Tuesday, February 10, 2009

Playing with power

(Warning: this is a fairly heavyweight post. If you're looking for wit or entertainment, or if you think "economics" is for people who find subjects like "law" and "politics" just too risque, you should probably skip it. I wouldn't blame you a bit.)

This morning, as occasionally happens, my job (don't ask) caused me to look at the Australian electricity market.

The market regulator, you'll be interested to learn, has recently redesigned its website. On its home page, now it shows graphs of electricity demand and spot price, by state, per five-minute interval over the past 24 hours.

There's something irresistible about data like this. Real, live numbers. They're completely public - you can check them out for yourself, if you share my morbid fascination. And clicking idly through the states, I noticed something that struck a chord with me from 15 years ago.

Unless you clicked on the link just now, you won't know what the graphs look like. This is one (click to see detail):

It shows demand (green line) and spot price (red line) for the state of Victoria for the 24 hours up to about 9 a.m. on 10 February. As you'd expect, demand follows a fairly smooth curve, peaking at about 6500 MW in the middle of the day (from around 10 a.m. to 5 p.m.), then dropping off to less than 5000 MW during the night. The price chart is much more spiky, but on the whole it approximates towards the demand curve, with the mid-day price being about two to three times the mid-night price.

(This price, I should explain, is the price at which electricity retailers - the companies that sell power to people like you and me - get to buy the stuff from the grid. It is not, directly, related to the price consumers pay. There are complications, of course - each retailer has its own set of contracts with various suppliers, which will set different rates and times in advance - so in practice the "spot price" only covers a tiny volume of the total trade - but there's always a fraction of their demand that can't be forecast quite accurately, and that's where this price comes in.)

With me so far? Wow, I'm quite - touched. Thank you.

Now for the exciting bit. Here's the graph for Queensland for the same period:

See the price spikes there? That's two times, during the 24 hours, when the spot price quite suddenly shot up from its normal average of around $50-60 per MWh, to around $10,000 per MWh. That's like paying $400 for a carton of milk.

Now, I have a whole argument leading on from this little observation, based on my unique perspective as a former journalist who covered the newly-privatised UK power industry in the early 90s - but I'm thinking that only my most hardcore fans would have the stamina to read it; and so far as I know, none of them have much influence over these things. So it'd pretty much be a waste of all our time. I'll summarise it in bullet-point form:
  • It's not just Australia; this goes on most everywhere. (What makes the Aussies unusual is that they publish the data in this easy-to-digest form. Try getting the equivalent graph for, say, Texas. See how far you get.)
  • There exist some ridiculously complex economic theories about what causes these spikes. (If you know what a "multivariate generalised autoregressive conditional heteroskedasticity model" is, feel free to take a look.)
  • A simpler theory is: "power companies are gaming the system". (Like Enron did, to cause the California power crisis of 2000-01.)
  • Said companies are robbing us, and the only way to stop them is to take away their market.

Now, good free-market idealogues will say: "All that proves, even if it's true - which incidentally you haven't shown - is that the market's not free enough! We need to improve transparency and remove the artificial constraints on trade!" And a few years ago I'd have listened to them.

But now I'm more inclined to think: it's impossible. We'll never make the market "free enough" for the "free-market" theory to work. Utilities - like banks - need much tighter and nastier controls than any mere regulator can provide.

Addendum: Thanks to Adam (see comments) for pointing me at this interesting analysis of the phenomenon - which, I'm amazed to see, confirms my prejudices.


Adam said...

Hi Vet,

I work in the industry and just thought I might point you to our blog were we have quite a detailed article on why these types of price spikes occur (using Summer 07-08 as an example):


There's quite a bit of commentary on that site. Which I thought you might find interesting.


vet said...

Adam - thank you! I knew there'd be people looking into it somewhere, but I didn't think any of them'd be reading me...

As I hinted - my own perspective is an old one. Back in the early 90s, when the UK industry was newly privatised, people were surprised and intrigued to see these kinds of price spikes. A colleague of mine, a serious industry wonk, spent some days researching them and trying to figure out where they came from. His best guess - and I don't think it was ever proved one way or the other - was that the generators were doing it on purpose to improve their own revenues.

I don't know if he was right, then, or how much of his analysis is applicable to the AU market today. But those spikes just took me right back...

Thank you again. I shall now go read that article, when I get time.

vet said...

I've looked at the study Adam points to. And slightly to my surprise, it actually supports my paranoid thesis. Although the author seems to think that extreme "price volatility" is, and should be, an acceptable part of the generators' manoeuvres to optimise their own income. "The cure", it says, "is worse than the disease."

But it doesn't say why. And although I'd have taken that sentiment as a matter of trust, not that long ago - I don't think I believe it any more.