THE SUBSTITUTION OF ICT CAPITAL FOR ROUTINE LABOR: TRANSITIONAL DYNAMICS AND LONG-RUN IMPLICATIONS

Transcription

1 THE SUBSTITUTION OF ICT CAPITAL FOR ROUTINE LABOR: TRANSITIONAL DYNAMICS AND LONG-RUN IMPLICATIONS Maya Eden (The World Bank) Paul Gaggl (UNC Charlotte)

2 Disclaimer 2! This paper reflects our own views and not necessarily those of the World Bank, its Executive Directors or the countries they represent.

3 Computers: Many Questions 3! Will they replace many workers?! What will happen to the labor share?! What are the limits to computerization?! What is their effect on the standard of Living?

4 Focus: One Narrow Channel 4! ICT: Information and Communication Technology! The substitution of ICT for routine labor! Routine Labor! Can, in principle, be performed by ICT! Examples! Routine: Factory worker, typist, secretary,! Non-Routine: CEO, Cab driver, janitor, nanny

5 Outline of Agenda 5! Document trends in income shares! Use trends to estimate production structure! Embed in standard growth model! Conduct counterfactual simulations

6 Labor s Income Share 6 Income Shares (%) CPS (R+NR) BLS: Non Farm Business NIPA: Compensation of Employees (BEA:A033RC0) NIPA: Wage & Salary Accruals (BEA:A034RC0) 1980q1 1985q1 1990q1 1995q1 2000q1 2005q1 2010q1 2015q1 Quarter : CPS MORG ( ) & authors computations

7 Routine vs. Non-Routine Income Share 7 (A) Labor s Income Share Income Shares (%) Routine Income Share (SA) Non Routine Income Share (SA) 1980q1 1985q1 1990q1 1995q1 2000q1 2005q1 2010q1 2015q1 Quarter : CPS MORG ( ) & authors computations

8 Who are routine workers? 8 (A) Labor s Income Share Income Shares (%) Routine Income Share (SA) Non Routine Income Share (SA) 1980q1 1985q1 1990q1 1995q1 2000q1 2005q1 2010q1 2015q1 Quarter : CPS MORG ( ) & authors computations! Routine Jobs:! Production! Operative! Clerical! Support! Sales! Non-Routine Jobs:! Professional! Managerial! Technical! Service

9 Price-Quantity Decomposition 11! What drives these trends?! Labor Share: s = wl/y! Price (wages)?! Quantity (employment)?

10 ensure that this wedge is not simply driven by a changing composition of characteristics 12 Non-Routine Wage Premium tine and non-routine workers, we estimate the following set of cross-sectional wage regressio arately for each quarter, q: ln w i,q = 0,q + 1,q NR i,q + 2,q X i,q + i,q for q 2 {1979q1,...,2013q4}, (1 (A) Routine vs. Non-Routine Wages (B) Non-Routine Wage Premium Log Real Wage Routine Non Routine 1980m1 1985m1 1990m1 1995m1 2000m1 2005m1 2010m1 2015m1 Month 10 Non Routine Wage Premium (%) s.e. Non Routine Wage Premium Trend 1980q1 1985q1 1990q1 1995q1 2000q1 2005q1 2010q1 2015q1 Quarter Notes: Panel A plots the unconditional quarterly mean real wage in each occupation group. Panel B graphs the coefficients from quarterly regressions of individual level real wages on a non-routine dummy and a host of demographic control variables including flexible: functional CPS forms MORG in industry, age, and education. Occupation and individual specific wages are based on the monthly CPS merged outgoing rotation group (MORG) extracts provided by the NBER for the period 1979m1-2013m12. We deflate wage data

11 Price-Quantity Decomposition 14! Relative price AND quantity of NR increase! Non Routine Wage Premium (%) s.e. Non Routine Wage Premium Trend 1980q1 1985q1 1990q1 1995q1 2000q1 2005q1 2010q1 2015q1 Quarter NR/R Employment (%) NR/R Employment: Effective NR/R Employment: CPS in each occupation group. Panel B graphs the coefficients from e dummy and a host of demographic control variables including on and : individual CPS specific MORG wages & authors are basedcalculations the monthly CPS NBER for the period 1979m1-2013m12. We deflate wage data

12 Capital Income Share 15 Share of Income (%) Non Residential: ICT Non Residential: Non ICT Residential & Consumer Durables : BEA detailed asset accounts & author s computations

13 Capital Income Share 16 s c,t = MPK c,t K c,t Y t = s n,t = MPK n,t K n,t Y t s h,t = MPK h,t K h,t Y t Share of Income (%) Non Residential: ICT Non Residential: Non ICT Residential & Consumer Durables : BEA detailed asset accounts & author s computations

14 Measurement Strategy 17! How should we define ICT?! How can we measure income shares?! CRS:! Return on a unit of capital type i:! At profit-maximizing choice: Needs to equal 1+r!? s K,t = MPK c,t K c,t Y t + MPK n,t K n,t Y t + MPK h,t K h,t Y t =1 s L,t P t MPK i,t + P i,t+1 (1! BEA s estimates from detailed asset accounts P i,t i,t)

15 Information & Communication Technology (ICT) 18 Share of Aggregate Capital (%) Average Growth in Share (%) ICT Assets EP20: Communications ENS3: Own account software ENS2: Custom software EP34: Nonelectro medical instruments EP36: Nonmedical instruments ENS1: Prepackaged software EP35: Electro medical instruments EP1B: PCs RD23: Semiconductor and other component manufacturing RD22: Communications equipment manufacturing EP31: Photocopy and related equipment EP1A: Mainframes EP1H: System integrators RD24: Navigational and other instruments manufacturing EP1D: Printers EP1E: Terminals EP1G: Storage devices EP12: Office and accounting equipment RD40: Software publishers RD21: Computers and peripheral equipment manufacturing RD25: Other computer and electronic manufacturing, n.e.c EP1C: DASDs EP1F: Tape drives Notes: The data are drawn from the BEA. Assets are ranked by their average share in aggregate capital during

16 Depreciation & Relative Price of ICT 19! Implicit Price Deflators for capital! BEA s estimate of depreciation Price Index (1979=1) Non ICT Capital Price (Deflator, 1979=1) ICT Capital Price (Deflator, 1979=1) NIPA: GDP (Deflator, 1979=1) NIPA: Consumption (Deflator, 1979=1) Month Annual Depreciation Rate (%) Non Residential: ICT Non Residential: Non ICT Residential & Consumer Durables : BEA detailed fixed asset accounts & author s computations

17 21 Chain Index (1979=1) following expressions:! Neoclassical dynamics!! ICT Quantity (index) increases! ICT Price (rental rate) declines (A) Capital Stock Relative to 1979 ICT Capital Non ICT Capital after the 1982 recession. Panel B of Figure 6 graphs the respec Price-Quantity ofdecomposition capital. These estimates then allow us to compute the income sha : BEA detailed fixed asset accounts & author s computations Marginal Product (% Output Units) s i,t = MPK i,t K i,t Y t = (B) Marginal Product of Capital Non Residential: ICT Non Residential: Non ICT Residential & Consumer Durables Pt MPK i,t P i,t for each capital type i 2 {c, n, h}. Notice that the first brack responds to the respective RHS expressions of equations (9) contains nominal shares. We can therefore compute the cap nominal values at current cost, as well as a measure of inflati Figure 1 illustrates the resulting estimates for these inco the income shares of non-ict capital do not show any signi

18 Production Structure 22! Production Structure Y t = (A nc,t K nc,t ) nc,t (A l,t L nr,t ) nr,t X 1 nc,t nr,t t X t = ( (A c,t K c,t ) +(1 )(A l,t L r,t ) ) 1! Exogenous factor intensities! Trade, complementarity with ICT,

19 Estimation Strategy 24! Equilibrium Relation ln sr,t s c,t 1 = ln! Assumptions on A: + ln Al,t A c,t + ln Lr,t K c,t A i,t A c,0 =1 = A 0,i e i t+ i,t

20 GMM Estimates 28 Exogenous Technical Progress A. Labor Augmenting B. Labor & Capital Augmenting C. Labor & ICT Augmenting 0.191*** 0.190*** 0.190*** (0.0104) (0.0105) (0.0103) 0.470*** 0.465*** 0.468*** (0.0498) (0.0513) (0.0494) ln(a l,0 ) *** *** *** (0.0664) (0.0691) (0.0663) l ** ** ** ( ) ( ) ( ) k Implied Elasticity of Substitution ( ) ( ) Lower 95% CI Point Estimate Upper 95% CI Obs Notes: The estimates are based on a Newey-West HAC robust weighting matrix and Newey-West standard errors are reported in parentheses. Significance levels are indicated by * p<0.1, ** p<0.05, and *** p<0.01. For each specification we also report the implied elasticity of substitution, " c,r =1/(1 ), evaluated at the respective point estimate, lower, and upper 95% confidence interval.

21 29 Neoclassical Growth Model max c t,k nc,t+1,k c,t+1,l r,t,l nr,t,l s r,t,ls nr,t 1X t=0 X t (A tc t ) 1 1 c t +(1+ )(1 + p ) X such that p i,t k i,t+1 = t +(1+ t )w r,t lr,t s + w nr,t lnr,t s + T t i=nc,c t = y t + X i=nc,c (1 i,t)p i,t k i,t w r,t l r,t w nr,t l nr,t y t = k nc,t nc,t l nr,t nr,t ( k c,t +(1 )l r,t ) 1 nr,t nc,t Units: per effective worker (1/AL) l s r,t + l s nr,t =1 A t = A 0 e t

22 Model Parameters 30 A. Fixed Parameters & Initial Values IES parameter 2 Discount factor 1/1.03 p Population growth k nc,0 Initial non-ict capital stock k r,0 Initial ICT capital stock B. Long-Run Values nc,1 Depreciation rate: non-ict capital r,1 Depreciation rate: ICT capital nr,1 Non-routine labor share nc,1 Non-ICT capital share p 1 ICT capital price Non-routine wage premium C. GMM Estimates CES scale parameter CES substitution parameter Growth rate of labor-augmenting productivity

23 Exogenous Paths 31 1 (A) Relative ICT Price (p c,t ) (B) ICT Capital Depreciation ( c,t ) p 0.6 deltar (C) Non-ICT Capital Intensity ( nc,t ) (D) NR Wage Premium ( t ) alphak nrp

24 Simulation Figure 8: Simulated Paths of Income Shares () (A.1) Routine Labor (A.2) Non-ICT Captial (A.3) ICT Captial lr slr Simulation kn Simulation kr Simulation (B.1) Routine Share (B.2) Labor Share (B.3) ICT Share Simulation 0.22 sl Simulation skr (C.1) Wages (C.2) Consumption (C.3) Output Simulation Simulation Simulation w c 0.3 y Routine: Simulation Routine: Non Routine: Simulation Non Routine:

25 : No Figure Labor 11: : Wedge No Labor Market Frictions (A.1) Routine Labor (A.2) Non-ICT Captial (A.3) ICT Captial lr slr kn kr (B.1) Routine Share (B.2) Labor Share (B.3) ICT Share sl skr (C.1) Wages (C.2) Consumption (C.3) Output w c y Routine: Routine: Non Routine: Non routine:

26 : Fixed Pattern of Specialization (A.1) Routine Labor (A.2) Non-ICT Captial (A.3) ICT Captial lr slr kn kr (B.1) Routine Share (B.2) Labor Share (B.3) ICT Share 0.22 sl skr (C.1) Wages (C.2) Consumption (C.3) Output w c y Routine: Routine: Non Routine: Non routine:

27 : Fixed ICT stock/ict Price (A.1) Routine Labor (A.2) Non-ICT Captial (A.3) ICT Captial lr slr Fixed ICT Price Fixed ICT kn Fixed ICT Price Fixed ICT kr Fixed ICT Price Fixed ICT (B.1) Routine Share (B.2) Labor Share (B.3) ICT Share Fixed ICT Price Fixed ICT 0.22 sl Fixed ICT Price Fixed ICT skr Fixed ICT Price Fixed ICT (C.1) Wages (C.2) Consumption (C.3) Output 0.65 w c y Fixed ICT Price Fixed ICT Routine: Routine: Non Routine: Non routine: 0.45 Fixed ICT Price Fixed ICT

28 Conclusions 36! Decline in Labor Share: due to ICT! Decline in Routine Labor: due to increase in routine intensity! Gains from pure ICT accumulation exhausted?

29 THANK YOU!!