It has long been recognized that inventory investment plays a large role in explaining fluctuations in real GDP, although it makes up only a small fraction of the latter. Blinder and Maccini (1991) document that in a typical recession in the U.S., the fall in inventory investment accounts for 87% of the decline in output despite being only one half of 1 percent of real GDP. A lot of research has been trying to explain how this seemingly insignificant component of GDP has such a disproportionate role in business cycle fluctuations. However, surprisingly few studies have focused on the conduct of monetary policy when firms can invest in inventories. In this paper we attempt to fill this gap by investigating how inventory investment affects the design of optimal monetary policy.
We employ the simple New Keynesian model which has become the benchmark for analyzing monetary policy from both a normative and a positive perspective. We introduce inventories into the model by assuming that the inventory stock facilitates sales, as suggested in Bils and Kahn (2000). We first establish that the dynamics, and therefore the monetary transmission mechanism, differ between the models with and without inventories for a given behavior of the monetary authority. Monetary policy is then endogenized by assuming that policymakers solve an optimal monetary policy problem.
We first compute the optimal Ramsey policy. A Ramsey planner maximizes the welfare of the agents in the economy taking into account the private sector’s optimality conditions. By doing so, the planner chooses a socially optimal allocation. While this does not necessarily bear any relationship to the typical conduct of monetary policymakers, it provides a useful benchmark. Subsequently, we study optimal policy when the planner is constrained to implement simple rules. That is, we specify a set of rules that let the policy instrument, the nominal interest rate, respond to target variables such as the inflation rate and output. The policymaker chooses the respective response coefficients that maximize welfare. Optimal rules of this kind may be preferable to Ramsey plans from an actual policy maker’s perspective since they can be operationalized and are easier to communicate to the public.
Our most interesting finding is that Ramsey-optimal monetary policy deviates from full inflation stabilization in our model with inventories. This stands in contrast to the standard New Keynesian model. In the latter framework, perfectly stable inflation is optimal since movements in prices represent deadweight costs to the economy. Introducing inventories modifies that basic calculus since holding inventories allows firms to smooth sales over time with concomitant effects on consumption. This change in the economy’s propagation mechanism can require, however, movements in labor input. Moreover, output and consumption need no longer coincide as firms can invest in inventory holdings, which is similar to capital in that it provides future consumption opportunities. Changes in prices serve as the equilibrating mechanism for the competing goals of reducing consumption volatility and of avoiding price adjustment costs. The inventory specification therefore contains something akin to an inflation-output trade-off. Consequently, the optimal policy no longer fully stabilizes inflation. The second important finding concerns the efficacy of implementing simple rules. Similar to most of the optimal policy literature, we show that simple rules can come exceedingly close to the socially optimal Ramsey policy in welfare terms.
Our paper relates to two literatures. First, the amount of research on optimal monetary policy in the New Keynesian framework is very large already, and we do not have much to contribute conceptually to the modeling of optimal policy. Schmitt-Grohé and Uribe (2007) is a recent important and comprehensive contribution. A main conclusion from this literature is that optimal monetary policy will choose to almost perfectly stabilize inflation. In environments with various nominal and real distortions, this policy prescription becomes slightly modified, but nevertheless perseveres. We thus contribute to the optimal policy literature by demonstrating that the results carry over to a framework with another, previously unconsidered modification to the basic framework in the form of inventories.
