What would it cost Australia to go nuclear? Good question. Excellent question, in fact, because no one really knows.

Estimating the cost of building a nuclear power plant is difficult even in countries with established nuclear industries. For a start, not too many have been built in recent years. Further, very few — in fact, not really any — have been built in an open, competitive, private power-supply market where the plant would face competition from alternative electricity sources. Governments indirectly or indirectly play a huge role in offsetting costs, transferring risk from company balance sheets into governments programs, tax offsets and implicit guarantees.

Worst of all, the construction of nuclear power plants is notoriously prone to cost overruns. It wasn’t Three Mile Island and Chernobyl that smashed the US power industry in the 1980s — it was the fact that from the 1960s to the end of the 1970s the final construction cost of US nuclear plants was on average twice to four times the original cost.

Regularly blowing out your costs by 400% is not a way to engender investor and government confidence.

Currently we’re seeing the same thing in Finland and France, where costs for new nuclear plants (initially €3.3 billion, or $A5.3 billion) are currently 50-70% over budget, years out from completion.

Overruns and delays are toxic to the balance sheets of nuclear plant builders and their customers because of the cost of capital, already high due to the extended construction period. That’s why so many nuclear power advocates say low-interest government loans are necessary to make building new reactors financially viable — they want taxpayers, rather than capital markets, to provide their capital.

There is also uncertainty over the construction costs of new generation “III+” reactors, which have been developed to the design stage since 2000. And construction costs have also escalated rapidly in recent years, partly because the lack of plant construction has meant bottlenecks and reduced capacity among manufacturers of the specialised equipment required. That, however, would be remedied if the nuclear industry started to develop real momentum.

There’s also the problem of accurately costing the entire lifecycle of a nuclear plant. The costs of running nuclear power plants, much more so than other types of power plants, are mainly fixed costs derived from initial construction — between 60-70% of overall costs. Operating costs (such as fuel, although the cost of uranium has spiked more than 600% in the latter part of this decade), tend to be comparatively lower than for, say, coal-fired plants. But there are also decommissioning costs, which need to be built into revenue while the plant is operational or picked up by a government. And there’s the the vexed problem, still unresolved, of the storage of nuclear waste (similar to the problem of carbon capture and storage), which will be a cost for the sorts of periods beyond even the most forward-looking economic modelling, let alone power company balance sheets.

And there’s the insurance cost of dealing with a nuclear accident, which is so large that several governments have simply capped the liability of plant owners, with the assumption taxpayers will pick up the cost of mopping up a major radiation leak.

So, with enough caveats to sink an industry, how much do nuclear power plants cost? The industry and analysts talk of “overnight” costs — that is, what it would cost to build one overnight and connect it to the grid. Couched in terms of cost per kilowatt capacity, the industry itself variously proposes costs ranging between $US1500-3500 per KW. Actual power plants proposed in the US since 2006, however, have all cost more, with cost estimates starting at $2500/kW, averaging about $4500/kW, up to more than $6000/kW.

In May last year, ratings agency Moody’s estimated a cost for US nuclear power plants “potentially exceeding” $7000/kW. More extreme cost estimates — from the Wall Street Journal, not green groups — put the figure as high as $8000/kW.

At the time the contract for the Finnish plant was signed in 2005, the cost was about $2400/kW. With overruns and delays, the cost is now $4000/kW and rising. The equivalent figures for the Flamanville plant under construction in France are $2600/kW and $3260/kW.

How does that compare with other energy sources? Last year financial advisory giant Lazard published an analysis of US energy costs. Lazard put the construction cost of nuclear power at $5750-7550/kW — twice as expensive as coal, four times as expensive as gas and more expensive than coal with CCS. It was also much more expensive than wind, geothermal and slightly more expensive than even the most expensive solar technologies. (click here or on the image to enlarge)

When it comes to operational costs, nuclear has a greater advantage given its high proportion of fixed costs. Even then, based on Lazard’s figures, it is not much cheaper than coal with carbon capture and storage, and still more expensive than renewable wind or geothermal and the same cost as some solar technologies. (click here or on the image to enlarge)

Remember, these are US figures. Our operating costs for gas and coal are probably lower since we’ve got so much of the stuff.

Wind and solar power have the advantage of much shorter construction times, and no decommissioning costs. Neither requires taxpayers to take on risk — either through lending capital to nuclear reactor builders for the decade-plus construction and 200% budget blow-out history shows they require, or through requiring high electricity prices for consumers to guarantee a return on capital, or through the acceptance of safety and storage risks by the taxpayer into infinity, or at least the next 200,000 years.

Some of nuclear power’s construction costs will come down if reactor construction significantly expands globally. By the time Australia has developed the basic regulatory infrastructure and skills base required to seriously consider a nuclear power industry, we may be able to take advantage of it. But construction costs will have to fall a long, long way before nuclear power can be remotely considered a viable economic option compared to renewables or even gas-fired power.

So next time a nuclear proponent tells us we should be “having a debate” about nuclear power, ask them a simple question: why should taxpayers fund the most expensive and slowest energy option when so many alternatives are significantly cheaper and pose less financial risk?