We need renewable energy, such as solar power. Because of that, companies are working hard on making it cheaper. They can do this either by developing new, cheaper to manufacture technologies, cheaper ways of installing or by simply getting economies of scale as demand and production increase. They haven’t managed to follow Moore’s law, though some new-technology developers predict they someday will.

However, there is a disturbing paradox in these activities. Unlike computers, it does not make financial sense to buy solar (or any other low or zero operating cost energy technology) if you have reasonable confidence the price is going to improve at even modest rates.

Imagine you have an energy technology with effectively zero operating cost, like PV panels. Let’s say that it’s reached the point that it can match the price of grid power over a 20 year lifetime. That means that, if it costs $10,000, it costs $72 per month or $872 per year at a 6% cost of funds. (Since $872 buys 9688 kwh at the national average grid price of 9 cents, that means you need a 4800 watt PV system to match the grid which is hard to do for $10,000 but someday it won’t be.)

But here’s the problem? Let’s say that it’s very reasonable to predict that the cost of solar will drop by more than 9% over the coming year. That’s a modest decrease, entirely doable just with increased production, and much less than people hope from new technology. That means that your $10,000 system will cost you $9,100 to buy a year down the road. Since we are talking about a grid-equivalent price system, the cost of grid power in this example is $872. So you can buy the power from the grid, wait a year, and save money. The more you expect the price of solar to drop, the more it makes financial sense to delay. (Note that at this lower price the system is now beating the grid. What matters really is whether the dollar cost reduction of the solar system exceeds the dollar cost of the grid electricity purchased.)

Indeed, if you predict the cost-drops will continue for many years, it sadly does not make sense to buy for a long time. Effectively until your predictions show that the cost decrease of the system no longer exceeds the cost of the power generated by it. That has to eventually come some time, since as it gets very cheap it can’t really drop in price by more than the cost of grid power, especially while there are physical install costs to include in the mix. But it can certainly drop by 6% per year for a decade, which would take it down to half its original cost. Possibly longer.)

Now, you will note I speak of the financial cost. This ignores any motivations based on trying to be greener. This is the analysis that would be done by somebody who is simply looking for the best price on power. This is frankly how most people think. This can be altered by both government incentives to buy solar and by externality taxes on polluting grid power.

This also applies not simply to solar, but any technology where you invest a lot of money up front, and have close to zero operating cost. Thus wind, geothermal and certain other technologies face the same math. Even nuclear to some extent.

All of this also depends in your confidence in your predictions. The more uncertain your predictions of price drops, the more you might be pushed in other directions to obtain certainty.

The paradox is this. We may be in a situation where solar is competing with grid power, and many are poised to buy it. If many do buy it, economies of scale will drive the price down. Thus, nobody should buy it, as they should wait for that price decrease! But if nobody buys it, it won’t decrease in price as much, creating a chaotic system. Some will buy it (to be green, for example) so it’s not a total loss, but it becomes harder to understand.

We’re used to dealing with computers, which reduce in price not just 6% a year but 40 to 50%. We’ve all felt the dilemma over whether to buy a computer or other electronic device that will lose its value so quickly, or whether to wait. However in that case, if we wait, we don’t get our nice new computer or camera, and thus lose out. You can wait forever which makes no sense. This is not the same logic with power. With power, we’re talking about a commodity that you can buy elsewhere, and get all the benefit — for less than the depreciation on what you considered buying.

What can keep the market for solar going if it looks like it will drop in price? Well, first of all, many people want to buy solar other than to save money. (Indeed, only a few people today with high local electricity prices and fat government rebates can save money with it.) Secondly, it seems that few people, even if their goal is to save money, think this way. And if they do, they are uncertain of their predictions, and would rather get the solar now than risk grid power going up or solar going down. But curiously, it remains the case that if people make predictions of cheap solar in the near term, and they are believed, it should kill most sales of solar in the present term.