International Trade and Labour Market in Luxembourg

International Trade and Labour Market in Luxembourg

Xi Chen 13 October 2017 STATEC - Research Division

“No extension of foreign trade will immediately increase the amount of value in a country, although it will very powerfully contribute to increase the mass of commodities, and therefore the sum of enjoyments.”

1

Ch.7, On the Principles of Political Economy and Taxation published on the April 19th 1817

2

“In order to promote economic and national security and to help stimulate economic growth, create good jobs at decent wages, strengthen our middle class, and support the American manufacturing and defense industrial bases, it shall be the policy of the executive branch to maximize, consistent with law, through terms and conditions of Federal financial assistance awards and Federal procurements, the use of goods, products, and materials produced in the United States.”

3

Sec. 2. (a), Presidential Executive Order launched on the April 18th 2017 (the eve of 200th anniversary of Ricardos Principles)

4

Mercantilism (or Protectionism)

Jean-Baptiste Colbert (1619-1683)

Saint-Gobain (founded in 1665)

”Buy French - Hire French” 5

Open Trade vs. Protectionism and Labour Implications

• OECD: “Many people are concerned about the effects of open trade on employment where they live. In fact, OECD analysis shows that liberalised trade is an engine for job creation in all countries, [. . . ] However, trade liberalisation must be accompanied by appropriate employment and social policies, [. . . ]” • ILO: “Globalization can contribute to employment growth but open markets alone are unlikely to create enough good-quality jobs.”

6

Outline

• Part I: Importing Intermediate Inputs and Labour Demand

• Part II: Offshoring, Fair Wage and Labour Demand

7

Part I: Importing Intermediate Inputs and Labour Demand (Intensive margin)

Part I: Importing Intermediate Inputs and Labour Demand US.

Figure 1 from Autor, Dorn and Hanson (2013). The China syndrome: Local labor market effects of import competition in the US. American Economic Review 8

Part I: Importing Intermediate Inputs and Labour Demand

0.13 0.12 0.11 0.10 0.09

Manufacturing/Total hours worked

0.015 0.010 0.005

Chinese imports penetration

0.020

0.14

Luxembourg

2000

2005

Chinese Imports penetration

Chinese imports penetration =

2010

2015


Imported intermediate inputs from China Total intermediate consumption

9


2005

2010

0.15 0.13

0.03 0.01

2015

0.11

0.11 0.09

0.015 0.005 2000

0.05

Belgium 0.13

Luxembourg

2000

2010

2015


0.200

0.04

0.190

0.03 0.02

0.180

0.11 0.10 2005

2015

0.01

0.12

0.13

0.025 0.015 0.005 2000

2010

Germany 0.14

0.035

France

2005

2000

2005

2010

2015


10


Research question: How imports of intermediate inputs affect local labour demand?

11

Part I: Importing Intermediate Inputs and Labour Demand Luxembourg

2005

2010

2015


2000

2005

2010


0.010 0.005

0.0022 2000

0.020 0.025 0.030 0.035 0.040

0.015

Basic and fabricated metal industry

0.000

0.0026

0.08 0.06 0.04 0.02 0.00

Chinese imports penetration

0.10

Chemical and pharmaceutical industry

2015


The effects of imports on employment cannot be summarized by a single stylized fact! 12


How imports of intermediate inputs affect local labour demand? • Feenstra and Hanson (1996, 1997): purchasing an input from a foreign source must replace a task previously done by a domestic worker, which would suggest displacement and lower wages ,→ Labor-substitution effect

• Grossman and Rossi-Hansberg (2007, 2008): the ability to use foreign inputs may lower a firms costs and raise its productivity, allowing it to expand output and employment and raise wages ,→ Cost-reduction effect

13


Theory: • The trade model (`a la Melitz, 2003) - Trade in intermediate inputs - Additional heterogeneity: absorptive capacity - Nested CES production function

,→ Open trade increases relative productivity of intermediate input (factor biased technical change)

14


Empirical strategy: • A structural econometric model - Estimation of elasticity of substitution (ρ) between labour and intermediate input at the industry-level , which captures the production technology - Estimation of factor-augmenting productivity (zit ) at the firm-level, which captures the biased technical change

,→ The effects of imports on labour demand is regulated by the production technology

15


Table 1: Industrial grouping based on estimated ρ Substitution

Manufacturing Industries

#Firms

#Obs.

Low: ρ < 6

Wood Products Paper Products Chemical and Pharmaceutical Products Computer, Electronic and Optical Prod. Motor vehicles, Trailers and Semi-Trail. Other Transport Equipments Other Manufacturing, Repair and Inst.

61

508

Medium: ρ ∈ [6, 12]

Food Products, Beverages and Tobac. Printing and Recorded Media Other Non-Metallic Mineral Products Machinery and Equipment

97

751

High: ρ > 12

Basic and Fabricated Metal Products

52

440

16


Table 2: Partial correlation with factor-augmenting productivity, zit a Imp. Intensity. 0.178

#employees 0.055

High substitution group

0.082

-0.094

Medium substitution group

0.172

0.055

Low substitution group

0.246

0.193

Full Sample

a Note:

The partial correlations are controlled for industry and year fixed-effects.

17


Table 3: Partial correlation with factor-augmenting productivity, zit a Imp. Intensity. 0.178

#employees 0.055

High substitution group

0.082

-0.094

Medium substitution group

0.172

0.055

Low substitution group

0.246

0.193

Full Sample

a Note:

The partial correlations are controlled for industry and year fixed-effects.

18

Part II: Offshoring, Fair Wage and Labour Demand (Intensive/extensive margin)

Part II: Offshoring, Fair Wage and Labour Demand

In September 2010, Sergio Marchionne, CEO of Fiat, explicitly threatened to pull production out of Italy and offshore it to lower-cost plants located in Serbia and Poland if there is no major concessions from the unions.

19


Research question: How the wage setting process influences the local economy when firms may choose to offshore certain activities in order to minimize their domestic labor costs

20


Theory of Gift Exchange (Akerlof, 1982): • Workers have a preference for fairness and provide more efforts in exchange for a wage above some reference level that is considered as fair • Profit maximizing firms under the fair wage consideration may find it optimal to offer a wage that exceeds the market clearing level • In the context of heterogeneous firms and open economy, a fair wage condition can be expressed as wd (φ) = φθ (λwf )1−θ

21


Main contribution: This paper develops a general equilibrium model with monopolistic competition that incorporates (i) A sector of heterogeneous firms that differ in their productivity (Melitz, 2003) (ii) A fair wage which exceeds the market clearing level and varies with firms productivity (Egger and Kreickemeier, 2009) (iii) An open economy in which a part of production process can be moved to foreign countries (Groizard et al., 2014)

22

Part II: Offshoring, Fair Wage and Labour Demand Part I:

23

Part II: Offshoring, Fair Wage and Labour Demand Part II:

24

Part II: Offshoring, Fair Wage and Labour Demand Part II:

25


Effect of increasing fairness on local labour demand ∂M ∂l d ∂Ld = ld + M ∂θ ∂θ ∂θ • The extensive margin:

∂M ∂θ

• The intensive margin:

∂l d ∂θ

26


Effects of increasing fairness on labour demand:

Extensive margin

Intensive margin

27

• Hidden Champion: “48% of the mid-sized world market leaders come from Germany”; “Despite their medium or small size, they are true global players” • Investing in workers: “Middle-class manufacturing jobs”; “The Hidden Champions invest 50% more in vocational training than the average German company” • Globalizing:“Hidden Champions are present in their target markets with 30 subsidiaries on average”

28

The labour-substition effect 29

The cost-reduction effect 30

Technical appendix

The basic Melitz (2003) model

• World is comprised symmetric countries • Consumer’s preference over a continuum of final good varieties is characterized as in Dixit and Stiglitz (1977) • Final good producer pays a sunk cost of entry and draw their productivity • Productivity is the only source of heterogeneity

31

The basic Melitz (2003) model Equilibrium in a close economy • Zero cutoff profit: π(φ∗ ) = 0 • Free entry: π(φ) = π

32

Production function in the open economy

Following Groizard et al. (2014), the production of final goods is characterized by a continuum of tasks in the interval [0, 1]: "Z q(φ) = φ

α ˆ

qf (α)

ρ−1 ρ

Z

1

qd (α)

dα +

0

ρ−1 ρ

ρ # ρ−1

dα

α ˆ

where α ˆ reflects the offshoring intensity

33

Part I: Empirical model • The empirical version of the production function is in the form of Yi = φi Kiβ F (M0i , Li )1−β that is similar to Goldberg et al. (2010). Yi =

φi Kiβ

1 B(τi∗ ) ρ

[A(ai )M0i ]

ρ−1 ρ

+ (1 −

ρ−1 ρ

1 τi∗ ) ρ Li

ρ(1−β) ρ−1

• Maximizing the output with respect to the cost provides the first order conditions for every production factor. Dividing the FOC for labor with that for intermediate inputs provides the following. ρ 1 ρ−1 Li B (τi∗ ) ρ−1 w i Li = A (ai ) M0i 1 − τi∗ pMi M0i

• The factor-augmenting productivity Zi is B (τi∗ ) Zi = 1 − τi∗

1 ρ−1

A (ai ) 34

Part I: Empirical model

• The estimating equation: sit = β0 + βnit + zit + εit where β0 is the vector of constant, time and industry fixed effects, β is the coefficient of interest and εit is the i.i.d error term. • Estimating the above equation has various advantages over estimating a production function. • The capital stock does not enter into the estimating equation. • The unobserved output prices issue is eliminated. • The number of parameters to be identified is reduced, so that no need for a large number of observations.

35

Part I: Estimation Methodology • Standard control function approach used to estimate production functions (e.g. Olley and Pakes, 1996; Levinsohn and Petrin, 2003) has 2-steps. • Step 1: qit = α0 + αL lit + φ(mit , kit ) + it ∗ ˆ it , kit ) − αM ω (mit , kit ) = φ(m mit − αK∗ kit • Step 2: ω (mit , kit ) = g (ω (mit−1 , kit−1 )) + νit qit = δ0 + α ˆ L lit + αM mit + αK kit + gˆ (ω (mit−1 , kit−1 )) + νit + it

• Wooldridge (2009) reduces the 2-step approach into a single step by estimating a system GMM. • Petrin and Levinsohn (2012) simplifies Wooldridge’s approach by estimating only the second stage. We follow the simplified approach and estimate the following by the GMM using nit−1 and nit−2 as the instruments, while iit is the log import intensity as a proxy. sit = β0 + βnit + z (iit−1 , kit−1 ) + it + εit

36

Part I: Robustness testing Table 4: Employment Regressions Dependent Variable: Employment (1) (2) zit × GroupL 0.971*** 0.968*** (0.217)

zit × GroupM zit × GroupH Firm Age

(0.224)

0.186

0.209

(0.278)

(0.269)

-0.849***

-0.978***

(0.308)

(0.322)

0.012** (0.005)

Firm Age 2

0.000 (0.973)

Export Status

0.397*** (0.107) 37

Part II: Marginal cost function with offshoring and fair wage

• The dual representation of production technology is c(φ) =

1 i 1−ρ 1h 1−ρ K (ˆ α) (λwf ) + (1 − α ˆ )wd1−ρ , φ wd α ˆ = log λwf

where wd and wf denotes the domestic and foreign wages; λ is an offshoring cost • Fair wage condition wd (φ) = φθ (λwf )1−θ

38

Part II: Effects of fair wage, wd (φ) = φθ (λwf )1−θ , on the equilibrium

• Result 1. An increase in the fixed offshoring cost λ causes a decline in the threshold productivity level φ∗ in the new equilibrium

• Result 2. The effect of an increase in the fairness parameter θ on the equilibrium threshold productivity φ∗ is non-monotonic and determined by the offshoring level and the elasticity of substitution between domestic and foreign labor inputs

39

Part II: Effects of increasing fairness on local economy

φ∗

1.6

1.10

1.15

1.7

1.20

1.8

1.25

1.9

1.30

2.0

1.35

2.1

1.40

φ

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.1

θ

(a) Threshold productivity

π/π (θ=.1)

0.2

0.3

0.4

0.5

0.6

(b) Average productivity

0.7

θ

0.9 0.8 0.4

0.4

0.5

0.5

0.6

0.6

0.7

0.7

0.8

0.9

1.0

1.0

M/M(θ=.1)

0.1

0.2

0.3

0.4

0.5

0.6

0.7

(c) Average profit (baseline at θ = .1)

θ

ρ = 0.5

0.1

0.2

0.3

0.4

0.5

0.6

0.7

θ (d) Equilibrium mass (baseline at θ = .1) ρ = 15

40