RDP 8709: A Note on Aggregate Investment in Australia 3. Australian Studies

Warwick J. McKibbin and Eric S. Siegloff

October 1987

In the early 1970s, empirical work on investment behaviour in Australia was strongly influenced by standard neo-classical theory.[7] Although this approach was theoretically more appealing than the popular accelerator models of the 1960s, it proved to be a poor explanation of investment behaviour in Australia. A broader study by Higgins et.al. (1976) sought to improve empirical understanding of investment by employing a more refined neo-classical model as well as a variety of different models. As in the earlier studies, Higgins et.al. found that the neo-classical model performed poorly; furthermore, they found that it was noticeably inferior to simpler investment functions based on the accelerator and securities value models[8]. Subsequent development of the neo-classical model has been taken up by Australian macro-econometric models such as the RBII[9] model, while the AMPS[10] and NIF88[11] models claim to use a variant of Tobin's q theory to determine business fixed investment. In a recent study, Rider (1987) tests the q theory using a tax-adjusted q series. However, the results do not support the q theory.

Footnotes

See Mackrell et.al. (1971) and McLaren (1971) for example. [7]

For a useful survey article of the econometric contributions of the 1970s see Hawkins (1979). See also Stegman (1982) for a novel approach to estimating an accelerator-type investment function, and Kohli and Ryan (1986) for a more refined neo-classical model. [8]

See Edey et.al. (1987). [9]

See Murphy et.al. (1986). [10]

See Simes (1987). [11]