Nicholas Sheppard
Research
Programming languages

About six months ago, I wrote about Peter Cappelli's arguments concerning the existence or otherwise of skills shortages, and what to do about them. Cappelli was writing primarily about the United States, and about ten years ago, so I decided to look into the evidence that exists in Australia now, with particular focus on computing professions.

I reviewed articles referring to skills shortages published between 2019 and 2021 (inclusive) in the Australian Computer Society's Information Age and the Australian Broadcasting Corporation's News site. In each case I followed the links given to studies addressing the existence of skills shortages. Sometimes these studies themselves referred to earlier studies, which I also collected.

Out of these studies, I identified ten reports making claims about the supply and demand for skills based on evidence original to that study (i.e. not simply repeating a claim from an earlier study). Of these, I was able to download eight:

RMIT University commissioned another study for which a summary of results is available on the university's web site, but I wasn't willing to sign up to SMS notifications of RMIT course offerings in order to download the full report.

The reports vary in their level of detail, the body of skills that they studied, and their methods for ascertaining what skills might be in short supply. The remainder of this article makes a few broad observations on the studies, noting their differences where necessary, but I won't attempt to review the methodology or results of each one in detail.

What did the studies find?

The studies report at least three distinct but related kinds of findings:

  • that the demand for technology workers is growing, as in the Digital Pulse and RMIT reports;
  • that the demand for technology workers is forecast to outstrip supply, as in the Accenture and Digital Pulse reports; and
  • that a shortage exists now, as in the Skills Priority List, the APS' Workforce Strategy, and the Hays report.

That the demand for technology workers is growing is not of itself an indicator of shortages, since the supply of technology workers is also growing (which the Accenture and Digital Pulse reports both acknowledge), and a shortage will only exist if the former exceeds the latter for a significant period of time.

The forecasts are problematic in that none of reports reviewed here provided enough information to definitively identify the projections for demand that they used. The National Skills Council produces employment projections by extrapolating current trends according to a statistical model but acknowledges that all such exercises are "subject to an inherent degree of uncertainty".

Studies differ on whether and what shortages exist now. The Skills Priority List classes occupations including "developer programmer" and "software engineer" as being in shortage, whereas Deloitte's Path to Prosperity says that the supply of "coding/programming" skills closely matches demand and Digital Pulse says that the supply of "technology workers" increased during the period of its study while the number of job postings for the same workers declined. In the rest of this article I'll try to identify what the studies have in common and look at some possible explanations for why they might differ.

Differences in perspective

One of the most striking observations when reading news articles about skills shortages is that articles reporting the views of employers invariably report that skills shortages exist, while articles reporting the views of workers invariably report that they're unable to find work. Virtually all of the data analysed in the studies reviewed here comes from employers, either by surveying employers or by reviewing job advertisements, with only the National Skills Council making a broad reference to consulting with unions.

Since I searched for articles specifically referring to shortages, I wasn't likely to find articles about employers being able to find the workers they need, or workers being able to find jobs, even if news organisations published such "news". But several explanations exist for why employers might report shortages at the same time as workers report being unable to find work.

Firstly, employers presumably prefer workers to have more skills rather than fewer, while workers presumably prefer to think of themselves as being highly skilled. So employer reports may be biased towards looking for more skills, while worker reports may be biased towards skills being available.

Alison Pennington and Jim Stanford (2019) point to a more subtle phenomenon: after a period of high unemployment, employers become used to having their pick of the labour available; but as unemployment decreases, employers find getting the skills they want more difficult, which they perceive as a "shortage", especially if they're resistant to raising wages to meet what the new market demands. In Cappelli's analysis, this corresponds to employers who find that skills are available, but who are not willing or able to pay the going rate for the workers who possess them, thus experiencing a shortage of skills at the price they are willing offer even where skills actually exist in the market. (Conversely, workers might perceive increases in unemployment as a "skills oversupply", but unemployment was decreasing at the time Pennington and Stanford were writing.)

Another obvious explanation is that the skills employers are seeking are not the same as the ones that workers possess. To assess this, we need to look more carefully at what skills are being referred to in each study—but this turns out to be very difficult.

What skills are in shortage?

Studies like the Australian Industry Group's Skilling report say things like "employers continue to report difficulties recruiting employees with STEM skills" (p. 15), based on an analysis of occupations ranging from "manager" and "professional" to "technicians" and "sales workers". But even within a category like "professionals", vast differences exist in the "STEM skills" required by lawyers, engineers, health professionals, and so on, so it's hard to say precisely what the employers are finding difficult to recruit.

The Skills Priority List does try to evaluate shortages on an occupation-by-occupation basis, but its output is very coarse: any given occupation is either said to be in shortage or that it is not, so we don't know if the occupation needs only a few more university graduates or if a vast gulf exists between demand and supply. The National Skills Council's definition of a "shortage"—that skills can't be found at the price employers are willing to pay—is also vulnerable to the criticisms made by Pennington and Stanford above, and by Peter Cappelli in my previous article.

One useful distinction that appears in a number of studies, however, is the one between "digital skills" (or "digital literacy") and what we might call "STEM skills" proper. The former refers to skills required by all sorts of workers who use computer software, electronic devices, and so on, but who are not themselves involved in the design or construction of computer hardware and software. The studies I reviewed generally agree that workers do not have enough skills of this sort; but (as already mentioned) they differ on whether there's a shortage of engineering and mathematics skills.

Another useful distinction (particularly for someone who works in a university) is between broad character traits typically developed through experience versus specific disciplinary skills learned in formal education. Deloitte's Path to Prosperity, for example, says that the biggest shortages are in "customer service" and "organisation and time management", while little or no shortage exists in "coding/programming". The Accenture report specifically says that shortages exist in experienced workers (my emphasis) while Pennington and Stanford say that shortages exist in "complex problem solving skills" that (as I interpret their meaning) require mastery of a vast body of knowledge acquired over many years of working in field.

The problem with skills learned by experience is that they can't be easily manufactured. Formal education teaches students codified knowledge like computer programming, data analysis techniques, and so on, but it cannot give them the two or three (or five or ten) years' experience often asked for in job advertisements. A difficult question is, who's going to pay for someone to gain two or three years experience?

Finally, one might conjecture that the reason skills like "time management" and "problem solving" appear at the top of skills lists is simply that all but the most routine jobs require managing time and solving problems, whereas programming is required only by a relatively small set of specialised workers. So the demand for "problem solvers" will always greatly exceed the demand for programmers in terms of absolute numbers—and someone good at solving computer programming problems, say, might not be so good at solving chemical engineering problems.

References

A. Pennington and J. Stanford (2019). The Future of Work for Australian Graduates, The Centre for Future Work at the Australia Institute. Available from https://www.futurework.org.au/the_future_of_work_for_australian_graduates.