Machine Replacement and the Business Cycle: Lumps and Bumps

Abstract

This paper explores investment fluctuations due to discrete changes in a plant's capital stock. The resulting aggregate investment dynamics are surprisingly rich, reflecting the interaction between a replacement cycle, the cross-sectional distribution of the age of the capital stock, and an aggregate shock. Using plant-level data, lumpy investment is procyclical and more likely for older capital. Further, the predicted path of aggregate investment that neglects vintage effects tracks actual aggregate investment reasonably well. However, ignoring fluctuations in the cross-sectional distribution of investment vintages can yield predictable nontrivial errors in forecasting changes in aggregate investment.

Re-Investment and the Survival of Foreign-Owned Plants

“…Potential omissions of data on re-investments cannot be known, but the extent to which this is an issue it seems more likely to occur in the less well-known smaller plants (which would certainly be the case if the census data were used), so that as a robustness check we later disaggregate the results by plant size. 3 Nevertheless, in the extensive checking of the data nothing came to light to make us question its reliability, and we believe it gives a comprehensive account of foreign-owned investment in a UK region, so that it is therefore representative As a further point, the paper's focus on projects treats investment as intermittent and lumpy, but this is consistent with the literature that analyses investment episodes (e.g., Cooper et al, 1999). These other studies tend to use census data, but even here there is an element of subjectivity as an investment episode is defined if the investment rate exceeds some multiple of the median rate, e.g.…”

“…13 Further, the lognormal imposes negative duration dependence, but this is rejected by the log-logistic function in favor of an increasing and then decreasing hazard rate. Like Cooper et al (1999), it suggests an episode of investment exhibits positive duration dependence, although unlike this other study it becomes less likely after a period of time. This is plausible and it is supported by the empirical hazard rates in Table 4.…”

Strategic Investments in the Pulp and Paper Industry: A Count Data Regression Analysis

“…Cooper et al (1999) use duration regressions in analysing the time since the previous investment spike. 12 …”

“…Cooper et al (1999) de…ne "investment spikes" as years when the gross investment rate exceeds 20 percent. In the present study, an "investment" is de…ned as the installation of a new processing unit, i.e., a new paper machine or a new pulp line.…”

Non-convexities in the adjustment of different capital inputs: A firm-level investigation

“…2 Non-convexities alter investment dynamics both at firm and aggregate level, making optimal investment a non-linear function of its fundamentals. For the U.S., the empirical evidence on non-convexities has been growing starting from the explorative work by Doms and Dunne (1998) up to the investigation by Cooper, Haltiwanger, and Power (1999) and to the "gap methodology" approach proposed by Caballero, Engel, and Haltiwanger (1995) and Caballero and Engel (1999). 3 While these papers typically use plant level data but do not estimate structural investment equations, more recent contributions adapt Tobin's Q models to take into account non-convex costs and estimate their extent using firm-level panel data.…”