ON LAGGED ADJUSTMENT, PERMANENT INCOME, EXPECTATIONS FORMATION AND THE DEMAND FOR MONEY*
| Author | R. MacDonald,D. A. Peel |
| DOI | http://doi.org/10.1111/j.1468-0084.1986.mp48001004.x |
| Date | 01 February 1986 |
| Published date | 01 February 1986 |
OXFORD BULLETIN OF ECONOMICS AND STATISTICS, 48, 1(1986)
0305-9049 $3.00
ON LAGGED ADJUSTMENT, PERMANENT
iNCOME, EXPECTATIONS FORMATION AND
THE DEMAND FOR MONEY
R. MacDonald and D. A. Peel
INTRODUCTION
A large number of researchers have estimated, or proposed in theoreti-
cal models, demand for money functions in which permanent income is
specified as one argument of the function (see, e.g. Artis and Lewis,
1976; Brunner etal., 1982;Feige, 1967;Fisher, 1968;Friedman, 1959,
Laidler, 1966; Spinelli, 1978). Permanent income, following Friedman's
(1957) work on the consumption function, is determined empirically as
a weighted average of current and past real incomes. Defining permanent
income as a geometrically declining lag of current and past incomes
(Koyck lag) offers one interpretation of the long adjustment lags which
Chow (1966) and Goldfield (1973) found (modelled econometrically
by including a lagged dependent variable) when using measured income
as the scale variable in their estimates of the demand function for
money. Indeed, the paper cited by Fiege (1967) was designed to dis-
criminate between alternative interpretations of a lagged dependent
variable in the demand for money function and supported the proposi-
tion that adjustment lags are apparently negligible once proper account
is taken of the role of adaptive expectations in determining permanent
income.
It follows from the innovative paper by Hall (1978) and the further
analysis by Flavin (1981) that if households hold rational expectations
of the path of future incomes then under certain circumstances, notably
that the real rate of interest is constant, households are not liquidity
constrained and the households' utility function is separable in its
arguments, planned consumer expenditure and permanent income will
follow first order autoregressions.
A large number of empirical studies have been conducted to test the
validity of the Hall hypothesis as applied to consumer behaviour (see
e.g. Bilson, 1980; Davidson and Hendry, 1981; Flavin, 1981; Muellbauer,
1983). The empirical results are perhaps suggestive that the simple Hall
hypothesis gives a good description of the time series representation of
consumer expenditure but that consumption is sometimes not ortho-
* We are grateful to the editors for helpful comments on an earlier draft.
61
To continue reading
Request your trial