Australia's $40 billion of education exports is a statistical trick

You've heard Australia exports billions in education services, but you won't believe where this number comes from

Thank you to the 3,000 subscribers who read Fresh Economic Thinking.

I want to ask a small favour. If you enjoy these articles, the best thing you can do to support me is to share them on your social media and forward the emails to your friends.

And don’t forget to subscribe. It’s free after all.

---

You’ve heard about Australia’s whopping great “education exports”, right?

“Education is our largest services export and the biggest product we don’t source from the ground,” Universities Australia Chief Executive Catriona Jackson said.

“The export income our universities help generate pays for essential services and underpins a higher standard of living for all Australians, regardless of where they live. Covid-19 halved the value of education as an export, but we are well on our way back to reaching, and hopefully surpassing, the $40 billion mark we recorded in 2019.”

But is it true?

There are many hidden measurement decisions in every economic statistic and education exports are no different.

Salvatore Babones is one of the few to have noticed some peculiar measurement issues around this education “export” figure.

International students are clearly important for Australia’s universities, but their importance to the economy as a whole is frequently overstated. One oft-quoted statistic is that educational exports have risen to become Australia’s third-largest export after iron and coal. That doesn’t really capture the full story, since exports in different sectors are reported at different levels of granularity.
…

Educational exports overtook receipts from all other travel (tourism, family, and business combined) in 2008, but are still smaller than Australia’s exports of agricultural or manufactured goods. Moreover, more than half of Australia’s reported educational exports (53.7% in higher education and 57.2% for the education sector as a whole) consists not of student fees, but of goods and services bought by students while in Australia. Since this spending is at least partly generated by income that students earn from working in Australia while studying, the true net value of education exports to the Australian economy is likely lower than the headline figures reported by the ABS and DET.

Leith van Onselen at MacroBusiness is another who has reservations.

I thought it was worth finding out for myself where the $40 billion education exports number comes from.

International students are the exception to the rule

The Australian Bureau of Statistics (ABS) has the near-impossible task of estimating education exports.

You might think it’s easy. But defining boundaries is difficult, especially for services exports that don’t see a physical product leave the country. There are often hidden but important assumptions hidden in our economic measurements, which is why I regularly comment on economic measurements.

---

Mysteries of inflation measurement

---

The ABS methodology explains how the ‘centre of predominant economic interest’ is used to determine whether a person is a tourist buying export services or a local engaging in the normal economic activity of a resident.

This is their explanation.

The concept of residency is vital for international trade statistics – including for international trade in services.

Residency is not based on location, nationality or legal status (e.g. citizenship). Instead, residency is defined by ‘the centre of predominant economic interest’ of an organisation or individual.

It is not always easy to determine the residency for an organisation or individual. Typically, residency is determined based on (or intent to engage in) significant economic activity. Economic activity can be deemed significant either by value, or by time (usually a minimum of one year).

The only exception to this residency guideline is for international students, who are deemed residents of their home economies for the duration of their study.

For example, a tourist from New Zealand who visits Australia for a month is classed as a non-resident and therefore their travel would contribute to international trade in services. If that tourist decided to stay in Australia to undertake study, they would continue to be classed as a non-resident and their education related travel would continue to contribute to international trade in services statistics. If instead, they chose to extend their holiday to Australia and stayed for longer than twelve months, they would transition from a non-resident to a resident for the purposes of international trade statistics. Their expenditure would no longer be captured in international trade in services statistics as it is assumed that they have set up a household within Australia and are contributing to the economy domestically.

Notice that for everyone but international students, if they are working and living in Australia the value of goods and services they consume is not an export. Only international students are treated this way, despite most of them working locally.

This is the statistical trick at the heart of the education exports figure.

How are these exceptional numbers estimated?

The ABS explains how education services exports (education-related travel credits) are estimated here.

For education-related travel credits, the number of students is determined by the number of people, in Australia, on a primary student visa (as reported by the Department of Home Affairs). The number of students is then multiplied by an average spend estimate from Tourism Research Australia (as per the other travel sub-categories). This estimate is supplemented by the addition of the total expenditure on course fees (of international students studying in Australia, sourced from the Department of Education, Skills, and Employment).

Let me convert this to a calculation.

Education exports = (# international students x average tourist spend) + education fees

It is not clear to me that applying an estimate of the average non-education tourism spending is a meaningful way to estimate the annual spending of international students residing and working locally.

Maybe it’s fine. I don’t know. It is unusual.

I emailed the ABS to see if I could find out the average spending estimate that goes into this calculation. They responded as follows.

We utilise detailed data provided by Tourism Research Australia that is not available to the public and are not able to release these details further.

That’s a shame. The key number that this widely cited education export statistic is based on is a secret.

I also provided the following scenario to clarify that local work incomes are not subtracted from the export estimate.

Are work incomes of students simply overlooked in this estimate because of the residency assumption? I.e. if students spend $30k per year but earn $40k per year while studying, the work income is ignored and the $30k spend is treated as an education-related export.

They responded.

The incomes of students in Australia are not considered in the calculations for student expenditure on goods and services. In the example provided, only the $30k spend on goods and services is treated as an education related export.

Which is exactly as described in the methodology publication. Incomes don’t matter, as students are the exception to the ‘centre of predominant economic interest’ principle in services trade statistics.

Reverse engineering the per-student spend

I’m going to look back to 2019, pre-COVID, to see if we can get a feel of the numbers used. Department of Home Affairs reports that there were 480,000 student visa holders in Australia at the end of that year. Tourism Research Australia reports that total spending was $45 billion that year for 276 million international tourist nights or about $160/night.

The ABS measure of goods and services spending by international students that year was $22 billion, so that’s $45,000 per student per year. If we assume they reside in Australia for 300 days a year, the nightly spend is $150, which is close to the Tourism Research Australia figure.

University fees for international students can be up to $40,000 (at least at The University of Sydney), which seems right as the fees in the ABS data are $33,000 per student.

What this means is that wherever the ABS figures are coming from, it amounts to assuming that the average international student is spending about $78,000 per year, which is a touch more than the average Australian full-time employee can with their after-tax income.¹ And all of this assumed spending is counted as an export, even if the international students work locally and stay for many years.

What about international students working?

Around 52% of international student work.

If that makes Australia their ‘centre of predominant economic interest’ then none of their spending would normally count as a services export, including their university fees. Only a strange statistical assumption has created a “$40 billion export industry” out of one that is probably closer to a $15-20 billion one. Or less. We don’t know.

It probably doesn’t matter whether international students are measured as exports for the substance of the policy debate around international students and the education sector.

But the halo the education sector gets from the line that “exports are good and we export $40 billion” offers enormous political advantages. With billions in fees on the line, the higher education sector has a huge incentive to sell the idea politically that this is a massive export boom for the country rather than an immigration policy where they clip the ticket.

1. The average full-time salary is about $100,000 and the tax on that is about $24,000, so that’s about $76,000 of spending possible from such an employee.