English
abstract of chapter 3 of Jacques Friggit’s book “Home
prices, property derivatives and risk management” (Economica
publisher, 2001)
A.
The missing market
Several thousand
years after the first leases indexed on the volume of crops transferred risks
between peasants and landowners, real estate assets remain the only large
family of assets which has not generated a modern, active derivative market.
Three
prerequisites condition the emergence of a real estate derivative market:
-
a proper underlying asset, which in the case of real estate can only be
an index, since real estate assets are not fungible,
-
potential markets for real estate risk trading,
-
a launch market
large enough to provide a first "critical mass" of reasonably liquid
instruments.
Nowhere
have all of these conditions been fulfilled up to now.
Indices
robust enough to be used as underlying asset can be found in only a few
countries. Until recently, France was
not part of them. It should be stressed
that real estate companies' stocks, since their price is little correlated with
the price of physical real estate, cannot be used as underlying assets for
proper real estate derivatives.
The size of
the potential market has been reduced by the general illiquidity of physical
real estate assets, itself a consequence of their lack of fungibility. Geographical heterogeneity has played a role:
there are not necessarily many real estate markets small enough to allow price
hedging by one index but large enough to provide sufficient trading
volumes. The fact that real estate risk
has been perceived traditionally as low in many countries has further reduced
the demand for derivatives. Another
impediment has been the traditionally huge cultural differences between real
estate and financial markets professionals.
All these
factors explain that, up to now, no liquid real estate derivative market has
come to existence anywhere[1].
The
United Kingdom provides two recent experiments.
The first
experiment resulted in a failure. In May
1991, the London Futures and Options Exchanges launched an organized market on
four real estate indices, both in commercial and residential real estate. Confronted from the start to very low volumes
of transactions, the organizers of the market created artificial
liquidity. The regulators stepped in,
closed the market, fined several participants.
As (Roche, 1995) remarks in his thorough analysis of this experiment,
one of the consequences was that "property futures acquired an unpleasant
taint of sleaze, not properly understood anywhere, and the whole cause of
property derivatives was set back over a decade". One major reason of this failure probably was
that the FOX attempted to create from scratch an organized market, without
waiting for OTC instruments to first familiarize the markets with real estate
derivatives.
The
second experiment, Barclays' property index certificates, had a better
outcome. Even though the secondary
market for these securities has not been liquid, this experiment provides the
first example of significant price indexed real estate derivatives.
In
France, until now, no such real estate derivative market could develop since no
price index robust enough to be used as underlying asset existed.
B.
The emergence of robust databases and indices in France
Since the
seventies, the French notaries have been building a comprehensive database of
real estate transactions. Their database
contains 5 million records in October 2000 and incorporates 0.8 million new
records every year. The coverage rate is
around 70% for home resales. An excerpt of the database containing the
transactions in Paris and its suburbs posterior to 1990 is sold publicly.
Based on
this database, the notaries and INSEE (the National Institute of Statistics)
publish home price indices, denominated "indices
Notaires-INSEE". These indices are
based on transaction prices (as opposed to appraisal prices) of home resales. They use a
hedonic methodology which tends to filter out quality effects. INSEE provides the methodology and controls
the quality of the indices.
The
indices cover 87% of the territory at the end of 2000. The remaining 13% should be covered before
2002. Their reactivity is low for now:
they are calculated quarterly but their base (length of the period on which the
price is calculated) is 6 or 12 months.
Nevertheless, there are enough data in the database to calculate indices
on much shorter bases, so that true quarterly and monthly indices should be
available soon.
The
quality of the notaries' database and the scientific expertise of INSEE make
the indices robust enough to be used as underlying assets for derivative
financial instruments. Thus, the first
precondition mentioned above for real estate derivative markets to take off -
the availability of proper price indices - looks solved in France for residential
real estate.
The
production of similar commercial real
estate price indices has been impaired by the fact that in France many
commercial real estate transactions used not to be registered by notaries
since, to avoid real estate transaction taxes, they were completed as stock
transactions. A reform of real estate
transaction taxes in 1999 has significantly reduced this form of tax evasion,
but commercial real estate indices on the model of the Notaires-INSEE home
price indices will not be calculated from the notaries' database before several
years, if ever. The IPD-France indices
are based on appraisal values, as the IPD-UK indices. They have not reached the same level of
consensus as their British counterparts yet.
C.
Home price risk and return in France for the last 50 years
Three
fields of the notaries' database describe the previous transaction: its date, price and type (purchase, inheritance,
etc.). They are particularly important
since they allow a reconstitution of price indices on past periods (back to the
1930s) during which the database did not even exist. Such long price series are crucial to
describe the past price process with a decent enough confidence interval and
thus to facilitate derivative pricing and to reduce the risk premia inevitably associated with new financial
products. In addition, in Paris, yearly
home prices can be reconstituted back to 1840.
The main
results on the last 50 years are the following:
-
from a very low level in 1948 induced by wartime rent controls and
inflation, average home prices in France have grown very fast from 1948 to
1965; since 1965, they have fluctuated around the disposable income per
household (chart 1);
-
home price local differentiation has been real but home prices have been
highly correlated (R>0.8) on most of the territory; the main cause of local
differentiation has been that the real estate crisis of 1987-1995 took place
only in the Paris region and in a few other places.
-
home price returns do not follow a lognormal process: volatility grows
faster than the square root of the time scale, first order autocorrelation is
positive for 1 year returns and negative for 5 year returns.;
-
home price volatility has been lower than stocks volatility but
equivalent to bonds volatility and credit cost volatility;
-
home prices have not been highly correlated with bonds and stocks
(including real estate companies' stocks);
-
the gross domestic product (GDP) has been much better correlated with
home prices than with bonds and stocks;
-
the same has been true for the disposable income per household (which
evolves in parallel with
GDP).
chart 1 - Home price divided by the
disposable income per household, basis 1965=1
Source: after the sources mentioned in appendix 3.
D.
Potential markets
Derivatives
can be used for hedging, exposure or arbitrage (table 1).
table 1 - Potential
markets for home price index derivatives and order of their size
|
Motivation |
Index position |
|
|
|
short |
long |
|
|
|
Households
which save in order to buy a home.
Tens of billions € |
|
|
Builders,
their shareholders and their lenders during a construction project. Several billion €? |
|
|
|
Investors
disvesting from housing. A few billion €? |
Investors
investing into housing. A few billion €? |
|
|
Holders
of mortgage loan portfolios in quest of a hedge against credit risk. A few billion €? |
|
|
Hedge against home price risk |
Financial
firms in quest of a better insertion of real estate into their assets and
liabilities management. Several
billion €? |
Financial
firms in quest of a better insertion of real estate into their assets and
liabilities management. Several
billion €? |
|
|
|
Insurance
firms which insure risks contingent on home prices. Small? |
|
|
Local
and national governments in quest of budget balance smoothing. 15 billion €? |
|
|
|
Hedge
of tax and non tax income from infrastructure projects. Small |
|
|
|
Investors
in quest of diversification outside of stocks and bonds. Tens of billions € |
Investors
in quest of diversification outside of stocks and bonds. Tens of billions € |
|
Exposure to home price risk |
Participants
in quest of a long term hedge against slow economic growth. Tens of billions €? |
|
|
|
Directional
trading. Size? |
Directional
trading. Size? |
|
Arbitrage |
Small
at first |
Small
at first |
Hedging
As far as
hedging is concerned France seems to present a specific opportunity compared to
countries such as the USA: French households accumulate huge savings in the
perspective of the purchase of a home.
These savings, which amount to between 80 and 150 billion € (depending on how one evaluates
them) are invested either in regulated savings products, in stocks or in
bonds. They are exposed to the risk of
home price increase, and none of these three forms of investment provides a
proper hedge of this risk (as mentioned above, the correlation of stocks and
bonds with home prices is low). Since
households, as far as their housing projects are concerned, are very sensitive
to security, one may suspect that savings plans indexed on the price of
residential real estate, if they existed, would meet a huge success.
On the
other hand, as in other countries, real estate derivatives could have many
other applications:
-
hedging of new housing construction projects,
-
hedging of the disvestment strategy of an
institutional investor,
-
hedging of the credit risk of a porfolio of
mortgages (highly correlated, one may suspect, with home prices and with GDP,
itself well correlated with home prices),
-
better integration of real estate into the general risk management and
ALM of financial institutions,
-
etc..
The
geographic differentiation of home price risk makes that several indices will
have to be used to optimize the quality of the hedge of individual projects
(savings plan of a given household, construction project in a given city). Long period correlation analyses show that
combinations of two indices, of home prices in Paris and home prices in
Province(France minus the Paris region), provide a hedge at 5 years with
R>0.9 in the Paris region (which represents 30% of the amount of
transactions in the whole country) and R>0.8 in most of Province. Large financial firms with a customer basis
spread over the whole territory could mutualize the
geographic component of the risk.
Exposure
The
properties of home prices underlined above show that financial instruments
indexed on them would provide portfolio managers with a huge diversifying
power:
-
their correlation with stocks and bonds would be low;
-
they would provide a diversification by the "real economy"
(GDP in constant currency) as opposed to the "financial sphere";
-
they would provide a geographic diversification which other instruments
will not provide or will provide to a lesser extent: after European currency
unification, the risk on bonds is unified, and the systematic risks of the
various European stock markets will get closer, whereas home prices will remain
geographically differentiated throughout Europe (one reason being that they are
highly correlated with GDP, which will converge only slowly throughout Europe);
-
they would
provide a sectorial diversification, while presently
no financial instrument replicates properly home prices.
In
addition, since home prices are significantly correlated with GDP, a short
position in home price index would provide a long term hedge again slow
economic growth, which may be of interest to, for instance, pension funds. This hedge would be partial, but would be
better than the hedge provided by other existing financial instruments,
including so-called cyclical stocks.
Arbitrage
Arbitrage
provides most of the liquidity on large markets. Transaction taxes and the general illiquidity
of the physical market will impair arbitrages between physical assets and
financial instruments indexed on home prices.
Therefore, arbitrage will develop only after a first issue of simple,
liquid instruments such as bonds indexed on home prices has provided a stock of
assets which replicate the index and which can be arbitraged.
E.
Launch market
For the
launch of a derivative market to succeed, the first market plays a critical
role, by:
-
enabling a market price of indexed instruments to emerge through liquid
enough trading,
-
building a minimum stock, a "critical mass" of basic indexed
instruments liquid enough to be used as building blocks for more complex
indexed instruments,
-
making possible
arbitrage strategies, which will ensure the coherence of the price system and
increase liquidity.
Whereas
one can suspect that the main hedging market would be the market of households
savings, such a market could probably not be the launch market: few financial
firms would launch a retail product without being sure to find counterparts to
which they could resell the risk they would have collected whatever its amount,
and without being reasonably sure of the price at which this would be possible.
This
implies a preexisting market of reasonably liquid instruments. Several possibilities can be listed.
-
Many French institutional investors own not only commercial real estate
but also huge residential real estate assets, particularly in and around
Paris. On the long run, they have been
and are expected to remain net sellers of residential real estate. The illiquidity of physical real estate
assets forces them to sell slowly. Such
investors, by issuing bonds indexed in capital on a price index, could raise
capital and sell their residential real estate risk by anticipation.
-
Local
governments could issue bonds indexed on home prices to smooth their budget
balances, which are highly correlated with home prices.
These
bonds (or equivalent indexed instruments) would be bought and traded by
investors looking for risk diversification or hedging. This would supply a basis for further
developments.
The real
estate crisis of 1987-1995 may contribute to the constitution of a launch
market. On one hand, it reminded market
participants that real estate prices fluctuate and may even decrease (which had
been forgotten in France since the thirties), and thus increased home price
risk awareness and potential market interest in hedging it. On the other hand, by generating an acute
banking crisis, it narrowed the cultural gap between real estate and finance
professionals inside financial institutions, forcing them to get together and
solve the crisis.
F.
Conclusion
New
powerful databases and the Notaires-INSEE indices have solved the problem of the
underlying asset, which up to now was an insurmountable obstacle to real estate
derivatives in France.
Specificities
of French households' savings habits provide a huge potential hedging market,
and the diversifying power of real estate derivatives should be of great
interest to investors.
The
existence in France of economic agents which are long in home price risk and
may wish to disvest it creates an opportunity for the
constitution of a first "critical mass" of liquid indexed
instruments, on the basis of which a home price index derivative market could
develop further.
The
recent real estate crisis has helped narrow the cultural gap between real
estate and finance professionals.
Therefore
perspectives for real estate derivatives in France look better than ever.