the age of digital disruption: the future of employment · the age of digital disruption: the...

The Future of Employment / 1

The Age of Digital Disruption: the Future of Employment III Jornadas de Postgrado en Iberoamérica "El Futuro del Empleo” Sevilla, January 24, 2019


The consequences for human welfare involved in questions like these are simply staggering: once one starts to think about them,

it is hard to think about anything else

Robert E. Lucas, “On the Mechanics of Economic Development," 1988

In the same vein, the effects of the digital revolution on employment, productivity, inequality and welfare could be so

stunning that, once we start to think about them, it is hard to think about anything else

  Economic progress and social welfare depend on technical progress in the long run. Technological and digital transformation represents an opportunity in the history of mankind, but also enormous challenges

  The digital revolution is having disruptive effects on employment, occupations, required skills, the wage premium, inequality and polarization, although so far there are no grounds to claim that it affects unemployment at aggregate level

  It is crucial for societies (public sector, firms and workers) to anticipate and manage actively the digital revolution, using a broad array of policies that

•  ensure equal opportunities,

•  enhance the long-term positive effects of an inclusive technical and digital progress to bring this age of new opportunities to everyone, and

•  reduce transition costs in the short and medium term

Key messages


Introduction 01

02 Overview and historical evidence of the effects of technical progress

The effects of the digital revolution Will this time be different?

03

Policies for the digital revolution 04

Contents

Conclusions 05


01 Introduction


Introduction: The conventional wisdom

  Long-run progress and welfare in societies are determined by technological change, which boosts productivity, wages and income per capita

  Over the last two centuries, technological change has allowed for increasing employment without raising unemployment rates

  On an individual basis, workers with complementary skills to new technologies have traditionally benefited from better employment opportunities and higher wages (greater productivity and division of labor, Adam Smith)

  On the contrary, workers with skills that are partial or perfect substitutes for automation have experienced wage losses and even unemployment during the transition to new occupations


Introduction: Is the digital revolution different?

  Is the current technological and digital transformation different from previous industrial revolutions?

  To what extent will the automation of routine and non-routine tasks materialize?

  How will the impact of technologies such as computerization, robotisation, big data and artificial intelligence differ from that of previous technologies?

  We can attempt to forecast some trends by analyzing the immediate impact and scope of emerging technologies

  Growing interest academia, think tanks, international institutions and many other: a new line of research at BBVA Research


02 Overview and historical evidence

of the effects of technical progress


Overview: The pessimistic vision of a world without labour

PESSIMISTS

  Luddites, who destroyed industrial looms between 1811 and 1816, and the Swing Riots of 1830, where threshing machines were destroyed

Marx (1867): machines replace workers, whose value falls. This contradiction will mark the end of capitalism

Frey and Osborne (2013): 47% of employment in the US is under threat from computers

Brynjolsson and McAfee (2014): Digital innovations are contributing to the stagnation of average incomes in the United States and the disappearance of many middle-level jobs

Piketty (2014): Capital (in the hands of only a few) grows more than GDP and exacerbates inequality

Benzell et al (2015): A long-term decrease in labor's share in the redistribution of income from losers to winners. Smart machines can mean long-term misery for all

De Stefano (2016): risks for workers from new forms of work

Milanovic (2016): Technological progress, sectoral reallocation of labor, globalization and current policy are generating a second Kuznets curve that will not disappear any time soon

Avent (2016): New technologies will create new and good jobs, but they will not be enough to absorb the over-abundance of workers


Overview: A more optimistic vision of technical progress

  Technical progress has been a constant feature in history. At the height of the Luddite movement A. Lovelace, the creator of the first programming algorithm, was born in 1815

  Progress in some sectors increase income growth, which raises the demand for production in other sectors, and prompts the appearance of new goods and services, which in turn increases employment

Moretti (2010): A job created in high-tech sectors creates further 4.9 jobs in non-trade Goods sectors

Mokyr (2014): the future holds occupations that will seem as strange to us as many of those today seem to our grandparents. Our lack of imagination is largely responsible for much of today’s pessimism

Arntz et al (2016): When one considers the various different tasks in each occupation, only 9% of employment is capable of being automated on average in 21 OECD countries, which is far below the number estimated by Frey and Osborne (2013)

Gregory et al (2016): Technical change that substitutes routine work has positive net effects on overall employment in a sample of 27 countries between 1999 and 2010, as the externalities predominate which compensate the substitution of certain jobs by capital

Conseil d’Orientation pour l’Emploi (2017): Estimates do not take account of the fact that current jobs will change or of direct and indirect job creation deriving from technological change

  The “gig economy” can improve employment match-ups and labour market efficiency

OPTIMISTS


Today’s unemployment in the US and the UK are at rates similar to those at the start of the 20th century

GDP per capita in the United States, United Kingdom and Spain 1900-2018 (1900 = 100)

Unemployment rate in the US, Spain and the United Kingdom, 1900-2018

Source: Andrés and Doménech (2019) from Lebergott (1957), BLS, BoE, OECD. INE, Alcaide (2007) and de la Fuente (2017)

Source: Andrés and Doménech (2019) based on The Maddison Project (2018) and OCDE. Data in logarithm form

For over a century, technical progress has not destroyed jobs in aggregate terms, despite population growth and the increase of women labor participation rates

85

231

628

1900 1915 1930 1945 1960 1975 1990 2005

USA x 8,8 UK x 5,1

Spain x 10,8

100 0

5

10

15

20

25

1900 1915 1930 1945 1960 1975 1990 2005

USA Spain

UK


A negative correlation in the OECD between productivity growth and changes in the unemployment rate

  In the last six decades there has been a negative correlation in the OECD between productivity growth, due to innovation, and changes in the unemployment rate

  Nevertheless there are differences across countries and other factors explain these differences and the negative relationship between both variable (correlation is not causation)

Productivity growth and unemployment changes, OECD, 1960-2018

Source: Andrés and Doménech (2019) based on OECD.

AUS

AUT

BEL

CAN

CHE

DEU

DNK

ESP

FIN

FRA

GBR

GRE

IRL ISL

ITA

JAP

LUX

NLD

NOR

NZL PRT SWE

TUR

USA

-2

0

2

4

6

8

10

12

14

1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

Cha

nge

in th

e un

empl

oym

ent r

ate,

201

8-19

60

Annual rate of growth of GDP per hour worked


A continuous process of creative destruction, new occupations, and sectoral reallocation

Sectoral distribution of full-time equivalent employment, Spain, 1900-2015

Employment in services and per capita income in OECD countries, 1840-2000

Source: Andrés and Doménech (2018) based on Herrendorf et al (2014) Source: Andrés and Doménech (2018) based on Prados de la Escosura (2017)

Sectoral development due to technical progress, increasing globalization and changes in consumer preferences (Baumol, 1967)

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

100%

1900 1915 1930 1945 1960 1975 1990 2005 Primary sector Manufacturing Construction Services

0%

20%

40%

60%

80%

6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 GDP per capita (in logs.)

OECD (10) Spain USA UK


Thanks to technical progress, working hours have voluntarily fallen over time, contradicting Keynes

Weekly hours worked in the United States, the United Kingdom and Spain, 1870-2000

Hours worked and productivity in the OECD, 2015

Source: Andrés and Doménech (2018) based on OECD Source: Andrés and Doménech (2018) based on Huberman and Minns (2007)

30

35

40

45

50

55

60

65

70

1860 1880 1900 1920 1940 1960 1980 2000

Wor

king

hou

rs p

er w

eek

Spain USA UK

IRL

MEX

NOR

ESP UK

USA

7.2

7.3

7.4

7.5

7.6

7.7

7.8

2.9 3.1 3.3 3.5 3.7 3.9 4.1 4.3 4.5

Wor

king

hou

rs p

er y

ear,

2015

(lo

gs)

GDP per hour, 2015 ($I and 2010 PPP, logs)

Keynes (1930) predicted that in the long run the working week would be 15 hours to keep employment stable, which has not been the case so far


Technical progress in health has increased life expectancy

  Since 1960 life expectancy has increased by 1.9 years per decade

  Life expectancy at age 65 years is increasing more than one year per decade

  The increase in life expectancy is one of the determinants of the great improvement in well-being

  A challenge for the pension system. Retirement age (65 years) has remained virtually unchanged since 1916

Life expectancy at birth, 1900-2011

Source: Own work based on www.clio-below.eu EU5: Denmark, Finland, Germany, the Netherlands and Sweden

30

40

50

60

70

80

1900 1920 1940 1960 1980 2000

Year

s

USA

Spain

UK

EU5


Inequality: shifting trends in the last century

Income share of the top 1%, 1900-2012

-0.5

0.6

1.7

2.7

3.8

4.9

6.0

4

8

12

16

20

24

28

1900 1920 1940 1960 1980 2000

0.1% share 1%

sha

re USA

UK Spain

(0,1%,right)

EU5

Spain (1%)

Source: Own work based on www.wid.world EU5: Denmark, Finland, Germany, the Netherlands and Sweden

  Inequality increased at the end of the 19th century

  The Great Levelling 1920-1970

  Inequality (personal and functional) has increased since 1980, mainly in major Anglo-saxon countries … … affected by the interaction between technical progress, demography, educational policies, globalization, competition in product and labor markets, and the response of the welfare state


03 The effects of the digital revolution

Will this time be different?


Is automation destroying employment? An open debate

Correlation between exposure to robots and employment for metropolitan areas of the US

Source: Acemoglu y Restrepo (2017)

Detroit, MI

Lansing, MI

Muncie, IN Beaumont, TX

Cleveland, OH Toledo, OH

Exposure to robots between 1993 and 2007

Cha

nge

in e

mpl

oym

ent (

Cen

sus)

, 199

0-20

07

The race against the machine Acemoglu y Restrepo (2017):

  One additional robot (autonomous machines with multiple reprogrammable applications) per thousand employees reduces the employment rate by 0.18 - 0.34pp and wages by 0.25 - 0.5pp

  Greater effect among more exposed industries, manual occupations and workers without a university education

Graetz y Michaels (2016)   In 14 industries in 17 countries

(1993-2007), robots boost productivity and wages, and reduce prices but not aggregate employment, although they do that among the least skilled


Is automation destroying employment? An open debate

Digital intensity and unemployment in 40 countries, 2016-2017

Source: BBVA Research based on data from Eurostat (2018), European Commission (2017), International Telecommunication Union (2017), and the International Federation of Robotics (2017)

UE8

KOR

DNK

GBR

NLD

JAP

SWE

DEU USA

SGP

AUT

FIN

BEL

ESP GRE (21.5)

2

4

6

8

10

12

14

16

18

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5

Une

mpl

oym

ent r

ate,

201

7

Digital intensity and robotics, 2016-2017

The race against the machine

  There is a negative correlation between the digital and robotics intensity and unemployment is observed

  The digital revolution is fostering new activities, while boosting demand for less innovative sectors


Skill biased technological change

Relative supply of human capital and relative salary in 18 OECD countries, 2000-16

Source: Andrés and Doménech (2018) based on the OECD (2018)

1.81

1.83

1.85

1.87

1.89

1.91

1.93

1.95

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2000 2004 2008 2012 2016 W

age

prem

ium

(ter

tiary

/low

sec

onda

ry)

Rel

ativ

e su

pply,

adu

lts w

ith te

rtiar

y ed

ucat

ion

Wage premium (right)

Relative supply (left)

The race against the machine

Goldin and Katz (2008) and Acemoglu and Autor (2011). Despite the increase of workers with higher education, their relative wage has risen compared to those with lower educational levels

  Technical progress is complementary to skilled workers, increasing their demand faster than their supply


Routine tasks and occupations

Share of routine tasks by occupations ranked by wages

Source: BBVA research based on Aum et al (2018), and Díaz, Doménech and Neut (2018)

0%

10%

20%

30%

40%

50%

60%

70%

0 20 40 60 80 100

Sha

re o

f rou

tine

task

s

Accumulated employment by occupations, sorted by average wage

Spain

USA

Technical progress biased against routine work

Autor, Katz y Kearney (2006) find that automation and computerization complement workers who perform non-routine and abstract tasks, substitute those who carry out routine work, and do not affect those who undertake manual and non-routine activities

  Polarization has also occurred in Europe in the last two decades


Job polarization

Change in employment between 1993 and 2010 in 16 European countries

Source: Andrés and Doménech (2018) based on data from Goos, Manning and Salomons (2014)

-15

-10

-5

0

5

10

15

NO

R

AU

T P

RT

DE

U

FRA

ITA

NLD

G

RE

G

BR

S

WE

D

NK

B

EL

ES

P IR

E

FIN

LU

X

NO

R

AU

T P

RT

DE

U

FRA

ITA

NLD

G

RE

G

BR

S

WE

D

NK

B

EL

ES

P IR

E

FIN

LU

X

NO

R

AU

T P

RT

DE

U

FRA

ITA

NLD

G

RE

G

BR

S

WE

D

NK

B

EL

ES

P IR

E

FIN

LU

X

Four occupations with lower salaries

Nine occupations with lower salaries

Eight occupations with lower salaries

Technical progress biased against routine work

Autor, Katz y Kearney (2006) find that automation and computerization complement workers who perform non-routine and abstract tasks, substitute those who carry out routine work, and do not affect those who undertake manual and non-routine activities

  Polarization has also occurred in Europe in the last two decades


Skill biased technological change and job polarization are having heterogeneous effects across countries

Per capita income, inequality and unemployment OECD, 2016

Source: BBVA Research based on the OECD and Eurostat. The size of each circle is in proportion to unemployment

AUS

AUT

BEL

CAN

CHE

DEU

DNK

ESP FIN FRA

GBR

GRE

IRL

ISL

ITA

JAP

NLD

NZL PRT

SWE

USA

40

50

60

70

80

90

100

22 24 26 28 30 32 34 36 38 40

Gro

ss N

atio

nal I

ncom

e (U

SA

)=10

0

Gini Index, 2015

An efficient and fair transition

  There are big disparities among advanced economies as regards to per capita income, unemployment and inequality

  The challenge is to manage the technological and digital transformation by spurring growth while reducing both inequality and the unemployment rate.

  Certain countries, such as Spain, are far from the frontier, allowing for higher-growth policy strategies that do not raise unemployment and inequality


The future of employment

  It is much easier to estimate which jobs are going to be affected than to guess which ones are going to be created

  The Baumol effect: Richer societies demand more labour-intensive services, increasing wages: health and personal care (ageing), education, leisure and tourism activities, personal services, …

  The challenge is a fair transition from jobs destroyed to jobs created: Protecting people rather than jobs (Tirole, 2017)


Workers affected by the digital revolution

Percentage of workers in occupations with high risk of automation in the U.S. and Spain*

Source: Andrés and Doménech (2018) based on data from Fey and Osborne (2017), Arntz et al (2016) and Domenech, García, Neut and Montañez (2018)

0

10

20

30

40

50

Frey y Osborne (2017)

Arntz et al (2016) Doménech et al (2018)

Arntz et al (2016)

EE.UU. España USA Spain

  According to Frey and Osborne (2017), 47% of jobs in the U.S. face a high risk of automation

  With the same methodology, in Spain the percentage is 36%

  When you take into account the various tasks of each occupation the risk is around 10%


Workers affected by the digital revolution

Characteristics of workers in jobs most at risk of being automated in Spain*

Source: Domenech, García, Neut and Montañez (2018)

Low level of education

Women

In the sector: •  Primary •  Manufacturing •  Trade •  Hotels, restaurants

and cafés •  Finance •  Property

Young people

Do not undertake non-formal education

In small firms

Not telecommuting

Without positions of responsibility

Seeking a new job

Have come out of unemployment

Private sector

With temporary contract

  The probability of automation of individuals is determined by variables that differentiate between each employee with regard to personal characteristics, as well as employment and the firm

  Technological transformation is both an opportunity and a challenge

  Although the aggregate impact is positive, it may have very different effects for different groups of workers

  It is essential to lead the change with policies that smooth the transition, cushion the costs and boost the benefits


04 Policies for

the digital revolution


Policies for managing technological change: Education

Investing in human capital is crucial to achieving skills complementary to technical progress, even for less-skilled tasks

01 Raising the quality of jobs in the services sector and improving society’s views towards many of them

02 Improving professional (e.g. languages), social, managerial and personal skills which are required to meet society’s growing needs

03

Continuous training and flexibility to change occupations over working life. New tasks for public employment services and in collective bargaining

04 Adapt the educational system and training to new demands tanking into account skill biased technological progress

05 The educational system must evolve in step with society, fomenting creativity, boosting non-cognitive skills and improving social intelligence

06


Policies for managing technological change: Education

Percentage of the population aged between 25 and 34 years by education level in 2016

Source: Doménech, García, Neut and Montañez (2018) based on data from the OECD (2017)

  Skill-biased technical progress and new jobs that might be hard to imagine at the present time not only require more but, above all, better and more flexible training

  The dual distribution of educational levels in Spain means that approximately one third of the younger population may lack the skills needed for the digital transformation

KOR POL

DEU

FRA

ITA POR COL ARG

ESP BRA

UE8

0

10

20

30

40

50

60

70

10 20 30 40 50 60 70

Bel

ow u

pper

sec

onda

ry e

duca

tion

(%)

With tertiary education (%)


Labor market policies

Removing barriers to job creation, investment and firm growth: finance, start-ups, taxes, red tape, regulations in product markets, quality of institutions, etc,

01 Better labor market regulations, and active and passive policies: A more efficient and equitable

02 Improving the matching process between vacancies and the unemployed using big data + AI. Additional information to improve skills

03

A better tax structure that allows for redistribution without harming employment or investment in new technologies

04 Adapt labor regulations to the gig economy and the needs of independent workers and new employers in the labor market

05


Policies for inclusive growth and redistribution

Education and employment policies are necessary conditions, but may be not enough

01 Greater welfare for all in the long run, but substantial transition costs for many workers in the medium term

02 First, ensure equality of opportunities and later insure individuals against adverse situations (ex-post redistribution)

03

Efficiency of public policies: ensure inclusive growth at the lowest possible cost in terms of employment and investment in innovation

04 The challenge is to distribute wealth and not to curb its creation, with taxes on automation (what is a robot?), artificial intelligence or big data

05 The welfare state still has a lot of scope for improvement in boosting employment, income and equality before proposing a basic universal income

06


Multiple transitions: Winners or losers?

Low productivity and high unemployment

(Inequality with low redistribution)

Low productivity and low unemployment

(Ageing problem)

High productivity and high unemployment

(Inequality with redistribution)

High productivity and low unemployment

(A new Great Levelling)

Employment creation

Pro

duct

ivity

Countries can end up in very different equilibria, depending on how they manage the digital revolution. Better policies will allow more successful outcomes.

A new international division of labor and wealth (strategic map)

+

+


05 Conclusions

Key messages

  Economic progress and social welfare depend on technical progress in the long run. Technological and digital transformation represents an opportunity in the history of mankind, but also enormous challenges

  The digital revolution is having disruptive effects on employment, occupations, required skills, the wage premium, inequality and polarization, although so far there are no grounds to claim that it affects unemployment at aggregate level

  It is crucial for societies (public sector, firms and workers) to anticipate and manage actively the digital revolution, using a broad array of policies that

•  ensure equal opportunities,

•  enhance the long-term positive effects of an inclusive technical and digital progress to bring this age of new opportunities to everyone, and

•  reduce transition costs in the short and medium term


The Age of Digital Disruption: the Future of Employment III Jornadas de Postgrado en Iberoamérica "El Futuro del Empleo” Sevilla, January 24, 2019

the age of digital disruption: the future of employment · the age of digital disruption: the...

Documents