Job Polarization and Structural Change

Size: px

Start display at page:

Download "Job Polarization and Structural Change"

Wilfred Todd
5 years ago
Views:

1 Job Polarization and Structural Change Zsófia L. Bárány and Christian Siegel Sciences Po University of Exeter October 2015 Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 1

2 Introduction Job polarization is a widely documented phenomenon in developed countries since the 1980s: employment has been shifting from middle to low- and high-income occupations average wage growth has been slower for middle-income occupations than at both extremes Main explanation: routinization; ICT substituting for middle-skill occs In this paper 1 we document a set of facts ICT routinization is not the sole driving force behind this phenomenon 2 based on these facts we propose a novel perspective on the polarization of labor markets one based on structural change (reallocation across sectors) Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 2

3 Roadmap 1 Empirical evidence 2 Model 3 Quantitative Results Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 3

4 Two new facts, plus one 1. polarization started as early as 1950/ it is present across broadly defined sectors +1 between industry shifts important for occupational employment Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 4

5 Polarization in terms of occupations 30-Yr Change in ln(real wage) Change in Log Wage Occupation's Percentile in 1980 Wage Distribution Yr Change in Employment Share Change in Employment Share Occupation's Percentile in 1980 Wage Distribution Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 5

6 Polarization in broad occupational categories Change in Log Median Wage Median Log Wage Change in Employment Share Median Log Wage epanechnikov kernel fractional polynomial epanechnikov kernel fractional polynomial Change in Log Median Wage Change in Employment Share Median Log Wage Median Log Wage 1980 epanechnikov kernel fractional polynomial epanechnikov kernel fractional polynomial classification Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 6

7 Polarization for broad occupations a m Relative average wages compared to routine jobs Year manual abstract Share in Employment r a m Employment shares of occupations Year manual routine abstract classification Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 7

8 Two new facts, plus one 1. polarization started as early as 1950/ it is present across broadly defined sectors +1 between industry shifts important for occupational employment Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 8

9 Polarization for broad industries splitting services in two driven by production & consumption side S L Relative residual wages compared to manufacturing jobs year low-skilled services high-skilled services Share in Employment M S L Employment shares of industries Year low-skilled serv. manufacturing high-skilled serv. classification regression gender and age effects Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 9

10 Two new facts, plus one 1. polarization started as early as 1950/ it is present across broadly defined sectors +1 between industry shifts important for occupational employment Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 10

11 Shift-share decomposition E ot = λ oi E it + λ oit E i, i i }{{}}{{} Eot B Eot W Decompose the change in an occupation s employment share to a between industry component a within industry component Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 11

12 Shift-share Decomposition of Employment Shares 3 x 3 10 x Manual Total Between Within Routine Total Between Within Abstract Total Between Within Average Total Between Within only within shifts Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 12

13 Summary of key observations 1. polarization started as early as 1950/1960 before ICT or increased trade 2. it is present across broadly defined sectors: low-skilled services, manufacturing, high-skilled services +1 between industry shifts important for occupational employment Observing that 1 polarization seems to be a long-run phenomenon 2 middle earning jobs are in manufacturing 3 the structural shift from manufacturing to services started in the s structural shift of the economy might be the driving force behind the polarization of the labor market how much of the polarization of sectors can a (parsimonious) model of structural change explain? Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 13

14 Roadmap 1 Empirical evidence 2 Model 3 Quantitative Results Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 14

15 Model Overview the most parsimonious model that allows for the joint analysis of wages and employment multi-sector growth model, similar to Ngai & Pissarides (2007) with a Roy-type self-selection mechanism: workers with heterogeneous skills optimally select sector of work three sectors: manufacturing, low- and high-skilled services as goods and services are complements in consumption, when relative manufacturing productivity increases: labor reallocates to both service sectors wages in expanding sectors have to increase manufacturing jobs tend to be in the middle polarization pattern Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 15

16 Production - perfect competition Low-skilled service goods Y l = A l L l ω l = p l A l Manufacturing goods Y m = A m N m ω m = p m A m N m efficiency units of labor ω m wage per efficiency unit of labor High-skilled service goods Y s = A s N s ω s = p s A s Note: * in producing M and S efficiency units of labor matter income of someone with a efficiency units in M/S is aω m /aω s * in L raw labor is used and income is ω l if working in L Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 16

17 Labor Supply every individual works full time in one of the market sectors continuum of different types heterogeneity in innate ability (a m, a s ) R 2 + for simplicity assume: am : efficiency units of labor in M earn a m ω m if in M as : efficiency units of labor in S earn a s ω s if in S all individuals equally productive in L earn ωl if in L each agent chooses, given ability, the sector that provides the highest income Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 17

18 Sector of work decision: endogenous sorting Optimal sector choice characterized by two cutoff values: a s â m ω l ω m and â s ω l ω s â s â m a m S â s M L â m a m Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 18

19 Demand The stand-in household solves: ( [ θ l C ε 1 ε l max u C l,c m,c s s.t. + θ m C ε 1 ε m ] + θ s C ε 1 ε ) ε 1 ε s p l C l + p m C m + p s C s ω l L l + ω m N m + ω s N s The household s optimal consumption bundle has to satisfy: ( C l pl = C m C s C m = ) ε θ m, p m θ l ( ) ε ps θ m. p m θ s Unbalanced technological progress Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 19

20 Structural change Proposition When manufacturing goods and the two types of services are complements (ε < 1), then faster productivity growth in manufacturing than in both types of services (da m /A m > da s /A s = da l /A l ), leads to a change in the optimal sorting of individuals across sectors. In particular â m = ω l /ω m and â m /â s = ω s /ω m unambiguously increase, â s = ω l /ω s can rise or fall. This leads to an increase in employment in L, an increase in efficiency labor in S, a fall in effective and raw labor in M. Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 20

21 Structural change relative average wages Low-skilled service relative to manufacturing: w l w m = ω l ω mn m L m = ω l ω m 1 N m L m = âm a m. High-skilled service relative to manufacturing: w s w m = ω sn s L s ω mn m L m = ω s ω m N s L s N m L m = âm â s a s a m. rel. value added Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 21

22 Roadmap 1 Empirical evidence 2 Model 3 Quantitative Results Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 22

23 Calibration Strategy data targets (US Census/ACS): we categorize workers as low-skilled service, manufacturing, or high-skilled service based on their industry code (ind1990) four key moments of the US economy in 1960 sectoral employment shares (in terms of hours worked) relative average sectoral wages all parameters are time-invariant, chosen to match 1960 Details incl. ability distribution (a bivariate uniform distribution) Details only exogenous change over time is productivity growth akin Ngai and Petrongolo (2014), calculate labor productivity by dividing sectoral value added output data from Herrendorf, Rogerson, Valentinyi (2013) with sectoral employment data from the BEA due to data limitations we cannot break the labor productivity growth of services into low- and high-skilled possibilities: raw/adjusted, average/decennial Details on Adjustment Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 23

24 Calibrated parameters Description Value [ã m, ã m ] range of manufacturing efficiency [0.40, 1.60] [ã s, ã s ] range of high-skilled service efficiency [0.02, 1.98] ε CES b/w L, M and S in consumption τ l relative weight on M 0.49 τ s relative weight on S 0.91 Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 24

25 Annual average labor productivity growth Based on raw labor Adjusted by average efficiency Manufacturing Services Manufacturing Services Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 25

26 Transition under baseline productivity growth Index of sectoral productivities Sector of work cutoffs A m âm A l, A s M 1.2 S 1.15 âs L, S M Employment shares: data vs model 0.45 M S 0.3 Relative average wages: data vs model wl/wm 1.2 ws/wm S 0.9 L 0.25 L l L m L s L value addded Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 26

27 Transition under selection-adj. productivity growth M S L 0.25 Employment shares: data vs model L l L m L s Relative average wages: data vs model S L w l/w m w s/w m Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 27

28 Transition under decennial productivity growth Index of sectoral productivities Sector of work cutoffs A m A l, A s M L, S S M âm âs Employment shares: data vs model 0.45 M S L 0.25 L l L m L s Relative average wages: data vs model S L wl/wm ws/wm Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 28

29 Summary from data 1. polarization started as early as polarization present also across sectors 3. between industry shifts important for occupational employment structural change possible force driving polarization the model introduce heterogeneous labor via Roy-type selection into a multi-sector growth model unbalanced technological change affects not only employment and expenditure shares, but also sectoral average wages if productivity growth is highest in jobs that are in middle of the distribution, it leads to polarization quantitatively, simple model does very well over the last 50 years around 70% of the relative average wage gain of high- and low-skilled services compared to manufacturing about 75% of changes in employment shares Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 29

30 Thank you! Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 30

31 Occupation categories The 10 occupational codes are: 1. personal care; 2. food and cleaning services; 3. protective services; 4. operators, fabricators and laborers; 5. production, construction trades, extractive and precision production; 6. administrative and support occupations; 7. sales; 8. technicians and related support occupations; 9. professional specialty occupations; 10. managers. back Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 31

32 1 Manual: low-skilled non-routine housekeeping, cleaning, protective service, food prep and service, building, grounds cleaning, maintenance, personal appearance, recreation and hospitality, child care workers, personal care, service, healthcare support 2 Routine construction trades, extractive, machine operators, assemblers, inspectors, mechanics and repairers, precision production, transportation and material moving occupations, sales, administrative support, sales, administrative support sales, administrative support 3 Abstract: skilled non-routine managers, management related, professional specialty, technicians and related support back Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 32

33 3 High-skilled services: professional and related services, finance, insurance and real estate, communications, high-skilled business services (advertising, services to dwellings and other buildings, personnel supply services, computer and data processing services, detective and protective services, business services not elsewhere classified), communications, utilities, high-skilled transport (railroads, U.S. Postal Service, water transportation, air transportation), public administration Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 33 Industry classification 1 Low-skilled services: personal services, entertainment, low-skilled transport (bus service and urban transit, taxicab service, trucking service, warehousing and storage, services incidental to transportation), low-skilled business and repair services (automotive rental and leasing, automobile parking and carwashes, automotive repair and related services, electrical repair shops, miscellaneous repair services), retail trade, wholesale trade 2 Manufacturing: mining, construction, manufacturing

34 Regression of log hourly wages: industry effects low s (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) high s (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Obs R Standard errors in parentheses p < 0.05, p < 0.01, p < controls: a polynomial in potential experience (defined as age - years of schooling - 6), dummies for gender, race, and born abroad back Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 34

35 Descriptive statistics low-sk. serv. manuf. high-sk. serv. Highschool Dropout 20.66% 27.54% 8.27% Highschool Graduate 36.76% 37.57% 24.36% Some College 28.33% 21.19% 29.05% College Degree 11.20% 10.37% 23.00% Postgraduate 3.05% 3.34% 15.32% Mean Years of Education Female Share 44.35% 23.33% 51.37% Foreign-Born Share 12.05% 11.21% 8.97% back Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 35

36 Gender and age effects in employment shares Counterfactual exercise: only changes in the gender-age composition of the labor force Share in Employment M S L Year L actual M actual S actual L cntrf M cntrf S cntrf fix industry shares of gender-age cells at their 1960 level, let age and gender shares to follow their actual path back Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 36

37 How important are within-industry shifts in occ shares? fix industry shares at 1960 level, let within-ind occ shares follow the actual path Share in Employment r a m Year manual routine abstract cntrf cntrf cntrf back Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 37

38 Structural change Using market clearing conditions in household demands: A l A m L ( l ωl = N m A s N ( s ωs = A m N m ω m A m A l }{{} =p l /p m ω m A m A s }{{} =p s/p m θ m θ l ) ε, θ m θ s ) ε. a change in relative productivities has two direct effects: on supply and demand Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 38

39 Structural change Using optimal sector of work cutoffs: L l (â m, â s ) )âε m = N m (â m, â s ) ε N s (â m, â s ) (âm = N m (â m, â s ) â s ( Am A l ( Am A s ) 1 ε ( ) ε θm, (1) θ l ) 1 ε ( ) ε θm. (2) θ s These two equations implicitly define â m, â s, which fully characterize the equilibrium. back Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 39

40 Structural change optimal sorting a s â s â m a m â s â m a m a s â s â m a m â s â m a m â s â s â s â s â m â m a m â m â m a m back Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 40

41 Structural change relative value added Proposition When manufacturing goods and the two types of services are complements (ε < 1), then faster productivity growth in manufacturing than in both types of services (da m /A m > da s /A s = da l /A l ), increases the relative value added in both high- and low-skilled services compared to manufacturing: d p sy s p m Y m > 0 and d p ly l p m Y m > 0. Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 41

42 Structural change relative value added Since p i Y i = p i A i N i = ω i N i, relative value added shares can be expressed as: p s Y s = ω s N s = âm p m Y m ω m N m â s p l Y l = ω l p m Y m ω m N s N m, L l N m = â m L l N m. Moreover, since ω i N i = w i L i, relative VA can be expressed as: p i Y i p j Y j = w i w j L i L j. back Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 42

43 Calibration of utility function and initial productivities parameters of the utility function: ε, θ l, θ m, θ s initial productivities: A l (0), A m (0), A s (0) back take ε, the elasticity of substitution from the literature ε estimated by Herrendorf, Rogerson, Valentinyi (2013); when sectoral output is measured in value added terms, ε = Ngai and Pissarides (2008) find that plausible estimates are in the range [0, 3] ) 1 ε ( ) ε ) 1 ε ( ) ε θ m θl and θ τs m θs calibrate τ l ( Am(0) A l (0) ( Am(0) A s(0) to match 1960 relative average wages and employment shares Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 43

44 Calibration of ability distribution assume f (a m, a s ) is uniform (which requires a minimal choice of parameters) normalize (w.o.l.g.) the mean of a m and a s to be unity (not separately identified) need to find a m and a s calibrate these such that the observed employment shares and relative average wages are consistent with each other given f (a m, a s ), the observed labor shares uniquely identify the sector-of-work cutoffs, the sector-of-work cutoffs in turn imply relative average wages pin down a m, a s such that when matching the raw employment shares in 1960, the model also matches the relative average wages still have to calibrate other parameters to ensure that in equilibrium we are matching these moments in the first place back Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 44

45 Adjustment for average labor efficiency changes due to the self-selection of individuals into sectors expanding sectors increase by soaking up relatively less efficient workers contracting sectors decrease by shedding relatively less efficient workers average efficiency of labor in expanding sectors fall, while in contracting sectors it increases manufacturing productivity growth might be overestimated; services productivity growth might be underestimated when calculating from raw employment data pointed out in the context of measuring productivity growth across sectors by Young (2014 AER), estimated for the bias in skill premium estimates by Carneiro and Lee (2011 AER) Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 45

46 Productivity growth adjustment based on our calibration use calibration for efficiency distribution, f (a m, a s ) take raw employment shares from the data given cutoff structure in our model calculate the change in average labor efficiency in each sector overall efficiency gain in manufacturing: 4.8% overall efficiency loss in services: 3.4% adjust the annual change in raw employment by calculated annual labor efficiency gain/loss in the sector adjusted labor productivity growth back Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 46

47 Value added shares 1 Relative manufacturing value added: data vs model back Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 47

Job Polarization and Structural Change

Job Polarization and Structural Change Zsófia L. Bárány and Christian Siegel Sciences Po University of Exeter May 2015 Bárány and Siegel (Sciences Po, Exeter) Job Polarization and Structural Change 1 Introduction