Methodology for Housing Market Ranking

The methodology for The Beracha and Johnson Housing Market Ranking (Top 100 U.S. Housing Markets for Over and Under Pricing) is tentatively schedule to be published in the Journal of Housing Research (JHR) under the title A Note on the Estimation of the Degree of Over- or Under-Pricing of Housing Markets Relative to their Long-Term Pricing Trend in Volume 1 - 2022. JHR is dedicated to publishing scientifically reviewed articles on the state of housing and housing markets around the world.  It is a member of the suite of journals published by the American Real Estate Society (ARES). 

A Note on the Estimation of the Degree of Over- or Under-Pricing
of Housing Markets Relative to their Long-Term Pricing Trend

Denise Gravatt, DBA
Florida Atlantic University

Eli Beracha, Ph.D.
Florida International University

Ken H. Johnson, Ph.D.
Florida Atlantic University

09-06-2021

Abstract

This note outlines the estimation methodology of the degree of over- or under-pricing of a given housing market relative to its long-term pricing trend. The purpose of this effort is to provide buyers, sellers, real estate professionals, and policy makers a tool that estimates the premium (over-pricing) or discount (under-pricing) a local market is experiencing at a given moment. In general, this model is designed more for greater public rather than academic consumption.

Introduction

Academic research is driven very much by faculty evaluations, which weights ever more rigorous academic review over practical usefulness to consumers, industry professionals, and policy makers. This note departs from this pattern and develops a simple methodology to estimate the premium or discount being paid for an average residential transaction in a defined metropolitan market at a particular time. Armed with this estimate, market participants can make more informed real estate decisions.

In this note are additional sections outlining the methodology, a sample analysis using data from the Miami metropolitan area, and concluding remarks.

Modeling

Using open-source data from third-party housing data providers, a simple Housing Index (HI) time trend can be developed from the data:¹

LnHI_t = β₀ + β₁ x_t + ε_t                      (I)

Where t represents the numeric month for data with t = 1 representing the initial month of the data and  is a time trend code in terms of months or quarters depending on the data source.² The natural log of HI rather than HI is utilized as the left-hand side variable in order to capture the index percentage change rather than its nominal change in the given housing market over time. As a result, a change in index value from 100 to 110 and from 300 to 330, for example, over a given time-period, are associated with the same price-time trend.

The HI estimation equation provides for the development of a fundamental long-term pricing trend established by the data:                                                  

E(LnHI_t) = β₀ + β₁ x_t                      (II)

Here E(LnHI_t) is the expected HI score in natural log terms. For practical reasons and in order to make the estimation easier to visualize in index or dollar terms, the expected natural log of HI is then converted back to nominal terms:

E(HI_t) = e^E(LnHI_t)                        (III)

The degree that a local housing market is above or below its fundamental long-term pricing trend can then be calculated.  Since data is typically available monthly, the calculation is:                                                  

Premium_t = HI_t - E(HI_t) / E(HI_t)                      (IV)

Equation IV is simply the percentage difference between where housing prices should be (the expected value of HI) and the market’s actual HI scores at time t. Positive values for IV represent a premium or the degree of overpricing, while negative values represent a discount or the degree of underpricing.³

Sample Analysis of Miami Metro

For our sample analysis, we will investigate the metro Miami housing market. The specific primary data employed is the Zillow Home Value Index (ZHVI). This measure is a smoothed and seasonally adjusted housing index that is reported monthly with a three-week lag. The ZHVI does not include the tails of reported transactions. Instead, it only reports closings in the 35^th to 65^th percentile range, which allows for a more accurate estimate of any local housing price trend.  Finally, the ZHVI has been reported monthly since January of 1996.  A sufficiently long enough time span for trend development that includes periods of housing booms and busts.

Results for Miami Metropolitan Housing Market – July 2021

Executing on equations I through IV reveals that the current Miami-Fort Lauderdale Metro housing market is 12.91% overpriced as of July 2021. Said another way, the average home in the Miami-Fort Lauderdale metropolitan area is selling for a premium of 12.91% with respect to the area’s long-term pricing trend.

Figure 1

Figure 1 reveals the path of actual vs. estimated pricing for the Miami-Fort Lauderdale metropolitan housing market. 

Clearly, the market is presently above its long-term pricing trend depicted by the red line in Figure 1. It is also clear that the current Premium for the market, while noticeable, is significantly less than the Premium paid in September 2006 when the market achieved its largest Premium of 79.9%. In contrast, the largest discount relative to its estimated trend was 28.9% and recorded in April 2012. 

Concluding Remarks

This note provides a description of estimation methodology of the degree of over-pricing (Premium) or under-pricing (Discount) in a housing market relative to its long-term trend. The execution of this methodology is dependent upon the existence of a housing price index being available for the market in question. 

Executing on Equations (I) through (IV) provides an estimate for the Premium/Discount in a market and delivers information that creates more informed decision making by market participants (buyers, sellers, real estate professionals, and policy makers).

This note is an extension of the now well accepted Case-Shiller style modeling. Housing price indices are now relatively common with many being available in open-source forums. Housing price indices allow for the recognition of where current prices for a given housing market are located and facilitate accurate calculation of property appreciation. Unfortunately, they do not allow for a method of estimating if markets are particularly over- or under-priced. This note lays out a simple and direct method for making these critical calculations.

Endnotes

¹ Some providers of Housing Indices include, but are not limited to, the Federal Housing Finance Agency, Zillow, and Realtor.com.
² This analysis employs monthly data.
³ This methodology could be employed to deliver regular periodic updates of the Premium/Discount scores for metropolitan areas around the country.