exchange. However, if the taxpayer fails to identify replacement properties in a timely

fashion, or close the exchange according to the guidelines issued by the IRS, no

replacement exchange takes place. Also, it is common for the replacement property to be

of a different type than the original (relinquished property). Finally, replacement

properties may also be located in different markets than the original properties.

 Table 4 shows that direct exchanges are quite rare. Out of 9,450 sales in Los

Angeles only 24 transactions represented direct swaps. This is also the largest number of

direct exchanges recorded for any market. The small number of direct swaps is not

surprising given how difficult it is to match the requirements of both taxpayers and

complete a direct swap of properties.

 For the purposes of the regression analysis that follows in a later chapter, I exclude

all direct exchanges to eliminate the possible bias that their inclusion may induce in the

model. "Other" exchanges, which mostly include exchanges whose type has not been

confirmed, can introduce similar problems as with direct exchanges; hence, these

transactions are also eliminated from the sample. I also exclude sales that are associated

with any other "special condition" that is not of interest for this study. Examples of such

special conditions are sales that are part of an auction, bankruptcy or sales that involve

building contamination, natural disaster damage, or the threat of contamination. In total,

CoStar delineates more than 30 such unusual conditions, which potentially have an effect

on observed transaction prices. I therefore eliminate all observations that contain a sale

condition which is not analyzed by this study. The only special conditions that I allow are

related to condominium conversions and portfolio sales. Since, as noted previously, the

number of sales in the top 20 apartment markets accounts for 86 percent of all apartment