Aggregate Investment in South Africa: A Model with Implications for Political Reform

Published date01 August 1997
Date01 August 1997
DOIhttp://doi.org/10.1111/1468-0084.00070
AuthorDavid Fielding
OXFORD BULLETIN OF ECONOMICS AND STATISTICS, 59, 3 (1997)
0305-9049
AGGREGATE INVESTMENT IN SOUTH
AFRICA: A MODEL WITH IMPLICATIONS FOR
POLITICAL REFORM
David Fielding*
I. INTRODUCTION
A number of recent papers have attempted to identify the main factors
determining the evolution of fixed capital expenditure in South Africa,
and so predict the likely consequences of the recent political liberaliza-
tion on investment and growth (see for example Bleaney, 1994). This
paper is intended to provide more evidence on the determinants of
investment by estimating the parameters of a structural macroeconomic
model.
Political liberalization might have a number of consequences for the
macroeconomy. On the one hand, the removal of economic sanctions may
reduce import costs and increase export prices, leading to an improved
terms of trade. On the other, democratic government may lead to higher
expenditure on public goods, increasing aggregate demand. In addition,
political stability may reduce the perceived risks of investment in the
South African economy. In many respects, such changes resemble those
which accompany external shocks resulting from international commodity
price movements or the discovery of a natural resource. Our theoretical
model therefore bears a strong resemblance to those which have been
constructed to analyse the impact of external shocks in developing coun-
tries (for example, Devarajan and de Melo, 1987; Pinot, 1987; Devarajan,
1991; Bevan et al., 1992). The relatively high quality and frequency of
South African data permit us to go further than these papers by estimat-
ing the parameters of the model.
In the next section, we sketch out a theoretical framework, introducing
a system of equations which will provide a context for interpreting the
results of the econometric model estimated in the following section.
Finally, we outline the policy implications of our results.
*I am grateful to participants at seminars at Oxford and Monash and to two anonymous
referees for comments and suggestions on previous versions of this paper. The usual
disclaimers apply. 349
© Blackwell Publishers Ltd, 1997. Published by Blackwell Publishers, 108 Cowley Road, Oxford
OX4 1JF, UK & 350 Main Street, Malden, MA 02148, USA.
II. THE THEORETICAL FRAMEWORK
Existing empirical work on aggregate private investment functions in
LDCs, and the theory on which they might be based, are discussed by
Serv´en and Solimano (1993) and Rama (1993). In a standard neoclassical
model the representative firm maximizes the rise in its market value
(operating profits plus capital gains less capital adjustment costs) over its
time horizon (t,t+1) subject to a production function and an equation of
motion for the capital stock (Kt):
It\d·Kt+DKt+1(1)
where Itis gross investment and dthe rate of depreciation. If output
depends on the firm’s own capital stock, publicly owned capital (KG
t) and
employment, then one can derive an equation for the firm’s planned
capital stock with competitive goods and factor markets:
Kt+1\g(E[Ct+1], E[Wt+1], E[KG
t+1], Kt) (2)
where Wis the real wage and Cthe real user cost of capital. The user
cost of capital is equal to Q·[rµp+d], where Qis the real price of capital
goods and rµpthe real interest rate (ris the domestic nominal interest
rate and pthe rate of inflation). E[] is an expectations operator, and the
presence of Ktreflects the existence of adjustment costs. Gross investment
is given by the equation of motion:
It\d·Kt+DKt+1\f(Ct, E[Ct+1], Wt, E[Wt+1], KG
t, E[KG
t+1], Ktµ1) (3)
If expected future KG, W, Qand [rµp] depend on present and past levels
(up to an horizon, n) and if capital stocks (both public and private) can
be expressed as a function of past investment levels (with weights depend-
ing on the rate of depreciation), then one can write:
It\h(Qt,..., tµn, [rµp]t,..., tµn, Wt,...,tµn, IG
t,..., tµn, Itµ1,...,tµn) (4)
Such a model can be extended or modified in various ways: with imper-
fectly competitive goods markets the expected future rate of growth of
aggregate demand, and hence past and present income levels, will affect
present investment, which is an excuse for including GDP in the regres-
sion equation and estimating an accelerator model. Under credit ration-
ing investment may no longer be demand-determined, but depend instead
on the quantity of credit supplied by financial markets.
Rama summarizes the alternative model specifications which have been
estimated on LDC data sets and notes their differing degrees of success.
Most models manage to find a significant negative coefficient on some
proxy for C; some include a proxy for KG, but not all find a positive
coefficient, even when controlling for credit market crowding out effects.1
1Contrast Gupta (1984) with Musalem (1989).
© Blackwell Publishers 1997
350 BULLETIN

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