Import Competition and Employment Dynamics
Import Competition and Employment Dynamics Hˆale Utar University of Colorado
July 31, 2007; NBER International Trade and Investment Program
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Introduction
Objectives Develop and estimate an industrial evolution model with monopolistic competition, so that: Describe responses to heightened import competition in terms of: 1 2 3
job creation and destruction patterns productivity distributions entry and exit patterns
isolate the roles of different macro environments and labor regulations
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Introduction
Motivation Trade openness affects job flows in two ways: - sectoral reallocation of labor (e.g.Wacziarg and Wallack (JIE, 2004)) - within sector reallocation (e.g. Pavcnik (RESTud, 2002)) This paper quantifies, from the perspective of a single industry - Transition to greater openness: net job creation or destruction, aggregate productivity gain or loss - Intra-industry reallocations in the new ”long run” Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Introduction
Motivation The correlation between trade and labor market outcomes does not address the question of causality since both trade and employment could be responding to other factors ..Deardoff(1994) To characterize effects of openness on intra-industry reallocations, it is desirable to allow for: - Labor market frictions — policy-induced and others - Expectations on macroeconomic conditions, such as the possibility of regime change - Dynamic responses of firms, including entry/exit decisions Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Introduction
Related Literature
Dynamics of job flows under firing costs — e.g.Hopenhayn and Rogerson (JPE, 1993)
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Introduction
Related Literature
Dynamics of job flows under firing costs — e.g.Hopenhayn and Rogerson (JPE, 1993) Heterogeneity-based trade models — e.g. Melitz (Econometrica, 2003), Bernard, Eaton, Jensen and, Kortum (AER, 2003)
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Introduction
Simulation Experiments Simulation experiments quantify effects of transitionary changes in industry-wide environment (short-run response of industry to heightened import competition) role of severance payments during the transition to the low import price regime
Industry-wide environments - stochastic process for the real exchange rate and wages - tariff policy
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Data – Colombian Metal Products Industry, 1977-1991
Why Colombia?
Open, developing economy Change in policy regime (tariff, exchange rate, labor) Volatility & Policy Reversal Detailed plant-level panel data (77-91), macroeconomic data (77-98)
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Data – Colombian Metal Products Industry, 1977-1991
Why Metal Products?
Moderate to small scale producers (approximately 160 enterprises) High entry and exit rate — 22% and 21% High Import-Penetration Ratio — above 25% Insignificant Export-Orientation — 6%
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Basic Features of the Model
Monopolistically competitive product market Small, open industry — firms take prices of imported varieties and wages as given Heterogeneous firms in terms of productivity produce uniquely differentiated varieties Changing the level of employment requires adjustment costs Starting up a business is costly
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Basic Features of the Model
Monopolistically competitive product market Small, open industry — firms take prices of imported varieties and wages as given Heterogeneous firms in terms of productivity produce uniquely differentiated varieties Changing the level of employment requires adjustment costs Starting up a business is costly
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Basic Features of the Model
Monopolistically competitive product market Small, open industry — firms take prices of imported varieties and wages as given Heterogeneous firms in terms of productivity produce uniquely differentiated varieties Changing the level of employment requires adjustment costs Starting up a business is costly
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Basic Features of the Model
Monopolistically competitive product market Small, open industry — firms take prices of imported varieties and wages as given Heterogeneous firms in terms of productivity produce uniquely differentiated varieties Changing the level of employment requires adjustment costs Starting up a business is costly
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Basic Features of the Model
Monopolistically competitive product market Small, open industry — firms take prices of imported varieties and wages as given Heterogeneous firms in terms of productivity produce uniquely differentiated varieties Changing the level of employment requires adjustment costs Starting up a business is costly
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Environment
Infinite Horizon – Discrete Time Model Two Types of Firms: - Incumbents * Production and Exit decision
- Potential Entrants * Entry decision
State of the Industry: Distribution of firms on employment and productivity, Γt .
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Sequence of Actions beginning of time t incumbent firms decide to stay in the industry or exit observe the current realization of exogenous aggregate shocks observe their realization of productivity shocks observe the current state of the economy, Γt , and make their employment/production decisions potential entrants decide to enter or stay out for the next period end of time t
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Sequence of Actions beginning of time t incumbent firms decide to stay in the industry or exit observe the current realization of exogenous aggregate shocks observe their realization of productivity shocks observe the current state of the economy, Γt , and make their employment/production decisions potential entrants decide to enter or stay out for the next period end of time t
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Sequence of Actions beginning of time t incumbent firms decide to stay in the industry or exit observe the current realization of exogenous aggregate shocks observe their realization of productivity shocks observe the current state of the economy, Γt , and make their employment/production decisions potential entrants decide to enter or stay out for the next period end of time t
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Sequence of Actions beginning of time t incumbent firms decide to stay in the industry or exit observe the current realization of exogenous aggregate shocks observe their realization of productivity shocks observe the current state of the economy, Γt , and make their employment/production decisions potential entrants decide to enter or stay out for the next period end of time t
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Sequence of Actions beginning of time t incumbent firms decide to stay in the industry or exit observe the current realization of exogenous aggregate shocks observe their realization of productivity shocks observe the current state of the economy, Γt , and make their employment/production decisions potential entrants decide to enter or stay out for the next period end of time t
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Demand System Demand System quasi-linear preferences on horizontally differentiated products, domestic or imported.
U(qo , q1, q2 , .., qN ) = qo + α
N X i=1
qi = (
N
N
i=1
i=1
1 X 2 1 X 2 qi − γ qi − η( qi ) 2 2
α 1 ηN 1 − pi + P) ηN + γ γ ηN + γ γ
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Demand System
Pt
=
ND,t P D,t + NF ,t P F ,t ND,t + NF ,t
where P D,t denotes the average price among the domestic varieties; P F ,t denotes the average price of imported varieties, determined by the exchange rate and tariffs; ND,t and NF ,t are the number of domestic and foreign varieties, respectively.
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Production
qit = e µit litθ ,
0<θ≤1
µit = a0 + a1 µit−1 + εµ ,
εµ ∼ N(0, σµ2 )
The transition density for the firm specific productivity is denoted by M(µit+1 |µit ).
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Costs
Firms are price takers in the factor market, and wages are denoted by wt . Firing costs, cf per dismissed employee. Fixed per period costs, f .
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Exogenous Aggregate Shocks
Average imported prices, P F ,t The number of imported varieties, NF ,t Wages, wt st = (P F ,t , wt ) evolves according to a first order Markov Process with transition density Φ (st+1 |st ) . NF ,t = N F + t , with t ∼ N(0, σ2 )
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Incumbents’ Decision Problem Incumbents’ Problem is to choose labor and whether to exit or continue next period V (µit , lit−1 ; Γt , st ) = max Pi (Γt , lit , µit )e µit litθ − wt lit − c(lit , lit−1 ) − f lit
+β max(EV (µit+1 , lit ; Γt+1 , st+1 |µit , st ), −c(0, lit ) + x(lit )) subject to Γt+1 = H(Γt , st ), and c(lit , lit−1 ) = Max{0, cf (lit−1 − lit )}, where Pi (Γt , lit , µit ) is the inverse demand function. Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Potential Entrants’ Decision Problem Potential entrants draw entry costs, F , from a commonly known distribution, Ψ They choose to start-up a business or stay out. V E (Γt , st |M0 ) = βEV (µi,t+1 , 0; Γt+1 , st+1 ) subject to Γt+1 = H(Γt , st ) where M0 is the productivity distribution for potential entrants. They enter if the value of entering is higher than the cost associated with starting up a business, V E (Γt , st |M0 ) > F . Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Decision Rules
The optimization problems will generate policy functions, one for employment, lit = e(µit , lit−1 ; Γt , st ) one for exit, 0 if EV > −c(0, lit ) + x(lit ) χ(µit , lit−1 ; Γt , st ) = 1 otherwise one for entry rule for potential entrants, ENt .
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Model
Equilibrium Given transition densities for firm specific productivity M, M0 , for exogenous aggregate shocks,Φ, for the distribution of firms, H, and the distribution of entry costs,Ψ, an equilibrium is a value function V for incumbents, a value function V E for potential entrants, and a set of decision rules e(.) , χ(.) and ENt such that 1
2
3
Given M, Φ, and H each incumbent solves her decision problem and the resulting decision rules are given by e(.) and χ(.). Given V , M0 and H, V E characterizes the problem of potential entrants. H is consistent with firm’s optimal decision rules.
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Estimation
Estimation Methodology
Solving the model (Krusell and Smith,1998) Profits depend on [P t , Nt ], which depends on [Γt , st ] So agents use Γt+1 = H(Γt , st ) to solve the optimization problem But an agent who knows [P t , Nt ] process behaves the same as an agent who knows Γt process e t , st ) where mt = [P D,t , ND,t ]. Approximate H with H(m
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Estimation
Estimation of Aggregate Shock Processes Using Data between 1980-1998 The Markov Switching VAR: st = βor + β1r st−1 + rt 0
where E (rt rt ) = Σr .Switches between regimes are governed by the transition matrix p p Π = 11 12 , p21 p22 where pij , i ∈ {1, 2} is the probability of moving to regime j, given that the economy is currently in regime i. Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Estimation
Regime Switching VAR Processes Intercept β01 (regime 1) Intercept β02 (regime 2) AR coefficients β11 (regime
1)
AR coefficients β12 (regime 2) Covariance matrix Σ1 (regime 1) Covariance matrix Σ2 (regime 2) Switching probabilities Π
Utar
Wage 1.9229 (0.3296) 0.0520 (0.1520) 0.5131 (0.0819) -0.0149 (0.0083) 0.9905 (0.0410) -0.0035 (0.0050) 4.5240e-4 -1.5667e-5 1.2329e-4 6.0883e-5 0.9842 0.0185
Price 0.4470 ( 0.1493) -0.8942 (0.5112) -0.0847 (0.0376) 0.9700 (0.0037) 0.2892 (0.1377) 0.9538 (0.0175) -1.5667e-5 9.5084e-5 6.0883e-5 1.5470e-3 0.0158 0.9815
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Estimation
Estimation of Structural Parameters Embed the dynamic model in a method of moments estimator Choose the set of parameters, δ = (FH , f , cf , x, α, η, γ, Nf , a0 , a1 , σµ2 , z, θ) that minimizes a measure of distance between moments implied by model simulations and their sample counterparts, i.e. X (δ) = (d − m(δ))0 W (d − m(δ)) d : moments based on industry data m : simulated moments based on model W : matrix of weights
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Estimation
Identification Strategy
1 fixed cost, entry cost, scrap value → entry/exit rate, variance of log employment, job creation and destruction through entry and exit 2 demand parameters (η, α) →operating profit, entry rate, import-penetration; demand parameter (γ) →variances of log profit, number of firms 3 firing costs→ mean size, job destruction from exit, covariance moments
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Estimation
Identification Strategy 4 mean of entrants productivity→job creation from entry; variance of productivity→variance of employment, variance of employment growth 5 persistence of productivity, returns to scale→percentage of firms with no change in employment, covariance terms, employment growth, job flows from existing firms, mean size 6 variance of number of foreign varieties→variance of number of firms, variance of entry and exit; intercept of productivity distribution→mean employment
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Estimation
Estimation Results Cost Parameters
Sunk Entry Cost FH Fixed Cost, f Scrap Value, x Firing Cost, cf ∗ In thousand 1977 peso.
Parameters 7370∗ 1032∗ 45∗ 23.69∗
Utar
(100,218) (28,066) (1,223) (644) (1977 USD)
Standard Errors 97.1423 7.6165 5.2782 0.4193
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Estimation
Estimation Results
Production & Productivity Parameters
Parameters Returns to scale,θ Persistence parameter, a1 Variance, σµ2
Estimated Values 0.489 0.8987 0.229
Utar
Standard Errors 0.0062 0.0005 0.0115
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Simulation Experiments
Simulation Experiments
1
Using the two identified macro regimes in Markov-Switching VAR estimation, characterize the transition from the relatively closed regime to the low import price regime
2
Characterize the role of severance payments in the transition
Simulate the economy over 100 periods with 20 periods burn-in, repeat 20 times and average
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Simulation Experiments
The Effect of regime switching I
Total Employment Mean Gross Job Flows Variance Gross Job Flow Mean Productivity of Incumbents Mean Size Weighted Productivity Mean Number of Firms Mean Entry Rate Mean Exit Rate
Relatively Open (Regime 1) 2997 0.3338
Relatively Closed (Regime 2) 4508 0.3749
0.8687 1.2110 125.37 0.1549 0.1864
0.7876 1.1636 151.30 0.1828 0.1475
Weighted productivity increases by about 4.8 percent; employment decreases by about 33 percent. Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Simulation Experiments
The Effect of regime switching II
Mean Log Operating Profit Variance Log Operating Profit Mean Log Size Variance Log Size Mean Demand Elasticity Mean Import Price Variance Average Import Price
Relatively Open (Regime 1) 6.9291 1.8293 3.0712 0.5791 9.4325 62.621 0.2638
Relatively Closed (Regime 2) 6.9733 2.9626 3.1337 0.9162 10.954 89.359 9.7003
Firms contract by about 6 percent; profits decrease by about 4 percent with 35 percent reduction in import prices.
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Simulation Experiments
Simulation Experiments
1
Using the two identified macro regimes in Markov-Switching VAR estimation, characterize the transition from the relatively closed regime to the low import price regime
2
Characterize the role of severance payments in the transition
Simulate the economy over 20 periods with 10 periods burn-in, repeat 20 times and average
Utar
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Simulation Experiments
Transition: Number of Firms 190
firing costs; 3 months wages (benchmark) firing costs; 2 months wages
180
170
Number of Firms
160
150
140 regime change 130
120
110
100
2
4
6
8
Utar
10 12 Year
14
16
18
20
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Simulation Experiments
Transition: Aggregate Employment 5500
firing costs; 3 months wages (benchmark) firing costs; 2months wages
5000
Total Number of Jobs
4500
4000
3500
3000
2500
2000
2
4
6
8
Utar
10 12 Year
14
16
18
20
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Simulation Experiments
Transition:Covariance Between Size and Productivity 1.5 firing costs; 3 months wages (benchmark) firing costs; 2 months wages 1.4
Covariance of Size and Productivity
1.3
1.2
1.1
1
regime change
0.9
0.8
0.7
2
4
6
8
Utar
10 12 Year
14
16
18
20
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Simulation Experiments
Transition: Size Weighted Productivity 1.35 firing costs; 3 months wages (benchmark) firing costs; 2 months wages
size−weighted productivity
1.3
1.25
1.2
1.15
1.1
regime change
2
4
6
8
Utar
10 12 Year
14
16
18
20
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Simulation Experiments
Transition:Productivity of Exiting Firms −0.2
firing costs; 3 months wages (benchmark) firing costs; 2 months wages
un−weighted productivity of exiting firms
−0.3
−0.4
−0.5
−0.6
−0.7 regime change −0.8
−0.9
2
4
6
8
Utar
10 12 Year
14
16
18
20
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Simulation Experiments
Transition: Productivity of Contracting Firms 1.2 firing costs; 3 months wages firing costs; 2months wages
Un−weighted productivity of contracting firms
1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2
2
4
6
8
Utar
10
12
14
16
18
20
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Simulation Experiments
Transition: Productivity of Expanding Firms 1.6 firing costs; 3 months wages firing costs; 2 months wages
Un−weighted productivity of expanding firms
1.4
1.2
1
0.8
0.6
0.4
0.2
2
4
6
8
Utar
10
12
14
16
18
20
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Simulation Experiments
Long-Run: Severance Payments Mean Log Size Variance Size Mean Profit Mean Number of Firms Mean Total Employment Mean Entry Rate Mean Exit Rate Mean S-W Productivity Mean Covariance Mean Total Layoff Costs † In thousand 1977 pesos.
(Benchmark) 3.0998 0.7061 6.9484 136.97 3632.6 0.1655 0.1694 1.1941 0.9842 14318†
Utar
(Lower Firing Costs) 3.1424 0.6281 7.2590 138.91 3806.2 0.1694 0.1730 1.2172 1.0634 11867†
Import Competition and Employment Dynamic
Import Competition and Employment Dynamics Simulation Experiments
Concluding Remarks Developed and estimated a dynamic structural model with firm heterogeneity and employment frictions Established a link between macroeconomic environment and the benefits of openness. The response of industry to tariff policy particularly depend on - underlying labor market policies - associated changes in aggregate volatility - expectations on the regime sustainability and time horizon
Utar
Import Competition and Employment Dynamic