Inequality and Technology:

The Future of Jobs and Social Policy

Omar Arias, World Bank European Investment Bank , Luxembourg 2017

115

years

198

years

Technological breakthroughs are speeding up

Mobile

Internet

First phone

call 1876

First website

1991

First iPhone

2007

16

years

Hargreaves’

Jenny 1764 GM’s Unimate

1962

Google’s

Schaft 2010

48

years

SOURCE: McKinsey Global Institute analysis

Internet of

things

Machine

intelligence

Advan-

ced

robotics

Alternatives

An unprecedented pace of penetration

0

20

40

60

80

100

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012

improved water

improved sanitation

secondary school

enrollment

mobile phone

internet

mobile broadband

Note: Mobile phone and mobile broadband subscriptions, internet users, improved water and sanitation are per 100

individuals. Net secondary school enrollment is the percent of the relevant age group. Sources: World Bank, WDR on Internet and Development Team based on World Development Indicators and ITU data.

High, and growing use …

2.95 billion

42%

3.63 billion

50%

2.03 billion

28%

Source: wearesocial.sg, September 2014

… Across developed and developing countries

101

%

89

% 67

%

129

%

151

%

112

%

124

%

90

%

72

%

92

%

109

%

94

%

Mobile Penetration, January 2014

Source: wearesocial.sg, January 2014

Although access and use of digital technologies, especially the internet, is

very unequal

18

12

23

10

20

8

11

4

Men Women Urban Rural Age 15-24 Age 45+ Upper 60% * Bottom 40%*

Africa: Percentage of individuals who report ever using the internet (%) (2012)

Source: WDR 2016, based on Research ICT Africa RIA survey.

Improving welfare and reducing poverty?

Exacerbating inequality?

Growing Opportunities Threats of Disruption and Exclusion

The increasing use of digital technologies is creating jobs …

Source: OECD 2017.

Contribution of ICT sector to total employment

(circa 2012)

But still a small share of employment, while ICT intensity of

jobs is larger

Contribution of ICT to employment

(2012-2013)

Source: WDR 2016, based on STEP household surveys. OECD definitions for ICT sector and occupations.

0

5

10

15

20

25

30

Sh

are

of

em

plo

ym

en

t

(%)

ICT Sector ICT Occupations ICT Intensive

Impact on employment and earnings inequality is a balance

between two forces

Technology

complements

some

Workers (skill-

biased)

But technology

can substitute

others (labor-

saving)

What matters is whether the task is ROUTINE (and can thus be

automated) or NON-ROUTINE

Work is becoming more intensive in non-routine skills, and labor markets

are polarizing

Skills-intensity of Employment (simple cross-country average by type of occupation)

(2000-2012)

20

25

30

35

40

45

50

Sh

are

in t

ota

l em

plo

ym

en

t (%

)

OECD countries

Non-routine cognitive or inter-personal

Routine cognitive or manual

Non-routine manual

20

25

30

35

40

45

50

Sh

are

in t

ota

l em

plo

ym

en

t (%

)

Developing countries

Non-routine cognitive or inter-personal

Routine cognitive or manual

Non-routine manual

Source: WDR 2016, based on ILO KILM data. Skills classification follows Autor (2014).

Demand of skills is shifting towards jobs requiring both non-routine cognitive/technical

and socio-emotional skills

Source: Deming (2015), “The Growing Importance of Social Skills in the Labor Market”, NBER WP No. 21473

Labor markets in the developing world are also becoming polarized (shift towards non-

routine skills)

Source: WDR 2016, based on ILO KILM data. For China, data from the Population Census for 2000 vs 2010.

Change in employment shares in selected developing countries: 1993-2010

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

An

nu

al a

ve

rag

e c

ha

ng

e in

em

plo

ym

en

t sh

are

(pe

rce

nta

ge

po

ints

)

High-skilled occupations (intensive in non-routine cognitive and interpsersonal skills)

Middle-skilled occupations (intensive in routine cognitive and manual skills)

Low-skilled occupations (intensive in non-routine manual skills)

The one notable exception to labor market polarization trends – China- is

no longer…

Changes in the Skills-intensity of Employment in China (2000-2015)

Source: Park and du Yang 2017, forthcoming for China’s Sources of Growth Study.

-.06 -

.05-

.04-

.03-

.02-

.01

0

.01

.02

.03

.04

.05

.06

stand

ard

devia

tion

from

200

0 m

ean

2000 2005 2010 2015year

-.06

-.05-

.04-

.03

-.02-

.01

0

.01

.02

.03

.04

.05

.06

stan

dard

dev

iatio

n fro

m 2

000

mea

n

2000 2005 2010 2015year

-.06 -

.05-

.04-

.03-

.02-

.01

0

.01

.02

.03

.04

.05

.06

stan

dard

dev

iatio

n fro

m 2

000

mea

n

2000 2005 2010 2015year

-.06

-.05

-.04

-.03

-.02

-.01

0

.01

.02

.03

.04

.05

.06

stan

dard

dev

iatio

n fro

m 2

000

mea

n

2000 2005 2010 2015year

Non-routine cognitive analytical

Non-routine interpersonal

Routine cognitive Routine manual

Use of robots in China is up 60% between 2010-15

Digital technologies are expected to take on

or transform many jobs

Source: World Bank 2016, based on household surveys, the Income Distribution Database (I2D2),ILO

Laborsta database, China’s Population Census, Frey and Osborne 2013, and Comin and Hobjin (2010).

0

20

40

60

80

100

HR

V

CY

P

LVA

MLT

LTU

CH

N

OEC

D

ALB

THA

RO

U

EC

U

CR

I

MY

S

MU

S

ZA

F

SR

B

PA

N

GTM

AR

G

SLV

BG

R

SY

C

ETH

UK

R

PSE

GEO

KSV

Sh

are

of

em

plo

ym

en

t th

at

ca

n b

e a

uto

ma

ted

(%)

Adjusted (technological feasability + adoption time lags)

Estimated share of employment that is susceptible to automation (%)

50%: Probability that a child in the developing world will find a job in an occupation

as they exist today

Newer technologies (e.g, robots) can have net negative labor impacts by

displacing workers altogether

Source: WDR 2016 team, based on household surveys, the Income Distribution Database (I2D2),ILO Laborsta database,

China’s Population Census, Frey and Osborne 2013, and Comin and Hobjin (2010).

Estimated impacts on employment (left) and wages (right)

of exposure to robots in the US (%)

Technological change is one key driver of the fall in labor shares in

output across the world

Source: Karabarbounis, L. and B. Neiman (2013) Note: The figure shows estimated trends in the labor share for all countries in data set with at least 15 years of data starting in 1975. Trend coefficients are reported in units per 10 years (i.e., a value of 5 means a 5 percentage point decline every 10 years).

-15

-10

-5

0

5

10

PO

L

MEX

HU

N

EST

BH

R

SV

N

LTU

ZA

F

NO

R

LUX

FSM

NA

M

LVA

NZL

CH

N

FIN

TUN

AR

G

SV

K

GER

AU

T

SW

E

FR

A

ITA

AU

S

TWN

CA

N

JP

N

DN

K

CH

E

USA

NLD BEL

CZE

ESP

MA

C

SG

P

GB

R

PR

T

BO

L

TUR

AR

M

CO

L

KEN

THA

CR

I

ISL

BLR

MD

A

KO

R

UK

R

BR

A

Lab

or

sha

re tre

nd

s, p

erc

en

tag

e p

oin

ts e

ve

ry 1

0 y

ea

rs

Trends in labor shares in output every 10 years

since 1975

Policies have to adapt to new realities…

Technology changes the skills required to succeed in a modern economy.

Technology also accelerates the pace of change, making skills obsolete more quickly and opening up new opportunities.

Technology further changes the world of work, introducing new forms of work and allowing for more flexible work arrangements but also eroding traditional employer-employee and social protection schemes.

Policy Implications: Strengthening life long-learning and training programs

Equip future workers with the skills that are complementary to

technology: foundational (cognitive and socio-emotional) skills,

digital skills;

Schools need to shift from rote learning to nurturing “learning to learn”

Refocus training programs to equip workers with both foundational

and technical skills

Improve incentives for life-long learning

For individuals

Training accounts

For industries

Work with sector-wide trade and employer unions to co-finance training

and retraining in sector-specific, but not firm-specific training

For firms

Subsidies for firms to provide non-firm, non-sector specific training

Policy Implications: Rethinking social protection schemes –shift to protections

delinked from the job

Technology can make too strict labor regulations more

binding (need for workforce reorganization)

All individuals should be registered in the same social

insurance system, regardless of where they work, with subsidies

for the poor or low-wage earners.

Strengthen the link between employment services, post-

secondary educational institutions and the private sector

using technology throughout the service chain

Raises several policy issues:

How to move away from an insurance system designed with little careers

disruptions and stable formal employment in mind?

How to support workers that are not able to work or earn enough to afford

a basic standard of leaving and coverage? Universal basic income?

How to finance social insurance if labor taxation becomes less desirable?

Source: The New Yorker; Cartoon by Zachary Kanin.

And hopefully we can avoid this!

Thank you!

CHECK OUT THE

WORLD DEVELOPMENT REPORT 2016