Skip to main content

How Much Did Goldman Screw The City Of Oakland Anyway?

  • Author:
  • Updated:

Oakland has been fighting with Goldman Sachs over an interest rate swap for a while and I've always thought it's a little embarrassing to talk about. Obviously Oakland's theory - "we entered into a bet, and we lost, so we want to pretend it never happened" - is pretty silly, but it's like, yeah, Oakland has it rough, and Goldman has it less rough, so just give them some money, no?

Oakland is trying to get out of a Goldman-brokered interest rate swap that is costing the cash-starved city some $4 million a year. The swap, entered into 15 years ago as part of a bond sale to hedge against rising interest rates, has turned sour for Oakland now that interest rates are near zero. ... Oakland is paying 5.68 percent on debt associated with the swap, even with interest rates at record lows. Getting out of the contract would cost the city $16 million in termination fees, it says; it wants Goldman to waive the termination fees. ...

But at Tuesday's protest, civic leaders said the bank had benefited from a government rescue package during the 2008 financial crisis, and now it should give a break to cities like Oakland.

This is not a way to run a railroad, of course: Why Oakland? Why Goldman? Maybe AIG could give Oakland some of its bailout money? It got more than Goldman did, after all.

To really buy into Oakland's case you have to think that this swap deal was somehow unfair when it was entered into.* But it doesn't sound like Oakland is saying that: they entered into a synthetic fixed-rate deal with Goldman in 1998, and now they're sad that they didn't enter into a floating-rate deal. Since rates are down since 1998, that is a sensible thing to be sad about, but it's hardly Goldman's fault. Oakland's synthetic fixed-rate deal with Goldman has a rate of 5.6775% for a 23-year bond thingy; from Bloomberg it looks like the same Oakland agency issued a 15-year fixed-rate bond in December 1997 at 5.65%, and you don't see Oakland complaining to those bondholders just because rates subsequently went down.

Still, Oakland's city council seem to be the last people who could figure out if this swap is unfair, so maybe someone should do it for them. (HINT: lawyers?) One place to start would be Oakland's disclosure of the swap:

On January 9, 1997, the City entered into a forward-starting synthetic fixed rate swap agreement (the “Swap”) with Goldman Sachs Mitsui Marine Derivatives Products, U.S., L.P. (the “Counterparty”) in connection with the $187,500,000 Oakland Joint Powers Financing Authority (the “Authority”) Lease Revenue Bonds, 1998 Series A1/A2 (the “1998 Lease Revenue Bonds”). Under the swap agreement, which effectively changed the City’s variable interest rate on the bonds to a synthetic fixed rate, the City would pay the Counterparty a fixed rate of 5.6775% through the end of the swap agreement in 2021 and receive a variable rate based on the Bond Market Association index. The City received an upfront payment from the Counterparty of $15 million for entering into the Swap.

The Bond Market Association index, incidentally, is the SIFMA index, so Oakland avoided Libor manipulation. For a little while:

On March 21, 2003, the City amended the swap agreement to change the index on which the Swap is based from the Bond Market Association index to a rate equal to 65% of the 1-month London Interbank Offer Rate (“LIBOR”). This amendment resulted in an additional upfront payment from the Counterparty to the City of $5.975 million.

Of course it did. For no reason at all I figured I'd reconstruct this swap. You can find my work here. I am not a swapper of swaps so this is just for fun but by my math there's no smoking gun, though there's maybe a dagger smudged with a little bit of blood. Go play around if you'd like but some conclusions:

  • Oakland has paid about $42mm to Goldman under the swap so far in excess of what it's gotten back from Goldman, and will pay another $17mm over the next ten years,** for a total loss under the swap of $59mm.
  • That sucks.
  • But it's because rates went down.
  • It's not, incidentally, because of Libor nastiness: my guesstimate is that Oakland would have lost almost $64mm if it had used SIFMA instead of Libor.
  • But is part of Oakland's loss due to Goldman screwing them on pricing? Unclear. For giggles, I ran some numbers pretending that, instead of all those random up-front payments, Goldman and Oakland had just entered into a plain-vanilla Libor swap, on January 9, 1997 (when they did the real swap), at 65% of the publicly observable mid-market swap rate.*** That would mean a fixed rate for Oakland of about 4.55%, but it wouldn't have gotten its free $15mm in 1997 or its free $5.975mm in 2003. But it would have lost a total of $55mm on the swap, or almost $4mm less than it actually did.
  • You can call that $4mm GS's profit, though I'm actually eliding a bunch of real costs (it was a forward-starting swap, I don't know how the first payments worked and so may be off there, there's credit risk to Oakland, etc.) that make that look less outlandish than it was. (And probably some benefits that make it look more outlandish?) $4mm on a $187.5mm bond, or over 2% of notional, would be very pleasant for Goldman; sadly 2% is probably well within the margin of error of my stupid spreadsheet so there's no real reason to think that that's the right number. In fact, if you take the present value of these numbers as of 1997, discounting at a 20-year swap rate (and, I mean, maybe don't), Goldman seems to have done worse on its complicated structure than it would have on a vanilla mid-market swap, by almost $2mm. Perhaps Oakland drove a hard bargain.

Anyway. I'm not sure what the moral here is. Probably: don't enter into fixed-rate debt if rates will go down. Another possibility is: if you have the choice between doing something very plain-vanilla and at observable market rates, or doing something more customized that pays you $15mm up front, the one that pays you $15mm up front is always a worse idea. But, of course, if you're the City of Oakland, you wanted that $15mm in 1997. And you want it now! But, Goldman being Goldman, you only get it once.

Oakland leaders urge broad battle with Goldman Sachs [Reuters]
City of Oakland Comprehensive Annual Financial Report [massive PDF, I wouldn't click if I were you]
Here be spreadsheets [Google Docs & my own two hands]

* Or, I suppose, have a fanciful view of swap markets in which it doesn't cost Goldman anything to tear up this swap, so they're just being dicks asking for a $16mm unwind payment. This view is wrong. You can be sure Goldman hedged this swap with other parties who would charge it a similar amount to unwind.

** So a $16mm to unwind fee is a plausible fair value for that remaining $17mm.

*** Important: is that very wrong? Should a swap for 65% of Libor actually cost 65% as much as a swap for 100% of Libor?

**** Floating unattached footnote: no involvement with anything that happened here.