Is Deep Research deep? Is it research?

I’m working on a first draft of a book arguing against pro-natalism (more precisely, that we shouldn’t be concerned about below-replacement fertility). That entails digging into lots of literature with which I’m not very familiar and I’ve started using OpenAI’s Deep Research as a tool.

A typical interaction starts with me asking a question like “Did theorists of the demographic transition expect an eventual equilibrium with stable population”. Deep Research produces a fairly lengthy answer (mostly “Yes” in this case) and based on past interactions, produces references in a format suitable for my bibliographic software (Bookends for Mac, my longstanding favourite, uses .ris). To guard against hallucinations, I get DOI and ISBN codes and locate the references immediately. Then I check the abstracts (for journal articles) or reviews (for books) to confirm that the summary is reasonably accurate.

Read More »

If something can’t go on forever, it will stop

My latest in Inside Story

The most striking observation in Dean Spears and Michael Geruso’s new book, After the Spike, is summed up by the cover illustration, which shows a world population rising rapidly to its current eight billion before declining to pre-modern levels and eventually to zero. As the authors observe, this is the inevitable implication of the hypothesis that fertility levels will remain below replacement level indefinitely into the future.

Read More »

Stop the free ride: all motorists should pay their way, whatever vehicle they drive

My latest in The Conversation


\A new road charge is looming for electric vehicle drivers, amid reports Treasurer Jim Chalmers is accelerating the policy as part of a broader tax-reform push.

At a forum in Sydney this week, state and federal Treasury officials are reportedly meeting with industry figures and others to progress design of the policy, ahead of next week’s economic reform summit.

Much discussion in favour of the charge assumes drivers of electric and hybrid vehicles don’t “pay their way”, because they are not subject to the fuel excise tax.

This view is based on an economic misconception: that fuel taxes are justified by the need to pay for the construction and maintenance of roads.

This is incorrect. In a properly functioning economic system, fuel taxes should be considered a charge on motorists for the harmful pollution their vehicles generate.

That leaves the problem of paying for roads. To that end, a road-user charge should be applied to all motorists – regardless of the vehicle they drive – so no-one gets a free ride.

Real science from real scientists.

Get free newsletter

A traffic jam on highway surrounded by bush.
A road-user charge should be applied to all motorists. NSW government

What is the fuel excise?

The fuel excise in Australia is currently about 51 cents a litre and is rolled into the cost of fuel at the bowser.

Some, such as the Australian Automobile Association claim revenue from the excisepays for roads. But it actually goes into the federal government’s general revenue.

The primary economic function of the fuel tax is that of a charge on motorists for the harmful pollution their vehicles generate.

A man in a black cap and top walks through petrol station.
Fuel excise is rolled into the cost of fuel at the bowser. FLAVIO BRANCALEONE/AAP

Paying the cost of pollution

Vehicles with internal combustion engines – that is, those that run on petrol or diesel – create several types of pollution.

The first is carbon dioxide emissions, which contribute to human-caused climate change. Others include local air pollution from particulates and exhaust pollution as well as noise pollution.

In economic terms, these effects are known as “negative externalities”. They arise when one party makes another party worse off, but doesn’t pay the costs of doing so.

How big are the costs to society imposed by polluting vehicles? Estimates vary widely. But they are almost certainly as large as, or larger than, the revenue generated from fuel excise.

Let’s tease that out.

A litre of petrol weighs about 0.74 kg. But when burned, it generates 2.3 kg of CO₂. That’s because when the fuel is combusted, the carbon combines with heavier oxygen atoms.

Before the re-election of United States President Donald Trump, the nation’s Environmental Protection Agency estimated the social cost of carbon dioxideemissions at about US$190 (A$292) per metric tonne.

So in Australian terms, that means CO₂ emissions from burning petrol costs about 67 cents a litre, more than the current excise of 51 cents per litre.

Even using a more conservative estimate of US$80 a metric tonne, CO₂ emissions generate costs of around 28 cents a litre, more than half the fuel excise.

A spotlight on health impacts

Motor vehicles are a major cause of air pollution. Air pollution is causally linked to six diseases:

  • coronary heart disease
  • chronic obstructive pulmonary disease (COPD)
  • stroke
  • type 2 diabetes
  • lung cancer
  • lower respiratory infections.

Estimates of the deaths associated with air pollution in Australia range from 3,200 to more than 4,200 a year.

Even the lower end of that range is far more than the roughly 1,200 lives lost in car crashes annually.

University of Melbourne analysis in 2023 landed at an even higher figure. It suggested vehicle emissions alone may be responsible for more than 11,000 premature deaths in adults in Australia a year.

Putting a dollar value on life and health is difficult – but necessary for good policy making.

The usual approach is to examine the “statistical” reduction in deaths for a given policy measure. For example, a policy measure that eliminates a hazard faced by 1,000 people, reducing death risk by 1 percentage point, would save ten statistical lives.

The Australian government ascribes a value of $5.7 million per (statistical) life lost or saved. So, hypothetically, a saving of 2,000 lives a year would yield a benefit of more than $10 billion.

This is more than half the revenue collected in fuel excise each year.

A woman wearing a mask walks out of a hospital.
Putting a dollar value on life and health is difficult – but necessary for good policy making. DIEGO FEDELE/AAP

The best road forward

Given the harms caused by traditional vehicles, society should welcome the decline in fuel excise revenue caused by the transition to EVs – in the same way we should welcome declining revenue from cigarette taxes.

If we assume fuel excise pays for pollution costs, then who is paying for roads?

The cost of roads goes far beyond construction and maintenance. The capital and land allocated to roads represents a huge investment, on which the public, as a whole, receives zero return.

Vehicle registration fees make only a modest contribution to road costs. That’s why all motorists should pay a road-user charge.

The payment should be based on a combination of vehicle mass and distance travelled. That’s because damage to roads is overwhelmingly caused by heavy vehicles.

Then comes the question of Australia’s emissions reduction. The switch to electric vehicles in Australia is going much too slowly. A road user charge targeting only electric and hybrid vehicles would be a grave mistake, slowing the uptake further.

Mitigating the productivity damage from Covid-19: the case for improved ventilation standards

I wrote this for the Cleaner Air Collective, who used it as an input to their submission to the Productivity Roundtable

Cleaner Air Submission here

Given the purpose of the exercise, the discussion is framed in terms of productivity though many of the issues are broader

Covid-19 is a serious economic problem for Australia, not only as a major cause of death, but because of serious impacts in productivity.

Although most Covid-19 deaths occur among people over 80, there were over 200 deaths from Covid among people aged 40-64. This is a mortality rate comparable to that of road trauma (377 deaths in this age group in 2022) As of 2023, excess mortality remained high at 5 per cent

With the effective abandonment of most forms of reporting, it is hard to assess the prevalence and impact of Covid-related morbidity. However, there is substantial global evidence of increased worker absenteeism associated with both acute Covid-cases and post-Covid conditions (long Covid). Evidence also suggests cumulative damage to various organs associated with repeated infection.

Read More »

A billion people would be plenty to sustain civilisation …

… as long as they are healthy, well fed and well educated

Much of the panic about falling birth rates can be dispelled once we realise that (barring catastrophe) there will almost certainly be more people alive in 2100 than there were in 2000. But what about the distant future? Dean Spears, co-author of After the Spike has kindly provided me with projections showing that with likely declines in fertility the world population will decline by half each century after 2100, reaching one billion around 2400. Would that be too few to sustain a modern civilisation ?


We can answer this pretty easily from past experience. In the second half of 20th century, the modern economy consisted of the member countries of the Organization For Economic Co-Operation and Development (OECD). Originally including the countries of Western Europe and North America, and soon extended to include Australia and Japan, the OECD countries were responsible for the great majority of the global industrial economy, including manufacturing, modern services, and technological innovation.

Except for some purchases of raw materials from the “Global South”, produced by a relatively small part of the labour force, the OECD, taken as a whole, was self-sufficient in nearly everything required for a modern economy. So, the population of the OECD in the second half of last century provides an upper bound to the number of humans needed to sustain such an economy. That number did not reach one billion until 1980.

Read More »

No more anti-wokism

I can’t be bothered engaging with people who use “woke” as a pejorative, or similar Trumpist rhetoric. I’ve therefore banned a couple of commentators (Jack Strocchi and Svante) who persisted in using it, despite warnings. There are plenty of other places where they can bloviate if they choose. Anyone offended by my decision on this is welcome to a full refund on the way out.

Whatever happened to Romney Republicans?

Have they changed, or just become their worst selves

While Trump is unpopular with a majority of Americans, his support among Republicans remains solid. That’s despite blatant corruption, fascist policies and a failure to deliver any of the economic benefits he promised. Faced with this depressing fact, the standard New York Times response has been to send an intrepid reporter to “Trump Country” (rural Kentucky or Midwestern diners) to find out what is going on.

But it would be far more instructive to send them to Long Island, where Trump won both counties in 2024. Long Island voters have given solid support to Republicans at all levels. Even as he was crushingly defeated in New York as a whole, Mitt Romney got close to half the vote in Suffolk and Nassau counties. Trump did a few percentage points better in 2024, winning both. But he would have gone nowhere if not for the solid support of Romney voters

This doesn’t fit at all with the usual stories about Trump voters. The residents of Long Island are not the “left-behinds” routinely described in explanations of Trump’s appeal. The average income is over $100 000 and unemployment rates have long been around 3 per cent. Like most New Yorkers, Long Islanders have been beneficiaries of the globalised economy of which Romney was a symbol. And, if you were to believe Jonathan Haidt’s The Righteous Mind they did so because they valued honor, loyalty and purity, qualities Trump routinely trashes.

Democrats from Hillary Clinton on assumed that these contradictions would lead suburban Republicans to abandon Trump in numbers large enough to offset any losses of Democrats attracted by Trump’s racism and misogyny. Evidently this is not the case. Not only have the Republicans who once voted for Romney maintained their support for Trump but they have preferred him to any Republican alternative. And, with few exceptions, they have embraced Trump’s racist and fascist policies, even as he approaches outright Nazism.

What has happened here? Has Trump, as Walter Olson suggests, radicalised his followers leading them to support positions they would once have rejected? Or has he simply allowed them to reveal themselves (or at least their worst selves) as the racists and fascists they always were?

The answers to these questions are academic, in the pejorative sense of the term, as regards the US. Romney-Trump voters have made their choice, and there is no going back to old-style Republicanism. Perhaps, if enough of them realise that their choices have been both evil and disastrous for the US as a whole, the regime might collapse relatively quickly. But there is no sign of that.

The big question for those of us living outside the US is whether it could happen here. As long as the far-right remains essentially a protest party for low-education voters who are mostly disengaged and disaffected, like Pauline Hanson’s One Nation in Australia, its occasional flare-ups can be expected to fade, as appears to have happened with Geert Wilders in the Netherlands. But if the middle class and business base of the mainstream conservative parties goes the same way, democracy is in trouble.

Read More »