What sells in a crisis? Determinants of sale probability over a cycle and through a crash
| Pages | 619-637 |
| DOI | https://doi.org/10.1108/JPIF-02-2017-0013 |
| Date | 04 September 2017 |
| Published date | 04 September 2017 |
| Author | David Scofield,Steven Devaney |
| Subject Matter | Property management & built environment,Real estate & property,Property valuation & finance |
What sells in a crisis?
Determinants of sale probability
over a cycle and through a crash
David Scofield
Department of Real Estate, Ted Rogers School of Management,
Ryerson University, Toronto, Canada, and
Steven Devaney
Cass Business School, University of London, London, UK
Abstract
Purpose –The purpose of this paper is to understand what affects the liquidity of individual commercial
real estate assets over the course of the economic cycle by exploring a range of variables and a number of time
periods to identify key determinants of sale probability.
Design/methodology/approach –Analyzing 12,000 UK commercial real estate transactions (2003 to 2013)
the authors use an innovative sampling technique akin to a perpetual inventory approach to generate
a sample of held assets for each 12 month interval. Next, the authors use probit models to test how market,
owner and property factors affect sale probability in different market environments.
Findings –The types of properties that are most likely to sell changes between strong and weak markets.
Office and retail assets were more likely to sell than industrial both overall and in better market conditions,
but were less likely to sell than industrial properties during the downturn from mid-2007 to mid-2009.
Assets located in the City of London more likely to sell in both strong and weak markets. The behavior of
different groups of owners changed over time, and this indicates that the type of owner might have
implications for the liquidity of individual assets over and above their physical and locational attributes.
Practical implications –Variation in sale probability over time and across assets has implications for real
estate investment management both in terms of asset selection and the ability to rebalance portfolios over
the course of the cycle. Results also suggest that sample selection may be an issue for commercial real estate
price indices around the globe and imply that indices based on a limited group of owners/sellers might be
susceptible to further biases when tracking market performance through time.
Originality/value –The study differs from the existing literature on sale probability as the authors
analyzed samples of transactions drawn from all investor types, a significant advantage over studies based
on data restricted to samples of domestic institutional investors. As well, information on country of origin for
buyers and sellers allows us to explore the influence of foreign ownership on the probability of sale. Finally,
the authors not only analyze all transactions together, but the authors also look at transactions in five
distinct periods that correspond with different phases of the UK commercial real estate cycle. This paper
considers the UK real estate market, but it is likely that many of the findings hold for other major commercial
real estate markets.
Keywords Turnover, Liquidity, Ownership, Foreign investment, Market cycles, Sale probability
Paper type Research paper
1. Introduction
Real estate assets are lumpy and heterogeneous, and their returns vary based on their
physical and spatial characteristics. Investors in commercial real estate exchange rights to
properties in private, decentralized markets and through a process that is often lengthy and
involves significant transaction costs. These factors reduce liquidity in real estate as
compared with many other investment assets. In this paper, we try to understand what
affects the liquidity of individual commercial real estate assets over the course of the
economic cycle, focusing on the probability of sale as a proxy for liquidity. Identifying and
analyzing the variables that affect likelihood of sale during different market periods can Journal of Property Investment &
Finance
Vol. 35 No. 6, 2017
pp. 619-637
© Emerald PublishingLimited
1463-578X
DOI 10.1108/JPIF-02-2017-0013
Received 12 February 2017
Revised 28 April 2017
Accepted 9 May 2017
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/1463-578X.htm
The authors thank Real Capital Analytics/Property Data for access to data used in this research.
Any errors or omissions are the responsibility of the authors alone.
619
Sale
probability
inform investment strategy and is important for understanding real estate market
conditions. Our study explores a range of variables and a number of time periods in order to
identify key determinants of sale probability and how these change over time. The results
provide insights into investment behavior over a commercial real estate cycle.
In previous work, Fisher et al. (2004) examined the probability of a commercial real estate
sale as a function of market, property and owner characteristics, finding that each group of
factors displayed roughly equal significance to sale probability. A similar array of factors is
studied here. However, while their US data set reflected the activity of domestic institutional
investors,our UK data set includes both institutionaland non-institutionalreal estate investors,
as well as domestic and foreign investors. We also extend previous analysis by splitting our
time periodinto sub-periodsthat correspondwith different phasesof the commercialreal estate
cycle. We explore the effects of factors such as real estate price movements, economic growth
and changes in the flow and cost of funds, in addition to property related attributes such as
sector, size and location. Meanwhile, the presence of ownership variables in our data set
enables us to studythe influence of (equity)ownership type and nationality on the probability
of sale. These additions represent significant advances on previous research.
We find that the types of properties that are most likely to sell change between strong
and weak markets. For example, office and retail assets were more likely to sell than
industrial ones both overall and in better market conditions, but they were less likely to sell
than industrial properties during the downturn from mid-2007 to mid-2009. However, other
factors were more enduring in increasing the probability of sale, with assets located in the
City of London more likely to sell in both strong and weak markets. Meanwhile, the behavior
of different groups of owners changed over time. Nonetheless, private investors, REITs and
REOCs were more active than institutions, while European owners were less active than
domestic owners. This indicates that the type of owner might have implications for the
liquidity of individual assets over and above their physical and locational attributes.
This study marks the first time that research has sought to determine likelihood of sale
across different property and investor types (including foreign and domestic) at different
points in the market cycle. This matters as it is during periods of market instability that
liquidity is arguably most important.Understanding which assets are most likely to trade at
such times is informative to investorsin general, but of particular interestto types of investors
whose need to maintain or access capital in such markets is greatest. For instance,
Forbes (2017) suggests that selection of properties by UK open-ended funds is influenced by
their perceived saleability in the event of high redemption requests. Investor sensitivity to
liquidity shocks will affect the types of investors and types of stock that sell in different
phases of the realestate cycle. This has knock oneffects for the availability andinterpretation
of market evidence at different times in tasks like appraisal and market analysis.
To determine what sells and who sells investment grade commercial real estate before
during and following a period of acute financial crisis, we examine data obtained from Real
Capital Analytics/Property Data (RCA/PD) on over 12,000 transactions in the UK
commercial real estate market between 2001 and 2013. We model probability of sale in this
market for the period 2003 H2 to 2013 H1, examining the whole period and five sub-periods
which capture distinct market states. The rest of the paper proceeds as follows. The next
section discusses previous research on sale probability and related research on holding
periods. The following section then details the data that are used and the modeling
approach. After this, our empirical results are presented. The paper then concludes with a
discussion of the importance and ramifications of the findings.
2. Background and review
Transaction volume, turnover and the probability that a particular asset will trade are
interrelated measures of transaction activity that are connected to the concept of liquidity.
620
JPIF
35,6
To continue reading
Request your trial