I happen to know a fair amount about the problem of swapped muni VRDB’s as I had a two clients threatened by similar circumstances. Unfortunately, the NYT article makes it sound like the municipalities were betting on interest rates which simply isn’t the case. So I feel compelled to tell the world what’s really going on here. Bear in mind that I can’t speak to the situation in Tennessee specifically, because every situation can be a little different, but this should give you the general picture.
First, let’s say its five years ago and you are one of these poor unsuspecting municipal authorities. Let’s assume you are the authority who manages the local airport, the Bumpkin Airport Authority. You’d like to issue debt, and like any responsible financial steward, you want to minimize your interest cost.
Your banker suggests that a variable rate bond would lower your expected interest cost, because demand for short-term bonds is extremely strong. In 2004, the typical rate on variable rate muni debt (either auction rate or VRDN) was around 1.5%. (There are some additional fees involved, which we’ll get to in a minute.)
First, a quick lesson on muni variable rate bonds. In a VRDN, the investor has the option to “put” the bond back to the municipality on any interest rate reset date, usually every 7 days, at par value. With an auction rate, investors can choose to “sell” at any auction, assuming the auction doesn’t fail. Remember that until 2007, auctions almost never failed, so this wasn’t seen as a big risk.
In both cases, the interest rate isn’t based on some reference index, like LIBOR, but whatever interest rate clears the market.
But you, as the municipal airport authority, aren’t interested in taking variable interest rate risk, as you don’t have any natural variable rate assets. You’d rather lock in a certain interest rate today and have a known cost for whatever you are selling the debt to construct.
Your friendly banker has a solution. Sell the debt variable rate, and at the same time enter into a pay fixed, received floating swap. On its face, this can hardly be called creative finance. Its really finance 101. You have a floating liability, you want a fixed liability, just enter into a swap. Simple.
The Sith Lord was in the details. First of all, in order to do a VRDN, you needed to get a letter of credit from a bank. See, investors needed to know that the municipality had the cash to fund that put option I described above. The bank LOC allowed for that. So let’s say the bank was charging 0.25% for the LOC. In the case of an airport authority, the bank would probably require that the municipality also buy a monoline insurance (e.g. Ambac) policy to protect the bank in the event the municipality defaults and the bank gets hit with a wave of puts. Let’s say that costs about 0.10%.
But even in the face of those extra fees, the issue floating/swap to fixed still saves you a lot of money, because the fixed side of the swap is actually below where you could sell fixed rate debt. Everything is peachy.
The only remaining hitch is that, as I said above, muni VRDNs don’t reset based on a specific index, but on whatever rate clears the market. This left the possibility that issuer A might pay a slightly higher rater than issuer B one week, but then issuer B would be higher the next week. Not because of anything about the issuers themselves, but just because of random variations in supply and demand at any point in time.
Unfortunately, the floating side of the swap had to be based on some predetermined index. Bankers usually picked one of two options. Either the SIFMA index, which is a published index of muni VRDN rates. Or they used 67% of LIBOR. E.g., if 1-week LIBOR was 3%, the the swap rate would be 2%. The 67% number was intended to reflect the typical gap between taxable and tax-exempt money market instruments. I believe the LIBOR version was more popular than the SIFMA version, and I have also heard swaps struck at 80% of LIBOR.
Right there was the red flag. What happens if the VRDN rate set by market forces isn’t equal to the 67% of LIBOR level? This is known as basis risk, and it did happen under normal times. But it was always short lived. For example VRDN rates always rose during times when retail investors were pulling money out of muni money market funds, such as tax time. But those periods of elevated rates was always short-lived. The huge savings from the synthetic fixed rate structure overwhelmed these short-term costs.
Let’s go back to the bank providing the LOC. Remember they required you to have a monoline insurance policy from Ambac to protect themselves. The actual legal language probably says something to the effect of…
“ABC Bank requires that Bumpkin Airport Authority acquire an insurance policy from a monoline insurer rated in the top ratings category from Standard & Poors and Moody’s Investor Service. Should the authority be unable to acquire such a policy or should the monoline insurer be downgraded below Baa3/BBB- ABC Bank may withdraw the letter of credit.”
Of course, don’t need to worry about Ambac being downgraded right? Er… From the investor’s perspective, you didn’t wait around for Ambac to actually be downgraded. You were allowed to put these bonds back to the issuer at par! You hit that bid as hard as you could as fast as you could.
So now what happens? Remember that the interest rate that the Bumpkin Airport Authority actually pays is set by supply and demand. Now that the LOC is threatened, there is no demand, all supply. In order to actually entice some buyers, they had to set the rate at 7%, 8%, 9%, etc. Note that these weren’t the failing auction rate bonds we heard so much about, although a similar story would apply have Bumpkin decided to go ARS.
Now Bumpkin is paying 9% on their VRDN, while the floating end of the swap is only paying you 67% of LIBOR, currently a glorious 0.25%. On top of the 9% you are paying investors, you are also paying your swap provider whatever the fixed leg of the swap is, probably something in the 4% area. Ugly.
But wait… it get worse. The interest rates are actually set by some dealer, called the remarketing agent. In normal times, the dealers would set the rate at something reasonable, and if they couldn’t sell all their bonds right away, they’d just inventory them. So if it happened to be that a big holder of the Bumpkin Airport bonds wanted to put their bonds back on a given day, it was no big deal. The investment bank was willing to just hold the bonds waiting for the right investor to come along. It was considered a good use of balance sheet because it justified the remarketing fees the bank was collecting.
Once dealer balance sheets became crunched, nicities like this went right out the window. Instead of holding the unsold inventory, the dealers were exercising their rights to push bonds they couldn’t remarket back to the LOC bank. These then because so-called bank bonds, and Bumpkin was charged some pre-determined rate on these, I think it was set off Prime.
But wait… it gets worse. Remember that the swap was intended to be a hedge against rising interest rates. It is therefore effectively a short position on long-term fixed rate bonds. In fact, long-term bonds have skyrocketed in value. Thus your swap is getting crushed. A 30-year swap struck on January 1, 2008 for $10 million notional value would currently be down $3 million in market value. Put another way, if you want out of this swap, you need to pay the investment bank $3 million.
Had the swap remained an effective hedge, this wouldn’t be a problem, because Bumpkin Airport would be saving an equivalent amount of money on plummeting short-term rates. But in fact, Bumpkin is paying a usurious 9%.
So the VRDN itself is killing you. The swap is killing you. Basically, you’re dead unless something changes.
What most municipalities did was refinance the Ambac-backed deal with a new VRDN without that stipulation. Except for a brief period in September and October 2008, the VRDN market has been pretty healthy. So once you refinance the VRDN, then the swap goes back to being a decent hedge. Everything works out just fine.
But even if you do a new VRDN deal, you still need a LOC from a bank. Guess what? Banks aren’t so keen on tieing up their capital to make 25bps on muni LOCs. Instead, they’ve been picking carefully who they deal with, and charging a lot more to do it.
Even if the municipality can restructure, it isn’t out of the woods entirely. If the swap is deeply underwater in nominal market value, the municipality probably has to post additional collateral. Think of it similar to margin posting on a futures contract. In some cases, this is no big deal, because the municipality has a decent sized general fund and simply must set aside certain securities as collateral. But in other cases, the municipality has little safety net. In fact, its more likely an issuer like Bumpkin Airport Authority has a sizeable investment portfolio compared with some county or school district which collects taxes directly. A lot of times, issuers with full taxing authority keep less in general funds. Politically, if the voters see that their county has a big investment balance they start wondering why tax rates aren’t being lowered and/or why the money isn’t being spent on new projects. An issuer with more volatile revenue, like a airport, toll road, hospital, etc., is more likely to build a reserve. It tends to be less politically sensitive if there aren’t any direct taxes involved.
If you are an investor in munis, the best thing to do is hunt down how much VRDN exposure your bond issuers have, whether they have any monoline contracts attached, and what their plan for dealing with both is. You will probably find that you have nothing to worry about, but if you are sloppy, you could wind up with the next Jefferson County.