On Choosing Among House Price Index Methodologies

This paper compares housing price indices estimated using three models with several sets of property transaction data. The commonly used hedonic price model suffers from potential specification bias and inefficiency, while the weighted repeat-sales model presents potentially more serious bias and inefficiency problems. A hybrid model combining hedonic and repeat-sales equations avoids most of these sources of bias and inefficiency. This paper evaluates the performance of each type of model using a particularly rich local housing market database. The results, though ambiguous, appear to confirm the problems with the repeat sales model but suggest that systematic differences between repeat-transacting and single-transacting properties lead to bias in the hedonic and hybrid models as well.     

Graduated Flood Risks and Property Prices in Galveston County

Our hedonic property analysis approach in Galveston County, Texas aims at estimating the impacts of flood risks and water‐related amenities in a more systematic way. First, we interact distance to the nearest coastline and flood risk in order to account for these impacts acting together on housing sales prices in our coastal community. Second, we use more granular flood risk measure in the analysis compared to the existing literature. Results show that the hedonic price effect is dependent upon the distance to the nearest coastline, and as expected the distance effect varies by flood risk type. We find that in this coastal housing market properties located in the highest risk flood area, for up to nearly a quarter mile from the nearest coastline, actually command a price premium. A recent movement toward risk‐based flood insurance premiums in the United States was deeply opposed by the real estate sector for fear of causing property values to steeply decline. This analysis sheds some further light on this depressed property value assertion highlighting its sensitivity to distance to the water.     

Repeat Sales as a Matching Estimator

The most common approaches for constructing house price indices—hedonic price functions and the repeat sales estimator—focus on changes over time in mean prices. Though the hedonic approach is less wasteful of data than the repeat sales estimator, it relies on an accurate specification of the underlying econometric model. I suggest using a matching estimator as an alternative to the hedonic and repeat sales approaches. Like the repeat sales approach, a matching estimator uses pairs of sales from different dates to estimate the mean difference in sales prices over time. The matching approach preserves much larger sample sizes than the repeat sales estimator while requiring less preimposed structure than the hedonic approach. The matching approach makes it easy to characterize changes in the full distribution of house prices.     

Hedonic Property Valuation Models: Are Subjective Measures of Neighborhood Amenities Needed?

Common applications of hedonic residential property valuation models have relied on proxy measures of neighborhood amenities rather than directly measured subjective evaluations. While this approach simplifies data collection, there is reason to suspect that the use of proxy measures has degraded the accuracy of hedonic valuation models. The extent to which the use of proxy measures may decrease predictive accuracy, and fail to capture the effect of neighborhood amenities, has not been reported previously. The results presented here show that direct subjective measures yield only modest improvements in price prediction. Careful use of proxy measures for neighborhood amenities appears to be cost efficient for large-scale appraisal work.     

Dwelling Age Heteroskedasticity in Repeat Sales House Price Equations

Several authors have attributed the heteroskedasticity observed in repeat sales house price equations to the length of time between sales. Recently, Goodman and Thibodeau (1995) developed a theoretical model that relates heteroskedasticity in hedonic house price equations to dwelling age. Using data for nearly 2,000 repeat sales in Dallas, Texas, this research examines whether repeat sales heteroskedasticity is related to dwelling age, to the length of time between sales, or to both. An iterative generalized least squares procedure that explicitly models the residual variance is used to obtain robust parameter estimates and to increase the efficiency of the usual repeat sales price indices.     

A Varying Parameters Approach to Constructing House Price Indexes

Conventional housing price index models assume interperiodparameter stability and typically employ either repeat sales or hedonic methodologies. This paper introduces a method of index construction that combines multiple sales observations with single sale transactions while permitting characteristics prices from hedonic regressions to vary over time. A test for interperiod parameter stability is provided. Each period's data are arranged by location and repeat sales are matched by rows. This construction allows greater use of sample information and acknowledges the unique contribution of repeat sales to the estimation process. It also produces intertemporal error correlations that can be beneficially exploited by the seemingly unrelated regressions (SUH) technique. The paper also demonstrates a significance test for error correlation and discusses the treatment of unequal numbers of observations among index periods.     

GMM Repeat Sales Price Indices

Illiquid assets are widely spread within the economy but their indices are difficult to measure. This paper proposes a Generalized Method of Moment (GMM) repeat sales regression for estimating illiquid asset price indices. This method has estimators that are arithmetic averages of individual asset returns. This method is able to estimate custom-weighted indices, including equal- and value-weighted indices. It can incorporate hedonic variables to improve estimation accuracy, and it can work with a reweighting technique to mitigate a biased sample problem. Simulations based on artificial markets indicate that the method is more accurate than some alternatives in both efficient and sluggish markets, with and without temporal aggregation. As an application, we use this method to estimate a commercial property price index.     

Gentrification Trends in New Transit-Oriented Communities: Evidence from 14 Cities That Expanded and Built Rail Transit Systems

Over 25 billion dollars were spent between 1970 and 2000 in 14 major cities in the United States on the construction of new rail transit lines. This massive investment in rail transit construction and expansion allows me to study the consequences of local public goods improvements for communities nearby new stations. This article uses a 14-city census tract–level panel data set covering the years 1970 to 2000 to document significant heterogeneity in the effects of rail transit expansions across the 14 cities. Communities receiving increased access to new "Walk and Ride" stations experience greater gentrification than communities that are now close to new "Park and Ride" stations.     

Redevelopment along arterial streets: The effects of light rail on land use change

We use difference-in-differences approaches and parcel-level data from Minneapolis to estimate the effects of light rail on land use change using alternate definitions of treatment area. Results using circular buffers corroborate previous findings that light rail has virtually no effect on land use change in our study area. In contrast, light rail increases the likelihood of land use change along arterial streets that cross the line at station areas. To accurately model the effects of public transit projects on urban land use, one must consider how potential riders access station areas, rather than assuming accessibility improves radially around a station.     

Commercial Real Estate Rental Index: A Dynamic Panel Data Model Estimation

Using the actual quarterly rental income generated in the years between 2001 and 2010 by over 9,000 NCREIF commercial properties, we construct a commercial real estate rental index and estimate the time series properties (e.g., mean‐reversion speed and volatility) of market‐wide rental growth using a dynamic panel data model. The dynamic panel data model has several advantages over a standard hedonic regression. In addition, we incorporate age effects into our panel data model, and by doing so we correct the age bias in the repeated sales method and in the simple average method. Our estimates show that rental growth is cyclical but it generally lags behind broader economic growth. Surprisingly, the long‐term average rental growth is significantly lower than what is usually perceived, and the volatility of rental growth can be significantly under estimated when the conventional methods are adopted. We also find significant cross‐property type and cross‐region variations in the rental adjustment process. In contrast to the existing literature, we find a strong negative relation between rental growth and cap rate, and that this relation is significantly stronger than that between NOI growth and cap rate. Finally, we establish an empirical relation between price return and rental growth in the commercial real estate market.     

Financing Residential Development with Special Districts

This paper empirically examines the extent to which the property tax liability created by financing residential infrastructure using special district bonds is capitalized in house prices. We compare house prices for single‐family detached homes built within development districts to similar properties located outside development districts. Our hedonic specification includes the usual housing characteristics and controls for the influence of spatial attributes using Census Block Group “neighborhood” fixed effects. The preferred empirical specification restricts the data to neighborhoods that have numerous sales of recently constructed single‐family detached homes located both within and outside development districts. The empirical results indicate that house prices for homes located within development districts are lower than house prices for similar homes located outside of development districts, but the amount of property tax capitalization is significantly less than full. Results depend on our Generalized Methods of Moments estimator, which instruments property tax rates using the characteristics of development districts. We identify valid instruments by restricting transactions to properties located in rapidly growing suburban developments.     

Estimating the Gains from New Rail Transit Investment: A Machine Learning Tree Approach

Urban rail transit investments are expensive and irreversible. As people differ with respect to their demand for trips, their value of time, and the types of real estate they live in, such projects are likely to offer heterogeneous benefits to residents of a city. Defining the opening of a major new subway in Seoul as a treatment for apartments close to the new rail stations, we contrast hedonic estimates based on multivariate hedonic methods with a machine learning (ML) approach. This ML approach yields new estimates of these heterogeneous effects. While a majority of the “treated” apartment types appreciate in value, other types decline in value. We cross‐validate our estimates by studying what types of new housing units developers build in the treated areas close to the new train lines.     

Amenity‐Based Housing Affordability Indexes

The recent slump notwithstanding, substantial increases in house prices in many parts of the United States have served to highlight housing affordability for moderate‐income households, especially in high‐cost, supply‐constrained coastal cities such as Boston. In this article, we develop a new measure of area affordability that characterizes the supply of housing that is affordable to different households in different locations of a metropolitan region. Key to our approach is the explicit recognition that the price/rent of a dwelling is affected by its location. Hence, we develop an affordability methodology that accounts for job accessibility, school quality and safety. This allows us to produce a menu of town‐level indexes of adjusted housing affordability. The adjustments are based on obtaining implicit prices of these amenities from a hedonic price equation. We thus use data from a wide variety of sources to rank 141 towns in the greater Boston metropolitan area based on their adjusted affordability. Taking households earning 80% of area median income as an example, we find that consideration of town‐level amenities leads to major changes relative to a typical assessment of affordability.     

Robust Repeat Sales Indexes

Using single‐family sales data for Louisville, Kentucky, we show the benefits of applying robust methods to down‐weight problematic transactions in a repeat sales context. Robust estimators reduce the influence of outliers in repeat sales price changes that are due to data entry errors, quality changes or nonmarket transactions. In addition to comparing conventional and robust indexes, we also use simulated data, where the correct index is known, to show that robust methods control for the impacts of contaminated data. Finally, we demonstrate that robust methods reduce the magnitude and volatility of index revisions.     

Explaining the Variability of Apartment Rents

The research reported here uses regression analysis to analyze rent variations in a sample of apartment data from the Phoenix metropolitan area. Many of the variables used in hedonic price studies of houses are found to be significant in explaining variations in apartment rent. There are differences between hedonic studies of houses and apartments particularly with respect to common area features or amenities. The analysis of various submarkets also produced interesting results. Various uses can be made of the results of this and similar studies by appraisers (market-derived adjustments), property managers (setting rents) and feasibility analysts (the design of apartment projects).     

Output‐Based Performance Pay,Performance‐Support Bias,and the Racial Pay Gap within a Large Retail Stock Brokerage

We measure sources of racial inequality in stockbroker pay. Pay differences arise from sales differences. We measure the extent to which sales differences are due to performance‐support bias, whereby African American brokers receive weaker firm supports, or to forces outside the firm, including client access, selection, and consumer discrimination. Data on firm policies are matched to sales results. Data on self‐generation of accounts measure access to wealthy clients. Sales generated from accounts with sales histories show racial differences in sales arising from selection or consumer discrimination.     

A Semiparametric Method for Estimating Local House Price Indices

Spatial autoregressive hedonic models utilize house prices lagged in space and time to produce local house price indices, for example, the spatial and temporal autoregressive (STAR) model might be used this way. This paper complements these models with a semiparametric approach, the Local Regression Model (LRM). The greater flexibility of the LRM may allow it to identify space–time asymmetries missed by other models. The LRM is fitted to 49,511 sales from 1972Q1 to 1991Q2 in Fairfax County, Virginia. The local price indices display plausible and significant variations over space and time. The LRM price indices in selected neighborhoods are shown to differ significantly from those in some other neighborhoods. A new method for estimating standard errors addresses an overlooked problem common to all local price indices: how to evaluate the amount of noise in the estimates. Out-of-sample prediction errors demonstrate that LRM adds significant information to the hedonic model.     

The Value of Access to Rail Transit in a Congested City: Evidence from Housing Prices in Beijing

Using unique data sets of Beijing's congestion patterns and housing prices, I find that consumers are willing to pay significantly more for access to rail transit in more congested areas. Transit accessibility, however, offers little travel advantage outside of dense urban areas. The expansion of the metro network mitigates the costs of road congestion, creating both private and social benefits. Two policy initiatives aimed at reducing congestion are found to have achieved positive value effects. Further analysis reveals heterogeneous demand for accessibility, with wealthier residents and those owning fewer cars paying a higher premium for access to rail transit.     

Is Selection Bias Inherent in Housing Transactions? An Equilibrium Approach

We develop an equilibrium model for residential housing transactions in an economy with houses that differ in their quality and households that differ in their planned holding horizon. We show that, in equilibrium, a clientele effect persists, with long‐horizon buyers overwhelmingly choosing higher quality properties and short‐horizon buyers settling for lower quality properties. This clientele effect creates a sample selection bias: the properties that are on the market are predominantly of lower quality. Since these are the preferred choice of short‐horizon buyers, they demonstrate a faster turnover. Both the clientele effect and the selection bias are more pronounced with an increase in the variance of house quality and in the variance of the planned holding horizon. Our theoretical model supports empirical evidence on the existence of such bias in home price indices and explains it by the differences in ex ante holding horizons.     

Temporal Aggregation in Real Estate Return Indices

Temporal aggregation is defined as the use of spot valuations of properties occuring over an interval of time to impute the spot value of a property or of a real estate value index as of a single point in time. Temporal aggregation may characterize not only appraisal-based indices but also indices based directly on transaction prices, such as the National Real Estate Index (NREI) and regression-based indices such as hedonic or repeat-sales indices. This paper analyzes the effect of temporal aggregation on the smoothing of the time series second moments in the resulting real estate return index. Assuming true spot returns are uncorrelated, temporal aggregation-induced smoothing will cause the empirically observed real estate index to understate the own-variance by one-third and the beta by one-half. This amount of bias in the second moments can have major implications for the real estate share in an optimal portfolio. Thus, empirical-based investment analysis could be led astray by smoothing even if the real estate return index is "transaction-based" rather than "appraisal-based."