LEAPFROGGING INTO THE UNKNOWN
SOUTHEAST ASIA AND THE FUTURE OF JOBS
Lukas Schlogl
12 September 2019, Bangkok
University of Vienna @LukasSchlogl
University of Vienna @LukasSchlogl
Outline
1. Trends in structural transformation (ST)Tertiarization and convergence
2. Future employment challengesTechnological catch up during late development
3. Conclusions
University of Vienna @LukasSchloglUniversity of Vienna @LukasSchlogl
University of Vienna @LukasSchlogl
Employment across sectors
0
25
50
75
100
1990 2010 2030 20500
25
50
75
100
1990 2010 2030 2050
0
25
50
75
100
1990 2010 2030 2050
HIC LIC LMC UMC
(ILO modelled estimates. Dotted lines are projections based on the linear trend. Country aggregates are based on population-weighted averages.)
Agriculture Industry Services
University of Vienna @LukasSchlogl
Value addition across sectors
0
25
50
75
100
1990 2010 2030 20500
25
50
75
100
1990 2010 2030 2050
0
25
50
75
100
1990 2010 2030 2050
HIC LIC LMC UMC
(ILO modelled estimates. Dotted lines are projections based on the linear trend. Country aggregates are based on population-weighted averages.)
Agriculture Industry Services
University of Vienna @LukasSchlogl
Structural convergence
• Expansion of service-sector jobs at expense of industrial (HICs) and agricultural (DCs) jobs. Little potential left in HICs for cross-sector ST.
• Industrial employment shares increasingly detached from development level. Congo, Bangladesh, Mexico, Finland have ~20%-25% employment in industry.
• In OECD, industrial employment peaked in 60s/70s at ~30%. If path is followed, UMICs reach Clark turning point in 2030s, LMICs in 2040s.
• Structural change fastest in LICs (Merotto, Weber, & Aterido, 2018)
University of Vienna @LukasSchlogl
The trend in Southeast Asia
Movement of labour across sectors following similar pattern:
• Employment in agriculture contracting
• Service sector expanding
• Industrial employment mixed• Most of SEA still shows expanding employment in industry
• Rich SEA contracting: Malaysia reached ‘Clark turning point’ in late 90s, Brunei in early 90s, Singapore earlier. Singapore service sector share now higher than UK or US.
So, business as usual?
University of Vienna @LukasSchlogl
Enter the Robots.
University of Vienna @LukasSchlogl
Automatable work across SEA
Indonesia
Malaysia
Philippines
Singapore
Thailand
0
30,000
60,000
90,000
40 50 60
GN
I pe
r ca
pit
a
Malaysia
Thailand
Cambodia
0
10,000
20,000
30,000
50 60 70 80 90
Share of the labour force susceptible to automation
McKinsey (2017)World Bank (2018)
University of Vienna @LukasSchlogl
% of automatable jobs (estimates)
University of Vienna @LukasSchlogl
WB (2016) MGI (2017)Cambodia 78%Indonesia 52%Malaysia 68% 51%Philippines 48%Singapore 44%Thailand 72% 55%
University of Vienna @LukasSchlogl
Industrial robot density low but growing fast
Indonesia
Malaysia
Philippines
Singapore
Thailand
R² = 0.6
0.6
0.7
0.8
0.9
1
0 200 400 600 800
Hu
man
Dev
elo
pm
ent
Ind
ex
Robots per 10k manufacturing workers (IFR estimate for 2016)
University of Vienna @LukasSchlogl
Robot sales: prices↓ demand↑
3.1 million industrial robots in operation globally by 2020
University of Vienna @LukasSchlogl
A trend towards service automation
• Indonesian SOE Jasa Marga: phased in e-toll system in 2017 (GOI mandated it)
• Unions feared loss of 20,000 jobs –JSMR denies layoffs
• “Although we do not expect the majority of workers to leave JSMR (…), personnel expense growth should still be limited, as the A-Life [task rotation] program should reduce the need for additional recruitment in the coming years.” (Mirae Asset 2018)
University of Vienna @LukasSchlogl
Meanwhile, 40 miles from Silicon Valley...
Toll Gates at Oakland Bay Bridge“Gutierrez and other collectors sometimes worry that the FasTrak lanes could replace tollbooth operators altogether. Gutierrez says it’s a question that has been coming up repeatedly at union meetings over the last three months. “The union claims that if it does come to that, they’ll help us get to another job site,” Gutierrez says. But CalTrans officials say—and Gutierrez agrees—that it is just a rumor. No official proposal has been laid out or discussed.” (Waheed 2012)
University of Vienna @LukasSchlogl
Late developers, early adopters
McDonald’s e-Kiosks in Indonesia
• Weak labour protection might make adoption of labour-displacing technology easier in countries where labour costs would otherwise provide little incentive for it
University of Vienna @LukasSchlogl
Late developers, early adopters
SEA: Rapid rise of gig economy HICs: Banning Uber…
University of Vienna @LukasSchlogl
What’s the trouble?
University of Vienna @LukasSchlogl
Lots of agricultural work, despite ST
OECD ASEAN-5
0
25
50
75
100
1991 1996 2001 2006 2011 2016Agriculture Industry Services
(%, OECD, Employment share)
0
25
50
75
100
1991 1996 2001 2006 2011 2016Agriculture Industry Services
(%, ASEAN-5 (simple ave.), Employment share)
University of Vienna @LukasSchlogl
University of Vienna @LukasSchlogl
Premature deindustrialization?
0
5
10
15
20
25
30
35
40
6.5 7 7.5 8 8.5 9 9.5 10 10.5
Manufacturing (% of value added, constant prices)
Korea(1960-2011)
Malaysia(1970-2011)
China(1960-2010)
Indonesia(1960-2012)
GDP per capita (2011 PPP dollars, log)
Philippines(1971-2012)
Thailand(1951-2011)
Source: GGDC 10 Sector Database
University of Vienna @LukasSchlogl
University of Vienna @LukasSchlogl
R&D expenditure
0
1
2
3
OECD (2015) Malaysia(2015)
Thailand(2015)
Vietnam (2013) Philippines(2013)
Indonesia(2013)
(%, R&D expenditure/GDP)
University of Vienna @LukasSchlogl
University of Vienna @LukasSchlogl
Skills
0
10
20
30
40
50
60
70
80
OECD (2016) Thailand(2015)
Malaysia(2016)
Philippines(2017)
Vietnam (2016) Indonesia(2016)
(%, Tertiary school enrolment rate)
University of Vienna @LukasSchlogl
Source: World Bank
University of Vienna @LukasSchlogl
Demographic dividend?
HICs SE Asia
Source: UN World Population Prospects
University of Vienna @LukasSchlogl
University of Vienna @LukasSchlogl
But: stable labour force participation
Source: ILO
0
20
40
60
80
100
1990 1995 2000 2005 2010 2015
Indonesia Malaysia Philippines Singapore Vietnam Cambodia Thailand
University of Vienna @LukasSchlogl
Technological catch-up or automation creep?
• Gerschenkron’s (1962) ‘advantages of backwardness’: “It makes sense for latecomers to utilise all the resources from the advanced world that they can acquire” (Mathews 2006). Today’s parlance: “successful economic development involves a leapfrogging process” (Burlamaqui and Rainer Kattel 2014). Also a major theme in NSE (Lin).
• Robotisation pressing in tradeable sectors where GVC suppliers face international competition and risk of Reshoring
But…
• Rodrik (2018, p. 14) argues that “The evidence to date, on the employment and trade fronts, is that the disadvantages [of new technologies] may have more than offset the advantages” for developing countries due to erosion of low-cost labour advantage.
• Brynjolfsson and McAfee (2014, p. 184) argue that the “biggest effect of automation is likely to be on workers (…) in developing nations that currently rely on low-cost labor for their competitive advantage”.
University of Vienna @LukasSchlogl
Late Development, Early Adoption
1. Directed technological change• Labour saving innovation driven by firms innovating in a high-skill, high-cost,
high-social-protection environment of HICs
2. Innovation as global public good• Innovation hard to contain in a hyper-globalised, digitally connected world
(‘automation creep’) unless concentrated political resistance
3. Undirected early adoption• Selective early adoption during late development despite low-skilled, low-
cost, low-social protection environment.
4. Specialisation towards comparative disadvantage?• Less inclusive development, a smaller middle class, less democracy?
University of Vienna @LukasSchlogl
The near term future of ST in SEA
… will be characterised by:
• Continued expansion of service sector employment
• MICs reaching Clark turning point, starting job deindustrialisation
• More technological convergence unless labour protectionism
… will raise the following questions:
• Should technological adoption be commensurate to level of econ development? Can you automate prematurely?
• Can the labour force be upskilled fast enough?
• Does automation require a specific (social) policy context?
Thank you!
Forthcoming with Palgrave as OA:
University of Vienna @LukasSchlogl
Disrupted Development. The Future of Industrialisation and Inequality in the Age of Automation (Schlogl & Sumner 2019)