Banque nationale de Belgique SA · EURONEXT:BNB
Banque nationale de Belgique SA — presentatie bartelsman
40 slides from Banque nationale de Belgique SA's investor presentation, Nov 30, 2012
Slide 1: Employment Protection, Technology Choice, and Worker Allocation
Introduction Model and Calibration Results Employment Protection, Technology Choice, and Worker Allocation Eric J. Bartelsman1 , Pieter Gautier1 , and Joris de Wind2 1 Vrije Universiteit Amsterdam; Tinbergen Institute; IZA Bonn 2 CPB Netherlands November 30, 2012 – NBB, Brussels Bartelsman, Gautier,
Slide 2: Outline
Introduction Model and Calibration Results Outline 1 Introduction Motivation Stylized Facts 2 Model and Calibration Theory Calibration and Simulation 3 Results Simulation Empirical Results Bartelsman, Gautier, de Wind EPL and Technology Choice
Slide 3: Productivity Divergence Puzzle
Introduction Motivation Model and Calibration Stylized Facts Results Productivity Divergence Puzzle EU productivity almost caught up to US by 1995. Since then, the gap has widened again Labor productivity EU15 vs US 1.1 1 US = 1 0.9 0.8 0.7 1985 1990 1995 2000 2005 year source: EUKLEMS, Market secto
Slide 4: Divergence continues in recent data
Introduction Motivation Model and Calibration Stylized Facts Results Divergence continues in recent data 1996-2005 2006-2010 US 2.3 1.3 EU-15 1.5 .7 UK 2.3 .05 Source: The Conference Board Total Economy Database Bartelsman, Gautier, de Wind EPL and Technology Choice
Slide 5: Main explanations productivity divergence
Introduction Motivation Model and Calibration Stylized Facts Results Main explanations productivity divergence Macro/industry: EU productivity almost caught up to US by 1995. Since then, the gap has widened again Main explanation: In the US, ICT capital deepening and TFP increase likely associated w
Slide 6: Potential Storyline on ICT investment
Introduction Motivation Model and Calibration Stylized Facts Results Potential Storyline on ICT investment Investments in ICT and related intangibles increases the variance of firm outcomes ICT-related innovations require firm to experiment in the market Markets becomes more contestable and the inno
Slide 7: ICT and Increase in Performance Spread
Introduction Motivation Model and Calibration Stylized Facts Results ICT and Increase in Performance Spread Source: Brynjolfsson et al. Bartelsman, Gautier, de Wind EPL and Technology Choice
Slide 8: ICT and Increased Job Flows
Introduction Motivation Model and Calibration Stylized Facts Results ICT and Increased Job Flows Gross Job Flows Entry-Exit Job Flows 1986–94 1995–04 1986–94 1995–04 Hi ICT Inds 17.5 23.1 6.8 10.4 Lo ICT Inds 17.5 18.6 8.1 8.1 Source: Authors’ calculation from micro-aggregated file from LBD Bartelsm
Slide 9: An Aside: Micro-aggregated EU indicators
Introduction Motivation Model and Calibration Stylized Facts Results An Aside: Micro-aggregated EU indicators Distributed micro data research Researcher Policy Question Research Design Program Code Publication Metadata Network Cross-country Tables Network members Provision of metadata. Approval of a
Slide 10: Productivity Heterogeneity and Allocation
Introduction Motivation Model and Calibration Stylized Facts Results Productivity Heterogeneity and Allocation Manufacturing Excluding ICT, 2003-2009, Finland and Sweden. Labor productivity, Employment Share and Employment Growth by Productivity Quartile Bartelsman, Gautier, de Wind EPL and Technolo
Slide 11: Productivity Heterogeneity and Allocation
Introduction Motivation Model and Calibration Stylized Facts Results Productivity Heterogeneity and Allocation Manufacturing Excluding ICT, 2003-2009, Italy and the United Kingdom. Labor productivity, Employment Share and Employment Growth by Productivity Quartile Bartelsman, Gautier, de Wind EPL an
Slide 12: Productivity Heterogeneity and Allocation
Introduction Motivation Model and Calibration Stylized Facts Results Productivity Heterogeneity and Allocation Market Services Excluding ICT, 2003-2009, Finland and Sweden. Labor productivity, Employment Share and Employment Growth by Productivity Quartile Bartelsman, Gautier, de Wind EPL and Techno
Slide 13: ICT and Dispersion
Introduction Motivation Model and Calibration Stylized Facts Results ICT and Dispersion σc,i,t = α + γBBIc,i,t + FE + ε c,i,t (1) FE : country, industry, time fixed effects Std. Deviation of firm-level productivity distribution regressed on Broadband intensity Levels First-differences γ 0.47 .28 (5.
Slide 14: Employment Adjustment by ICT Intensity
Introduction Motivation Model and Calibration Stylized Facts Results ICT and Dispersion Employment Adjustment by ICT Intensity ‘Hockey Stick’ Chart, Data appear similar to real stuff Bartelsman, Gautier, de Wind EPL and Technology Choice
Slide 15: ICT and Dispersion
Introduction Motivation Model and Calibration Stylized Facts Results ICT and Dispersion Table: Measures of Dispersion by ICT intensity Prod. (TS) Output (TS) Prod. (XS) Output (XS) Country ICT=0 ICT=1 ICT=0 ICT=1 ALL ICT=1 ALL ICT=1 DK .037 .044 .057 .068 .23 .24 .29 .32 FI .036 .079 .043 .097 .25 .
Slide 16: Risky versus Safe Sectors, US vs EU
Introduction Motivation Model and Calibration Stylized Facts Results Risky versus Safe Sectors, US vs EU Employment share risky sector Relative productivity risky sector 0.6 2.1 US US EU EU 0.55 1.8 0.5 1.5 0.45 1.2 1990 1995 2000 2005 1990 1995 2000 2005 year year Source: EUKLEMS: ppp-adjusted outp
Slide 17: Risky versus Safe Sectors, EU by EPL
Introduction Motivation Model and Calibration Stylized Facts Results Risky versus Safe Sectors, EU by EPL Employment share risky sector Relative productivity risky sector 0.6 2.1 EU low EPL EU low EPL EU high EPL EU high EPL 0.55 1.8 0.5 1.5 0.45 1.2 1990 1995 2000 2005 1990 1995 2000 2005 year year
Slide 18: Selection of Related literature
Introduction Theory Model and Calibration Calibration and Simulation Results Selection of Related literature ICT and Productivity van Ark, O’Mahony, and Timmer (2008); Bloom, Sadun, Van Reenen (2012);Brynjolfsson, McAfee (2011) Employment Protection Autor, Kerr, and Kugler (2008); Samaniego (2006);
Slide 19: A model with firing cost and choice of technology: Setup
Introduction Theory Model and Calibration Calibration and Simulation Results A model with firing cost and choice of technology: Setup 2 technologies/ sectors compete for workers 0: safe sector, known technology and productivity (i.e. Pissarides, 2000) 1: risky sector, gets draws from prod. distribut
Slide 20: A model with firing cost and choice of technology: Setup
Introduction Theory Model and Calibration Calibration and Simulation Results A model with firing cost and choice of technology: Setup Firms pay a fixed fee (c0 , c1 ) to post a vacancy in either the safe or risky sector, until profit opportunities are exhausted In the safe sector, productivity, y ,
Slide 21: Vacancies and Jobs
Introduction Theory Model and Calibration Calibration and Simulation Results Vacancies and Jobs Value of vacancy m0 rV0 = −c0 + [J0 − V0 ] = 0 θ0 m1 rV1 = −c1 + [J1 (0) − V1 ] = 0. θ1 Value of a job for the firm rJ0 = y − w0 − δJ0 . R xu [J1 (z ) − J1 (x )] dF (z ) rJ1 (x ) = y + x − w1 (x ) − δJ1 +
Slide 22: Employment and unemployment
Introduction Theory Model and Calibration Calibration and Simulation Results Employment and unemployment Value of unemployment rU = b + m0 [max [0, E0 − U ]] + m1 [max [0, E1 (0) − U ]] . Value of employment rE0 = w0 − δ [E0 − U ] rE1 (x ) = w1 (x ) − (δ + λF (xd )) [E1 (x ) − U ] Z xu +λ [E1 (z ) −
Slide 23: Equilibrium
Introduction Theory Model and Calibration Calibration and Simulation Results Equilibrium Nash bargaining Free entry Beveridge curve, Steady state flow conditions Endogenous job destruction threshold Bartelsman, Gautier, de Wind EPL and Technology Choice
Slide 24: Equilibrium
Introduction Theory Model and Calibration Calibration and Simulation Results Equilibrium Free entry conditions (no arbitrage) Job creation condition sector 0, ( m0 θ0 ) m1 Job creation condition sector 1, ( θ1 ) Participation constraints: Job destruction sector 1: J1 (xd ) = −k Value of marginal job
Slide 25: Parameters from the literature
Introduction Theory Model and Calibration Calibration and Simulation Results Parameters from the literature Parm. Value Description Motivation y 1 productivity safe sector normalization r 0.004 monthly interest rate Pissarides (2009) β 1−η Nash bargaining share worker Hosios condition b 0.4 unemploy
Slide 26: Parameters matching US labor market stocks and flows
Introduction Theory Model and Calibration Calibration and Simulation Results Parameters matching US labor market stocks and flows Parm. Value Description Motivation l 0.77 size labor force size labor force (OECD LFS) δ 0.026 Poisson rate ex. job destr. ex. job destr. (JOLTS, EUKLEMS) c0 0.2092 vacan
Slide 27: Target output
Introduction Theory Model and Calibration Calibration and Simulation Results Target output yb is average output in the risky sector (1 − s ) workers have not received a shock and produce y Rx s workers produce on average y + 1−F1(x ) xdu zdF (z ) d solve for s using the steady state flow equation λ
Slide 28: Target variance and endogenous job destruction
Introduction Theory Model and Calibration Calibration and Simulation Results Target variance and endogenous job destruction The variance of output per worker in the risky sector is Z xu 1 σ2 = s (y + z − yb)2 dF (z ) + (1 − s ) (y − yb)2 1 − F (xd ) b xd Z xu Z x 2 ! λ λ u 2 = z dF (z ) − zdF (z ) .
Slide 29: Target risky sector job creation and destruction
Introduction Theory Model and Calibration Calibration and Simulation Results Target risky sector job creation and destruction λF (xd ) = s risky − δ. m1 (r + λ + δ) c1 = θ1 −(1 − β) (σxd + (r + λ + δ) k ) Bartelsman, Gautier, de Wind EPL and Technology Choice
Slide 30: Matching US cross-sectional productivity distribution
Introduction Theory Model and Calibration Calibration and Simulation Results Matching US cross-sectional productivity distribution Solve 4 equations (b σ2 , yb, mθ11 , λF (xd )) for 4 unknowns {λ, µ, σ, k } for median worker Parameter Value Description λ 0.1410 Poisson rate productivity shock µ 0.06
Slide 31: Increase in Riskiness, under different Firing Costs
Introduction Simulation Model and Calibration Empirical Results Results Increase in Riskiness, under different Firing Costs Share risky Productivity 1 1.2 benchmark benchmark high cf = 6 high cf = 6 0.8 1.15 0.6 1.1 0.4 1.05 0.2 1 0 0.95 0.2 0.3 0.4 0.5 0.6 0.2 0.3 0.4 0.5 0.6 σ σ Bartelsman, Gautie
Slide 32: Model Simulation: Comparative Statics
Introduction Simulation Model and Calibration Empirical Results Results Model Simulation: Comparative Statics Benchmark σ = 0.50 High σ = 0.75 k = 1.25 k = 3 k = 7 k = 1.25 k = 3 k = 7 xd -0.729 -1.010 -1.688 -0.781 -1.048 -1.693 λF (xd ) 0.008 0.002 0.000 0.018 0.010 0.001 ŷ 1.100 1.071 1.055 1.18
Slide 33: Cross country / industry Regressions
Introduction Simulation Model and Calibration Empirical Results Results Cross country / industry Regressions Employment share, employment growth, or productivity is regressed on industry riskiness indicator interacted with exit costs depc,i,t = α + βkc,t + γkc,t R (σ)i + FE + ε c,i,t X Where kc,t is
Slide 34: Regression Results
Introduction Simulation Model and Calibration Empirical Results Results Regression Results Labor share Labor growth TFP growth γ −1.08 −.86 −1.12 (2.94) (3.00) (2.48) R2 .84 .20 .07 D.F. 5025 4979 4396 Fixed effects ind mean+trends ctry,ind,time ctry,ind,time t-statistic in parenthesis. Period: 1995
Slide 35: Robustness γ: by sub-period
Introduction Simulation Model and Calibration Empirical Results Results Robustness γ: by sub-period sub-period Sample 1995-2000 2000-2005 1995-2005 EU −.71 −.84 −.77 (5.04) (5.49) (7.15) EUN −0.87 −.94 −0.91 (6.72) (6.89) (9.27) EURO −.63 −.63 −.62 (3.32) (7.15) (4.62) OECD −0.85 −0.94 −0.89 (7.85)
Slide 36: Robustness γ: by riskiness and exit cost
Introduction Simulation Model and Calibration Empirical Results Results Robustness γ: by riskiness and exit cost Industry sub-sample Country sub-sample Low risk High risk Low firing cost −2.27 −4.39 (2.71) (3.62) High firing cost −.98 −2.71 (.94) (2.81) t-statistic in parenthesis. Period: 1995-2005;
Slide 37: Robustness γ: by k and R(σ)
Introduction Simulation Model and Calibration Empirical Results Results Robustness γ: by k and R (σ) Riskiness indicator Exit Cost DSLpct P4/P Variance Exitloss% −4.99 −3.55 −3.16 (2.63) (1.83) (1.67) Exitcost% −22.76 −16.20 −13.03 (3.00) (1.87) (1.62) Firerule −.42 −.36 −.32 (.66) (.70) (.69) Firec
Slide 38: Employment Share: Entry Costs vs Exit Costs
Introduction Simulation Model and Calibration Empirical Results Results Employment Share: Entry Costs vs Exit Costs Entry Cost Indicator only γentry γentry γexit Starting a Business - # of procedures −.20 −.09 −.94 (1.64) (.97) (2.85) Starting a Business - time (days) −.17 −.11 −.76 (2.27) (1.37) (1
Slide 39: Robustness γ: by random R(σ)
Introduction Simulation Model and Calibration Empirical Results Results Robustness γ: by random R (σ) Random rankings 1.5 γ 95% CI 1 0.5 0 −0.5 −1 −1.5 0 200 400 600 800 1000 1200 Estimates of γ and confidence interval from 1200 regressions using random draws from all possible permutations of indust
Slide 40: Long term effects of the crisis? Productivity
Introduction Simulation Model and Calibration Empirical Results Results Long term effects of the crisis? Productivity Employment share risky sector Relative productivity risky sector 1 1.2 0.8 1.15 0.6 1.1 0.4 1.05 0.2 1 0 0.95 0.3 0.4 0.5 0.6 0.7 0.8 0.3 0.4 0.5 0.6 0.7 0.8 c1 c1 Bartelsman, Gautie
Banque Nationale De Belgique — presentatie bartelsman on 6ix