A Muni CDS Market Primer

With increasing confusion over the cash muni bond market, very little has so far been said about the even more confusing muni CDS market. However, as municipal bankruptcies are likely about to take the country by storm, it is really the synthetic market that should be occupying investors' attentions. This is especially true with yesterday's disclosure that the bankrupt city of Vallejo is offering recoveries of only 5-20 cents to its sub creditors: it means that muni insolvencies will be not only a "survival" issue but one of recovery as well, considering assumptions embedded in cumulative loss forecasts that predict 80% recoveries by default. Below we present the most comprehensive report we have read so far on the matter of muni CDS, which should serve as a primer to anyone who wishes to be abreast not only of events in the muni cash space (where cash outflows are now comparable to what happened to equities following the flash crash), but in the wonderful world of synthetic paper.

From Santhosh Bandreddi at Bank of America

The Muni CDS Market

Municipal bond spreads have been under pressure since the advent of the European sovereign crisis because of the similar negative credit headlines around the fiscal woes of muni issuers. Investors use muni CDS (MCDS) and MCDX indices to hedge muni bond exposures by buying protection or, alternatively, monetize this risk premium in the muni CDS market by selling protection.

CDS contracts on Munis started trading back in 2003-04 but traded more actively since 2007-08, after the downgrade of the monoline insurers during the crisis. Some of the municipal bonds were insured by these monoline insurers and their downgrade meant higher counterparty risk for those muni bond holders.

Volumes in MCDS, although still low compared to corporate CDS, have been picking up this year with the increased interest in this market. Figure 2 shows the most actively traded MCDS names, since Jul ’10 when DTCC started publishing CDS trade volume data. Figure 3 aggregates the total volume across these 8 names for each week. Because of the lower liquidity in muni CDS, a similar notional trade can result in a bigger move in spreads in MCDS versus corp CDS.

Muni CDS still trades with a floating coupon or as a “Par CDS” i.e. if an investor sells/buys protection at 120 bps, the investors gets paid/pays 120 bps per annum carry with no upfront cash exchanged. Both US/EU corporate and sovereign CDS currently trade with a fixed coupon. However, efforts are currently on to design a “Big Bang” style protocol that includes fixed coupon convention and an auction settlement process for MCDS. Figure 4 shows a sample run of some of the liquid MCDS names, which are most liquid in the 10 year tenor (5y is the most liquid tenor in corporates).

Muni CDS Basics

Similar to corporate CDS, a muni CDS contract transfers the risk of a muni default from one party (the protection ‘buyer’) to another (the ‘seller’). The protection buyer pays a periodic premium to the protection seller, in return for providing protection against a credit event of the municipality.

If there is no credit event over the life of the CDS contract, the premiums are the only cash flows exchanged. However, if there is a credit event, the protection seller compensates the protection buyer for the loss on the underlying debt (exchanges par for the delivered bond). The protection buyer has the option to deliver the cheapest bond with a maximum maturity of 30 years, similar to a
sovereign CDS even though sovereign CDS incorporates cash settlement via an auction mechanism. Figure 5 and Figure 6 recap the cash flow exchanges that happen in a CDS contract.

Different types of MCDS

A MCDS contract may have either a Full Faith & Credit obligation (GO) or a revenue obligation as the reference obligation. While most single-name MCDS contracts currently traded are on U.S. states and reference a GO, some of the entities in the MCDX index reference a revenue obligation.

General Obligations

A general obligation (GO) is generally secured by a full faith and credit pledge of the issuing entity. Typically, the entity is a state or local government which pledges its full taxing power to payment of the bonds. While this pledge is strong based on the taxing power of the entity, it does not create a lien on any revenues or property of the public entity. In the event of failure to pay, the bondholder can pursue a lawsuit to seek payment. In addition to pledging its full taxing power, issuers may also define a priority of payments for general obligation bond debt service or provide strong remedies in the event that funds appropriated for debt service prove to be inadequate. The general obligation pledge for an issuer including debt limits, priority of payments, and bondholder remedies is typically defined under the issuer’s state constitutional or statutory laws.

Revenue Obligations

In contrast to general obligations, revenue obligations are secured by a discrete revenue stream in accordance with a Resolution or Indenture. The Resolution or Indenture creates a lien on the pledged revenues and may include liens on property as well. It also defines a priority of payments. A senior lien revenue bond can have a gross revenue or net revenue pledge. In case of a net revenue pledge, payments towards senior lien debt service come second only to operations and maintenance although this is not always the case. On the other hand, in the case of a gross revenue pledge the total revenues are applied for debt service on the bond.

The general obligation pledge is generally assumed to be stronger than the revenue pledge given the powers and resources of the GO issuers. This translates to lower probabilities of default (Figure 7) as well as higher assumed recovery values (see Figure 8). While this is true in the case of, for example, revenue bonds on toll roads, it may not necessarily be true for revenue bonds on water and sewer utilities.

Chapter 9 Municipal Bankruptcy

Municipal bankruptcy is covered under Chapter 9 of the U.S. Bankruptcy Code versus Chapter 7 and Chapter 11 that are resorted to on the corporate/taxable side of the markets. The greatest distinction between municipal bankruptcy and corporate bankruptcy is that municipal bankruptcy is voluntary on the part of the municipality. No other parties may put a municipal entity into involuntary
bankruptcy. From the outset, the municipal entity under consideration has a considerable amount of control in the proceedings. Since municipalities generally do not cease to exist, the motivation for a filing is often to seek relief and to invoke an automatic stay. The ultimate goal of the municipality is to work on a plan for recovery while seeking relief. In many municipal bankruptcies, but certainly not in all, existing bondholders have been made whole on principal (this circumstance tends to be more the case in the tax-backed sector versus some of the other sectors). There may exist other pressing matters that need to be resolved away from any debate as to whether or not to pay bondholders. Most issuers will do their utmost to see to it to pay bondholders because bondholders in general and for subsequent issuances are often part of the resolution of the circumstances that have contributed to the filing. 

Although filings by municipal entities are voluntary, state law determines whether that municipal entity is authorized to file or not. States, themselves, may not file for bankruptcy. The law varies considerably in this regard from state to state. According to research done by James Spiotto at Chapman & Cutler, LLP, the status of the ability to file or not must be broken down into several groups of states.

  • According to this legal research, a total of 16 states specifically authorize municipal entities to file. The list is as follows: AL, AZ, AR, CA, FL, ID, KY, MN, MO, MT, NB, NY, OK, SC, TX, and WA (We would state that after the New York City fiscal crisis in the 1970’s, the law was modified to require entities in NY to seek permission to do so).
  • Once again, according to the same legal research, a total of 7 states conditionally authorize municipal bankruptcies including CT, LA, MI, NJ, NC, OH, and PA. In the early 1990’s Bridgeport, CT, after filing for bankruptcy, was forced to withdraw the petition as the state maintained that the city was not authorized due to some clauses embedded in a bond transaction and due to the determination that the city was not insolvent.
  • A total of three states grant limited authorization under specified circumstances and in some cases for only specified credits: CO, OR, and IL.
  • Two states have an outright prohibition: GA and IA. However, concerning IA, a filing is permitted if insurance proceeds are not sufficient to cover existing debt in an involuntary circumstance. Presumably, the circumstance would be due to a natural hazard of some sort.
  • According to the law firm, the remaining 22 states either do not possess clarity or have no stated authorization in regards to the ability of a municipal entity to file.

We conclude from the legal analysis that in many states it is not a straightforward matter whether a municipal entity has the right to file or whether permission must be sought from the state or the courts in the state where the filing is contemplated before any such intended filing.

Another important aspect in the discussion of municipal bankruptcy is the matter of the “special revenue” doctrine. The bankruptcy code has been changed on occasion in important ways that pertain to municipal filings. Starting with the bankruptcy modifications made in 1988, a revenue that is pledged to the repayment of outstanding bonds may not be clawed back or intercepted by the host issuer. Let’s take the case of a hypothetical city, viz., Star City. Let’s consider the case where Star City has faced consequential financial difficulties that have forced the city to file. In the case where Star City has an enterprise system with revenue bonds outstanding that has the security pledge of the revenue stream generated by that specific enterprise, said revenues may not be diverted to pay Star City’s other obligations that are subject to the bankruptcy filing. If we say that the enterprise is a water system, revenues of the water system may be applied to the payment of outstanding water revenue bonds even while the host community, Star City, is in receivership. This doctrine is very important due to the fact that some sixty to seventy percent of issuance each year in the municipal market is in the form of revenue bonds. (Last year, revenue bonds accounted for 62% of the volume.) Conversely, it is also true that general fund moneys may not be forced to pay the revenue bonds unless the revenue bonds feature some kind of back up pledge from Star City or Star City elects to pay the revenue bonds from its own funds on a voluntary basis.

Finally, we would like to emphasize the role of a reference obligation in a MCDS contract. A reference obligation represents the  seniority of the deliverable obligations. Any bond that is pari passu or senior to the reference obligation can be delivered into the contract upon a credit event, as long as the bond has a tenor less than 30 years at the time of settlement. It is also important to note that both tax-exempt and taxable bonds can be delivered as long as they reference the same entity and have similar seniority as described above.

Credit Event Triggers for Muni CDS

  • Failure to pay - A muni issuer misses a payment of interest or principal on an obligation, beyond any grace period allowed on the obligation indenture (not more than 30 calendar days). If no grace period is specified on the indenture, a grace period of 3 business days is used. However, similar to most corporate CDS except Latin American ones, the protection buyer is not paid anything if the CDS maturity falls within the grace period.
  • Restructuring - Alters the principal amount, coupon, currency, maturity, or the ranking in priority of repayment of an obligation as a consequence of deterioration in credit worthiness or its financial condition. Unlike corporate CDS, bonds up to 30 year maturity can be delivered in the case of a muni default. This provision is usually called cum-Restructuring or old Restructuring.

We think Bankruptcy is not a credit event in an MCDS contract due to the following reason. Bankruptcy does not apply to U.S. state GOs because, as mentioned above, a state entity typically does not have the legal ability to declare/file for bankruptcy. While the non-state municipal entities can file for bankruptcy, chapter 9 gives them the power to modify or abrogate debt agreements which would be classified as a “Restructuring” event and hence should be covered by the MCDS contract.


Currently, all MCDS contracts are expected to settle physically i.e. upon a credit event, the buyer of protection can deliver any bond up to a maximum maturity of 30 years and receive par from the protection seller. The biggest hurdle to a physical settlement process is typically the existence of a higher notional of CDS contracts than the total eligible deliverable bond notional. However,  there haven’t been any credit events so far on MCDS since the inception of this market, and hence the settlement process is yet to be completely stress-tested.

An auction process, similar to the one in the corporate CDS world is in the process of being designed to enable cash settlement and avoid any bond squeeze in case of the CDS notional exceeding the outstanding deliverables. Market participants are still discussing the possible risks of implementing a cash settlement process as the protection buyers could deliver low/zero coupon 30 year bonds that trade at low dollar prices. They fear that this may result in a much lower recovery than the current recovery assumption for muni CDS of 80%. While the Western European sovereign CDS has similar issues, that market uses a recovery assumption of 40%, and that reduces the downside risk for a protection seller.

CDS-Cash Basis in Munis

Theoretically, MCDS is equivalent to a financed purchase of a muni bond with an interest rate hedge and, therefore, the MCDS spread should track the muni bond spread. However, like in the corporate market, there could be a basis between the two spreads based on demand/supply and transaction cost considerations. One other difference specific to the muni market that could contribute to a wider CDS-Cash basis in the MCDS market is that investors cannot short tax-exempt muni bonds by law and there isn’t an efficient repo market available for taxable BAB bonds.

Tax exempt muni bonds are usually quoted as a spread over the MMD curve, which is the AAA muni curve. For the purposes of comparison to a CDS, however, we need to first gross up the yield to a taxable bond equivalent yield level i.e. tax-exempt bond yield divided by 65% (or 1-tax rate). On the other hand, the BAB bonds are typically quoted directly on a taxable basis. Once we have a taxable yield level for the muni bond, we can then compute the spread over Libor (z-spread or CDS Par equivalent spread) to compare with the MCDS spread.

Figure 11 shows the current CDS-Cash basis for select state/muni CDS w.r.t both tax-exempt bonds (T/E) and Figure 12 shows the same versus BAB bonds. The T/E basis for Illinois and Massachusetts CDS is a big negative as the reference bonds used were 30 yr bonds while the CDS had a 10y tenor. The BAB bonds, on the other hand, should have had a big negative basis as all the Bab bonds  used were roughly 30y bonds. But the very small basis in the BAB bonds reflects the demand for these bonds that drove their spreads much tighter.

MCDX Index

MCDX is a portfolio of 50 equally-weighted MCDS referencing both GO and revenue obligations. The MCDX index was launched in May 2008 and trades in 5y and 10y maturities. While the index traded as different series over time, from S10 in Mar 2008 to S15 now, the constituents have not changed in this index over time - like the SOVX WE index. Among the 50 constituents in the index, 26 reference GO and 24 reference Revenue obligations. Figure 13 shows the growth in MCDX notionals over time but it should be noted that these are extremely small when compared to the CDX/iTraxx index notionals outstanding. For example, the total outstanding notionals across all CDX IG series is about 257x on a gross basis and 81x on a net basis, relative to the notionals outstanding on all  MCDX series. On the other hand, the gross and net notionals outstanding on global SOVX indices in comparison to MCDX indices is 17x and 4x respectively, which is more comparable.

MCDX is quoted on spread like the CDX IG index and can be valued through the CDSW screen on Bloomberg (Figure 14). Unlike MCDS, MCDX trades with a fixed coupon (100 bps) like the CDX index family.

As only half of the reference entities in the MCDX index (the 26 GO names) trade liquidly in single-names (MCDS), it is not easy to determine a fair value for the index. This also precludes index basis arbitrage trades and makes it difficult to accurately gauge if the index is wider or tighter to single names.

Figure 15 below shows the list of constituents of the MCDX index and if each CDS refers a GO (state) or revenue bond.