Business cycle accounting (BCA) is a method to investigate the important sources of business cycles. The procedure has two parts. The first part involves measuring distortions using a prototype model with time-varying wedges-which resemble aggregate productivity, labor and investment taxes, and government consumption-such that the prototype model exactly accounts for the observed data. The second part investigates the importance of each wedge in business cycles through counterfactual simulations. BCA has become a popular method of business cycle analysis and has been applied to many countries.
For the theoretical justification of BCA, Chari, Kehoe, and McGrattan (2007a) (hereafter,CKM) show what they called the "equivalence results." A model with frictions, which they called a "detailed model," is equivalent to (covered by) the prototype model if the allocation realized in the detailed model is replicated by the prototype model. For example, a sticky wage model is equivalent to a prototype model with labor wedge. The equivalence holds through an appropriate adjustment of wedges. CKM show the equivalence results without restricting the classes of the stochastic process for wedges.
In many application of BCA, the VAR(1) specification is employed as the stochastic process of wedges. However, recent papers by Baurleand Burren (2007) and Nutahara andInaba (2008) show that the equivalence results do not hold in many models if the stochastic process of the wedges of the prototype model is VAR(1). Then, there might be misspecifications of the stochastic process of the wedges in the applications of BCA. The main purpose of the present paper is to quantitatively investigate the significance of the distortions in the measurement of wedges and counterfactual simulations caused by the misspecification of the stochastic process of wedges.
In this paper, in order to assess the empirical usefulness of BCA, we apply BCA to a medium-scale dynamic stochastic general equilibrium (DSGE) model that is not covered by the prototype model but widely used for policy analyses. Since we knowa true data generating process, we can compute the true wedges that are consistent with ourDSGEeconomy. Our economy is based on that of Smetsand Wouters (2007) without investment-specific technology shocks. We find that the measured wedges capture the properties of the true wedges almost correctly. We also apply BCA to the economy with investment-specific technology shocks. In this case, the performance of BCA is a little worse. However, the business cycle implications of wedges are captured almost correctly. Therefore, BCA is empirically useful.
The rest of this paper is as follows. Section 2 introduces the prototype economy for BCA andourdetailed economy. Section 3 presents our main results. We also apply BCA to our detailed economy in this section. In Section 4, we apply BCA to the economy with investment-specific technology shocks. Section 5 contains the concluding remarks.
Download
An Application of Business Cycle Accounting with Misspecified Wedges
