Posted on

Whither baseload demand?

John Quiggin has a post in which he argues that, if baseload demand exists in any meaningful sense, it is much lower than current offpeak demand.  I want to paraphrase and expand on what he said.

There is no such thing as a “natural” or baseload level of demand.  There is a demand curve that plots quantity demanded as a function of price (or if you’re trained as an economist, the other way around).  There is a 3rd dimension of “time of day” (or more strictly, time of week, if I can say that): the curve of quantity-versus-price shifts in and out over the day.  The entire thing then shifts out slowly over time as population and the economy increase.

At most, we might say that there is a region of the demand curve for the offpeak period that is highly inelastic with respect to price.  Quiggin is arguing that that region would only be for quite small amounts of power, distinctly less than we currently see in offpeak load figures.

The reason lies in the economics of our current electricity supply through coal-fired power stations. (Side note:  I’m not 100% certain of these points – if anyone can confirm or deny them, I’d be glad to hear from you):

  • There is some range in the thermal output of a single furnace (it’s not simply all or nothing), but real variation comes from switching entire furnaces off and on.
  • The cost of moving within the output range of a given furnace is essentially just the fuel cost; the concurrent manpower required and the maintenance needs accrued are unchanged.
  • There are economies of scale in concurrent manpower when increasing the number of furnaces.  Moving from one furnace to two does increase the staff requirement, but it does not double it.
  • There are significant one-time costs associated with starting (and possibly also with shutting down) a furnace, largely due to accruing future maintenance costs.  This means that once you start a furnace, you want to keep it running as long as possible so as to amortise that cost over the greatest amount of output.

The upshot of these points (and all of them point in the same direction) is that a cost-minimising coal-fired power station is one with many furnaces that are shut down as rarely as possible.  In other words, they ideally want to supply a large and constant amount of power to the grid.

But the demand curve at 3pm is a lot further out than at 3am.  The coal powered stations can handle this a little bit by scheduling all non-emergency maintenance overnight, but ultimately, they face a conundrum:  the demand simply doesn’t exist — at any price — to meet their cost-minimising supply in the dead of night.  So they compromise by shutting down some furnaces (which raises the average cost of the remaining power generated) and lowering the offpeak price by half (which lowers the average revenue they receive for that power) in order to raise the quantity demanded.

Quiggin is contesting that the increase in quantity demanded during offpeak is significant compared to the “true baseload” demand, the quantity that would be demanded at 3am at just about any price.

In contrast, solar power, in particular, would have supply shifting in and out over the day along with demand.

Posted on

Richard Freeman, WorkChoices and the dead hand of government

Richard Freeman is continuing his assault on WorkChoices:

[T]he new Australian labour code is such a massive break with Western labour traditions that it merits [global] attention. It was enacted in the midst of prosperity, without union or management excesses that endangered the economy, or public support. From the perspective of social science, we cannot get much closer to the ideal random assignment experiment at the national level than WorkChoices – an extreme change in law with no economic rationale or cause.

… Downloading the Workchoices legislation, I found a 687-page law with 565 pages of accompanying memorandum, all amending [i.e. not replacing] the government’s previous 861-page labour act …

… Parts of the law made so little economic sense that it seemed as if the Howard government had found a new band of whigged judges and labour lawyers to write it, on behalf of management. Which, in fact, I learned, was more or less how the law was developed. Writing the law was outsourced to the major Australian law firms that represented management …

… If re-elected this fall, the government will stay the course with Workchoices and we will see the results of this extraordinary effort to destroy collective action by workers. For the sake of social science, it would be great to see the experiment carried through to completion. For the sake of Australia, it would be great to see the election end the experiment.

He has managed to attract the attention of Justin Wolfers, guest-blogging on Marginal Revolution:

This is what happens when conservative governments confuse decentralization and deregulation.

Professor Freeman visited Australia back in September, speaking at the University of Sydney (I can’t seem to find a transcript online; only the event details and the press release) and on the ABC. He is not without his critics on the topic, but I think his points are valid. Even if you hate the unions, you’ve got to oppose Workchoices for the sheer weight of it. Where are the small-government Liberals in Australia?

Cam Riley wrote on this a while back:

When I read through the Workchoices legislation a while ago it was a brain dulling experience. The bill was long, boring and complex. It recently received a one hundred and eleven page amendment to add to the Workplace Relations Amendment Act, the Workplace Relations Amendment Bill, the Explanatory Memorandum, the Supplementary Explanatory Memorandum and the Second Reading Speech. Human Resources just got job security in the same way accountants do with the complex tax system.

Have a look at the graphs on Cam’s page. Make sure you take note of the scale on the vertical axes.

Meanwhile, John Quiggin has a suggestion for the Labor party in their campaign:

If I were running Labor’s campaign, I’d take the government’s total ad spending this term (around $750 million, IIRC) and convert that into around $5 million per electorate. Then find, for each electorate, $5 million of spending effectively foregone (two extra teachers at X High School, a local road project etc). Finally, promise to create a fund for worthwhile local projects like these, to be funded by a cessation of large-scale government propaganda.