Yes, a carbon tax

Some very smart people, from Bill Gates on down, are skeptical of the prospects for a stiff carbon tax and have embraced the concept of subsidies for alternative energy and energy conservation. Americans' hatred of taxation, the theory goes, makes carbon taxes a dead-end proposition, politically speaking. It may be, for the time being. The politics are a discussion for another day. Distinct from the political calculus, it is worth elucidating the reasons a carbon tax is vastly superior to R&D subsidies as a means of reducing carbon emissions.

First, carbon taxes are path independent. Whereas a subsidy for a particular technology is like forcing your way through concrete with a hammer, changing the relative pricing of the inputs is like forcing your way with water under pressure; it seeks the easiest path, it can follow many paths at once.

A carbon tax creates a financial incentive to reduce carbon emissions in any way possible, and the more economically inefficient (wasteful) the emissions, the more powerful the incentive to save. While a solar panel research subsidy reduces emissions (in the best case) only by increasing the amount of solar energy, a carbon tax creates pressure to improve energy efficiency, reduce consumption, upgrade infrastructure, and develop new technology -- all simultaneously. It does not pick winners, does not depend on identifying promising technologies a priori -- tasks the market is better suited for. It allows for experimentation and heterogeneous strategies in a way no government subsidy or set of subsidies ever could, inasmuch as any subsidy, by definition, is targeted at a predetermined solution, and further restricted in how entrepreneurs can pursue that solution by regulatory requirements to minimize waste and fraud.

Second, the power of the incentives offered by a carbon tax is orders of magnitude greater than that available via subsidies. Consider a "neutral" environment, in which the only benefit from (for example) developing a new solar cell is the profit to be derived from selling it. Now add, for example, a billion dollars in subsidies. You have spent one billion dollars of the taxpayers' money, and you have added one billions dollars to the incentive to develop a new solar cell.

Now imagine a carbon tax of $100/ton. Total emissions per year in the US are about 6 billion tons. So imagine the new solar cell, if completely successful, would reduce emissions by 20%. This would save $120 billion dollars a year, every year. In a stable regulatory environment, the total incentive to develop such a solar cell would be equivalent to $120 billion dollars annuity over the life of the patent. If we estimate roughly ten years of useful life for the patent and an overall interest rate of 6%, you get a total value, and hence a total incentive, of about $880 billion. The incentives differ by orders of magnitude. Yet the incentive costs nothing from a budgetary standpoint, and, assuming it is structured to be revenue neutral (as with James Hanson's tax-and-dividend approach) the only cost to the economy is the actual cost of foregoing "cheap" fossil fuel energy in favor of alternative sources which are more "expensive." The economic impact of this is projected to be small.

One might argue that we could make the direct subsidy larger, but there are several problems with this. Unlike profits earned reducing emissions, subsidies must be awarded via some sort of application process. At the most basic level, the incentive is to win the subsidy, not to innovate or reduce carbon emissions -- we need the regulators to set the terms of the subsidies in such a way that these objectives happen. But the larger you make the subsidies, the greater the incentive to game the system -- to profit by winning the subsidy, not by reducing emissions. With a carbon tax, one pays only for success.

Finally, subsidies, unlike a carbon tax, have inherent structural problems that discourage the kind of large-scale changes in our infrastructure that they are meant to encourage. If a subsidy promotes the development of new technology, there is no guarantee that the technology will be cost effective or that it will be widely adopted. If a subsidy is for adoption of a particular technology, it may do nothing to encourage further refinement of that technology, and may become a positive barrier to replacing the "new" subsidized technology with a better solution.

Yet the biggest problem with subsidies is the "activation energy" problem. Both subsidies and carbon taxes are intended to overcome the problem of the cost advantage, in the current regulatory environment, of fossil fuel energy. If low-carbon sources of energy and increased efficiency were much cheaper, they would be widely adopted. At current prices, only a minority will adopt these strategies, either out of moral obligation or because in a minority of circumstances, the low-carbon solutions are cost-saving. The switch to low-carbon solutions can be compared to a chemical reaction with a high activation energy:


Subsidy and carbon taxes are two strategies to attempt to accelerate the forward reaction. The subsidies, by pushing particular technologies or particular plans of adoption, try to push us over the "hill" -- the substantial (let's not kid ourselves) difference between the price(*) of fossil fuels and the price of most alternatives. The carbon tax, by permanently altering the cost comparison, flattens the hill, reducing the upward slope innovators need to overcome, and increasing the rewards on the downward slope. A subsidy is like a factory furnace, needing to be continually fed and stoked in order to supply the energy to drive a reaction. The carbon tax is a catalyst.