Enter another crazed ANU immigration booster

By Leith van Onselen

Regular MB readers would be well aware of the mass immigration propaganda flowing from the Australian National University (ANU) via Dr Liz Allen and Professor Peter McDonald.

Now another ANU immigration spruiker, Professor Glenn Withers, has entered the fray claiming that Australia risks “big economic losses” if it dares to lower immigration from current turbo-charged levels. From The Australian:

…reducing the migrant intake by 30,000 a year would prompt “big economic losses” for the nation and Australians individually.

“One measure of such a policy is we’d lose a Wollongong or a Gold Coast — 300,000 people over a decade, or 600,000 over 20 years. That’s a big economic loss compared to what we would have otherwise had,” Professor Withers told The Australian yesterday.

“Remember, migrants are more skilled and younger on average; broadly, GDP would be about $50 billion a year lower by 2040,” he added…

“It would absolutely affect incomes per person — it isn’t just about the size of the whole economy,” Professor Withers added, estimating GDP per capita would be 1.2 per cent lower by 2040 if immigration were cut by 30,000 a year…

“Studies suggest for every extra two million people your city average incomes go up by 8 per cent,” Professor Withers said, referring to the benefits in rates of innovation that come with denser populations.

This is what proper economists call selective analysis. Below are my responses to Withers’ biased claims.

First, Australia would not “lose a Wollongong or a Gold Coast” over a decade because migrants are not going to the regions or creating new cities. Rather, they are piling into Sydney and Melbourne, stretching infrastructure further and making congestion and housing affordability worse:

The fact of the matter is that both current immigration levels and settlement patterns into Sydney and Melbourne are unsustainable, as noted by former Treasury Secretary Ken Henry:

The Australian population is growing by something like 400,000 a year. Think of it: a new Canberra every year between now and the end of the century. Or, put it this way, every five years building a brand new city from scratch in Australia for 2 million people.

Or put it this way: building a whole new city the size of Melbourne every decade between now and the end of the century…

My observation in Sydney and Melbourne today, is that people already think, with very good reason, that the ratio of population to infrastructure is too high. But we have set ourselves on a journey that implies an increase in that ratio. An increase in that ratio that is associated with more congestion, longer commute times to work, increasing problems with respect to housing affordability…

As well as George Megalogenis:

“If most of the population growth that’s already in train for the next 10, 20, 30 years ends up in Sydney and Melbourne, we’ve got a problem…

“You look at Sydney’s topography and it can’t fit another million people easily. And you look at Melbourne’s, and it will fit in another million but at the expense of livability because they just keep pushing the boundary out…

“The default setting to me could potentially be catastrophic for the country over the next 20 years if people just end up in Melbourne and Sydney”.

Too much immigration is the problem, not tiny cuts.

Second, Withers’ claim that migrants are “more skilled than average” is false and completely shot to pieces by the data.

According to the Productivity Commission (PC), only 30% of Australia’s permanent migrant intake is actually “skilled”:

…within the skill stream, about half of the visas granted were for ‘secondary applicants’ — partners (who may or may not be skilled) and dependent children… Therefore, while the skill stream has increased relative to the family stream, family immigrants from the skill and family stream still make up about 70 per cent of the Migration Programme (figure 2.8)…

Primary applicants tend to have a better fiscal outcome than secondary applicants — the current system does not consider the age or skills of secondary applicants as part of the criteria for granting permanent skill visas…

The PC also showed that while primary skilled migrants have marginally better labour market outcomes than the Australian born population in terms of median incomes, labour force participation, and unemployment rates, secondary skilled visas, and indeed all other forms of migrants, have worse labour market outcomes:

A recent major survey from the Bankwest Curtin Economics Centre found that 53% of skilled migrants in Western Australia said they are working in lower skilled jobs than before they arrived, with underemployment also rife:

Whereas Dr Bob Birrell from the Australian Population Research Institute has revealed that most recently arrived skilled migrants (i.e. those that arrived between 2011 and 2016) cannot find professional jobs, with only 24% of skilled migrants from Non-English-Speaking-Countries (who comprise 84% of the total skilled migrant intake) employed as professionals as of 2016, compared with 50% of skilled migrants from Main English-Speaking-Countries and 58% of the same aged Australian-born graduates:

Finally, the Australian Bureau of Statistics (ABS) most recent Characteristics of Recent Migrants report, released last June, revealed that migrants have generally worse labour market outcomes than the Australian born population, with recent migrants and temporary residents having an unemployment rate of 7.4% versus 5.4% for the Australian born population, and lower labour force participation (69.8%) than the Australian born population (70.2%):

Third, the claim that migrants are “younger on average” is irrelevant, with migrants having a tiny and temporary impact on Australia’s population age distribution.

On this point, Withers should consult his own colleague, Peter McDonald, who in a 1999 parliamentary research paper explicitly noted that it is “demographic nonsense to believe that immigration can help to keep our population young”, while also recommending “a population of 24-25 million within 50 years” as well as “annual net migration… in the order of 80 000“. McDonald also stated “that there were difficulties in the late 1980s when net migration rose for just two years to over 150 000 per annum” and that “a sustained net migration level of 120 000 per annum is at the high end of what Australia seems to be able to manage”.

Fourth, the claim that cutting immigration by 30,000 “would absolutely affect incomes per person” as well as “GDP per capita” is spurious and does not tell the full story.

The PC has conducted two detailed modelling exercises on the economic impacts of immigration, both of which were less than flattering.

The PC’s Migrant Intake Australia report, released in September 2016, compared the impact on real GDP per capita from:

• Historical rates of immigration, whereby population hits 40 million by 2060; and
• Zero net overseas migration (NOM), whereby the population stabilises at 27 million by 2060.

The PC’s modelling did find that GDP per capita would be 7% ($7,000) higher by 2060 under current mass immigration settings. However, all the gains are transitory and come from a temporary lift in the employment-to-population ratio, which will eventually reverse once the migrants age (i.e. after the forecast period):

The continuation of an immigration system oriented towards younger working-age people can boost the proportion of the population in the workforce and, thereby, provide a ‘demographic dividend’ to the Australian economy. However, this demographic dividend comes with a larger population and over time permanent immigrants will themselves age and add to the proportion of the population aged over 65 years.

The PC also explicitly acknowledges that per capita GDP is a “weak” measure of economic welfare:

While the economywide modelling suggests that the Australian economy will benefit from immigration in terms of higher output per person, GDP per person is a weak measure of the overall wellbeing of the Australian community and does not capture how gains would be distributed among the community. Whether a particular rate of immigration will deliver an overall benefit to the existing Australian community will crucially depend on the distribution of the gains and the interrelated social and environmental impacts.

It is worth pointing out that the PC’s modelling unrealistically assumed that Australia’s infrastructure stock would keep pace with the extra population, which is vital if economy-wide productivity is not to diminish:

Specifically, the expansion in labour supply through migration is projected to lead roughly to the same proportional growth in capital and output in most industries including infrastructure industries. That is, the modelling broadly assumes that there are constant returns to scale in production…

As the modelling broadly assumes that there are constant returns to scale in production, the economy-wide modelling results are broadly linear. Hence, while the modelling provides insight into the economic impact of NOM, in practice limits on Australia’s absorptive capacity (including environmental factors) mean that constant returns to scale are unlikely to hold for very high rates of immigration.

Clearly, this assumption is at odds with the Australian economy’s ‘lived experience’, whereby massive infrastructure deficits have accumulated over the last 15-years of hyper immigration, particularly in the major cities.

Most importantly for incumbent Australian workers, the PC’s modelling finds that labour productivity and real wages are projected to decrease under current mass immigration settings versus zero net overseas migration (NOM):

Compared to the business-as-usual case, labour productivity is projected to be higher under the hypothetical zero NOM case — by around 2 per cent by 2060 (figure 10.5, panel b). The higher labour productivity is reflected in higher real wage receipts by the workforce in the zero NOM case…

With zero NOM, real wages are projected to increase over time, and at a rate greater than in the business-as-usual scenario. That is, in the zero NOM scenario labour is relatively scarce which puts upwards pressure on real wages and causes a substitution towards capital, contributing to the marginally higher labour productivity relative to the business-as-usual scenario (figure 10.5, panel b). Higher rates of labour force participation through immigration in the business-as-usual case is projected to moderate such wage pressures.

Therefore, according to the PC’s 2016 modelling, high immigration improves per capita GDP by 2060 by temporarily boosting the proportion of workers in the economy, but this comes at the expense of lower labour productivity and lower real wages.

Moreover, beyond the forecast period (2060), the migrants will age and retire, thus dragging down future growth – classic ‘ponzi demography’.

As noted by the PC above, its 2016 modelling also did not take account of the distribution of gains to per capita GDP, which is vitally important. Thankfully, it’s 2006 major study on the Economic Impacts of Migration and Population Growth did, and the results were poor.

Here, the PC modelled the impact of a 50% increase in the level of skilled migration over the 20 years to 2024-25 and found that “the incomes of existing resident workers grow more slowly than would otherwise be the case”. Below is the money quote:

The increase in labour supply causes the labour / capita ratio to rise and the terms of trade to fall. This generates a negative deviation in the average real wage. By 2025 the deviation in the real wage is –1.7 per cent…

Broadly, incumbent workers lose from the policy, while incumbent capital owners gain. At a 5 per cent discount rate, the net present value of per capita incumbent wage income losses over the period 2005 – 2025 is $1,775. The net present value of per capita incumbent capital income gains is $1,953 per capita…

Owners of capital in the sectors experiencing the largest output gains will, in general, experience the largest gains in capital income. Also, the distribution of capital income is quite concentrated: the capital owned by the wealthiest 10 per cent of the Australian population represents approximately 45 per cent of all household net wealth…

The PC’s Migrant Intake Australia report also went to great lengths to stress that there are many costs associated with running a high immigration program that are not captured in the modelling but are borne by incumbent residents and unambiguously lowers their welfare:

High rates of immigration put upward pressure on land and housing prices in Australia’s largest cities. Upward pressures are exacerbated by the persistent failure of successive state, territory and local governments to implement sound urban planning and zoning policies…

Urban population growth puts pressure on many environment-related resources and services, such as clean water, air and waste disposal. Managing these pressures requires additional investment, which increases the unit cost of relevant services, such as water supply and waste management. These higher costs are shared by all utility users…

Immigration, as a major source of population growth in Australia, contributes to congestion in the major cities, raising the importance of sound planning and infrastructure investment …governments have not demonstrated a high degree of competence in infrastructure planning and investment. Funding will inevitably be borne by the Australian community either through user-pays fees or general taxation.

…there will be additional costs for the community where environmental services that are currently ‘free’ have to be replaced with technological solutions…

Accordingly, the PC explicitly requested that these costs be considered as part of any cost-benefit analysis on the immigration intake, rather than blindly following the partial results of its modelling.

A prime example of these costs is infrastructure. In its Migrant Intake Australia report, the PC pulls no punches about the higher cost of living imposed on incumbent residents from mass immigration, particularly in the big cities:

…where assets are close to capacity, congestion imposes costs on all users. A larger population inevitably requires more investment in infrastructure, and who pays for this will depend on how this investment is funded (by users or by taxpayers). Physical constraints in major cities make the costs of expanding infrastructure more expensive, so even if a user-pays model is adopted, a higher population is very likely to impose a higher cost of living for people already residing in these major cities.

Similarly, in its latest Shifting the Dial: 5 year productivity review, the PC explicitly noted that infrastructure costs will inevitably balloon due to our cities’ rapidly growing populations:

Growing populations will place pressure on already strained transport systems… Yet available choices for new investments are constrained by the increasingly limited availability of unutilised land. Costs of new transport structures have risen accordingly, with new developments (for example WestConnex) requiring land reclamation, costly compensation arrangements, or otherwise more expensive alternatives (such as tunnels).

In short, there is little hope of achieving the level of investment required to sustain current levels of mass population growth.

Finally, Whithers’ claim that “studies suggest for every extra two million people your city average incomes go up by 8 per cent” is ridiculous.

If this was the case, then mega cities like Mexico City, Buenos Aires, Moscow, Sao Paulo, Rio de Janeiro, Manila, Jakarta, Istanbul, Bangkok, Lagos, and Tehran would all be economic powerhouses with wealthy populations.

If immigration falls then there will corresponding rises in areas of the economy that are suppressed by it. We’ll see greater efficiency, a lower currency, cheaper houses, higher wages, better per capita income and access to services over time.

The facts say that immigration is an interest group not a panacea for growth.

[email protected]