digital technologies: economic growth and the future of work · digital technologies: economic...
TRANSCRIPT
Digital Technologies: Economic Growth
and the Future of Work
OECD/BEIS Conference, November 2018
John Van Reenen
Massachusetts Institute of Technology
Productivity
Jobs and skills
Labor Share & Superstar Firms
Policy implications
What are the new digital technologies?
OUTLINE OF TALK
3
INDUSTRIAL REVOLUTIONS
• First Industrial Revolution: 1760-1840
• Second Industrial Revolution: 1870-1914
• Third Industrial Revolution: 1996-2004; Digital
• Fourth Industrial Revolution: ???
4
DIGITAL INDUSTRIAL REVOLUTION POWERED BY MOORE’S LAW
Stable 35% p.a growth in semiconductor productivity required 18x
growth in # researchers
Source: Bloom, Jones, Van Reenen & Webb (2017)
0
1
0
1
1
0
1
1
1
0
1
1
1
1
0
0
1
0
1
0
0
1
0
1
1
1
1
1
0
0
1
0
1
0
1
0
1
0
1
0
0
1
0
1
0
0
0
0
1
0
1
1
0
1
1
1
0
1
1
1
1
0
0
1
0
1
0
0
0
1
1
1
1
1
0
0
0
0
1
1
1
0
1
0
1
0
0
1
0
1
0
0
1
0
1
1
0
1
1
1
0
1
0
1
1
0
1
1
1
1
0
0
1
0
1
1
1
1
1
0
0
1
1
1
1
0
1
0
1
0
0
1
0
1
0
1
0
0
0
1
0
1
1
0
0
1
1
0
1
1
1
1
0
0
1
0
1
0
0
0
1
1
1
1
1
0
0
1
0
1
0
1
0
1
0
1
0
0
1
0
1
0
0
1
1
1
1
1
0
0
1
0
1
0
1
0
1
0
1
0
0
1
0
1
0
0
1
0
1
1
0
1
1
1
0
1
1
1
1
0
0
1
0
1
0
0
1
0
1
1
1
1
1
0
0
1
0
1
0
1
0
1
0
1
1
1
1
0
1
0
0
0
0
1
0
1
1
1
1
1
1
0
1
1
1
1
0
0
1
0
1
0
0
0
1
1
1
1
1
0
0
0
1
1
1
1
0
1
0
1
0
0
1
0
1
0
0
1
0
1
1
1
1
1
1
0
1
0
1
1
0
0
1
0
1
0
0
1
0
1
1
1
1
1
0
0
1
1
1
1
1
0
1
0
1
0
0
1
0
1
0
0
0
0
1
0
1
1
0
0
1
1
1
1
1
1
1
0
0
1
0
1
0
0
0
1
1
1
1
1
0
0
1
0
1
1
1
1
1
0
1
0
0
1
0
1
0
0
1
1
1
1
1
0
0
1
0
1
0
1
0
1
1
1
1
1
0
1
0
1
0
1
0
1
1
0
1
1
1
0
1
1
1
1
0
0
1
0
1
1
1
1
0
1
1
1
1
1
0
0
1
0
1
0
1
0
1
0
1
0
0
1
0
1
0
0
0
0
1
0
1
1
0
1
1
1
1
1
1
1
1
0
0
1
0
1
0
0
0
1
1
1
1
1
0
0
0
0
1
1
1
0
1
0
1
1
1
1
0
1
0
0
1
0
1
1
1
1
1
1
0
1
0
1
1
0
0
1
0
1
0
0
1
0
1
1
1
1
1
0
0
1
0
1
1
1
1
1
0
1
0
0
1
0
1
0
0
0
0
1
0
1
1
0
0
1
1
0
1
1
1
1
0
0
1
0
1
0
0
0
1
1
1
1
1
0
0
1
1
1
0
1
0
1
0
1
0
0
1
0
1
0
0
1
1
1
1
1
0
0
1
0
1
0
1
0
1
0
1
0
0
1
1
1
1
0
1
1
1
1
1
0
1
1
1
1
0
1
0
1
0
1
0
0
1
0
1
0
0
0
0
1
0
1
1
0
0
1
1
0
1
1
1
1
0
0
1
0
1
0
0
0
1
1
1
1
1
0
0
1
0
1
0
1
0
1
0
1
0
1
1
1
1
0
0
1
0
1
1
1
1
1
1
0
1
0
1
1
0
0
1
0
1
0
0
1
Digital
7
INDUSTRIAL REVOLUTIONS
• First Industrial Revolution: 1760-1840
• Second Industrial Revolution: 1870-1914
• Third Industrial Revolution: 1996-2004; Digital
• Fourth Industrial Revolution: ???
“Heavy Truck Drivers” is the 13th largest U.S. occupation. Light Trucks is 37th. >4 million driving-related jobs in the U.S. in 2014.
Productivity
Jobs and skills
Labor Share & Superstar Firms
Policy implications
What are the new digital technologies?
OUTLINE OF TALK
Growth Rate of GDP per capita, G7, 1976-2016
Note: Annual average over decade; G7 = Canada, France, Germany, Italy,
Japan, UK and US
Source: OECD (2017) http://stats.oecd.org/Index.aspx?DataSetCode=PDB_LV#
0
0.5
1
1.5
2
2.5
3
1976-1986 1986-1996 1996-2006 2006-2016
US TFP Productivity Growth (“Frontier”) weak in last decade
Note: Total Factor Productivity (TFP); Annual average growth over different
periods
Source: Fernald (2016)
0
0.5
1
1.5
2
2.5
1947-1973 1974-1995 1996-2004 2005-2016
Large Literature looking at impact of adopting digital
technologies at firm level
• Case Studies
– Fascinating, but hard to generalize
• Statistical evidence
– Look at firm performance (productivity, profitability,
growth, etc.) before and after introduction of technology
– Control for other factors that could generate spurious
correlation (industry, area, other investments, etc.)
– Always issue that purely experimental variation is rare
• My Summary of findings
– On average positive effect on firm performance
– But impact is highly variable; e.g. organizations can
spend huge amounts on ICT for zero benefit
The bill for abortive plan, described as 'the biggest IT
failure ever seen', was originally estimated to be £6.4bn
An abandoned NHS patient record system has so far
cost the taxpayer nearly £10bn
“Abandoned
NHS IT system
has cost £10
billion”
Sept 17, 2014
When does technology successfully raise firm performance?
• Key to getting most out of new technologies is also having
other “complementary” organizational factors
– Early work by Bresnahan, Brynjolfsson & Hitt (2002) on
US; Caroli & Van Reenen (2001) on EU
– Mirko Draca to discuss later
• Management is critical
– Firm organization
– Skills
• True at macro as well as micro level (e.g. Historian Paul
David on electricity and computers)
– Impacts takes time
Economic Evidence on management is
limited
“No potential driving factor of
productivity has seen a
higher ratio of speculation to
empirical study”.
Chad Syverson (2011,
Journal of Economic
Literature)
1) Developing management questions
• Scorecard for 18 monitoring (e.g. lean), targets & people (e.g.
pay, promotions, retention and hiring). ≈45 minute phone
interview of manufacturing plant managers
2) Obtaining unbiased comparable responses (“Double-blind”)
• Interviewers do not know the company’s performance
• Managers are not informed (in advance) they are scored
3) Getting firms to participate in the interview
• Official Endorsement: Bundesbank, Bank of England, RBI, etc.
• Run by 200 MBA types (loud, assertive & business experience)
WORLD MANAGEMENT SURVEY (WMS); BLOOM & VAN REENEN (2007)
Examples of tracking performance – Car Plant
World Management Survey (~12,000 firms, ~20k managers
in 4 major waves: 2004, 2006, 2009, 2014; 34 countries)
Medium sized manufacturing firms(50-5,000 workers, median≈250)
Now extended to Hospitals, Retail, Schools, etc.
http://worldmanagementsurvey.org/
Average Management Scores by Country
Note: Unweighted average management scores; # interviews in right column (total = 15,489); all waves pooled (2004-2014)
2.0272.2212.225
2.2542.316
2.3722.397
2.5162.549
2.5782.6082.611
2.6842.6992.7062.7122.720
2.7482.7522.762
2.8262.8392.8512.861
2.8872.899
2.9782.9973.0153.033
3.1423.188
3.2103.230
3.308
1.5 2 2.5 3 3.5Average Management Scores, Manufacturing
MozambiqueEthiopia
GhanaTanzania
ZambiaMyanmar
NicaraguaNigeriaKenya
ColombiaVietnam
IndiaBrazil
ArgentinaTurkeyChina
GreeceSpainChile
Republic of IrelandPortugal
Northern IrelandNew Zealand
SingaporePolandMexico
ItalyAustralia
FranceGreat Britain
CanadaSweden
GermanyJapan
United States
Africa
Asia
Oceania
Europe
Latin America
North America
1564
178
749
404
419
1540
780
473
632
406
525
364
151
137
410
161
611
214
585
763
332
568
1151
151
170
937
185
118
97
147
69
150
108
131
109
Interviews
Fra
cti
on
of
firm
s
Firm level average management scores, 1 (worst practice) to 5 (best practice)
0.5
11.5
0.5
11.5
0.5
11.5
0.5
11.5
0.5
11.5
0.5
11.5
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
Total Argentina Australia Brazil Canada Chile
China Colombia Ethiopia France Germany Ghana
Great Britain Greece India Italy Japan Kenya
Mexico Mozambique Myanmar New Zealand Nicaragua Nigeria
Northern Ireland Poland Portugal Republic of Ireland Singapore Spain
Sweden Tanzania Turkey United States Vietnam Zambia
Management also varies heavily within countries
MOPS: Survey run with the US Census Bureau
~47,000 manufacturing
plants in 2011 & 2016
(US ASM).
This was quick and easy to
fill out - and mandatory - so
~80% of plants responded
MOPS also now done in
many other countries
(Australia, Canada, China,
Finland, Germany, Japan,
Mexico, Pakistan, UK, etc.)
Management score decile
Pro
du
ctivity
Pro
fit
Ou
tpu
t g
row
th
Exp
ort
ers
R&
D p
er
em
plo
ye
e
Pa
ten
ts p
er
em
plo
ye
e
Management scores positively correlated with many
measures of firm performance
Source: Bloom, Brynjolfsson, Foster, Jarmin, Patnaik, Saporta-Eksten & Van Reenen (forthcoming, AER).
“Americans do I.T. better” (Bloom, Sadun and Van Reenen,
AER, 2012)
• Use management data + IT data (ONS & Harte-Hanks)
• What happens to establishment productivity after changes in
ICT investment?
• Firms with better people management, don’t just spend more
on IT, but enjoy bigger productivity boost from each $ of IT
spend
– Well managed firms get double the productivity boost from
IT compared to poorly managed
– Accounted for half of the faster productivity growth in US
compared to Europe in decade since mid 1990s
• Similar findings on more recent data (e.g. Pelligrino & Zingales,
2018)
Productivity
Jobs and skills
Labor Share & Superstar Firms
Policy implications
What are the new digital technologies?
OUTLINE OF TALK
Will new technology make our lives better?
Or is it “Robo-calypseNow?”
Déjà vu all over again…
Not Running out of Jobs – U.S. Added
19.4 Million Jobs Between Jan 2010 – Sep 2018
129.7
Million
149.1
Million
Is Automation Labor Displacing?
Four countervailing forces against the employment-reducing effect of
automation
1. Uber effects
2. Walmart effects
3. Business-to-Business effects
4. Creation of new work / new tasks
‘Uber’ Effects – Produce a Cheaper,
Better Product, and Employment May Rise,
Taxi Trips
Taxi Trips
0,00
100,000
200,000
300,000
400,000
500,000
600,000
700,000
800,000
900,000
2015 2018
410,000
295,000
Ride Hailing Trips in New York City, 2015 and 2018
2015 2018
‘Uber’ Effects – Produce a Cheaper,
Better Product, and Employment May Rise,
Taxi Trips
Taxi Trips
Uber Trips
Uber Trips
Lyft Trips
0,00
100,000
200,000
300,000
400,000
500,000
600,000
700,000
800,000
900,000
2015 2018
475,000
820,000
2015 2018
Ride Hailing Trips in New York City, 2015 and 2018
Walmart Effects – A Fall In the Cost of
Necessities Frees Income for Luxuries
Business-to-Business Effects – There’s Been a Lot of Productivity
Growth in Steel!
Business-to-Business Effects –
486.9
955.4
1,100
1,500
377.40,00
0,500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
Metal Making Jobs(1,000s)
Metal Using Jobs(1,000s)
MetalManufacturing
Fabricated MetalProducts
Machinery
Motor Vehicles
AerospaceProducts
< 400K Jobs
> 4,000K Jobs
New technology destroys old tasks, but creates new tasks
• Acemoglu and Restrepo (2017,2018)
– Automation technologies reduce labor share and may
reduce overall labor demand.
– The “reinstatement effect” generated by new tasks
counterbalances this effects
• No trend in unemployment in long-run (but hours worked
have fallen)
• A bigger problem than the number of jobs is the quality of
jobs. Wages and other aspects of the desirability of work
Biased Technical Change → Shrinking Middle:
The ‘Barbell’ Labor Market (“Job Polarization”)
2016Low Skill
18.2%High Skill
38.6%Medium Skill
43.2%
1979Low Skill
13.7%High Skill
25.2%
Medium Skill
61.1%
Source: US data Autor (2018)
Source: KLEMS data for OECD
countries 2004-1980. Michaels,
Natraj & Van Reenen (2014)Agri
Food
TextilesWood
PaperChemicals
RubberMinerals
Metals
Machinery ElectricalTransport
Manufacturing
-30
-25
-20
-15
-10
Cha
ng
e L
ow
-Ski
lled W
age
bill
Sha
re
0 .01 .02 .03 .04Change ICT/VA
Figure 5B: Growth of Low-Skilled ShareTraded Industries Only
Agri
FoodTextilesWood
Paper
Chemicals
RubberMineralsMetals
Machinery
Electrical
Transport
Manufacturing
05
10
15
20
Cha
ng
e H
igh
-Ski
lled
Wag
ebill
Sh
are
0 .01 .02 .03 .04Change ICT/VA
Figure 3B: Growth of High-Skilled ShareTraded Industries Only
High Skill Agri
Food
TextilesWood
Paper
Chemicals
Rubber
MineralsMetals
Machinery
Electrical
Transport
Manufacturing
51
01
52
02
5
Cha
ng
e M
ed
ium
-Skill
ed
Wa
ge
bill
Sh
are
0 .01 .02 .03 .04Change ICT/VA
Figure 4B:Growth of Medium-Skilled ShareTraded Industries Only
ICT GROWTH REDUCES DEMAND FOR MIDDLE SKILL JOBS, 2004-1980
Low Skill
Medium Skill
New Jobs are Not Primarily STEM!
(US 2012 – 2000)
Source: Deming (2018)
Many Growing Occupations Combine Interpersonal
with Technical Skills
Source: Deming (2018)
FIGURE I
Each row presents 100 times the change in employment share between 2000 and 2012 for the indicated
occupation. Consistent occupation codes for 1980-2012 are updated from Autor and Dorn (2013) and
Autor and Price (2013) and consolidated to conserve space – see the Data Appendix for details.
-.2 0 .2 .4 .6
Engineers (All)
Drafters And Surveyors
Engineering And Science Technicians
Architects
Physical Scientists
Biological Scientists
Pilots/Air Traffic Control
Math/Stats/Actuaries
Medical Scientists
Operations Researchers
Comp. Sci./Programming/Tech Support
STEM Occupations
-.2 0 .2 .4 .6
Writers, Editors & Reporters
Marketing, Advertising & Pr
Arts & Entertainment, Athletes
Social Scientists And Urban Planners
Dentists
Dental Hygienists
Pharmacists
Legal Assistants & Paralegals
Physicians' Assistants
Other Business Support
Lawyers & Judges
College Instructors
Physicians
Social Workers, Counselors & Clergy
Economists & Survey Researchers
Accounting And Finance
Health Therapists
Health Technicians
Nurses
Managers (All)
Teachers (K-12)
All Other Managerial or Professional Occupations
Source: 2000 Census and 2011-2013 ACS
100 x Change in Employment Share
Change in Relative Employment for Cognitive Occupations, 2000-2012
Productivity
Jobs and skills
Labor Share & Superstars
Policy implications
What are the new digital technologies?
OUTLINE OF TALK
Falling Labor Share of Corporate Sector
Value-Added
Why has labor share fallen?
‘Superstar Firms’ hypothesis (Autor, Dorn, Katz,
Patterson & Van Reenen, 2018)
• Large firms tend to have lower labor shares
• Rising prevalence of “winner take most” competition
• Small set of large firms capture increasing share of
market, aggregate labor share falls due to reallocation
The Rise of Superstar Firms
Source: Autor, Dorn, Katz, Patterson & Van Reenen (2017), Compustat
Global Sales of Top 500 US Firms tripled from
$4 trillion in 1972 to $12 trillion in 2015
Top3 in 1985
Top3 in 2015
The Rise of Superstar Firms doesn’t just reflect
US GDP growth
Source: Autor, Dorn, Katz, Patterson & Van Reenen (2017), Compustat
US Sales of Top500 US Firms / US GDP
Change in economic environment
• Change in environment which reallocates more market share
to superstar firms will tend to (i) increase concentration and (ii)
reduce aggregate labor share. Examples:
• Increased importance of platform competition (network
effects, especially in digital markets)
• Larger firms better at exploiting intangible capital; e.g.
Walmart ICT – Besson ’17; Lashkari & Bauer ’18
• Falling competition? Döttling, Gutierrez & Philippon ‘18 on
weaker US anti-trust, greater regulation & licensing.
• Rising Competition? Globalization; online comparison sites;
liberalization. A “Matthew Effect” allocates more output to
more efficient firms - Melitz, ’03
Summary of Evidence
1. A rise in sales concentration within four-digit
industries across US private sector
2. Industries with larger increases in concentration
see larger falls in labor share
3. Labor share fall largely due to reallocation of
activity between firms, not primarily a general fall
within all firms
4. Reallocation component of falling labor share
largest in industries with rising sales concentration
5. These patterns are seen internationally, not just in
US
Rising Concentration in the US across 4 digit SIC
industries - Manufacturing and Retail
A. Manufacturing Sector B. Retail Trade
Notes: Weighted average of 4 digit industries within each large sector. Manufacturing:
388 inds; Retail: 58.
Source: Autor, Dorn, Katz, Patterson & Van Reenen (2017)
CR20
CR4
Rising Concentration in the US -
Services and Wholesale Trade
C. Services D. Wholesale Trade
Notes: Weighted average of 4 digit industries within each large sector. Wholesale: 56.
Services: 95.
Source: Autor, Dorn, Katz, Patterson & Van Reenen (2017)
Rising Concentration in the US -
Utilities/Transport & Finance
E. Utilities + Transportation
Sector
Notes: Weighted average of 4 digit industries within each large sector. Utilities &
Transport: 48; Finance 31
Source: Autor, Dorn, Katz, Patterson & Van Reenen (2017)
F. Finance Sector
Fall of Labor Share mainly reallocation between firms U.S
Notes: MP decomposition over 5 year periods, aggregated to two 15 year periods
Manufacturing: Payroll over Value Added
Reallocation
between survivors
Within firms
Reallocationvia Exit
Reallocation via Entry
Productivity
Jobs and skills
Labor Share & Superstar Firms
Policy implications
What are the new digital technologies?
OUTLINE OF TALK
Implications
• New technologies create challenges and opportunities
– For business leaders & policy-makers
• Making the most of these opportunities is not automatic
– Technology can create huge value requires
complementary changes in organizational & skills
• How to improving management?
– Information/training; Ownership/governance; Competition.
Policy moving in wrong direction right now
• Competition/Anti-trust policy
– Even if firms become superstars via legitimate
competition, it does not mean that they will always use
their power in the best interests of consumers
THANKS!
Score (1): Measures
tracked do not
indicate directly
if overall
business
objectives are
being met.
Certain
processes aren’t
tracked at all
(3): Most key
performance
indicators
are tracked
formally.
Tracking is
overseen by
senior
management
(5): Performance is
continuously
tracked and
communicated,
both formally and
informally, to all
staff using a range
of visual
management tools
MONITORING – e.g. “HOW IS PERFORMANCE TRACKED?”
62
Note: All 18 questions and over 50 examples in Bloom & Van Reenen (2007)
http://worldmanagementsurvey.org/
POLARIZATION: OCCUPATION JOB SHARES CHANGE (PAY TERCILE) 1993-2010
Source: Goos, Manning & Salomons (2014)