Recently, Smets and Wouters (2003) and Christiano Eichenbaum and Evans (2005) (CEE) have estimated empirical New Keynesian DSGE models to evaluate the role of shocks and frictions for the business cycle. These large scale DSGE models feature enough internal propagation mechanisms to fit the data reasonably well and they have become the workhorse to conduct monetary policy analysis. A robust conclusion of these studies is that labor market frictions are crucial in accounting for the behavior of output and inflation. Yet, with the exception of Trigari (2005) and Walsh (2005b), the current vintage of empirical New Keynesian models does not include the modern theory of unemployment due to Diamond, Mortensen and Pissarides. An empirically grounded labor market block is, however, crucial to uncover the sources of business cycle fluctuations and to understand the dynamics of marginal cost and inflation. Trigari and Walsh both find that labor search reduces the elasticity of marginal cost with respect to output and therefore helps to replicate the sluggish response of inflation to monetary shock without relying on an excessive degree of nominal price stickiness.
However, these two authors do not adress the "volatility" and the "persistence" puzzles recently highlighted by Shimer (2005) and Yashiv (2005). Shimer shows that, conditional on productivity shocks of plausible magnitude, the standard Search and Matching model can not replicate the observed high volatility of vacancies and unemployment. He also suggests that the model’s wage is too volatile and therefore annihilates most of firms’ incentives to hire new workers. Yashiv (2005) establishes the importance of convex hiring cost to match the high persistence of the vacancy and unemployment rates. The first task of this paper is therefore to update the labor market block of Trigari and Walsh’s models by following the recent advises of Shimer (2005) and Yashiv (2005). Then, this paper also aims at shedding new lights, using a DSGE approach, on some central empirical macroeconomic relationships that involve the unemployment rate such as the Phillips curve, the Okun’s Law and the Beveridge curve.
This paper introduces investment adjustment costs, variable capital utilization, convex hiring costs and real wage rigidity in a New Keynesian business cycle model with labor search and exogenous job destruction. Price stickiness is introduced by assuming that monopolistically competitive firms face quadratic price adjustment costs à la Rotemberg. One advantage of "Rotemberg pricing" over "Calvo pricing" is to avoid price dispersion and, thereby, issues related to firm-specific inputs. The model contains seven shocks. The neutral technology shock is permanent and implies that output, consumption, investment and real wage share a common stochastic trend. The remaining six shocks, investment-specific-technology, preference, cost push, wage bargaining, exogenous spending and monetary policy are all stationary.
To handle misspecification, I estimate the model with a limited information approach that focuses on the business cycle facts and consists in matching model with data spectra. The spectrum matching estimation technique was proposed by Watson (1993), implemented by Wen (1998) and formalized and extended by Diebold, Ohanian and Berkowitz (1998).
In the model, output is demand driven because of sticky prices, and labor is a predetermined, firm-specific, input because of search and matching frictions. Hence, in the aftermath of a demand shock, the only ways for firms to adjust output and to satisfy demand on impact is by varying the degree of capacity utilization or by affording a costly jump in prices. These features make capital utilization and the rental rate of capital excessively volatile at very high frequencies. The DSGE model being clearly misspecified at high frequencies, I downweight those frequencies in the estimation and aim at maximizing the agreement between the model and the data at low and, especially, at business cycle frequencies. I study the implications for parameter estimates, impulse responses and goodness of fit of two alternatives weighting schemes. Downweighting very high frequencies drastically helps recovering reasonable estimates and meaningful impulse responses to the various shocks.
Download
PDF Ebook An Estimated New-Keynesian Business Cycle Model with Equilibrium Unemployment
