17 September 2011

I continue to rave like a crazy person


Last post I talked about the externalised cost of insurance and what that insurance should cost. I don't know if it came across but I put the insurance cost in the *best possible* light.

This post I'll look quickly at a more likely cost to insure and what the cost of going uninsured would really be.

Firstly I halved the number of nuclear meltdowns. I ignored the reported accidents, didn't mention the fact that accidents in Japan and communist countries are generally never reported. It's unlikely that any insurance company would be so generous and I estimated the number of required plants assuming that the load is constant when it actually fluctuates and that they produce power 100% of the time when it's actually about 75%. I also ignored the new failure mode of nuclear power plants that has only existed for the past couple of years. Software attack of the control systems. http://en.wikipedia.org/wiki/Stuxnet Previous accidents have relied on humans to make errors. For instance turning off all the safety systems (Chernobyl) or failing to take into account likely natural conditions (Fukushima) or setting fire to the control systems (Three Mile Island). Now dedicated self replicating software has been written and released specifically targeted at nuclear facilities. The likelihood an attack of this type is difficult to estimate. However, given that there have already been successful attacks carried out by this method in the 2 years that it has existed then one might conclude that the likelihood is quite high, at least as high as human error.

So just running those more realistic figures. Rather than 30 plants, probably 70 are really needed. Rather than 2 containment breaches lets assume 4. Add another 4 by software attack. 15000 reactor years, divided by 70 plants, divided by 8 breaches comes to a major containment breach predicted to average every 27 years. We're still ignoring the 25 odd reported accidents that didn't catastrophically breach containment but  did release radioactive materials. To carry the risk an insurer is going to want to get at least twice the insured amount during that 27 year period. Previously I estimated the cost of evacuating Sydney (or Melbourne by inference) at 4 T$. That's a pretty conservative estimate really. New York lost one city block 10 years ago. Despite hanging their national pride on rebuilding, not a single office space has been completed in the intervening decade. I think it's safe to assume that the rebuilding of an entire city would take at *least* 10 years. During which time the workers of the city should have their lost wages covered. The average wage in NSW is 65 000 per year and roughly half the population are wage earners. So that's 2 million times 65 000 times 10 years. 1.3 T$ for lost wages alone. Compensation for lost business for companies would have to be at least that much, another 1.3 T$. The people would need to be housed and fed during that time as well, another T$ gone there. Then you'd need to build a new city, 4 T$ there or compensate people for their lost homes and businesses, and you'd still need to build them new infrastructure in whatever places they ended up, so the cost would be similar. So that's near as makes no difference to 8 T$. To cover that risk the insurers would want around 16 T$ every 27 years. About 600 G$ a year or 60 000 dollars a year for every wage earner in Australia. Alternatively you could charge an extra 3 dollars per kWh consumed. Almost all that money would go to off shore insurers and Australia would go bankrupt quick smart as it's about 2/3rds of our GDP.

The other alternative is to externalise the risk to the public of Australia. So we know roughly what the risk is. Somewhere between my two estimates of once every 27 years and once every 250 years. Personally I think it's closer to the 27 years than the 250. So what is the hazard? Well evacuating a capital city. But what does that mean to the average person in Australia? Let's say for a moment you're not actually evacuated... All good then? Somebody Else's Problem? Ok, for a start every business in Australia is connected in some intimate way with the capital cities. If you suddenly evacuated Sydney for example, all the major local corporations would go broke all at once. The international ones would leave. The banks have huge amounts of money tied up in property and business loans. About half the property they have lent on would instantly be worth nothing and the people who owed them the money would have no jobs. So all the banks would fold and take all your savings with them. The stockmarket would fall over as most corporations went into receivership, so your superannuation would be worth nothing. No-one would be able to borrow money and millions of people would put their properties up for sale as they had no money or job, so your property value would fall to virtually nothing. It's almost certain that the government, suddenly unable to raise tax and gifted with 4 million homeless unemployed people, would start to print money like mad. Expect hyperinflation to destroy any money you have under the mattress and reduce your wage (should you still have a job) to virtually zero.

Fun! But it gets better...

Food in Australia has some of the longest distances from farm to plate and some of the most oil intensive and corporate owned agriculture in the world. With the Australian dollar in freefall the corporations all broke and no foreign exchange with which to buy oil, food production would stop as would food transport. Expect food riots.

That's if you're not one of the evacuees. Imagine how much fun they would be having.

16 September 2011

Why can't people calculate cost?


I try to be funny on this blog, but now for something completely different.  I had the most annoying "conversation" at work today. I'm wanting to talk about work but there's already a conversation going about how coal or nuclear are the only options and renewables can't ever play a part. So as I couldn't work I "joined in". I tried to say that the only way either could compete with renewables on price was by externalising the costs. So their way of discussing that was to simply tell me I was wrong and talk over the top of my reasoning.

Since I can type here and no-one can talk over the top of me, I'm going to climb on my soap box and talk about externalised costs. The externalised cost of coal is well known: possibly making the planet uninhabitable. Also taking assets owned by all and using them for the benefit of a few without compensation. Nuclear is different. The upfront cost is known. It's between 4 and 10 dollars per watt, but lets call it 4 dollars per watt. Australia draws 222 billion kWh a year (CIA fact book). That's an average of about 25 GW. So that would cost about 100 G$ to install nuclear replacement. The interest payments on that would be about 5 G$/yr. Replacement would be another 5 G$/yr to keep them up to date. So that's 10 G$/yr. Can't externalise that, but what is externalised is the risk. If one blows up then everything downwind of it is rendered uninhabitable for about 100 years or so. (Which in economic terms is forever). So say it made a city like Sydney uninhabitable. Any normal business would have to have insurance to cover any public liability costs. So what would be the cost of public liability to cover that? First you have to have some idea of the risk. There have been around 25 nuclear accidents in the 15000 reactor years that we've racked up so far in all the reactors around the world. There have been really 4 reactor meltdowns that breached the containment completely. However 3 happened together so lets call them 2 meltdowns. So say you get 2 meltdowns (really 4) per 15000 reactor years. To supply Australia you'd need about the same number of reactors as Russia, about 30. 15000/2/30 is 250. So on average you would expect a major meltdown every 250 years. So to make it worth while for an insurance company to cover the “worst case” you should expect to pay somewhat more than 1/250 of the total worst case cost. Say 1/100th of the cost. So what is 1/100th of the cost of building a replacement city and compensating all the business that have lost money from the disruption. It's hard to say, but given that you'd need to build new factories, schools, roads, hospitals, railways, ports, dams, sewers, houses, opera houses etc lets say a million dollars per person. There's 4 million people in Sydney so that's 4 T$. So the externalised cost is 1/100th of that per year. 40 G$/yr. Add the non externalised cost and you get say 50 G$/yr to make 222 billion kWh. This ignores the cost of mining and enrichment.

Compare that to renewables. Powerplant level Photovoltaics are down to 1 dollar per watt, but the sun doesn't shine all the time. Say 10 dollars a watt effectively (surely a generous multiplier!). That's 250 G$ fully installed. Using the same maths as nuclear to estimate cost of borrowings and replacement after 20 years comes to 25 G$ a year. That's half the price! (even ignoring the cost of mining and enriching uranium and disposing of the waste and decommissioning the plants which is thought to be similar in cost to building them in the first place)