RDP 8501: Neoclassical Theory and Australian Business Investment: A Reappraisal 1. Introduction

It has long been recognised that business investment is an important yet very volatile component of aggregate demand. As such it plays an important role in determining the level of real activity and the cyclical behaviour of the economy. Naturally, business investment is also crucial in shaping the growth path of the economy.

Because it is so volatile, business investment is difficult to model. Many theories of investment behaviour have been proposed in the literature, and have been applied with some success to Australian data. In recent years, however, business investment in Australia has eluded attempts to explain it. As shown in Figure 1, business investment has stagnated during most of the first half of the seventies. (It actually fell as a ratio to GDP.) It then fell sharply in the aftermath of the first oil shock; it shot up during 1980 and 1981, only to drop sharply again in 1982–83.

Existing models of investment behaviour have been at a loss to explain the ups and downs of Australian business investment. Flexible accelerator models have generally done best in tracking Australian investment, but their performance has deteriorated significantly recently.^[1] As for neoclassical models, they have performed poorly quite consistently in Australia.^[2] Moreoever, it is sometimes argued that the evidence of the 1970's does not support neoclassical investment theory.^[3] The seventies were characterised in Australia by a substantial increase in real wages (Figure 2) and in the relative rental price of labour. This should have led to an increase in the desired capital/labour ratio, and it should have triggered, the argument goes, an increase in investment. Yet it is the opposite that occurred as noted above.

This paper attempts to reconcile neoclassical theory with the facts. We argue that it is not neoclassical theory that is at fault, but rather the use to which it has been put. However, the standard neoclassical model may be too simple to explain the Australian facts, and we extend it in a number of directions. In particular an attempt is made to integrate the investment decision with other related decisions.

The paper proceeds as follows. In Section 2 we briefly review neoclassical inuestment theory, and an interpretation of the poor performance of the model is suggested. In Section 3 we construct an integrated model of investment behaviour that seems consistent with the facts. Empirical implementation of the model is undertaken in Section 4. Section 5 reports a number of simulation results, and Section 6 contains our conclusions.

Footnotes

See Higgins et al. (1976) for instance. [1]

See Higgins et al. (1976), Norton and Henderson (1972), and Hawkins (1979). [2]

Hawkins (1979). [3]