Blending a Proper Mix
By Samuel Solomon and Peter Shapiro
A major snowstorm that hit in January 2008 seemed somewhat symbolic of the situation taking place in the credit market. Widely used debt management techniques were freezing up, and at Northeastern University, Boston, we realized that we needed to do some significant restructuring or suffer soaring debt costs, pressure from bondholders, and other negative consequences.
Using a deliberate approach to debt restructuring, we've come up with a mix that results in a number of benefits: a more robust balance sheet, lower borrowing costs, and increased flexibility. Overall, we've learned that there continue to be a number of reasons for many institutions to issue variable or shorter-maturity debt.
Background on the Debt Scramble
Before the market crash of 2008, two forms of floating rate debt were widely used: variable rate demand bonds (VRDBs) and auction rate securities (ARS). These were simple, inexpensive, and easy to issue, and all types of institutions routinely used both forms of debt. The issuance was easy because of a number of factors: availability of municipal bond insurance; ready access and the low cost of bank-supported letters of credit (LOCs); and underwriters providing support for the debt that they issued by holding an inventory of bonds during short periods of insufficient investor demand.
These characteristics made floating rate debt quite popular. Issuers could access the market, without independent ratings if they had insurance, and could use swaps to synthetically create fixed rate debt at costs substantially below conventional fixed rate bonds. Since the support costs of floating rate debt were cheap and predictable, institutions that used swaps usually assumed that the all-in cost would equal the swap's fixed rate plus a nominal amount of 25 to 35 basis points in support costs.
The variable rate debt meltdown began before the larger market crisis in fall 2008. In late November 2007, auctions supporting the ARS market began to fail. As borrowers flocked to convert their ARS programs to bank-backed VRDBs, the availability of bank LOCs to support such bonds began to shrink; many banks exited the LOC market as their balance sheets became bloated with subprime assets. Those that remained were able to take advantage of the situation to increase the cost of the LOC and demand additional restrictive covenants from the institutions that were able to procure LOCs at all. At the same time, the bond insurers were suffering calamitous credit downgrades, making the debt they had supported even less attractive to the market and more expensive for the borrower.
As the floating rate bond market was melting down, other factors began affecting colleges and universities, adding to the already-significant financial pressures. These included declining endowment values that put stress on operating budgets and weakened the balance sheet; collateral calls that took effect when the mark-to-market on swaps became very high, as treasury interest rates plummeted; and the risk of violations of bank lending covenants.
Adding to the stress was the bankruptcy of Lehman Brothers, which had a large presence in the college and university bond-underwriting business and was a remarketing agent for many variable rate municipal bonds. The bankruptcy removed Lehman as a counterparty on thousands of swap transactions and left institutions scrambling to replace their relationship with the company.
One University's Response
Prior to 2008, almost all Northeastern University's outstanding debt was issued as auction rate securities (ARS). We overlaid this debt with swaps to create synthetic fixed rate debt. On the bonds, the university paid the floating auction rate, which on average mirrored the SIFMA Municipal Swap Index, or 65–70 percent of one month LIBOR, plus a 25 basis point dealer fee. On the swap, the university paid the fixed rate and received a floating rate designed to offset the floating rate on the bonds (70 percent of one-month LIBOR).
Thus, the university's net debt cost was the fixed swap rate plus the 25 basis points dealer fee. All-in, this structure provided the university with a significant reduction in costs as compared to issuing conventional fixed rate bonds. This relationship holds because the benefit of the tax exemption for municipal debt is normally greater for shorter maturity debt, as illustrated in Figure 1.
The synthetic fixed rate debt was "cheaper" than long-term fixed rate debt and had saved the university millions of dollars in interest expense over the preceding five years and beyond. On average, the Northeastern debt issues were 0.6 percent less than comparable fixed rate debt issued at the same time. On an outstanding debt portfolio of $590 million of par amount, this equated to $3.5 million
[0.6% x $590M=$3.54M] per year in interest savings. Over the five-year period ending in FY07, it is estimated that the university saved $12.7 million on less than $400 million of outstanding debt.
In addition, the ability to issue ARS with bond insurance from MBIA, Inc. meant that the university had never had to pursue a public rating from the rating agencies. The insurance provided all that the investors needed to be comfortable with the debt, and MBIA did its own analysis in setting the price for the insurance, which was very favorable for Northeastern.
Shaky ground. In late fall 2007, all this began to change. Just after Thanksgiving, we noticed that the auctions were not clearing as they had in the past. At that time, we made a decision to investigate the kinds of alternative forms of variable rate debt we could pursue.
Our first choice was to replace the ARS with VRDBs, supported by liquidity facilities issued by good, creditworthy banks. We began to solicit requests for proposals for liquidity facilities, with the hope of an early January ability to replace the ARS with bank-backed VRDBs. Early in January 2008, in the middle of that aforementioned snowstorm, it became apparent that even those banks that had originally submitted bids for liquidity facilities, to support a conversion of the debt from Auction Rate to VRDB, were backtracking and that there would not be a simple way to resolve the situation.
Unless changes were made, the university would be saddled with increasing variable rate debt costs, misaligned swap hedges, and pressure from bondholders who could not sell their ARS. It was decision time. We asked ourselves the following questions:
- Should we terminate the existing swaps, which would be costly because rates had dropped sharply? If so, this would require issuing taxable debt, in some cases.
- Could we replace variable rate debt with fixed rate debt? If so, what about the swaps? Having long-term fixed debt, with a swap that was not aligned with the debt structure, would create a total debt cost composed of fixed rate debt and a swap payment that would be significantly more expensive than the original synthetic fixed rate debt. The all-in cost of debt would be higher as the swap would remain in place, and the new fixed rate debt would be more expensive than the synthetic fixed rate debt being replaced.
- Would we be able to quickly acquire a public rating? We knew that any debt issuance or reissuance would require an independent rating, now that the bond insurers were in financial trouble.
A studied response. Facing this maelstrom, Northeastern took a measured approach; the university did not rush into refinancing debt. The first step was to quickly meet with the rating agencies and receive a public rating-and then to examine multiple refinancing options including: fixed rate debt, variable rate demand bonds, tender bonds, and short-term bullet maturities. We then scrutinized the swap portfolio to determine which of them, if any, could be terminated by using tax-exempt debt.
In order to issue VRDBs, we approached multiple banks to provide letters of credit. To obtain these letters, additional business and covenant requirements had to be met. The resulting restructuring left the university with a diversified debt structure (see Figure 2).
The covenants related to expendable financial resources to direct debt required in the letters of credit later caused additional stresses, as the decline in endowment market values—reducing expendable resources—created the possibility of breaching terms of the LOCs. The expected response of the banks was unknown in stressful economic times, and this uncertainty also led the rating agency to place the outlook for the university rating as negative.
That said, shortly thereafter, Northeastern was able to refinance all of its variable rate debt at favorable terms, thereby eliminating covenants, reissuance risk, LOC renewal risk, and bank LOC rating-change risk. We accomplished this by buying out all the VRDB debt and associated swaps and replacing the debt with long-term fixed rate debt.
The restructuring created a much stronger balance sheet and a ratings upgrade to a stable outlook, despite continued pressure on the broader market.
Tactics for Today's Market
Why would an institution that is issuing debt want to consider floating rate debt? This is a question that some may ask now, when long-term municipal and taxable debt are at or near historic lows—and given the disruptions that occurred in the municipal debt market in 2008-09.
Despite these questions, we discovered reasons why floating rate debt can be attractive and why it led Northeastern to maintain 20 to 30 percent of outstanding debt in a tender bond mode. Such an approach can:
Achieve lower cost. Variable rate debt is also at historic lows and is less expensive in interest costs than long-term fixed rate debt. The SIFMA Index, the market benchmark for variable rate debt, averaged 0.18 percent for 2011, by far the lowest in its history.
Make more efficient use of tax exemption. The short end of the yield curve has greater spreads to taxable debt than the long end of the curve. Today, on the short end, the tax-exempt floating rate benchmark SIFMA Index is 31 percent of taxable one-month LIBOR. On the long end, 30-year, fixed rate, single-A higher education bond yields are 145 percent of 30-year treasury yields.
Create a natural hedge. Most higher education institutions hold significant amounts of cash on their balance sheets, providing a natural hedge for floating rate debt. When rates are low, the return on the cash suffers, but the institution benefits from the lower cost of debt. If rates rise, the increased debt cost is offset by the higher interest earnings on the cash.
In a good economy, rates may rise, but other factors affecting universities will also become stronger, such as the ability of students to afford tuition or the availability of donations and research support.
Contribute to flexibility for readily and easily refinancing debt. Variable rate debt is callable at any time, or, in a tender mode is regularly remarketed, providing the opportunity to adjust structure to economic conditions.
Allow you to change modes if desired, such as from 1-year to 3-year tender, or even to fixed rate mode, if rates are attractive enough.
Be used to create synthetic fixed rate debt. Often synthetic fixed rate debt is at a lower rate than long-term fixed rate debt.
All of these factors are reasons why there is still a great benefit for many institutions to issue variable, or shorter-maturity debt.
Overall, the Northeastern experience taught the value of a proper mix of debt and the need for an A-rated institution to carry a lower percentage of variable rate debt than would an institution with a higher rating. There is value to variable rate debt as a method of reducing interest rate expense. However, the amount must be sized appropriately for the financial capacity of the institution, considering both the asset side of its balance sheet and the confidence it has in its ability to gain market access when needed. Clearly, the confidence factor is something that has become much more uncertain in the years since the financial crisis.
SAMUEL SOLOMON is director of finance and treasurer, Northeastern University, Boston; and PETER SHAPIRO is managing director, Swap Financial Group, an independent adviser on swaps and interest rate risk management.