Caveat artifex

By Robert Hickson 09/02/2014

The title is my pig latin for let the worker beware.

In the US The Atlantic picked up on the US Bureau of Labor Statistic’s projections of the fastest growing jobs over the next decade. The latter conclude that the fastest growing jobs will be in healthcare, office administration, retail sales, and the food workers. These are driven by demographic changes and economic expectations.

US Bureau of Labor Statistic's job projections 2012-2022
US Bureau of Labor Statistic’s job projections 2012-2022

[Click on image for a larger view]

But The Atlantic notes that healthcare spending in the US is already growing more slowly than expected, which will affect the job projections if that trend continues. In a separate article they point out that the Bureau’s projections a decade ago were wide of the mark, so the results need to be viewed as indicative only

The projections, according to the Atlantic, also seem to take insufficient note of recent robotic and computational trends that could make some of the “fast growing” jobs obsolete for humans. They refer to the modelling by Benedikt & Osborne [Pdf] from the Oxford Martin School who predict retail, office, and service-related jobs have high probabilities of being automated “over the next decade or two.”

“How susceptible are jobs to automation?”. Benedikt & Osborne (2013)

These projections conform to other predictions of a hollowing out of the middle classes.

Benedikt & Osborne are uncertain of the timeframe for their automation impacts, so the Bureau of Labour Statistic’s projections for the next 10 years may not be that greatly influenced by them. And the former’s model may be just as fraught as the standard economic ones.

As I’ve noted previously, while some job types may be taken over by robots or computers, other types of work are usually created. The big issue is whether they’ll be meaningful jobs, and ones that pay reasonably well.

What about the future job market in New Zealand? The Ministry of Business, Innovation & Employment did a 10 year employment outlook in 2012 . It doesn’t drill down as deep into job types as their US counterpart. They list the top 25 most rapidly growing jobs in five year blocks (2011 to 2016, and 2016 to 2021), so I’ve calculated the overall growth projections from 2011 to 2021 (Bubble size gives an indication of overall number of jobs in 2021).


Three main things to note. Firstly, many of the job types also fall into the potentially automatable class. If Xero’s business model succeeds, for instance, there will be much less demand for accountants. Similarly, if even more people shift to buying goods online.

Secondly, the health care sector is less a feature in the NZ projections.

Thirdly, the projected rates of growth for NZ are at least an order of magnitude less than in the US, so our economic growth is projected to be pedestrian by comparison.

I haven’t found an analysis of previous NZ projections versus actual employment, but my guess is that they are probably not that close. The dairy boom, and Christchurch earthquakes, for instance, were unlikely to be factored into them.

The Treasury has a paper comparing projected vs actual economic performance [Pdf]. They generally only look a couple of years ahead but even then the error rates increase the further out they project.

So don’t rely too much on government projections in times when significant change is going on.

What types of jobs will I be suggesting to my children? That will depend on their skills and interests of course. But ones that are likely to be challenging and interesting, and involve great interpersonal relations, judgement and/or creativity would be better bets.