FactCheck: Demographer Dr Liz Allen gets facts wrong on optimum level of immigration

Dr Liz Allen is a demographer at ANU and a regular commentator on population issues for the ABC. The ABC recently announced Dr Allen as one of their ‘top 5 humanities researchers‘, which apparently means she gets to work at the ABC for a few weeks and presumably gets more gigs as an expert commentator.

Dr Allen is a supporter of the trend toward a ‘Big Australia’, which implies continued rapid growth from our current population of 25 million to somewhere between 40 and 45 million by 2060, and presumably continuing to grow further after that. Currently 62 percent of Australia’s annual population growth is driven by immigration.

In one of her latest articles for The Conversation web site, ‘Migration helps balance our ageing population – we don’t need a moratorium‘, Dr Allen argues against what she claims is a proposal by WA Senator Dean Smith for a ‘moratorium’ on immigration. The title of the article is itself misleading, because Dean Smith has not called for a moratorium. What he has called for is a year-long Senate inquiry into population which would give ordinary people a chance to have a say on the direction in which Australia is heading. The article by Michelle Grattan referenced by Dr Allen states that:

Smith said “moderation” of the intake was important. “We need to perhaps give ourselves some time to breathe, some time to pause and reflect, to make sure the predictions are the best they can be and if they’re not – let’s correct that. Importantly, to make sure the infrastructure spending and public confidence is maintained”.

A ‘moderation’ of the migrant intake is not the same as a moratorium – the latter implies a complete (temporary) halt or suspension. So Dr Allen’s reference to a moratorium is a misquote and a straw man.

Even more significantly, Dr Allen gets her facts wrong when she goes on to claim that Australia’s current level of immigration is ‘about right’. After presenting some population growth scenarios from the Australian Bureau of Statistics showing different levels of immigration, Dr Allen concludes that our current level of immigration is pretty much at the optimal or ‘goldilocks’ point. Dr Allen bases this judgement on modelling conducted by Professor Peter McDonald and Jeremy Temple, for the Department of Immigration and Citizenship in 2010. According to Dr Allen:

The latest available research suggests immigration levels at “about 160,000 and 210,000 seem to have the ‘best’ impact by 2050 on ageing of the population and the rate of growth of GDP per capita”.

and

Any changes to the migration program should be considered alongside evidence. Available evidence shows Australia’s current migration program intake (190,000) is about right for Australia.

This is not the first time Dr Allen has used the findings from this report to argue her point. In an ABC Radio National Drive interview with Patricia Karvelas (and William Bourke) on 19 February 2018, Dr Allen said the following:

Interviewer:    Liz, what’s your take on all of this? Because we’re hearing increasing numbers of politicians now raising issues around Australia’s immigration numbers, I think about 190,000 migrants come into the country, 18,000 refugees on top of that every year. Is that a sustainable number?

Liz Allen: Look, I’m going by the evidence and the evidence that we have at the moment suggests that what we’ve been running at in the last period of time in terms of immigration intake has been at an optimal level to maximise the benefits of having migrants come to Australia. But more importantly I do concede that there is a point at which we get to diminishing return and that is we’ve not got to that point yet, that’s over 210,000 immigrants each year so we’ve not exceeded that place where we’ve got too many people coming in and I caution that when we have these conversations we’re often having these conversations based on emotion in a vacuum where evidence is not at all used or considered in – with regards to the population where we are now and where we’re going.

Interviewer: So Liz, you said 210,000 a year is the figure that – which you think is too high, is that what you’re saying?

Liz Allen: That’s based on research that was done by Peter McDonald and Jeremy Temple. The range they’ve suggested is between about 160 to 210, 220 annual intake and that’s based on as I said the idea of a diminishing return but anything above that upper limit or anything below that limit, lower limits actually doesn’t benefit the country in a way that we would hope it would and that’s in terms of the age structure, we’re an ageing population and so we’re getting to a point now where we want to avoid any adverse consequences of having a lower proportion of the population in the workforce contributing to income tax. So immigration has an ability to offset those adverse circumstances so – but as I said there are diminishing returns either side of that window. We’ve got to get it right, yes but let’s be led by evidence.

Similarly, in a news.com.au article on 13 March 2018, entitled What would Australia look like with a population of 36 million?, Dr Allen is reported as saying:

Evidence shows that the optimal level for Australia, given the population characteristics, is between 160,000 and 210,000.

So Dr Allen’s reliance on this ‘evidence’ is not a one-off. She has repeatedly made this claim to back her view that our current level of immigration is optimal.

Let’s look more closely at what the cited reference for these statements (McDonald and Temple 2010) actually says. The figures that the authors were citing (160,000 to 210,000) are for Net Overseas Migration (NOM). The figure cited by Dr Allen for our current migrant intake, of 190,000, is the ‘migration program planning level’, or permanent migrant intake quota, from the Department of Home Affairs.

However, the permanent migrant intake quota is not the same as the NOM. According to the ABS, NOM is defined as “the net gain or loss of population through immigration to Australia and emigration from Australia.” It is larger than the permanent migrant intake quota because NOM includes humanitarian visas (lately 18,000-20,000 per year) as well as temporary immigrants (for example, students, 457 visa holders) who reside in Australia for 12 months or more. In fact, the latest estimate for the NOM to end December 2017 is 240,400.

Dr Allen is therefore incorrect to cite the 190,000 quota figure as a basis for comparison with McDonald and Temple’s range of 160,000 to 210,000. In fact, the correct figure to use is 240,400, which is well above the claimed ‘optimal’ range.

Australia is already taking well above the claimed optimum annual number of immigrants as identified in the modelling cited by Dr Allen. Dr Allen has simply got it wrong with her conclusion that the current level of immigration is in the sweet spot suggested by McDonald and Temple (2010).

As well as getting the numbers wrong, Dr Allen places far too much credence on the claim that 160,000 to 210,000 NOM is somehow ‘optimal’. The authors of the report (McDonald and Temple 2010) themselves make a very clear disclaimer:

At the same time, increases to migration add constantly to the population and this increases the burdens associated with the provisioning and servicing of a growing population. This gives rise to the question of balance. At what point do the disadvantages of increased population outweigh the advantages to the economy of increases in immigration? This is a very large question and is beyond the scope of this report. Instead, this report examines one component of this question. Is there a point where further increases in immigration lead to substantially lower marginal increases in the growth of GDP per capita? (p. 19,  emphasis added)

The authors clearly state they are not making a complete ‘economic’ assessment of all the costs and benefits involved, which is what would be needed in order to reach a judgement about the optimum level. The authors are looking only at very specific parameters, notably GDP per capita, which is notoriously imprecise in capturing all of the costs, impacts, values and benefits of particular societal outcomes.

In particular GDP does not measure environmental and ecological outcomes, which must surely be factored in to any serious consideration of what is an optimum level of immigration. Dr Allen happens to overlook the fact that, in the very same year as the McDonald and Temple (2010) report was completed, the Department of Immigration and Citizenship commissioned a report by a number of scientists on the Long-Term Physical Implications of Net Overseas Migration, the lead author of which was Dr Jonathon Sobels. The Department did its best to bury the study by releasing it on Christmas eve. Leith van Onselen from Macrobusiness (whose post has additional useful commentary on Peter McDonald, Liz Allen and the question of the ‘optimum’), provides a summary of some of the report’s findings:

This report concluded that “higher levels of NOM impose greater adverse impacts on the quality of our natural and built environments” and that the “geographical concentration… within Sydney, Melbourne and Perth… substantially increases their environmental impact”. The report also found that “decreased urban water supply is a significant environmental constraint exacerbated by higher levels of NOM”. In particular, “modelling shows the vulnerability of Sydney, Melbourne, Brisbane and Perth to deficits in water supply.”

More recently, the Australian State of Environment Report (SoE) (2016) says that there are some areas where the condition of the environment is poor and/or deteriorating, including “the more populated coastal areas and some of the growth areas within urban environments, where human pressure is greatest (particularly in south-eastern Australia)…” The SoE also notes that “the key drivers of environmental change are population growth and economic activity.”

Dr Allen further claims that the 2010 paper by McDonald and Temple is the “latest available research.” Not only does she ignore the already-cited major environmental studies, she also does not refer to the more comprehensive and more recent study by the Productivity Commission in its 2016 report on Migrant Intake into Australia, which uses the same formal model as McDonald and Temple (2010), to do the same calculations (with slightly different assumptions).

The Productivity Commission study considers impacts on wages and productivity (both lowered by higher population growth), acknowledging that the benefits of growth would go to the few: big businesses and migrants themselves, while ordinary working people would be worse off. The Commission also mentions that its formal economic modelling does not take into consideration effects on the environment and lifestyles, and recognizes that these factors must be considered in any overall assessment.

On the question of an ageing population, which Dr Allen claims is the reason that we need continued high immigration, the Commission concludes that immigration is an ineffective way to address ageing in the longer term, and that efforts to address ageing would be better focused on managing it than mitigating it.

The Productivity Commission also usefully models a scenario of NOM at 100,000 per year, which is an intermediate step between the zero and 200,000 levels that are modelled in the ABS projections used by Dr Allen. This lower growth scenario, provided in Box 10.1 of the Commission’s report, shows a population of 35 million by 2060 with a NOM of 100,000, compared to 44 million by 2060 with a NOM of 250,000. This is a significant difference and the lower scenario would likely be less stressful on the environment and infrastructure.

Interestingly McDonald and Temple (2010) also model a scenario based on an ultimate NOM of 100,000 per annum. Comparing this with their ‘preferred’ scenario (NOM = 180,000), the magnitude of the differences, in particular a difference of just 0.05 per cent a year to the projected fall in the rate of per capita economic growth, is very small. As remarked by Bob Birrell et al in a 2011 report by the Centre for Population and Urban Research, the difference between the scenarios is minute compared to the overall projected long run decline in economic growth assumed by Treasury forecasts at the time.

Given all the uncertainties, omissions and restrictive assumptions in the McDonald and Temple (2010) modelling exercise, it is clear that it does not provide a sound basis for concluding that the optimal level of immigration is between 160,000 and 210,000, rather than, say, 100,000.

Moreover, the final paragraph of McDonald and Temple’s (2010) report concludes with a warning that can only be seen as prescient but seemingly ignored by Dr Allen:

While this report argues that immigrants will be important to the construction of productive infrastructure in Australia, if increased immigration proceeds without investment in new infrastructure, especially urban infrastructure, the result could be reductions in productivity through increased congestion and inefficiency. Thus, a plan relating to Australia’s future levels of immigration must be coordinated with policy for urban infrastructure especially housing, transport, water and appropriate energy supply. With constant fertility and net migration at 180,000 per annum, Australia’s population would rise to 35.9 million by 2050. This is a large increase and most of the additional population would be settled in the existing cities all of which are already under strain from infrastructure shortages. (p. 45)

In summary, Dr Allen has repeatedly gotten her facts wrong, comparing apples with grapefruit (permanent migrant intake quota versus NOM). Not only has she drawn facile conclusions based on inappropriate comparators, she has ignored the caveats of McDonald and Temple (2010) and incorrectly concluded that their study offers a valid basis for defining an optimum level of immigration. Further, Dr Allen has ignored the more comprehensive and more recent work of the Productivity Commission, as well as major environmental studies, which wisely point to a wide range of values, costs, impacts and benefits (not all of them commensurate) that need to be weighed in deciding what is an optimal level of immigration and population. It is incumbent upon Dr Allen to demonstrate why, once all the relevant factors and values are considered, a 100,000 NOM would be any less beneficial for the future of Australia than a 200,000 or 250,000 NOM. Indeed, the lower 100,000 NOM scenario could well be far closer to an optimum. So far, Dr Allen’s argument misconstrues or overlooks much of the best available evidence and is not rigorous or convincing.

(Thank you to Dr Jane O’Sullivan and Jenny Goldie for comments on an earlier draft of this post).