The Use of Financial Derivatives in State and Local Government Bond Refinancings: Playing with Fire or Prudent Debt Management?
Luby, Martin J., Journal of Public Budgeting, Accounting & Financial Management
ABSTRACT. The esoteric area of financial derivatives has become quite salient in light of the financial crisis of the last few years. In the public sector, state and local governments have increasingly employed derivatives in their bond financings. This paper analyzes state and local governments' use of a specific type of municipal derivative instrument (a floating-to-fixed interest rate swap) in a specific type of transaction (bond refinancing). The paper provides a case study of an executed bond refinancing transaction that employed a floating-to-fixed interest rate swap quantifying the substantial long-term costs financial derivatives can impart on state and local governments. The paper concludes with some specific lessons learned about debt-related derivative usage for public financial managers and offers some suggestions for further empirical and theoretical research in this area of public financial management.
The esoteric area of financial derivatives has become quite salient in light of the financial crisis of the last few years. While much of the recent discussion in this area has been in relation to corporations', especially banks', use of these financial instruments and their contribution to the global financial collapse, state and local governments have also increasingly employed derivatives in their debt financings. Some estimate that the supply of municipal derivatives increased to over $500 billion in the years leading up to the financial crisis (Lucchetti, 2010). In general, debt-related derivatives are attractive to state and local governments because of the potential reduction in borrowing costs associated with the use of such financial products (Stewart & Cox 2008). For example, the execution of a financial derivative instrument in connection with a project financing or debt refinancing offers the possibility of significantly lower interest costs (especially in the short-term) vis-à-vis a financing that did not utilize such a financial instrument. In fiscally constrained times, such debt service savings can provide substantial operating budget relief for cash-strapped state and local governments. Thus, sub-national governments have justified debtrelated derivative use as a means of more efficiently managing their bond portfolio interest costs.
However, derivatives carry significant financial risks that have unfortunately manifested themselves to state and local governments since the onset of the recent financial crisis. As such, elected officials, the media and taxpayers are now just beginning to focus on the derivative practices of sub-national governments and their impact on government finances. For example, a March 22, 2010, Wall Street Journal front page story detailed how hundreds of state and local governments were losing money on interest rate swaps, a type of financial derivative, executed in recent years (Lucchetti, 2010) and that such losses were greatly exacerbating the operating budget deficits of these governments. Rolling Stone magazine, not a publication known for reporting on state and local government finances, detailed Jefferson County's use of interest rate swaps and the pay-for-play activities of several finance consultants and its impact on the county's much publicized troubled finances (Taibbi, 2010). As more information comes to light on state and local governments' use of these derivative instruments and their financial impacts, lawmakers have begun to propose restrictions on such risky financial instruments. For example, the Los Angeles city council recently instructed the city finance department to renegotiate or terminate an interest rate swap agreement that helped fund its water system while the State of Tennessee is attempting to restrict all of its local governments from using derivatives (Ackerman, 2009; Luchetti, 2010). Unfortunately, for many state and local governments, modifying or "getting out from under" these financial instruments is not easy and often can be quite expensive as evidenced by Jefferson County's scheduled $647 million termination payment on its swaps and the Bethlehem Pennsylvania school district's recent $12.3 million payment to terminate an interest rate swap (Lucchetti, 2010).
This paper analyzes state and local governments' use of a specific type of municipal derivative instrument (a floating-to-fixed interest rate swap) in a particular type of transaction (bond refinancing). It provides a case study of an actual bond refinancing that employed an interest rate swap as a means of quantifying the potential benefits and costs that financial derivatives can impart on state and local governments. The paper concludes with some specific lessons learned about debt-related derivative usage in bond refinancings for public financial managers and offers some suggestions for further empirical and theoretical research in this area of public financial management.
STUDY CONTEXT: DERIVATIVES AND BOND REFINANCING
The context of this study is the use of financial derivatives in municipal bond refinancings. Over the last couple of decades, state and local governments have increasingly used innovative bond financing structures in their debt management programs. One such innovative structure is the interest rate swap, a type of financial derivative. A financial derivative is an instrument whose value is derived from another asset or an asset index. The most common types of financial derivatives in municipal finance include futures, options, contracts, caps, floors, collars, rate locks, and interest rate swaps. As President Obama's economic advisor Austan Goolsbee creatively explains, a derivative can be likened to a taco coupon given out to attendees of a basketball game that entitles the attendee to a free taco if the home team scores at least a certain amount of points. That is, the coupon only pays off if some external event comes true, in this case, the home team scoring more than a certain amount of points, and you can trade it because it carries value. The taco coupon is not part of the game and does not have any impact on the game but it derives its value from its outcome but not the most fundamental outcome, who wins or loses (Javers, 2010). Financial derivatives are similar in that they are transactions that derive their value from other transactions yet do not impact the value of the transaction they derive value from.
An interest rate swap is an agreement between two parties to exchange a series of interest payments based on a notional amount of debt (usually a specified amount of outstanding bonds) without exchanging the underlying debt. The economic benefit from an interest rate swap is a result of the principle of comparative advantage. This comparative advantage emanates from imperfections in different money and capital markets. Bicksler and Chen (1986) clearly illustrate how a U.S. corporation and a foreign bank can enter into an interest rate swap agreement and capitalize on these market imperfections for the mutual benefit of each entity. For example, U.S. corporations often pay lower short-term interest rates but higher long-term interest rates than foreign banks that often pay higher short-term interest rates but lower long-term interest rates than US corporations. By entering into an interest rate swap agreement, a U.S. bank could issue short-term variable rate securities and a foreign bank could issue long-term fixed rate securities and swap interest rate payments whereby the net effect is that the US Corporation will effectively have a borrowing cost equal to the foreign bank's long-term fixed rate and the foreign bank will have a borrowing cost effectively equal to the U.S. corporation's short-term variable rate (Bicksler & Chen, 1986).
As it relates to the context of municipal bond refinancing, state and local governments sell debt obligations for many purposes. Most state and local government debt provides capital financing for infrastructure projects including the construction of schools, roads, government buildings, airports, water and wastewater systems. This debt is often structured with long-term maturities, often between 20 and 30 years, and annual principal amortizations that allow for a yearly incremental retirement of the debt. Due to the long-term nature of this debt, it is common for the state or local government to experience many changes from the issuance date until final retirement of the bonds. These changes, which include changes related to both bond market conditions and the state/local government's own financial and operational position, often provide opportunities for these government entities to consider refinancing all or a portion of their outstanding debt portfolios. The most common reason for a state or local government to refinance or refund its debt is to take advantage of lower interest rates1. In this case, the state or local government sells lower interest rate bonds and uses the proceeds from these bonds to pay off the higher interest rate bonds. This substitution of higher interest rate debt for lower interest rate debt provides operating budget savings for the state or local government.
State and local governments use quantitative metrics in the refinancing decision. These metrics often takes the form of a minimum present value savings threshold percentage. Present value savings involves calculating the annual interest savings associated with the refinancing (i.e., the difference between the principal and interest payments of the refinanced bonds and refinancing bonds each year) and present valuing each year's savings back to the current date using the refinancing bond's true interest cost as the discount rate. This present values savings amount is divided by the principal amount of the refinanced debt to get a present value savings percentage. Historically, it has been common for subnational governments to execute a refinancing if its present value savings exceeded a certain present value savings threshold, often between 3 percent and 5 percent (i.e., rule of thumb model) (Wood, 2008).
The rule of thumb model is commonly utilized but does not represent a very sophisticated way to make such a critical public finance decision (Kalotay & May, 1998). In short, the rule of thumb model is lacking because it does not provide any information on opportunity cost. That is, the opportunity cost of not being able to execute the refinancing at a future date when interest cost savings could be greater. Specifically, bond market conditions could change whereby future interest rates could decline from current rates and thus make a future refinancing much more economically beneficial to the state or local government than the refinancing being contemplated at the current date. Thus, there is value to exercising the call option today on the refinanced bonds (i.e., the intrinsic value) and there is value to exercising it later (i.e., the time value option). The time value option is foregone any time a state or local government executes a refinancing. In order for a state government to make a full analysis of whether to refinance or not, it needs to know the expected amount of both of these values. Some state governments have begun to use option models to make such calculations as a way to help them better evaluate their refinancing options (Kalotay & May, 1998; Zhang & Li, 2004). However, based on bond documents provided by the City of Chicago, a rule of thumb model calculating the present value savings discounted at the new bond's true interest cost was utilized in making the decision to refinance in this case (Morgan Stanley, 2004).
This paper will also seek to rely on and add to two different strands of applied empirical literature. First, this study attempts to contribute to the empirical literature on bond refinancing of municipal debt. The empirical academic literature on municipal bond refinancing is somewhat limited and mostly focused on the efficiency and timing of bond refinancings (Babad & Speer, 1978; Vijayakumar, 1995; Kalotay & May, 1998; Zhang Li, 2004). This paper will add to the literature by advancing the field's understanding of an emerging bond refinancing structure (e.g., floating-to-fixed interest rate swap refinancing) and the benefits and costs of using such a structure. The use of derivatives in the US state and local government finance sector is the second empirical literature this paper will seek to advance. Like the bond refinancing literature, most of the studies on municipal financial derivatives have been relatively limited in scope and mostly with a "best practices" focus (Gray, 1993; MacManus, Pfiel and Zibit 2003; Taub, 2005; Buchanan, 2005; Lamb, 2008).
This study is most connected to other municipal derivative studies related to valuation, use, and efficiency. In light of the increasing interest in municipal derivative instruments in the early to mid-1990s, several researchers developed models to value and evaluate financial derivatives. Gray and Engebretson (1992) created a synthetic asset pricing model to value the structural aspects of municipal inverse floating rate bonds (Gray and Engebretson, 1992). Gray and Cusatis (1993, 1995) developed the derivative asset pricing model (DAP) as a means of valuing five different types of derivatives by breaking down these securities into more basic securities that are more easily valued (Gray and Cusatis, 1993, 1995). Stewart and Cox (2008) empirically analyzed the debt-related derivative usage of the US states and the largest US municipal governments as of 2003 and found considerable derivative usage by state and local governments with 23 states and 23 of the these municipalities using derivatives in their financings totaling $32 billion in size according to their 2003 comprehensive annual financial reports. Cusatis (2006, 2008, 2009) has empirically evaluated the effectiveness and efficiency of certain financial derivatives and their ability to hedge interest rate risk and has concluded that the municipal swap market is desperately in need of a more effective and efficient hedging product (Cusatis). This study will add to Cusatis' research by showing the actual financial impacts on a sub-national government due to the lack of an effective swap hedging instrument. It will also lend empirical support to the import of Gray and Cusatis' valuation research that stresses the need for state and local governments to utilize sophisticated models in evaluating the risks of derivative instruments.
SYNTHETIC FIXED RATE BOND REFINANCING OVERVIEW
In recent years, many state and local governments have utilized interest rate swaps to create what is known as "synthetic fixed rate debt" for the purpose of more efficiently managing their bond-related interest costs. This type of structure entails that the government entity issue variable rate bonds and enter into a floating-to-fixed interest rate swap agreement with a counterparty, usually a large investment bank. The floating-to-fixed interest rate swap agreement consists of the state and local government receiving from the counterparty a variable interest rate based on some bond market index such as the one-month London Inter-Bank Offered Rate (LIBOR) (or a percentage of one-month LIBOR) or one-month SIFMA, the municipal bond swap index, and using such receipt to make the interest payment on the government's variable rate bonds. The state or local government then pays a fixed rate to the counterparty per a negotiated or competitive bidding. Figure 1 visually shows the cash flow exchanges in a typical floating-to-fixed interest rate swap. The floating-to-fixed interest rate swap agreement is structured with the goal that the payment received by the government entity from the counterparty will completely offset the interest paid on the state and local government's variable rate bonds leaving the fixed rate paid to the counterparty as the net payment made by the state and local government. Based on the principle of comparative advantage as described above, it is expected that this fixed rate paid by the state or local government to the counterparty would be lower than the fixed rate it would pay if it sold traditional fixed rate bonds without entering into the interest rate swap agreement. Thus, while the state or local government sells variable rate debt, the interest rate swap agreement attempts to convert the financing into a fixed rate obligation (this is why it is called a synthetic fixed rate debt) at a fixed rate lower than what government entity would have currently paid in the traditional fixed rate bond market. Figure 2 compares the cash flows between a traditional fixed rate financing and a synthetic fixed rate transaction.
Interest rate swap agreements, like all financial derivatives, carry significant risks that a state or local government must consider. These include risk related to basis, termination, counterparty, and enforceability (Buchanan, 2005, p. 37). Secondary but no less important risks relate to changes in the availability, cost, and/or creditworthiness of the credit enhancement needed in these types of municipal derivative transactions. In the Chicago case, these secondary risks resulted in a relatively substantial increase in the costs of the refinancing over initial expectations. Because of these sizable risks, the Government Finance Officers Association (GFOA) cautions that governments must thoroughly understand these risks in order to properly evaluate derivatives as financial tools and recommends that all governments that utilize interest rate swaps employ a comprehensive derivatives policy that will guide prudent financial decision-making in this area (GFOA, 2010). A discussion of all these risks is beyond the scope of this paper but basis risk is central to this study. Basis risk refers to the concern that the index used for the swap does not match the interest due on the underlying debt obligation. In the case of a municipal floating-to-fixed interest rate swap agreement, basis risk entails that the payment from the counterparty, as based on the LIBOR or SIFMA indices, does not totally offset the interest paid on the state or local government's underlying variable rate debt. For instance, the counterparty's payment in a typical floating-to-fixed interest rate swap is often based on a percentage of one-month LIBOR, oftentimes 67 percent of LIBOR. If one-month LIBOR was 3 percent and the government entity's bonds were bearing interest at 2.60 percent, the mismatch to the state or local government between the swap agreement receipt and the underlying variable rate bond interest cost (also known as basis mismatch) would be 0.59 percent (2.01 percent (i.e., 67 percent times 3 percent) minus 2.60 percent).
In the example above where there is significant basis mismatch, the use of a floating-to-fixed interest rate swap agreement essentially serves only as a partial hedge to interest rate changes. That is, the state or local government's effective interest cost includes the fixed rate payment to the counterparty plus the basis mismatch, in this case an additional 0.59 percent. However, the state or local government can also benefit from basis mismatch in the event that the floating swap payment is higher than the variable rate bond interest. Unfortunately, state and local governments will never know upfront the future extent and direction of this potential basis mismatch since its exposure will contract or expand according to bond market conditions. Moreover, the basis mismatch changes frequently usually in line with how often the variable interest rate on the underlying bond changes (usually daily, weekly or monthly).
If the purpose of the bond issue is "new money" (i.e., to fund new capital projects), theoretically at least, this basis mismatch may not be problematic in that the state and local government is effectively utilizing a variable rate financing structure - one that is partially fixed and partially variable (due to the potential basis mismatch). The use of variable rate debt to fund capital projects can be more cost effective due to the historical cost advantage of short-term, floating rate debt, and, as such, some public finance experts recommend its use, at least on a limited basis (GFOA, 1997). The reason that variable rate debt often can provide a lower borrowing cost over longterm fixed rate bonds is due to the term structure of interest rates whereby short-term interest rates are often lower than long-term interest rates. However, the floating rate component in a floating-tofixed interest rate swap agreement is not derived from accessing the short-end of the yield curve but rather from changes in the relationship between two short-term indices and possibly changes in the general credit quality of the state and local government's underlying variable rate bonds. Nevertheless, variable rate bond structures including structures utilizing floating-to-fixed interest rate swap agreements would generally not violate prudent debt policy in the case of financing new capital projects if they were used to finance a relatively small part of the capital plan. For example, the GFOA recognizes the potential interest cost benefits of debt-related derivative structures while at the same time cautioning state and local governments to carefully consider the risks involved with these financial instruments (GFOA 2010).
The use of synthetic fixed rate debt for the purpose of economically refinancing other fixed rate debt may be more problematic. Up until the last ten years, state and local governments generally advance refunded (i.e., refinanced bonds in advance of maturity to take advantage of low interest rate environments) their outstanding fixed rate debt using only fixed rate refinancing bonds. The use of fixed rate refinancing bonds made logical financial sense in that it "locked in" the interest savings on the refinancing through the final maturity of the bonds. Since the issuer knew the fixed rate principal and interest payments on both the refinanced and refinancing bonds, the interest payment savings were determined at the pricing of the bonds. Conversely, if the state and local government had attempted to use variable rate debt to refund the fixed rate bonds, the government would not be able to know the savings (or dissavings also known as negative savings) until the final maturity of the refinancing bonds since the refinancing bonds interest payments would not be known at closing since they were sold in a floating rate interest payment mode. Since the sole purpose of an economic refinancing was to provide interest payment savings and knowing this savings amount at pricing determined whether to proceed or not with the transaction, it was only natural to restrict that sale of refinancing bonds to a fixed rate interest mode2.
Over the last ten years, the use of interest rate swaps in refinancings became more prevalent in the municipal bond market. Based on a state bond refinancing database provided by Thomson Reuters and their corresponding official statements, 10 state governments executed floating-to-fixed interest rate swaps on over $6.6 billion in refinancing bonds between 2002 and 2009. According to other estimates, local governments executed swaps on potentially billions more of refinancing and new money bonds (Stewart & Cox, 2008). Issuers began to explore the use of derivatives as a means of "enhancing" refinancing transactions3. That is, if financial derivatives such as floating-to-fixed rate interest rate swap agreements ostensibly reduced borrowing costs over traditional fixed rate bonds in new money transactions, some government entities believed that these derivative products could be used just as successfully in refinancing transactions. However, the basis mismatch component of using synthetic fixed rate debt for advance refundings (as described above) essentially changed the refinancing paradigm from one where locking-in savings was primary to a paradigm where accepting some risk that refunding savings could change over the life of the refinancing was justified by a lower refinancing bond synthetic fixed rate as offered through the swap market. State and local government finance managers rationalized their acceptance of the basis risk inherent in synthetic fixed rate refundings by utilizing a counterparty swap payment that historically (i.e., over the last 3, 5 or 10 years) approximated the interest rate that they expected to pay on their underlying variable rate bonds thus attempting to address the potential for basis mismatch.
Unfortunately for many state and local governments, interest rate history did not repeat itself as the global financial crisis in late 2007 and 2008 caused the tax-exempt floating rate bond markets to collapse. Two types of floating rate securities issuers commonly sold for the underlying debt in synthetic fixed rate refinancings were auction rate securities (ARS) and variable rate demand obligations (VRDOs). Unfortunately, both of these security types were greatly affected by the financial crisis whereby issuers saw their weekly resets on these securities skyrocket up to sometimes as high as 20 percent while one-month LIBOR and one-month SIFMA rates (i.e., the indices that the swap payments received by the government usually were based upon in a swap) stayed in the 1 percent to 5 percent range. As state and local governments struggled for months to find cost effective ways to restructure their variable rate debt exposure, the sizable basis mismatch accruing in these transactions effectively eliminated a portion and sometimes all of the government entity's expected refinancing savings since the effective interest rate now consisted of the fixed swap rate plus the substantial basis mismatch payment. Thus, at the extreme, one could view, in retrospect, synthetic fixed rate refinancings where all of the refinancing savings were eliminated as serving no other purpose than increasing the supply of tax-exempt securities in the municipal bond market since these bonds did not end up providing any interest cost savings to the government entity selling the securities. Consequently, in addition to the impact on the state or local government's finances, the use of interest rate swaps in bond refinancing was costly to the federal government due to an increase in federal tax expenditures associated with a greater than necessary amount of tax exempt bond volume outstanding.
Another problem with using synthetic fixed rate debt vis-à-vis traditional fixed rate bonds is that interest rate swaps make future refinancing of the refunding bonds cost prohibitive. This is due to the fact that swaps must be terminated at market value if the state or local government wants to refinance and retire the bonds early. If a government entity wanted to refinance the refunding bonds at the call date in advance of maturity due to interest rates being lower, to get out of restrictive bond covenants, or because the swap was not as economical as expected (as was the case for many governments in the last couple years during the financial crisis), the government may have to pay a termination payment to the counterparty that could be substantial depending on current market interest rates. In the case of a potential economic refunding, this termination payment would likely offset any savings from refunding the bonds thus making a refinancing not economical. In the case of non-economic refinancings, the termination payment could still be expensive depending on the interest rate environment at the time of the proposed refinancing and budget constrained governments may not have available funds to make such payment. In sum, the termination at market value feature of interest rate swaps effectively provides the government little economic flexibility in managing its "swapped" debt in the future. The potential short-term benefits and long-term costs from using interest rate swaps in a bond refinancing will be detailed in the next section through a case study of a City of Chicago debt refinancing.
CITY OF CHICAGO CASE STUDY
The City of Chicago issued $500,000,000 in Second Lien Water Revenue Refunding Bonds (the "bonds") on August 5, 2004. The bonds advance refunded a portion of the City's Series 1995, Series 1997, Series 2000 and Series 2001 senior lien water revenue bonds. The total principal amount of refunded bonds was $460.744 million. The transaction was designed to produce economic savings, reposition debt from a senior lien status to a second lien status, update certain key terms of the bond documents and enable the City to better manage water rates in future years. Net present value "savings" were in excess of $31 million representing 6.72 percent of the bonds being refunded. The refinancing savings were primarily taken upfront with interest cost savings of approximately $10.7 million expected in each fiscal year between 2004 and 2006 with the remaining years producing no greater than $25,000 in savings each year. The true interest cost was estimated at 4.1889 percent for a bond offering with an average life of 18.8 years. Table 1 details the expected savings from the bonds.
When the City first announced the water refunding transaction in the spring of 2004, the derivative market afforded greater "savings" opportunity than the conventional bond market. A strategy of a synthetic fixed rate refunding (i.e., the use of a floating-to-fixed interest rate swap) was quickly adopted. The City decided on a BMA conversion swap whereby the swap provider would pay the City the BMA bond index during the escrow period (the period between the bonds' issuance date and the call date of the refunded bonds) and 67 percent of one-month LIBOR index thereafter. The City would pay the counterparty a fixed rate, issue variable rate bonds and use the counterparty payment to offset the payment on the underlying bonds thus leaving the fixed rate payment to the counterparty as the net payment on the transaction. The City competitively bid out the interest rate swap with RBC Dain Rauscher winning with a bid of 3.8669 percent (UBS matched this bid and received 50 percent of the interest rate swap transaction). Thus, the City would pay RBC and UBS 3.8669 percent and RBC and UBS would pay the City the BMA floating rate index for the first seven years of the transaction and 67 percent of the one-month LIBOR index during the remaining term of the bonds.
The City chose to use 7-day variable rate demand bonds, which it hoped would deliver an interest rate that tracked closely with the BMA index. The City opted for a blended credit enhancement structure, where municipal bond insurance was provided by MBIA and the liquidity facility provided by Dexia Bank. The municipal bond insurance policy cost the City $1,946,000 at closing. Liquidity facility and remarketing fees were 17.5 and 7.5 basis points (0.175 percent and 0.075 percent), respectively, of the outstanding bonds paid quarterly throughout the term of the bonds. The cost of the credit enhancement, other closing costs and the fixed swap rate produced an estimated true interest cost of 4.1889 percent. From closing until about January 2008, the interest rate swap performed effectively. That is, the variable rate swap receipt mostly matched the underlying variable rate bond interest resulting in the City having an effective borrowing cost of the fixed swap rate of 3.8669 percent plus liquidity and remarketing fees. Chart 1 demonstrates this initial lack of basis mismatch by comparing the weekly SIFMA (formerly BMA) variable rates and the bonds interest reset rates.
Unfortunately, the effectiveness of the interest rate swap would not last as the recent financial crisis significantly affected the municipal variable rate bond market. By late 2007, as the sub-prime mortgage crisis began to fully reveal itself, the bond markets began to question the financial soundness of most of the monoline bond insurers and many letter of credit banks that insured many of the municipal variable rate securities due to their exposure to the subprime mortgage crisis. As these insurers and banks began to be downgraded, the value of the variable rate debt insured by these companies dropped, which was evidenced by substantially higher variable rate bond interest resets. In many cases, these higher resets exceeded the maximum interest rate thus leading to many failed bond remarketings and auctions.
The financial crisis and its impact on the creditworthiness of the monocline bond insurers and letter of credit banks directly affected the bonds. The bonds were insured by MBIA which was officially downgraded in May 2008. This downgrade resulted in a substantial lack of investor interest in the City's bonds resulting in several failed remarketings which caused the City to draw on its liquidity facility and pay a much higher interest rate once such draw was made. This can be seen in Chart 1 which shows the spread between the weekly SIFMA interest rates and the bonds weekly resets interest rates. As shown in the chart, from January 2008 through November 2008, the interest rate spread between SIFMA and bonds was substantial. This widening produced significant basis mismatch whereby the City was paying substantially more on its variable rate bonds than the corresponding SIFMA interest rate it was receiving from its swap counterparty. Since the bonds refinancing plan relied on minimal basis mismatch, the interest cost savings began to swiftly evaporate as the City continued to pay more on its underlying variable rate bonds than it was receiving from its swap counterparty.
To address this basis mismatch and try to maintain as much of the originally expected refinancing savings as possible, the City executed two transactions. First, in April 2008, it refunded $100 million of the bonds with a fixed rate bond offering. Second, in November 2008, the City purchased letters of credit from two highly rated entities, CALPERS and State Street Bank, while terminating its bond insurance policy with MBIA and its liquidity facility agreement with Dexia on the bonds. Both of these actions helped minimize the impact of the basis mismatch both in actual dollars (from reducing the amount of variable rate bonds outstanding through the $100 million refunding) and in percentage terms (from the replacement of MBIA and Dexia with higher rated banks CALPERS and State Street Bank). However, these actions carried their own costs. First, the $100 million refunding essentially converted the variable rate bonds to fixed rate bonds with such fixed rate bonds carrying a higher interest rate than the synthetic fixed rate. This can be seen in Table 2 which shows the savings analysis from this refunding. The $100 million fixed rate refunding resulted in over $11 million in negative savings compared to the expected debt service costs on these bonds when they were originally issued in 2004 in synthetic fixed rate mode. Second, since $100 million of the bonds were refunded and retired, the City needed to terminate the portion of the swap that was executed in connection with these bonds. Since interest rates had declined since the execution of the swap, this termination resulted in a $2.9 million termination payment from the City to the counterparties. Third, the City had to pay a much higher LOC fee in the new credit constrained financial market environment than it was scheduled to pay Dexia under the original liquidity facility agreement. Specifically, the replacement of MBIA and Dexia with the letters of credit from CALPERS and State Street Bank increased the City's annual liquidity facility costs by 0.45 percent. That is, the original liquidity facility fees were 0.175 percent annually based on the total amount of outstanding bonds. The new letters of credit carried an annual fee of 0.625 percent of the total amount of the outstanding bonds.
The net effect of the basis mismatch from January 2008 through November 2008, the higher interest costs from the $100 million fixed rate refunding, the swap termination payment, and higher annual LOC fees can be seen in Table 3. The table provides a savings analysis for the bonds to date taking into account actual swap and interest payments, the $100 million fixed rate refinancing, the swap termination payment and the increase in LOC fees. This analysis assumes there will not be any basis mismatch in the future between the interest rate swap and the City's underlying variable rate debt. Moreover, it assumes that the letters of credit which expire in 2011 will be renewed at the same cost which is hard to predict since such renewal will be subject to the City's credit rating and the future availability of bank letters of credit. Based on these assumptions and the performance of the interest rate swap and underlying bonds to date, the analysis shows that the bonds will produce refinancing savings of -$11.785 million or -2.56 percent of the par amount of the refunded bonds with a $20.8 million increase in interest costs in 2008 compared to what was expected on the closing date of the bonds. In addition, the increase in interest costs from "fixing" $100 million of the 2004 water bonds and the increased LOC fees results in between $4 million and $4.9 million in additional interest costs from 2010 to 2023 which will put additional pressure on the City's water fund finances for the foreseeable future.
The updated present value savings of -$11.785 million or -2.56 percent of the par amount of the refunded bonds is obviously far lower than the present value interest savings expected at closing of $30.5 million or 6.64 percent of the par amount of the refunded bonds. Assuming the arbitrary (and flawed) minimum 3 percent to 5 percent present value savings refinancing rule of thumb that the City used as a benchmark of when to undertake a refinancing, in retrospect, the City clearly would not have executed this refinancing using a financial derivative structure given the savings results to date. However, one needs to be careful when evaluating this type of analysis for a couple of reasons. First, depending on market conditions, it may not be accurate to say the City should have just used a non-swap structure since it may not have been economically worthwhile to execute the refinancing if the City could only avail itself to traditional tax-exempt bonds. That is, a traditional fixed rate bond structure may not have produced the requisite savings (e.g., 3 percent present value savings) to warrant the transaction.
Second, it is not accurate to say that the City has lost $42 million in interest savings (i.e., the difference between the initial savings analysis of $30.5 million and the current savings analysis of - $11.785 million) for the sole reason that the initial analysis relied on the expected performance of the swap and underlying variable rate debt. The City really had no idea what its present value savings were at the time of the transaction. The initial analysis essentially said that if past performance is a proxy for future performance, the City will realize $30.5 million in savings. The current analysis assumes that the swap and underlying bonds perform as expected at closing taking into account actual performance to date and current costs calculating the savings as -$11.785 million. Thus, the City effectively bet that the swap and bonds would perform in line with historical performance and, if so, it would realize $30.5 million in savings. If it did not, it would realize a present value savings number less than $30.5 million of which it is now calculated to be -$11.785 million.
Unfortunately, this analysis may even be conservative since the City of Chicago bonds may not trade "flat" to the SIFMA index or 67 percent of LIBOR going forward given its past trading performance and the City's current deteriorating fiscal condition which will impact its future trading performance (Spielman, 2010). That is, there may be more basis mismatch to the negative detriment of the City in the future. Based on the potential for such a great change in the efficacy of a transaction utilizing derivative instruments, it is clear state and local governments should use more sophisticated models (similar to the ones discussed in the literature review) that employ Monte Carlo simulation calculating the expected future cash flows taking into account volatility risk related to interest rates, historical relationships between bond indices, and bond credit rating changes. This model should also evaluate the potential cash flow impacts related to future changes in the availability, cost or creditworthiness of the credit enhancement necessary for municipal derivative transactions.
In sum, the primary results of the 2004 water bonds refinancing have turned out to be (1) a significant amount of expected negative present value savings, (2) an increase in tax-exempt bond volume that served no subnational government purpose since the refinancing savings have been depleted, (3) significant short-term budgetary savings from 2004 through 2006 (since almost all the interest savings were structured to occur in the first 3 years when the interest rate swap performed effectively) and (4) substantial budgetary cost to the water fund from 2008 through 2024 most notably in 2008 when the refinancing resulted in $20.8 million in negative savings.
LESSONS LEARNED FOR PUBLIC FINANCIAL MANAGEMENT
The financial problems caused by the use of derivatives in municipal bond refinancings provide some important lessons for public financial managers as such use relates to future debt management practices4. The overriding lesson learned is that bond refinancings should only be executed when the government entity can guarantee that the savings are "locked in" over the life of the bond issue. This implies that the use of bond structures that effectively result in variable rate debt, such as floating-to-fixed interest rate swaps, is inappropriate when the sole purpose of the transaction is to produce economic interest cost savings. "Drilling down" from this overriding lesson, this paper's analyses provide the following more technical lessons.
Lesson 1: Financial Market Stability is Not a Given
State and local government issuers of variable rate bonds believed that this floating rate market would always remain stable. Unfortunately, credit markets froze and investors and investment banks didn't have the cash to bid on securities; Bondholders could not liquidate their investments, and issuers were stuck paying much higher interest rates on their underlying variable rate than what they were receiving from the swap counterparty - thus producing a sizable net interest cost premium over their fixed swap rate. Financial markets will not always be strong and stable, and issuers should prepare for times of uncertainty and not execute transactions (such as bond refinancings) whose success depends solely on this uncertainty being minimal or non-existent.
Lesson 2: Financial Intermediary Risk Is Real
State and local governments should not assume that the credit quality of their capital market partners such as large investment banks, commercial banks, and bond insurers (i.e. financial intermediaries) is constant and not subject to dramatic change. The financial events of the last few years clearly show the inaccuracy of that assumption. The downgrading of the City of Chicago's original letter of credit bank and bond insurer made its water bonds much less attractive to investors, thus substantially increasing the bonds' risk premium and corresponding interest rate. Bank credit is not infinite, may change in value, and could cost significantly more in the future.
Lesson 3: Interest Rate Risk Should Be Shifted Permanently
In the case of bond refinancings, the risk that interest rates could change in the future is the determining factor in whether the financing is successful or not (i.e., does the bond issue provide interest cost savings to the government or not and how much). Thus, it is critical to make sure risks transferred at the time of bond sale stay shifted throughout the life of the contract. In the case of the City of Chicago's water bond refinancing, changes in interest cost were borne by the City due to basis mismatch on the interest rate swap as well as to changes in liquidity costs. In the case of a bond refinancing, interest rate risk should truly be eliminated and often the only way to accomplish this is by using a traditional fixed rate bond structure. For some governments, creating a variable rate stabilization account or having money on hand that serves as a natural hedge on the economics of the swap and the underlying variable rate bonds can also mitigate interest rate risk. Unfortunately, in this era of fiscal stress, most state and local governments do not have such large pots of money to act as natural hedges for swaps and variable rate instruments.
Lesson 4: Manager Should Stress Test Contracts and Craft Bond Structures with Reasonable Flexibility
Public financial managers need to understand the precise economic impact on its finances from changes in value of its financial contracts in the event of specific market situations. This should entail significant amounts of financial stress testing or sensitivity analysis prior to executing a financial instrument such as a municipal derivative. Public financial managers should consider formally analyzing the fixed components of financial derivatives (e.g., fixed swap payment, up-front credit enhancement fees) that will not change with market factors and the variable components (e.g., risks related to basis, termination, counterparty, enforceability, credit enhancement, etc.) which will tie both components together in a cost of funding analysis. Ideally, such analyses should include a probability distribution of simulated cash flows scenarios but, at a minimum, include at least best, worse, and expected case scenarios.
In addition, the fair market value requirement associated with early termination of a swap severely limits the future financial flexibility of the government to refinance in the future. One never knows if there will be a need to refinance in the future either to take advantage of lower interest rates or to retire the bonds early due to legal requirements (as in the case of a privatization of an asset). Using such bond structures significantly "ties the hands" of future public managers and could threaten their ability to effectively respond to changing market or policy conditions. The collapse of the municipal floating rate bond market and the challenges the City of Chicago faced in reducing exposure to this market in a cost effective manner shows the benefit of using bond structures that provide for a reasonable amount of flexibility should dramatic changes occur in the market. While traditional fixed rate bonds are the strongly preferred structure for refinancing debt, state and local governments could address some of the aforementioned concerns with more flexible instruments (such as callable swaps).
Lesson 5: Managers Should Recognize That Historical Relationships between Bond Indices May Not Hold in the Future
Almost all equity market information provided by brokerage firms is prefaced with the warning "past performance is not a guarantee of future performance." Such is the case with the relationship between different interest rate indices. Over the last decade or so, the relationship between SIFMA index and 1-month LIBOR was between 65 percent and 70 percent (i.e., the SIFMA index was 65 to 70 percent of one month LIBOR). As such, many of the interest rate swaps in bond refinancings were structured so that the swap counterparty would make a payment in this range and this would offset the variable rate paid on the underlying bonds. As the financial crisis deepened, this relationship changed dramatically with the relationship exceeding 100 percent for a significant period of time in 2008. During this time of severe basis mismatch, the economic savings of the refinancing transaction evaporated. State and local governments should never assume that historical relationships between bond indices are stable and not susceptible to sharp and dramatic changes.
Lesson 6: Up-front Refinancing Savings Structures Should Be Avoided
To produce maximum operating budget relief, the City of Chicago structured its refinancing savings to occur primarily in the first three years between 2004 and 2006, effectively using the refinancing to produce a short-term budget benefit (described as a "one shot" revenue in the budgeting literature). This structure entailed that the city somewhat "back-load" the amortization of the refunding bond principal. Because of such back-loading, the City had more variable rate debt outstanding subject to basis mismatch at the time of the financial crisis (i.e., 2007 and 2008) than if it had used a more conservative refinancing saving structure (i.e., greater amortization of bond principal in 2004, 2005, 2006 and 2007) whereby the savings would be taken more evenly each year throughout the maturity of the bond issue. In addition to creating or exacerbating structural budget deficits, up front savings structures should be avoided given uncertainties related to future interest rate changes.
CONCLUSION AND FUTURE RESEARCH DIRECTIONS
In light of the never ending quest to maximize financial resources in an era of limited appetite for increased taxes and lower spending, state and local governments turned to innovative financial management practices to accomplish such a goal. The use of financial derivatives, and specifically floating-to-fixed interest rate swaps, in bond refinancings represents a prime example of such an approach. The potential benefits from these instruments in the form of greater expected interest costs savings (which directly translate into lower operating budget costs) vis-à-vis using more traditional refinancing bond structures offered state and local governments greater flexibility in avoiding spending cuts (or increased spending) and/or tax increases (or tax reduction). However, these benefits can come at a cost in the form of potential future increases in interest costs in the event there are changes in the bond market as it relates to liquidity, bank credit, investor appetite, and/or relationships between bond indices. The financial crisis of the last few years has shown how these changes could actually come to fruition producing significant financial hardships to state and local governments from their use of such derivative instruments as acutely evidenced by this paper's City of Chicago case study. With respect to bond refundings, the overriding lesson learned is that transactions should only be executed when the government entity can guarantee that the interest savings are "locked in" over the life of the bond issue and, thus, state and local governments should avoid financing structures that include any interest rate, liquidity, financial intermediary, or counterparty risks when refinancing its debt. Floating-to-fixed rate swaps may not be as problematic for new money bond transactions as they effectively produce a variable rate bond structure which may be appropriate to employ in some state and local governments' capital financing programs.
This paper empirically demonstrates through case study analysis the potential problems of derivative usage in bond refinancings. Given that the derivative market is unregulated and there is not a repository of information on these financial instruments, further research is needed to determine the extent of derivative usage by state and local governments over the last ten years not only in refinancing transactions but also in new money financings to determine the significance of this area of financial management. Some have estimated this market to be quite large while others have downplayed its size (Stewart & Cox, 2008; Mysak, 2009; Luchetti, 2010). Empirical research detailing the use of debt-related derivatives by state and local governments during the last ten years as investment banks accelerated their marketing of these instruments would be helpful in understanding the significance of this public financial management area.
A second strain of future research is needed to determine the factors associated with sub-national governments' usage of these derivative instruments as a means for enhancing our understanding of public financial management decision-making. Statistical analysis on the decision to use swaps and its various explanatory factors would likely shed greater light on this area of debt management decision-making by providing insights into the factors impacting the decision to use or not use derivatives. In addition, such analysis may allow further theory testing in this area of public financial management. For example, an analysis of interest rate swap usage in debt refinancing taking into account the amount of state and local governments who were statutorily allowed and actually executed derivative transactions may provide a good test of agency theory and/or stewardship theory in public financial management. That is, these financial instruments involved information asymmetry and delegated bureaucratic decision-making whereby the usage benefits may be short-term or may not surface at all (and may only accrue to the benefit of the elected official, the agent while the long-term costs could potentially be substantial (to the detriment of the taxpaying public, the principal) thus allowing for the prospect of moral hazard. Pervasive use of these instruments may provide support for agency theory while limited use could support some semblance of stewardship theory whereby the decision not to use swaps by public financial managers' would show evidence that their interests were aligned with the goals of their ultimate principal, the public, and that these interests and goals were the long-term cost efficient management of debt.
The author would like to thank Craig Johnson, Kurt Zorn, David Reingold, Sergio Fernandez, Sharon Kioko, William Kingore, and Charles Wise for their very helpful comments on a previous version of this manuscript.
1. In the parlance of municipal finance, a refinancing is often referred to as a refunding and those terms are used interchangeably in this paper.
2. Thus, one can see the difference in risk from using synthetic fixed rate bond structures for a new money transaction compared to an economic refinancing transaction. If the swap is not cost effective in the new money transaction compared to a traditional fixed rate bond structure, the financing may not be optimal but the primary purpose of the transaction will still be satisfied (i.e., the sole purpose is to provide financing for a capital project). If the swap is not cost effective in the economic refinancing transaction compared to a traditional fixed rate bond structure, in the extreme the primary purpose of the transaction may not be satisfied (i.e. the sole purpose is to provide interest cost savings to the state and local government).
3. My analysis of interest rate swaps for refinancing purposes does not include the use of forward starting swaps as that type of derivative product is primarily used to hedge interest rate risk in the future using a forward starting swap to lock in today's interest rates to refund bonds not allowed to be advance refunded.
4. These lessons are also applicable to new money bond transactions as the same risks exist for refinancing and new money transactions.
Ackerman, A. (2009, May 4). "Tennessee Seeks to Curb Derivative Deals." Bond Buyer: 1.
Babad, Y., & Speer, P. (1978). "Optimization of the Arbitrage Process in Advance Refunding of Municipal Bonds." Management Science, 24 (10): 987-1000.
Bicksler, J., & Chen, A. (1986). "An Economic Analysis of Interest Rate Swaps." The Journal of Finance, 41 (3): 645-655.
Buchanan, A. B. (2005). "Dealing with Municipal Swap Risks." Municipal Finance Journal, 26(3): 35-43.
City of Chicago (2004). Official Statement. $500,000,000 Second Lien Water Revenue Refunding Bonds Series 2004. [On-line]. Available at http://emma.msrb.org. [Retrieved March 31, 2010].
City of Chicago (2008a). Official Statement. $549,915,000 Second Lien Water Revenue Refunding Bonds Series 2008. [On-line]. Available at http://emma.msrb.org. [Retrieved March 31, 2010].
City of Chicago (2008b). Tax Compliance Certificate. $549,915,000 Second Lien Water Revenue Refunding Bonds Series 2008. [Online]. Available at http://emma.msrb.org. [Retrieved March 31, 2010].
City of Chicago (2008c). Official Statement. $395,925,000 Second Lien Water Revenue Refunding Bonds Series 2008. [On-line]. Available at http://emma.msrb.org. [Retrieved March 31, 2010].
Cusatis, P. (2006). "Municipal Bond Basis." Municipal Finance Journal, 27 (3): 19-33.
Cusatis, P. (2008). "An Analysis of the Failed Municipal Bond and Note Contracts." Journal of Futures Markets, 28 (7): 656-679.
Cusatis, P. (2009). "Hedge Effectiveness and Efficiency in the Municipal Market." Municipal Finance Journal, 30 (2): 49-69.
Government Finance Officers Association (GFOA) (1997). Using Variable Rate Instruments (1997). [On-line]. Available at www.gfoa.org. [Retrieved January 28, 2008].
Government Finance Officers Association (GFOA) (2010). Use of Debt- Related Derivative Products and the Development of a Derivatives Policy (2010). [On-line]. Available at www.gfoa.org. [Retrieved May 20, 2010].
Gray, G.J., & Engebretson, K. (1992). "Residual Interest Bonds (RIBS)" Municipal Finance Journal, 13 (1): 1-29.
Gray, G.J., & Engebretson, K. (1993). "An introduction to municipal derivative products." Government Finance Review, 9 (2): 17-19.
Gray, G.J., & Cusatis, P.K J. (1993). "Understanding and Valuing Municipal Derivative Securities." Municipal Finance Journal, 14 (2): 1-41.
Gray, G. J., & Cusatis, P. J. (1994). Municipal Derivatives Securities: Uses and Valuation (1st ed.). New York: Irwin Professional Pub.
Javers, E. (2010, April 20). "Debate Test Members' Financial IQs." Politico: 1.
Kalotay, A., & May, W. (1998). "The Timing of Advance Refunding of Municipal Bonds" Municipal Finance Journal, 19 (3): 1-15.
Kalotay, A., Yang, D., & Fabozzi, F. (2007). "Refunding Efficiency: A Generalized Approach" Applied Financial Economics Letters, 3 (3): 141-146.
Lamb, R. (2008). "Asset-Liability Management: Managing Swaps." Municipal Finance Journal, 29 (1): 23-26.
Luchetti, A. (2010, March 22). "Interest Rate Bets Bite Cities, States." Wall Street Journal: 1.
McManus, K., Pfeil, K., & Zibit, T. (2003). "Guidelines for the Effective Uses of Swaps in Asset-Liability Management." Government Finance Review, 19 (3): 34-38.
Morgan Stanley (2004, August 5). Final Pricing Information: $500,000,000 City of Chicago, Second Lien Water Revenue Refunding Bonds, Series 2004. (Online). Available at www.morganstanley.com.
Mysak, J. (2009, November 24). "Municipal Derivative Mythology Turns on 500 Swaps." Bloomberg. [On-line]. Available at www.bloomberg.com. [Retrieved May 17, 2010].
Spielman, F. (2010, August 5). "City faces higher borrowing costs as bond rating is downgraded." Chicago Sun Times: 8.
Stewart, L. J., & Cox, C. A. (2008). "Debt Related Derivative Usage by U.S. State and Municipal Governments and Evolving Financial Reporting Standards." Journal of Public Budgeting, Accounting & Financial Management, 20 (4): 466-483.
Taub, D. L. (2005). "Understanding Municipal Derivatives." Government Finance Review, 21 (4):18-23.
Taibbi, M. (2010, April 15). "Looting Main Street." Rolling Stone: 38- 41.
Vijayakumar, J. (1995). "An Empirical Analysis of the Factors Influencing Call Decisions of Local Government Bonds." Journal of Accounting and Public Policy, 14 (3): 203-231.
Wood, W. H. (2008). "Municipal Bond Refundings." In S. Feldstein and F. Fabozzi (Eds.), The Handbook of Municipal Bonds (Vol. 1, pp. 235-246). Hoboken, NJ: Wiley and Sons Publishing.
Zhang, W. D., & Li, J. (2004). "The Value of the Advance Refunding Option and the Refunding Efficiency of Tax-Exempt Municipal Bonds." Municipal Finance Journal, 25 (1): 17-31.
Martin J. Luby*
* Martin J. Luby, Ph.D., is an Assistant Professor, School of Public Service, DePaul University. His teaching and research interest are in state and local government capital markets, public financial management, and organizational behavior. Prior to pursuing his academic career, he spent 13 years in the public finance industry as an investment banker and finance consultant to state and local governments.…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: The Use of Financial Derivatives in State and Local Government Bond Refinancings: Playing with Fire or Prudent Debt Management?. Contributors: Luby, Martin J. - Author. Journal title: Journal of Public Budgeting, Accounting & Financial Management. Volume: 24. Issue: 1 Publication date: Spring 2012. Page number: 1+. © PrAcademics Press, Florida Atlantic University Spring 2007. Provided by ProQuest LLC. All Rights Reserved.
This material is protected by copyright and, with the exception of fair use, may not be further copied, distributed or transmitted in any form or by any means.