On the Aggregate Housing Market Implications of Labour Market Change
| Date | 01 September 1998 |
| Author | Mark Andrew,Geoffrey Meen |
| DOI | http://doi.org/10.1111/1467-9485.00105 |
| Published date | 01 September 1998 |
II
NTRODUCTION
Given the important changes that have taken place in labour markets in the UK
in recent years, it would be surprising if there were no knock-on effects to
housing markets. Even basic microeconomic models of household behaviour
indicate that labour supply, consumers’ expenditure, housing demand and
household financial asset accumulation are all simultaneously determined. It is
with the interactions between markets that this paper is concerned and, in
particular, how the interrelationships may have changed during the nineties.
However, although a large literature exists on how housing markets altered in
response to structural changes originating in financial markets during the
eighties, to date, no corresponding literature exists on the implications o f labour
market liberalisation except through survey evidence.
The labour market reforms of the eighties and nineties were primarily
microeconomic in orientation, designed to improve the workings of market
mechanisms. In terms of analysing their effects on housing, theref ore, one
natural approach might be to use cross-section or panel data to estimate joint
models of housing choice and labour supply behaviour. However, our
orientation is rather different. We are concerned with the aggregate housing
market implications of these micro labour market changes. Just as financial
liberalisation in the eighties was a micro measure, but had far-reaching
macroeconomic effects, so labour market reform might also be expected to
generate macroeconomic housing market changes.
Our focus means that special attention has to be paid to the issue of
aggregation. Aggregation would not be a problem if (i) all agents had the same
housing demand functions or (ii) if incomes (and other regressors) changed at
the same rate for all individuals and over all spatial units. In f act, neither is the
case. The literature suggests that the income elasticity of housing demand varies
by income level; furthermore the income distribution in this country has
changed considerably since the late seventies. Both violate standard aggregation
conditions so that aggregation from micro models of household behaviour to
aggregate models of housing demand is not straightf orward.
At the aggregate level, however, there is empirical evidence that housing
demand changed during the nineties. This appears most clearly in estimates of
393
Scottish Journal of Political Economy, Vol. 45, No. 4, September 1998
© Scottish Economic Society 1998. Published by Blackwell Publishers Ltd, 108 Cowley Road, Oxf ord OX4 1JF, UK and
350 Main Street, Malden, MA 02148, USA
ON THE AGGREGATE HOUSING MARKET
IMPLICATIONS OF LABOUR MARKET
CHANGE
Geoffrey Meen and Mark Andrew*
*University of Reading
the coefficients of house price equations which, under some conditions, may be
interpreted as inverted owner-occupier housing demand functions. Although
house price equations exhibited considerable coefficient stability through the
seventies and eighties, major structural changes to the coefficients appear to
have occurred from around 1991 when prices began to f all in nominal
terms– the first sustained fall since the fifties. Typically equations have heavily
overpredicted prices and their tracking record has not improved with the partial
recovery in prices since the end of 1996. There are a number of possible
explanations in terms of f actors omitted from standard models, of which the
following are most commonly cited:
(i) the failure is evidence of a negative speculative bubble;
(ii) the failure represents changes to labour market behaviour;
(iii) the errors at the aggregate level represent a form of aggregation bias, both
over individuals and over space;
(iv) attitudes towards risk have changed;
(v) the errors reflect changes in tax policy.
In this paper we concentrate on (ii) although , as we shall see, (ii) and (iii) are
not independent of each other. However, econometrically, discriminating
between the different hypotheses is not straightforward because the number of
periods of predictive f ailure against which to assess the evidence is still limited.
Nevertheless, since the policy implications are important, we feel it worthwhile
to make a first attempt at analysing the evidence. In terms of the f uture outlook
for demand and prices (i ) has very different implications from (ii ) or (iii). The
latter look for explanations in terms of changes to ‘fundamentals’ and might
suggest that demand will be permanently lower whereas (i) might indicate that
the capital losses of the early nineties were purely temporary and a catch-up in
the future might be in prospect.
II STRUCTURAL CHANGE AND THE NATURE OF LABOUR AND CREDIT
MARKETCONSTRAINTS
Since the early eighties, mortgage and labour markets have undergone major
structural changes. But income and the cost and availability of credit are central
variables in most housing demand functions and, therefore, the interactions
between housing, labour and credit markets and the constraints which each face
are key issues.
Consider the following lif e-cycle model which, with one addition, lies behind
most modern work on housing demand. In the standard model, there are two
goods, housing services and a composite consumption good ( C). Assuming that
the flow of housing services is directly proportional to the housing stock (H)
and, given a real discount rate ( β), lifetime utility is described by equation (1).
c
0
∞ −
eβtµ(H(t), C(t), L(t))dt.
(1)
The budget constraint is given by (2) and technical constraints (3) and (4)
394 GEOFFREY MEEN AND MARK ANDREW
© Scottish Economic Society 1998
To continue reading
Request your trial