International Economics: Lecture 10 & 11

Transcription

1 International Economics: Lecture 10 & 11 International Economics: Lecture 10 & 11 Trade, Technology and Geography Xiang Gao School of International Business Administration Shanghai University of Finance and Economics Oct. 13, / 4

2 International Economics: Lecture 10 & 11 News The 2011 Nobel Prize in Economics From left, Christopher Sims and Thomas Sargent 2 / 4

3 International Economics: Lecture 10 & 11 News What They Had Studied? I How temporary changes in economic policy, such as raising interest rates by a central bank or cutting taxes by the government, a ect the economy in terms of macroeconomic variables such as GDP and in ation. 3 / 4

4 International Economics: Lecture 10 & 11 In-Class Question 1 Light Switches I There is a light bulb inside a room and four switches outside. All switches are currently at o state and only one switch controls the light bulb. You may turn any number of switches on or o any number of times you want. How many times do you need to go into the room to gure out which switch controls the light bulb? 4 / 4

5 Technology, Geography, and Trade Eaton and Kortum (Econometrica, 2002) March 10, 2009 Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

6 Introduction Trade diminishes with distance Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

7 Introduction Trade diminishes with distance Prices vary across locations (US-Canada versus US-New Zealand) Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

8 Introduction Trade diminishes with distance Prices vary across locations (US-Canada versus US-New Zealand) Factor rewards are far from equal across countries Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

9 Introduction Trade diminishes with distance Prices vary across locations (US-Canada versus US-New Zealand) Factor rewards are far from equal across countries Countries relative productivities vary across industries Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

10 Introduction A Ricardian Model (as opposed to Heckscher - Ohlin model) Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

11 Introduction A Ricardian Model (as opposed to Heckscher - Ohlin model) Incorporates Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

12 Introduction A Ricardian Model (as opposed to Heckscher - Ohlin model) Incorporates absolute technological di erences Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

13 Introduction A Ricardian Model (as opposed to Heckscher - Ohlin model) Incorporates absolute technological di erences heterogeneity within industries Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

14 Introduction A Ricardian Model (as opposed to Heckscher - Ohlin model) Incorporates absolute technological di erences heterogeneity within industries geographical barriers Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

15 Model Index of goods j 2 [0, 1] ;Consumers: U = h R i σ 1 0 (q 1 (j))1 σ 1 σ Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

16 Model h R i σ 1 Index of goods j 2 [0, 1] ;Consumers: U = 0 (q 1 (j))1 σ 1 σ Each country draws a productivity for each j from F i (z) = Pr [Z i z] = e T i z θ Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

17 Model h R i σ 1 Index of goods j 2 [0, 1] ;Consumers: U = 0 (q 1 (j))1 σ 1 σ Each country draws a productivity for each j from θ = 8, T = 1 and 5 F i (z) = Pr [Z i z] = e T i z θ Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

18 Model T = 5, θ = 5 and 8. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

19 Model T = 5, θ = 5 and 8. Technology and pricing P ni = c i d ni Z i where c i is unit input cost; d ni is the iceberg cost; Z i is the productivity draw for that good Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

20 Model The probability that any good in country n from source i has a price less than p : G ni (p) = Pr [P ni p] = Pr Z i c i d ni p = 1 Pr z i < c i d ni = 1 e T i (c i d ni ) θ p θ p Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

21 Model The probability distribution of prices in country n: the probability that the lowest price of a good in country n is less than p requires that at least one of the prices is less than p: Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

22 Model The probability distribution of prices in country n: the probability that the lowest price of a good in country n is less than p requires that at least one of the prices is less than p: G n (p) = 1 Π N i =1 (1 G ni (p)) {z } probability that all prices exceed p {z } probability that at least one of them is below p = 1 e pθ Σ N i=1 T i (c i d ni ) θ = 1 e Φ np θ Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

23 Model The probability distribution of prices in country n: the probability that the lowest price of a good in country n is less than p requires that at least one of the prices is less than p: G n (p) = 1 Π N i =1 (1 G ni (p)) {z } probability that all prices exceed p {z } probability that at least one of them is below p = 1 e pθ Σ N i=1 T i (c i d ni ) θ = 1 e Φ np θ Φ n distills the parameters of the e ciency distributions, input costs, and trade costs around the world into a single term to summarize the distribution of lowest costs available to country n Φ n Σ N i=1 T i (c i d ni ) θ Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

24 Model The derivation of π ni : the probability that country i o ers the lowest price of a good to country n is that all other countries prices are above it. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

25 Model The derivation of π ni : the probability that country i o ers the lowest price of a good to country n is that all other countries prices are above it. π ni = Pr [P ni (j) min [P ns (j) ; s 6= i]] Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

26 Model The derivation of π ni : the probability that country i o ers the lowest price of a good to country n is that all other countries prices are above it. π ni = Pr [P ni (j) min [P ns (j) ; s 6= i]] = Z 0 s6=i (1 G ns (p)) dg ni (p) Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

27 Model The derivation of π ni : the probability that country i o ers the lowest price of a good to country n is that all other countries prices are above it. π ni = Pr [P ni (j) min [P ns (j) ; s 6= i]] = Z 0 s6=i (1 G ns (p)) dg ni (p) = Z 0 e Φ np θ θp θ 1 T i (c i d ni ) θ dp Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

28 Model The derivation of π ni : the probability that country i o ers the lowest price of a good to country n is that all other countries prices are above it. π ni = Pr [P ni (j) min [P ns (j) ; s 6= i]] = Z 0 s6=i (1 G ns (p)) dg ni (p) = Z 0 e Φ np θ θp θ 1 T i (c i d ni ) θ dp = T i (c i d ni ) θ Φ n Z 0 e Φ np θ θp θ 1 Φ n dp Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

29 Model The derivation of π ni : the probability that country i o ers the lowest price of a good to country n is that all other countries prices are above it. π ni = Pr [P ni (j) min [P ns (j) ; s 6= i]] = Z 0 s6=i (1 G ns (p)) dg ni (p) = Z 0 e Φ np θ θp θ 1 T i (c i d ni ) θ dp = T i (c i d ni ) θ Φ n Z 0 e Φ np θ θp θ 1 Φ n dp π ni = T i (c i d ni ) θ Φ n Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

30 Model The distribution " of prices of # goods that a country n actually buys from i: (1 G ns (q)) dg ni (q) denotes the pdf as before; divide s6=i by π ni to get the conditional distribution Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

31 Model The distribution " of prices of # goods that a country n actually buys from i: (1 G ns (q)) dg ni (q) denotes the pdf as before; divide s6=i by π ni to get the conditional distribution " # Z 1 p π ni (1 G ns (q)) dg ni (q) 0 s6=i = 1 e Φ np θ = G n (p) Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

32 Trade shares and price level Because the country n faces an identical distribution of prices from each country and the exports of country i equals its total mass of minimum prices it o ers, that is π ni. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

33 Trade shares and price level Because the country n faces an identical distribution of prices from each country and the exports of country i equals its total mass of minimum prices it o ers, that is π ni. X ni X n = π ni = T i (c i d ni ) θ Φ n Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

34 Trade shares and price level Because the country n faces an identical distribution of prices from each country and the exports of country i equals its total mass of minimum prices it o ers, that is π ni. X ni X n = π ni = T i (c i d ni ) θ Φ n Price index: the unit mass of product space transformed into a probability space. Then the price index essentially is p n = E p 1 σ h σ (1 σ) ln = E he pii 1 σ = γφ 1 θ n Φ n = γ θ pn θ γ is a function of σ and θ. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

35 Gravity equation Total export sales from country i Q i = N T i (c i d mi ) θ X m or T i ci θ = m=1 Φ m N m=1 Q i dmi θ Φ m X m Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

36 Gravity equation Total export sales from country i Q i = N T i (c i d mi ) θ X m or T i ci θ = m=1 Φ m Imports from i in country n X ni = T i (c i d ni ) θ X n Φ n N m=1 Q i dmi θ Φ m X m Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

37 Gravity equation Total export sales from country i Q i = N T i (c i d mi ) θ X m or T i ci θ = m=1 Φ m Imports from i in country n X ni = T i (c i d ni ) θ X n Φ n X ni = Q i d θ ni Φ n N m=1 X n dmi θ Φ m X m N m=1 Q i dmi θ Φ m X m Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

38 Gravity equation Gravity equation X ni = N m=1 dni p n θ Xn dmi p m θ Xm Q i Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

39 Gravity equation Gravity equation X ni = N m=1 dni p n θ Xn dmi p m θ Xm Q i The geographic barrier d mi is de ated by the price level in m; a sti er competition in m reduces p m reducing i s access in the same way as d mi does. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

40 Gravity equation Gravity equation X ni = N m=1 dni p n θ Xn dmi p m θ Xm Q i The geographic barrier d mi is de ated by the price level in m; a sti er competition in m reduces p m reducing i s access in the same way as d mi does. θ One can think of dmi p Xm m as the market size of destination m from the perspective of source country i. dmi p m θ Xm denotes the total market size whereas the numerator denotes n s share of i s total market. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

41 General equilibrium Input cost c i = w β i p 1 i β Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

42 General equilibrium Input cost Labor market equilibrium c i = w β i p 1 i β w i L i = β N n=1 π ni X n Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

43 General equilibrium Input cost Labor market equilibrium c i = w β i p 1 i β w i L i = β N n=1 π ni X n Either have wages determined in another non-manufacturing sector and let the labor be mobile. Or x the labor in the manufacturing sector that determines wages. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

44 General equilibrium When d ni = 1 for all n and i, Φ n = Φ for all n, and p n = p i for all n and i. Then π ni = T i (c i ) θ = T i w β i p 1 β θ i Φ Φ Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

45 General equilibrium When d ni = 1 for all n and i, Φ n = Φ for all n, and p n = p i for all n and i. Then π ni = T i (c i ) θ = T i w β i p 1 β θ i Φ Φ Then w i L i = T i w βθ i w j L j T j w βθ j Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

46 General equilibrium When d ni = 1 for all n and i, Φ n = Φ for all n, and p n = p i for all n and i. Then π ni = T i (c i ) θ = T i w β i p 1 β θ i Φ Φ Then or w i L i = T i w βθ i w j L j T j w βθ 1 w i Ti L 1+βθ j = w j T j L i j Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

47 General equilibrium The share of country 1 s output in the total expenditure of all N countries depends inversely on country 1 s marginal cost. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

48 General equilibrium The share of country 1 s output in the total expenditure of all N countries depends inversely on country 1 s marginal cost. If all countries are identical (equal labor) and there are no iceberg costs, each country is essentially supplying 1/N th of its output to all including itself. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

49 General equilibrium The share of country 1 s output in the total expenditure of all N countries depends inversely on country 1 s marginal cost. If all countries are identical (equal labor) and there are no iceberg costs, each country is essentially supplying 1/N th of its output to all including itself. Now increase the labor supply of one country. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

50 General equilibrium The share of country 1 s output in the total expenditure of all N countries depends inversely on country 1 s marginal cost. If all countries are identical (equal labor) and there are no iceberg costs, each country is essentially supplying 1/N th of its output to all including itself. Now increase the labor supply of one country. Intuitively, it should produce more. To do so, it must increase its relative share. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

51 General equilibrium The share of country 1 s output in the total expenditure of all N countries depends inversely on country 1 s marginal cost. If all countries are identical (equal labor) and there are no iceberg costs, each country is essentially supplying 1/N th of its output to all including itself. Now increase the labor supply of one country. Intuitively, it should produce more. To do so, it must increase its relative share. In equilibrium, this occurs through lower wages, lower marginal costs, and thus a higher share. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

52 General equilibrium The share of country 1 s output in the total expenditure of all N countries depends inversely on country 1 s marginal cost. If all countries are identical (equal labor) and there are no iceberg costs, each country is essentially supplying 1/N th of its output to all including itself. Now increase the labor supply of one country. Intuitively, it should produce more. To do so, it must increase its relative share. In equilibrium, this occurs through lower wages, lower marginal costs, and thus a higher share. Put di erently, the increase in labor goes to less e cient production thus pushing wages down. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

53 results Counterfactual Results: Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

54 results Counterfactual Results: 1 Trade bene ts all; smaller countries gain more Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

55 results Counterfactual Results: 1 Trade bene ts all; smaller countries gain more 2 As geographical barriers decrease, manufacturing shifts to larger countries where intermediates are cheaper, but at some point further decreases shift it back to smaller countries who can also now buy cheaper intermediates Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

56 results Counterfactual Results: 1 Trade bene ts all; smaller countries gain more 2 As geographical barriers decrease, manufacturing shifts to larger countries where intermediates are cheaper, but at some point further decreases shift it back to smaller countries who can also now buy cheaper intermediates 3 An increase in a country s technology bene ts all, the more proximate ones bene t more. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

57 Bernard, Eaton, Jensen, and Kortum (2003) Exporters are in minority Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

58 Bernard, Eaton, Jensen, and Kortum (2003) Exporters are in minority They tend to be more productive and larger Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

59 Bernard, Eaton, Jensen, and Kortum (2003) Exporters are in minority They tend to be more productive and larger Yet, they export only a small fraction of their output Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

60 Paper s contribution Link the micro and macro level data. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

61 Paper s contribution Link the micro and macro level data. Aggregate output and bilateral trade volumes to estimate aggregate technological di erences and geographical barriers. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

62 Paper s contribution Link the micro and macro level data. Aggregate output and bilateral trade volumes to estimate aggregate technological di erences and geographical barriers. Plant level data to estimate heterogeneity of production and consumption Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

63 Model Similar to Eaton and Kortum (2002). Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

64 Model Similar to Eaton and Kortum (2002). But each good produced by one rm (in equilibrium) in each country. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

65 Model Similar to Eaton and Kortum (2002). But each good produced by one rm (in equilibrium) in each country. For any good j they draw productivity Z (j) from a distribution, such that for any country the highest Z 1 and the second highest Z 2 are distributed in the following way: F (z 1, z 2 ) = Pr [Z 1i z 1, Z 2i z 2 ] i h = h1 + T i z2 θ z1 θ exp i T i z2 θ Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

66 Firms pricing Bertrand competition σ P n = min σ 1 C 1n, C 2n Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

67 Firms pricing Bertrand competition σ P n = min σ 1 C 1n, C 2n π ni similar to what we saw before Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

68 Firms pricing Bertrand competition σ P n = min σ 1 C 1n, C 2n π ni similar to what we saw before First and second lowest cost distributions for any country are identical to that o ered by any source Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

69 Firms pricing Bertrand competition σ P n = min σ 1 C 1n, C 2n π ni similar to what we saw before First and second lowest cost distributions for any country are identical to that o ered by any source mark-ups o ered for any country follows a truncated Pareto distribution; identical for every country Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

70 Firms pricing Bertrand competition σ P n = min σ 1 C 1n, C 2n π ni similar to what we saw before First and second lowest cost distributions for any country are identical to that o ered by any source mark-ups o ered for any country follows a truncated Pareto distribution; identical for every country price index similar Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

71 Firms pricing Bertrand competition σ P n = min σ 1 C 1n, C 2n π ni similar to what we saw before First and second lowest cost distributions for any country are identical to that o ered by any source mark-ups o ered for any country follows a truncated Pareto distribution; identical for every country price index similar results mimic stylized facts. Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21

72 Counterfactual experiment Experiment: A 5% drop in geographical barriers increases trade by 15% 3.3 % of US plants exit. But the 5% of existing who sold only to domestic markets begin exporting Exit of marginal rms, drop in the cost of intermediates, reallocation of market shares to bigger more e cient plants raise productivity Eaton and Kortum (Econometrica, 2002) () Econ 657 Spring 09 March 10, / 21