It is often stated that, with perfect information and neglecting administrative burdens, the following two policies are economically equivalent:
- a carbon tax of $T$ that results in a equilibrium quantity $Q$ of carbon being emitted, and
- a cap-and-trade system that caps the quantity of emitted carbon at $Q$ and results in a allowance price of $T$, in which all of the allowances are initially auctioned by the government and then freely traded.
The practical differences between the two policies are higher-order effects such as:
- different sensitivities to uncertainty in the supply and demand curves,
- differing administrative costs, and
- differing political economy.
How does the ability to freely bank and borrow allowances across years under a cap-and-trade system change this story? Is a straightforward carbon tax equivalent to a cap-and-trade system with or without banking and borrowing? Which version is more efficient?
My guess is that a fixed carbon tax is equivalent to a cap-and-trade system in which allowances cannot be banked or borrowed, but must be repurchased every year. I also suspect that in principle, a cap-and-trade system with unlimited banking and borrowing is more efficient than a system in which allowances must be used in the year they are purchased. By analogy with the result that (in the absence of transaction costs) the equilibrium distribution of allowances across firms is independent of the initial allocation, I even suspect that (in the absence of borrowing constraints and intertemporal uncertainty), the distribution of allowances across firms and across time is independent of the auctioning schedule, but depends only on the total number of allowances that are auctioned over the entire lifetime of the cap-and-trade system. Are these guesses correct?
