Infrastructure cannot outrun the population ponzi


By Leith van Onselen

One of the key reasons why Australia’s high population growth (immigration) is lowering the living standards of existing residents is because of the strain that it places on infrastructure, which inevitably leads to more congestion on roads, public transport, as well as more expensive housing.

Basic math (and commonsense) suggests that if you double the nation’s population, you need to at least double the stock of infrastructure to ensure that living standards are not reduced (other things equal).

In practice, however, the solution is not that simple. In already built-up cities like Sydney and Melbourne, which also happen to be the major magnets for new migrants, the cost of retrofitting new infrastructure to accommodate greater population densities can become prohibitively expensive because of the need for land buy-backs, tunnelling, as well as disruptions to existing infrastructure.

We have seen these diseconomies of scale time and time again. For example, projects like Melbourne’s now defunct East West Road Link was expected to cost 18 billion, whereas Sydney’s North West Rail Link would cost $8 billion. That’s an astounding $350 million to $1 billion per kilometre.


Hence, running a high immigration program becomes increasingly costly for existing residents. The huge infrastructure costs also force unpopular asset sales, increased debt borrowings and austerity – none of which is a desirable outcome.

In November 2013, the Productivity Commission (PC) released its final report on An Ageing Australia: Preparing for the Future, which projected that Australia’s population would balloon to 38 million people by 2060 (mostly via immigration) and warned that total private and public investment requirements over the 50 year period are estimated to be more than 5 times the cumulative investment made over the last half century:

Australia’s population is projected to increase to more than 38 million by 2060, more than 15 million above the population in 2012. While significant variations are possible around that estimate, it is unlikely that the population would be less than 34 million or more than 42 million (figure 1)…

ScreenHunter_15677 Oct. 25 14.33

The likely population growth will place pressure on Australian cities. All of Australia’s major cities are projected to grow substantially. Sydney and Melbourne may grow by around 3 million each over the next 50 years (figure 2). In response to the significant increase in the size of Australian cities, significant investment in transport and other infrastructure is likely to be required. This is true both within the cities themselves and for the links between regional and major cities. Policies will be needed to reduce congestion problems, and to ensure adequate infrastructure funding and investment efficiency…

ScreenHunter_15678 Oct. 25 14.34

Total private and public investment requirements over this 50 year period are estimated to be more than 5 times the cumulative investment made over the last half century, which reveals the importance of an efficient investment environment…
ScreenHunter_15679 Oct. 25 14.39


The report also warned that without such massive investment, multifactor productivity – the key driver of living standards – would fall:

With MFP growth projected to be 0.7 per cent per year under the base case, the remaining share of labour productivity is driven by the accumulation of capital. Given assumptions about the capital share of income, this study estimates that the capital/labour ratio would increase by around 1.8 per cent per year over the projection period, only slightly less than the long-run growth rate from 1974-75 to 2012-13 (figure 4.7).

ScreenHunter_15680 Oct. 25 14.45

The implied level of investment to drive such capital accumulation is large — estimated at around $38 trillion dollars over the projection period in constant 2011-12 prices (table 4.4). To put that in context, in the more than fifty years from 1959-60 to 2012-13, total investment in Australia has been around $8.2 trillion. While different assumptions about capital income shares, multifactor productivity growth and depreciation affect the projections, they all produce qualitatively similar outcomes: Australia will be buying and building a large amount of physical capital. Without the efficient allocation of that capital, the achievable labour productivity growth rate would be considerably lower.

Now let’s fast forward to the PC’s Migrant Intake into Australia report, released in September. This report revised Australia’s population projection upwards to 40 million by 2060 (i.e. an increase on 2 million people), while also expressing little confidence that Australia’s governments can achieve the necessary investment required to accommodate such growth and maintain productivity:


Assuming that net overseas migration (NOM) continues at the long-term historical annual average rate of 0.6 per cent of the population, the Australian population is projected to grow to 40 million by 2060 — some 13 million larger in 2060 compared to natural increase alone. Over the past decade, however, NOM has averaged around 1 per cent of the population annually. If NOM continued to grow at that higher rate, the population projection would reach close to 50 million by 2060, or an additional 23 million people…

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. While a larger population offers opportunities for more efficient use of, and investment in, infrastructure, 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 are also impacts on the price of land and housing particularly in metropolitan areas. While this is beneficial to property owners, it increases costs and thereby reduces the living standards for those entering the property market…

In determining the migrant intake, the Australian Government should give greater consideration to the implications for planning and investment in infrastructure…

In 2010, Jane O’Sullivan, a research fellow at the School of Land Crop and Food Sciences at The University of Queensland, penned a great piece also questioning Australia’s ability to provide the infrastructure necessary to accommodate a rapidly growing population, which contained sobering conclusions:

To explain how these diseconomies [from population growth] work, I’m going to use some plausible ball-park figures. These are not precise costings – they are only illustrating a rationale by which the costing could be done. My point is that Treasury has not done such analyses, and this omission is unleashing a disaster on Australia.

Let’s start with infrastructure. In a stable population, our infrastructure needs would include replacing worn-out facilities and modernising items whose technology or design has been superseded. Different items of infrastructure have differing useful lifespans, but a cost-weighted average must surely be at least 50 years, once we remove growth as a reason for facilities being rendered obsolete. That would imply the need to replace no more than 2 per cent of all infrastructure annually.

If population is growing at 2 per cent, we need to expand the capacity of our entire stock of infrastructure by 2 per cent per annum, or else we start building up an infrastructure deficit, and service access and quality declines. That means doubling the annual requirement for creating infrastructure, compared with a stable population: 2 per cent replacement plus 2 per cent additional. Given that the useful life of much existing infrastructure is reduced by growth outstripping its capacity, the rate of replacement must increase on top of the requirement for addition.

Based on this conservative estimate, we can see that the Howard and Rudd governments have imposed a 33 per cent or greater hike in infrastructure requirement by doubling our population growth within a decade. No wonder local governments are squealing. Evidently, they have not actually increased spending by that much, and this is why we are all suddenly feeling the infrastructure stress.

How much does this actually cost? Estimates vary, but they are in the hundreds of thousands of dollars per added person. Using USA data, MIT economist Lester Thurow estimated that it requires 12.5 per cent of GDP to expand capacity at 1 per cent per year. For the developed world this was over $200,000 per person of net population growth. Australian estimates would suggest that figure is right in our ball-park too.

So, if we’re currently growing at 2 per cent per year, then 25 per cent of our GDP is currently being used to expand capacity to accommodate the people who are not yet here (or will have to be spent eventually to catch up). This means that the GDP available per capita to serve current residents is 25 per cent less than the advertised per capita GDP…

Growth is a virtually insurmountable challenge, becoming ever more costly as resources are spread thinner, pushing an ever increasing burden on future generations, while diluting their wealth base and inflating their living costs… Can we really be so stupid?


The bottom line is that running a high immigration program requires massive investment and costs a lot. Australia’s governments have failed dismally on this front, preferring to take the sugar hit from added demand while leaving the problems to be solved down the track on somebody else’s watch (i.e. never).

The obvious solution is to significantly scale back immigration and forestall the need for costly new infrastructure investment in the first place.

[email protected]

About the author
Leith van Onselen is Chief Economist at the MB Fund and MB Super. He is also a co-founder of MacroBusiness. Leith has previously worked at the Australian Treasury, Victorian Treasury and Goldman Sachs.