Jefferson County sewer swaps

A consent decree and a debt problem

Jefferson County, Alabama, contains Birmingham and, from 1996, a federal court order. That year a federal judge directed the county to rebuild its ageing sewer system, after raw sewage had been discharging into the Black Warrior and Cahaba rivers in breach of the Clean Water Act. Rebuilding a sewer system for a large, densely populated county is expensive, and the county borrowed heavily to pay for it. By the early 2000s it had around $3.2bn of fixed-rate sewer debt outstanding.

The refinancing, the hedge, and what was not disclosed

In 2002 and 2003, advised by JPMorgan, the county refinanced almost all of that fixed-rate debt into variable-rate and auction-rate bonds. On their own, variable-rate bonds are cheaper in normal times, but they leave the borrower exposed to rising rates. To manage that exposure, the county also entered a large book of interest-rate swaps. The structure was simple in principle: pay a fixed rate to a bank, receive a floating rate back, and let that floating receipt offset the floating rate owed to bondholders. The intention was to convert a floating liability into a synthetic fixed one and lock in a lower long-run borrowing cost. This is the standard swap-based hedge covered in Lecture 7, and on paper it read as sound financial engineering.

The deal-making around these bonds and swaps was corrupt. Between 2002 and 2007, Larry Langford, who served as president of the county commission and later as mayor of Birmingham, solicited and accepted more than $241,000 in cash, loan payoffs and gifts from a bond dealer and a lobbyist, in exchange for steering the county's lucrative bond and swap business their way. He was one of more than 20 people eventually convicted in connection with the county's sewer financing.

The 2008 downgrade

The structure held together as long as short-term rates behaved normally and the county's own borrowing costs tracked the broad market. In March 2008, Moody's downgraded the county's sewer revenue warrants, then around $3.2bn outstanding, from B3 to Caa3, citing a heightened chance of default. The monoline insurers that had guaranteed more than 90% of the sewer debt, XL Capital and FGIC, were downgraded too. Auctions on the county's auction-rate securities began failing outright, and attempts to remarket its variable-rate demand debt failed as well, which pushed the county's own borrowing costs to penalty levels written into the bond documents. At the same time, the downgrade triggered collateral-posting requirements on the swaps of around $184m, money the county said it did not have.

Prosecution and settlement

The gap between what the county now owed on its bonds and what it was receiving on its swaps did not close. It widened. The county spent the rest of 2008 trying and failing to renegotiate its way out, while state and federal investigations into the underlying corruption moved forward in parallel. In November 2009, the SEC settled its case against JPMorgan over undisclosed payments made to win the county's business: a $25m civil penalty, a $50m payment to the county, and forgiveness of more than $647m in termination fees that JPMorgan had claimed the county owed on the swaps. Two former JPMorgan managing directors, Charles LeCroy and Douglas MacFaddin, were charged separately. In March 2010, Langford was sentenced to 15 years in federal prison.

Bankruptcy and exit

None of that was enough to fix the county's finances. On 9 November 2011, Jefferson County filed for Chapter 9 municipal bankruptcy, listing about $4.2bn of total debt, of which roughly $3.14bn was the sewer debt. At the time it was the largest municipal bankruptcy in US history, a record it held until Detroit filed in July 2013 with an estimated $18bn to $20bn of debt. The bankruptcy court confirmed Jefferson County's plan of adjustment on 22 November 2013, and the county exited bankruptcy the following month, closing on about $1.8bn of new 40-year sewer warrants to refinance roughly $3.2bn of old sewer debt. Bondholders took a write-down of just over 40%, and sewer rates for county residents rose sharply for several years afterwards.

Two threads run through this case, and they teach different lessons. The swaps did not fail because the county made a wrong-way bet on the level of interest rates. They failed because the floating rate the county received from its banks and the floating rate it owed its own bondholders were never the same thing, and once the two came apart, nothing in the structure could bring them back together.

The intended hedge

This is a Lecture 7 case, and it turns on basis risk, not on a directional bet on interest rates. Jefferson County entered plain-vanilla interest-rate swaps. The county was the fixed-rate payer, paying a fixed rate to its bank counterparties (led by JPMorgan) and receiving a floating rate back, on a notional larger than the $3.14bn of sewer debt it was meant to hedge.

The intended structure was the textbook "synthetically fixed" hedge taught in Lecture 7. The county's underlying liability was variable-rate and auction-rate sewer bonds, on which it paid whatever rate the market, or after 2008 the auction-failure penalty rules, demanded. The swap was supposed to cancel this out: receive floating on the swap, let it offset the floating rate owed to bondholders, and be left paying a net fixed cost.

Where the two floating legs diverged

The case turns on the two floating rates involved. The rate the county received from its swap counterparties was tied to a broad market index, a fraction of one-month LIBOR. The rate it paid its own bondholders was the county's own auction-rate and variable-rate cost, which depended on the auctions clearing and on the monoline insurers keeping their credit rating. These two floating rates usually move together, loosely, but they are not the same index. Once the monoline insurers were downgraded in 2008 and the auctions failed outright, the rate the county paid jumped to contractual penalty levels, while the rate it received on the swaps, tied to a market index that was falling as the credit crisis unfolded, did not rise to match. The two legs, which were supposed to cancel each other out, moved in opposite directions instead. That gap between two related but distinct floating indices is exactly what basis risk means.

Size and termination risk

Size made the mismatch worse. The swap notional was large relative to the underlying sewer debt, so a given move in rates produced an outsized swing in the swaps' mark-to-market value. As rates generally fell after 2008, the swaps moved into the money for the banks and out of the money for the county, meaning the county would have owed large termination payments to exit them, at precisely the moment it could least afford to pay anything.

Complexity as cover

Product complexity sits alongside the basis-risk mechanism as a separate theme. The deal was arranged through undisclosed payments to intermediaries connected to county commissioners, so this was never simply a technical hedging failure. Multiple swap counterparties, auction-rate resets, monoline wraps and swap-collateral triggers made the structure hard for elected officials and the public to monitor, and that opacity is exactly where the undisclosed payments were hidden. Complexity, in this case as in others, was a price paid to whoever designed the product.