demographic and technological change: two megatrends ... · 5/18/2014 · sgp 0 10 20 30 40 50 60...

Demographic and technological change: Two megatrends shaping the

labour market in Asia

Rafal Chomik

John Piggott

1. Will ageing inhibit productivity growth

2. Is technology change age-biased in labour markets?

3. How can technologies help with ageing pressures?

4. Which policy and research can help?

But first, some context

Life exp. increasing, in Asia at lower levels of development

JPN

CHNVNM

QTR

SGP

30

40

50

60

70

80

300 3000 30000

2015

1950

1950 1950

19501950

2015

2015

2015

2015Life exp. at birth

GDP per capita PPP

And large declines in fertility

JAP

KORIDN

PHL

CHN

0

1

2

3

4

5

6

7

8

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050

TFR, 1950-2050

Which means ageing workforces

CHN

JPN

KOR

IDN

MYS

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

1950 1965 1980 1995 2010 2025 2040

Population aged 50-64 as a share of 20-64, 1950-2050

At the same time, Asia has more tech innovation…

JPN

KOR

CHN

USA

SGP

MYS

THA

IDN3

30

300

3000

1980 1985 1990 1995 2000 2005 2010 2015

Patent applications per million population, by origin, 1980-2016

…and technology diffusion

CHN

KOR

IDN

MYS

AUS

SGP

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1990 1995 2000 2005 2010 2015

Share of population using the internet,

1990-2016

Ageing prospects and tech-readiness in Asian countries

Bangladesh

Cambodia

China

India

Indonesia

Japan

Malaysia

Nepal

Philippines

Korea

Singapore

Sri Lanka

Thailand

Viet Nam

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

7

0.0 0.5 1.0

Tech

dif

fusi

on

20

15

50+/20-49 pp chg 2015-2050

US

High-ageing-

high-diffusion

High-ageing-

low-diffusion

low-ageing-

high-diffusion

BangladeshCambodia

ChinaIndiaIndonesia

Japan

Malaysia

Nepal

Philippines

Korea

Singapore

Sri Lanka Thailand

Viet Nam

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

0.0 0.5 1.0

Inn

ova

tio

n 2

01

5

50+/20-49 pp chg 2015-2050

US

High-ageing-

high-innovation

High-ageing-

low-innovation

low-ageing-

high-innovation

1. Will ageing inhibit productivity growth?

2. Is technology change age-biased?



Productivity growth depends on several components

Δ Physical

capital

Δ Technology

Infrastructure capital

(e.g., transport, communications)

Production capital

(e.g., computers, factories)

Embodied (e.g., better

computer, better factory)

Δ Productivity Δ Human

capital quality

(e.g., Health, Education,

Training, Experience)

Abstract processes (e.g.,

better legal, scientific,

management, org)

Invention (when first

discovered)

Innovation (when

application developed)

Diffusion (when adopted

across economy)

Ageing and productivity: Physical capital channel

1 Life-cycle saving more investment 2nd demog. dividend +

2 Fewer workers per existing capital higher productivity per worker +

3 Slower growth less attractive for FDI -

4 Ageing related spending higher taxes less attractive for FDI -

Ageing and productivity: Human capital channel

1 Edu, health, cognition lower for older avg. quality of labour lower -

2 Greater experience of older quality of the labour input +

3 Demography moving into most productive peak ~ age 40 +

4 Longer lives investment in own HC +

5 Fewer children spending per child is higher avg. HC higher +

6 Healthy life expectancy ageing retarded in successive cohorts +

Ageing and productivity: Technology channel

1 If creativity lower by age lower invention and innovation -

2 More 2nd careers + access to credit more entrepreneurs +

3 If older more risk averse less venture capital -

4 If older adoption tech slower Lower diffusion -

5 Fewer new workers fewer new ideas invented and diffused -

Ageing and productivity: Combination of channels

1 Smaller workforce induce investment in capital and technology +

2 Consumption composition low or high productivity industries? ?

3 Ageing less mobility? Or innovators move to high growth areas? ?

4 If older less skilled skill-intensity of production in older region declines -5 Age-diversity in workplace synergistic or polarising? ?

So, what does the empirical literature say?

Productivity by age at individual level: Psychometric tests

0 10 20 30 40 50 60 70 80 90

25

30

35

40

45

50

55

60

65

70

Age

Sco

re

Chomik et al. (2018)

Productivity by age at individual level: Skill tests

170

190

210

230

250

270

290

310

Engl

and

(U

K)

Gre

ece

Un

ited

Sta

tes

Slo

vak

Rep

.N

ew

Zea

lan

dSw

ede

nN

orw

ayD

enm

ark

Jap

anEs

ton

iaA

ust

ria

Au

stra

liaC

anad

aC

zech

Rep

.G

erm

any

Po

lan

dIr

elan

dJa

kart

a (I

DN

)N

eth

erl

and

sIt

aly

Be

lgiu

mFr

ance

Isra

el

Spai

nSl

ove

nia

Turk

ey

Fin

lan

dC

hile

Ko

rea

Sin

gap

ore

Mean numeracy

25-34 year-olds 55-65 year-olds

0

10

20

30

40

50

60

70

80

Turk

eyG

reec

eU

nit

ed S

tate

sC

hile

Slo

vak

Rep

.P

ola

nd

Isra

el

Engl

and

(U

K)

Au

stra

liaIr

ela

nd

Ne

w Z

eala

nd

Can

ada

Slo

ven

iaEs

ton

iaC

zech

Rep

.G

erm

any

Be

lgiu

mN

eth

erla

nd

sA

ust

ria

No

rway

Swed

en

Jap

anD

en

mar

kK

ore

aSi

nga

po

reFi

nla

nd

IT / computer skills

25-34 year-olds 55-65 year-olds

But older people today are not older people tomorrow

2010

2015

2005

20102015

2000

2005

20102015

1995

2000

20052010

2015

1990

1995

20002005 2010 2015

1990

19952000 2005 2010 2015

19901995

2000 2005 2010 2015

1990 19952000 2005

20102015

0%

10%

20%

30%

40%

50%

60%

70%

80%

25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65+

2015 Tertiary educational attainment by age group, Singapore,

1990-2015,

Cross-section v Cohort

Productivity by age at individual level: Work assessments

SupervisorPeers

ObjectiveSelf-rated

Supervisor or objectiveSelf-rated

Supervisor or objective

General - Supervisor or peersGeneral - self-rated

Directed at others - Superv. or peersDirected at others - Self-rated

Directed at org - Superv. or peersDirected at org - Self-rated

Directed at tasks - Supervisor or peersDirected at tasks - Self-rated

Compliance - Self-ratedInjuries - ObjectiveInjuries - Self-rated

Supervisor or peersSelf-rated

Self-rated

Self-rated

Supervisor or peersSelf-rated

General - ObjectiveGeneral - Self-ratedSickness - ObjectiveSickness - Self-rated

Non-sickness - ObjectiveNon-sickness - Self-rated

-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

Core tasks

Creativity

Training

Org.Citizen Behaviour

Safety

Counter-productive behaviour

Workplace aggression

Workplace substance abuse

Tardiness

Absenteesm

…an

d le

ss li

kely

to

d

o t

he

bad

th

ings

Old

er

peo

ple

mo

re li

kely

to

do

go

od

th

ings

…

Correlation with age (meta-analysis 400+ studies)

Productivity by age at team, plant, and firm level

Aubert and Crepon 2004 (71k French firms)

Crepon et al. 2002 (12k US manuf. firms)

Gelderblom and de Kooning 2002 (78k French…

Grund & Westergard-Nielsen 2008 (30k Danish…

Haegeland and Klette 1999 (7k Norwegian firms)

Hellerstein and Neumark 1995 (1k Israel firms)

Hellerstein and Neumark 2004 (3k US firms)

Hellerstein et al. 1999 (3k US firms)

Ilmakunnas et al. 2004 (4k Finnish firms)

Prskawetz et al. 2005 (95k Swedish firms)

Prskawetz et al. 2007 (34k Austrian firms)

Schneider 2006 (1k German manuf. firms)

Lallemand and Rycx 2009 (500 Belgian ICT firms)

Ours and Stoeldraijer 2011 (14k Dutch manuf.…

20 30 40 50Age

Peak productivity age, by study

Productivity by age at the aggregate level

Study Coverage Result

Lindh and Malmberg (1999) OECD 1950-1990 50-64 age group affects labour productivity positively

Feyrer (2007, 2008) 87 countries 1960-1990 40-49 age group affects TFP positively

Liu and Westlius (2016) Japanese prefectures 1990-2007 40-49 age group affects TFP positively

Maestas et al (2016) US states 1980-2010 60+ age group affects labour productivity negatively

Ozimek et al (2017)US states-industries, 2000-2015

Teams in HR company. 2013-2016

Share of 65+ affects productivity of state-ind negatively

Share of 65+ affects wage of team negatively

Acemoglu & Restrepo (2017) 169 countries 1990-2015Ageing (ratio of 50+/20-49) affects GDP per capita growth

positively

In fact, we may be seeing induced technological change

VNMIDN

PHL

MYS

SWE

GBR

BRA

ESP

POL

AUS

DNK

PRT

USA

NZLTHA

FRA

ITA

DEU

KOR

NDL

HKG

FIN

SGP

0

10

20

30

40

50

60

0 5 10 15 20 25 30 35 40 45 50

Change in r

obots

per

mill

ion h

rs w

ork

ed 1

993 -

2015

Change in ratio of population age 50+ to 20-49, 1990-2015

Workforce ageing and use of robots,

1990-2015

And we’ve seen health (education & IQ) improvements

69

JPN, 73

67

SGP, 74

60

CHN, 68

63

KOR, 71

50

KHM, 60

56

IDN, 63

46

LAO, 59

63

MYS, 66

1990 1995 2000 2005 2010 2015

Healthy life expectancy at birth, both sexes, 1990-2016

Interim conclusions

1.Asia is experiencing rapid ageing and

technological change

2.These can affect productivity growth via physical

and human capital and technological change

3.Evidence of productivity by age is mixed and

confounded by changes in cohorts and selection

Skill-biased technological change

TU

R

FR

A

UK

SW

E

FIN

DN

K

DE

U

NZ

L AU

S

US

A

PH

L MY

S KO

R

TH

A

IND L

KA BT

N

KA

Z

MN

G

PA

K

CH

N

-1.5

-1

-0.5

0

0.5

1

Pe

rce

nta

ge

po

ints

Low-skilled occupations (intensive in nonroutine manual skills)

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

High-skilled occupations (intensive in nonroutine cognitive and interpersonal skills)

Annual average change in employment share, by occupation skill, 1990-2012

Age-biased technological change: Empirical findings

Study Coverage Effect of technological change

Peng et al

(2017)

Employee-employer data

8 EU countries

1970-2007

Older (and younger) high skill gain wage-share

Older low skill gain empl-share

Collective bargaining reduces age-bias of tech

Hujer and Radic

(2005)


West Germany

mid-1990s

Older high skill lose employment share

Older low skill gain employment share

Beckmann

(2007)


West Germany

mid-1990s

Older lose empl-share

Aubert et al

(2006)


France

mid-1990s

Older more likely to leave, less likely to be hired

Older lose wage-share

Ahituv and Zeira

(2010)

HRS individual data

USA

mid-1990s

Drives many older people to retire

Drives those with wage gains to retire later

People affected by

skill biased tech

change

People that find

skill updating

more difficult

Older people

Age biased

tech change?

Conceptualising age-biased technological change

Skill types overall: More non-routine; less routine

Routine manual

Routine cognitive

Non-routine analytic

Non-routine interactive

Non-routine manual

70

80

90

100

110

120

130

140

19

60

19

67

19

74

19

81

19

88

19

95

20

02

20

09

USA,

1960-2009

Routine manual

Routine cognitive


Non-routine manual


94

95

96

97

98

99

100

101

102

103

104

19

60

19

67

19

74

19

81

19

88

19

95

20

02

JPN,

1960-2005



Routine cognitive

Routine manual

Non-routine manual

94

95

96

97

98

99

100

101

102

103

104

20

00

20

02

20

04

20

06

20

08

20

10

20

12

20

14

CHN,

2000-2015

Skill intensity changes by cohort

Non-routine

cognitive

Routine cognitive

Manual Skills

75

80

85

90

95

100

105

110

115

2001 2002 2003 2004 2005 2006 2007 2009 2010

Evolution of skills

intensity of jobs for

cohort born <1955,

MYS 2001-2010

Non-routine

cognitive

Routine cognitive

Manual Skills

75

80

85

90

95

100

105

110

115

2001 2002 2003 2004 2005 2006 2007 2009 2010

Evolution of skills

intensity of jobs for

cohort born >1974,

MYS 2001-2010

Employment automation potential by age

CHN

HKG

IDN

JPN

MYS

KOR

44%

46%

48%

50%

52%

54%

56%

15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60+

Share of employment subject to potential automation, by age, 2016 or latest

Interim conclusions

1. Relationship between technology and employment is complex

2. The overall adverse effect of technology on employment is

overstated

3. Technology often impacts the employment and wages of middle

and lower skilled workers

4. The pattern of employment among older cohorts suggest that

they will be disproportionately caught up in this impact

5. Non-adaptability of a relatively aged workforce may slow growth

and lead to displacement

Specific technologies and ageing pressures

Long Term Care• Increasing demand, highly labour intensive

• Tech could free up formal & informal labour

• Examples include AT, ICT, Robots

• Functions include Monitoring, Connectivity, Perform tasks

• Tech inducement externalities to contribute to productivity growth

Healthcare• Health tech and ageing intimately related

• Fiscal challenge

• Example includes Telehealth in NCD (chronic disease) management

Digital Identification Technology• Cheaper/easier/safer than traditional centralised record keeping

• Induced by social protection/insurance requirements

• Likely large externalities via formalisation

Policy needs

Production and infrastructure• Sources of capital

• Technology infrastructure

• Narrow infrustructure inv. gaps in Asia

Human resource policies• Education, training, and lifelong learning

• Getting foundation right

• Targeting lifelong learning

• Migration policy

Addressing inequality• Life course approach to intervention

• Social protection

A new program:

Technology

Adjustment

Assistance

Research needs

Data and modelling capability• Asia lacks the data available in OECD

• Examples include employee-employer data, skills data by age

• New modelling needed on demog. effects on production, consumption, trade

Workplace and job design• Work-related factors promoting/inhibiting successful ageing

• Job/task design and cognitive ability

Links between demography and productivity• Strength of inducement effect

• Strength of consumption composition and endowment effect (eg. G-Cubed)

• Effectiveness of re-training programmes to offset age-biased tech change

Cost and benefit analysis of specific tech• Physical and tech infrastructure (e.g., DIT)

In association with:

demographic and technological change: two megatrends ... · 5/18/2014 · sgp 0 10 20 30 40 50 60...

Documents