Robert S. Pindyck, January 1990

The explanation of aggregate and sectoral investment behavior has been one of the less successful endeavors in empirical economics. Existing econometric models have had little success in explaining or predicting investment spending. This may be because most such models fail to account for the irreversibility of most investment spending. With irreversibility, changes in the riskiness of future cash flows or interest rates should in theory dramatically affect the decision to invest - more so than, say, a change in the levels of interest rates. Here I survey some of the empirical support for this proposition, and discuss the implications for investment modelling.