AN ECONOMETRIC MODEL OF UNITED KINGDOM BUILDING SOCIETIES*
| Date | 01 August 1984 |
| DOI | http://doi.org/10.1111/j.1468-0084.1984.mp46003001.x |
| Author | Gordon J. Anderson,David F. Hendry |
| Published date | 01 August 1984 |
OXFORD BULLETIN
of
ECONOMICS and STATISTICS
OXFORD BULLETIN OF ECONOMICS AND STATISTICS, 46, 3 (1984)
0305-9049 ¶3.00
AN ECONOMETRIC MODEL OF UNITED
KINGDOM BUILDING SOCIETIES*
Gordon J. Anderson and David F. Hendry
I. INTRODUCTION
Uñited Kingdom Building Societies are non-profit-rn aking financial
intermediaries with Friendly Society status, which confers on them
various privileges, including tax advantages. They dominate the UK
mortgage market with their assets representing about an 80 per cent
share of housing finance and their liabilities comprising over 40 per cent
of personal sector liquid assets. The former fact alone makes a good
understanding of their behaviour very important since housing
represents over half of personal sector net wealth in the UK and
mortgage finance has a major impact on housing transactions. However,
since their liabilities are highly liquid (mostly obtainable on presenta-
tion of a deposit book, for far longer hours than banks are open) but
oddly are excluded from the widely quoted definitions of money, the
demand function for their deposits is also important. Mortgages are lent
(and deposits borrowed) on a variable interest rate basis which can be
changed as desired by Building Societies; although over 270 exist, the
largest five own over 55 per cent of the total assets of the movement,
and until recently almost all societies quoted the same borrowing and
lending interest rates. Consequently, changes in their interest rates
can have large and immediate effects on housing demand, and on the
* This research was supported in part by grants HR8789 and BOO 220012 from the Economic
and Social Research Council, whose financial assistance is gratefully acknowledged over the
ten years during which this research was in progress. We arc indebted to Frank Srba for
invaluable assistance with the preliminary model estimates and associated computer program-
ming. The research is based on work undertaken in part for the UK Department of the
Environment in 1980. Helpful discussions with John Muellbauer and John Knight significantly
improved the paper.
185
Volume 46 August 1984 No. 3
186 BULLETIN
financial well-being of more than 5 million borrowers as well as the
majority of UK households (there are 32 million deposit accounts,
relative to 21 million households). A recent detailed discussion is
provided by Boléat (1982).
Unsurprisingly, Building Societies' unique and important status have
attracted considerable interest both from governments and from
economists. The former have intervened on several occasions and
applied pressure at other times to influence the level of lending (which
affects the extent of the notorious and chronic 'mortgage rationing'
prevalent in the UK, as well as the rate of change of house prices) and
to reduce rates of interest (which enter the retail price index as well as
having implications for political popularity). Governments have also
altered the tax deductibility of interest payments. Presently, mortgage
interest on an owner-occupied dwelling is essentially the only
deductible interest rate for the personal sector; it may be set against an
individual's marginal tax rate up to a principal of £30,000. All interest
receipts, however, are liable to tax.
Economists have studied Building Societies from many viewpoints.
Reveil (1973) analysed their structure and functioning, Clayton et al.
(1 974) investigated the multiple objectives they claimed to have, Ghosh
and Parkin (1972) and Ghosh (1974) their portfolio and debt behaviour,
and O'Herlihy and Spencer (1972), Riley (1974), Renton (1975, Ch. 1),
Hadjimatheou (1976), Hendry and Anderson (1977), Hewitt and
Thom (1977), Mayes (1979) and Pratt (1980) all provided econo-
metric models. Our previous paper (denoted HA below), proposed a
general theoretical model of dynamic decision-taking rules of Building
Societies which accorded with the available institutional evidence and
sought to reproduce rationing endogenously. Although the model was
not estimated, it was tested in a novel way: the theory was used to
predict precisely which specification errors were likely in the O'Herlihy
and Spencer empirical model. Testing these predictions produced sub-
stantial corroboration, thus endowing the theory with some credibility.
In addition, the decision rules of the theory embedded as special cases
most econometric models then existing.
Since that earlier paper, many of the econometric studies have
implemented variants of the theory model, including an interesting
comparative study with Australian Building Societies by Meen (1982).
Consequently, the theory model remains both relevant and well tested.
This paper reports estimates of a revised version of the original theoreti-
cal formulation, using a more homogeneous objective function as
presaged in HA. The organization of the remainder of the paper is as
follows: Section II briefly describes the basis of the econometric
modelling strategy, Section III discusses the formulation of the theory
model, Section IV the empirical estimates and their evaluation and
interpretation, and Section V concludes the analysis.
To continue reading
Request your trial