تاثیر هزینه های مبادله و طول مدت انتظار اقامت بر صاحب خانه شدن

Economic theory suggests that an increase in the expected length of stay in a dwelling increases the probability of a household choosing to own a house rather than rent. This hypothesis is derived from recognition that there are substantial transaction costs associated with homeownership and increased expected time in the home reduces the annualized transaction cost. Using a military data set, we confirm that expected length of stay in a dwelling and the transaction costs of selling are very important to the ownership decision. Our best estimate of the transaction costs of selling a home are the sum of 3% of house value and 4% of household earnings.

Recent increases in the aggregate rate of homeownership has renewed interest in determining the factors that influence a household’s tendency to own rather than rent 45, Table 27 . There is substantial interest in determining the future course of housing demand and the effects of existing and proposed housing policies. There is also substantial interest in the prospects of homeown- ership for low income and minority households.Our study of the determinants of homeownership focuses on two factors that have received relatively little attention in the literature: transaction costs and a household’s planned duration of stay in a dwelling. Economic factors typically hypothesized to be important to the tenure choice decision include a household’s real income and wealth, the price of owning compared with renting, and mortgage lender requirements. Although both transaction costs and planned stay have been hypothesized as being important to tenure choice, prior empiri- cal work has omitted these factors or has been plagued by measurement error. Further, sorting out the reciprocal impact of decisions about ownership and the duration of stay has been difficult. Through the use of a new data set and the unique employment situation of the respondents in our military sample, we are able to substantially overcome these problems and show that transaction costs are important to the tenure choice decision. Many researchers have explicitly or implicitly ignored the transaction costs of selling a home when determining the cost of homeownership; however, such costs are not trivial. Most easily measurable is the real estate broker’s fee, often 5to7 % of the sale price of the house in the United States. Other costs include the implicit cost of maintaining the house in showroom condition during the time when it is on the market, the lengthier search process for homeowners Ž . compared with renters, transfer taxes, and legal fees Linneman 34 . Also, there is the risk of bearing the cost of simultaneously owning two houses in two localities or delaying the purchase of a new house while attempting to sell the current house. Once the transaction costs of homeowning are recognized, the expected length of stay in the home becomes a key variable in determining the expected annualized cost of owning. Besides the transaction cost of selling a home, other costs are incurred when homeowners relocate including the loss of attachment to a neighborhood, the loss of information about local private and public services, and the possible loss of benefits associated with a mortgage interest rate that is lower than the current market rate. Although important, our data do not permit us to estimate the size of these costs. Measuring the size of transaction costs associated with homeowning is important for a number of research questions involving the housing market other than the tenure choice decision. Linneman 34 needed an estimate of transaction costs to study housing market efficiency. Both Venti and Wise 46 and Weinberg et al . 47 study renters who participated in the Housing Allowance Demand Experiment. These researchers needed an estimate of transaction costs to study respondents’ residential mobility and whether the market was in equilibrium. Transaction costs are an important component of Ž intertemporal housing adjustment models Harmon and Potepan 20 , Edin and . Englund 11 , and an exact estimate is needed for simulations of these models Ž . Ioannides and Kan 30 . For example, Goodman 16 assumes the transaction costs for a relocating homeowner are 5 to 10 % of current income. The transaction cost associated with selling a home is key to the explanation of why some homeowners substantially renovate their current home rather than relocate Ž . Potepan 39 , Montgomery 38 . Transaction costs affect expected future mobility which in turn affects financial choices associated with home mort- Ž . gages including the FRM ARM choice Brueckner and Follain 5 and the Ž . mortgage points interest rate trade-off Brueckner 4 . Three major difficulties are encountered when trying to account for the impact of transaction costs on tenure choice: determining the size of the transaction costs, estimating the initial planned length of stay, and separating the length of stay decision from the simultaneous decision of housing tenure choice. The literature concerning these difficulties is discussed next. Transaction costs associated with homeownership are measured infrequently. Also, authors discussing transaction costs are not always clear about whether they are describing only the transaction cost of selling or the full transaction cost of moving, this including the cost of leaving a neighborhood. Chambers and Simonson 6 argue that transaction costs are at least 6 % of house value. Linneman 34 assumes that the selling costs for homeowners is 12 % of house value and Cunningham and Hendershott 7 argue that it is at least that amount. In a careful study, Rosenthal 42 assumes that transaction costs are 7 % of the home’s value when sold, but notes that this value must be discounted to the present to compute the user cost at the time of house purchase. He finds that the component of user cost attributable to transaction costs is statistically significant in his tenure choice estimation. Unlike the prior studies, Malatesta and Hess 36 measure transaction costs. Using a sample of only 100 observa- tions, they find that transaction costs average 12 to 13 % of house value. A second difficulty is that the planned length of stay is not often observed; however, this variable is important because annualized transaction costs depend upon the time over which they can be amortized. Unless a survey question asking the purchaser about the expected length of stay is administered at the time of the house purchase, a proxy for this variable must be used. Zorn 48 used proxies including the size of the previous dwelling, the duration of occupancy in the previous dwelling, and whether the prior dwelling was owned or rented. Other studies based on panel data used the ex post observed length of stay to proxy for the planned length of stay. However, the observed length of stay measures the planned length of stay with error, possibly large, because unplanned events can cause changes in desired location, desired house size, and the ability of a household to meet mortgage payments. Henderson and Ioan- nides 28 used the predicted value of length of stay derived from a hazard model estimation of observed stays to proxy for the expected stay. The third difficulty is that the planned length of stay is a choice variable Ž . whose value is simultaneously determined with tenure choice Ioannides 29 . One part of the relationship is that the length of stay directly affects the relative cost of homeownership, which then affects tenure choice. The reciprocal relationship is that once the choice of owning or renting is made, owners recognize their greater transaction costs of relocating, causing their lengths of stay to be longer. Thus, small deviations in a household’s optimal spatial location or optimal housing quantity are less likely to cause homeowners to relocate than renters. Chambers and Simonson 6 found a reciprocal relation- ship between tenure choice and length of stay through the impact of length of stay on user cost. Boehm 3 was the first to estimate a simultaneous equations model. However, his data were limited to two measures of the expected length of stay, one being the realized stay and the other being an indicator measured at the time of home purchase of whether a household expected to move in the next few years. He found that shorter stays reduce the likelihood of homeownership. Henderson and Ioannides 28 developed a joint tenure choice and length of stay model. However, they did not allow the length of stay to impact tenure choice or allow it to change owners’ user cost. In general, modeling this process has proven difficult given that one choice variable is dichotomous Ž.Ž tenure and the other choice variable is not observed expected duration of . stay . If the expected length of stay was exogenous and observed, then a straightforward test of its impact could be conducted. However, this test requires finding a sample where housing tenure is freely chosen and the length of stay in the dwelling is predetermined. Our study uses a survey of married military men to study tenure choice. Expected length of stay at the assigned base is known with a high degree of certainty at the time of assignment; thus, use of this data set makes possible a direct test of the link between transaction costs, length of stay, and tenure choice. Our results confirm the hypothesis that the longer the expected stay, the more likely is homeownership, ceteris paribus. Additional specification tests compare the predictive ability of a simple measure of the relative cost of owning used in much of the housing literature with other, more complex, expressions. We find that using a measure of relative cost that accounts for transaction costs and length of stay yields a better fit of the model to the data than using a simple relative cost variable. A final specification check supports the conclusion that the transaction cost associated with selling a home equals the sum of two components: 3 % of house value and 4 % of household earnings. It is important to clarify that our measure of transaction costs associated with owner-occupied housing does not include factors such as the value of the attachment to a neighborhood or mortgage lock-in effects. Venti and Wise 45 found in their study of renters that it took a 14 % increase in utility to induce the average low income renter to move to a new location. This value includes not only the utility value of out of pocket relocation expenses, but also the utility value of the loss of neighborhood. Quigley 39 studies the impact of mortgage lock-in on homeowner mobility and found quite large effects. 3 Thus if a homeowner has a mortgage with interest rate that is below the current market, the transaction cost of terminating that mortgage could add substan- tially to our estimate. In our sample, the respondents know they must relocate at a particular future time; thus, their ties to neighbors and knowledge about a neighborhood will be lost no matter whether they choose to rent or own. Our estimate of transaction costs is therefore limited to the cost of disposing of an owned dwelling.

We test the hypothesis that increased expected length of stay in a dwelling increases the probability of a household choosing to own a house rather than rent, this effect occurring because the transaction costs of owning are spread over a longer period. Prior tests of this relationship have been plagued by mismeasurement of the expected stay and have had to address the problem in a complicated simultaneous equations approach. We avoid these problems by using a survey of married military men where the length of stay at an assigned base is known with a high degree of certainty at the time of assignment, but the choice of owning or renting is made by the military member. Simulation results show that as the expected length of stay increases from 1 to 10 years, the impact on the user cost of housing is the same as a decrease in the nominal mortgage interest rate from 21.3 % to 5.3 % . Also, the same impact occurs when the income tax rate increases from zero to 53 % . Given the extensive prior analysis of the impact of taxes and mortgage interest rates on homeownership tendencies, it is surprising that so little analysis of transaction costs has occurred. Our results confirm the hypothesis that the longer the expected stay, the more likely is homeownership. We find that the standard measure of user cost found in much of the housing literature does not adequately capture the impact of variations in length of stay. Therefore, user cost measures such as the one derived by Hendershott and Shilling or by Chambers and Simonson that incorporate the impact of length of stay on tenure choice are superior. Our tests indicate that households behave as if transaction costs are a significant factor in their decision to own or rent. Our best results occur when the standard user cost formula for homeownership is amended to include a term equaling 3 % of house value and a term equal to 4 % of household earnings, this last term representing the time cost of selling. This result confirms conjectures in a number of prior studies that argued transaction costs were important to the tenure choice decision. Using the best measure of user cost, we find that the price elasticity of demand for homeowning is about 1.1. In our military sample, omission of the length of stay in the tenure choice equation results in a biased coefficient for household income. In civilian samples, we suspect that biases could occur in other coefficients because of the correlation of the planned length of stay in a home with age, race, ethnicity, marital status, income, and wealth. In our sample of military members, length of stay is exogenously determined and less likely to be correlated with the above factors.