After my post a few weeks ago proposing a form of emissions-linked government bond which I called a ‘carbon bond’, I’ve realised an unintended benefit that the carbon bonds would bring to any government that issues them: they would reduce long-term borrowing costs. This actually surprised me when I realised it, as it certainly wasn’t one of the intended consequences of issuing the bonds, but after considering the matter, it’s actually quite clear that borrowing costs would be lower for governments offering carbon bonds, subject to a couple of conditions.
I’m not going to go over a formal mathematical proof here (I like to think my blog hasn’t become quite that nerdy yet), but here’s the reasoning behind my claim:
Assume creditors are buying ten-year government bonds at an interest rate n. Now, if they also have the option of buying a ten-year carbon bond from the government, that carbon bond is going to have a base interest rate m, which would be increased by r if the government goes over its promised allocation of emissions permits by some given amount. We’ll assume that the market of creditors will place a probability p on the government actually having to pay out the higher m + r interest rate. The only reason that p would be zero would be if the markets had absolute and complete confidence that a government would never renege on its promises. We know that’s not true, so we can say that p is nonzero.
The expected interest payment for the carbon bond will be the base rate m plus the increment r times the probability p that the increment will have to actually be paid. That is, expected interest will be m + (p * r). Given the choice between either the regular government bond or the carbon bond, if one has an expected interest rate which is higher than the other, then creditors will simply all buy the one with the higher rate. This means that, in an efficient market, the expected rate for both should be equal, ie n = m + (p * r). Now, we know that both p and r are nonzero (if r was zero, then it would just be a regular bond), and are both positive. This means that m must be less than n for the market for both bonds to operate. That is, the base interest rate on the ten-year carbon bond has to be lower than the interest rate on standard government bonds. Which means that, so long as the government doesn’t actually deviate from its proposed emissions limits, issuing carbon bonds is going to be a definitively cheaper way to raise cash.
Of course, this is a simplified explanation; there would actually be a range of possible rs depending on how far over the emissions schedule the government went, and then a range of possible ps for the expected probabilities of each of those rs being paid out. However, the principle remains the same; markets are going to be willing to accept a lower base interest rate on carbon bonds so long as there is some non-zero probability that a higher interest rate will have to be paid out instead.
Interestingly, what a government would actually be doing with this scheme is betting on its own trustworthiness. The government is assuming with certainty that it will stick to its promises, whereas the markets don’t have the same faith. By issuing carbon bonds, a government can actually leverage this lack of trust in the form of lower interest rates. Of course, to benefit from these lower base rates, the government would actually have to stick to its promises, and can it really be that confident of itself?
Edit (30/03/2010): This post was originally written using the term ‘green bonds’. I’ve since changed all references to ‘carbon bonds’ instead. This is to bring the terminology closer to that used by Michael Mainelli and Jan-Peter Onstwedder in a proposal made last year, where they used the term ‘index-linked carbon bonds’ to describe effectively the same thing. The change should also remove any confusion with other uses of the term ‘green bonds’.