Content uploaded by Joaquim Montezuma
Author content
All content in this area was uploaded by Joaquim Montezuma on Jan 30, 2014
Content may be subject to copyright.
Owner-occupied housing and
household asset allocation
A review of the issues
Joaquim Montezuma
Department of Urban Studies, University of Glasgow, Glasgow, UK
Keywords Assets management, Housing, Return on investment
Abstract This second of two related papers in this Journal, reviews empirical evidence available
from the literature on the problem of household’s optimal portfolio when owner-occupied housing
is included in the list of available assets, namely the risk-return performance of residential
investment, and its usefulness in efficient mixed-asset portfolios. The risk-return characteristics of
the housing asset is highly dependent on the type of perspective under analysis (household or
institutional investor’s perspective) and therefore, the two housing investment approaches could
lead to different conclusions about the role of housing investment in an portfolio context. The
consumption demand for housing together with the markets imperfections place serious constraint
on the household’s portfolio problem.
Introduction
This is the second of two related papers published in this journal. The first paper
(housing investment in an institutional portfolio content review of the issues) surveys
the role of residential property in an institutional portfolio context and finds evidence
that residential property is a more effective hedge against inflation than bonds and
shares. Additionally, that paper shows that unsecuritized housing investment not only
generates risk-adjusted returns comparable to those of bonds and shares, but also
exhibits low levels of correlation with classic institutional asset classes. This second
paper reviews empirical evidence available from the literature on the problem of a
household’s optimal portfolio when owner-occupied housing is included in the list of
available assets, namely the risk-return performance of residential investment, and its
usefulness in efficient mixed-asset portfolios (Englund et al., 2000; Flavin and
Yamashita, 1998; Goetzmann, 1993). There are a number of reasons why housing
investment in an institutional portfolio context is not entirely comparable to the
perspective of a household investor. For instance, portfolio constraint imposed by
housing consumption (“housing constraint”) has a critical influence on how households
allocate wealth amongst alternative assets. Additionally, the household is not able to
benefit from the risk and operational costs reduction created by holding large diversified
housing portfolios.
The paper is structured as follows. The next section reviews the literature on
owner-occupied housing in a household portfolio, identifies some weaknesses with the
housing diversification benefits argument and discusses future research directions.
The third section explores some theoretical explanations of property’s diversification
characteristics. Conclusions are drawn in the last section.
The Emerald Research Register for this journal is available at The current issue and full text archive of this journal is available at
www.emeraldinsight.com/researchregister www.emeraldinsight.com/0263-7472.htm
Thanks go to Kenneth Gibb for his helpful comments and suggestions. All remaining errors and
omissions are the sole responsibility of the author.
Housing and
household asset
allocation
267
Received July 2003
Property Management
Vol. 22 No. 4, 2004
pp. 267-275
qEmerald Group Publishing Limited
0263-7472
DOI 10.1108/02637470410558134
Residential property in a household portfolio context
As mentioned earlier housing investment in a multi-asset portfolio context is
considered from two substantially different perspectives: institutional investor’s
perspective, in the earlier paper and the household’s perspective. The risk-return
characteristics of the housing asset is highly dependent on the type of perspective
under analysis and therefore, the two housing investment approaches could lead to
different conclusions about the role of housing investment in a portfolio context.
Housing is a dominant asset category in a household’s wealth portfolio and one of
the most important components of household consumption is the expenditure. For
instance, the average US homeowner holds 88 percent of his non-pension wealth in
home equity and the average household in Western Europe and US spends between 25
to 35 percent of its income on housing services (Englund et al., 2000).
Housing services consumption is essentially determined by the household’s
ownership of residential property. Thus, the household’s demand for housing, which
could be optimal from the point of view of the consumption of housing services, may
differ from the optimal level of housing allocation in a purely portfolio investment
context. In other words, the portfolio constraint imposed by the housing consumption
(“housing constraint”) has a critical influence on how households allocate wealth
among alternative assets.
Households would like to consume an optimal mix of consumption good and
housing services and maintain a balanced wealth portfolio and, at the same time, to
reduce the residential transaction costs. However, housing markets have several
imperfections such as tax distortions, transaction costs and moral hazard issues.
According to Flavin and Yamashita (1998), those market imperfections prevent
households from holding a diversified portfolio and renting to satisfy their housing
services consumption. In that sense, rental housing is not a perfect substitute for
owner-occupied housing. Other market frictions that influence homeowners’ wealth
portfolio include uninsurable labour income risk and borrowing constrains (Cocco,
2000). The latter author states that labour market incompleteness avoid human capital
to be capitalized and its risk insured. The author finds empirical evidence that both
labour income and interest rate risk crowed out housing investment. Grossman and
Laroque (1990) show that the transaction costs reduces the percentage of household’s
wealth allocated to risky assets (e.g. shares) after he purchases a new house. Following
Grossman and Laroque (1990), Cocco (2000) finds that transaction costs decrease the
frequency of housing adjustment and restricts investor’s ability to take advantage of
serial correlation in house prices.
By investing in large housing portfolios, the institutional investor has, relative to
the household investor, a lower risk exposure and higher net returns. Diversification
(e.g. geographical diversification) that results from holding a larger residential
portfolio, can improve investor welfare by reducing the exposure to any specific risk
borne by any single house. Additionally, the institutional investor benefits from
considerable economies of scale. In other words, size enables institutions to have lower
operating costs per amount of residential investment, and hence, higher net returns.
Papers by Devaney and Rayburn (1988), Goetzmann (1993), Flavin and Yamashita
(1998), Englund et al., (2000), Eichholtz et al., (2000) and Gatzlaff (2000), using different
procedures to measure housing returns, have all analysed the problem of households’
optimal portfolio when owner-occupied housing is included in the list of available
PM
22,4
268
assets. Their findings also confirm the institutional portfolio studies that correlations
between housing and other assets are low. However, when they look at housing
investment within a mean-variance framework, the precise amount of residential
assets to include in optimal portfolio is not the same for both types of investors.
Devaney and Rayburn (1988) examine the role of residential property in the
household investment portfolio, with capital appreciation returns based on a
transaction price index for Memphis (US) over the period 1970-1984. They find
empirical evidence of low correlation between residential property investment and
financial assets (shares, short and long term bonds). Additionally, they suggest that
there are low incremental benefits of diversification across housing submarkets[1] in
the same city. One note of caution in reading the empirical results is that the authors
use a housing appreciation return instead of a total housing return.
Goetzmann (1993) also examines the role of single family homes in the investors’
portfolio, with capital appreciation returns based on a value-weighted repeated-sales
index[2] in four urban US markets over the period 1971-1985. The author finds that
there are potential allocations for home investment in optimal portfolios comprising of
shares and bonds, although housing is not the predominant asset class and the
proportion of those allocations increases with the risk aversion of the investor. The
potential diversification benefits result is due primarily to the weak correlation
between housing returns and those of financial assets. Additionally, the paper confirms
the findings of Goetzmann and Ibbotson (1990) about the geographical diversification
benefits within residential property. The author presents evidence that even by
diversifying across four houses, one in each regional market, it is possible to reduce the
risk significantly. Furthermore, the study suggests that owning four homes in different
parts of the country generates risk-adjusted returns comparable to owning a fraction of
a portfolio composed of thousands of properties within a single region. In other words,
regional diversification dominates local diversification. Goetzmann further observes
that given the short time interval of analysis, one must be cautious in generalising the
results obtained regarding suggestions for future investment. Nevertheless, according
to the author, allocation studies based upon inputs derived from longer time intervals
suggest that housing may command an even larger percentage of the portfolio.
Flavin and Yamashita (1998) using data from the Panel Study of Income Dynamics
(PSID)[3] on home values, compute real after-tax returns to owner-occupied housing
from 1968 to 1992. They report that owner-occupied housing returns and standard
deviation are slightly lower than those of shares. The authors also indicate that housing
returns are weakly correlated with those of financial assets (mortgages, short- and
long-term government bonds, and shares). Their empirical results show that households
hold different portfolios of financial assets because the constrain imposed by the
housing consumption varies considerably across households. In fact, the ratio of
housing investment to net worth declines over the household’s life-cycle, inducing a
life-cycle pattern in the household’s portfolio. Furthermore, they argue that
owner-occupied housing is an attractive asset for hedging fluctuations in financial
assets.
Recently, Englund et al., (2000), using a quarterly repeat sales house price index for
eight major Swedish regions between 1981 and 1993, examine the owner-occupied
housing investment in a multi-asset optimal portfolio (short- and long-term bonds,
market share index, and an index for real estate corporations traded on Stockholm
Housing and
household asset
allocation
269
share exchange) over different holding periods. The authors report that returns to
housing are positively correlated with real estate shares and negatively correlated with
bonds, and all correlations with housing are stronger at longer periods of time. The
correlation with market share index is nearly zero. Moreover, they point out that for
short-holding periods, housing allocation in an optimal portfolio is nearly zero. On the
other hand, for longer periods, low risk portfolios contain 15 to 50 percent on housing.
In disagreement with the Hoesli and Hamelink (1997) findings for Switzerland, the
authors indicate high correlations between housing returns across the different regions
of Sweden during the sample period. As a consequence, the diversification benefits
from holding a multi-regional housing portfolio within Sweden appear to be negligible.
Using a quarterly value-weighted repeated-sales index for Florida during the
1970-1999 period, Gatzlaff (2000) finds that housing returns are negatively correlated
with those on long-term government bonds, shares, and real estate investment trusts
and positively weakly correlated with equity REIT total returns. Moreover, the paper
shows evidence that housing assets allocation has potential benefits for mixed-asset
portfolios. They find that the efficient allocation of housing in these portfolios ranked
from 20 to 45 percent. Furthermore, the paper argues that housing assets reduce
the efficient institutional portfolios’ allocations to non-residential property. Thus, the
institutional investor should optimise not the plan’s investment portfolio, but rather
the wealth of individuals’ personal portfolio (members of the plan). This argument,
however, seems not valid for defined-benefit pension funds and life insurance
companies, because their liabilities tend to be defined in nominal terms.
Another recent study, Eichholtz et al., (2000), rely upon monthly value-weighted
repeat-sales index during the period 1980-1997 for five major US’ cities, to analyse the
performance of owner-occupied residential property in portfolios relative to shares and
bonds. The study reports that the housing property returns (measured by the
appreciation return component) and standard deviation have been lower than those of
shares and bonds. Additionally it shows that correlations between housing and
financial assets are small. In concordance with Hoesli and Hamelink’s (1997) results
(for Switzerland) and in disagreement with the findings of Englund et al., (2000)
(for Sweden), the paper shows evidence of diversification benefits from holding a
multi-regional housing portfolio within the US, since the relationship between the
different regional housing markets is not significant. In contrast to the Gatzlaff
argument, the study shows that the demand for shares and bonds is not strongly
affected by the issue of home ownership, as shares and bonds do not provide a good
housing property hedge.
In spite of the research evidence, there are several empirical weaknesses with the
housing diversification benefits argument.
First, the severe measurement problems of housing values. Thus, the results
obtained might actually be an artifact of the house price index used rather than a true
feature of the housing market. Second, the heterogeneity of the procedures to calculate
the housing returns. Some of the reviewed studies rely on appreciation returns as a
proxy for total returns and others use total returns. In addition, the income component
of the latter could be based either on a residential rent index or constant rent-to-value
ratio. Therefore, one must be careful to collate findings based on heterogeneous
housing returns.
PM
22,4
270
Third, the results are limited by the use of ex post returns instead of ex ante returns.
The previous studies assume that historic data are a good estimation of future values
on returns, variances and correlations. Sharpe (1990) however, suggests that financial
historical data appears to be useful for predicting future variances, reasonably useful
for forecasting correlations and virtually useless for estimation of expected returns.
Furthermore, MacGregor and Nanthakumaran (1992) report that the correlation of
property returns with those of financial assets is not stable over time. The authors
argue that the correlation is dependent on the economic fundamentals registered
during the time period analysed.
Fourth, the former empirical analysis was based on short-term transaction period
returns (that is, annual or quarterly holding periods). Recent research has indicated
that over long intervals there is a positive relationship between share returns and
commercial property price changes (Quan and Titman, 1999). Additionally, Geltner
and Miller (2001) suggest that when asset returns are correlated across time, the
optimal long-term portfolio could differ from the optimal short-term portfolio.
Fifth, the traditional minimum variance optimisers ignore the special features that
characterize some types of assets like property (illiquidity, marketability costs, and
measurement issues).
The statistical measure of risk, standard deviation, does not capture the illiquidity
of direct property, and thus standard optimisers tend to over weigh the portfolio
allocation into property. However, it is possible to account for these liquidity risk, by
subtracting illiquidity premiums from the average return on property. For instance,
MacGregor and Nanthakumaran (1992) use a quarterly illiquidity premium of 50 basis
points for UK property and Hoesli and Hamelink (1997) use illiquidity premiums of 50,
100 and 150 basis points for Swiss residential property. The illiquidity issue appears to
be more important in the commercial market than in the housing market, since the
latter, as mentioned before, exhibits a greater number of participants and transactions.
The traditional MPT ignores the marketability costs (Ibbotson and Siegel, 1984).
The marketability costs are related to the buying and selling process and comprise
information, transaction, and divisibility costs.
The information costs are, those related with the asset valuation process. The
unsecuritized property market, contrarily to the security markets, is characterized by
the absence of a transparent marketplace where it is easy to learn about expected asset
returns. Thus, the direct property should have higher before-cost expected returns than
assets that are easier to learn about (e.g. shares and bonds) (Ibbotson and Siegel, 1984).
The same authors also state that these type of costs tend to benefit the larger
investments, due to economies of scale.
The transaction costs (including the legal costs associated with the transaction,
costs of advertising, costs of brokers used to effect the transaction) in buying and
selling property are much greater than those in securities market. The transaction
costs could remove part of the potential residential investment profit from establishing
trading rules that consistently yield above-normal returns (Cho, 1996). Even so, the
transaction costs could be relieved by holding the property assets during longer
periods of time (Geltner and Miller, 2001). The transaction costs are extremely
important for investment strategies that require frequent trading, but less critical for
the long-term asset allocation decisions that are those usually faced by the institutional
investors.
Housing and
household asset
allocation
271
Divisibility costs emerge from the large size of some investments, such as property
(Ibbotson and Siegel, 1984). The indivisibility could cause some investors to hold a
non-optimal quantity of a particular investment. For instance, the smaller investors
may have difficulties in eliminating the systematic risk of property portfolios (Hoesli
and Macgregor, 2000) and, consequently, the portfolio theory methodology must be
applied to large portfolios only.
The mentioned special features that characterize the property assets comprise some
of the explanations like why the standard MPT optimisers tend to over weigh the
portfolio allocation into property. Despite the fact that these characteristics have an
effective influence on the “true” risk and return of property assets, they are not
incorporated in the traditional mean-variance approach.
Finally, Portfolio Theory assumes that investors derive utility from consumption,
which in turn is supported by financial wealth. The utility function takes a standard
form in which relative risk aversion is constant. In contrast, the behavioural finance
followers postulate that some investors have non-standard preferences. In a similar
vein, the prospect theory of Kehneman and Tversky (1979) models a utility function
that has a kink at the reference point and different patterns of curvature above and
below the point.
The prospect theory has been applied to housing price cycles with some interesting
results. For instance, Genesove and Mayer (2001) and Engelhardt (2001) suggest that
nominal loss aversion – whereby house owners are particularly reluctant to realising
nominal price losses– is an important explanation (along with housing equity
constraints) for some housing market puzzles, like the strong positive price-volume
correlation observed over the housing cycle and the negative correlation between house
prices and time on the market. Given the nominal loss aversion effect, one can
hypothesize that, during downward income cycles, the potential house purchasers may
prefer renting to buying houses at inflated prices. Such comportment would have a
positive effect over the demand in the rented housing sector during the downward
income cycles. These are times in which the investors in the private rented housing
would especially like their investments not to perform poorly. Additionally, Bernardes
and Montezuma (2002) consider the (indirect) effects that loss aversion in consumption
decisions might have on the housing market via demand behaviour. They argue that
house prices will be much more responsive to negative changes in income than to
positive changes, even at the start of the price cycle.
Others cite behaviour theory, include mental accounting, too much emphasis on
recent experience, overconfidence, amongst others. Behaviour finance appears to be a
promising research area with the potential to explain some types of investor behaviour
and possibly some patterns in asset pricing (Campbell and Viceira (2002)).
One can identify several possible directions for future research. The first involves a
consideration of the allocation factors that are not taken into account in the simplistic
modern portfolio theory framework, namely, the liability matching issue in the
institutional investment strategies and the special housing attributes (e.g. illiquidity,
indivisibility, political risk). One can speculate that the former factor may have a
positive impact on the proportion of institutional funds to residential property, since
the housing returns usually track household growth and incomes.
The inclusion of foreign-based equities as an institutional eligible asset could also
be a novel addition in relation to reviewed research. One can expect the inclusion of
PM
22,4
272
foreign equities to have a negative impact on the residential allocation, as a result of the
typically lower correlations between returns on foreign and those on national shares.
An additional direction involves the utilization of a multifactor asset pricing model
in order to produce forward-looking instead of historical-based risk and return
estimates (that is, ex post risk and returns). Given the cyclical behaviour of residential
property, the forward-based approach to estimate both risk and returns on housing
may well be advantageous for the analysis of residential property as an investment.
Developments in the study of the risk-return characteristics of housing property and
its role in a mixed-asset portfolio may also include the utilisation of residential data
pooled across countries, instead of data from a single country. Other research
extensions on this topic include the use of multi-asset time series over longer holding
periods (like five years) and the use of lags in time-series regressions.
Theoretical arguments for the property’s diversification characteristics
Different theoretical explanations are presented for the empirical results on the
correlations between property and financial assets. When the property returns are
affected essentially by the same economic factors (or systematic influences) that affect
the financial assets, property diversification benefits to multi-asset portfolios will be
minimal. In other words, when common economic factors strongly affect returns on
both property and financial asset classes, diversification across property may not be
cost justified. On the contrary, if the property returns also respond to property-specific
factors (non-systematic influences), diversification will be beneficial.
For instance, Quan and Titman (1999) hypothesize that positive correlations
between long-term share returns and those of commercial property are related to the
fact that commercial property and shares are both driven either, by changing
expectations (rational or irrational) of future economic growth or, by current economic
fundamentals.
Conversely, MacGregor and Nanthakumaran (1992) hypothesize that the low
correlations between property and financial assets are, to some extent, a result of the
property-specific factors rather than an artifact of the data. The authors’ argument is
that the property market supply responses to unexpected changes in the economic
fundamentals are slow. The production of new property is slow, not only because the
planning and building process is time-consuming, but also because the development
decision process itself is slow, i.e. the developers, in order to minimise their
investments risk, wait until the demand’ and prices’ increase is completely clear. The
short-term supply rigidity associated with a deficient demand forecast cause an
oversupply of property when demand is beginning to fall.
One can speculate that there are residential property-specific features able to reduce
the correlation between residential property returns and those of financial assets. For
example, foreign competition may lead to decreases in domestic wage rates, which in
turn leads to increased corporate profits and higher share prices. However, if
households’ income decreases, the housing demand will also decrease, and the
residential property values will fall. Changes in transport and communication network
features could also induce a negative relation between property and share values. For
instance, a more efficient transport and communication network may lead to increases
in corporate productivity and higher share prices, which in turn cause land rents to
decline and, consequently, decrease land values.
Housing and
household asset
allocation
273
From the empirical studies reviewed and the theoretical arguments offered, one
cannot reject, at least in the short-term, the hypothesized ability of housing to generate
improved risk-adjusted returns when added to share and bond portfolios.
Conclusions
The risk-return characteristics of the housing asset is highly dependent on the type of
perspective under analysis and therefore, the two housing investment approaches
could lead to different conclusions about the role of housing investment in a portfolio
context. The household’s demand for housing, which could be optimal from the point of
view of the consumption of housing services, may differ from the optimal level of
housing allocation in a purely portfolio investment context, the consumption demand
for housing together with the market imperfections places a constraint on the
household’s portfolio problem.
The empirical studies generally indicate that housing returns are weakly correlated
with those of financial assets (mortgages, short- and long-term government bonds, and
shares). They also suggest that housing is able to diversify a multi-asset portfolio and
that the proportion of housing allocations increases with the risk aversion of the
investor. However, the optimal amount of housing varies across the studies.
There is some theoretical and empirical evidences that the household’s optimal
portfolio is constrained by the ratio of housing investment to net worth, and this ratio
tends to be related to age. The households tend to hold different portfolios of financial
assets over their life-cycle. For example, a young household compared with a mature
household tends to hold an optimal portfolio with lower allocation on shares.
Because the housing returns tend to exhibit positive autocorrelation and the
transaction costs are high, the optimal household’s portfolio is dependent upon the
holding period (Englund et al., 2000). For instance, for short-holding periods, housing
allocation in an optimal portfolio is nearly zero. On the other hand, for longer periods,
low risk portfolios contain 15 to 50 percent on housing. Additionally, the transaction
costs decrease the frequency of housing adjustment and restrict investors’ ability to
take advantage of serial correlation in house prices. Finally, labour income and interest
rate risk seem to crowd out housing investment (Cocco, 2000).
Notes
1. The residential submarkets are based on Memphis City Planning Commission “planning
districts”. Planning districts were drawn to preserve the architectural integrity and to
incorporate natural and manmade boundaries.
2. The author uses estimated return series provided by Case and Shiller (1987).
3. On the PSID, the housing prices result from the owners’ own assessments of house values.
References
Bernardes, L. and Montezuma, J. (2002), “Loss aversion in consumption behavior and housing
price dynamics”, paper presented at the American Real Estate and Urban Economics
Association – 2002 Annual Conference, Atlanta, GA.
Campbell, J. and Viceira, L. (2002), Strategic Asset Allocation – Portfolio Choice for Long-Term
Investors, Oxford University Press, London.
Case, K.E. and Shiller, R. (1987), “Prices of single-family homes since 1970: new indices for four
cities”, New England Economic Review, pp. 45-56.
PM
22,4
274
Cho, M. (1996), “House price dynamics: a survey of theoretical and empirical issues”, Journal of
Housing Research, Vol. 7 No. 2, pp. 145-72.
Cocco, J.F. (2000), “Hedging house price risk with incomplete markets”, working paper, London
Business School, London.
Devaney, M. and Rayburn, W. (1988), “When a house is more than a home: performance of the
portfolio household portfolio”, The Journal of Real Estate Research, Vol. 3 No. 1, pp. 75-85.
Eichholtz, P., Koedijk, K. and De Roon, F. (2000), “The portfolio implications of home ownership”,
working paper, Maastricht University, Maastricht.
Engelhardt, G. (2001), “Nominal loss aversion, housing equity constraints, and household
mobility: evidence from the United States”, mimeo, Center for Policy Research, Syracuse
University, Syracuse, NY.
Englund, P., Min-Hwang and Quigley, J.M. (2000), “Hedging housing risk”, working paper
N. W00-07, University of California, Berkeley, CA.
Flavin, M. and Yamashita, T. (1998), “Owner-occupied housing and the composition of the
household portfolio owner the life cycle”, working paper No. 6389, National Bureau of
Economic Research.
Gatzlaff, D.H. (2000), “The effect of single-family housing on multi-asset portfolio allocations”,
working paper, Florida State University, Tallahassee, FL.
Geltner, D. and Miller, N. (2001), Commercial Real Estate Analysis and Investments,
South-Western Publishing, Cincinnati, OH.
Genesove, D. and Mayer, C. (2001), “Loss aversion and seller behavior: evidence from the housing
market”, Quarterly Journal of Economics, Vol. 116 No. 4, pp. 1233-60.
Goetzmann, W.N. (1993), “The single family home in the investment portfolio”, Journal of Real
Estate Finance and Economics, Vol. 6 No. 3, pp. 201-22.
Goetzmann, W.N. and Ibbotson, R. (1990), “The performance of real estate as an asset class”,
Journal of Applied Corporate Finance, Vol. 3 No. 1, pp. 65-76.
Grossman, S.J. and Laroque, G. (1990), “Asset pricing and optimal portfolio choice in the presence
of illiquid durable consumption goods”, Econometrica, Vol. 58, pp. 25-51.
Hoesli, M. and Hamelink, F. (1997), “An examination of the role of geneva and zurich housing in
swiss institutional portfolios”, Journal of Property Valuation & Investment, Vol. 15 No. 4,
pp. 354-71.
Hoesli, M. and MacGregor (2000), Property Investment – Principles and Practices of Portfolio
Management, Longman – Pearson Education Limited, Harlow.
Ibbotson, R. and Siegel, L. (1984), “Real estate returns: a comparison with other investments”,
Journal of American Real Estate and Urban Economics Association, Vol. 12 No. 3, pp. 219-42.
Kahneman, D. and Tveraky, A. (1979), “Prospect theory: an analysis of decision under risk”,
Econometrica, pp. 263-91.
MacGregor, B. and Nanthakumaran, N. (1992), “The allocation to property in multi-asset
portfolio: the evidence and the theory reconsidered”, Journal of Property Research, Vol. 9,
pp. 5-32.
Quan, D. and Sheridan, T. (1999), “Do real estate prices and stocks prices move together? an
international analysis”, Real Estate Economics, Vol. 27 No. 2, pp. 183-207.
Sharpe, W. (1990) in Magin, J. and Tuttle, D. (Eds), Asset Allocation, in Management Investing
Portfolios – A Dynamic Process.
Housing and
household asset
allocation
275
