Globalization, Market Structure and Inflation Dynamics

Transcription

1 Discussion of Globalization, Market Structure and Inflation Dynamics by Sophie Guilloux-Nefussi Oleg Itskhoki Princeton University NBER Summer Institute ITM July / 10

2 Introduction Exciting paper! Fact: Inflation has become less responsive to output gaps interpreted as a flattening of the Phillips curve Can globalization flatten the Phillips curve? To answer this question, the paper brings together three literatures: 1 New Keynesian monetary model with sticky prices 2 Oligopolistic competition and incomplete pass-through 3 Heterogenous-firm trade model and globalization shock The answer is yes, if: large firms exhibit great strategic complementarities and have lower pass-through of the cost shocks globalization favors large firms (vs. competition!) 2 / 10

3 What are we talking about? New Keynesian Phillips Curve (NKPC): π t = βe t π t+1 + κ x t + ε t inflation π t = p t p t 1 output gap x t = γ(mc t p t ) 3 / 10

4 What are we talking about? New Keynesian Phillips Curve (NKPC): π t = βe t π t+1 + κ x t + ε t inflation π t = p t p t 1 output gap x t = γ(mc t p t ) slope of the NKPC: κ = λ 1 + Γ, where λ = (1 θ)(1 βθ) decreases in price stickiness θ θ Γ 0 is real rigidity ( strategic complementarities, round-about production, local input markets) 3 / 10

5 What are we talking about? New Keynesian Phillips Curve (NKPC): π t = βe t π t+1 + κ x t + ε t inflation π t = p t p t 1 output gap x t = γ(mc t p t ) slope of the NKPC: κ = λ 1 + Γ, where λ = (1 θ)(1 βθ) decreases in price stickiness θ θ Γ 0 is real rigidity ( strategic complementarities, round-about production, local input markets) Can globalization increase Γ and reduce κ? 3 / 10

6 ... or is there a Phillips curve? Flattening of the Phillips Curve Gopinath, Itskhoki and Neiman (IMFER 2012): Trade Prices and Global Trade Collapse of Diff exports Diff imports Non-diff imports 0.2 Non-diff exports :1 2007:7 2008:1 2008:7 2009:1 2009:7 4 / 10

7 ... or is there a Phillips curve? Flattening of the Phillips Curve Gopinath, Itskhoki and Neiman (IMFER 2012): Trade Prices and Global Trade Collapse of Diff exports Diff imports Non-diff imports 0.2 Non-diff exports :1 2007:7 2008:1 2008:7 2009:1 2009:7 Flat Phillips curve or missing inflation/deflation? 4 / 10

8 Mechanism I: Flex Price General flex-price model (Amiti, Itskhoki and Konings, 2016): p i = Γ i }{{} Pass-through elasticity mc i + Γ i p 1 + Γ }{{} i Strategic complementarity Often, Γ i = Γ(s i ), an increasing function of market share In particular, it is the case in Atkeson and Burstein (2008), but also in many monopolistic competition models 5 / 10

9 Mechanism I: Flex Price General flex-price model (Amiti, Itskhoki and Konings, 2016): p i = Γ i }{{} Pass-through elasticity mc i + Γ i p 1 + Γ }{{} i Strategic complementarity Often, Γ i = Γ(s i ), an increasing function of market share In particular, it is the case in Atkeson and Burstein (2008), but also in many monopolistic competition models Aggregate inflation in a flex-price environment: π = 1 S 1+Γ + S 1+Γ ( S 1 + Γ mc + S ) 1 + Γ mc S and S = 1 S are home and foreign expenditure shares Γ and Γ are average real rigidities of home and foreign firms respectively 5 / 10

10 Calvo price setting: Solution: π t = βe t π t+1 + λ Mechanism II: Sticky Price p it = (1 βθ) l=0 (βθ)l E t p i,t+l, p t = θp t 1 + (1 θ) p it di [ S 1 + Γ (mc t p t ) + S ] 1 + Γ (mc t p t ) 6 / 10

11 Calvo price setting: Solution: π t = βe t π t+1 + λ Mechanism II: Sticky Price p it = (1 βθ) l=0 (βθ)l E t p i,t+l, p t = θp t 1 + (1 θ) p it di [ S 1 + Γ (mc t p t ) + S ] 1 + Γ (mc t p t ) Globalization shock: S and S, and Γ With heterogeneous firms, it is possible than Γ Then, it is possible that κ (relevant for mc t = mc t shocks) 6 / 10

12 Calvo price setting: Solution: π t = βe t π t+1 + λ Mechanism II: Sticky Price p it = (1 βθ) l=0 (βθ)l E t p i,t+l, p t = θp t 1 + (1 θ) p it di [ S 1 + Γ (w t a t ) + S ] 1 + Γ (w t at + q t ) Globalization shock: S and S, and Γ With heterogeneous firms, it is possible than Γ Then, it is possible that κ (relevant for mc t = mc t shocks) Over-id tests for additional empirical discipline: 1 Differential response to a t and a t 2 Response to RER q t 3 Globalization with Europe vs China ( dollarized exports) (see Gopinath, 2015) 6 / 10

13 Technical challenge Sticky prices with large firms: Menu costs: multiplicity of equilibria (whether to adjust) (Neiman, 2010; 2011) Calvo: large state space (who adjusted when) Two solutions in the literature: 1 Sbordone (2010): drop large firms, adopt non-ces demand 2 Benigno and Faia (2010): large firms with Rotemberg pricing This paper: Beningno and Faia (2010) + heterogenous firms Alternative modeling approach (Mukhin, in progress): 1 Calvo pricing with large firms market power 2 Each firm is a collection of a mass of products, each subject to iid Calvo arrival, so LLN applies within a firm tractability 7 / 10

14 Pro-competitive effects of globalization? Large closely-related literature in trade Recent surge in interest: EMX (2015) and ACDR (2016) Pro-competitive effect suggests Γ stronger than Γ? fall in domestic markups and increase in importer markups The running null is that pro-competitive effect is a wash? fall in domestic markups is nearly fully offset by increase in importer markups 8 / 10

15 Other implications Cyclicality of markups Sungki Hong (2016) 1 Both small and large firms have countercyclical markups 2 Small firms are substantially more countercyclical 9 / 10

16 Conclusion Fascinating topic! expect more work in this area Very useful framework: multiple additional applications 10 / 10