07.12.2015 Views

TECHNOLOGY AT WORK

1Oclobi

1Oclobi

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

February 2015<br />

Citi GPS: Global Perspectives & Solutions<br />

61<br />

respectively. 83 By contrast, occupations such as Computer Network Specialists and<br />

Web Developers, of which many work in Information, already employ 143,000 and<br />

141,000 workers, respectively, and are expected to add another 94,200 jobs before<br />

2022, according to the US Bureau of Labor Statistics (BLS) occupational<br />

projections. This illustrates an important challenge facing companies in most<br />

industries: as low-skill jobs are being replaced and new high-skill jobs created, they<br />

will need to invest substantially in up-skilling their workforce.<br />

No single industry is completely immune to<br />

the expanding scope of automation<br />

Outside of the US, studies have found that<br />

upwards of 54% of EU jobs are at risk<br />

However, these estimates are likely to be<br />

systemically upward-biased<br />

Other industries that are at low-risk of automation include Management of<br />

Companies and Enterprise, and somewhat surprisingly Agriculture, Forestry, Fishing<br />

and Hunting, most likely reflecting that most jobs that can be automated in<br />

agriculture already have been. The potential scope for further automation is<br />

substantially larger in Manufacturing, where 62% of jobs are still at risk.<br />

No single industry is completely immune to the expanding scope of automation.<br />

Even in some relatively skilled industries such as Finance and Insurance, 54% of<br />

jobs are at risk. Traditional low productivity industries such as Healthcare, Education<br />

and Government are facing future transformations too. At a time when government<br />

budgets are under pressure, such productivity gains would be a blessing.<br />

Countries at Risk<br />

To be sure, some countries are better positioned to adapt to the expanding scope of<br />

automation than others. Because different occupational classifications exist across<br />

countries, however, meaningful direct comparisons are notoriously difficult to make.<br />

In a recent study, for example, Bruegel translated our findings for the US to 28<br />

European Union (EU) countries, using 22 (instead of 702) more aggregated<br />

International Standard Classification of Occupations (ISCO) job categories for which<br />

the International Labor Organization (ILO) provides consistent data. Doing so, they<br />

find that 54% of EU jobs are at risk of automation, spanning 47% in Sweden and<br />

62% in Romania (see Figure 49).<br />

Nevertheless, while the ILO data used by Bruegel is comparable across the<br />

countries they examine, the relatively crude occupational classification being used<br />

is likely to be systematically upward biased. Because many occupations with very<br />

different probabilities of automation often fall into the same broader category, and<br />

the employment share in occupations that have a higher probability of automation is<br />

likely to be relatively low, the Bruegel study most likely overstates the share of jobs<br />

at risk relative to the US. This is also suggested by other studies using more<br />

detailed occupational classifications for single European countries. According to a<br />

study by The Research Institute of the Finnish Economy (ETLA), examining 410<br />

detailed occupations, about 37% of jobs in Finland are at risk of automation.<br />

Similarly, when translating our findings to the United Kingdom, analysing 369<br />

occupations, we found that 35% of UK jobs are highly susceptible to automation,<br />

and in skilled places like London that figure is even lower (29.5%).<br />

83 Based on BLS employment figures for 2012.<br />

© 2015 Citigroup

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!