1. PUTTING FACES TO
THE JOBS AT RISK OF
AUTOMATION
3rd Annual Conference of the
Global Forum on Productivity
28-29 June 2018, Ottawa
Glenda Quintini – Senior Economist
Skills and Employability division
OECD Directorate for Employment, Labour and Social Affairs
Funded by:
2. Plan of the presentation
Based on Automation, Skills Use and Training
Joint work with Ljubica Nedelkoska
The potential for automation
What it means concretely
What we can do about it
3. The scary numbers: US 2013-2017
47%
27%
US 47%
Ethiopia 85%
51%
62% of workers =
computers
4. To what extent will
technology change
the way jobs are
performed?
Who will be the
most affected
and how could
adult learning
policies help?
1 Understand 2 Policy
2 OBJECTIVES
5. Measuring the risk of automation
with PIAAC
BASELINE: Frey and Osborne
Canadian PIAAC sample to
exploit 4-digit ISCO
Identify the same occupations and
similar bottlenecks
Out-of-sample prediction for jobs
in different countries
6. Key bottlenecks to the risk of automation
Perception
and
manipulation
• Finger dexterity
• Manual dexterity
• Cramped space
Creative
intelligence
• Originality
• Fine arts
Social
intelligence
• Social
perceptiveness
• Negotiation
• Persuasion
• Assisting/caring
Potential over-estimate at low endWarning!!!
7. 0
10
20
30
40
50
60
70
Fears of mass technological unemployment are
likely exaggeratedTHAN THOUGHT
Source: Survey of Adult Skills (2012, 2015)
Share of jobs at significant risk (50-70%) and of high risk (>70%) of automation
SIGNIFICANT RISK
HIGH RISK
However, many jobs will experience significant change
%
8. Why are jobs in other countries more/less
automatable than in Canada? (1)
(0.06)
(0.04)
(0.02)
-
0.02
0.04
0.06
0.08
0.10
0.12
0.14
PositivevaluesmeanhigherautomatabilitythanCanada
btw wth
A shift share analysis of industries
Within-industry
variation in tasks
(70%) more
important than
differences in the
industrial structure
(30%)
Source: PIAAC, all countries, own calculations.
9. Why are jobs in other countries more/less
automatable than in Canada? (2)
(0.06)
(0.04)
(0.02)
-
0.02
0.04
0.06
0.08
0.10
0.12
0.14
PositivevaluesmeanhigherautomatabilitythanCanada
btw wth
A shift share analysis of occupations
Equal importance of
differences within and
between occupations
Source: PIAAC, all countries, own calculations.
10. In the era of AI, the risk of automation is
highest for low-skilled low-paid workers
Highest risk in routine jobs
with low skill and education
requirement BUT low risk
applies to a broad range from
professionals to social workers
Automation mostly affects
manufacturing industry and
agriculture BUT some service
sectors are highly automatable
too.
The risk of automation falls
monotonically with hourly
wages
The risk of automation also falls
with educational attainment
No evidence of higher risk for
middle-skilled or rising risk at
the high end: automation risk
declines with skills, education
and hourly wages
Young people are the most at
risk of automation, followed by
older workers, with
disappearing student jobs and
entry positions.
11. Timing? demand already surpasses supply
in high-level cognitive and social skills
11
-0.015
-0.01
-0.005
0
0.005
0.01
0.015
0.02
JudgmentandDecisionMaking
SocialPerceptiveness
Troubleshooting
EquipmentSelection
VerbalAbilities
ReasoningAbilities
PhysicalStrength
Endurance
HealthServices
EducationandTraining
ArtsandHumanities
ManufacturingandProduction
Skills Abilities Knowledge
United States EU unweighted average
<-surplusshortage->
Source: OECD Skills for Jobs Database
12. Source: OECD Skills for Jobs Database- Unweighted average for European Countries
…and this is intensifying over time
-0.020
-0.015
-0.010
-0.005
0.000
0.005
0.010
0.015
0.020
0.025
WrittenExpression
WrittenComprehension
OralExpression
DeductiveReasoning
InductiveReasoning
FluencyofIdeas
OralComprehension
ProblemSensitivity
Originality
MathematicalReasoning
SpeechClarity
PeripheralVision
GlareSensitivity
SpatialOrientation
ResponseOrientation
GrossBodyEquilibrium
DepthPerception
RateControl
SpeedofLimbMovement
ReactionTime
Arm-HandSteadiness
GrossBodyCoordination
DynamicStrength
ControlPrecision
Stamina
ManualDexterity
MultilimbCoordination
TrunkStrength
ExtentFlexibility
StaticStrength
Initial year Final year
Shortages in cognitive
skills have become
more pervasive
Surpluses in routine and
physical abilities have become
more evident
13. Job content has changed significantly in
the past decades
United Kingdom
manual tasks (between
occupations)
social skills (within
occupations)
analytical skills (within
and between occupations)
Germany
manual tasks (between
occupations)
social and analytical skills
(within and between
occupations)
Overall
❖Shift away from occupations with high manual task content
❖Rise of social and analytical tasks: more important in existing
jobs, more jobs that use these intensively
14. Sectors differ in:
Their human capital endowment
Their structure and organisation
of production
The extent to which develop
and adopt new technologies
From potential to actual
employment effects TALES…
Job creation!!!
15. Penetration and adoption of technology
vary across sectors (and countries)
Dispersion of sectors in each considered dimension of digitalisation, 2013-15
Source: OECD Science, Technology and Industry Scoreboard 2017 , Statlink: http://dx.doi.org/10.1787/888933618669
16. Context matters too
Source: PIAAC, all countries, own calculations.
Ongoing work to
account for these
factors
Rigidities
• Collective
bargaining
• Job protection
Labour
costs
• Labour costs
• Minimum
wages
Social
protection
• Unemployment
benefits
• Social norms
17. The cost of inaction is high
Vacancies remain unfilled
for too long
Delays in production
Slow adoption of
technologies
Higher turnover and re-
training costs
17
Lower wages
Risk of jobs loss
and skills
obsolescence
For individuals For employers
18. Skills mismatch and productivity
0
2
4
6
8
10
12
0
5
10
15
20
25
30
35
40
Percentage of workers with skill mismatch (LHS)
Gains to labour productivity from reducing skill mismatch (RHS)
19. What can we do about it?
Education
Teach kids the
relevant skills
Adult
learning
Develop future-
ready inclusive
programmes
Social
protection
Income support and
re-employment
assistance
20. Re-training for new jobs is challenging,
especially for the low-skilled
Source: PIAAC, all countries, own calculations.
0
5
10
15
20
25
30
35
40
45
0
10
20
30
40
50
60
70
Annual incidence of job-related training, average (left-hand scale)
Annual incidence of job-related training, low-skilled adults (left-hand scale)
Annual hours of job-related training (median; right-hand scale)
21. Risk of automation and training
Workers in
fully
automatable
jobs get:
4 times less job-
related training
3 times less on-
the-job training
30 hours less
training per year
22. What do inclusive adult learning systems
look like?
Source: OECD, Future-ready adult learning (forthcoming, Q4 2018)
Information
and
guidance
Public awareness
campaigns
Career guidance
Online databases
Lower
participation
barriers
Recognition of
prior learning
Flexible learning
provision
Training leave
Financial
incentives
Target the
low-skilled
Higher subsidies
for low-skilled
Targeted
information
campaigns
Engage
social
partners
Provision
subsidies to
employers
Training levies
paid by employers
Social partners
management of
training funds
Unions outreach