Understanding International Prices:Customers as Capital

Transcription

1 Understanding International Prices: Customers as Capital Lukasz A. Drozd 1 Jaromir B. Nosal 2 1 University of Wisconsin-Madison 2 Columbia University

2 Fundamental features of international price data Aggregate data: - Real export and import prices of a country positively correlated - Both positively correlated with the real exchange rate Disaggregated data shows evidence of pricing to market - export price domestic price for the same commodities - vary systematically with the real exchange rate

3 Fundamental features of international price data Aggregate data: - Real export and import prices of a country positively correlated - Both positively correlated with the real exchange rate Disaggregated data shows evidence of pricing to market - export price domestic price for the same commodities - vary systematically with the real exchange rate Puzzle for a large class of models

4 Outline Illustrate why these observations are puzzling Document correlation of aggregate prices Document pricing-to-market using disaggregated data Propose model with customers as capital Show how model consistent with prices and quantities

5 Illustrate the Puzzle for Standard Theory Assumes country specific tradable goods; possibly non-tradable goods consumption baskets biased towards the home tradable good law of one price for each tradable good

6 Illustrate the Puzzle for Standard Theory Assumes country specific tradable goods; possibly non-tradable goods consumption baskets biased towards the home tradable good law of one price for each tradable good inconsistent with pricing-to-market observations

7 Illustrate the Puzzle for Standard Theory Assumes country specific tradable goods; possibly non-tradable goods consumption baskets biased towards the home tradable good law of one price for each tradable good inconsistent with pricing-to-market observations show also inconsistent with aggregate data

8 Illustrate the Puzzle for Standard Theory DOMESTIC COUNTRY Produces domestic goods CPI mostly comprised of domestic goods P f FOREIGN COUNTRY Produces foreign goods CPI mostly comprised of foreign goods P d

9 Illustrate the Puzzle for Standard Theory DOMESTIC COUNTRY Produces domestic goods CPI mostly comprised of domestic goods P f FOREIGN COUNTRY Produces foreign goods CPI mostly comprised of foreign goods P d p m p x P f CP I P f (P d ) ω (P f ) 1 ω = (P f ) ω P d P d CP I P d (P d ) ω (P f ) 1 ω = ( P d ) 1 ω P f corr(p x, p m ) = 1

10 Illustrate the Puzzle for Standard Theory DOMESTIC COUNTRY Produces domestic goods CPI mostly comprised of domestic goods 1 2 < ω < 1 P f P d FOREIGN COUNTRY Produces foreign goods CPI mostly comprised of foreign goods p m p x x P f CP I P f (P d ) ω (P f ) 1 ω = (P f ) ω P d P d CP I P d (P d ) ω (P f ) 1 ω = ( P d ) 1 ω P f CP I CP I = (P f ) ω (P d ) 1 ω (P d ) ω (P f ) 1 ω = (P f P d ) 2ω 1 corr(p x, p m ) = 1 corr(p x, x) = 1

11 Evidence for Correlations of Aggregate Prices 12 major OECD countries Statistics refer to detrended quarterly series, Statistic Country corr(p x, p m ) corr(p x, x) OECD median where p x = EP I CP I, IP I pm = CP I CP I x = CP I EP I = export price index, IP I = import price index Data has opposite signs as the standard model

12 Non-tradable Goods? With non-tradable goods, the CPI is given by CP I = (v(pd ω P 1 ω f ) µ 1 µ +(1 v)(cp I N ) µ 1 µ µ ) µ 1, where: v - share of N goods, µ - elasticity between T & N

13 Non-tradable Goods? With non-tradable goods, the CPI is given by CP I = (v(pd ω P 1 ω f ) µ 1 µ +(1 v)(cp I N ) µ 1 µ µ ) µ 1, where: v - share of N goods, µ - elasticity between T & N From the definition of p x and p m, obtain p x p m P d CP I P f CP I P d (v(pd ωp 1 ω f ) µ 1 µ + (1 v)(cp IN ) µ 1 µ µ ) µ 1 P f (v(pd ωp 1 ω f ) µ 1 µ + (1 v)(cp IN ) µ 1 µ µ ) µ 1

14 Non-tradable Goods? With non-tradable goods, the CPI is given by CP I = (v(pd ω P 1 ω f ) µ 1 µ +(1 v)(cp I N ) µ 1 µ µ ) µ 1, where: v - share of N goods, µ - elasticity between T & N From the definition of p x and p m, obtain p T x p T m [ 1 v ( P d CP I ) 1 µ (1 v) µ ( v [ 1 v ( P f CP I ) 1 µ (1 v) µ ( v P d CP I N ) 1 µ µ P f CP I N ) 1 µ µ corr(p T x, p T m) = 1 ] µ 1 µ = ( P d P f ) (1 ω) ] µ 1 µ = ( P f P d ) ω

15 Non-tradable Goods? 9 major OECD countries Statistics refer to detrended quarterly series, Statistic Country corr(p T x, p T m) corr(p T x, x) OECD median where [ p T 1 x 1 µ v p µ x ] µ (1 v) (p N x ) 1 µ 1 µ µ v [ p T 1 m 1 µ v p µ m ] µ (1 v) (p N m) 1 µ 1 µ µ v Data has opposite signs as the standard model

16 Non-tradable Goods? 9 major OECD countries Statistics refer to detrended quarterly series, Statistic Country corr(p T x, p T m) corr(p T x, x) OECD median where [ p T 1 x 1 µ v p µ x ] µ (1 v) (p x) 1 µ 1 µ µ v [ p T 1 m 1 µ v p µ m ] µ (1 v) (p m) 1 µ 1 µ µ v Data has opposite signs as the standard model

17 Non-tradable Goods? 9 major OECD countries Statistics refer to detrended quarterly series, Statistic Country corr(p T x, p T m) corr(p T x, x) OECD median where p T x = p x p T m = p m Data has opposite signs as the standard model

18 Disaggregated Evidence For Pricing-To-Market

19 Disaggregated Evidence For Pricing-To-Market Chosen OECD country: Japan Strong patterns on macro level (all correlations close to 1) Conveniently tabulated data (domestically-produced & sold goods)

20 Disaggregated Evidence For Pricing-To-Market Chosen OECD country: Japan Strong patterns on macro level (all correlations close to 1) Conveniently tabulated data (domestically-produced & sold goods) Disaggregated data for Japan suggests Export price movements attributable to pricing-to-market (PTM) PTM may be all we need to understand aggregate prices

21 Disaggregated Evidence For Pricing-To-Market Disaggregated wholesale price data for Japan Quarterly frequency, detrended using HP-filter 31 manufacturing commodities: copying machines, computers, etc...

22 Disaggregated Evidence For Pricing-To-Market Disaggregated wholesale price data for Japan Quarterly frequency, detrended using HP-filter 31 manufacturing commodities: copying machines, computers, etc... - Domestic Price : DP i price of goods produced and sold at home - Export Price : EP i price of goods produced at home but sold abroad

23 Micro Evidence For Pricing-To-Market Decomposing movements in real export price of commodity i p i x EP i CP I i EP i } DP {{ i } PTM PTM. Pricing-To-Market DP i } CP {{ I i } Res deviations of export price from domestic price for the same good Res. Residual term deviations of domestic price of the good from CPI

24 Micro Evidence For Pricing-To-Market Decomposing movements in real export price of commodity i p i x EP i CP I i EP i } DP {{ i } PTM Decomposing volatility: var(p i x)/var(x) 88% PTM = Res = DP i } CP {{ I i } Res var( EP i DP i ) var( EP i DP i ) + var( DP i CP I ) 93% var( DP i CP I ) var( EP i DP i ) + var( DP i CP I ) 7%

25 Micro Evidence For Pricing-To-Market Decomposing movements in real export price of commodity i p i x EP i CP I i EP i } DP {{ i } PTM Decomposing volatility: var(p i x)/var(x) 88% PTM = Res = DP i } CP {{ I i } Res var( EP i DP i ) var( EP i DP i ) + var( DP i CP I ) 93% var( DP i CP I ) var( EP i DP i ) + var( DP i CP I ) 7% Volatility attributable to pricing-to-market (PTM)

26 Micro Evidence For Pricing-To-Market Decomposing movements in real export price of commodity i p i x EP i CP I i EP i } DP {{ i } PTM Decomposing correlation with x: corr( p i x, x ) = 0.82 PTM = corr( EP i DP i } CP {{ I i } Res DP i, x ) = 0.84 Res = corr( DP i CP I, x ) = 0.15

27 Micro Evidence For Pricing-To-Market Decomposing movements in real export price of commodity i p i x EP i CP I i EP i } DP {{ i } PTM Decomposing correlation with x: corr( p i x, x ) = 0.82 PTM = corr( EP i DP i } CP {{ I i } Res DP i, x ) = 0.84 Res = corr( DP i CP I, x ) = 0.15 Correlation attributable to pricing-to-market (PTM)

28 Data - Summary Aggregate data real export and import prices positively correlated real export price positively correlated with the real exchange rates Disaggregated data suggests export price movements can be attributed to pricing-to-market for more evidence see the survey by Goldberg and Knetter (1997)

29 This Paper Building market shares is costly and time consuming as argued by Dornbusch (1987) and Krugman (1986) Leads to variable markups and pricing-to-market

30 This Paper Building market shares is costly and time consuming as argued by Dornbusch (1987) and Krugman (1986) Leads to variable markups and pricing-to-market Contribution formalize this idea into a micro-founded GE model - model marketing frictions in search and matching environment - make relations with customers valuable

31 This Paper Building market shares is costly and time consuming as argued by Dornbusch (1987) and Krugman (1986) Leads to variable markups and pricing-to-market

32 This Paper Building market shares is costly and time consuming as argued by Dornbusch (1987) and Krugman (1986) Leads to variable markups and pricing-to-market Formalize this idea in a micro-founded model model marketing frictions in a search and matching environment make relationships with customers valuable

33 This Paper Building market shares is costly and time consuming as argued by Dornbusch (1987) and Krugman (1986) Leads to variable markups and pricing-to-market Formalize this idea in a micro-founded model model marketing frictions in a search and matching environment make relationships with customers valuable Key quantitative discipline different S-R and L-R price elasticity of trade flows

34 This Paper Building market shares is costly and time consuming as argued by Dornbusch (1987) and Krugman (1986) Leads to variable markups and pricing-to-market Formalize this idea in a micro-founded model model marketing frictions in a search and matching environment make relationships with customers valuable Key quantitative discipline different S-R and L-R price elasticity of trade flows Show promising in accounting for price data

35 Related Literature Direct Evidence Hakansson (1982), Egan & Mody (1993) Turnbull & Cunnigham (1981), Ruhl & Willis (2008)

36 Related Literature Direct Evidence Hakansson (1982), Egan & Mody (1993) Turnbull & Cunnigham (1981), Ruhl & Willis (2008) Models of Pricing-to-Market (Dornbusch 1987, Krugman 1986) Sluggish market shares: Froot & Klemperer (1989), Alessandria (2004) Consumer Search: Alessandria (2005) Vertical Structure of Industry: Atkeson & Burstein (2008) Local Nontradable Component: Dedola & Corsetti (2002, 2004)

37 Related Literature Direct Evidence Hakansson (1982), Egan & Mody (1993) Turnbull & Cunnigham (1981), Ruhl & Willis (2008) Models of Pricing-to-Market (Dornbusch 1987, Krugman 1986) Sluggish market shares: Froot & Klemperer (1989), Alessandria (2004) Consumer Search: Alessandria (2005) Vertical Structure of Industry: Atkeson & Burstein (2008) Local Nontradable Component: Dedola & Corsetti (2002, 2004) Short-Run/Long-Run Elasticity Puzzle Sunk Cost of Entry: Ruhl (2008) Evidence: Eaton & Kortum (2002), Head & Ries (2001), Hummels (2001), Reinert & Roland-Holst (1992) Incomplete Pass-Through Literature Goldberg & Campa (2005,06), Goldberg & Knetter (1997), Marston (1990)

38 Model

39 Basic Structure Symmetric world with two-countries and country-specific goods d good produced in the domestic country f good produced in the foreign country

40 Basic Structure Symmetric world with two-countries and country-specific goods d good produced in the domestic country f good produced in the foreign country Composite consumption and investment good domestic country: c + i = G(d, f) foreign country: c + i = G(f, d )

41 Basic Structure Symmetric world with two-countries and country-specific goods d good produced in the domestic country f good produced in the foreign country Composite consumption and investment good domestic country: c + i = G(d, f) foreign country: c + i = G(f, d ) d and f the only tradable goods Physical capital and labor immobile across countries

42 Agents Producers Produce goods d at home (f abroad), match with retailers Retailers Match with producers, intermediate in trade between producers and households Households Buy goods from retailers, accumulate physical capital, supply labor and capital to producers, trade assets

43 Flow of Goods Domestic Country Foreign Country Producers Producers d d f f Retailers Retailers d f f d Households Households c+i = G(d, f) c +i = G(f, d )

44 Flow of Goods Domestic Country Producers Foreign Country Producers d d Retailers Retailers Households Households

45 Flow of Goods Domestic Country Producers d f Foreign Country Producers Retailers Retailers Households Households

46 Flow of Goods Domestic Country Producers Foreign Country Producers Retailers Retailers d f Households c+i = G(d, f) Households

47 Two Levels of Trade Domestic Country Producers Foreign Country Producers Wholesale trade search and matching Wholesale trade search and matching Retailers Local retail trade competitive market Retailers Local retail trade competitive market Households Households

48 Two Levels of Prices Domestic Country Foreign Country Producers Producers p d p d p f p f Retailers Retailers P d P f P f P d Households Households

49 Producers (domestic country perspective) Measure one of producers Produce good d according to: zk α l 1 α Subject to country specific productivity shock log(z t ) = ψ log(z t 1 ) + ε t log(z t ) = ψ log(z t 1) + ε t

50 Producers (domestic country perspective) Measure one of producers Produce good d according to: zk α l 1 α Subject to country specific productivity shock log(z t ) = ψ log(z t 1 ) + ε t log(z t ) = ψ log(z t 1) + ε t By CRS, summarize production by marginal cost v (given w and r) v = min{wl + rk zf (k, l) = 1} k,l

51 Producers Face Marketing Friction Each has a list of customers H d, H d and marketing capital m d, m d Can only sell to customers from the list (a fixed amount per period) Marketing capital brings new customers to the list

52 Producers Face Marketing Friction Each has a list of customers H d, H d and marketing capital m d, m d Can only sell to customers from the list (a fixed amount per period) Marketing capital brings new customers to the list h searching retailers (potential new customers) m d m d + m f h searching retailers who become new customers

53 Producers Face Marketing Friction Each has a list of customers H d, H d and marketing capital m d, m d Can only sell to customers from the list (a fixed amount per period) Marketing capital brings new customers to the list h searching retailers (potential new customers) m d m d + m f h searching retailers who become new customers

54 Producers Face Marketing Friction Each has a list of customers H d, H d and marketing capital m d, m d Can only sell to customers from the list (a fixed amount per period) Marketing capital brings new customers to the list h searching retailers (potential new customers) m d m d + m f h searching retailers who become new customers Customer list evolves according to the law H d = (1 δ H )H d, 1 + m d m d + m f h

55 Producers Face Marketing Friction Each has a list of customers H d, H d and marketing capital m d, m d Can only sell to customers from the list (a fixed amount per period) Marketing capital brings new customers to the list h searching retailers (potential new customers) m d m d + m f h searching retailers who become new customers Customer list evolves according to the law H d = (1 δ H )H d, 1 + Marketing capital evolves according to the law m d m d + m f h m d = (1 δ m )m d, 1 + a d φm d, 1 ( δ m ) 2 m d, 1 a d

56 Summary of Producer Problem Maximize expected present value of Π subject to Π = (p d v)d + (xp d v)d va d xv a d sales constraints laws of motion d H d H d = (1 δ H )H d, 1 + m d m d + m f h m d = (1 δ m )m d, 1 + a d φm d, 1 ( δ m ) 2 m d, 1 analogous constraints apply in the foreign market a d

57 Search by Retailers (Other Producers) Search to match with producers (at cost χv) meet local producer with probability π= m d m d + m f meet foreign producer with probability 1 π= m f m d + m f The match gives opportunity to trade one unit of output per period The match dissolves with per period probability δ H

58 Search by Retailers (Other Producers) Search to match with producers (at cost χv) meet local producer with probability π= m d m d + m f meet foreign producer with probability 1 π= m f m d + m f The match gives opportunity to trade one unit of output per period The match dissolves with per period probability δ H

59 Endogenous Measure of Searching Retailers h Measure of searching retailers h is endogenously determined by where: πv d + (1 π)v f χv with = whenever h > 0 V d = max{0, P d p d } + (1 δ H )E t [QV d ] V f = max{0, P f p f } + (1 δ H )E t [QV f ]

60 Determination of Wholesale Prices Producer & retailer bargain for the wholesale price p d (or p f ) At each history s t prices satisfy the Nash Bargaining problem p d (s t ) argmax p {J d (s t ; p) θ V d (s t ; p) 1 θ } where J d (s t ; p) = max{0, p v(s t )} + (1 δ H)E tq(s t+1 s t )J d (s t+1 ; p d (s t+1 )) - value from the match for the producer V d (s t ; p) = max{0, P d (s t ) p} + (1 δ H)E tq(s t+1 s t )V d (s t+1 ; p d (s t+1 )) - value from the match for the retailer

61 Determination of Wholesale Prices Proposition The solution results in instantaneous surplus splitting p d = θp d + (1 θ)v p f = θp f + (1 θ)xv Intuition: from tomorrow on the trade surplus split in proportion θ, 1 θ from today on the trade surplus split in proportion θ, 1 θ Implication: today s instantaneous surplus split the same way

62 Households Maximize E t t=0 βt u(c, 1 l) subject to Armington aggregation c + i = G(d, f) = (ωd γ 1 γ law of motion for physical capital k(s t ) = (1 δ)k(s t 1 ) + i Numeraire normalization: price of final good is one + (1 ω)f γ 1 γ ) γ γ 1 standard budget constraint under complete markets P d d + P f f + Q(s t+1 s t )b(s t+1 s t )µ(ds t+1) = b(s t ) + wl + rk(s t 1 ) + Π S

63 Market Clearing and Feasibility Meeting probability consistency condition π = m d m d + m f Representativeness m d = m d, m f = m f Production feasibility d + d + a d + a f + χh = zf (k, l) Definition of equilibrium is standard

64 Intuition and Qualitative Features

65 Parameterization: Qualitative Features Model parameters are such that domestic and foreign goods close substitutes market shares are sluggish in the short-run Justified quantitatively by: estimates of low short-run and high long-run price elasticity of trade flows trade responsive to trade liberalizations trade unresponsive to price changes in time-series

66 Primitive Shock 1 Percentage Deviation from SS z * z Quarters After the Shock Positive productivity shock in the domestic country

67 Key Feature: Producers Price To Market Percentage Deviation from SS x p x -0.1 p d Quarters After the Shock Markups on exported goods go up when real exchange rate depreciates!

68 Comovement of Prices in the Models Benchmark model 0.8 Standard model Percentage Deviation from SS x p m p x Quarters After the Shock corr(p x, p m ) = +1 sd(p)/sd(x) = 0.26 Percentage Deviation from SS p m x p x = p d Quarters After the Shock corr(p x, p m ) = 1 sd(p)/sd(x) > 1

69 Why Do Producers Price To Market? Percentage Deviation from SS x p x -0.1 p d Quarters After the Shock p x = θxpd + (1 θ) v p d = θp d + (1 θ) v

70 Why Do Producers Price To Market? Wholesale prices 0.6 Retail prices Percentage Deviation from SS x p x p d Quarters After the Shock Percentage Deviation from SS * xp d P d Quarters After the Shock p x = θxpd + (1 θ) v p d = θp d + (1 θ) v xp d > P d p x > p d not arbitraged away due to marketing friction

71 Why xp d rises relative to P d? A. Retail prices (P d, P d) change slowly and little B. Real exchange rate x depreciates: xp d goes up relative to P d

72 Why xp d rises relative to P d? A. Retail Prices Change Slowly and Little Retail prices depend on relative scarcity of foreign to domestic goods P d = ω[ω + (1 ω) f d γ 1 γ ] 1 γ 1 Relative scarcity sluggish due to sluggish market shares in the S-R f d = H f = (1 δ H)H f, 1 + H d m f m d + m f h (1 δ H )H d, 1 + m d m d + m f h Domestic and foreign goods closely substitutable (high γ)

73 Why xp d rises relatively to P d? B. Real Exchange Rate Depreciates Real exchange rate = price of foreign consumption basket in terms of the domestic consumption consumption basket

74 Why xp d rises relatively to P d? B. Real Exchange Rate Depreciates Real exchange rate = price of foreign consumption basket in terms of the domestic consumption consumption basket Following the shock, delivering consumption at home costs less than delivering consumption abroad

75 Real Exchange Rate Depreciates Market share of good d Market share of good f Domestic Country Foreign Country

76 Real Exchange Rate Depreciates Domestic Households Foreign Households Retailers Retailers h h * Market share of good d Market share of good f Domestic Country Foreign Country

77 Real Exchange Rate Depreciates Domestic Households Foreign Households Retailers Retailers h h * Productivity shock = additional supply of cheaper d- goods Market share of good d Market share of good f Domestic Country Foreign Country

78 Real Exchange Rate Depreciates Domestic Households Foreign Households Retailers Retailers h h * How is the additional supply channeled to households? Market share of good d Market share of good f Domestic Country Foreign Country

79 Real Exchange Rate Depreciates Domestic Households Foreign Households Retailers Retailers h h * Market share adjustment (convex marginal cost) Market share of good d Market share of good f Domestic Country Foreign Country

80 Real Exchange Rate Depreciates Domestic Households Foreign Households Retailers Retailers h h * More intensive search by retailers (flat marginal cost) Market share of good d Market share of good f Domestic Country Foreign Country

81 Real Exchange Rate Depreciates Domestic Households Foreign Households Retailers Retailers h h * Asymmetric market shares = search more efficient at home Market share of good d Market share of good f Domestic Country Foreign Country

82 Real Exchange Rate Depreciates Domestic Households Foreign Households Retailers Retailers h h * Real exchange rate depreciates! Market share of good d Market share of good f Domestic Country Foreign Country

83 Real Exchange Rate Depreciates Domestic Households Foreign Households Retailers Retailers h h * Equilibrium response of search asymmetric Market share of good d Market share of good f Domestic Country Foreign Country c c *

84 Real Exchange Rate Depreciates Domestic Households Foreign Households Retailers Retailers h h * Market share of good d Market share of good f Domestic Country c Foreign Country Under complete markets MRS equalized with depreciating real exchange rate: * u'( c ) x = u'( c) c *

85 Recap Wholesale prices 0.6 Retail prices Percentage Deviation from SS x p x p d Quarters After the Shock Percentage Deviation from SS * xp d P * d P d Quarters After the Shock Retail prices sluggish Real exchange rate depreciates: xp d > P d Bargaining leads to p x > p d Marketing frictions make p x > p d sustainable in S-R

86 Parameterization and Quantitative Results

87 Key Quantitative Discipline Account for the short run vs. long run price elasticity puzzle trade responsive to tariff reductions in the long run trade unresponsive to price fluctuations in time-series

88 Key Quantitative Discipline Account for the short run vs. long run price elasticity puzzle trade responsive to tariff reductions in the long run price elasticity of trade high 8 trade unresponsive to price fluctuations in time-series price elasticity of trade low 1

89 Key Quantitative Discipline Account for the short run vs. long run price elasticity puzzle trade responsive to tariff reductions in the long run price elasticity of trade high 8 trade unresponsive to price fluctuations in time-series price elasticity of trade low 1 Pins down two parameters: elasticity of substitution in preferences: γ G(d, f) = (ωd γ 1 γ + (1 ω)f γ 1 γ ) γ γ 1 γ = 8 gives high long-run elasticity market expansion friction: φ m d = (1 δ m)m d, 1 + a d φ ( ad m d, 1 δ m ) 2 md, 1 φ gives low short-run elasticity

90 Details: Market Expansion Friction φ m d = (1 δ m )m d, 1 + a d φm d, 1 ( δ m ) 2 m d, 1 Set jointly with other parameters to match our measure of short-run empirical elasticity of substitution volatility ratio: σ( DA f )/σ( p f P DA ) = 0.71 (12 OECD) where: DA domestic absorption in constant prices where: f imports in constant prices where: p f deflator price of imports where: P DA deflator price of domestic absorption a d Theoretical justification: in the frictionless model volatility ratio is γ

91 Parameterization Overview Step 1: Select the following parameters independently γ = 7.9, β = 0.99, α = 0.36, σ = 2, δ = 0.025, δ H = 0.1 (arbitrary) Step 2: Select remaining parameters jointly φ = 18.4, δ m = 0.2, θ = 0.4, χ = 1.38, η = 0.34, ω = 0.56, shocks to hit the following targets from the data Data Target Value 1. Import to GDP 12% 2. Producer markups 10% 3. Volatility of p x relative to x 37% 4. Volatility ratio Market activities in time endowment 30% 6. Share of marketing expenditures in GDP 7.0% 7. Moments of TFP process

92 Quantitative Results

93 Quantitative Results State results and contrast with standard theory Benchmark ( γ = 7.9, φ = 18.4 ) Standard model ( γ = 0.71, no φ ) - worse statistics for international prices - similar statistics for quantities

94 Quantitative Results: International Prices Statistic A. Correlations Model Economies Benchmark Standard Benchmark FA γ = 7.9 γ = 0.7 γ = 7.9 Data φ > 0 no φ φ > 0 p x, p m p x, x p, x B. Volatility relative to x p x p m p std(x)

95 Quantitative Results: International Prices Statistic A. Correlations Model Economies Benchmark Standard Benchmark FA γ = 7.9 γ = 0.7 γ = 7.9 Data φ > 0 no φ φ > 0 p x, p m p x, x p, x B. Volatility relative to x p x p m p std(x)

96 Quantitative Results: International Prices Statistic A. Correlations Model Economies Benchmark Standard Benchmark FA γ = 7.9 γ = 0.7 γ = 7.9 Data φ > 0 no φ φ > 0 p x, p m p x, x p, x B. Volatility relative to x p x p m p C. Volatility of Real Exchange Rate std(x)

97 Quantitative Results: International Prices Statistic A. Correlations Model Economies Benchmark Standard Benchmark FA γ = 7.9 γ = 0.7 γ = 7.9 Data φ > 0 no φ φ > 0 p x, p m p x, x p, x B. Volatility relative to x p x p m p C. Volatility of Real Exchange Rate std(x)

98 Quantitative Results: Quantities Statistic A. International Comovement Model Economies Benchmark Standard Benchmark FA γ = 7.9 γ = 0.7 γ = 7.9 Data φ > 0 no φ φ > 0 Output Consumption Investment Employment B. Volatility relative to GDP Consumption Investment Employment Net Exports

99 Comparison to Disaggregated Data Consider our previous decomposition: p i x EP i CP I i EP i } DP {{ i } PTM Volatility of p i x relative to x: DP i } CP {{ I i } Res Data Benchmark : 93% from PTM : 87% from PTM Correlation with x: Data : PTM= 0.84 Res= 0.15 Benchmark : PTM= 1.00 Res= 1.00

100 Comparison to Disaggregated Data Consider our previous decomposition: p i x EP i CP I i EP i } DP {{ i } PTM Volatility of p i x relative to x: DP i } CP {{ I i } Res Data Benchmark Standard : 93% from PTM : 87% from PTM : 0% from PTM Correlation with x: Data : PTM= 0.84 Res= 0.15 Benchmark : PTM= 1.00 Res= 1.00 Standard : PTM= 0.00 Res= 1.00

101 Conclusions Model can account for prices without hurting quantities Same frictions account for different SR and LR elasticities of trade flows Important to integrate international macro with static trade theory

102 Conclusions Model can account for prices without hurting quantities Same frictions account for different SR and LR elasticities of trade flows Important to integrate international macro with static trade theory Other interesting features of the theory positive relation between trade and comovement of business cycles Long-Run Price Elasticity of Trade and the Trade-Comovement Puzzle positive relation between trade and volatility of real exchange rates Trade Intensity and Real Exchange Rate Volatility

103 Backup Slides

104 Volatility Ratio in the Standard Model Standard model adopts Armington 69 model of trade G(d, f) = (ωd γ 1 γ + (1 ω)f γ 1 γ γ ) γ 1 d domestic good, f foreign good, γ Armington elasticity Step 1: demand relations: p d = G d (d, f), p f = G f (d, f) Step 2: derive from demand relations log( f d ) = γ log( p d ) + γ log( ) p f 1 ω t Step 3: independent ω shocks + standard deviation of both sides σ[log( f d )] γσ[log( p d p f )] ω t 1

105 Details: Market Expansion Friction φ Logged quarterly data Volatility Ratio Country HP-1600 HP-10 6 US Canada Japan UK OECD median Standard Model = γ = γ 1

106 Robustness Price index used to construct p x, p m, x CPI all-items CPI tradables WPI or PPI None (nominal) Country p x, x p m, x p x, x p m, x p x, x p m, x p x, e p m, e Belgium Canada France Germany Italy Japan Netherlands Switzerland US Australia n.a n.a Sweden n.a n.a UK n.a n.a Median