November 2001 |
The 2002 Frisch Medal "Explaining Investment Dynamics in U.S. Manufacturing: A Generalized (S,s) Approach." Econometrica 67(4), July 1999, 783-826. The authors develop a model and estimator that enables them to bridge the gap between micro data on the investment of firms, and data on aggregate investment. Firm level data on investment indicate that investment projects are lumpy. To rationalize this phenomenon, the authors adopt a micro model with random, fixed, adjustment costs. Their approach is shown to generate an (S,s) investment rule and can be aggregated in a way that allows estimation of its parameters from either aggregate, or firm level, data. The framework allows the history of micro and aggregate shocks to generate bunching of firms, and therefore can replicate the sharp increases in investment activity seen during economic upturns. Using a panel of data on two digit industries, they show that their underlying micro model outperforms standard time series investment models with respect to both the fit of the investment relationship and the (out of sample) predictions of future investment. This paper develops techniques that combine recent advances in the availability of data, in econometrics, in theory, and in computing power to provide a better understanding of an important aggregate. |