that purchasers make pr perty
impr vements that are n t adequately reﬂected in the available data.
We als ﬁnd that estimated standard deviati ns f the disturbance term sh w a marked decrease as the fre-
quency f transacti n increases. Since frequently transacting pr perties are n t f und t be systematically m re
h m gene us than seld mly transacting pr perties, we d n t attribute this t any increase in h m geneity f r
frequently transacting pr perties, but rather t the length f time elapsed between transacti ns f pr perties.
The ﬁndings f this article suggest that repeat-sales price m dels may need t be adjusted t acc unt f r cr ss-
secti nal variati n in transacti n pr babilities—that is, the selectivity f the subsample f pr perties that transacted
( r transacted repeatedly) during any ﬁnite study peri d.
price indices, repeat sales, hed nics