sales, I focus in the empirical analysis on property sales in these markets only. In

addition, I exclude markets for which there is not a sufficient number of exchanges to

generate meaningful statistical tests. These excluded markets are Miami, Ft. Lauderdale,

Washington, DC, Dallas, and Colorado Springs. This leaves us with 15 markets, which

represent 75 percent of all apartment sales. The final sample has 23,640 apartment

transactions over the 1999-2005 time period.

Office Properties

 Table 5 presents information on office property sales by markets and shows that in

this sample the largest office market is Los Angeles with 8.5 percent of all sales. Phoenix

is the second largest market with 7.2 percent of all transactions, while Washington, DC is

the third largest market with 6.8 percent of all office transactions. The smallest market

(San Antonio) contains only 22 sales observations.

 The table reveals a different picture than with apartments. First, there is less

concentration of transactions in the largest markets. Consequently, the top 20 markets

account for 75 percent of all sales. Second, the distribution of exchanges across markets

is also quite different. Although the use of exchanges varies substantially across major

metropolitan areas, in none of the markets do exchanges represent the dominant form of

property transaction.2 The observation that markets located in the Western U.S. have the

highest number of exchanges remains unchanged. The distribution of relinquished and

replacement types of exchanges also varies across the sample. In approximately one half

of the markets, replacement exchanges are either equal to or as much as twice the number

of relinquished exchanges. In some markets, sharp contrasts are observed. For example,


2 Salt Lake City is the only market that has a share of exchanges that is above 50 percent, but this market
can be ignored since it only has 26 transactions. Therefore, percentages may be biased due to the small
sample.