Labor Market Rigidities, Trade and Unemployment

Transcription

1 Labor Market Rigidities, Trade and Unemployment Elhanan Helpman Harvard and CIFAR Oleg Itskhoki Princeton Chicago Booth May / 30

2 Motivation Institutional differences as a source of comparative advantage Conversly, trade may affect differently countries with different institutions Substantial differences across countries in labor market institutions: e.g., hiring costs, firing costs, and flexibility of hours Types of questions we want to address: How is a country affected by a change in its trade partner s labor market institutions? How does reduction in trade costs impact countries with different labor market frictions? Propose a framework to think about trade and labor market outcomes 2 / 30

3 Related Literature Institutions and Trade: Legal institutions: Levchenko (2004), Nunn (2006), Costinot (2006) Labor market: Cuñat and Melitz (2007) Long-run structural unemployment: Blanchard and Wolfers (2000) and Nickell et al. (2002) Trade and Unemployment: Minimum wages: Brecher (1974) Implicit contracts: Matusz (1986) Efficiency wages: Copland (1989) Fair wages: Kreickemeier and Nelson (2006) Search and matching: Davidson, Martin and Matusz (1999) Trade and unemployment with heterogenous firms: Egger and Kreickemeier (2006), Davis and Harrigan (2007), Mitra and Ranjan (2007), Felbermayr et al. (2007) 3 / 30

4 New Theoretical Framework Two countries with different degrees of labor market frictions Long-run (static) model; countries are linked only via trade Two sectors: DMP labor market frictions and wage bargaining Outside sector produces homogenous good under CRS Differentiated-good sector as in Melitz: monopolistic competition, increasing returns, heterogenous firms Labor market frictions: hiring/firing costs, productivity of matching, unemployment benefits, etc. Two types of shocks : labor market frictions and trade impediments 4 / 30

5 Main Results Differences in labor market frictions constitute a source of comparative advantage Both countries gain from trade, but country with a more flexible labor market benefits the most policy complementarity within country A country benefits from improvements in its labor market flexibility, but this harms its trade partner terms of trade vs. competitiveness effects However, both countries gain from a simultaneous proportional improvement in labor markets policy complementarity across countries Flexible labor market leads to higher productivity both at the firm level and selection at the aggregate Interactions of trade and labor market frictions generate rich patterns of unemployment 5 / 30

6 Extensions 1 Optimal labor and product market policy Helpman, Itskhoki and Redding (2011) Trade and Labor Market Outcomes 2 Political economy of labor market rigidity 3 Dynamics of adjustment to trade 4 Trade, inequality and redistribution Helpman, Itskhoki and Redding (2010) Inequality and Unemployment in a Global Economy with Marc Muendler: Empirics of trade and inequality Itskhoki (2010) Optimal Redistribution in an Open Economy 6 / 30

7 Preferences and Demand Utility: U = q ζ Qζ, 0 < ζ < 1 Differentiated Good Aggregator: Q = [ ] 1/β, ω Ω q(ω)β dω ζ < β < 1 Indirect utility function: V = E + 1 ζ ζ P ζ 1 ζ = E + 1 ζ ζ Qζ 7 / 30

8 Production and Market Structure Labor is the only factor of production Homogenous good requires one unit of labor: q 0 = H 0 Differentiated product market is as in Melitz (2003): f e is a fixed cost of entry with a variety ω upon entry productivity θ is drawn from G θ (θ): y = θh f d is the fixed cost of production f x is the fixed cost of entry into the foreign market τ > 1 is the variable iceberg trade cost Revenue is [ R j (θ) = Q β ζ 1 β j + I xj (θ) τ β β ζ 1 β Q 1 β ( j) ] 1 β (θh) β 8 / 30

9 Labor Market Homogenous sector Upon matching and production, surplus is split equally: Cost of hiring: Free entry of firms: π 0 = w 0 = 1 2 b 0 = 1 2 a 0x α 0, x 0 = H 0 N 0, a 0 2ν 0 m 1+α 0 π 0 = b 0 = 1 2 Expected worker income: x 0 = a 1/α 0 ω 0 = x 0 w 0 = b 0 x 0 9 / 30

10 Hiring cost: C(h) = bh, Labor Market Differentiated sector b = 1 2 ax α Stole and Zwiebel (1996) bargaining: Inequality w(h, θ) = β 1 + β R(h, θ) = b h Expected workers income: ω = xw = bx Indifference of workers (ω 0 = ω): b 0 x 0 = bx b = b 0 ( a a 0 ) 1 1+α 10 / 30

11 Labor Market Digression: Dynamic Model Wages: w = (r + s)b + (b u + xb) Labor market equilibrium: 2(r + s 0 )b 0 + x 0 b 0 = 1 b u, b 0 = a 0 x α 0 Comparative advantage: b = b 0 ( a / a0 ) α 1+α Φ = (r + s)b (r + s 0 )b 0 = (s s 0 )b if b = b 0 11 / 30

12 Differentiated Sector Firm s problem Profit maximization π(θ) = max h,i x {0,1} { 1 } 1 + β R(h, I x; θ) bh f d I x f x Exit if π(θ) < 0 Free entry: max{π(θ), 0}dG(θ) = f e 12 / 30

13 Operating profits: Equilibrium Conditions where Θ θ β 1 β π dj (Θ) = φ 1 φ 2 b π xj (Θ) = φ 1 φ 2 b Two cutoff conditions: β 1 β j Q β ζ 1 β j Θ f d, β 1 β j τ β 1 β Q β ζ 1 β ( j) Θ f x π dj (Θ dj ) = 0 and π xj (Θ xj ) = 0 Free entry condition: Θ dj π dj (Θ)dG(Θ) + Θ xj π xj (Θ)dG(Θ) = f e 13 / 30

14 Production and Export Cutoffs Figure: Cutoffs in a trading equilibrium 14 / 30

15 Welfare and Gains from Trade Proposition (i) Welfare is higher in the flexible country; (ii) An improvement in labor market institutions in one country raises its welfare and reduces the welfare of its trade partner terms of trade vs. competitiveness effects (iii) A simultaneous proportional improvement in labor market institutions in both countries raises welfare in both countries; (iv) A reduction in trade impediments raises welfare in both countries, however, the flexible country benefits proportionately more Proposition Both countries gain from trade 15 / 30

16 Trade Patterns Proposition (i) There are more firms and a larger fraction of firms export in the flexible country (ii) The flexible country exports differentiated products on net and imports homogenous good (iii) The share of intra-industry trade is smaller the larger the proportional gap in labor market institutions (iv) Assuming Pareto-distributed productivity, total volume of trade increases in the proportional gap in labor market institutions and decreases in the trade costs 16 / 30

17 Consider the following measure of productivity: [ TFP j M j H j Proposition Θ dj Θ 1 β β h dj (Θ)dG(Θ) + Productivity Θ 1 β β Θ xj (i) TFP j does not depend on b j in the closed economy; h xj (Θ)dG(Θ) (ii) TFP j is higher in any trading equilibrium than in autarky. Proposition Assuming individual productivity draws are distributed Pareto: (i) TFP j is higher in the flexible country; (ii) An improvement in b j raises TFP j and reduces TFP ( j) ; (iii) A reduction in τ raises TFP j in both countries. ] 17 / 30

18 Unemployment Unemployment rate: u = (1 x 0 ) L N L + (1 x) N L Two effects: Sectoral labor market tightness, x and x 0 Compositional effect across sectors, N/L Symmetric countries, b A = b B = b: 1 hump-shaped response of u to b 2 trade raises unemployment iff x < x 0 18 / 30

19 Unemployment Response to Labor Market Frictions Figure: Unemployment as a function of labor market friction 19 / 30

20 Unemployment Response to Trade Frictions Figure: Unemployment as a function of τ: b A = 0.7 > b B = / 30

21 Labor Market Institutions Unemployment Benefits: b u if not matched w/p 1 x Firing Costs: with probability σ match is unproductive firm bears administrative cost of firing b f and severance pay b p fired worker receives b p + ηb u > b u Labor market equilibrium: b = ax α + σ(b f + b p ) 1 σ w = b u, β R 1 + β h b u = b b u, x [ (1 σ)w + σ(b p + ηb u ) ] + (1 x)b u = ω 0 21 / 30

22 Labor Market Institutions Sufficient statistic (x, b): b = ax α + σ(b f + b p ) 1 σ b u, (C) x [ (1 σ)b + σ(b p + ηb u b u ) ] = ω 0 b u (I) Comparative statics: a, b f, b p and b u reduce x, i.e. slacken the labor market a and b f raise b, i.e. reduce competitiveness b p and b u may raise or reduce b, that is by slackening the labor market they may increase competitiveness 22 / 30

23 Distortions: Optimal Policy I Constrained efficiency 1 Search externalities 2 Monopolistic distortion vs. Stole-Zweibel overhiring Policy in labor market: Hosios condition: λ = 1/α (relative bargaining power of firm) Vacancy posting subsidy: s b = 1 αλ 1 + λ Unemployment benefits: b u = αλ 1 (1 + α)λ Policy in product market (with s b in place): s r = 1 β β λ 1 + λ and s f = λ 23 / 30

24 1 B Figure: Welfare gains and losses from b u 24 / 30 Optimal Policy II Second best policy: unemployment benefits only < 9;: '2 & ( * &.'/0, * -& ()+*, %'& =?>A@ =?>A@ =C>A@ED FHGI BFJI "!$#

25 Optimal Policy II Second best policy: unemployment benefits only ! Figure: Effects of b u 24 / 30

26 Additional extensions Political economy of regulatory gridlock Why keep inefficient labor market institutions? Standard answer: unions, insiders and labor parties Our answer: large firms support rigid labor markets as effective entry barriers for smaller firms Result: π = π ( θ; A(b), b ) Consider linear demand. For large enough θ, full effect of b on π is positive. 25 / 30

27 Additional extensions Large firms and the Shimer puzzle Wage rate w = b u + φ ( ) r + s + x b 1 φ Optimal employment of firms: ( (1 φ) β (1 φ) + φβ A(zθh) β ) b u = (r + s + φx)b }{{ h } =p Elasticity of b with respect to p does not depend on β, independently of free entry and other details Caveats: (1) b dynamics, and (2) non-hiring firms 26 / 30

28 Additional extensions Labor market dynamics of adjustment to trade Reallocation both across and within sectors Empirically most reallocation is within Tentatively, search frictions affect volume of trade but have little affect on welfare 27 / 30

29 Summary and Policy Implications Interaction of labor market rigidities and trade frictions generate rich patterns of unemployment Trade may raise unemployment in both countries Both countries gain from trade, but more flexible country gains proportionately more Improvements in a country s labor market institutions raises its welfare and hurts the trade partner Yet, a simultaneous proportional improvement in both labor markets raises welfare in both countries Trade liberalization and labor market deregulation are complementary Labor market reforms in one country are likely to encourage labor market reforms in its trade partners 28 / 30

30 Trade and inequality Helpman, Itskhoki and Redding (2010) Wage equals replacement cost of a worker: w(θ) = b h(θ)/n(θ) = ba c(θ) k Additional ingredients: unobservable worker heterogeneity (match effects) production complementarities costly imperfect screening back to model 29 / 30

31 Trade and inequality Helpman, Itskhoki and Redding (2010) w(θ) Wage rate, w(θ) w a (θ) 1 θ d a θ d θ x Productivity, θ Figure: Wage schedule: size and export wage premia 30 / 30

32 Trade and inequality Helpman, Itskhoki and Redding (2010) Theil Index, T w µ ln(1+µ) Trade Openness, ρ = θ d /θ x Figure: Trade openness and inequality 30 / 30

33 Trade and inequality Helpman, Itskhoki, Muendler and Redding (2011) % change over Total Within Between Figure: Wage inequality in Brazil: Within vs Between Occupations 30 / 30

34 Trade and inequality Helpman, Itskhoki, Muendler and Redding (2011) 0.12 % change over Total Within Between Figure: Wage inequality in Brazil: Within vs Between Sectors 30 / 30

35 Trade and inequality Helpman, Itskhoki, Muendler and Redding (2011) Between-firm % change over Within-firm Total (within s o) Figure: Wage inequality in Brazil: Within vs Between Firms 30 / 30