Description
This paper draws from a wider research programme in the UK undertaken for the Investment Property Forum examining liquidity in commercial property. One aspect of liquidity is the process by which transactions occur including both how properties are selected for sale and the time taken to transact.
LIQUIDITY IN COMMERCIAL PROPERTY MARKETS
DECONSTRUCTING THE TRANSACTION PROCESS
Neil Crosby and Patrick McAllister
The University of Reading Business School
PO Box 219,
Whiteknights
Reading
RG6 6AW
Tel No : 0118 931 6657
Fax No : 0118 931 8172
Email : [email protected]
Abstract
This paper draws from a wider research programme in the UK undertaken for the Investment
Property Forum examining liquidity in commercial property. One aspect of liquidity is the
process by which transactions occur including both how properties are selected for sale and
the time taken to transact. The paper analyses data from three organisations; a property
company, a major financial institution and an asset management company, formally a major
public sector pension fund. The data covers three market states and includes sales completed
in 1995, 2000 and 2002 in the UK. The research interviewed key individuals within the three
organisations to identify any common patterns of activity within the sale process and also
identified the timing of 187 actual transactions from inception of the sale to completion.
The research developed a taxonomy of the transaction process. Interviews with vendors
indicated that decisions to sell were a product of a combination of portfolio, specific property
and market based issues. Properties were generally not kept in a “readiness for sale” state.
The average time from first decision to sell the actual property to completion had a mean time
of 298 days and a median of 190 days. It is concluded that this study may underestimate the
true length of the time to transact for two reasons. Firstly, the pre-marketing period is rarely
recorded in transaction files. Secondly, and more fundamentally, studies of sold properties
may contain selection bias. The research indicated that vendors tended to sell properties
which it was perceived could be sold at a ‘fair’ price in a reasonable period of time.
2
1. Introduction
Despite the fact that nearly all market participants would cite low liquidity as a (problematic)
characteristic of commercial property as an asset class, its complex nature and precise
consequences are rarely analysed. Lizieri and Bond (2004) review the definitions of liquidity
provided within both finance and real estate markets and conclude that liquidity is more than
simply sales rates or turnover of transactions, there are cost and price dimensions. However,
they also suggest that amongst other issues, information on the time taken to sell real estate
helps in the development of understanding of liquidity of property. Therefore, as part of the
wider IPF Liquidity of Commercial Property Markets scoping project, it was necessary to
begin to identify the different elements of the sale process and the timing of each element.
The ability to enter and exit property markets at specific times is constrained by the time
transactions take, any difficulties in identifying and bringing specific properties to the market
and uncertain prices, including changes to prices over the transaction period. Time to transact
has important implications for risk and return. Delay in realisation of capital value will
reduce total return. Uncertainty about timing of receipt of capital value adds to the volatility
of expected returns
1
with long delays being associated with increased uncertainty. Issues
include differences in transaction times between property as an asset and competing asset
classes and between different types of property, differentiated by, for example, type, size,
number of tenancies, etc. Other questions include the factors that determine transaction time
and whether any changes in those factors can be observed through time.
The overall aim of this paper is to carry out a preliminary examination of the property
transaction process to begin to answer some of these questions for the UK commercial
property market. In order to achieve this, three case studies were undertaken during October
and November 2003. They provide benchmark information on practice in terms of both
process and time to transact. Before setting out the details and results of the case studies, a
review of literature related to these two aspects is set out below and related to the interviews
carried out in the fieldwork.
1
That is, the a priori risk of the asset – see Lizieri and Bond (2004).
3
2. Liquidity and the Transaction Process
Bond and Lizieri (2004) present a comprehensive discussion of liquidity drawing upon a
range of literature from real estate and financial economics. We draw upon this work in this
section. They identify a number of dimensions of liquidity including:
• the rate of turnover/transactions and the time taken to transact;
• the costs associated with transacting (both formal costs – buy or sell fees – and
information costs);
• the impact of the decision to transact on the price of the asset and the prices of similar
assets; and
• uncertainty as to achieved price or return at the time of the decision to transact.
Many of the ‘standard’ approaches to liquidity focus on the importance of time to transact and
the consequences for certainty about price.
A well-known risk in property investment is that asset managers may be unable to rebalance
portfolios, may be unable to acquire the type of property required, or, due to lack of potential
buyers, may be unable to obtain a “fair” price for an owned asset. Consequently Key et al.
(1998) suggest that, for the property asset manager, liquidity is:
(a) being able to buy/sell when I want;
(b) being able to buy/sell what I want;
(c) being able to sell at the price that I want.
McNamara (1998) offered a somewhat different perspective. His starting position is that
liquidity is ‘the ability of an investor to trade assets into a cash form or vice versa. It is used
more loosely to describe the speed and/or volume of transacting in a given market’.
In property market text books, a similar set of definitions emerge focussing on ease of sale.
Baum and Crosby (1995) define liquidity as ‘the ease and certainty with which an asset can
be converted to cash at, or close to, its market value’. However, their definitions and
explanations are all in terms of time to sale and the barriers faced by potential purchasers.
Ball et al. (1998) note that ‘difficulties in trading property add a timing risk to uncertainties
surrounding the cash-flow and cause problems in implementing an active portfolio
management strategy.’ The length of time taken to transact is an associated disadvantage’.
4
Hoesli & MacGregor (2000) identify two consequences: ‘low liquidity creates two problems:
first, it takes longer to realise an asset’s market value and, secondly, there is a risk that the
market price will change between the decision to sell and a sale being implemented. Thus the
actual return may differ from the expected.’ Influenced by Lin and Vandell’s (2001) work in
this area, Bond and Lizieri (2004) sum up these strands succinctly
“A potential seller of real estate faces uncertainty as to the correct “price” for the
asset, uncertainty as to potential buyers and uncertainty as to the likely sale date.
These extra dimensions of uncertainty may not be fully reflected in ex-post
measures of property market performance. “
The length of the transaction is a central, if partial, factor influencing the level of risk. The
lengthier the sale period, the more likely that market conditions will change and the investor
is less certain about the cash flow.
3. Literature Review: Property Transactions: The Process and the Time to Sale
3.1 The Transaction Process
A number of studies have examined whether there are systematic differences between sold
and unsold properties. These studies have raised interesting questions concerning temporal
and cross-sectional variations in saleability. For example, are certain assets more saleable
than others at their market value; does saleability vary between time periods; does location or
type make assets more saleable and does uncertainty concerning price at the decision to sell
stage create a reluctance to sell
2
? Although there has been limited research on the
determinants of sale of individual properties in the UK, there are strong a priori expectations
drawn from this previous research and (albeit often anecdotal) market observation.
The first study related to this topic was carried out by Guilkey, et al (1989). They
investigated whether there were systematic differences between sold and unsold properties.
Using relatively small sample in the US, they test four hypotheses concerning the impact of
information asymmetries, liability matching, economies of scale associated with large lot
sizes and geographical remoteness. Supporting agency and information asymmetry effects,
they found that managers tended to sell assets that did not maximise manager compensation
and properties located in markets with strong current demand but rapid recent increases in
new supply that were not continuing. They also found that lease maturity, holding period,
5
tenant quality, capitalisation rate, income per square foot, age and a range of economic drivers
had significant explanatory power.
In related work, Collett et al (2003) focused on the holding periods of commercial property
assets in the UK. Using the IPD transaction data, they examined hypotheses concerning the
effect of size, returns and market conditions in acquisition and sale period. They found that
good market performance was associated with higher sale rates. Further, they identify a lot
size effect with small lot sizes having a higher propensity to sell than large lots.
In recent research, Fisher et al (2003) examine the determinants of transaction frequency and
the underlying factors that affect the probability of property sales occurring from period to
period. They draw an important distinction between liquidity and transaction frequency. This
is an interesting issue since properties may not transact because they are difficult to sell or
because the owner does not wish to sell. A decision not to sell may be associated either with
negative or with positive asset attributes. For instance, the low transaction frequency
identified by Collett et al (2003) for retail warehouses is almost certainly due to positive
attributes rather than negative factors and does not suggest that they are less liquid for
owners. Conversely, studies which find that small lots sizes are sold more commonly than
larger lots sizes do not indicate differential liquidity. Rather they may imply differences in
motivation to sell rather than ability to sell.
A priori, Fisher et al (2003) hypothesise that a range of owner specific (gearing, fund type,
historic performance, previous valuation) and property specific (holding period, voids, size
and age) variables together with market factors (cost and flow of funds, employment, capital
growth, and equity returns) affects sales activity. In line with Collett et al (2003), they point
to a strong positive correlation between capital growth and turnover. Overall, whilst bearing
in mind that sale probability and liquidity are separate, they find that their a priori
expectations are confirmed and that the factors identified provide significant explanatory
power of sale probability.
This research suggests that there are both systematic or market and specific factors that affect the
probability that an asset is selected for sale. This point will be developed later in this paper
2
That additional uncertainty is, as in the financial definitions in Lizieri and Bond (2004), a liquidity
cost. It is not the additional time taken to transact, but the uncertainty as to final achieved return that is
critical.
6
In 1995, the Investment Property Forum reported on the results of a working party
investigating the streamlining of the property transaction process. The objective was to make
property “more liquid” by the identification of areas of the transaction process that could be
improved, thereby quickening the sale process (IPF, 1995). This work did not identify any
specific time frames for selling property but did identify the process. It included some
element of preparation by the seller and also identified the period of marketing and
negotiation; including the agreement of heads of terms, negotiation of the documentation and
the undertaking of surveys and environmental investigations.
The working party concluded that the system in England and Wales was capable of being
flexible and a great deal could be done to decrease the time taken to transact and to reduce
difficulties in the system. They also concluded that advance preparation of materials
necessary to affect a sale and, in appropriate circumstances, the use of alternative methods of
due diligence and disposal could speed up transactions significantly.
Following this report, a supporting document was produced by the Investment Property
Forum setting out a Code of Practice to implement a streamlining of the transaction process
(IPF, 1996). It set out the information that a prospective seller should have available to show
prospective purchasers including management information (service charge accounts, rent
arrears, etc), documents and plans, replies to normal pre-contract enquiries, and an “informal”
inspection and survey.
3
It also suggests that an environmental audit should be undertaken
prior to offering for sale to identify possible problems which may abort a sale.
After heads of terms are agreed, IPF (1996) accepts that there will be a normal “ritual dance”
around these terms by legal advisors. It suggests that timetables for negotiation and contract
exchange are agreed at the same time as heads of terms to limit the open-ended nature of
these negotiations.
3.2 Time to transact
There appears to be little work in the UK which identifies how long transactions take.
McNamara (1998) identified the three periods as the time from initial decision to dispose to
the point at which draft heads of terms were agreed (marketing period), to exchange of
contracts (due diligence) and final transfer of monies (settlement). He carried out a survey of
around 30 property professionals and asked them for estimates of the average time taken to
7
transact typical property types measured across the three basic events identified above. The
property types and the time taken in weeks for the three events are set out in Table 1.
Table 1 : Time taken in weeks to transact
Marketing Period Due Diligence Settlement Total
Mean Stan Dev
Cathedral City retail unit 4.2 1.5 5 1 10.2
Large town retail unit 4.7 1.7 4 1 9.7
Small town retail unit 5.8 2.6 6 1 12.8
Major city shopping centre 8.6 4.5 12 1 21.6
Large town shopping centre 7.7 3.4 12 1 20.7
Small town shopping centre 7.6 3.6 12 1 20.6
Retail warehouse 5.0 2.7 4 1 10.0
Retail warehouse park 5.2 3.0 6 1 12.2
City office 7.4 3.1 8 1 16.4
West End office 6.2 2.2 8 1 15.2
Provincial city centre office 6.8 2.9 6 1 13.8
Business park 6.0 3.2 6 1 13.0
Standard industrial shed 5.4 2.5 6 1 12.4
Distribution warehouse 5.7 2.1 6 1 12.7
Source : McNamara (1998)
Generally, free-standing retail units reportedly took the least time to transact, with retail
warehouses at 10 weeks and standard shop units in large towns and cathedral cities also
around 10 weeks. Small town standard units took around 13 weeks. Standard industrial units
and distribution warehouses also took around 12/13 weeks and offices outside London took
around 13/14 weeks. City and West End offices were longer at over 15 weeks. Shopping
centres took the longest at around 20/21 weeks. Due diligence ranged from 4 weeks for the
retail warehouses and some standard shop units to 12 weeks for shopping centres. Marketing
periods ranged from 4/5 weeks for the shop and retail warehouse units to over 8 weeks for
shopping centres. Offices were around 7 weeks. The completion period was 1 week in all
cases although this seems very low and anecdotal comment would suggest 1 month to be
more likely.
However, this evidence is drawn from surveys of agents and “typical” periods and the
standard deviations suggest that there are some significant differences of opinion between
respondents. The case studies reported here will provide some real data on actual transactions
and how long they took to complete.
3
The IPF suggestion of an informal survey and inspection being made available to the prospective
8
4. The Case Studies
The case studies were based on the data from three different funds; one large financial
institution running a variety of general, long term and pension funds, one pension fund and
asset management company and one large property company. Between them, they administer
or manage a wide variety of different portfolios, including some monthly valued funds, and a
mixture of property only and mixed asset portfolios.
The research included two strands. First, an interview was carried out with a number of
representatives of each fund to discuss the processes involved in the different organisations
leading to decisions to sell. The actual sale process was then followed through in each
interview.
The second aspect of the research entailed the detailed investigation of actual transactions. In
order to address the issue of different market states, transactions in the calendar years 2000
and 2002 were collected. This normally entailed sales which were completed in the calendar
year but occasionally the data related to completion dates which went into 2001. In order to
gain all three states of rising, falling and stable markets, it was originally decided to attempt
to get 1995 in addition (as this was the last time all three property sector capital growth
indices fell). However, the files of properties so far back proved difficult to access, especially
in the tight time frame of this preliminary study, and data for 1995 (with occasionally
completions into 1996) was only available from one fund. Whilst providing information about
the assets sold, two of the organisations allowed access to the actual sale files in order to
extract relevant dates. This involved visits to their offices in order to read the files. The other
organisation provided pre-analysed data.
Data were obtained from over 187 properties across the three main commercial property
sectors. A proportion of the assets were sold by auction (approximately 10%). The majority
of the assets were sold by private treaty, often through a ‘best bids’ process. In a number of
cases, owners had been approached regarding individual buildings or portfolios and made
acceptable offers so that the “decision to sell” was made after the offer was received.
It should further be noted that one of the organisations had a policy of disposing of small,
non-core assets in this period and a substantial proportion of the sales involved this type of
property – most commonly ‘High Street’ shops in market towns. For 182 properties, the basic
property sector was identified within five segments; Office, Industrial, Standard Shop, Retail
purchaser raises issues of liability and whether such a thing as an “informal” survey is possible.
9
Warehouse and Shopping Centre. Some of these segments are very small; for example, only
five shopping centres and 12 retail warehouses. As the data is only from three companies, it
cannot be assumed to represent any sort of sample of the institutional and quoted property
company sector; the results are indicative only.
Discussion with the interviewees took place around a “model” transaction. Before analysing
the data on transaction times the transaction process, as identified by the interviewees, is set
out below.
5. The Transaction Process
5.1 A Model Transaction
McNamara (1998) breaks the sales process down into three parts; marketing, due diligence
and settlement. The available case study data does record the time taken from the decision to
market the property to completion which marries well with the McNamara survey. However,
a typical transaction as identified by the case study interviewees includes a pre-marketing
process. Therefore, transaction time commencing with the marketing of the particular asset
underestimates the total time for the sale process.
The interviewees from the three funds described a typical transaction as involving a number
of the key stages. These are illustrated diagrammatically in Appendix One.
The pre-marketing period where decisions to transact are made could be split into three stages
encompassing four decisions. The first is the general portfolio decision to sell property as an
asset - this strategic process is similar for all the competing assets. This triggers the sale
process. The first stage runs from this decision and the decision as to which sector or sub-
sector the particular asset to be sold will come. The second stage is the decision to sell a
particular asset within that sector. Finally, having decided to sell the property, the process of
getting the property ready to market takes time.
This third stage, between decision to sell and marketing, usually involves an instruction to
agents to prepare an assessment of value and marketability. Often, but not always, solicitors
are simultaneously instructed to identify any potential legal obstacles to sale. This can take
one to two weeks. It is possible that agents and solicitors may identify market factors (agents)
or asset specific factors (solicitors) that might need to be addressed before marketing.
10
Following receipt of marketing report from agents, formal marketing occurs
4
involving
production and distribution of a brochure, advertising etc. Best bids are then invited from
interested purchasers. Typically, this can take three to four weeks, according to the case
study interviewees.
The bids received are assessed and Heads of Terms agreed with the selected bidder. At this
point, solicitors are instructed to proceed towards exchange of contract and go through the
due diligence process. Due diligence can take another three to four weeks. However, it was
at this stage that transactions are most likely to be delayed, sometimes dramatically, due to
four main factors listed below.
Previously unknown or ignored inherent problems;
Changes in the asset e.g. tenant default;
Change in market conditions;
Changes in the circumstances of the purchaser, for example:
- Difficulty of funding. Increasing use of debt was said to sometimes result in
an additional due diligence process which could cause delay;
- Re-assessment of offer price.
Exchange of contracts takes place at the end of this period. This is the point at which at
which the sale becomes certain. For properties sold at auction, price agreement and exchange
of contract occur ‘when the hammers falls’.
Legal completion is the final act in the process. This is the date on which ownership rights
are transferred to the purchaser and cash is transferred to the vendor. Anecdotal evidence
suggests that simultaneous exchange of contract and completion has become more common.
However, the norm is for a gap of two to four weeks between exchange of contracts and
completion.
5.2 Variations to the model transaction
The interviewees identified a number of variations within all parts of the transaction process.
For fund managers, the initial selection process may be generated at a strategic asset
allocation level. This would then be followed by a tactical analysis of the sectors and regions
from which to sell property assets. At the individual asset level, assets would be ranked
4
In practice, agents may well have already marketed the asset with some of the contacts.
11
according to their estimated future performance. Performance analysis may be both backward
and forward-looking. The assessment would focus on issues such as bad debts, voids, the
outcome of rent reviews and achieved growth. The forward-looking analysis would
essentially involve an assessment of worth. Finally, assets would be selected that could be
sold in the time period to generate the funds required. This leads to an important finding.
Where funds need to generate cash in a specific time period, only properties which can be
reasonably expected to find a buyer in that specific time period could be selected.
There were a number of other ‘routes’ to sale
- Certain organisations may focus on specific regions, sectors or lot sizes. Non-
conforming assets were more likely to be sold.
- In some cases attractive unsolicited offers are received. Where acceptable, this
tends to speed up the disposal process dramatically since the marketing and
negotiation phases are bypassed.
- For open-ended funds such as unit trusts, there may be an urgent requirement to
liquidate assets to match unit redemptions. This increased pressure to sell could
force managers to consider selling any asset.
Property-specific factors which can delay disposal can be categorised into problems that are
either solvable or temporary but intractable. Solvable problems are issues which can be
addressed over a period of time but would render a property non-saleable, or unattractive to a
significant proportion of potential purchasers if marketed prior to problem resolution. The
consequence is that price achieved may be significantly below the perception of market value
with the problem resolved. Such issues include title problems, outstanding rent reviews,
disputes with tenants, tenant insolvency, non-compliance with fire regulations inter alia.
Theoretically, all inherent obstructions to sale can be resolved in advance of any decision to
sell.
However, this is not necessarily the case with temporary intractable factors. Although these
are often predictable and will disappear over time, crucially they tend to be outside the control
of the owner. Imminent rent reviews and potential lease terminations are the main problems.
The additional risk associated with unknown future income due to imminent rent reviews or
potential lease expiries can reduce the pool of potential buyers and hence the price obtained.
Whilst, these issues can be anticipated, they are not easy to resolve in advance.
12
In addition, there are a number of other ways in which the implementation of the decision to
sell may be delayed. The decision-making process may identify ways in which value can be
added to an asset at relatively low cost e.g. by redecoration or refurbishment. Where third
parties are involved e.g. in a head lease, or limited partnership, there may be delays associated
with permissions to assign or pre-emption rights. Associated delays can occur during the
selling process, as well as affecting the decision to sell due to largely unpredictable events.
For instance, tenants can become insolvent, seek to assign or be in breach of the lease
covenants.
It was also interesting to note that, contrary to expectation, very few of the interviewees
claimed experience of abortive transactions. This could be because of the filtering process by
which there is a tendency to only bring forward for sale assets which can be sold. Further,
transactions tend to acquire a momentum so that when a problem occurs with a sale, the
agents, vendors and other interested buyers have both financial and psychological reasons to
proceed.
We turn now to analysis of the transactions data obtained.
6. The Case Study Results
In order to examine the validity of this typical transaction, transaction data was collected and
analysed to validate the approximate timings of the typical transaction outlined above and in
McNamara (1998). The transactions were scrutinised for the following base data.
The date of decision to sell. In practice, this proved extremely difficult to identify.
Sale files often commenced with an instruction to agents. Rarely could we find any
evidence of the precise date when the organisation had decided to sell an asset.
The date of commencement of marketing. As noted above, in the ‘idealised’
transaction, the formal marketing would occur two to three weeks after instruction of
the agent to prepare an assessment of value and marketability.
The date of final price agreement. This was usually easily identified since Heads of
Terms could be found on the file. It is specifically termed final price agreement
since, in a number of transactions, price agreement could occur only for the
transaction to break down.
Exchange of contracts.
Completion.
13
The interviews with the representatives of the three funds suggested that there are distinct
periods in which the decision making process moves from the decision to sell property as an
asset class, to the decision to sell from a particular sectors and finally to the identification of
individual assets for disposal. These periods are extremely difficult to identify
chronologically and are rarely formally recorded. First sale records usually commence after
that process has been completed and the agent is about to be instructed. In the 154 instances
where the date of 1
st
record and the date agent appointed are both known, 100 occur at the
same time. We should also note that the database will almost certainly exclude some
properties that were withdrawn from sale and never brought back to the market
5
. As a result,
the data presented here will tend to understate the total length of the sales process.
The overall transaction time as set out in Table 2 is therefore the time from the first record of
the proposed sale, the date the sale file was started, which often coincide with the date the
agent was instructed. The average transaction time for the 184 transactions where this
information was recorded is 298 days, over 9 months. However, this average is skewed by a
small number of very long transactions: the median transaction time is 190 days, or just over
6 months.
The longest period is for negotiation. The average time is 178 days but again this is heavily
skewed and the median is 88 days, nearly 3 months. The due diligence process identified by
McNamara (1998) between sale agreed and contract averages 83 days and, although less
heavily skewed, the median is lower at 62 days or 2 months. The contract to completion
period averages 19 days or nearly 3 weeks.
Table 2 : Overall Transaction Times
Overall
Transaction
Time
Exchange to
completion
Price to
exchange
1st record to
price
Average 298 19 83 178
Median 190 19 62 88
Standard Dev 381 19 82 325
Skewness
6
4.07 1.43 2.25 5.39
Number 184 185 178 179
Figures 1 to 4 set out the distributions of the periods identified above. Figure 1 illustrates that
very few transactions take less than 50 days. The largest tranche of transactions (around 25%)
5
Technically, the data is “right censored”.
6
Normalised to zero: large positive numbers indicate positive, upside, skewness.
14
take between 50 and 100 days, with another 15% taking between 100 and 150 days. Well
over 60% take no more than 250 days or 8 months.
Figure 1 : Total Transaction Time
Tot al l engt h of t ransact i on
0
5
10
15
20
25
30
35
40
45
50
5
0
1
5
0
2
5
0
3
5
0
4
5
0
5
5
0
6
5
0
7
5
0
8
5
0
9
5
0
1
0
0
0
0
Days
F
r
e
q
u
e
n
c
y
Figure 2 illustrates that over 20% of transactions take between 10 and 50 days to market
while another 15% take less than 10 days. A further 15% take between 50 and 100 days.
Marketing in around 60% of cases takes three months or less. However, that still leaves
around 30% of cases taking between 100 and 300 days; over 3 months to nearly 10 months to
market.
Figure 2 : First Record to Price Agreement
Fi r s t r e c o r d t o f i n al p r i c e ag r e e m e n t
0
5
10
15
20
25
30
35
40
45
10 50 100 150 200 300 400 More
Day s
F
r
e
q
u
e
n
c
y
15
Figure 3 illustrates that in just less than 10% of the transactions monitored price agreement
and exchange appear to be simultaneous, due to some properties being sold at auction (and
also, we suspect, a few recording errors). The majority of transactions had a time from
agreement to contract of less than 100 days (nearly 60%) and a further 15 % took no more
than 150 days or just over 5 months. The due diligence periods are therefore both shorter (the
median being nearly a month less) and less variable than the marketing period.
Figure 3 : Price Agreement to Contract Exchange
Pr ice agreement to exchange
0
10
20
30
40
50
60
0 50 100 150 200 250 More
Days
F
r
e
q
u
e
n
c
y
Figure 4 illustrates that over 30% of transaction have simultaneous exchange and completion
but the largest group (around 25%) take between 26 and 30 days, or four weeks. Another 25%
approximately take less than four weeks leaving relatively few transactions taking more than
a month to complete.
Tables 3 to 5 set out the breakdown of the above figures for the three time periods of
completions in 1995/96, 2000/01 and 2002. The mean times suggest that transaction times
have increased rather than decreased despite the Forum’s attempts to streamline the process
(Investment Property Forum, 1995; 1996). In 1995, the average transaction time was 165
days: this rises to 272 days in 2000 and 339 days in 2002. However, 1995 is a very small
sample and the 2000 and 2002 results are influenced by skewness. The median times for 2000
are higher at 235 days than for 2002 at only 144 days, under 5 months, suggesting that the
“typical” time to sale is shorter.
16
Figure 4 : Contract Exchange to Completion
Exchange to completion
0
10
20
30
40
50
60
70
0
0
.
15
1
0
1
5
2
0
2
5
3
0
3
5
4
0
4
5
5
0
1
0
0
M
o
r
e
Days
F
r
e
q
u
e
n
c
y
One major difference between 2002 and the earlier transactions is that the completion period
has dropped from around four weeks in 1995 to three weeks in 2000 and to two weeks in
2002. Both medians and averages tell a similar story. The marketing period has a more
variable trend. The small number of transactions in 1995 suggest a short period of around one
and a half to two and a half months increasing significantly in 2000 to around five months. In
2002 the median falls back to less than two months, similar to 1995 but the average increases
significantly on the back of a few very long transactions. Price to exchange, the due diligence
period remains virtually identical in 2000 and 2002 suggesting no improvements in this part
of the transaction.
Table 3: 1995/96 Transaction Times
Overall
Transaction Time
Exchange to
completion Price to exchange 1st record to price
Average 165 24 53 75
Median 76 28 0 44
SD 141 14 110 65
Skewness 1.13 -0.15 1.92 2.25
Number 16 17 15 15
17
Table 4: 2000/01 Transaction Times
Overall
Transaction Time
Exchange to
completion Price to exchange 1st record to price
Average 272 21 86 166
Median 235 22 62 151
SD 143 19 83 122
Skewness 0.87 1.01 3.39 1.11
Number 70 69 69 70
Table 5: 2002 Transaction Times
Overall
Transaction Time
Exchange to
completion Price to exchange 1st record to price
Average 339 16 85 203
Median 144 14 66 51
SD 500 20 77 434
Skewness 3.17 1.90 1.64 4.20
Number 98 99 94 94
The sale price and all details of the different parts of the transaction were available for around
half (93 of the 187) transactions. These transactions were analysed to test whether the higher
value transactions took longer and whether the time taken for specific parts of the transaction
process changed with higher value properties.
Table 6 sets out the correlation matrix of price and transaction times. It appears that the value
of the property has very little effect on how long it takes to sell. The relationship between
price and total transaction time is not significantly different from zero with a correlation
coefficient of 0.06. The highest positive relationship between price and transaction time is for
the marketing period but this correlation coefficient is only 0.18
7
. Of equal interest is the fact
that the various components of the process are not correlated: a long marketing period is not
followed by a long due diligence or completion period. This suggests that a long transaction
is not a simple function of value: is may be a function of a long marketing period or a long
due diligence period, but not rarely both together.
7
Weakly significant at the 10% level.
18
Table 6: Correlation Matrix of Price and the Different Parts of the Transaction Process
Time exchange
to
completion
Time sale
agreed to
exchange
1st record to
agreement
1st record to
completion Price
Time exchange to completion 1.000
Time sale agreed to exchange -0.104 1.0000
1
st
record to agreement 0.087 -0.022 1.000
1
st
record to completion 0.106 0.477 0.707 1.000
Price -0.035 -0.065 0.179 0.060 1.000
Analysis of the 182 transactions where the property segment was known was undertaken, but
given the time constraints and the small number of observations in many of the segments,
only total transaction time was identified. It does not show the trends picked up in the survey
of professionals by McNamara (1998); shopping centres and standard shops have the same
median similar to the shopping centre times identified by McNamara but double the time
identified by him for standard shops. and in this sample retail warehouses have the highest
mean and the highest median again over double the time identified by McNamara. The office
and industrial median times seem closer to those of McNamara’s survey.
Table 7 : Transaction Total Time by Property Segment
Industrial Office
Retail
Warehouse
Shopping
Centre
Standard
Retail Total
Mean 215 197 292 232 219 219
Median 133 119 231 202 203 172
Standard deviation 248 184 271 102 147 176
Skewness 3.16 2.66 1.5 2.02 0.88 2.1
Number of
transactions 20 35 12 5 110 182
Maximum 1140 921 920 411 693 1140
Minimum 52 25 36 156 31 25
But, in order to progress any disaggregation, the sample size needs to increase and the range
of ownerships also needs to increase.
19
7. Conclusions
It is clear that time to transact is an important dimension of liquidity risk for property
investors. Prior to decision to sell, investors are uncertain about the amount of cash to be
received and the period until receipt. In addition, it is also clear that the longer the period
until receipt of cash, the greater the uncertainty about the amount receivable. Measures of
performance which neglect this risk will tend to underestimate the volatility of the asset class.
This preliminary analysis of the transactions data indicates that a typical transaction has six
separate stages. The available data identifies the timing of last three of these stages for 187
transactions in 1995/96, 2000/01 and 2002. The six stages are bounded by seven separate
decisions: The portfolio decision to sell property starts the process; the stages which follow
are:
Stage 1 - Property portfolio decision to sell particular sector or sub-sector
Stage 2 - Decision to sell particular asset
Stage 3 - Pre-marketing period
Stage 4 - Marketing period
Stage 5 - Due diligence period
Stage 6 - Exchange to Completion
The interviews give some insight into the sale decision. They suggest that many assets are
sold for portfolio reasons (such as a decision to sell smaller properties or a particular sub-
sector). However, the specific stock selection decision often relates to a notion of readiness
for sale. They implied that a relatively small number of property specific problems might
inhibit a sale. Properties are sold because they can be: those which, for example, have
imminent rent reviews and lease expiries are not considered saleable at an acceptable price.
A large number of prospective problems with specific properties, which might inhibit sales,
are identified in Stages 2 and 3 of the decision making process - therefore the number of
aborted sales in the database appears low.
If sales were a sample of all properties in portfolios, it would be expected that they would
include more properties with attributes which inhibit sale and, therefore, potential time to sale
would extend well beyond those observed average transaction times for the actual sales.
Consequently, the study also suggests that transaction frequency or probability of sale
provides only a partial indicator of asset liquidity. Sale probability depends upon whether the
20
seller is motivated to sell and whether the seller is able to sell. Proxy liquidity measures
based on time to an actual sale are driven by the latter, which may be misleading.
The preliminary analysis of the 187 transactions for transaction time over the last three stages
of the process suggests that very few generalisations can be made concerning the causes of
longer and shorter transactions times. The only apparent trend is the continuing reduction in
the time from exchange to completion, which now appears to average just over two weeks.
However, of the three stages, this is the least variable and the least lengthy so it does not
significantly reduce the overall transaction time. Over the whole data the average transaction
time is 298 days, over 9 months. However, this average is skewed by a small number of very
long transactions and the median transaction time is 190 days, or just over 6 months. Around
25% of transactions take between 50 and 100 days and 60% get completed within 8 months.
These figures need to be placed in the context of the time to transact in securities markets
(even for small capitalisation stocks and those with low free floats).
Given the length of time between exchange and completion is around two to three weeks, the
vast majority of time to sale is in the marketing and due diligence periods. Marketing
(median 88 days) is slightly longer than due diligence (median 62 days). No clear downward
trend through time in either of these two periods is observable from the data, despite the
efforts of the Investment Property Forum in promoting the streamlining of property
transactions (although we should stress that the sample for the 1995/6 period was limited).
Perhaps more surprisingly, there appears to be no reason or relationship between the length of
these two periods. A long marketing period does not lead necessarily to a shorter or longer
due diligence period. This may be because for some complex properties they both take longer
to complete while, in others, some of the due diligence may be undertaken before the final
price is agreed. If a prior offer had been received and later withdrawn, this would possibly
increase the marketing period to final agreed price, but reduce the time to exchange. Reasons
for purchasers withdrawing can be very specific; such as a tenant defaulting or changes to
market conditions in the due diligence period.
Value of property has no apparent effect on length of transaction. Property sector
disaggregation does not validate (or refute) the estimates of McNamara’s (1998) respondents
regarding different transaction times for the different property types and locations, but the
largest sample, standard retails does appear to take far longer than those estimates. However,
the sample also appears to include a number of small, secondary units being cleared out of
portfolios.
21
Overall, the case study interviews provide some insights into the transaction process and the
transactions data gives some indications of the timings of the last three stages of the
transaction process; marketing, due diligence and completion. The key outstanding issue is
the factors that cause extended transaction times. Are these simply ‘liquidity shocks’ that can
occur randomly and are essentially unpredictable? Are certain categories of asset more prone
to such liquidity shocks? Would unsold assets have taken longer to sell? The data could be
examined further for sector differences but without a larger number of transactions the ability
to drill down and disaggregate remains poor. Extending the data collection to more funds and
companies could give extended insight in to selection bias and the drivers for transactions and
deeper analysis of the source files could reveal and categorise the property specific issues
which cause transactions to vary so much in time taken to completion.
22
Bibliography
Ball, M., Lizieri, C. & MacGregor, B. (1998) The Economics of Commercial Property
Markets London, Routledge
Baum, A. & Crosby, N. (1995) Property Investment Appraisal London, Routledge
Bond, S and Lizieri, C (2004) Defining liquidity in property, Working Paper 1, Investment
Property Forum
Collett, D., Lizieri, C. and Ward, C. (2003) Timing and Holding Periods of Institutional Real
Estate, Real Estate Economics, 31,2, 205-222
Fisher, J . Gatzlaff, D. Geltner, D. and Haurin, D. (2003) An analysis of the determinants of
transaction frequency of institutional commercial real estate investment property, Paper
delivered at Skye real estate conference, August 2003.
Guilkey, D. Miles, M. and Cole, R. (1989) The Motivation for Institutional Real Estate Sales
and Implications for Asset Class Returns, AREUEA Journal, 17, 1, 70-82
Hoesli, M. & MacGregor, B. (2002) Property Investment Harlow: Longman
IPF (1995) Report of the Working Party on Streamlining Commercial Property Transactions,
London: Investment Property Forum
IPF (1996) “Readiness for Sale” The Code of Practice for Streamlining Commercial
Property Transaction, London: Investment Property Forum
Key, T., Durkin, S. & McBurney, D. (1998) Liquidity, Trading and Performance. Paper
presented to 7
th
IPD Investment Strategies Conference, Brighton
Lin, Z. & Vandell, K. (2001) Illiquidity and Real Estate Risk, Mimeo, School of Business,
University of Wisconsin – Madison.
McNamara, P. (1998) Exploring liquidity: Recent Survey Findings. Paper to the 7
th
Investment Property Databank Conference, Brighton, November.
23
Strategic allocation process
generates sale requirements
Tactical asset allocation identifies
preferred sectors and regions.
Worth assessment identifies
individual assets preferred
for sale
Unsolicited approach
from (special)
purchaser.
Individual asset recommended for sale
Asset does not
conform to target
portfolio holdings
Internal assessment of sale feasibility
Abort
Delay
Prepare for
sale
Wait for improved
market conditions
Minor preparations for sale
Instruct agents to
assess
marketability and
to estimate sale
price.
Instruct solicitors
to assess legal
constraints to
sale and carry out
preliminary work
External assessment of sale feasibility
Abort
Delay
Prepare for
sale
Wait for improved
market conditions
Minor preparations for marketing and
sale
Comparison of ‘book value’ with Market
Value
Formal marketing
Receipt and assessment of offers/bids
Purchaser’s due diligence
Lender’s due diligence, negotiation
of funding and funding approval
procedures?
Re-negotiation of price?
Exchange of contracts
Completion.
Pressure to
generate cash e.g.
unit redemptions
Major sale constraints
identified
No major sale constraints identified
Major sale constraints identified
No major sale constraints identified
Go back to second/third bidders?
Appendix One :
Taxonomy of the
transaction sale
process
24
25
doc_241395487.pdf
This paper draws from a wider research programme in the UK undertaken for the Investment Property Forum examining liquidity in commercial property. One aspect of liquidity is the process by which transactions occur including both how properties are selected for sale and the time taken to transact.
LIQUIDITY IN COMMERCIAL PROPERTY MARKETS
DECONSTRUCTING THE TRANSACTION PROCESS
Neil Crosby and Patrick McAllister
The University of Reading Business School
PO Box 219,
Whiteknights
Reading
RG6 6AW
Tel No : 0118 931 6657
Fax No : 0118 931 8172
Email : [email protected]
Abstract
This paper draws from a wider research programme in the UK undertaken for the Investment
Property Forum examining liquidity in commercial property. One aspect of liquidity is the
process by which transactions occur including both how properties are selected for sale and
the time taken to transact. The paper analyses data from three organisations; a property
company, a major financial institution and an asset management company, formally a major
public sector pension fund. The data covers three market states and includes sales completed
in 1995, 2000 and 2002 in the UK. The research interviewed key individuals within the three
organisations to identify any common patterns of activity within the sale process and also
identified the timing of 187 actual transactions from inception of the sale to completion.
The research developed a taxonomy of the transaction process. Interviews with vendors
indicated that decisions to sell were a product of a combination of portfolio, specific property
and market based issues. Properties were generally not kept in a “readiness for sale” state.
The average time from first decision to sell the actual property to completion had a mean time
of 298 days and a median of 190 days. It is concluded that this study may underestimate the
true length of the time to transact for two reasons. Firstly, the pre-marketing period is rarely
recorded in transaction files. Secondly, and more fundamentally, studies of sold properties
may contain selection bias. The research indicated that vendors tended to sell properties
which it was perceived could be sold at a ‘fair’ price in a reasonable period of time.
2
1. Introduction
Despite the fact that nearly all market participants would cite low liquidity as a (problematic)
characteristic of commercial property as an asset class, its complex nature and precise
consequences are rarely analysed. Lizieri and Bond (2004) review the definitions of liquidity
provided within both finance and real estate markets and conclude that liquidity is more than
simply sales rates or turnover of transactions, there are cost and price dimensions. However,
they also suggest that amongst other issues, information on the time taken to sell real estate
helps in the development of understanding of liquidity of property. Therefore, as part of the
wider IPF Liquidity of Commercial Property Markets scoping project, it was necessary to
begin to identify the different elements of the sale process and the timing of each element.
The ability to enter and exit property markets at specific times is constrained by the time
transactions take, any difficulties in identifying and bringing specific properties to the market
and uncertain prices, including changes to prices over the transaction period. Time to transact
has important implications for risk and return. Delay in realisation of capital value will
reduce total return. Uncertainty about timing of receipt of capital value adds to the volatility
of expected returns
1
with long delays being associated with increased uncertainty. Issues
include differences in transaction times between property as an asset and competing asset
classes and between different types of property, differentiated by, for example, type, size,
number of tenancies, etc. Other questions include the factors that determine transaction time
and whether any changes in those factors can be observed through time.
The overall aim of this paper is to carry out a preliminary examination of the property
transaction process to begin to answer some of these questions for the UK commercial
property market. In order to achieve this, three case studies were undertaken during October
and November 2003. They provide benchmark information on practice in terms of both
process and time to transact. Before setting out the details and results of the case studies, a
review of literature related to these two aspects is set out below and related to the interviews
carried out in the fieldwork.
1
That is, the a priori risk of the asset – see Lizieri and Bond (2004).
3
2. Liquidity and the Transaction Process
Bond and Lizieri (2004) present a comprehensive discussion of liquidity drawing upon a
range of literature from real estate and financial economics. We draw upon this work in this
section. They identify a number of dimensions of liquidity including:
• the rate of turnover/transactions and the time taken to transact;
• the costs associated with transacting (both formal costs – buy or sell fees – and
information costs);
• the impact of the decision to transact on the price of the asset and the prices of similar
assets; and
• uncertainty as to achieved price or return at the time of the decision to transact.
Many of the ‘standard’ approaches to liquidity focus on the importance of time to transact and
the consequences for certainty about price.
A well-known risk in property investment is that asset managers may be unable to rebalance
portfolios, may be unable to acquire the type of property required, or, due to lack of potential
buyers, may be unable to obtain a “fair” price for an owned asset. Consequently Key et al.
(1998) suggest that, for the property asset manager, liquidity is:
(a) being able to buy/sell when I want;
(b) being able to buy/sell what I want;
(c) being able to sell at the price that I want.
McNamara (1998) offered a somewhat different perspective. His starting position is that
liquidity is ‘the ability of an investor to trade assets into a cash form or vice versa. It is used
more loosely to describe the speed and/or volume of transacting in a given market’.
In property market text books, a similar set of definitions emerge focussing on ease of sale.
Baum and Crosby (1995) define liquidity as ‘the ease and certainty with which an asset can
be converted to cash at, or close to, its market value’. However, their definitions and
explanations are all in terms of time to sale and the barriers faced by potential purchasers.
Ball et al. (1998) note that ‘difficulties in trading property add a timing risk to uncertainties
surrounding the cash-flow and cause problems in implementing an active portfolio
management strategy.’ The length of time taken to transact is an associated disadvantage’.
4
Hoesli & MacGregor (2000) identify two consequences: ‘low liquidity creates two problems:
first, it takes longer to realise an asset’s market value and, secondly, there is a risk that the
market price will change between the decision to sell and a sale being implemented. Thus the
actual return may differ from the expected.’ Influenced by Lin and Vandell’s (2001) work in
this area, Bond and Lizieri (2004) sum up these strands succinctly
“A potential seller of real estate faces uncertainty as to the correct “price” for the
asset, uncertainty as to potential buyers and uncertainty as to the likely sale date.
These extra dimensions of uncertainty may not be fully reflected in ex-post
measures of property market performance. “
The length of the transaction is a central, if partial, factor influencing the level of risk. The
lengthier the sale period, the more likely that market conditions will change and the investor
is less certain about the cash flow.
3. Literature Review: Property Transactions: The Process and the Time to Sale
3.1 The Transaction Process
A number of studies have examined whether there are systematic differences between sold
and unsold properties. These studies have raised interesting questions concerning temporal
and cross-sectional variations in saleability. For example, are certain assets more saleable
than others at their market value; does saleability vary between time periods; does location or
type make assets more saleable and does uncertainty concerning price at the decision to sell
stage create a reluctance to sell
2
? Although there has been limited research on the
determinants of sale of individual properties in the UK, there are strong a priori expectations
drawn from this previous research and (albeit often anecdotal) market observation.
The first study related to this topic was carried out by Guilkey, et al (1989). They
investigated whether there were systematic differences between sold and unsold properties.
Using relatively small sample in the US, they test four hypotheses concerning the impact of
information asymmetries, liability matching, economies of scale associated with large lot
sizes and geographical remoteness. Supporting agency and information asymmetry effects,
they found that managers tended to sell assets that did not maximise manager compensation
and properties located in markets with strong current demand but rapid recent increases in
new supply that were not continuing. They also found that lease maturity, holding period,
5
tenant quality, capitalisation rate, income per square foot, age and a range of economic drivers
had significant explanatory power.
In related work, Collett et al (2003) focused on the holding periods of commercial property
assets in the UK. Using the IPD transaction data, they examined hypotheses concerning the
effect of size, returns and market conditions in acquisition and sale period. They found that
good market performance was associated with higher sale rates. Further, they identify a lot
size effect with small lot sizes having a higher propensity to sell than large lots.
In recent research, Fisher et al (2003) examine the determinants of transaction frequency and
the underlying factors that affect the probability of property sales occurring from period to
period. They draw an important distinction between liquidity and transaction frequency. This
is an interesting issue since properties may not transact because they are difficult to sell or
because the owner does not wish to sell. A decision not to sell may be associated either with
negative or with positive asset attributes. For instance, the low transaction frequency
identified by Collett et al (2003) for retail warehouses is almost certainly due to positive
attributes rather than negative factors and does not suggest that they are less liquid for
owners. Conversely, studies which find that small lots sizes are sold more commonly than
larger lots sizes do not indicate differential liquidity. Rather they may imply differences in
motivation to sell rather than ability to sell.
A priori, Fisher et al (2003) hypothesise that a range of owner specific (gearing, fund type,
historic performance, previous valuation) and property specific (holding period, voids, size
and age) variables together with market factors (cost and flow of funds, employment, capital
growth, and equity returns) affects sales activity. In line with Collett et al (2003), they point
to a strong positive correlation between capital growth and turnover. Overall, whilst bearing
in mind that sale probability and liquidity are separate, they find that their a priori
expectations are confirmed and that the factors identified provide significant explanatory
power of sale probability.
This research suggests that there are both systematic or market and specific factors that affect the
probability that an asset is selected for sale. This point will be developed later in this paper
2
That additional uncertainty is, as in the financial definitions in Lizieri and Bond (2004), a liquidity
cost. It is not the additional time taken to transact, but the uncertainty as to final achieved return that is
critical.
6
In 1995, the Investment Property Forum reported on the results of a working party
investigating the streamlining of the property transaction process. The objective was to make
property “more liquid” by the identification of areas of the transaction process that could be
improved, thereby quickening the sale process (IPF, 1995). This work did not identify any
specific time frames for selling property but did identify the process. It included some
element of preparation by the seller and also identified the period of marketing and
negotiation; including the agreement of heads of terms, negotiation of the documentation and
the undertaking of surveys and environmental investigations.
The working party concluded that the system in England and Wales was capable of being
flexible and a great deal could be done to decrease the time taken to transact and to reduce
difficulties in the system. They also concluded that advance preparation of materials
necessary to affect a sale and, in appropriate circumstances, the use of alternative methods of
due diligence and disposal could speed up transactions significantly.
Following this report, a supporting document was produced by the Investment Property
Forum setting out a Code of Practice to implement a streamlining of the transaction process
(IPF, 1996). It set out the information that a prospective seller should have available to show
prospective purchasers including management information (service charge accounts, rent
arrears, etc), documents and plans, replies to normal pre-contract enquiries, and an “informal”
inspection and survey.
3
It also suggests that an environmental audit should be undertaken
prior to offering for sale to identify possible problems which may abort a sale.
After heads of terms are agreed, IPF (1996) accepts that there will be a normal “ritual dance”
around these terms by legal advisors. It suggests that timetables for negotiation and contract
exchange are agreed at the same time as heads of terms to limit the open-ended nature of
these negotiations.
3.2 Time to transact
There appears to be little work in the UK which identifies how long transactions take.
McNamara (1998) identified the three periods as the time from initial decision to dispose to
the point at which draft heads of terms were agreed (marketing period), to exchange of
contracts (due diligence) and final transfer of monies (settlement). He carried out a survey of
around 30 property professionals and asked them for estimates of the average time taken to
7
transact typical property types measured across the three basic events identified above. The
property types and the time taken in weeks for the three events are set out in Table 1.
Table 1 : Time taken in weeks to transact
Marketing Period Due Diligence Settlement Total
Mean Stan Dev
Cathedral City retail unit 4.2 1.5 5 1 10.2
Large town retail unit 4.7 1.7 4 1 9.7
Small town retail unit 5.8 2.6 6 1 12.8
Major city shopping centre 8.6 4.5 12 1 21.6
Large town shopping centre 7.7 3.4 12 1 20.7
Small town shopping centre 7.6 3.6 12 1 20.6
Retail warehouse 5.0 2.7 4 1 10.0
Retail warehouse park 5.2 3.0 6 1 12.2
City office 7.4 3.1 8 1 16.4
West End office 6.2 2.2 8 1 15.2
Provincial city centre office 6.8 2.9 6 1 13.8
Business park 6.0 3.2 6 1 13.0
Standard industrial shed 5.4 2.5 6 1 12.4
Distribution warehouse 5.7 2.1 6 1 12.7
Source : McNamara (1998)
Generally, free-standing retail units reportedly took the least time to transact, with retail
warehouses at 10 weeks and standard shop units in large towns and cathedral cities also
around 10 weeks. Small town standard units took around 13 weeks. Standard industrial units
and distribution warehouses also took around 12/13 weeks and offices outside London took
around 13/14 weeks. City and West End offices were longer at over 15 weeks. Shopping
centres took the longest at around 20/21 weeks. Due diligence ranged from 4 weeks for the
retail warehouses and some standard shop units to 12 weeks for shopping centres. Marketing
periods ranged from 4/5 weeks for the shop and retail warehouse units to over 8 weeks for
shopping centres. Offices were around 7 weeks. The completion period was 1 week in all
cases although this seems very low and anecdotal comment would suggest 1 month to be
more likely.
However, this evidence is drawn from surveys of agents and “typical” periods and the
standard deviations suggest that there are some significant differences of opinion between
respondents. The case studies reported here will provide some real data on actual transactions
and how long they took to complete.
3
The IPF suggestion of an informal survey and inspection being made available to the prospective
8
4. The Case Studies
The case studies were based on the data from three different funds; one large financial
institution running a variety of general, long term and pension funds, one pension fund and
asset management company and one large property company. Between them, they administer
or manage a wide variety of different portfolios, including some monthly valued funds, and a
mixture of property only and mixed asset portfolios.
The research included two strands. First, an interview was carried out with a number of
representatives of each fund to discuss the processes involved in the different organisations
leading to decisions to sell. The actual sale process was then followed through in each
interview.
The second aspect of the research entailed the detailed investigation of actual transactions. In
order to address the issue of different market states, transactions in the calendar years 2000
and 2002 were collected. This normally entailed sales which were completed in the calendar
year but occasionally the data related to completion dates which went into 2001. In order to
gain all three states of rising, falling and stable markets, it was originally decided to attempt
to get 1995 in addition (as this was the last time all three property sector capital growth
indices fell). However, the files of properties so far back proved difficult to access, especially
in the tight time frame of this preliminary study, and data for 1995 (with occasionally
completions into 1996) was only available from one fund. Whilst providing information about
the assets sold, two of the organisations allowed access to the actual sale files in order to
extract relevant dates. This involved visits to their offices in order to read the files. The other
organisation provided pre-analysed data.
Data were obtained from over 187 properties across the three main commercial property
sectors. A proportion of the assets were sold by auction (approximately 10%). The majority
of the assets were sold by private treaty, often through a ‘best bids’ process. In a number of
cases, owners had been approached regarding individual buildings or portfolios and made
acceptable offers so that the “decision to sell” was made after the offer was received.
It should further be noted that one of the organisations had a policy of disposing of small,
non-core assets in this period and a substantial proportion of the sales involved this type of
property – most commonly ‘High Street’ shops in market towns. For 182 properties, the basic
property sector was identified within five segments; Office, Industrial, Standard Shop, Retail
purchaser raises issues of liability and whether such a thing as an “informal” survey is possible.
9
Warehouse and Shopping Centre. Some of these segments are very small; for example, only
five shopping centres and 12 retail warehouses. As the data is only from three companies, it
cannot be assumed to represent any sort of sample of the institutional and quoted property
company sector; the results are indicative only.
Discussion with the interviewees took place around a “model” transaction. Before analysing
the data on transaction times the transaction process, as identified by the interviewees, is set
out below.
5. The Transaction Process
5.1 A Model Transaction
McNamara (1998) breaks the sales process down into three parts; marketing, due diligence
and settlement. The available case study data does record the time taken from the decision to
market the property to completion which marries well with the McNamara survey. However,
a typical transaction as identified by the case study interviewees includes a pre-marketing
process. Therefore, transaction time commencing with the marketing of the particular asset
underestimates the total time for the sale process.
The interviewees from the three funds described a typical transaction as involving a number
of the key stages. These are illustrated diagrammatically in Appendix One.
The pre-marketing period where decisions to transact are made could be split into three stages
encompassing four decisions. The first is the general portfolio decision to sell property as an
asset - this strategic process is similar for all the competing assets. This triggers the sale
process. The first stage runs from this decision and the decision as to which sector or sub-
sector the particular asset to be sold will come. The second stage is the decision to sell a
particular asset within that sector. Finally, having decided to sell the property, the process of
getting the property ready to market takes time.
This third stage, between decision to sell and marketing, usually involves an instruction to
agents to prepare an assessment of value and marketability. Often, but not always, solicitors
are simultaneously instructed to identify any potential legal obstacles to sale. This can take
one to two weeks. It is possible that agents and solicitors may identify market factors (agents)
or asset specific factors (solicitors) that might need to be addressed before marketing.
10
Following receipt of marketing report from agents, formal marketing occurs
4
involving
production and distribution of a brochure, advertising etc. Best bids are then invited from
interested purchasers. Typically, this can take three to four weeks, according to the case
study interviewees.
The bids received are assessed and Heads of Terms agreed with the selected bidder. At this
point, solicitors are instructed to proceed towards exchange of contract and go through the
due diligence process. Due diligence can take another three to four weeks. However, it was
at this stage that transactions are most likely to be delayed, sometimes dramatically, due to
four main factors listed below.
Previously unknown or ignored inherent problems;
Changes in the asset e.g. tenant default;
Change in market conditions;
Changes in the circumstances of the purchaser, for example:
- Difficulty of funding. Increasing use of debt was said to sometimes result in
an additional due diligence process which could cause delay;
- Re-assessment of offer price.
Exchange of contracts takes place at the end of this period. This is the point at which at
which the sale becomes certain. For properties sold at auction, price agreement and exchange
of contract occur ‘when the hammers falls’.
Legal completion is the final act in the process. This is the date on which ownership rights
are transferred to the purchaser and cash is transferred to the vendor. Anecdotal evidence
suggests that simultaneous exchange of contract and completion has become more common.
However, the norm is for a gap of two to four weeks between exchange of contracts and
completion.
5.2 Variations to the model transaction
The interviewees identified a number of variations within all parts of the transaction process.
For fund managers, the initial selection process may be generated at a strategic asset
allocation level. This would then be followed by a tactical analysis of the sectors and regions
from which to sell property assets. At the individual asset level, assets would be ranked
4
In practice, agents may well have already marketed the asset with some of the contacts.
11
according to their estimated future performance. Performance analysis may be both backward
and forward-looking. The assessment would focus on issues such as bad debts, voids, the
outcome of rent reviews and achieved growth. The forward-looking analysis would
essentially involve an assessment of worth. Finally, assets would be selected that could be
sold in the time period to generate the funds required. This leads to an important finding.
Where funds need to generate cash in a specific time period, only properties which can be
reasonably expected to find a buyer in that specific time period could be selected.
There were a number of other ‘routes’ to sale
- Certain organisations may focus on specific regions, sectors or lot sizes. Non-
conforming assets were more likely to be sold.
- In some cases attractive unsolicited offers are received. Where acceptable, this
tends to speed up the disposal process dramatically since the marketing and
negotiation phases are bypassed.
- For open-ended funds such as unit trusts, there may be an urgent requirement to
liquidate assets to match unit redemptions. This increased pressure to sell could
force managers to consider selling any asset.
Property-specific factors which can delay disposal can be categorised into problems that are
either solvable or temporary but intractable. Solvable problems are issues which can be
addressed over a period of time but would render a property non-saleable, or unattractive to a
significant proportion of potential purchasers if marketed prior to problem resolution. The
consequence is that price achieved may be significantly below the perception of market value
with the problem resolved. Such issues include title problems, outstanding rent reviews,
disputes with tenants, tenant insolvency, non-compliance with fire regulations inter alia.
Theoretically, all inherent obstructions to sale can be resolved in advance of any decision to
sell.
However, this is not necessarily the case with temporary intractable factors. Although these
are often predictable and will disappear over time, crucially they tend to be outside the control
of the owner. Imminent rent reviews and potential lease terminations are the main problems.
The additional risk associated with unknown future income due to imminent rent reviews or
potential lease expiries can reduce the pool of potential buyers and hence the price obtained.
Whilst, these issues can be anticipated, they are not easy to resolve in advance.
12
In addition, there are a number of other ways in which the implementation of the decision to
sell may be delayed. The decision-making process may identify ways in which value can be
added to an asset at relatively low cost e.g. by redecoration or refurbishment. Where third
parties are involved e.g. in a head lease, or limited partnership, there may be delays associated
with permissions to assign or pre-emption rights. Associated delays can occur during the
selling process, as well as affecting the decision to sell due to largely unpredictable events.
For instance, tenants can become insolvent, seek to assign or be in breach of the lease
covenants.
It was also interesting to note that, contrary to expectation, very few of the interviewees
claimed experience of abortive transactions. This could be because of the filtering process by
which there is a tendency to only bring forward for sale assets which can be sold. Further,
transactions tend to acquire a momentum so that when a problem occurs with a sale, the
agents, vendors and other interested buyers have both financial and psychological reasons to
proceed.
We turn now to analysis of the transactions data obtained.
6. The Case Study Results
In order to examine the validity of this typical transaction, transaction data was collected and
analysed to validate the approximate timings of the typical transaction outlined above and in
McNamara (1998). The transactions were scrutinised for the following base data.
The date of decision to sell. In practice, this proved extremely difficult to identify.
Sale files often commenced with an instruction to agents. Rarely could we find any
evidence of the precise date when the organisation had decided to sell an asset.
The date of commencement of marketing. As noted above, in the ‘idealised’
transaction, the formal marketing would occur two to three weeks after instruction of
the agent to prepare an assessment of value and marketability.
The date of final price agreement. This was usually easily identified since Heads of
Terms could be found on the file. It is specifically termed final price agreement
since, in a number of transactions, price agreement could occur only for the
transaction to break down.
Exchange of contracts.
Completion.
13
The interviews with the representatives of the three funds suggested that there are distinct
periods in which the decision making process moves from the decision to sell property as an
asset class, to the decision to sell from a particular sectors and finally to the identification of
individual assets for disposal. These periods are extremely difficult to identify
chronologically and are rarely formally recorded. First sale records usually commence after
that process has been completed and the agent is about to be instructed. In the 154 instances
where the date of 1
st
record and the date agent appointed are both known, 100 occur at the
same time. We should also note that the database will almost certainly exclude some
properties that were withdrawn from sale and never brought back to the market
5
. As a result,
the data presented here will tend to understate the total length of the sales process.
The overall transaction time as set out in Table 2 is therefore the time from the first record of
the proposed sale, the date the sale file was started, which often coincide with the date the
agent was instructed. The average transaction time for the 184 transactions where this
information was recorded is 298 days, over 9 months. However, this average is skewed by a
small number of very long transactions: the median transaction time is 190 days, or just over
6 months.
The longest period is for negotiation. The average time is 178 days but again this is heavily
skewed and the median is 88 days, nearly 3 months. The due diligence process identified by
McNamara (1998) between sale agreed and contract averages 83 days and, although less
heavily skewed, the median is lower at 62 days or 2 months. The contract to completion
period averages 19 days or nearly 3 weeks.
Table 2 : Overall Transaction Times
Overall
Transaction
Time
Exchange to
completion
Price to
exchange
1st record to
price
Average 298 19 83 178
Median 190 19 62 88
Standard Dev 381 19 82 325
Skewness
6
4.07 1.43 2.25 5.39
Number 184 185 178 179
Figures 1 to 4 set out the distributions of the periods identified above. Figure 1 illustrates that
very few transactions take less than 50 days. The largest tranche of transactions (around 25%)
5
Technically, the data is “right censored”.
6
Normalised to zero: large positive numbers indicate positive, upside, skewness.
14
take between 50 and 100 days, with another 15% taking between 100 and 150 days. Well
over 60% take no more than 250 days or 8 months.
Figure 1 : Total Transaction Time
Tot al l engt h of t ransact i on
0
5
10
15
20
25
30
35
40
45
50
5
0
1
5
0
2
5
0
3
5
0
4
5
0
5
5
0
6
5
0
7
5
0
8
5
0
9
5
0
1
0
0
0
0
Days
F
r
e
q
u
e
n
c
y
Figure 2 illustrates that over 20% of transactions take between 10 and 50 days to market
while another 15% take less than 10 days. A further 15% take between 50 and 100 days.
Marketing in around 60% of cases takes three months or less. However, that still leaves
around 30% of cases taking between 100 and 300 days; over 3 months to nearly 10 months to
market.
Figure 2 : First Record to Price Agreement
Fi r s t r e c o r d t o f i n al p r i c e ag r e e m e n t
0
5
10
15
20
25
30
35
40
45
10 50 100 150 200 300 400 More
Day s
F
r
e
q
u
e
n
c
y
15
Figure 3 illustrates that in just less than 10% of the transactions monitored price agreement
and exchange appear to be simultaneous, due to some properties being sold at auction (and
also, we suspect, a few recording errors). The majority of transactions had a time from
agreement to contract of less than 100 days (nearly 60%) and a further 15 % took no more
than 150 days or just over 5 months. The due diligence periods are therefore both shorter (the
median being nearly a month less) and less variable than the marketing period.
Figure 3 : Price Agreement to Contract Exchange
Pr ice agreement to exchange
0
10
20
30
40
50
60
0 50 100 150 200 250 More
Days
F
r
e
q
u
e
n
c
y
Figure 4 illustrates that over 30% of transaction have simultaneous exchange and completion
but the largest group (around 25%) take between 26 and 30 days, or four weeks. Another 25%
approximately take less than four weeks leaving relatively few transactions taking more than
a month to complete.
Tables 3 to 5 set out the breakdown of the above figures for the three time periods of
completions in 1995/96, 2000/01 and 2002. The mean times suggest that transaction times
have increased rather than decreased despite the Forum’s attempts to streamline the process
(Investment Property Forum, 1995; 1996). In 1995, the average transaction time was 165
days: this rises to 272 days in 2000 and 339 days in 2002. However, 1995 is a very small
sample and the 2000 and 2002 results are influenced by skewness. The median times for 2000
are higher at 235 days than for 2002 at only 144 days, under 5 months, suggesting that the
“typical” time to sale is shorter.
16
Figure 4 : Contract Exchange to Completion
Exchange to completion
0
10
20
30
40
50
60
70
0
0
.
15
1
0
1
5
2
0
2
5
3
0
3
5
4
0
4
5
5
0
1
0
0
M
o
r
e
Days
F
r
e
q
u
e
n
c
y
One major difference between 2002 and the earlier transactions is that the completion period
has dropped from around four weeks in 1995 to three weeks in 2000 and to two weeks in
2002. Both medians and averages tell a similar story. The marketing period has a more
variable trend. The small number of transactions in 1995 suggest a short period of around one
and a half to two and a half months increasing significantly in 2000 to around five months. In
2002 the median falls back to less than two months, similar to 1995 but the average increases
significantly on the back of a few very long transactions. Price to exchange, the due diligence
period remains virtually identical in 2000 and 2002 suggesting no improvements in this part
of the transaction.
Table 3: 1995/96 Transaction Times
Overall
Transaction Time
Exchange to
completion Price to exchange 1st record to price
Average 165 24 53 75
Median 76 28 0 44
SD 141 14 110 65
Skewness 1.13 -0.15 1.92 2.25
Number 16 17 15 15
17
Table 4: 2000/01 Transaction Times
Overall
Transaction Time
Exchange to
completion Price to exchange 1st record to price
Average 272 21 86 166
Median 235 22 62 151
SD 143 19 83 122
Skewness 0.87 1.01 3.39 1.11
Number 70 69 69 70
Table 5: 2002 Transaction Times
Overall
Transaction Time
Exchange to
completion Price to exchange 1st record to price
Average 339 16 85 203
Median 144 14 66 51
SD 500 20 77 434
Skewness 3.17 1.90 1.64 4.20
Number 98 99 94 94
The sale price and all details of the different parts of the transaction were available for around
half (93 of the 187) transactions. These transactions were analysed to test whether the higher
value transactions took longer and whether the time taken for specific parts of the transaction
process changed with higher value properties.
Table 6 sets out the correlation matrix of price and transaction times. It appears that the value
of the property has very little effect on how long it takes to sell. The relationship between
price and total transaction time is not significantly different from zero with a correlation
coefficient of 0.06. The highest positive relationship between price and transaction time is for
the marketing period but this correlation coefficient is only 0.18
7
. Of equal interest is the fact
that the various components of the process are not correlated: a long marketing period is not
followed by a long due diligence or completion period. This suggests that a long transaction
is not a simple function of value: is may be a function of a long marketing period or a long
due diligence period, but not rarely both together.
7
Weakly significant at the 10% level.
18
Table 6: Correlation Matrix of Price and the Different Parts of the Transaction Process
Time exchange
to
completion
Time sale
agreed to
exchange
1st record to
agreement
1st record to
completion Price
Time exchange to completion 1.000
Time sale agreed to exchange -0.104 1.0000
1
st
record to agreement 0.087 -0.022 1.000
1
st
record to completion 0.106 0.477 0.707 1.000
Price -0.035 -0.065 0.179 0.060 1.000
Analysis of the 182 transactions where the property segment was known was undertaken, but
given the time constraints and the small number of observations in many of the segments,
only total transaction time was identified. It does not show the trends picked up in the survey
of professionals by McNamara (1998); shopping centres and standard shops have the same
median similar to the shopping centre times identified by McNamara but double the time
identified by him for standard shops. and in this sample retail warehouses have the highest
mean and the highest median again over double the time identified by McNamara. The office
and industrial median times seem closer to those of McNamara’s survey.
Table 7 : Transaction Total Time by Property Segment
Industrial Office
Retail
Warehouse
Shopping
Centre
Standard
Retail Total
Mean 215 197 292 232 219 219
Median 133 119 231 202 203 172
Standard deviation 248 184 271 102 147 176
Skewness 3.16 2.66 1.5 2.02 0.88 2.1
Number of
transactions 20 35 12 5 110 182
Maximum 1140 921 920 411 693 1140
Minimum 52 25 36 156 31 25
But, in order to progress any disaggregation, the sample size needs to increase and the range
of ownerships also needs to increase.
19
7. Conclusions
It is clear that time to transact is an important dimension of liquidity risk for property
investors. Prior to decision to sell, investors are uncertain about the amount of cash to be
received and the period until receipt. In addition, it is also clear that the longer the period
until receipt of cash, the greater the uncertainty about the amount receivable. Measures of
performance which neglect this risk will tend to underestimate the volatility of the asset class.
This preliminary analysis of the transactions data indicates that a typical transaction has six
separate stages. The available data identifies the timing of last three of these stages for 187
transactions in 1995/96, 2000/01 and 2002. The six stages are bounded by seven separate
decisions: The portfolio decision to sell property starts the process; the stages which follow
are:
Stage 1 - Property portfolio decision to sell particular sector or sub-sector
Stage 2 - Decision to sell particular asset
Stage 3 - Pre-marketing period
Stage 4 - Marketing period
Stage 5 - Due diligence period
Stage 6 - Exchange to Completion
The interviews give some insight into the sale decision. They suggest that many assets are
sold for portfolio reasons (such as a decision to sell smaller properties or a particular sub-
sector). However, the specific stock selection decision often relates to a notion of readiness
for sale. They implied that a relatively small number of property specific problems might
inhibit a sale. Properties are sold because they can be: those which, for example, have
imminent rent reviews and lease expiries are not considered saleable at an acceptable price.
A large number of prospective problems with specific properties, which might inhibit sales,
are identified in Stages 2 and 3 of the decision making process - therefore the number of
aborted sales in the database appears low.
If sales were a sample of all properties in portfolios, it would be expected that they would
include more properties with attributes which inhibit sale and, therefore, potential time to sale
would extend well beyond those observed average transaction times for the actual sales.
Consequently, the study also suggests that transaction frequency or probability of sale
provides only a partial indicator of asset liquidity. Sale probability depends upon whether the
20
seller is motivated to sell and whether the seller is able to sell. Proxy liquidity measures
based on time to an actual sale are driven by the latter, which may be misleading.
The preliminary analysis of the 187 transactions for transaction time over the last three stages
of the process suggests that very few generalisations can be made concerning the causes of
longer and shorter transactions times. The only apparent trend is the continuing reduction in
the time from exchange to completion, which now appears to average just over two weeks.
However, of the three stages, this is the least variable and the least lengthy so it does not
significantly reduce the overall transaction time. Over the whole data the average transaction
time is 298 days, over 9 months. However, this average is skewed by a small number of very
long transactions and the median transaction time is 190 days, or just over 6 months. Around
25% of transactions take between 50 and 100 days and 60% get completed within 8 months.
These figures need to be placed in the context of the time to transact in securities markets
(even for small capitalisation stocks and those with low free floats).
Given the length of time between exchange and completion is around two to three weeks, the
vast majority of time to sale is in the marketing and due diligence periods. Marketing
(median 88 days) is slightly longer than due diligence (median 62 days). No clear downward
trend through time in either of these two periods is observable from the data, despite the
efforts of the Investment Property Forum in promoting the streamlining of property
transactions (although we should stress that the sample for the 1995/6 period was limited).
Perhaps more surprisingly, there appears to be no reason or relationship between the length of
these two periods. A long marketing period does not lead necessarily to a shorter or longer
due diligence period. This may be because for some complex properties they both take longer
to complete while, in others, some of the due diligence may be undertaken before the final
price is agreed. If a prior offer had been received and later withdrawn, this would possibly
increase the marketing period to final agreed price, but reduce the time to exchange. Reasons
for purchasers withdrawing can be very specific; such as a tenant defaulting or changes to
market conditions in the due diligence period.
Value of property has no apparent effect on length of transaction. Property sector
disaggregation does not validate (or refute) the estimates of McNamara’s (1998) respondents
regarding different transaction times for the different property types and locations, but the
largest sample, standard retails does appear to take far longer than those estimates. However,
the sample also appears to include a number of small, secondary units being cleared out of
portfolios.
21
Overall, the case study interviews provide some insights into the transaction process and the
transactions data gives some indications of the timings of the last three stages of the
transaction process; marketing, due diligence and completion. The key outstanding issue is
the factors that cause extended transaction times. Are these simply ‘liquidity shocks’ that can
occur randomly and are essentially unpredictable? Are certain categories of asset more prone
to such liquidity shocks? Would unsold assets have taken longer to sell? The data could be
examined further for sector differences but without a larger number of transactions the ability
to drill down and disaggregate remains poor. Extending the data collection to more funds and
companies could give extended insight in to selection bias and the drivers for transactions and
deeper analysis of the source files could reveal and categorise the property specific issues
which cause transactions to vary so much in time taken to completion.
22
Bibliography
Ball, M., Lizieri, C. & MacGregor, B. (1998) The Economics of Commercial Property
Markets London, Routledge
Baum, A. & Crosby, N. (1995) Property Investment Appraisal London, Routledge
Bond, S and Lizieri, C (2004) Defining liquidity in property, Working Paper 1, Investment
Property Forum
Collett, D., Lizieri, C. and Ward, C. (2003) Timing and Holding Periods of Institutional Real
Estate, Real Estate Economics, 31,2, 205-222
Fisher, J . Gatzlaff, D. Geltner, D. and Haurin, D. (2003) An analysis of the determinants of
transaction frequency of institutional commercial real estate investment property, Paper
delivered at Skye real estate conference, August 2003.
Guilkey, D. Miles, M. and Cole, R. (1989) The Motivation for Institutional Real Estate Sales
and Implications for Asset Class Returns, AREUEA Journal, 17, 1, 70-82
Hoesli, M. & MacGregor, B. (2002) Property Investment Harlow: Longman
IPF (1995) Report of the Working Party on Streamlining Commercial Property Transactions,
London: Investment Property Forum
IPF (1996) “Readiness for Sale” The Code of Practice for Streamlining Commercial
Property Transaction, London: Investment Property Forum
Key, T., Durkin, S. & McBurney, D. (1998) Liquidity, Trading and Performance. Paper
presented to 7
th
IPD Investment Strategies Conference, Brighton
Lin, Z. & Vandell, K. (2001) Illiquidity and Real Estate Risk, Mimeo, School of Business,
University of Wisconsin – Madison.
McNamara, P. (1998) Exploring liquidity: Recent Survey Findings. Paper to the 7
th
Investment Property Databank Conference, Brighton, November.
23
Strategic allocation process
generates sale requirements
Tactical asset allocation identifies
preferred sectors and regions.
Worth assessment identifies
individual assets preferred
for sale
Unsolicited approach
from (special)
purchaser.
Individual asset recommended for sale
Asset does not
conform to target
portfolio holdings
Internal assessment of sale feasibility
Abort
Delay
Prepare for
sale
Wait for improved
market conditions
Minor preparations for sale
Instruct agents to
assess
marketability and
to estimate sale
price.
Instruct solicitors
to assess legal
constraints to
sale and carry out
preliminary work
External assessment of sale feasibility
Abort
Delay
Prepare for
sale
Wait for improved
market conditions
Minor preparations for marketing and
sale
Comparison of ‘book value’ with Market
Value
Formal marketing
Receipt and assessment of offers/bids
Purchaser’s due diligence
Lender’s due diligence, negotiation
of funding and funding approval
procedures?
Re-negotiation of price?
Exchange of contracts
Completion.
Pressure to
generate cash e.g.
unit redemptions
Major sale constraints
identified
No major sale constraints identified
Major sale constraints identified
No major sale constraints identified
Go back to second/third bidders?
Appendix One :
Taxonomy of the
transaction sale
process
24
25
doc_241395487.pdf