Homeowners' Repeat-Sale Gains, Dual Agency and Repeated Use of the Same Agent

Abstract

Previous studies of dual agency, where single agent serves both buyer and vender in a transaction, use hedonic patterns Repeat-sale methods can test for the price result of accepting dual agency. Dual agency does not display convincing effects on expected gain, which would arise if there was a systematic bias, or upon heteroscedasticity, which would occur if there are large results that are rare. Earlier researchers could not proof for the effect of an proprietor picking a listing agent who was the earlier selling agent. Consistently positive mean abnormal price gains tend hitherward from this choice, as well as significant heteroscedasticity.

Do inherent principal/agent conflicts cause the real estate brokerage industry to systematically misrepresent buyer and/or sellers? Dual agency in a transaction involves single agent representing both buyer and vender Consumer advocates provide anecdotal evidence that the incidence of representation point in disputes is not zero. The industry recognized several of the question s long ago. Brokerage professional associations rejoined with ethical guidelines that would eliminate the question if agents fully and faithfully adhered to the principles. More lately states have enacted regulations requiring that clients be to the full informed of the potential for conflict of interest.

A review of extensive contributions to the literature upon this important topic finds that cross-sectional statistical analysis has been the style of empirical tests as to whether transaction price is higher versus lower when buyer and vender are represent by alternative agency formats. This application of mind differs from earlier research in that it uses a repeat-sale price index pattern to test for the issue of agency choice on the percentage gain in price above the tenure of a given holder Bailey, Muth and Nourse (1963) and Archer, Gatzlaff and Ling (1996) provide alternative repeat-sale formats for statistical tests

For a data unit in this mould the buyer makes a choice upon dual agency in the first of a pair of transactions, then another choice at the extreme point of his tenancy in the house. In addition to the decision to accept dual agency, the possessor also chooses whether to get back to the same agent for the next to the first transaction. Using a data plant that lists agent names in each transaction, indicator variables can be defined to record dual agency upon the first or second of repeat sales of the same house. Also, a variable may be defined to indicate whether the same agent serv the buyer in the first transaction as a selling agent, and then exhibited that same owner as a listing agent in the next to the first transaction.

Positive mean abnormal go [i]or[/i] come backs may be anticipated for variables defined to indicate that an holder returned to an agent who sold him the house for repeat service as a listing agent. The possessor may believe that the agent's superior representation gave him the opportunity to purchase the house at a price lower than market bourns The owner may believe that these same agency skills will generate a resale price that is also in the owner's favor. The positive coefficient may also tend hitherward from another source. If an agent becomes familiar with a house while serving as the selling agent upon a transaction, then that familiarity may allow the agent to more favorably market the house, generating abnormal gains for the owner

In applying the repeat-sale regression analysis, sum of two units main contexts are asserted as indication of the existence of principal/agent puzzles Of course, if the estimated repeat-sale regression coefficients are statistically significant, then mean holdingperiod gains are abnormal for the clusters of transactions associated with dual agency and repeat use of the same agent.

In a next to the first context, the indicator variables may have no significant impact upon expected price gain, but still have measurable importance. For example, imagine such a large number of agents tread in the steps of ethical guidelines that the distorted transactions are rare anomalies. While there may be no significant regression coefficients that indicate an consequence on expected price gain, the agency variables ; may be associated with outliers and the variance of price gains. Thus, a next to the first set of tests-for heteroscedasticity-may indicate important influences from the owners' agency choices. Findings that agency variables are associated with heteroscedasticity would also mean that researchers estimating repeat-sale price indexes ne to add plane more variables to the list now known to create statistical problems

The main accrues of this paper and its organization are as tread in the steps ofs A review of the literature for agency results indicates a lack of consensus among empirical papers that use cross-sectional hedonic price archetypes Empirical results also differ across house price ranges. The nearest two sections adapt repeat-sale price analysis to this application and provide sample information and variable definitions. A section of statistical proceeds for alternative models deals first with ordeals for heteroscedasticity, then tests for differences in the wait fored price gain that a homeowner may derive pleasure from The concluding section asserts that the flows add new heteroscedasticity worries for general users of repeat-sale rules The section also claims that there is no compelling evidence that dual agency has either the consequences anticipated if there is a general bias in favor of either buyer or venders or the effects that might be anticipated if agent misrepresentation has large events but is rare. The final main conclusion regards the homeowner's agency choice of returning to the agent who sold him the house for employ as a listing agent on resale. This agency factor is a major source of heteroscedasticity and is associated with increased price gains above the period of ownership. These issues range from almost one-third to more than one-half more than the gain that could be anticipated from market price trends