RDP 2023-09: Does Monetary Policy Affect Non-mining Business Investment in Australia? Evidence from BLADE 5. Effect on Expectations with Survey Data

We now turn to the expectations data to better understand how monetary policy affects investment expectations. For the regressions, as well as looking at the effect of the shocks on investment at horizon h , we also look at the effect on expected investment take at time h for the future. More precisely we estimate regressions of the following form:

where k represents the period covered by the short- or long-term expectation measure (e.g. the next quarter or full financial year beginning in a quarter's time). In general, if expectations are rational and forward looking they should be based on expected future outcomes. So if contractionary monetary policy today is expected to weaken outcomes in one year's time, we should expect to see expectations of the investment that will be undertaken next year decline immediately, ahead of any response from actual investment. However, this is not the case.

Overall, it appears that monetary policy shocks affect both actual and expected investment. Somewhat surprisingly though, both appear to respond after around one to two years, rather than expectations falling ahead of actual investment. This is evident in the aggregate CAPEX data (Figure 9), as well as with the microdata, where the extensive margin appears to be the more important margin of adjustment (Figure 10). It also seems to be broadly consistent across smaller and larger firms, though there is some tentative evidence that medium and larger firms adjust their expectations ahead of their actual investment (Figure C8). Taken together, these findings suggest that firms' expectations are largely based on current investment and conditions, rather than reflecting future conditions and how interest rates will affect those conditions.

Whether or not this finding has implications for real economic outcomes will depend on whether these lower expectations feed through to lower future investment. If so, this could be a mechanism through which economic shocks are perpetuated, as occurs in various models with backward-looking expectations (e.g. Brassil, Gibbs and Ryan 2023). If not, the implications may mainly relate to the usefulness of these expectations for economic forecasting. The results would suggest that expectations may lag the cycle somewhat, with realisation ratios tending to be higher around the trough of the cycle, and lower towards the peak. This is consistent with the findings in Cassidy et al (2012). Similarly, these findings suggest that expectations-based investment forecasts may be slow to capture the effects of monetary policy.

Overall, these results suggest that the effects of monetary policy on investment are delayed, hard to predict for firms, and that firms are largely responding to current conditions. This combination of outcomes may provide some insights into which channels of pass through are most important. For example, changes in the marginal cost of capital should likely flow through to firms quickly. So if this was the most important channel, firms should quickly realise that certain projects are not profitable and adjust investment expectations accordingly, which is somewhat inconsistent with these findings. In contrast, firms will only experience the weaker demand stemming from contractionary policy after several periods and may only update their expectations then, which would be consistent with these findings. So the demand channel being particularly important would be consistent with these results. How these findings relate to the financial constraint channel is less clear. Changes in collateral values and therefore borrowing capacity should be evident immediately to firms. However, firms may only realise that they will face cash flow constraints once economic activity weakens, which would be consistent with the delayed response of expectations.