Tyler Cowen:
From Adam Ozimek, here are some very good points, which I had not previously pondered:
…what a carbon tax does is push the required cost threshold up. This would allow solar to become the more profitable source of energy in the US sooner and increase the speed of its dominance here.
However, a carbon tax would raise the threshold in the US relative to the threshold for developing countries. In other words, the race for solar companies in the U.S. becomes to be cheaper than dirty energy + a carbon tax, which is a higher threshold than being cheaper than dirty energy alone, which is the threshold in many developing countries.
It is easy to see how this could cause downward march in solar costs to slow, and as a result solar would reach the threshold for China, India, and other developing countries perhaps much much later.
If this is true, it would suggest that for clean energy to become globally dominant faster it’s better for the U.S. to just subsidize solar innovation and let the untaxed U.S. market price of dirty energy stand as a strong incentive for solar to drive costs lower.
To see this, consider a world where solar was already dominant in the U.S. with current technology and costs, perhaps via a total ban of dirty energy. The supply curve of the installed base of solar technology would be much more price inelastic than the supply curve of today’s installed base of dirty energy due to higher fixed costs and lower marginal costs. This means a steeper residual demand curve for marginal innovators that provides less market share rewards for marginal declines in price, and therefore lower rewards for marginal cost cutting.
In this way, a carbon tax could make global warming worse.
First, a subsidy for solar would act just like a tax on oil and coal in terms of timing. To see this consider a standard model of the "switch point" (below). The brown lines (dirty = brown, get it?) are the Hotelling price price and the price path with a carbon tax. The green lines are the prices of the backstop (clean = green, get it?) fuel with and without a subsidy. Point A is the switch point with no government policy. A carbon tax and a clean fuel subsidy leads to about the same switch point (points B and C).
Second, once the clean fuel grabs market share it still has the incentive to reduce costs. Just like any group of business firms that operate in a competitive market, each firm has an incentive to cut costs to grab market share from its competitors. It is not true that once the solar industry grabs the market share it is done innovating. I think this makes the developing country argument irrelevant.
Second Third, I have no idea what is going on in the paragraph beginning with "to see this." Why would clean fuel supply be more inelastic if there is a ban on dirty fuel? I don't think the number of demand side substitutes affects the cost structure of an industry. This result doesn't show up in the micro textbooks (including managerial economics for MBAs), is it in the industrial organization books? And why does the residual demand become more steep? With the strawman of a ban on dirty fuel this might make sense because there are fewer substitutes. But at the switch point there is no ban, the dirty fuel substitute is always there.
So, I don't see how a carbon tax could make climate change worse. What am I missing?
Also, Ozimek goes on to say:
There are some alternatives to this view though. First, is that taxing dirty energy and using all the money for subsidies of clean tech innovation is really more efficient than subsidies without taxes. And after all, the money for subsidies will need to come from somewhere. In addition, one can imagine that economies of scale and learning by doing are extremely important in this industry. This would mean solar companies taking over the U.S. market will accelerate the decline of costs and speed up the point where it becomes cost competitive for the rest of the world.
This is point D in the diagram. I don't know if point D is more "efficient" than B or C. It does hurry up the switch point but efficiency really depends on the benefit-cost analysis and determining the optimal switch point.
