Rio fails basic maths at the coalface
The decision was "significantly obstructing investment and job creation in New South Wales", the mining giant thundered.
It was "a blow to our plans for the Mount Thorley Warkworth mine and the jobs of the 1300 people who work there".
If only the judge had thought about the impact on jobs.
As it happened the judge spent a good deal of time considering the impact on jobs, especially Rio's far-fetched claim that its plan would have created an extra 44,600 jobs.
What he found will send shock waves through the ranks of economic consultants. It will never again be safe to come up with a big number for jobs created ("direct and indirect") expecting the decision-maker to give it a tick because it's the outcome of an economic model.
The Treasury has been on to the scam for a long time. In an economy with nearly full employment you can't create extra highly skilled jobs.
Here's its former secretary, Ken Henry, addressing troops in 2007:
"Consider, for example, recent commentary in the press which argues that the government should support a nuclear power sector because jobs would be created. Where will the nuclear scientists and technicians come from? Is it seriously being suggested that they will come from the dole queue or from indigenous community development employment projects?"
"The next time any of you get an opportunity to write a co-ordination comment on a cabinet submission that proposes a taxpayer-funded handout for some stunning new investment proposition - and I predict that some of you won't have to wait very long for such an opportunity - I suggest you draw attention to the submission's failure to identify the businesses that will lose labour and be forced to reduce output if the proposal is agreed to."
Rio Tinto used what is known as an input-output model to argue that if it spent more at its mine, more jobs would be created elsewhere in the Hunter. That's what an input-output model is likely to conclude. Mining's links to some industries (such as transport) are strong, its links to others (such as communications) are weak. The model assigns each link a "multiplier" and - voila! - out comes a disturbingly precise estimate of the total jobs created. Rio said expanding its mine would create an extra 44,675 jobs, full-time and lasting for one year.
It's a common (if flawed) technique. Pizza Hut used it to claim the introduction of the Big Foot pizza would create thousands of jobs. But the organisers of really big events such as the Sydney 2000 Olympics have shied away from it, perhaps fearing close scrutiny.
Rio's study by Andrew Searles of The Hunter Valley Research Foundation was a good example of a well-constructed input-output study (although the data he used was more than a decade old). The main problem is that it is or was an input-output study. As with all such studies it assumed there were unemployed resources on tap, ready to meet the firm's needs.
"I am sure Dr Searles will agree with me," the Australia's Institute's Richard Denniss told the court. "Quite explicitly the input-output data that is before you assumes the existence of what I refer to as a ghost workforce."
"The idea that there are unemployed skilled mining and manufacturing workers in the Hunter Valley at the moment sitting and waiting for these projects to go ahead, I just don't think is plausible, accurate or useful."
Indeed, the mining industry had been warning of chronic shortages and bidding up wages to grab workers from other industries.
Counting as jobs created the jobs workers went to without counting as jobs destroyed the jobs they came from was double counting. "What would happen if every industry were to commission a consultant to write the same report for them," Dr Denniss asked rhetorically.
"If every industry were to go and try and estimate the indirect jobs that flowed because of their industry's existence, what you find is that Australia would employ around 200 per cent of its current workforce," he said using calculations from Bureau of Statistics input-output tables.
Dr Searles said he thought there was enough unemployed skilled labour in the region to meet the needs of the expanded mine.
The court was told the rate for Singleton Council was 1.1 per cent.
Dr Denniss extolled the virtues of computable general equilibrium, which takes skill shortages into account. Dr Searles said he had never used it.
It wasn't Rio's finest day in court, but it would be wrong to claim the judge didn't seriously consider its claims about jobs.
Twitter: @1petermartin
Ross Gittins is on leave.
Frequently Asked Questions about this Article…
The court overturned the government approval for a large expansion of Rio Tinto’s Mount Thorley–Warkworth open‑cut coal mine in the Hunter Valley. The judge examined Rio’s employment claims in detail before reaching the decision; Rio criticised the ruling as “significantly obstructing investment and job creation in New South Wales.”
Rio Tinto’s study claimed the expansion would generate about 44,675 extra jobs (full‑time and lasting for one year). That estimate came from an input‑output economic model presented by Andrew Searles of the Hunter Valley Research Foundation.
An input‑output model uses industry 'multipliers' to estimate how spending in one sector ripples through the rest of the economy. The court found it flawed here because such models assume unemployed resources are readily available (a so‑called 'ghost workforce'), which can overstate job creation when local unemployment is low or there are skill shortages.
Critics, including Richard Denniss from The Australia Institute, argued the model counted jobs workers moved into without subtracting the jobs they left — effectively double counting. They also pointed out the mining sector has been facing labour shortages and rising wages, so adding skilled workers isn’t simply a matter of tapping an unemployed pool.
Computable general equilibrium (CGE) models were recommended because they account for labour supply constraints and price/wage adjustments across the economy. The article notes Dr. Searles hadn’t used CGE modelling in his study, while critics favoured it for analysing realistic job effects.
Investors should treat large job‑creation claims with caution and check the methodology behind them. Multipliers from input‑output studies can overstate benefits in tight labour markets. Legal and planning outcomes (like this court decision) can materially affect project timelines, costs and local economic assumptions.
Yes. The judge spent considerable time assessing Rio’s employment claims, scrutinising the input‑output study and the plausibility of the large job estimate. The coverage noted it wasn’t Rio’s finest day in court, but that the judge did seriously consider the submissions about jobs.
Yes. The article highlights that low local unemployment (for example, the Singleton Council rate cited at 1.1%) undermines the assumption that a ready pool of skilled workers exists. Low unemployment and industry skill shortages make it less likely that an expansion will generate wholly new jobs rather than displace workers from other sectors.
Call 1300 880 160
Start a chat
Schedule a call
