By James Kwak
Felix Salmon linked to an article by David Kotok on Build America Bonds (BAB), which reminded me that I’ve been meaning to write about them (now that they no longer exist). BAB were introduced in the 2009 stimulus bill. If a state or local government issues BAB, the federal government pays 35 percent of the interest on the bonds; the bondholder pays tax on all the interest, as usual for corporate bonds — but not for traditional state or local government bonds (“munis”). BAB were initially only authorized for two years, and were not extended in the recent tax cut compromise.
The Republican attack line on BAB is that they “subsidize states in more imprudent-type budget and debt scenarios” (Rick Santelli, quoted in Kotok’s article) or they are “a back-door handout for profligate state and local governments, allowing them to borrow more money while shifting some of the resulting interest costs to the federal government” (Daniel Mitchell). Well yes, BAB are a subsidy for state and local borrowing. But to criticize them for that without even mentioning the alternative is either uninformed or irresponsible.
Since the beginning of time (OK, the beginning of the income tax), interest on munis has been exempt from federal income tax; this is why munis are also known as tax-exempt bonds. In other words, this is a federal subsidy for state and local borrowing. There are various policy arguments for and against such a subsidy, but the basic fact is that we have a federal system in which power and responsibility are shared between the national, state, and local governments, and this is one way (not the only one) that the national government distributes money to state and local governments. The simplest alternative would be for the national government to simply hold onto its money and decide how to spend it, instead of funneling it to state and local governments to let them decide how to spend it. So the basic principle of the national government distributing money back to the states is one that Republicans should be favorable to.
But the problem is that, like most subsidies effected through the tax code, this one is inefficient. It works like this. Say that, absent the subsidy, a state would have to issue bonds with a yield of 5 percent. (That is, corporate bonds with otherwise equivalent terms issued by a company with an equivalent credit rating would yield 5 percent.) If the bond is tax exempt, however, a buyer in the 35 percent tax bracket would be willing to accept a yield of only 3.25 percent rather than 5 percent (because $5 before tax is equivalent to $3.25 after tax.) In that case, the federal government is giving up $1.75 (per year, assuming a $100 bond), and it’s all going to the state issuing the bond, which has lower interest costs; the buyer is indifferent between the two scenarios. That’s a subsidy.
In practice, however, it doesn’t work like that. The actual yield on tax-exempt bonds is higher than necessary for top-bracket bond buyers to break even. Historically, it’s been about 75 percent of the taxable yields (according to my tax casebook — Graetz and Schenk, 6th ed., p. 224). In the example above, that would be a tax-free yield of 3.75 percent. That means that the $1.75 subsidy is now being shared between the state and the bond buyer. The state gets $1.25 in lower interest costs, and the buyer gets $0.50 in interest she could not have gotten without the subsidy.
And who buys tax-exempt bonds? Rich people. So by funneling the subsidy through this tax exemption, part of it gets siphoned off by the rich. (The $1.25 in lower interest costs do benefit all state residents, who would otherwise have to pay higher state taxes; but they would still prefer $1.75 in lower interest costs.)
BAB were designed to solve this problem. Instead of making the bonds tax-exempt, they are taxable, so in the example above the state would issue them with a 5 percent yield and the buyer would get the same $3.25 after taxes as if she bought a corporate bond. This time, the federal government simply gives the state $1.75 in cash, and none of the subsidy gets siphoned off by rich people.
So it’s disingenous to analyze BAB without acknowledging the alternative; you have to compare them to traditional tax-exempt munis. And on a first-order analysis, the hit to the national fisc is the same ($1.75). The only difference is who gets that $1.75. With BAB, it’s the state as a whole; with tax-exempt munis, rich people get a cut.
Now, it is true that there are second-order effects. In particular, because the subsidy for BAB was set at 35 percent, that did low the effective cost of borrowing for states from $3.75 to $3.25 (in the above example). That means that states will borrow more than they would have otherwise. So while the federal subsidy is $1.75 on a $100 bond in either case, there will be more of those bonds in a world with BAB — so more borrowing, and more national debt. (This problem, if it is one, could have been solved by setting the percentage lower — say, at 25 percent — so that states’ effective borrowing costs would have remained the same.)
Also, because BAB exist, this reduced the supply of munis, increasing demand for munis (since there are still the same number of rich people who want tax-exempt bonds), which increases prices and reduces borrowing costs. So at the margin, BAB did increase state and local borrowing. That was actually the point — this was a stimulus bill, after all — but if you don’t like one government subsidizing another government’s borrowing, it was a bad thing.
But when people like Daniel Mitchell rail against BAB because they subsidize state and local borrowing, without mentioning tax-exempt bonds, what are we supposed to think? Do they not realize that this subsidy has always been part of the bedrock of the tax code? Or do they know it’s there, and are they attacking BAB because they want to preserve a tax exemption that disproportionately favors the rich? I’m not sure which is better.
(By the way, I think this example supports my interpretation of the tax code. BAB probably got slipped into the stimulus bill because tax policy wonks have for decades thought they would be more efficient than the existing subsidy mechanism. But the constituency for efficient tax policy is not as powerful as the constituency for investments that favor the rich.)