Presentation on theme: "International Trade, Firm Heterogeneity, and Intermediation"— Presentation transcript:
1International Trade, Firm Heterogeneity, and Intermediation Horst Raff, University of KielZhejiang University17-19 May 2011
2Syllabus International Trade with Heterogeneous Firms IntroductionTrade Model with Monoplistic Competition (Krugman)Monopolistic Competition with Heterogeneous Firms (Melitz, Ottaviano)Reciprocal Dumping Model (Brander)Reciprocal Dumping Model with Heterogeneous Firms (Long, Raff, Stähler)Trade and InnovationIntra-Industry Adjustment to Import CompetitionIntermediation in International TradeBuyer Power in International MarketsImports and the Structure of Retail MarketsManufacturers and Retailers in the Global Economy
3Introduction Basic Models of International Trade Comparative advantage (Ricardo, Heckscher-Ohlin)Economies of Scale (Krugman)Reciprocal Dumping (Brander)Trade Models Based on Economies of Scalemay explain why we observe a lot of intra-industry trade between (rich) countries with similar technology and factor endowments,provide a theoretical foundation for the gravity equation.Reciprocal Dumping Modelalso features intra-industry trade (even in identical goods),endogenous mark-ups,competitive effects of trade.
4Krugman Model I. Model Key elements: Preferences: Dixit-Stiglitz preferences (love of variety),L consumers/workers (labor is the only factor of production)Monopolistic competition, general equilibriumEconomies of scale in productionPreferences:elasticity of substitutionset of consumed varieties (endogenous)
5Krugman Model Demand for variety i: Constant price elasticity of demand:Technology: labor requirement to produce y units of good i:
6Krugman Model II. Equilibrium Each firm chooses price to maximize its profit:First-order condition (marginal revenue = marginal cost):Maximized profit:
7Krugman Model Free-entry condition (zero profits): Labor market clearing:Equilibrium price index:Indirect utility is inversely proportional to the price index.
8III. Free Trade (C integrated countries) Krugman ModelIII. Free Trade (C integrated countries)No change in consumer pricesNo change in output per firmGreater product variety as consumers gain access to foreign varieties:Greater welfare / lower price index due to variety effect:
9Melitz-Ottaviano Model I. ModelKey elements:Quadratic, quasi-linear preferences (love of variety),L consumers/workers (labor is the only factor of production)Monopolistic competition, partial equilibriumEconomies of scale in productionPreferences:
10Melitz-Ottaviano Model Demand for variety i:with average industry price:Price elasticity of demand:Technology:Constant marginal cost: cFixed market entry cost:
11Melitz-Ottaviano Model II. Equilibrium with homogeneous firmsEach firm chooses price to maximize its profit:First-order condition:With symmetric firms:
12Melitz-Ottaviano Model Free-entry condition (zero profits):Equilibrium allocation:Indirect utility:
13Melitz-Ottaviano Model III. Free Trade (C integrated countries)Consumer prices fall as competition gets tougherOutput per firm increases as firms need to sell more to break evenProduct variety rises as consumers gain access to foreign varieties:Greater welfare / lower price index due to lower prices and greater variety
14Melitz-Ottaviano Model IV. Firm HeterogeneityKey changes relative to homogeneous firm model:Firms draw their productivity/marginal cost from a distribution G(c)To obtain closed-form solutions assume a Pareto distributionFirms enter the market and pay their (sunk) entry cost before observing their productivity draw (expected zero-profit condition)
15Melitz-Ottaviano Model Optimal output of a firm has to satisfy:Define:Firms with a marginal cost draw greater than do not produce any outputThe threshold summarizes all the information required to describe the behavior of firms that produce positive output.
17Melitz-Ottaviano Model Free-entry condition (zero profits):Equilibrium allocation with Pareto distribution of productivity:where:Indirect utility:
18Melitz-Ottaviano Model V. Free Trade (C integrated countries)The threshold value of the marginal cost becomes smaller, i.e., import competition forces the least efficient firms shut downDemand becomes more elasticAverage prices and mark-ups fall (pro-competitive effect)Average output risesHigher average productivitySurviving firms are more profitableGreater product variety as consumers gain access to foreign varietiesGreater welfare
19Melitz-Ottaviano Model V. Costly TradeAdditional threshold value of the marginal cost below which firms export:Only the most efficient firms export, less efficient firms sell only in the domestic market, the least efficient firms produce zero output.Expected zero-profit condition:Threshold value of the marginal cost above which firms shut down:
21Brander Model – reciprocal dumping I. ModelKey elements:Quadratic, quasi-linear preferences2 countries: home, foreignSegmented markets (no consumer arbitrage)Oligopolistic competition, partial equilibriumPreferences:
22Brander Model Inverse demand in each country j =h,f: The home firm sells y units at home and exports x units.The foreign firm sells y* units in the foreign country and exports x* units to the home country.
23Brander Model Technology: Constant marginal cost: c Fixed cost: f Per-unit trade cost: t
24Brander Model II. Equilibrium Each firm chooses domestic sales and exports to maximize its profit:With segmented markets and constant marginal cost we can focus on first-order conditions in the home market:Best-response functions:
26Brander Model Equilibrium domestic sales: Equilibrium exports are positive as long as
27Brander Model Mark-up on domestic sales: Mark-up on exports: Reciprocal dumping: firms accept lower mark-ups on their exports.
28Melitz-Ottaviano Model III. Trade Liberalization (reduction in t)Reduces each firm’s domestic sales and raises exportsRaises total output in each country and lowers pricesHas non-monotonic effect on social welfare since profitspro-competitive effect due to import competitionprofits may fall due to trade cost (cross-hauling)With free entry welfare effect is unambiguously positive due to pro-competitive effect
30Brander Model Equilibrium with n firms in each country: Equilibrium profits:
31Long, Raff, Stähler I. Model Key elements: Brander model Firms draw their marginal cost from a distribution F(c) as in Melitz/OttavianoFirms observe only their own cost draw, but do not know the cost draws of their competitorsDerive the Bayesian Nash equilibrium
32Long, Raff, Stähler Inverse demand in each country j =h,f: Marginal cost drawn from distribution
33Long, Raff, Stähler II. Equilibrium Firm i chooses domestic sales and exports to maximize its profit:First-order condition for domestic sales:Threshold value of the marginal cost above which domestic sales are zero:Optimal domestic sales:
34Long, Raff, Stähler First-order condition for exports: Threshold value of the marginal cost above which exports are zero:Optimal exports:With symmetric countries every firm sells at home and exportsHence expected output of rivals at home and abroad is:
36III. Trade Liberalization (reduction in t) Long, Raff, StählerIII. Trade Liberalization (reduction in t)Reduces each firm’s expected domestic sales and raises expected exportsRaises total expected output in each country and lowers expected pricesSelection effect: the least efficient firms are forced to shut downHas non-monotonic effect on social welfare since profitspro-competitive effect due to import competitionprofits may fall due to trade cost (cross-hauling)With free entry welfare effect is unambiguously positive due to pro-competitive effect
37Innovation and Trade with Heterogeneous Firms Ngo Van Long (McGill University, Montreal)Horst Raff (IfW, University of Kiel)Frank Stähler (University of Würzburg)
38Innovation and TradeHow does trade liberalization affect productivity?Does trade liberalization raise or reduce the incentive to invest in R&D (specifically process R&D)?How do changes in R&D interact with selection effects whereby the least efficient firms are forced to shut down?Innovation incentives and tradeFirms face tougher import competition and lose market share to foreign competitors,but gain better access to export markets.Overall effect of trade on R&D incentives is non-trivial
39Innovation and TradeWhat does existing theory tell us about the link between trade liberalization and productivity?Models with homogeneous firms (Melitz/Ottaviano, Brander):market integration leads to fewer but bigger firms, lower average cost (scale effect)Models with heterogeneous firms (e.g., Melitz/Ottaviano):Trade liberalization raises productivity in two ways:Selection effect: least efficient firms are forced to exitMarket-share reallocation effect: efficient firms gain market share at the expense of less efficient firmsNo scale effect with CES preferences (Melitz).
40Innovation and Trade Problems with the existing literature: Our paper: Firm productivity is assumed to be exogenous: firms draw their cost from a given distribution.Fact is: firms may influence their productivity.Trade liberalization may raise exports and productivity simultaneously.Evidence that firms raise their productivity with a view to becoming exporters (e.g. Alvarez and Lopez, 2005).Our paper:Firms may invest in R&D to increase their chance of drawing a low marginal cost.
41Innovation and Trade Model: R&D investment through which firms can increase their chance of drawing a low marginal cost.Oligopolistic competition: big firms exercise market power.Cost draws and R&D spending remain private information.Possible adjustments by firms to trade liberalization:R&D decisionsOutput adjustmentEntry and exit
42Innovation and TradeHigher r imlies a greater chance of drawing a low cost.Cost distribution with R&D (r):Convex cost of R&D:
44Innovation and Trade Timing of the game Firms choose R&D and make entry decision.Each firm learns its own cost.Firms choose domestic and export sales (Bayesian game).
45Innovation and TradeExpected sales:Expected profit:R&D choice:.
46Innovation and TradeHow does trade liberalization affect the threshold values of the marginal cost?For both fixed and endogenous market structure we can prove:t
47Innovation and Trade Trade Liberalization with Homogeneous Firms Endogenous R&D in the Brander modelWith and without zero-profit condition
48Innovation and TradeTrade Liberalization with Heterogeneous Firms – no entry
49Innovation and TradeTrade Liberalization with Heterogeneous Firms – free entry and exit (endogenous market structure)
50Innovation and Trade Reduction in trade costs: raises expected export sales due tohigher probability that any given firm will be efficient enough to be able to exportthose firms that do export increase their shipments abroadreduces/raises expected local sales due togreater import competitiongreater probability that firm has to exithigher R&D if trade cost is low
51Innovation and Trade U-shaped relationship between trade cost and R&D: A firm selling only on the domestic market would have less incentive to invest in R&D, since trade liberalization reduces its market share.An exporter would have an incentive to increase R&D, since its expected output rises and hence the marginal benefit of cost reduction.High trade cost: probability that the firm will be able to export is small. Hence trade lib. reduces R&D.Low trade cost: almost all active firms will be exporters and raise output when trade is liberalized. Hence R&D increases.
52Robust results (for both fixed and endogenous market structure) Innovation and TradeRobust results (for both fixed and endogenous market structure)Trade liberalization:raises (reduces) aggregate R&D spending if trade costs are low (high)increases firm size provided that trade costs are highinduces least efficient firms to exitraises social welfare if trade costs are sufficiently low
53Innovation and TradeHow does our model match up with the stylized facts of trade liberalization?The model reproduces the stylized fact that trade liberalizationreduces price-cost marginslowers domestic sales of import-competing firms (at least provided that trade costs are high or that market structure is endogenous)expands markets for very efficient firmsincreases efficiency at the plant level (at least for low trade costs or endogenous market structure)leads to different adjustment patterns within industries depending on the level of sunk entry costs.Firms that export tend to be larger and more productive than firms that do not export.
54Intra-Industry Adjustment to Import Competition: Theory and Application to the German Clothing IndustryHorst Raff (IfW, University of Kiel)Joachim Wagner (University of Lüneburg)
55Intra-Industry Adjustment to Import Competition: Theory and Application to the German Clothing IndustryHorst Raff (IfW, University of Kiel)Joachim Wagner (University of Lüneburg)
56IntroductionHow do heterogeneous firms in an industry adjust to an import shock?How does this affect industry productivity and competition in the short and the long run?We address these questions by constructing a simple heterogeneous-firm model and testing its predictions using micro-data for the German clothing industry.
57Stylized FactsSignificant increase in import penetration in the German clothing industryDue in part to successive elimination of import quotas under the Multi-Fibre Arrangement.Second step: Jan. 1, 2002.Third and final step: Dec. 31, 2004.Large changes in output, employment, market structure.
63Economic Issues How does the industry adjust to import penetration? through changes in competition between heterogeneous firms in the industry,i.e., changes in outputs and market structure.What does this imply for competition and industry productivity?Do the competitive effects differ between the short and the long run?
64Theoretical Framework Long, Raff, Stähler (2009): oligopoly model with heterogeneous firms.Short run: fixed number of entrants in the industry.Long run: endogenous market structure, number of entrants determined by an expected-zero-profit condition.(Questions 3 and 4 of the Advanced International Trade Exam 2009)
65ModelKey features:Oligopoly model of a domestic industry facing import competition.Heterogeneous firms: firms draw their marginal cost after entry, cost draws remain private information.Firms play a Bayesian Cournot game.
66Model Home inverse demand: p=A – Q – M M: import quota n domestic firmsCosts:Entry cost: feMarginal production cost: cCumulative distribution function: F(c)Timing of the gameEntry decision.Each firm learns its own cost. Cost draws are private information.Firms choose domestic sales (Bayesian Cournot game).
67Model Cost level below above which a firm produces zero output: Output of firm i:Ex-post profit of firm i:Expected ex-ante profit:
68Model Nash Equilibrium: all firms have the same expected output: Short run:Long run:Productivity:
71DataPanel data for enterprises from the German clothing industry (all plants with >20 employees or belonging to an enterprise with >20 employees)Industry, sales, total employees, hours worked by blue-collar workers, gross wages and salaries,…Active firms in the clothing industry:2000: 6142005: 3102006: 274Number of employees in the clothing industry:2000: 66,8812006: 31,420
75ConclusionsSimple oligopoly model used to derive predictions about the adjustment of an industry to import competition.Pro-competitive effects in the short run disappear in the long run.Predictions of the model for the short run are supported by data for the German clothing industry:The least efficient firms exit the market.The output of survivors decreases.Industry productivity rises (economically small effect and not statistically significant)