A DATA BASED SIMULATION MODEL OF THE FINANCIAL ASSET DECISIONS OF UK, ‘OTHER’ FINANCIAL INTERMEDIARIES

AuthorD. G. Barr,K. Cuthbertson
DOIhttp://doi.org/10.1111/j.1467-9485.1992.tb00619.x
Published date01 August 1992
Date01 August 1992
Scorrish
Joimral
01
PoIirica/
Econos1.v.
Vol.
3Y.
No.
3.
Augusi
1992
:
1992
Scollirh
Economic
Socivly
A DATA BASED SIMULATION MODEL
OF
THE
FINANCIAL ASSET DECISIONS
OF
UK,
‘OTHER’
FINANCIAL INTERMEDIARIES
D.
G.
BARR*
Economics Division, Bank
of
England, Threadneedle Street, London
AND
K.
CUTHBERTSON’
Department
of
Economics, The University, Newcastle-upon- Tyne,
and Economics Division, Bank
of
England
I
INTRODUCTION
The ‘other’ financial institutions
(OFI)
referred
to
in this paper comprise
UK
pension funds, insurance companies, unit and investment trusts and finance
houses; building societies are excluded. The aim
of
the paper is to assess how
far a particular theoretical model based on consumer demand theory can
explain the behaviour
of
asset holdings
of
UK, OFIs and provide sensible
simulation properties.
There is a voluminous applied literature on the demand
for
financial assets
of
the non-bank private sector both single equation (e.g. Laidler,
1977, 1980;
Hall
et
al.,
1989;
Roley,
1985)
and systems approaches (Serletis and Robb,
1986;
Ennis and Fisher,
1984;
Perraudin,
1987;
Backus
et
al.,
1980;
Feige and
Pearce,
1977;
Christensen
et
al.,
1975;
Green,
1984)
but applied work has
largely neglected
a
systems approach
to
the asset decisions
of
OFIs (but see
Keating,
1985).
For
example, Cummins and Outreville
(1984)
claim that their
work represents the first attempt to build
a
structural model
of
US
pension
funds. They utilise Friedman’s
(1977)
optimal partial adjustment model on
a
three asset system. Results are mixed: some variables have wrong signs, there
are no symmetry
or
homogeneity restrictions in the model and the theory model
is
eclectic.
*The views expressed are those of the authors and do not necessarily represent those of the
Bank of England. Useful comments from two anonymous referees are gratefully acknow-
ledged. The usual disclaimer applies.
Date
of
receipt of hnal manuscript:
31
October
1991.
24
1
242
D.
G.
BARR AND
K.
CUTHBERTSON
There are a large number
of
unidentified parameters in the system and this
makes the model somewhat difficult
to
interpret. Attempts in the
UK
have also
yielded results that often conflict with the chosen theoretical model
or
intuitive
a
priori
views. Ryan (1973),
in
a very different model to that in this paper
examines the demand for
a
selection
of
assets by the UK life funds. Honohan
(1980) applies the mean-variance model to the
UK
insurance companies. He
imposes symmetry and and homogeneity restrictions but finds that ‘the model
does not perform well
...’.
Similarly Keating’s (1985) highly restrictive mean-
variance model for pension funds does not perform well (Courakis, 1988).
Weale (1985) utilises a four-asset model
of
the life assurance and pension funds
and attempts to identify speculative and hedging components. He finds
speculative activity is small relative to the hedging component. His model is
closest to that used here but he does not impose long-run symmetry and
homogeneity and asset shares depend
on
the aggregate price
level.
The above
examples are not meant to highlight specific deficiencies in model building but
merely to point out that obtaining acceptable results in this area has proved
extremely difficult, to date.
In this paper we use a systems approach where the theoretical structure is
based on the Almost Ideal Demand System (AIDS) (Deaton and Muellbauer,
1980). The long run asset demand functions are determined using co-
integration techniques (see Hendry, 1986; Granger, 1986) while
in
the ‘second
stage’ we establish
a
data coherent interdependent error feedback model. The
long-run and short-run parameters are identified in the model. The
methodology adopted allows one to ‘search over’ alternative short run
specifications independently
of
the theoretically acceptable long run
cointegrating relationships. We find that with a suitable flexible dynamic
structure we obtain demand functions that are intuitively plausible and exhibit
parameter stability, when the restrictions implied by the AIDS model are
imposed. Thus
our
main aim in this paper, namely to provide acceptable
numerical values for the parameters
of
a model which obeys the restrictions
implied by consumer demand theory (i.e. symmetry and homogeneity) is met.
We also simulate the model to ascertain whether the time path
to
equilibrium
is plausible. The simulation properties are important
if
the model is to be
embedded in a large scale econometric model of the financial system. Clearly
many different theoretical models are (broadly) consistent with consumer
demand theory and we cannot claim generality here. However, when we have
a large number
of
parameters to estimate and interpret and there is
a
need to
investigate simulation properties, the linearity
of
the AIDS model under
symmetry and homogeneity is a considerable practical advantage.
The rest
of
this paper is organised as follows. In Section
11
we outline the
theoretical model and in Section
Ill
we consider the modeling
of
short run
dynamics in a systems framework and associated econometric problems.
In
Section
IV
we discuss data problems and in Section
V
we present
our
empirical
results. We discuss simulation properties in Section
VI
and conclude with
a
brief summary.

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT