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.     

The Long-Run Performance of REIT Stock Repurchases

This study investigates the long-horizon performance of open-market stock repurchases for real estate investment trusts (REITs). We develop a new methodology to model the autocorrelation of monthly returns into long-horizon buy-and-hold abnormal return estimators. Serial correlation can introduce bias (autocorrelation bias) because the bid-ask bounce may affect monthly returns for sample firms and non-sample firms in a different fashion. Previous long-horizon event studies have overlooked this source of bias. There is compelling evidence that the market underreacts to the stock repurchase announcements. The evidence holds for different measures of the variance and the effects of cross-correlation of abnormal returns. Results are also robust to the traditional buy-and-hold abnormal return and the wealth relative estimators. We investigate the nature of the underreaction and find strong support for the undervaluation hypothesis.     

The Substitutability of Real Estate Assets

This paper investigates the degree of substitutability between securitized real estate assets and real estate assets whose prices are appraisal-based. Given the insensitivity of unsecuritized asset's returns to the returns on stock market indices, equilibrium asset pricing models cannot be used to compare these two avenues of investment. Two assets are deemed substitutable if the information sets underlying unbiased, minimum error variance estimates of their pricing parameters are identical. The empirical evidence shows that the prices of the transactions-based assets—real estate investment trusts and the stock price index of the home building industry—follow a random walk while the prices of the appraisal-based assets—FRC/NCREIF indices—do not. The variance decompositions of the vector autoregressions also show that the level of economic activity helps predict the price indices of appraisal-based assets while the stock market index and the term structure of interest rates are better predictors of the prices of transactions-based assets     

Transactions-Driven Commercial Real Estate Returns: The Panacea to Asset Allocation Models?

A transactions-driven commercial real estate return series is generated in this study to determine whether the reliance on appraised values in the estimation of real estate returns is the source of the reported underpricing of real estate relative to stocks, bonds, and bills when analyzed in a traditional mean-variance setting. The reported underpricing of commercial real estate would be rational if transactions-driven returns exhibit more variance than appraisal-driven returns. While we find that transactions-driven real estate returns have greater variance than appraisal-driven returns for individual properties, most of the individual property risk is idiosyncratic and diversified away at the portfolio level. Real estate continues to be a dominate asset class in mean-variance allocation models even when represented with transactions-driven indices.¹     

Optimal Valuation of Noisy Real Assets

We study the optimal valuation of real assets when true asset values are unobservable. In our model, the observed value cointegrates with the unobserved true asset value to cause serial correlation in the time series of observed values. Autocorrelation as well as total variance in the observed value are used to calculate an efficient unbiased estimate of the true asset value (the time–filtered value). The optimal value estimate is shown to have three time–weighted terms: a deterministic forward value, a comparison of observed values with previously determined time–filtered values, and a convexity correction for incomplete information. The residual variance measures the precision of the value estimate, which can increase or decrease monotonically over time as well as display a linear or nonlinear time trend. We also show how to revise time–filtered estimates based on the arrival of new information. Our results relate to work on illiquid asset markets, including appraisal smoothing, tests of market efficiency, and the valuation of options on real assets.     

Privately Versus Publicly Held Asset Investment Performance

Classic asset pricing is problematic as a method to assess privately held asset investment performance. We propose an alternative approach that involves adjusting the characteristics of assets constituting an index or portfolio to match the asset characteristics of a reference index or portfolio. This approach is applied to commercial real estate, where we create an index of REIT returns to compare to the NCREIF index. To enhance comparability, return indices are adjusted for partial-year financial data, leverage, asset mix and fees. Adjusted results over a 1980–1998 sample period show general convergence between the indices, although an annual return difference of over three percentage points remains in favor of public market asset ownership. Possible causes of the investment performance gap include liquidity and geography as missing risk factor adjustments, an unrepresentative sample period, and the form in which commercial real estate assets are held.     

Rental Expectations and the Term Structure of Lease Rates

We consider the term structure of lease rates in a general setting where both rents and interest rates are stochastic. The framework is applicable to any leasing market, but we focus on real estate. We find that the ``expectations hypothesis", that is, forward rates are unbiased estimators of future rents, requires similar assumptions as in interest rate theory to hold. To study bias magnitude, simulations are performed using a parameterization of the general framework. Different realistic values for risk aversion and interest rate stochastics can generate widely different shapes of the term structure, holding objective expectations constant. Thus an expected increase in rent is consistent with a downward-sloping term structure and vice versa.     

Estimating the Lagging Error in Real Estate Price Indices

Real estate indices based on appraisals or sale prices of properties are known for their slow response to market news. These indices can therefore be represented (in logarithm) as the sum of a latent "true" price index and a lagging error. We show that the latent appreciation return and the lagging error can be jointly estimated in a state–space model, which has two key features. First, it employs exogenous variables known to predict asset returns to predict the latent appreciation return. Second, it incorporates known sources of the lagging error, such as the partial adjustment in observed index to the latent appreciation return and the seasonality in reappraisal quality. We find that, after the estimated lagging errors are removed, the appraisal–based National Council of Real Estate Investment Fiduciaries returns become more informative and hence exhibit (i) greater variance, (ii) weaker auto correlation, (iii) higher correlation with the returns of the securitized real estate and (iv) more timely response to market news.     

A Further Examination of Appraisal Data and the Potential Bias in Real Estate Return Indexes: Comment and Clarification

This comment points out a flaw in Gau and Wang's recent empirical analysis of appraisal return bias and clarifies a point in Geltner's previous article that may have misled Gau and Wang as well as others. In examining appraisal bias in returns it is important to distinguish between cross-sectional versus time-series moments. In dealing with time-series data, both the arithmetic mean and the geometric mean returns offer useful information and can complement each other in correcting for appraisal bias in the first moment of returns series.     

An Analysis of the Cross Section of Returns for EREITs Using a Varying-Risk Beta Model

A dual-beta asset pricing model is employed to examine the cross-section of realized equity real estate investment trust ( EREIT ) returns over bull and bear markets. No significant relationship is found between EREIT returns and a constant beta. However, beta explains cross-sectional returns when betas are allowed to vary across bull markets. This positive relationship exists for both January and non-January months. During bear-market months, no significant relationship is found between REIT betas and returns. But, during such months, size and book-to-market ratio are found to be negatively related to returns.     

The Underlying Return-Generating Factors for REIT Returns: An Application of Independent Component Analysis

Multifactor approaches to real estate returns have emphasized a macro-variables approach in preference to the latent factor approach originally used in arbitrage pricing theory. Use of high-frequency data, trading strategies and growing emphasis on the risks of extreme events makes the macrovariable procedure problematic. This article explores an alternative to the principal components analysis approach: independent components analysis (ICA). ICA seeks independence and maximizes a chosen risk parameter. We apply an ICA procedure based on a kurtosis maximization algorithm to real estate investment trust (REIT) data. The results show that ICA successfully captures kurtosis characteristics of REIT returns, offering possibilities for developing of risk management strategies that are sensitive to extreme events and tail distributions, augmenting traditional mean–variance approaches.     

Estimating the Excess Returns to Housing at a Disaggregated Level: An Application to Sydney 2003–2011

The returns to housing are particularly important because this asset class makes up such a large fraction of household wealth. Yet they are not straightforward to calculate given both the heterogeneity in homes and the fact they sell only infrequently. We outline a methodology for constructing the excess returns to housing at a disaggregated level, essentially that of the individual home. Our approach explicitly takes account of the inherent risk in homeownership with regard to the capital gain or loss component of housing returns. This approach is applied to a rich data set for Sydney, Australia, from 2003Q1 to 2011Q2. Our findings indicate that the returns to housing are on average quite weak though they exhibit significant diversity across dwelling types and regions. Excess returns are also strongly influenced by assumptions regarding the level of risk aversion.     

Appraisal Smoothing: The Other Side of the Story

Appraisal smoothing has been widely accepted as an important factor to consider when analyzing real estate returns using appraisal-based data. In this paper, we demonstrate that the general applicability of the appraisal-smoothing arguments developed so far in the literature is limited by the assumptions upon which the arguments are based. We further show that the use of appraisal-based data can result in a higher (not lower) variance than that of true returns. Given this, it might be more fruitful to analyze the unique characteristics of real estate markets as possible explanations for the seemingly low variance observed in appraisal-based (or transaction-based) return indexes.     

Institutional Portfolios: Diversification through Farmland Investment

This article demonstrates that farmland can enhance the overall performance of institutional portfolios which are currently dominated by stocks, bonds, and business real estate. Unlike previous articles on farmland returns, this article addresses the issue of "smoothing bias" associated with appraisal-based farmland returns. Improved measures of income returns to farmland are also used in developing the estimates of optimal portfolios. Parametric testing revealed that farmland continues to enter the optimal portfolios even for large increases in the variance or for large reductions in the annual returns to farmland.     

Repeat‐Sales Indexes: Estimation without Assuming that Errors in Asset Returns Are Independently Distributed

This article proposes an alternative specification for the second stage of the Case‐Shiller repeat‐sales method. This specification is based on serial correlation in the deviations from the mean one‐period returns on the underlying individual assets, whereas the original Case‐Shiller method assumes that the deviations from mean returns by the underlying individual assets are i.i.d. The methodology proposed in this article is easy to implement and provides more accurate estimates of the standard errors of returns under serial correlation. The repeat‐sales methodology is generally used to construct an index of prices or returns for unique, infrequently traded assets such as houses, art and musical instruments, which are likely to be prone to exhibit serial correlation in returns. We demonstrate our methodology on a data set of art prices and on a data set of real estate prices from the city of Amsterdam.     

The Cross-Sectional Dispersion of Commercial Real Estate Returns and Rent Growth: Time Variation and Economic Fluctuations

We estimate the cross-sectional dispersions of returns and growth in rents for commercial real estate using data on U.S. metropolitan areas over the sample period 1986 to 2002. The cross-sectional dispersion of returns is a measure of the risk faced by commercial real estate investors. We document that, for apartments, offices, industrial and retail properties, the cross-sectional dispersions are time varying. Interestingly, their time-series fluctuations can be explained by macroeconomic variables such as the term and credit spreads, inflation and the short rate of interest. The cross-sectional dispersions also exhibit an asymmetrically larger response to negative economics shocks, which may be attributable to credit channel effects impacting the availability of external debt financing to commercial real estate investments. Finally, we find a statistically reliable positive relation between commercial real estate returns and their cross-sectional dispersion, suggesting that idiosyncratic fluctuations are priced in the commercial real estate market.     

A Comparison of the Recent Performance of Publicly Traded Real Property Portfolios and Common Stock

In this study we compare the returns earned by investments in publicly traded limited partnerships (PTLPs), finite life equity REITs, and traditional equity REITs with those resulting from investing in common stocks (proxied by closed-end mutual funds). Performance comparisons are made using generalized stochastic dominance (GSD). This tool avoids the joint hypothesis problem that arises when an asset pricing model is used as a performance benchmark. The results of the analysis indicate that the performance of the closed-end mutual funds was preferred to that of the individual equity REITs (both traditional and finite life) and PTLP securities by a wide array of risk-averse investors. This result was most pronounced following the passage of the Tax Reform Act of 1986 which severely restricted the tax deductibility of real estate losses. When the equity REITs were combined into portfolios, their performance dominated the mutual funds during the 1980–85 period. Further, the PTLP portfolio returns were preferred to several of the mutual funds even in the post-1985 period. These findings reflect the fact that the securitized real property portfolios studied are not as well diversified as mutual funds. However, the mutual funds remained the dominant investment alternative in the post-1986 period.     

Real Estate for the Long Term: The Effect of Return Predictability on Long-Horizon Allocations

We examine how the predictability of real estate returns affects the risk of, and optimal allocations to, real estate for investors of differing investment horizons. Returns to direct real estate are mean reverting, and risk decreases with horizon. This is driven by a tendency for property transaction prices to overshoot inflation. Mean reversion in real estate returns is weaker than that of equities, resulting in real estate having similar risk to equities for long-term investors. However, optimal portfolios have large allocations to direct real estate at all horizons, and the allocation increases with horizon. Finally, we find that real estate investment trusts are a redundant asset class for investors with access to direct real estate as an asset class, but they do have a role in optimal allocations when direct property investment is not feasible.     

There is no place like home: Information asymmetries,local asset concentration,and portfolio returns

Using a unique setting with significant cross‐market information asymmetries and a large sample of individual commercial property holdings, we provide robust evidence showing that local information plays a significant role in the linkage between local asset concentrations and return outperformance. We further document a significant positive relation between local asset concentration and portfolio returns in markets where information asymmetry is most severe. Two novel identification strategies that exploit a local lender's ability to price the local investor's information advantage and exogenous variation in sales price disclosure laws across states confirm an information‐based effect that is distinct from risk‐based or behavioral explanations.     

Wage Effects of On‐the‐Job Training: A Meta‐Analysis

A meta‐analysis is used to study the average wage effects of on‐the‐job training. This study shows that the average reported wage effect of on‐the‐job training, corrected for publication bias, is 2.6 per cent per course. The analyses reveal a substantial heterogeneity between training courses, while wage effects reported in studies based on instrumental variables and panel estimators are substantially lower than estimates based on techniques that do not correct for selectivity issues. Appropriate methodology and the quality of the data turn out to be crucial to determine the wage returns.