International Competition and Inflation: A New Keynesian Perspective. Luca Guerrieri, Chris Gust, David López-Salido. Federal Reserve Board.

Transcription

1 International Competition and Inflation: A New Keynesian Perspective Luca Guerrieri, Chris Gust, David López-Salido Federal Reserve Board June 28 1

2 The Debate: How important are foreign factors for domestic inflation? The Protagonists: 1. The Global Enthusiasts: Rogoff (23), Borio and Filardo (26), Fisher (27) 2. The Moderates: Kohn (26), Yellen (26), Ihrig et. at. (27) 3. The Global Skeptics: Mankiw (26), Ball (26) Ball (26): but... Research is needed on how and why relative prices influence inflation. And this work must improve on current techniques for estimating these effects. 2

3 Our Question In spirit of Dornbusch and Fischer (1985), we develop a model in which variations in desired markups arise in response to foreign competition. In an open economy, a reduction in the prices of foreign competitors can induce domestic firms to lower their desired markups and prices. Estimate model to quantify the importance of this foreign competitiveness channel for domestic price inflation. Our focus is on how the domestic price set by a U.S firm is affected by foreign competitors. 3

4 Our Approach To model variations in desired markups, we use a Kimball aggregator. Developed Modified New Keynesian Phillips curve for domestic price inflation (not CPI inflation). Show that domestic inflation depends not only real marginal cost but on the ratio of foreign (import) prices to domestic prices. Show how open economy can be useful in separately identifying variation in the markup coming from nominal rigidities and variation from real rigidities. 4

5 Our Empirical Approach Focus on inflation dynamics using limited information approach: GMM-IV using data from In spirit of Fukac and Pagan (27), choose limited information estimation to guard against model misspecification. Which Inflation Measure? Traded goods price is the appropriate measure. Why? 1. Theoretically consistent. 2. The main candidate to evaluate the effects of foreign competition. 5

6 Model Features Final Goods Producers 1. Produce final tradable good, A t, according to a Kimball (1995) aggregator. 2. Demand A Dt (i) from domestic good producers and A Mt (i) from foreign producers. Intermediate Goods Producers 1. Set price in domestic market subject to Calvo contracts with partial indexation to past inflation. 6

7 Percent Demand Curve NCES CES log(p (i)/p ) D D 4 Desired Markup Percent log(p (i)/p ) D D 7

8 Inflation Dynamics: Identifying Nominal and Real Rigidities [ ˆπ t δ Dˆπ t 1 = βe t [ˆπ t+1 δ Dˆπ t ]+κ D (1 Ψ)ŝ t + Ψω ɛ ] A ɛ ˆp Mt + ϕ γ t, - Real Rigidity from VES (Ψ): identify through variations in ˆp Mt - Nominal Rigidity Index: κ = (1 βθ)(1 θ) θ - Firm Specific Capital: κ D = κγ D γ D = 1 (1+ 1 α α ɛ(1 Ψ)) - Special Cases: CES: Ψ = Closed economy model: ω = No Indexation: δ D = 8

9 Empirical Methodology: Present-value Approach - Forward Solution: π t = δ D π t 1 + κ D k= [ β k E t (1 Ψ)ŝ t+k + Ψω ɛ ] A ɛ ˆp Mt+k + ϕˆγ t+k. - Let X t be a vector of variables that includes s t and p Mt. Write a VAR (companion form) as: X t = AX t 1 + u t, - We use VAR to forecast s t and p Mt. 9

10 Baseline forecast-model (restricted VAR): [ 1 Ψ ˆπ t = δ Dˆπ t 1 + κ D ŝ t + ω ɛ ] A Ψ(1 + βφ 2 L) 1 βρ s ɛ 1 βφ 1 β 2 Mt φ 2ˆp ŝ t = ρ s ŝ t 1 + u st ˆp Mt = φ 1ˆp Mt 1 + φ 2ˆp Mt 2 + u Mt + ɛ πt - ɛ πt, iid markup shock, generates endogeneity problem. - We jointly estimate this system using GMM allowing for correlation between ɛ πt, u st, u Mt - Approach avoids breaking cross-equation restrictions implied by a DGE model. - Robustness: Alternative VAR specifications and moment conditions. 1

11 Estimates of Open Economy Calvo Model VES with VES without CES with CES without indexation indexation indexation indexation θ (.3) (.3) (.3) (.4) Ψ (.11) (.12) δ D.1.34 (.8) (.8) Q-Statistic(4) [.42] [.14] [.2] [.] g min

12 Interpreting the Estimate of ψ When ψ =.73 - If price of good i is 2%, demand 14%, desired markup 6 PPt. - If price of good i is 5%, demand 45%, desired markup 12 PPt. - If import prices 1%, domestic desired markup 2 PPt. CKM use ψ =.98 - If price of good i is 2%, demand 78%. If price of good i is 2.3%, demand 1%. 12

13 Actual and Predicted Inflation 4 Observed Goods Price Inflation Predicted Inflation (VES, no indexation) Predicted Inflation (CES, no indexation)

14 Comparison of Benchmark Estimates and Calibrated Demand Curves in the Literature ɛ µ ɛ(i) 1 p D (i) ɛ Ψ Benchmark Estimates Chari et al. (2) Coenen et al. (24) Eichenbaum and Fisher (27) Dossche et al. (27) Dotsey and King (25) Gust et al. (26) Bouakez (25)

15 Monte Carlo Sampling Distribution of Estimates (ψ =.67).25 Indexation, δ D True Value =.16 True value of ψ =.67 Calvo Probability, θ True Value =.75.2 VES Specification mean =.17 std =.1 CES Specification mean =.26 std = CES Specification mean =.79 std = VES Specification mean =.71 std =

16 True value of ψ = Indexation, δ D.5.4 Calvo Probability, θ True Value =.16 True Value =.75 VES Specification mean =.17 std =.1 CES Specification mean =.79 std =.2.1 CES Specification mean =.26 std = VES Specification mean =.71 std = True value of ψ = Indexation, δ D True Value =.16 VES Specification mean =.16 std =.9 CES Specification mean =.47 std = True Value =.67 Calvo Probability, θ CES Specification mean =.8 std =.4 VES Specification mean =.59 std =

17 Alternative Estimation Approaches Single-Equation Estimation: ˆπ t δ Dˆπ t 1 = βe t [ˆπ t+1 δ Dˆπ t ]+κ D [ (1 Ψ)ŝ t + Ψω ɛ A ɛ ˆp Mt ] +ɛ πt, Closed-Form Estimation: [ 1 Ψ ˆπ t = δ Dˆπ t 1 + κ D ŝ t + ω ɛ ] A Ψ(1 + βφ 2 L) 1 βρ s ɛ 1 βφ 1 β 2 Mt φ 2ˆp ŝ t = ρ s ŝ t 1 + u st ˆp Mt = φ 1ˆp Mt 1 + φ 2ˆp Mt 2 + u Mt + ɛ πt 17

18 Small-Sample Properties of GMM Closed Form Estimates mean =.76 std =.17 Degree of Real Rigidity, ψ True Value = Indexation, δ p.3.2 True Value =.16 Closed Form Estimates mean =.17 std =.1 Single Equation Estimates mean =.2 std = Single Equation Estimates mean =.73 std =.9 Calvo Probability, θ True Value =.75 Closed Form Estimates mean =.71 std =

19 Conclusions Foreign competition plays an important role in accounting for inflation dynamics. Neglecting foreign competition: 1. 1 ppt. overestimate of goods price inflation over the period Changes in foreign competition account for 1 3 domestic goods price inflation. of volatility of Estimates provide support for calibrated values of VES demand curves used in literature. Future Work: Extend analysis to aggregate inflation dynamics. 19