Business investment is one of the main components in aggregate demand and net investment, in particular, contributes directly to the capital accumulation process and the supply-side development. As a consequence, the study of business investment is crucial for a thorough understanding of macroeconomic events, especially during the aftermath of the Great Recession when most advanced economies have been locked up in a slow and disappointing recovery era. This research is thus motivated to look for approaches used in analyzing business investment behaviour, thereby exploring alternative explanations for overall fluctuations in real US data.

Thi graph 1
This study focuses on the US business investment as it is the origin of subprime mortgage default, triggering financial meltdown and the subsequent global recession. Most data used in the analysis are obtained from the Bureau of Economic Analysis while other specific indices are extracted from professional computations. The methodology is based on specifications for the accelerator model and q-model described in Chirinko’s paper (1993) on “Business Investment Spending: Modelling Strategies, Empirical Results, and Policy Implications”. Data analysis is conducted using time-series regression on US quarterly data from 1999 to 2015 with the Newey-West standard error correction.

The selection of these specifications is derived from a consideration of actual data. A simple graph plotting the logs of real US business investment depicts a sharp fall during the financial crisis period. Net investment, on the other hand, is still currently below the pre-crisis level, resulting in a lower rate of capital stock growth at 1.5% annually – only half of the pre-crisis trend. Thus, the timing of business investment shortfall suggests that weak economic activity or output slowdown may be the dominant factor. This rather intuitive argument is, indeed, supported by both theory and empirical consensus about the pure accelerator mechanism in which real output changes significantly affect investment flows. One empirical difficulty arises from the reverse impact of investment on contemporaneous output changes, thus leading to upward-biased estimates on output changes. Because the usual OLS regression yields an insignificant coefficient estimate of contemporaneous output change, the paper chooses to follow conventional – ad hoc method of dropping that variable to avoid reverse causality. The result implies an approximately three constant-dollar increase in investment spending for each constant-dollar increase in output. This significant impact is further consolidated by comparing against the IMF study, which uses instrumental variable across advanced economies’ database and finds a strong positive effect of cumulative output changes on investment growth.

The analysis is then extended further to account for q-model widely used in economic textbooks. “q-theory” is a forward-looking model, which was developed by the economist James Tobin to capture investment behaviour in a single, micro-founded q variable defined as the ratio of marginal return from an additional unit of investment over its marginal cost. Hence, profit maximizing firms will adjust their investments to the optimal level at which marginal revenue equals marginal cost. The paper uses “q” data computed from Smithers&Co and finds small but statistically significant impact on investment decisions. The research also finds a significant, albeit economically small, impact of policy uncertainty on hindering aggregate investment incentive. This emphasis on uncertainty as a potential determinant may need to be developed further by discovering its impacts in more detailed firm-level analysis.

Overall, this paper has provided a clearer understanding of business investment behaviour since the onset of the financial crisis. Given the sharp slowdown in economic activity – the biggest recession in modern era after the Great Depression, a large investment shortfall is unusual from its historical relationship with real output changes. From the prediction of the accelerator model, policies aiming at maintaining strong and stable output may help stimulate investment in the future. Given the distributed lags in the q-model, investment incentives could rise if stock markets remains buoyant over a considerable period of time. Last but not least, policy uncertainty should also be reduced by clear plans and effective communication to markets. Business investment future, after all, may not be as pessimistic as we have imagined if the US economy continues to recover strongly and policies are supportive in reviving the “animal spirit”.
Thi Tran

Full Paper at:

A Study of Business Investment in the US Post-Crisis