The to Monetary Policy Analysis: Lessons and New Directions Jordi Galí CREI and U. Pompeu Fabra ice of Monetary Policy Today" October 4, 2007
The New Keynesian Paradigm: Key Elements Dynamic stochastic general equilibrium framework Monopolistic competition Nominal rigidities =) forward-looking price-setting Emphasis on endogenous component of monetary policy ("MP rules") Evaluation of alternative policy rules =) welfare-based =) overcomes the Lucas critique Flexible tool: open economy, imperfect information and learning, state dependent pricing, unemployment, credit frictions,... But are there any new insights or lessons? What are the remaining challenges?
The Basic New Keynesian Model: Key Blocks New Keynesian Phillips Curve π t = β E t fπ t+1 g + κ x t + u t where Dynamic IS Equation x t y t y n t x t = 1 σ (i t E t fπ t+1 g rt n ) + E t fx t+1 g Monetary Policy Rule Example: i t = ρ + φ π π t + φ y by t + v t
What Are the Main Novel Insights? The costs of in ation and the bene ts of price stability The role of expectations and the gains from commitment Importance of "natural" levels of output and interest rates as policy benchmarks The bene ts of a credible anti-in ationary policy
The Costs of In ation and the Bene ts of Price Stability In ation as an indicator of an ine cient activity level π t = β E t fπ t+1 g + κ (y t y n t ) π t = 0 all t =) y t = y n t all t In ation as a source of ine cient allocation across rms/sectors Implied optimal policy =) zero in ation at all times Quali cations: - several factors call for positive average in ation - existence of short-run tradeo s π t = β E t fπ t+1 g + κ (y t y n t ) + u t =) positive target for in ation, to be attained in the medium term
The Role of Expectations and the Gains from Commitment Assume fu t g and fr n t g are i.i.d.: π t = κ x t + κ k=1 β k E t fx t+k g + u t x t = 1 σ i t 1 σ k=1 E t fi t+k g + 1 σ E t fπ t+1+k gg + 1 σ r t n k=0 =) Commitment to future policy actions improves current tradeo s =) Importance of expectations management
The Importance of Natural Levels as Policy Benchmarks A Brief Detour: Traditional vs Model-based Output Gap measures Traditional Output Gap Measure by t = y t f (t) f (t) smooth trend (e.g. projection on time) Output Gap inthe NK Model x t y t y n t y n t exible price output Empirical Evidence (I): Galí 03: using x t _ log(labor income share) Empirical Evidence (II): Edge, Kiley and Laforte 07 using DSGE model
Source: Galí (2003)
Source: Edge, Kiley, and Laforte (2007)
Policy Implications: Pitfalls of Output-Based Rules Optimal policy problem (under discretion) min α xt 2 + π 2 t subject to π t = β E t fπ t+1 g + κ x t + u t Implied optimality condition ("targeting rule") x t = κ α π t Implied volatility: σ(π t ) = αψ σ(u t ) ; σ(x t ) = κψ σ(u t ) Implementation with output-based rule by t = κ α π t Implied volatility: σ(π t ) = αψ σ(u t ) + χ σ(by n t ) ; σ(x t ) = κψ σ(u t ) + ϑ σ(by n t )
Taylor-type interest rate rule i t = ρ + φ π π t + φ y by t Equivalent representation i t = ρ + φ π π t + φ y x t + v t where v t φ y by n t =) High φ y destabilizes both in ation and the output gap =) Desirability of focus on in ation
The Bene ts of a Credible Anti-In ationary Policy Interest Rate Rule Perceived Interest Rate Rule i t = ρ + φ π π t i t = φ π (1 δ) π t + v t δ 0 : "credibility gap" Equilibrium: - solve given an arbitrary fv t g - impose v t = δφ π π t Simulation: Volatility Frontiers (Fig.)
Policy Credibility and Macroeconomic Volatility 3.5 3 δ=0 δ=0.3 2.5 Inflation Volatility 2 1.5 1 0.5 0 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5 Output Gap Volatility
Sources of Policy Tradeo s: A Challenge The New Keynesian Phillips Curve π t = β E t fπ t+1 g + κ (y t y n t ) Remark: similar when sticky wages and prices Underlying Real Model Combining both equations: y e t y n t = δ π t = β E t fπ t+1 g + κ (y t y e t + δ) =) no trade-o between stabilization of in ation and stabilization of the welfare-relevant output gap Assessment - Too strong: no concern for output or employment losses - At odds with central banks practice Solutions in the existing literature: ad-hoc or limited relevance
Sources of Policy Tradeo s: An Alternative View Real imperfections =) ine cient responses to shocks, even in the absence of nominal rigidities ( Blanchard-Galí JMCB 07) =) y e t y n t endogenous Implied in ation equation: π t = β E t fπ t+1 g + κ (y t y e t ) + u t where u t κ (y e t y n t ) - fu t g responds endogenously to all kinds of (real) shocks - strict in ation targeting no longer optimal
Sources of Policy Tradeo s: An Alternative View Examples of Possible Applications Labor market frictions (Blanchard-Galí,...): - makes room for slow adjustment of real wages to labor market conditions - strict in ation targeting! (ine ciently) large unemployment uctuations - optimal policy: partial accommodation of in ationary pressures Credit market imperfections (BGG,...) - external nance premium decreasing in net worth - ine ciently large uctuations in response to shocks - room for countercyclical monetary policy, short run uctuations in in ation...
What the has Delivered... novel insights a useful framework for organizing our thinking a exible tool useful for quantitative analysis at CBs?