Estimating the Lagging Error in Real Estate Price Indices

Real Estate Economics, 2001

We explore the causes and extent of appraisal smoothing, defined as a temporal lag bias in appraisals, by analyzing how appraisers use the transaction price data available to them. We test the empirical validity of the partial adjustment model that underlies the traditional "unsmoothing" of benchmark return indexes. We reject the no-lag null hypothesis and find that the extent of bias-inducing behavior appears to vary over time in the manner suggested by rational appraisal behavior as the quantity and quality of contemporaneous transaction information changes. We find evidence that appraisers valuing the same property in consecutive periods anchor onto their previous appraised values, resulting in more lagging than first-time appraisals. An implied policy prescription is for investment managers to rotate appraisers so as not to allow the same appraisal firm to consecutively value the same property.

The Journal of Real Estate Finance and …, 2000

Previous studies on real estate smoothing have generally focused on the second moment of returns for individual properties. Although this body of research has developed plausible reasons for explaining the observed lower risk associated with real estate, no explanation has, however, been offered to account for the large difference in serial correlation at the individual property level compared with the index level. This article addresses this issue and also offers an explanation for the difference in serial correlation observed with different frequency real estate indices. Employing the framework developed by Holbrook Working (1960), we argue that the high levels of serial correlation typically observed in real estate indices results from a combination of random and sticky appraisals that induce cross-correlations between the component returns. Using the concept of sticky values we question the results of Lai and Wang (1998) in which they argue that the variance of appraisal-based returns should always be greater than true returns. We argue that a pragmatic conclusion regarding volatility should be conditioned on the underlying stochastic processes. We draw a distinction between serial cross-sectional and temporal sticky appraisal processes that in¯uence smoothing at the index and individual property levels. Our results indicate that smoothing does not appear to be a serious issue at the individual property level. However, when different appraisal processes are aggregated into an index the underlying cross-correlation between those processes can induce high levels of smoothing.

The Journal of Real Estate Finance and Economics, 1994

The purpose of this paper is to shed light on the history of commercial property values over the past decade, and to compare different methods of constructing commercial properly value indices and returns series. We examine three types of indices: (i) Indices that attempt to reconstruct property market values by "unsmoothing" the appraisal-based Russell-NCREIF Index; (ii) Indices that trace average ex post transaction prices of commercial property over time; and (iii) an index based on unlevering REIT share prices. By comparing the different historical pictures that result from the various index construction methodologies, one gains insight into the nature of commercial property price and valuation behavior. The REIT-based values lead the other indices in time but display greater short-run volatility. The transactions-based indices lag behind the other series in time, and are consistent with the idea that institutional investors attempt to hold onto properties until they can sell them for a price at least equal to the current appraised value, in effect trading off liquidity for reduced volatility.

Social Science Research Network, 2009

Based on behavioral finance and economics literature, we construct a theoretical framework in which consumers of newly constructed housing units perceive prices to follow a stochastic mean reversion pattern. Given this belief and the high carrying cost maintained by real estate developers, potential buyers opt to either exercise immediately or defer the purchase. We simulate the model within a real option framework by which we show that the optimal time to wait before exercising a purchase is positively related to the price level; hence, a negative (positive) correlation between transaction volume and price level (yield) emerges. Observing data on housing prices and new construction sales in Israel for the years 1998-2007, we apply an adaptive expectation regression model to test consumers' belief in both mean reversion and momentum price patterns. The empirical evidence shows that while consumers' demand pattern is simultaneously consistent with the belief in both momentum and mean reversion processes, the effect of the latter generally dominates. Moreover, while the data does not allow for testing the volume and price-level correlation, it does provide support to the positive volume-price yield correlation.

2011

Although the correlation between the public and private market pricing of real estate has generated considerable research effort, the methods utilized in previous studies have failed to capture the dynamic nature of this correlation. This paper proposes a new statistical method to address this issue. This method, known as the dynamic conditional correlation GARCH model, enables us to study the dynamics of the correlation between the two markets over time and enrich our understanding of the public and private market pricing of real assets. We find that the correlation between NAV returns and REIT returns is dynamic for all REIT types and there is a strong degree of persistence in the series of correlation. Our Granger-causality tests show that price discovery generally takes place in the securitized public market. However, we also find significant variations across property types and individual firms within each type. Our results indicate that constructing an optimal portfolio requires firm level analysis of causality and correlation between REIT returns and NAV returns.

2011

The two basic models used for constructing price indexes for durable assets (such us real estate assets) have been the hedonic and repeated sales models. Case and Quigley (1991)-CQ proposed a generalized least squares (GLS) procedure to estimate a combined (single and repeated sales information) model (CSRS). Hill, Knight and Sirmans (1997)-HKS proposed a maximum likelihood procedure to estimate the CSRS accounting for the autocorrelated error process in the hedonic model. Rambaldi, Hill and Doran (2003)-RHD proposed a methodology based on interpolating incomplete observations. The sale price of a particular asset, a house for instance, can be viewed as an incomplete time series since the price is observed only when a sale occurs. The method proposed by RHD can estimate a CSRS model with autocorrelated errors using a maximum likelihood estimator. It improves on HKS' in that it also produces time-space consistent interpolations (model's prediction of repeated sales always equal the actual observations) and explicitly estimates a cross-sectional correlation parameter. The present paper compares RHD to traditional estimators as well as the HKS estimator through a simulation experiment.

2011

We present empirical evidence using daily data for stock prices for 17 real estate companies traded in the Sao Paulo, Brazil stock exchange, from August 26, 2006 to March 31, 2010. We use the US house price bubble, financial crisis and risk measures to instrument for momentums and reversals in the domestic real estate sector. We find evidence of conditional premium persistence and conditional volatility persistence in the market.

2000

We examine two related hypotheses in this paper. Hypothesis One states that the degree of appraisal smoothing in the direct market depends on market liquidity, with smoothing highest in periods of low liquidity. When there are few transactions (i.e., low liquidity), a prudent appraiser will place more weight on past evidence. Other authors have advanced this hypothesis and, in our literature review, we present some indirect evidence that supports the hypothesis. Hypothesis

2000

Repeat-sales or, more appropriately, repeat-observation home-price indices are the most widely used measures of changes in home values. The two most widely cited of these indices are the Conventional Mortgage Home Price Index (CMHPI) and the OFHEO Home Price Index (OHPI), and both are based on valuations of properties backing loans purchased by Freddie Mac and Fannie Mae. These indices use repeated valuations of the same properties over time to gauge the average change in home prices and suffer from "revision volatility." Revision volatility is a tendency of previously estimated values for prior quarters to change with a new release and is the focus of our study.