Tuesday, December 21, 2010

Eight bond markets a-sellin’

Did you know? The recent bond market sell-off has seen a 100 basis point rise in 10-year US Treasury yields in just five weeks. And there’ve been only seven previous sell-offs of similar speed and magnitude, according to Danske Bank.

According to Danske then, those sell-offs usually last around two months and the average increase in 10-year Treasury yields is around 120 basis points. They also add that bond yields tend to decline by 20-30bps in the months after a sell-off ends.
They also tend to coincide with changes in monetary policy, as this table shows:

But there’s quite a bit of variation in terms of post-sell-off moves:
a) In the instances where the Fed switches from being on hold to enter tightening mode, the performance is rather mixed. Post the sell-offs in 1994 and 1999, yields are roughly flat or slightly higher. However, after the sell-off during spring 2004, a strong bullish trend resumed, marking the beginning of the period of “the bond yield conundrum”.
b) In the cases where the Fed switches from being in easing mode to go on hold, there is a evidence that the sell-off reverses and bullish sentiment resumes shortly after the sell-off ends.
c) Where the Fed introduced additional monetary easing via the QE1 scheme on 18 March 2009, yields were pushed sharply higher, after an initial drop on 50bp on the day of the announcement. After the sell-off ended, the bond markets reversed the bearish trend and ten-tear yields were pushed 65bp lower in just one month.
So where are we now? The Fed introduced QE2 five weeks ago and since then the market has pushed up ten-year bond yields by 100bp, in a reaction that resembles the one that occurred after the announcement of QE1.
Over the past two decades, there is only one case of a bearish sell-off that led to a prolonged bearish trend in long bond yields. This occurred in 1994, when the Fed initiated a long period of monetary tightening. We are hardly in this scenario right now, and we expect the Fed to deliver a first hike in mid-2012.

Based on this case study, we see a good case for some short-term reversal of the selloff.
Up, down, movin’ all around.
The only thing that seems certain right now is volatility.

Thursday, December 16, 2010

Default and bankruptcy in the municipal bond market (part one)

A caveat: Although I cite the academic work of some well-established bond counsels in this post, I am not an attorney.  I am just writing this post to demystify a process that evidently needs demystifying.  (If I use any unexplained jargon, or if you happen to be a bond counsel and notice any mechanical errors, please let me know and I will make the appropriate amendments.)

One of the more frustrating aspects of muni market coverage in the news and blogosphere is the tendency to talk about municipal debt as if only one type of bond is issued and traded.  There is actually considerable diversity among borrowers in the muni market (e.g., they are not all government entities), and by extension, the types of commitments that are made for the repayment of the debt.  Although the relative health of the muni market has macroeconomic consequences, this is in many ways a market that defies generalization.  (That’s one reason I find the muni market unusually interesting…)  The defaults that have taken place both before and during the economic downturn are what finance-types would refer to as storied credits.  I often see people describing Jefferson County, Alabama, as the “canary in the coal mine” of muni defaults.  Suggesting that Jefferson County, which was the center of a widely-publicized securities fraud case, is a typical muni credit is kind of like portraying Enron as a typical corporate credit.  (Besides, the largest bagholder on that transaction, by far, is the same bank that that orchestrated the municipality’s distress, which is comme il faut if you ask me.)  Another example would be Florida dirt bonds, which are backed by special assessments on property in a severely depressed market.  These are not borrowers that were forced to establish their spending priorities or were muddling through difficult times; these are borrowers that experienced sudden and catastrophic losses and derived their revenues from limited sources.

Types of municipal bonds
The obvious starting place on this topic is to explain the types of muni bonds that are issued.  Municipal bonds are broadly divided into two classes: general obligation (GO) and revenue bonds.  The difference between GO and revenue bonds is the specific security that is pledged to repay the debt.  (Bonds may also be issued with more than one kind of security and may involve a moral obligation pledge that implies contingent financial support from another entity with stronger credit.)  GO bonds are secured by the full faith and credit of the issuer, meaning that the borrower is committing to raise taxes and other revenues sufficient to cover the amount owed.

Revenue bonds are secured by a defined stream of revenues.  Whether the principal and interest on these bonds is paid in a timely manner depends upon: (1) the reliability of the specific revenues pledged; and (2) whether that revenue stream has been pledged toward other debt or is used for other purposes.

It is important that you understand what kind of bond you have.  There are actually fewer GO bonds issued in the market than revenue bonds.  (GO bonds seem to receive a disproportionate amount of attention, which is frankly somewhat amusing considering how they are ultimately treated in bankruptcy proceedings – but more on that later.)  As of December 7, 2010, $137.9 billion of GO bonds had been issued this year, versus $251.2 billion of revenue bonds (Thomson Reuters data).  Some governments (even some states) do not issue GO bonds.  (If that seems counterintuitive to you, consider the existence of referendum requirements.)
State and local borrowers also issue what are known as “conduit” revenue bonds.  In a conduit bond issue, the government issues the bonds on behalf of a third party, generally a nonprofit entity (such as a hospital or college), a developer of government-subsidized housing projects, or a for-profit corporate entity (called industrial revenue bonds).  Conduit bonds are not government debt; they are private debts.  The bonds are actually secured by the revenues generated by the project being financed, the credit of the conduit borrower, and/or a mortgage on the property – and that is it.  (It is worth noting that the increased use of public-private partnerships has somewhat blurred this distinction – at a high price to state and local governments – but I could write a book on that topic.  I probably should – it would be the most politically scandalous book that three people ever read.)

Why do governments issue conduit bonds?  The short version is that government officials believe that the project will benefit the public in some way, and accessing the municipal bond market results in a lower interest rate for the borrower than the alternative (taxable) forms of financing.  (Read: This is a form of subsidy.)  Since the Tax Reform Act of 1986, the federal tax code has permitted states to allocate a limited amount of tax-exempt bond issuance each year for projects that will support private business activity.  The limit did not exist before this legislation, so keep that in mind when you see market statistics that reach far back in time.  Just to give you a sense of the magnitude, the cap for all states combined was $30.9 billion in 2010.  Some of that amount will be unused and carried forward, and issues for nonprofits are excluded but subject to many other geeky tax rules.  (I will exercise some restraint here.)

Just judging from news articles and my encounters in the blogosphere, the difference between government and conduit debt causes a lot of confusion with respect to the number of defaults that take place.  Conduit borrowers do not operate within the same framework of incentives as a government borrower.  (Sometimes, conduit borrowers are even just shell entities established to finance a specific project.  Compare that to a government borrower that theoretically exists in perpetuity.)  These are generally far riskier credits, and a large fraction of the defaults that actually take place in the municipal bond market are defaults on conduit bonds.  Recoveries are also likely to be smaller.  (You are more likely to see things like multiple-lien structures with this kind of debt.)

For some historical background on this point, Moody’s released a report back in February (subscription required) cataloguing municipal defaults and recoveries on bonds that the agency had rated from 1970 to 2009.  (Yes, yes, I know…  Suspend your judgment for a moment; enumerating defaults ex post facto is hardly rocket science.)  Moody’s counted 54 defaults total during that period, which included, by sector: housing – 21; healthcare – 21; utilities – 3; higher education – 1; recreation – 1; non-GO government debt – 4; and GO – 3.  (The issuer-pays model is doubly problematic in project finance because these issues generally require that an economic consultant draft a feasibility study for the project.  When has an economic consultant ever said that a project would not cash flow or that a designated district will not experience rapid growth?)

So far, this pattern has more or less held up for 2010 as well.  Of the monetary defaults that have taken place, approximately one-third were for conduit bonds and more than another third were for projects tied to property assessments.

As far as recoveries are concerned, Moody’s found that the average historical 30-day post-default trading price for municipal debt was $59.91 (par $100), versus $37.50 for corporate senior unsecured debt over the same period.

If you would like to know what the specific security for a bond issue is, look at the disclosure documents.  Many are freely available via the Municipal Securities Rulemaking Board EMMA system – use the search function, locate the individual bond issue, and follow the link to the official statement.  There will be a section devoted to discussing the security.  You can also search for continuing disclosure reports there.  (Regulators have made some strides at reeling in issuers for not reporting and in trying to have reports filed in a timely manner.)  There are also some subscription services that report on distressed debt specifically.

Monetary and technical defaults

So what happens when a government borrower defaults?  (For those among us who enjoy thinking through worst-case scenarios…)  The first thing you have to understand is that “default” is not a straightforward concept.  When most people talk about default, they are talking about monetary default, which is when a borrower fails to make a debt service (principal and interest) payment in full and on time.  Events of default, however, are actually defined in bond documents and include situations that are referred to as technical defaults.  With muni bonds, technical defaults may include (among many other things): a change in the tax status on the debt; a failure to comply with rate covenants (such as raising user fees on a project when the borrower’s debt service coverage ratio – the ratio of pledged revenues to required debt service – falls below a certain threshold); if the project being financed is not fully constructed; or if the issuer files for bankruptcy.  (A borrower can file for bankruptcy protection without there being a monetary default on a bond issue. This has happened in the past with GO defaults.)  If the bonds have a debt service reserve fund, unscheduled draws on these funds in order to make a debt service payment would also be considered a technical default.  (A debt service reserve fund is an account that is capitalized from bond proceeds or other sources of funds and provides bondholders with an additional buffer if the issuer experiences a temporary period of financial distress.)

Although these covenants generally exist to protect bondholders, it is not uncommon to see covenants that protect or (practically speaking) may even effectively give priority to other stakeholders in the deal.  A potential landmine for a bondholder (assuming the bondholder is not a bank), for example, would be a covenant that defines an event of default as nonperformance on a related financial instrument, like an interest rate swap.  If a derivative contract is associated with a bond issue (this should be disclosed, but it may be disclosed in a vague manner), then another layer of due diligence is required to gauge default risk.
There are two reasons that I mention technical default.  The first reason is that many of the statistics that are cited in market commentary will lump monetary and technical defaults together as “defaults.”  (I have done that in this post, in fact.)  Although these statistics may capture a level of distress in the market, they may also exaggerate it to the extent that there may be no meaningful risk of the bondholder not getting paid in some cases.  The second reason is that even with technical defaults a bondholder has rights that can be exercised.  For example, many issuers / conduit borrowers scrambled to refund outstanding debt when the bond insurers were swiftly being downgraded because it would be considered a technical default on the bonds if the bond insurer’s rating fell below a certain threshold, and payments on the debt could be accelerated.  (It is worth noting that such conditions were also generally carried over as a termination trigger on associated swap agreements.  This could result in the borrower having to find money to make a termination payment, depending on how the debt was structured and how interest rates have behaved since the contract was executed.)  My point here is that, in some circumstances, what could be statistically classified as a credit concern might be easily resolved with another transaction that is structured differently or simply involves different deal participants.

As an aside, one of the problems the muni market will likely face with the expiration of the BAB program and increased bank purchase limit is that market access for smaller borrowers, especially those with lower credit ratings, will be diminished.  This could place some strain on their ability to manage their outstanding debt and give smaller government entities an incentive to seek assistance from higher levels of government.  Policymakers would then have some tradeoffs to make.  This will likely have a negative impact on nonprofit borrowers for whom that is not even really an option.

Rights and remedies

Just like events of default, bondholders’ rights and remedies are defined in bond documents and state law.  These rights will be addressed in the bond resolution, ordinance, or other action through which the issuance of the bonds was authorized.  If the bond issue is part of an established debt program, there will be a trust indenture.  A trust indenture is a contract between the issuer of the bonds and the trustee (acting on behalf of bondholders) that governs the application of available revenues to the borrower’s debt service and other expenses / accounts.  (You will find a summary of the trust indenture in bond disclosure documents.  Also look for the “flow of funds” if you want to know where you rank in terms of getting paid and what other buffers against a monetary default exist in the structure.)

Attorneys with Vinson and Elkins provided a good overview of bondholders’ rights and remedies in the Summer 2010 issue of the Municipal Finance Journal (subscription required).  A common remedy in an event of default is that payment on the debt is accelerated. This pretty much means exactly what you think it means – the principal and accrued interest on the debt becomes due ahead of when it is scheduled.  (It is highly unlikely that a bondholder could force the sale of assets or take similar actions with a government borrower.  Conduit financings may involve a mortgage on the property, however.)

If it is in the bondholders’ economic interest to cooperate with the borrower, the bondholder may begin with a forbearance agreement.  With a forbearance agreement, the bondholder basically agrees not to declare an event of default while the government and its creditors negotiate a settlement.  Such a settlement could give the government some space to muddle through a temporarily rough period and not waste government resources (that could ultimately be passed on to bondholders) by addressing matters in court.  (Investors who have purchased asset-backed debt in the past few years are likely very familiar with forbearance agreements.)  This is the most likely outcome if the default occurred due to something like the timing of cash flows and the funds are expected to be recovered.  This option may be less palatable to retail (individual) than institutional investors (who can afford to wait), however.  Retail investors have a very strong presence in the muni market.

Bondholders may also seek a writ of mandamus from a court to compel government officials to take some specific action to cure whatever problem caused the default.  (Seeking court mandates to make government officials do things is practically a form of recreation in California, for matters not necessarily related to debt.)  For GO bonds, this would involve collecting taxes.  For revenue bonds, this would involve, for example, raising rates in accordance with a rate covenant (assuming this covenant exists with a particular bond issue).  Creditors can and will make life an absolute living hell for government officials and cost taxpayers a lot of money in attorneys’ fees if it comes to this. Aside from losing access to the capital markets, this is a very good reason governments do not generally repudiate their debt.  The legal process of answering for defaults only exacerbates officials’ administrative, financial, and political problems; it does not solve them, as some people have suggested.

Some government issuers are subject to legal provisions that provide bondholders with extra protections; for example, a state’s constitution may give debt service payments priority over all or most other expenditures.  Also, in some states, a state may take over the administrative responsibilities of smaller political subdivisions (cities, counties, etc.) if they run into serious trouble.  (This process may be automatically triggered by an event, such as if the local government’s credit rating falls below a certain threshold.)  Conversely, some borrowers may be subject to legal provisions that limit their ability to raise revenues.  Although this makes taxpayers happy, such provisions are major liabilities from a credit perspective, and (ironically) these taxpayers may end up paying more in interest to borrow funds in the bond market.

Another alternative to bankruptcy would be for the borrower to make a tender offer to bondholders, which is an offer to purchase the bonds back at a price that is in-between the market price of the bonds and the par value (face amount).  Bondholders may take a haircut (depending on their purchase price), but this may be better than drawing out the process, depleting the municipality’s resources, and getting even less.  Municipalities may also try to exchange existing bonds with new obligations that carry new terms (this is the capital markets’ version of a loan modification).  For muni issuers, these are much less convenient choices than you may surmise, because they are complicated by state and federal securities laws (such as getting all of the bondholders to agree to the plan) and tax issues (understatement of the year).

Some history (because history is awesome)
Since discussing the possibility of state defaults is all the rage these days, readers might be interested to know that previous state defaults did, in fact, progress through the above stages.  (Note: Federal bankruptcy law does not apply to states because they are granted sovereignty in the US Constitution.)  Many people cite Arkansas’ 1933 default as establishing the precedent for a state default, but there were others during the 19th century.  Pennsylvania’s 1842 default on its debt prompted William Wordsworth, whose family had bought Pennsylvania bonds, to compose the colorful sonnet “To the Pennsylvanians.”  Wait a minute, you say, wasn’t Wordsworth a Brit?  While apparently quite controversial now (with respect to the BAB program), when Americans were developing the frontier, they relied on financing from abroad to do so.  Pennsylvania’s default followed the Panics of 1837 and 1839, a period wherein there were plenty of municipal defaults to go around (and an epic number of bank failures).  This turned out to be great for foreign relations.  Per James Spiotto (Chapman and Cutler), “George Peabody, an eminent financier, sought to be admitted to polite English Society only to be rebuffed, not due to his lack of social grace, but because his countrymen did not pay their debts.  It was the defaults by government bodies in the latter half of the 1800s and early 1900s which brought about the procedures that are now taken for granted, including debt limitations on municipal issues, bond counsel, and clearly defined bondholders’ rights” (Handbook of Municipal Bonds). And, of course, the Confederate states repudiated their debts following the Civil War.

Bloomberg columnist Joe Mysak provided some excellent background on Arkansas’ default in a column several months ago.  Arkansas engineered a situation wherein one of the poorest states in the nation became the most indebted.  (Remember my comments about debt capacity from the last post?  This is precisely why such policies exist today.)  The state, already very much in debt, assumed the revenue bonds sold by hundreds of road districts (backed by a first lien on automobile and gasoline taxes) so these issuers could avoid default.  (Also keep in mind that the concept of professional credit analysis was only around twenty years old at this point.)  Thereafter, when state finances were, shall we say, strained by the Great Depression, state officials tried to exchange the outstanding revenue bonds for GO bonds with a smaller coupon.  The banks and insurance companies holding the bonds were outraged.  (Yes, they preferred a specific revenue pledge to a GO pledge.)  The state defaulted and officials blamed the underwriters for allowing the state to sell too much debt.

Bondholders took the state to court and successfully obtained a permanent injunction blocking the use of automobile and gasoline taxes for anything beyond debt service and highway maintenance.  The state then caved to bondholders (because it had no choice), refunded the debt with a longer maturity, and (gasp) raised taxes.

Arkansas’ problems
continued, and the state planned to refund the bonds once more when the bonds were callable.  Per Mysak, a syndicate of 250 banks was interested in bidding on the bonds (yes, even after the default, there was investor interest in investing in Arkansas – at a steep premium), but the Reconstruction Finance Corporation (established by Herbert Hoover) surprisingly picked up the whole deal.  And that is how the federal government got involved in Arkansas’ financial woes.  The RFC later sold the bonds to Wall Street banks at a profit.

As Mysak put it, “in 1933, nobody thought Washington should get involved in a state bond default.  In 2010, that’s the first place we look for help.”  People underestimate the system we already have in place.  The moral of this story for federal policymakers today is that maintaining a robust muni market is in their best interests too.  If their plan is to force social change on state and local governments by eliminating support for the muni market (as some have suggested, and who knows if this is true) and driving the various stakeholders in public finance to court (something that I think is highly unlikely anyway), they might find state policymakers simply enacting Thank Congressional Republicans Tax Acts of 2011, which will no doubt go over well with their constituents.  But it’s not like politicians to cut off their nose to spite their face, right?

Lunch is for wimps

Lunch is for wimps
It's not a question of enough, pal. It's a zero sum game, somebody wins, somebody loses. Money itself isn't lost or made, it's simply transferred from one perception to another.