Real Estate Price Indices & Price Dynamics:
An Overview from an Investments Perspective
By David Geltner*
Submitted to the Annual Review of Financial Economics
October, 2014
Abstract:
This article reviews the state of the art in real estate price indexing and, related to that, the
current state of knowledge about real estate price dynamics, with a primary focus on investment
property, income generating commercial properties. Such assets form a large component of the
national wealth and of the capital markets, and represent a major investment asset class. They are
characterized by heterogeneity of various types, among assets, markets, and data sources, making
the study of real estate pricing uniquely challenging. Yet urban economists and econometricians
have pioneered major new price indexing methodologies in recent decades which, combined with
new types of data sources, are now shedding new light on the nature of commercial property price
dynamics, revealing both important commonalities as well as unique differences compared with
equities and fixed-income securities pricing.
Keywords: Commercial Property; Real Estate; Price Indexing; Price Dynamics; Asset Markets.
*Massachusetts Institute of Technology, Department of Urban Studies & Planning, Center for
Real Estate. Contact: [email protected].
1
Table of Contents:
1. Introduction & Background …………………………………………………….…….. 3
2. Some Considerations About Real Estate Asset Markets …………………………….. 6
3. Pricing and Price Indexing in the Property Asset Market ………………………… 13
4. Methodology of Property Price Indexing …………………………………………..... 19
5. Some Findings from Property Price Indices ………………………………………… 24
6. Conclusion …………………………………………………………………………….. 28
Literture Cited …………………………………………………………………………... 30
2
Real Estate Price Indices & Price Dynamics:
An Overview from an Investments Perspective
David Geltner
1. INTRODUCTION & BACKGROUND
This article will focus on asset price indices and price dynamics in real estate. Though we
will not ignore owner-occupied housing, the main focus will be investment real estate, that is to
say, income-producing properties of the type and scale that are widely traded among professional
and institutional investors. We will use the term “commercial property” or “commercial real
estate” (CRE), with the understanding that this includes multi-family residential rental income
properties held in the private sector. We will make some international observations, but the main
focus will be on the United States.
Commercial real estate equity has become a major asset class in professionally managed
investment portfolios such as those of pension funds, sovereign wealth funds (SWFs), life
insurance companies, and other financial institutions. The market value of the stock of all real
estate in the U.S. is often said to equal about one-third of all investible wealth assets, and to
exceed the value of all publicly-traded equities (stocks). (Miles & Tolleson, 1997.) As of 2012
the U.S. Bureau of Economic Analysis National Balance Sheet (BEA 2014) lists a total current
cost net value of over $32 trillion for structures alone. If we assume, consistent with Davis and
Palumbo (2008), that land value is on average at least equal to the value of the structures on the
land, this implies total real estate value of over $60 trillion, four times the annual GDP. To put
this in perspective, the 2013 Federal Reserve Bank Flow of Funds Matrix (FRB 2014) lists total
financial assets of $195 trillion, including $34 trillion of corporate equities, suggesting that real
estate would be about a third of total assets and almost twice the value of all corporate equities.
3
Approximately half of the total real estate asset value is owner-occupied housing, but that still
leaves over $30 trillion in commercial property of various types. Most commercial real estate is
owned by private or government enterprises for their own use (owner-occupied commercial
property, sometimes referred to as “corporate real estate”). A recent private enumeration of CRE
asset value by Florence et al. (2010) based on the CoStar Group Inc database which seeks to
capture all commercial properties in the U.S. found total CRE value (including multi-family
rental) of approximately $11 trillion as of 2009, the low point of CRE value in the Great
Financial Crisis (GFC). It is reasonable to assume that about half by value of the CoStar
enumeration was largely non-traded corporate and owner-occupied real estate (mostly smaller
properties). Considering that the FRB reports approximately $3.3 trillion CRE mortgage debt
outstanding in 2013, these findings are also consistent with an approximate magnitude of at least
$6 trillion for traded investment CRE assets in the U.S. (assuming average outstanding loan-to-
value of about 50%). This compares to 2014 NYSE market capitalization of $16.6 trillion, $9.7
trillion of U.S. corporate bonds outstanding, $3.7 trillion of municipal bonds (SIFMA 2014). The
point is, CRE is a major investment asset class by market value of the stock of assets.
In spite of this, relatively little academic research attention has focused on CRE. When,
beginning in the 1950s and 60s, financial economics emerged as a rigorous discipline from its
origins in general economics, its focus was largely on stocks and bonds and their derivatives as
the subjects for investment theory and capital market theory, and this still remains largely true
today. When, in the 1970s and 80s, real estate economics emerged as a rigorous field from its
origins in micro-economics and in urban and spatial economics, its focus was largely on housing
and the commercial space usage (rental) market rather than on the commercial property asset
market, and this still remains largely true today. It was not until the late 1980s and the 1990s that
4
modern financial economic theory began to be rigorously applied to the study of commercial
property asset markets. Some pioneering studies included those by Ibbotson & Siegel (1984),
Hartzell et al. (1986), Firstenberg et al. (1988), Geltner (1989), Liu et al. (1990), Ross & Zisler
(1991), Giliberto (1992), Gyourko & Keim (1992), Giliberto & Mengden (1996), and Ling &
Naranjo (1997). The pioneering textbook for this integration of financial economics and real
estate economics was Geltner and Miller (2001, presently Geltner at al. 2014), though earlier
strands of academic literature notably included publications coming out of the appraisal
profession (see Lusht 1988, 1997). Even today the quantity, quality, depth and breadth of CRE
asset market academic literature remains relatively lean. But it is in this modern fusion of
financial economics and urban economics that we find the important academic work on CRE
price indices and price dynamics, with the academic literature importantly supplemented by
industry research and development. This integrated field of literature forms the source from
which this review article is primarily drawn.
With the above background in mind, this article will proceed in four sections and a
conclusion. Section 2 will present some important considerations about investment property asset
markets. Section 3 will present basic conceptual and definitional issues about price indexing of
CRE and some implications regarding price dynamics. Section 4 will briefly review the major
methods and considerations for producing real estate price indices. Section 5 will discuss some
interesting results of CRE price indexing including a brief comparison to housing and
international results. Section 6 concludes.
5
2. SOME CONSIDERATIONS ABOUT REAL ESTATE ASSET MARKETS
To begin, it will be instructive to consider why CRE assets have remained largely outside
the spotlight of mainstream financial economics. An important reason is the nature of the
markets in which the assets are traded. Broadly, it might be said that there are three major types
of markets in the world. First, there are the markets for mobile or perishable goods and for
services and information, which are the major and most widespread markets studied in micro-
economics. Such markets have been around since ancient times, and are arguably a natural and
omnipresent institution of human culture, from the agora of ancient Greece and bazaars of the
Middle East to the Walmarts and internet of today. All of these goods markets share fundamental
commonalities across history and across cultures, enabling very general and widely applicable
conclusions to be drawn from studying them, about supply and demand, equilibrium pricing,
competition, and so forth. Second, there are public securities markets or exchanges, for stocks,
bonds, commodity contracts, and derivatives. Such institutions are a relatively new and
specialized invention in human history, dating largely only from the seventeenth or eighteenth
centuries. Public securities exchanges form the main focus of the capital markets and
investments theory branch of modern financial economics. Like goods markets, securities
markets share fundamental commonalities across history and countries, facilitating their
scientific study. Securities markets have a place in CRE price indexing and the study of CRE
price dynamics, as we will note in this article, but they are not our central or fundamental focus.
It is the third major type of market, the private real estate market (or “property market,”
meaning, the market for the ownership of real property assets, land parcels including the
permanent structures on them) that must be our main focus, and whose unique characteristics
explain why CRE has been relatively ignored by mainstream financial economics.
6
Real property asset markets are very widespread, and in many countries go back many
centuries. But they are not nearly as ancient as real property ownership itself and the transfer of
such ownership. Human beings are territorial creatures. Until the industrial revolution almost all
income and wealth, and therefore power, derived from land. The control, effectively the
ownership, of land was certainly exchanged since ancient times, but it was not usually done so
by the market mechanism. Blood was the exchange mechanism, whether by warfare, marriage,
or inheritance. As civilization progressed, more and more of the exchange of real property began
to occur voluntarily in return for payment, and markets arose. But in many countries they arose
out of elements of the local culture different from the local goods markets. Nor did real estate
generally find its exchange through public securities markets (even where these existed). As a
result, unique institutions, customs and procedures arose for real property markets that can differ
notably across countries. This heterogeneity makes scientific study more challenging.
In some places, the “markets” for exchanging real estate may involve pricing that still
defers more to traditional formulas than to what economists would regard as the free intercourse
of supply and demand that results in free market equilibrium pricing such as prevails in goods
markets and public securities exchanges. Freely functioning (sometimes referred to as “arm’s
length”) real estate markets that reasonably well reflect the economist’s concept of equilibrium
pricing have the most history and ubiquity in the Anglo-Saxon countries and some lands of
northern and western Europe, though modern real estate markets are becoming more widespread
around the world. But even in places like the United States, which has among the most active
and unfettered real estate markets in the world, there may be pricing influences from professional
appraisal practice which go back far in time and reflect more traditional procedures. Indeed, a
characteristic of real property markets is that there is often an entrenched specialized pricing
7
profession, referred to as “appraisal” or “assessment” or “valuation,” which has varying degrees
of influence in the actual conduct and operation of the market. Though this influence can
sometimes be a bit exogenous to internal market equilibrium forces, appraisals can often
nevertheless be tapped as a unique source of information about CRE prices and values over time.
The point is that the nature and functioning of real estate asset markets not only differs
from that of public securities markets but also is more heterogeneous across countries.
Furthermore, the nature of real estate markets has meant that it is much more difficult to observe
and obtain large quantities of pricing and trading data, compared to what is possible for the other
major types of markets. This is because assets are traded in deals that are private, generally
between one buyer and one seller. While some jurisdictions require the public recording of the
price and of some characteristics of the traded asset, this requirement is not ubiquitous, and, has
not generally resulted in a centralized, easily compiled database (though progress is being made
in this regard). These features have no doubt made CRE a less appealing subject for study by
academic financial economists.
Another feature of the property market is that it trades unique, whole assets each one of
which is traded (and therefore priced) only infrequently and irregularly through time. This poses
an intellectual challenge to the construction of price indices necessary for the study of the asset
class’ pricing dynamics. While this challenge may be regarded by some financial economists as a
barrier, it was, in effect, welcomed by urban economists as a fascinating problem to solve. And
solve it they have, to a considerable degree, as we will review in this article. The infrequent
trading of unique long-lived assets also gives rise to the need to value (that is, to evaluate, or
estimate the value of) CRE assets more frequently than they are traded priced in the market by
actual consummated ownership exchange transactions. This need is a major raison d’être for the
8
appraisal profession, which as previously noted can also be a source of information useful for our
purposes.
To make this consideration of real estate markets more concrete for our purpose of
understanding price dynamics, suppose you own a CRE asset, “123 Sesame Street,” and you
want to sell it. Unlike when you want to sell a stock or bond you might own, you can’t sell your
property immediately. You will hire a broker, who will put together an information package
about the property and disseminate it in various ways, probably requesting written sealed bids by
a certain date (likely several weeks or months in the future). You may or may not have posted a
suggested minimum bid, or alternatively, an offer price, as this depends on the selling strategy
you and your broker have decided on based on the nature of the property, the condition of the
market, and how urgently you desire to sell the property. You have probably recently engaged a
professional appraiser who has given you advice on what she thinks is the “most likely” or
“expected” price at which the property will sell. (This is referred to as a “market value”
appraisal.)
Your sale of 123 Sesame Street is in competition with other similar properties nearby and
not so nearby, competing for the investment dollars of various possible types of buyers, which
might include private taxable investors, investment managers working for tax-exempt pension
funds or various types of private equity funds, REITs, life insurance companies buying for their
own account, foreign investors, and so on. Some of these buyers might be counting on borrowing
money in the commercial mortgage market, and/or bringing in equity joint venture partners, in
order to come up with the cash they will need to make the purchase. The easier it is for your
potential buyers to raise their sources of capital, the more potential buyers you will have and,
ceteris paribus, the lower will be the opportunity cost of capital (OCC) and the higher will be the
9
price you are likely to be able to negotiate. The reason why 123 Sesame Street is in competition
not just with close-by properties but also with some that could be quite distant, is because many
of these potential buyers can place their capital anywhere they think they find the best
investment.
When you receive the sealed bids, you will probably select a few of the most appealing
ones, based both on the bid price and also on other terms and conditions of the bid, and also on
the reputation you can ascertain about the bidder’s ability and propensity to expeditiously close
the transaction without playing too many “games.” You will invite that small number of first-
round bidders to a second round of “best and final bids.” From among those you will select the
trading partner with whom you will attempt to negotiate the details of the sale, which, if all goes
well, will close within a few weeks or months after detailed due diligence has been performed by
the buyer.
Now suppose that just before, or during, this process, news arrives that is clearly relevant
to the value of 123 Sesame Street. It might be news about the local rental market, such as an
announcement of a large new tenant, perhaps a corporate branch headquarters, looking for space
in the market. Or it might be national or international news about interest rates, or the macro-
economy. Suppose in fact it is bad news. How will this news affect the sale price of 123 Sesame
Street? Will it affect the price at all as such, or perhaps instead the likelihood of the sale
happening, or happening within a given time frame? We cannot know, no one can, the answers
to these questions with great precision or certainty.
In contrast, if 123 Sesame Street were a vast asset carved up into millions of identical
common shares of equity ownership actively traded on a public stock exchange, we would
immediately (or almost immediately) be able to observe a current equilibrium asset price
10
reaction to the news. We would see exactly how much the liquid value of 123 Sesame Street
equity shares fell on the day of the news and subsequent days. This price drop would reflect a
market consensus among numerous buyers and sellers all trading exactly the same asset
(common shares in 123 Sesame Street) at immediately publicly-quoted prices. The speed of the
information aggregation and price discovery would be much faster than what must occur with
your actual 123 Sesame Street in the private property market. (See Gyourko & Keim (1992),
Barkham & Geltner (1995), Clayton & MacKinnon (2001), Geltner et al. (2003), Yavas &
Yildirim (2011), Ang et al. (2013), Bond & Chang (2013), among others.)
------------------------------- Insert Figure 1 about here. -------------------------------
Now consider Figure 1, which depicts the functioning of the CRE asset market, in effect
the market for 123 Sesame Street and other similar properties competing in the investment
marketplace. The chart on the left shows the frequency distributions of reservation prices, sellers
(property owners) on the right (dashed line), and buyers on the left (solid line). The reservation
price frequency distributions in the left-hand chart correspond to demand (solid line) and supply
(dashed line) functions in the property asset market depicted in the right-hand panel (cumulative
integrals under the reservation price frequency distributions). The potential participants on both
sides of the market are heterogeneous, as are the traded assets, and these heterogeneous agents
must search for and find trading partners for the heterogeneous assets in what is effectively a
double-sided private search market (Wheaton, 1990). There is thus a range or distribution of
reservation prices on both sides, which means that the supply and demand functions in the asset
market are sloped; there is no infinite elasticity as has been often assumed in the classical capital
market theories of financial economics. Thus, the flow of financial capital into and out of the
11
property market influences the pricing in that market. (Fisher et al., 2003.) As transactions can
only be consummated between buyers whose reservation price is at least as great as that of a
corresponding seller, actual sales will only be drawn from the overlapping roughly triangular
intersection region below both frequency distributions. The larger this region, the easier it will be
for willing partners to find each other and the greater will be the transaction volume in the
market. But in any case, the heterogeneity causes individual asset transaction prices to be
dispersed around a market equilibrium central tendency value. Thus, individual transaction price
observations are “noisy” (Case & Shiller, 1987, 1989).
In real estate markets demand (potential buyers’ reservation prices) tends to move sooner
or faster or farther than supply (property owners’ reservation prices, which seem to be more
“sticky”). Thus, when demand is increasing (decreasing) and equilibrium prices rising (falling),
the overlap between the two frequency distributions increases (decreases). This pattern, with
dispersed reservation prices and demand moving more or sooner than supply, results in, and
reflects, a salient feature of such markets, which is that trading volume is highly variable and
pro-cyclical: volume and price move together. Usually volume moves slightly ahead of price, or
is able to be tracked slightly ahead of price movements. (Fisher at al., 2007.) In real estate
markets, trading volume is viewed as reflective of “liquidity” in the market, the ease or speed
with which properties can be sold, reflecting the strength of demand relative to supply. An
implication is that consummated price levels alone do not provide sufficient information to fully
represent the state of the market. Volume also is important. A price index reflecting only the
movements in the observable transaction price levels will be to some extent “apples versus
oranges” between up-markets and down-markets. More property assets can be sold quicker and
12
more easily at the observed prices in the up-market than at the observed prices in the down-
market. (Goetzmann & Peng, 2006.)
3. PRICING AND PRICE INDEXING IN THE PROPERTY ASSET MARKET
With these fundamentals about market functioning in mind, and also considering the
nature of asset price discovery as described in our 123 Sesame Street example, consider Figure 2,
which depicts four different conceptual constructs of real estate price indices. The four indices in
the Figure are actual empirical CRE asset price indices in regular production for commercial
investment property in the United States, covering the period 2001-2014. Three of the four
indices in the Figure are based on the private property market, in particular, on the portfolio of
investment properties tracked by the National Council of Real Estate Investment Fiduciaries
(NCREIF), the first organization in the world to publish a regularly updated and widely used
index of commercial property values. The official NCREIF Property Index (NPI), which dates to
1984 (with index inception 1977), is based on appraised values self-reported by NCREIF’s
member firms, which represent most of the major tax-exempt real estate investing institutions in
the country, dominated by pension funds. Presently the NPI includes over 7,000 properties worth
over $350 billion. The fourth index in Figure 2, the FTSE-NAREIT PureProperty® Index, is
based on the stock market. The over 100 real estate investment trusts (REITs) in the
PureProperty Index presently hold over 30,000 properties worth over $800 billion. Both the
REITs and the NCREIF members hold broadly similar types of commercial properties, generally
relatively large, prime investment properties generating stabilized income, dominated by office
buildings, retail stores, multi-family apartment properties, and industrial warehouse type
13
properties, the so-called “core” sectors of the investment property market, located in major
metropolitan areas throughout the U.S.
------------------------------- Insert Figure 2 about here. -------------------------------
The four indices in Figure 2 depict different ways and conceptions for measuring and
defining investment property price change in the U.S. But before delving into the differences, it
is important to note that all four of the indices present a broadly similar picture of the almost 15
years of history beginning in 2000. They all show a huge boom (arguably ultimately a “bubble”)
beginning by 2004, peaking around 2007-08, followed by a tremendous crash bottoming around
2009-10, and then a strong recovery bouncing right out of the trough and continuing through
early 2014. The major differences across the indices are in the exact timing of the major turning
points in the prices, and to some extent in the magnitude of the cycle and/or in the apparent
volatility and inertia in the price movements. Let us consider each of the four indices in turn.
The first index in Figure 2 in terms of the timing of the major cyclical turning points is
the stock market based PureProperty Index. From our earlier discussion of the sale of our 123
Sesame Street example property and its contrast to a public securitized asset, it is not surprising
that the stock market based valuation of commercial property moves ahead of the private market
based valuations in time. The PureProperty® capital return index depicted in the Figure is
derived from daily price movements of REIT common equity share prices on the stock market,
with the index being de-levered to reflect the stock market’s implied valuation of the underlying
property assets held by the REITs. It is thus a daily-frequency, liquid price indicator for the CRE
assets held by REITs. By using the stock and bond markets (the latter to offset leverage in the
traded REIT equity shares), investors can actually trade in the stock market at prices reflected by
14
the PureProperty Index. As REITs are “pure play” type firms specialized in holding commercial
property assets, stockholders can be under no illusions that they are essentially investing in
commercial property, so it must be presumed that REIT share prices essentially reflect the stock
market’s valuation of those assets.
It may appear visually that the PureProperty Index is more volatile than the other indices
in the Figure, and real estate investment practitioners often claim that the stock market adds
volatility to real estate assets. But this is arguably an illusion of the different frequencies at
which the indices are reported. As noted in the table of statistics for the indices (included in the
Figure), the quarterly volatility for the PureProperty Index is no greater than two of the three
private market based indices in Figure 2. However, the REIT-based index does register some
short-run price movements that are not echoed by the private market based indices. During the
run-up in prices in the great bull market before the GFC the PureProperty Index registered at
least four sharp, but temporary, downturns, in spring 2004, early and mid-2005, and spring 2006,
each of which corresponded to an increase in FRB interest rate policy. The PureProperty Index
also registered a sharp (but again temporary) pull-back in the nascent price recovery in mid-
2011, at the time of the debt ceiling brinksmanship and S&P downgrade of U.S. Treasury debt.
The stock market is designed to enable prices to move to whatever level they need to in
order to maintain liquidity, and therefore the price movements depicted by the PureProperty
Index in Figure 2 reflect highly liquid values throughout the cycle. The other three indices in
Figure 2 reflect the private property market, in different perspectives, and therefore we must
consider the variable liquidity point raised earlier in our discussion of Figure 1.
The index that moves next, after the PureProperty Index, is the NCREIF NTBI Demand
Index, which represents “constant liquidity” price movements. This index tracks estimates of the
15
relative movements in the central tendency of potential buyers’ reservation prices on the demand
side of the investment market. It is an econometrically derived index, produced by the MIT
Center for Real Estate, based on both the sold and unsold properties in the NPI database, using a
procedure described in Fisher at al. (2003, 2007). (A similar approach is presented in Goetzmann
and Peng, 2006.) By modeling at the disaggregate level both the transaction prices at which
individual properties actually trade and the probability of their trading based on the same
observable characteristics of the properties, it is possible to infer the movements in the
reservation prices on both sides of the market (supply and demand). If sellers were willing to
exactly follow the movement in potential buyers’ reservation prices at all times in the cycle, then
it would be as easy to sell as many properties in the down-market as in the up-market, and the
property market would be characterized by constant liquidity. Thus, the demand-side reservation
price index depicted in Figure 2 represents how market prices would have had to have moved in
the private property market in order to keep constant liquidity in the market. This index therefore
attempts to collapse into a single price-dimensioned metric a more complete representation of the
condition of the asset market as reflected on both the dimensions of the consummated prices and
the trading volumes (liquidity). In some sense the constant-liquidity private market price index is
more directly comparable to the always-liquid stock market based PureProperty Index. With this
in mind it is interesting to note that the constant-liquidity private market index moves slightly
later than the stock market based index (perhaps a one to two quarter lag during the depicted
history).
The constant-liquidity index in Figure 2 moves slightly ahead, and in a more exaggerated
manner, than the next index in the Figure, which is the “Transactions Based Index” published by
NCREIF (the NTBI Price Index). The NTBI tracks the movements in the actually consummated
16
transactions prices of properties sold from the NPI database, prices that reflect both the supply
and demand sides of the market. The fact that the transaction price index moves slightly behind
and not as far as the demand-side index reflects the tendency of property owners to display
“sticky pricing,” particularly in the down direction. Both sides of the market react conservatively
in the face of uncertainty such as occurs around major downturns, meaning that buyers reduce
their reservation prices while sellers may actually increase theirs, causing or reflecting a drop in
“liquidity” (trading volume). The prices in the deals that do get done do not reflect as great a
drop as has occurred in most potential buyers’ reservation prices. Thus, the transaction price
index fell only 34% in the GFC and its aftermath, while the constant-liquidity index fell 42% and
the stock market based index fell 38%.
The last of the four indices in Figure 2 is actually the first in the timing of its historical
development and is still the most widely used in industry practice among the four indices in
Figure 2. As noted, the NPI is based on the self-reported appraisals of the properties held by the
NCREIF members. The version of the index shown here is equally-weighted, that is, counts the
appraisal-based percentage change in value for each property equally in computing the average
same-property value change within each quarter. The version of the NPI capital return shown
here also includes the effect of capital improvement expenditures during the period. (That is, the
period’s “capex” is not subtracted from the end-of-period valuation, as is done in the official NPI
appreciation index.) It is thus directly comparable to the other three indices in the Figure
described previously. The NPI is lagged behind the TBI in the major turning points, and is
smoother with a dampened cycle amplitude (the boom was +77% versus +85% and the crash was
-28% versus -34%). Alone among the four indices in Figure 2, the NPI has strong positive first-
order autocorrelation, over 80%, reflecting strong quarterly inertia. These differences between
17
the NPI and TBI are characteristic of appraisal-based versus direct transaction price based
indices of commercial investment property prices. They reflect two phenomena. First, the
appraisal of individual property values reflects procedures of professional appraisal practice that
tend to result in some temporal lagging bias, and properly so, as appraisers need to document
their valuation estimates based on historical transaction price evidence, and as they need to filter
out the noise that exists in individual transaction prices. (Quan & Quigley, 1989, 1991.)
Secondly, although pension funds and their property investment managers are required by law
and by NCREIF rules to reappraise each property at some frequency, this is rarely done for every
property every quarter. Yet the quarterly NPI includes all properties every quarter, including
those that are not effectively reappraised in the current quarter and thus are reported into the
index at a prior (“stale”) appraisal value. In spite of these problems, and the implications they
have for the apparent price dynamics revealed by the NPI, we see that in the big picture the price
history depicted by the stale appraisal-based NPI is not terribly different from that depicted by
the consummated transaction price based NTBI. Of course, the types of differences revealed in
Figure 2 can be important for some purposes.
To summarize this discussion of investment property pricing and price indexing, it is
clear that there are important differences between real estate and the securities markets that
financial economists are most used to studying. The differences are fundamental, and can be
important in many types of analyses and studies. Indeed, we see that some of the issues are
conceptual and definitional in nature, not merely reflective of less data for real estate markets.
But we also see that a variety of different types of price indices can be defined and measured, at
least in some contexts.
18
4. METHODOLOGY OF PROPERTY PRICE INDEXING
While stock market or appraisal based indices of property prices can be useful and offer
interesting perspectives, the most direct and fundamental source of information about property
asset prices are the consummated transaction prices for individual assets that occur in the private
property market. How do urban economists and econometricians use this type of information
directly to compute transaction based price indices?
The basic concept of a price index is that it compares the change over time in the price of
the same good. In computing such an index for real estate assets there are two major technical
challenges. The first is that the “good” that is traded (real property assets) is heterogeneous, as
each property is unique, and the individual assets are traded only infrequently and irregularly
through time. Indices cannot be computed directly from property market transactions data with
simple matched samples or with stock market price indexing procedures. The second challenge is
that there is often relatively sparse transaction data, especially for commercial property. This
results in small (and often variable) sample sizes in the relatively short intervals of time to be
tracked by a periodic price change index. This second problem is closely related to the
segmentation that exists in the space usage (rental) market, such that properties in different
locations can have different price dynamics, reflecting different supply and demand
characteristics in the space market. Volatility, cyclicality, and the role of structure depreciation
(hence, long-term price trend rates) can all differ across locations due to factors such as land
availability and local economic and political environments. This means that relatively
“granular,” segment-specific indices are required for some purposes. Compounding the sparse
data problem is the transaction price “noise,” or dispersion, noted in Section 2. Evidence
suggests that individual property price movements can be quite idiosyncratic (Geltner et al. 1994,
19
Hansz & Diaz 2001, Shimizu & Nishimura 2006 & 2007). Noise in the price index, which
directly reflects random statistical estimation error caused fundamentally by sparse data and
individual transaction price dispersion, is one of the two major problems that transaction price
index estimation methodology seeks to minimize. The other major problem is the potential for
bias, as transaction samples may not be fully representative of the type and location of property
one is interested in studying, or data may have omitted variables that bias the index. (See Case et
al. 1992, Gatzlaff & Haurin 1997 & 1998, and Guo at al. 2014 for index comparison
methodologies.)
Real estate transaction price indexing that attempts to control for the heterogeneity issue
in a statistically rigorous and sophisticated manner has become widespread primarily only in the
past generation, and largely focused on single-family housing. Investment property transaction
price indices have been developed primarily only since the turn of the present century. (Early
exceptions include Fisher et al. 1994, and Gatzlaff & Geltner 1998.) Two major different
approaches to controlling for heterogeneity have characterized the development of transaction
price indices. The first, hedonic indexing (based on hedonic price modeling), controls for
heterogeneity by modeling property assets as bundles of characteristics (hedonic variables). This
approach goes back to general economic price indexing of goods whose quality changes over
time (Court 1939, Griliches 1961), and was early related to real estate (Kain & Quigley 1970),
and developed further into equilibrium modeling and formal index number theory notably with
Rosen (1974). With hedonic data one can regress prices onto characteristics and then compute an
index by holding the property characteristics constant, revealing the pure price changes. The
basic hedonic techniques were honed by the 1980s for housing. A single definitive specification
and functional form was generally found to be illusive, though most models have tended to use
20
log prices as the depended variable. (See Halvorsen & Pollakowski 1981, Pollakowski 1982, Hill
2013.)
There are two major different hedonic specifications, the “chained” or “imputed,” and the
“pooled” or “time-dummy.” The former, which re-estimates the price model on each period’s
data, is the primal and theoretically more complete and flexible specification, the preferred
model in principle (for example, it allows rigorous Laspeyres or Paasche price indexing useful in
national accounts, and it satisfies “time-fixity” in that data updates do not generate backward
historical revisions). But the chained hedonic is the most data intensive, and requires specifying a
“representative property” to construct the index. The alternative is the pooled or time-dummy
specification, which runs a single regression on the entire history with the time-dummy
coefficients providing the price index. The pooled model is less data intensive but constrains the
relative implicit prices to be invariant, thereby providing less structural information and a less
“pure” price index (and making the index subject to backward revisions). A hybrid approach is
the “rolling window” imputed model (de Haan & Diewert 2011). Most hedonic indices in the
academic literature have been of the pooled dummy-variable type, and they are especially useful
for commercial property due to data scarcity.
The major alternative to the hedonic approach is the repeat-sales regression. With data on
properties that sell more than once, one can regress the log price changes over time within the
same properties onto time dummy-variables representing the historical periods in the index, thus
producing the repeat-sales index. The coefficients on the time-dummies trace out the price index
based on the price changes experienced by the investors between the “buy” and the “sell” of the
investment asset. This approach dates from Bailey et al. (1963), Bryan & Colwell (1982), and
Case & Shiller (1987, 1989). The repeat-sales specification can be derived as a special case of
21
hedonic price modeling, ignoring changes in the hedonic values of the property between the buy
and sell dates (e.g., depreciation is included in the price index unless explicitly addressed as in
Shimizu Nishimura & Watanabe 2010). But the repeat-sales index can also stand on its own as a
primal specification by viewing the object to be measured as the price-change experiences of the
homeowners or investors (the dependent variable in the regression).
Following Case & Shiller’s seminal publications, there was a flowering of interest in
repeat-sales indices for housing, stimulated in part by the idea of using them to form tradable
house price futures contracts (Shiller 1993). Some push-back to repeat-sales indices came from
those concerned about sample selection bias (Gatzlaff & Haurin 1997 & 1998, Meese & Wallace
1997, Munneke & Slade 2000 & 2001), and backward revisions in index histories as the repeat-
sales estimation sample is updated (Clapham et al. 2006). In general, repeat-sales samples tend to
show greater house price appreciation due to the nature of the repeat-sales sample (properties
that appreciate more tend to be sold more frequently). And the nature of the estimation
methodology causes repeat-sales indices to be particularly susceptible to historical revisions
(updates of the repeat-sales sample inherently bring into the index the effect of first-sales that
may have occurred far back in history).
To date, both in housing and commercial property, most of the regularly published
indices have been repeat-sales rather than hedonic, particularly in the case of commercial
property. (The only regularly published hedonic index of commercial property is the recently-
launched vdpResearch Property Price Index of the Association of German Pfandbrief Banks,
tracking property prices in Germany.) The popularity of repeat-sales indices in practice is
probably due to the fact that they do not need hedonic data, are relatively robust to specification
error and omitted variables, and they appear relatively transparent and easy to understand for
22
industry practitioners and the public. By the mid-1990s repeat-sales housing indices were being
published on an on-going basis by Case-Shiller-Weiss Inc (now the S&P/Case-Shiller Home
Price Index) and U.S. Federal housing regulators (now the FHFA House Price Index, see
Abraham & Schaumann 1991, Calhoun 1991, 1996). Nevertheless, hedonic indices have also had
substantial academic development, particularly in housing and urban studies, and they are also of
interest in national economic statistics.
In fact, the innovation and honing of real estate transaction price indexing methodology
during the past two decades has been quite impressive. Not only has index methodology been
perfected, but large scale transaction databases have been developed, first in housing and more
recently in commercial property. Hybrid specifications that combine elements of hedonic and
repeat-sales indices have been developed and advocated. (See Case & Quigley 1991, Case et al.
1991, Quigley 1995, Knight Dombrow & Sirmans 1995, Meese & Wallace 1997, Hill Knight &
Sirmans 1997, Clapp & Giacotto 1998.) At one point there was even a formal contest to see
which index methodology was “better,” based on a common dataset made available to all
academics. The result of the competition basically argued for a hybrid approach if sufficient data
is available (Case et al. 1991). By the mid-1990s review compendia were being published in
special issues of the leading real estate journals (Thibodeau 1997). Truly a plethora of tools and
techniques are now available. For example, hedonic indices have been combined with appraisal
or assessment data to deal with omitted variables or sparse hedonic data (Clapp & Giacotto 1992,
Fisher at al. 2003, Fisher et al. 2007, Devaney & Diaz 2011, Gatzlaff & Holmes 2013). Repeat-
sales techniques have been elaborated to allow value-weighted and arithmetic average indices
(Shiller 1991) and total return indices (Geltner & Goetzmann 2000). Various methods for dealing
with sparse data have been developed, including: Bayesian techniques (Goetzmann 1992,
23
Francke 2010); Pre-regression data manipulation methods forming matched samples that
combine hedonic and repeat-sales approaches or effectively expand the usable sample size
(McMillen 2012, Deng et al. 2011, Guo et al. 2014); Time parameterization specifications
(McMillen & Dombrow 2001, Francke & Vos 2004, Francke 2008); Estimation techniques such
as spatiotemporal autoregressoin (Pace et al. 1998, LeSage & Pace 2010, Chegut et al. 2014);
and Post-regression index construction techniques such as filtering and curve-fitting (e.g.,
Hodrick & Prescott 1997) and frequency conversion (Bokhari & Geltner 2012). Recently a
specialized hedonic specification has been proposed to allow the simultaneous estimation of
structure value and land value indices from property asset price data, consistent with economic
statistical national accounts criteria (de Hann & Diewert 2011, Diewert & Shimizu 2013).
5. SOME FINDINGS FROM PROPERTY PRICE INDICES
The developments in property price indexing described in the previous section have
helped to shed light on some interesting questions about real estate price dynamics. Some of
these have already been noted in Section 3. There we noted the tendency of private property
market asset prices to move more sluggishly than corresponding liquid public securities market
prices, although with volatility and cyclical amplitude of similar magnitude, and we noted that
appraisal-based price indices tend to be even more lagged, as well as somewhat smoothed and
damped. In this last section we will take a broader look at a few other points of interest.
An important characteristic of investment property pricing that is revealed in part by
price indexing is that real estate asset market pricing seems to share the characteristic of
securities markets asset pricing that much of the volatility and cyclical price movements are
driven more by capital market forces than by news or changes in the underlying operating cash
24
flows that fundamentally undergird the value of the capital assets. Apparently, at least in the case
of investment property, the relatively high friction in the private property market (search and
transactions costs), which in part causes the sluggishness noted in Figure 2, does not prevent
asset pricing from exhibiting this type of “excess volatility” first suggested by Shiller (1981).
Viewed from the perspective of the classical present value model, longitudinal changes in the
discounted present value of property assets in the U.S. are driven more by changes in the
discount rate (the opportunity cost of capital, or realistic expected total returns going forward)
than by changes in the expected future net cash flow stream. (Geltner & Mei 1995, Plazzi et al.
2010.)
A second important feature of investment property pricing that echoes major findings in
the stock market is that there exists a type of pricing factor that is not immediately related to risk
or volatility as these can be easily quantified and understood in classical financial economic
capital market theory, a type of real estate analog to Fama-French pricing factors in the stock
market (Fama & French 1992, 2004). In the case of commercial real estate, the non-risk factor is
not exactly the same as the market capitalization and book/market ratio factors made famous by
Fama and French, but it may have some commonality with some aspects of those. In real estate
the most obvious and persistent non-risk pricing factor is the distinction between what is often
termed “institutional” and “non-institutional” investment property. It has long been noted from
surveys and empirical transaction price data that smaller (by value) properties that tend to be
located in less prime locations or that have older or lower-quality structures trade at higher initial
cash flow yields. The gap is often in the neighborhood of 200 basis-points or even more.
Commercial property price index data now allows one to see whether this gap in cash yield is
offset by either a lower price growth trend or greater volatility in the “non-institutional”
25
properties, and in fact this does not appear to be the case in either respect, at least since around
the turn of the century when the commercial property transaction price based indices provide a
historical record (Geltner et al. 2014, Chapter 25). In particular, the CoStar Commercial Repeat-
Sales Index (CCRSI, launched in 2010, inception 1998) is based on all commercial property
sales in the U.S. and is reported separately for “Investment” (corresponding to “institutional”)
and “General” (corresponding to “non-institutional”) properties. From 1999 to 2014 the average
annual price growth rate was 3.0% for institutional property and 4.2% for non-institutional, while
the annual volatility was 10.3% for institutional versus 9.6% for non-institutional (CoStar Group
2014). The magnitude of the price crash in the GFC and subsequent recession was 40% (of peak
value) in the institutional index versus 36% in the non-institutional index. The greater total
returns provided by non-institutional properties do not seem to be justified by greater “risk” as
traditionally measured. This “non-institutional factor” may reflect segmentation in the financial
markets, as different sources of financing and different types of investors are typical of the two
different types of properties. There may also be an uncertainty premium in the smaller properties,
as information may be less transparent among them.
------------------------------- Insert Figure 3 about here. -------------------------------
A third interesting point that is revealed by real estate price indices regards the similarity
and difference between commercial (investment) property and single-family housing price
trends, including the relationship to the real economic business cycle. Figure 3 depicts a longer
term historical extension of the NCREIF NTBI index of commercial property prices described in
Figure 2 (Geltner 2013) along with Shiller’s long-run historical home price index and the
Consumer Price Index (Shiller 2014). Recessions are indicated by the shaded bars. Both of the
26
real estate price indices track nominal values of same-property prices (i.e., they include the effect
of both structure depreciation and capital improvement expenditures). It is interesting to note that
in the U.S., single-family home values have shown a greater long-term price growth trend than
commercial real estate with home prices generally slightly outpacing inflation, while investment
properties track slightly below inflation. At the broad-brush aggregate level shown here, home
prices (as distinct from sales volume) were not very cyclical until the great crash of 2007-09,
while commercial properties experienced two major crashes prior to the mid-2000s, one in the
early/mid-1970s and another in the late 1980s/early’90s. Investment property may have had
greater cyclical tendencies than single-family housing for most of the post-war history. On the
other hand, housing is a much larger sector of the economy, with a consumption multiplier effect
that CRE lacks, making housing much more influential than CRE in the business cycle.
Commercial real estate tends to be a lagged effect of the business cycle (particularly the space
market), while housing has often been a leading or causal factor, though until the GFC housing’s
influence was not primarily through price change.
The fourth and final point which CRE price indexing can suggest in a rather graphic way
relates to one of the initial points raised at the beginning of Section 2, about the nature of
property markets as highly heterogeneous cultural phenomena. Figure 4 displays two price
indices of commercial investment property assets in Germany, quarterly from the end of 2009 to
through 2012. The vpd Property Price Index is a hedonic transactions price based index of
commercial property produced by the German mortgage bankers association. The index labeled
“PureProperty” is based on publicly traded European REITs and listed property companies
holding properties in Germany, de-levered in the manner described in Section 3 regarding the
FTSE-NAREIT PureProperty® Index (Elonen 2013). In the case of Germany, the two indices
27
reflect not so much different types of data or index construction methods (though that is also the
case), but rather they highlight the different price dynamics of different types of asset markets,
trading fundamentally the same types of assets, income-producing commercial investment
property. The three year period covered by the chart were eventful ones in the German economy
and financial markets, including the initial recovery out of the GFC, followed by the crisis in the
euro in 2011, which by 2012 led to capital flight into Germany. The European stock markets on
which the “PureProperty” index is based are public securities markets, a type of market in which
prices reflect an equilibrium that moves very freely and quickly responding to news and shifts in
supply and demand. The private property market in which German investment institutions trade
the properties whose prices are reflected in the vdp Index is a much quieter and more traditional
arena. The resulting difference in price dynamics is obvious in Figure 4 (along with a little
statistical estimation noise in the vdp Index).
------------------------------- Insert Figure 4 about here. -------------------------------
6. CONCLUSION
The preceding discussion has covered a lot of ground. Real estate assets are a large and
very important component of the national wealth and the capital markets, and their price
dynamics present a fascinating and important phenomenon. Heterogeneity of various types,
among assets, markets, and data sources make the study of real estate pricing both challenging
and uniquely interesting, not least for the econometricians and urban economists who have
pioneered important new price indexing methodologies. Tremendous development in data
sources and price indexing over the past few decades, and still continuing apace both in the U.S.
28
and abroad, relating to both housing and commercial property assets, has already provided
important findings and insights about real estate pricing and price dynamics, and no doubt it will
continue to provide more.
29
Literature Cited
Abraham JM, Schauman WS. 1991. New evidence on home prices from Freddie Mac repeat
sales. Real Estate Econ.. 19:333-52
Ang A, Nabar N, Wald S. 2013. Searching for a common factor in public and private real estate
returns. J. Portf. Manag. Special Real Estate Issue:120-133
Bailey MJ, Muth RF, Nourse HO. 1963. A regression method for real estate price index
construction. J. Am. Statist. Assoc. 58: 933–42.
Barkham R, Geltner D. 1995. Price discovery in American and British property markets. Real
Estate Econ. 23:21–44
BEA. 2014. Table 2600. Balance Sheets for National Assets. Washington, DC: U.S. Bureau of
Economic Analysis. http://www.bea.gov/national/sna.htm
Bokhari S, Geltner D. 2012. Estimating real estate price movements for high frequency tradable
indexes in a scarce data environment. J. Real Estate Fin. Econ. 45:522–43
Bond S, Chang Q. 2013. REITs and the private real estate market. In Alternative Investments:
Insruments, Performance, Benchmarks and Strategies, ed. HK Baker, G Filbeck, pp.79-99.
Hoboken, NJ: John Wiley & Sons
Bryan T, and Colwell P. 1982. Housing price indices. In Research in Real Estate (2), Ed. CF
Sirmans. Greenwich, CT: JAI Press
Calhoun CA. 1991. Estimating changes in housing values from repeat transactions. Presented at
Intl. Meet. Western Econ. Assoc., Seattle
30
Calhoun CA. 1996. OFHEO House Price Indexes: HPI Technical Description. Washington, DC:
Federal Housing Finance Agency.
http://www.fhfa.gov/PolicyProgramsResearch/Research/PaperDocuments/1996-
03_HPI_TechDescription_N508.pdf
Case B, Quigley JM. 1991. The dynamics of real estate prices. Rev. Econ. Statist., 22:50–58.
Case B., Pollakowski H, Wachter S. 1991.On choosing between house price index
methodologies., Real Estate Econ. 19:286-307.
Case K, Shiller R. 1987. Prices of single family homes since 1970: New indexes for four cities.
Work. Pap. No. 2393. National Bureau of Economic Research
Case K, Shiller R. 1989. The efficiency of the market for single-family homes. Am. Econ. Rev.
79:125-37
Chegut A, Eichholtz P, Rodrigues P, Weerts R. 2015. Spatial dependence in international office
markets. J. Real Estate Finan. Econ. Forthcoming
Clapham E, Englund P, Quigley JM, Redfearn CL. 2006.Revisiting the past and settling the
score: index revision for house price derivatives. Real Estate Econ. 34:275-302
Clapp JM, Giaccotto C. 1992. Estimating price trends for residential property: A comparison of
repeat sales and assessed value methods. J. Real Estate Finan. Econ. 5:357-74
Clapp JM, Giaccotto C. 1998. Price indices based on the hedonic repeat-sale method:
Application to the housing market.,J. Real Estate Finan. Econ. 16:5-26
Clayton J, MacKinnon G. 2001. The time-varying nature of the link between REIT, real estate
and financial asset returns. J. Real Estate Portf. Manag. 7:43-54
31
CoStar Group. 2014. CoStar Commercial Repeat-Sale Indices. Washington, DC: CoStar Group
Inc. http://costargroup.com/costar-news/ccrsi
Court A. 1939. Hedonic Price Indices with Automotive Examples. The Dynamics of Automobile
Demand, General Motors Corp., Detroit, MI
Davis MA, Palumbo MG. 2008. The price of residential land in large US cities. J. Urban Econ.
63:352-84
de Haan J, Diewert WE, eds. 2011. Handbook on Residential Property Price Indexes,
Luxembourg: Eurostat.
http://epp.eurostat.ec.europa.eu/portal/page/portal/hicp/methodology/hps/rppi_handbook
Deng Y, Mcmillen D, Sing TF. 2011. Private residential price indices in Singapore: a matching
approach. Reg. Sci. Urban Econ. 42:485-94
Devaney S, Diaz RM. 2011. Transaction based indices for the UK commercial real estate
market: an exploration using IPD transaction data. J. Property Research, 28:269-82
Diewert WE, Shimizu C. 2013. Residential property price indexes for Tokyo. Work. Pap., Sch.
Econ., Univ. British Columbia
Elonen K. 2013. Tracking and trading commercial real estate through REIT-based pure-play
portfolios: The European case. MS Thesis, MIT., Cambridge, MA
Fama E, French K. 1992. The cross-section of expected stock returns. J. Finan. 47:427–65
Fama E, French K. 2004. The capital asset pricing model: Theory and evidence. J. Econ.
Perspectives 18:25–46
Firstenberg P, Ross S, Zisler R. 1988. Real estate: The whole story. J. Portf. Manag. 14:22-34
32
Fisher J, Gatzlaff D, Geltner D, Haurin D. 2003. Controlling for the impact of variable liquidity
in commercial real estate price indices. Real Estate Econ. 31:269–303
Fisher J, Geltner D, Pollakowski H. 2007. A quarterly transactions-based index of institutional
real estate investment performance and movements in supply and demand. J. Real Estate Finan.
Econ. 34:5-33
Fisher J, Geltner D, Webb RB. 1994. Value indices of commercial real estate: A comparison of
index construction methods. J. Real Estate Finan.Econ. 9:137–64
Florance AC, Miller NG, Peng R, Spivey J. 2010. Slicing dicing and scoping the size of the U.S.
commercial real estate market. J. Real Estate Portf. Manag. 16:101-18
Francke M. 2008. The Hierarchical Trend Model. In Mass Appraisal Methods: An International
Perspective for Property Valuers, ed. T Kauko, M Damato, pp. 164-180. Oxford: Wiley-
Blackwell.
Francke M. 2010. Repeat sales index for thin markets: a structural time series approach. J. Real
Estate Fin. Econ. 41:24–52
Francke M, Vos GA. 2004. The hierarchical trend model for property valuation and local price
indices. J. Real Estate Finan. Econ. 28:179-208
FRB. 2014. Z.1 Financial Accounts of the United States. Flow of Funds, Balance Sheets, and
Integrated Macroeconomic Accounts. Washington, DC: Board of Governors of the Federal
Reserve System
Gatzlaff D, Geltner D. 1998. A transaction-based index of commercial property and its
comparison to the NCREIF Index. Real Estate Finan. 15: 7–22
33
Gatzlaff D, Haurin D. 1997. Sample selection bias and bepeat sale index estimates. J. Real Estate
Finan. Econ. 14:33-50.
Gatzlaff D, Haurin D. 1998. Sample selection and biases in local house value indices. J. Urban
Econ. 43:199-222.
Gatzlaff D, Holmes C. 2013. Estimating Transaction-Based Price Indices of Local Commercial
Real Estate Markets Using Public Assessment Data. J. Real Estate Finan. Econ. 46:260-81
Geltner D. 1989. Estimating real estate’s systematic risk from aggregate level appraisal-based
returns.” Real Estate Econ. 17:463–81
Geltner D. 2013. Commercial Real Estate and the 1990-91 Recession in the United State. Real
Estate Driven Systemic Risk: Country Cases & Their Policy Implications, Korea Development
Inst.: Seoul, Korea
Geltner D, Goetzmann W. 2000. Two decades of commercial property returns: A repeated-
measures regression-based version of the NCREIF Index.” J. Real Estate Finan. Econ., 21:5–21
Geltner D, MacGregor B, Schwann G. 2003. Appraisal smoothing and price discovery in real
estate markets. Urban Studies. 40:1047-64
Geltner D, Mei JP. 1995. The present value model with time varying discount rates: Implications
for commercial property valuation and investment decisions. J. Real Estate Finan. Econ.
11:119–35
Geltner D, Miller N. 2001. Commercial Real Estate Analysis and Investments. Mason, OH:
Prentice Hall
34
Geltner D, Miller N, Clayton J, Eichholtz P. 2014. Commercial Real Estate Analysis and
Investments. Mason OH: OnCourse Learning. 3rd ed.
Geltner D, Young M, Graff R. 1994. Random disaggregate appraisal error in commercial
property: Evidence from the Russell-NCREIF database. J. Real Estate Research, 9:403-19
Giliberto SM. 1992. The allocation of real Estate to future mixed-asset institutional portfolios. J.
Real Estate Research 7:423–32
Giliberto SM, Mengden A. 1996. REITs and real estate: Two markets reexamined. Real Estate
Finan. 13:56–60
Goetzmann W. 1992. The accuracy of real estate indices: Repeat sale estimators. J. Real Estate
Finan. Econ. 5:5–54
Goetzmann W, Peng L. 2006. Estimating house price indexes in the presence of seller
reservation prices. Rev. Econ. Stat. 88:100-12
Griliches Z, Adelman I. 1961. On an index of quality change. J. Am. Statist. Assoc. 56:535–48.
Guo X, Zheng S, Geltner D, Liu H. 2014. A new approach for constructing home price indices:
The pseudo repeat sales model and its application in China. J. Housing Econ. 25:20-38
Gyourko J, Keim D. 1992. What does the stock market tell us about real estate returns? Real
Estate Econ. 20:457–86
Halvorsen R, Pollakowski H. 1981. Choice of functional form for hedonic price equations. J.
Housing Econ. 10:37-49
Hansz JA, Diaz III J. 2001. Valuation bias in commercial appraisal: A transaction price feedback
experiment. Real Estate Econ.29: 553-65
35
Hartzell D, Heckman J, Miles M. 1986. Diversification categories in investment real estate. Real
Estate Econ. 14:230-54
Hill R. 2013. Hedonic price indexes for residential housing: A survey, evaluation and taxonomy.
J. Econ. Surveys 27:879-914
Hill R, Knight JR, Sirmans CF. 1997. Estimating capital asset price indexes. Review Econ.
Statist. 79:226-233
Hodrick R, Prescott E. 1997. Postwar U.S. business cycles: an empirical investigation. J. Money
Credit Banking. 29:1–16
Ibbotson R, Siegel L. 1984. Real estate returns: A comparison with other investments. Real
Estate Econ. 12:219-42
Kain JF, Quigley JM. 1970. Measuring the value of housing quality. J. Am. Statist. Assoc.
65:532–48.
Knight JR, Dombrow J, Sirmans CF. 1995.A varying parameters approach to constructing house
price indexes. Real Estate Econ. 23:187-205
LeSage J, Pace RK. 2010. Introduction to Spatial Econometrics. Boca Raton, FL: Chapman &
Hall/CRC
Liu C, Grissom T, Hartzell D. 1990. The impact of market imperfections on real estate returns
and optimal investment portfolios. Real Estate Econ. 18:453–78
Liu C, Hartzell D, Grissom T, Grieg W. 1990. The composition of the market portfolio and real
estate investment performance. Real Estate Econ. 18:49–75
36
Ling D, Naranjo A.1997. Economic risk factors and commercial real estate returns. J. Real
Estate Finan. Econ. 14:283–307
Lusht KM. 1988. The real estate pricing puzzle. Real Estate Econ. 16:95-104
Lusht KM. 1997. Real Estate Valuation: Principles and Applications. Chicago, IL: Irwin
McMillen DP. 2012. Repeat sales as a matching estimator. Real Estate Econ. 40:745–772
McMillen D, Dombrow J. 2001. A flexible fourier approach to repeat sales price indexes. Real
Estate Econ. 29:207–225
Meese RA, Wallace NE. 1997. The construction of residential housing price indices: A
comparison of repeat-sales, hedonic-regression, and hybrid approaches. J. Real Estate Finan.
Econ. 14:51-73
Miles M, Tolleson N. 1997. A revised look at how real estate compares with other major
components of the domestic investment universe. Real Estate Finan. 14:11-21
Munneke H, Slade B. 2000. An empirical study of sample selection bias in indices of
commercial real estate. J. Real Estate Finan. Econ. 21:45-64.
Munneke H, Slade B. 2001. A metropolitan transaction-based commercial price index: A time-
varying parameter approach. Real Estate Econ. 29:55-84.
Nishimura K, Shimizu C. 2006. Biases in appraisal land price information: The case of Japan. J.
Property Investment Finan. 24:150-75
Nishimura K, Shimizu C. 2007. Pricing structure in Tokyo metropolitan land markets and its
structural changes: Pre-bubble, bubble, and post-bubble periods. J. Real Estate Finan. Econ.
35:475-96
37
Pace RK, Barry R, Clapp J, Rodriquez M. 1998. Spatiotemporal autoregressive models of
neighborhood effects. J. Real Estate Finan. Econ. 17:15–33
Plazzi A, Torous W, Valkanov R. 2010. Expected returns and expected growth in rents of
commercial real estate. Rev. Financ. Studies 23:3469-3519
Pollakowski H. 1982. Urban Housing Markets and Residential Location. Lexington, MA:
Lexington Books
Quan D, Quigley J. 1989. Inferring an investment return series for real estate from observations
on sales. Real Estate Econ. 17:218–30
Quan D, Quigley J. 1991. Price formation and the appraisal function in real estate markets. J.
Real Estate Finan. Econ. 4: 127–46
Quigley JM. 1995. A simple hybrid model for estimating real estate price indexes. J. Housing
Econ. 4:1–12.
Rosen S. 1974. Hedonic prices and implicit markets: product differentiation in pure competition.
J. Political Econ. 82:34–55.
Ross S, Zisler R. 1991. Risk and return in real estate. J. Real Estate Finan.Econ. 4:175–90
Shiller R. 1981. Do stock prices move too much to be justified by subsequent changes in
dividends?,” Am. Econ. Rev. 71:421–36
Shiller R. 1991. Arithmetic repeat sales price estimators. J. Housing Econ. 1:110–26
Shiller R. 1993. Macro Markets: Creating Institutions for Managing Society’s Largest Economic
Risk. Oxford, UK: Oxford University Press
38
Shiller R. 2014. Online Data Robert Shiller. New Haven, CT:
http://www.econ.yale.edu/~shiller/data.htm
Shimizu C, Nishimura K, Watanabe T. 2010. Housing prices in Tokyo: A comparison of hedonic
and repeat sales measures. Jahrbücher f. Nationalökonomie u. Statistik. 230:792-813
SIFMA. 2014. US Bond Market Outstanding. Securities Industry and Financial Markets
Association. New York.
Thibodeau TG. 1997. Special issue on house price indices – Introduction. J. Real Estate Finan.
Econ. 14:5-9
Wheaton W. 1990. Vacancy, search, and prices in a housing market matching model. J. Political
Econ. 98:1270-92
Yavas A, Yildirim Y. 2011. Price discovery in real estate markets: A dynamic analysis. J. Real
Estate Finan. Econ. 42:1-29
39
Figures:
Figure 1: Property Asset Market Functioning
Num
ber o
f age
nts
Value ($/SF)
Buyer & Seller Reservation Price Frequency Distributions (as of a single point in time)
Sellers Buyers
Sellers(Supply)
Buyers(Demand)
P
Num
ber o
f age
nts
will
ing
to tr
ansa
ct
Value ($/SF)
Buyer & Seller Reservation Price Cumulative Distributions:Demand & Supply Functions
Sellers Buyers
Buyers Sellers
P*
Q*
40
Figure 2: Four Definitions & Measures of U.S. Investment Property Prices, 2000-2014:1
a Sources: FTSE, NCREIF. Figure updated and modified from Geltner et al. 2014, Chapter 25.
900
950
1000
1050
1100
1150
1200
1250
1300
1350
1400
1450
1500
1550
1600
1650
1700
1750
1800
1850
1900
1950
2000
2050
2100
12/3
1/19
99
6/28
/200
0
12/2
5/20
00
6/23
/200
1
12/2
0/20
01
6/18
/200
2
12/1
5/20
02
6/13
/200
3
12/1
0/20
03
6/7/
2004
12/4
/200
4
6/2/
2005
11/2
9/20
05
5/28
/200
6
11/2
4/20
06
5/23
/200
7
11/1
9/20
07
5/17
/200
8
11/1
3/20
08
5/12
/200
9
11/8
/200
9
5/7/
2010
11/3
/201
0
5/2/
2011
10/2
9/20
11
4/26
/201
2
10/2
3/20
12
4/21
/201
3
10/1
8/20
13
4/16
/201
4
1999
Val
ue =
100
0*
PureProperty daily, other indices quarterly
(1) REIT-based: FTSE-NAREIT PureProperty (daily)(2) Constant-Liquidity: NCREIF NTBI Demand(3) Trasaction Prices: NCREIF NTBI(5) Stale-appraisal-based: NCREIF NPI
(*NTBI Demand index starting value set to equalize average historical price and demand indices value levels.)
2001-14 Qtrly PP Dem NTBI NPIGeometric Mean: 0.79% 1.13% 1.04% 1.00%Volatility: 5.50% 6.83% 5.55% 2.83%1st-order Autocorrelation: 8.52% -10.98% -13.14% 83.16%Boom: 59% 81% 85% 77%Bust: -38% -42% -34% -28%Recovery: 60% 70% 38% 34%
41
Figure 3: CPI & Nominal CRE & Home Prices Since 19702
2 Source: Geltner 2013.
CRE Same-Property Transaction Prices, Shiller Nominal Home Price Index, CPI1969=100 shaded bars = GDP recessions:
0
100
200
300
400
500
600
700
800
900
1000
1969
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
2011
2013
Consumer Price Index CRE Same-Property Prices Nominal Home Prices
42
Figure 4: Stock Market Based & Transactions Based CRE Price Indices for Germany3
3 Source: Elonen 2013, vdpResearch 2014.
95
97
99
101
103
105
107
109
111
113
Q4-09 Q1-10 Q2-10 Q3-10 Q4-10 Q1-11 Q2-11 Q3-11 Q4-11 Q1-12 Q2-12 Q3-12 Q4-12
Germany Commercial Property Capital Value Indices:4Q1999 = 100.
REIT-based (PureProperty) Transactions-Based Hedonic (vdpResearch)
43