White smoke rises in the east! Later today (Friday), it is expected that the House-Senate Conference Committee will release a final draft of the Tax Cuts and Jobs Act. Consistent with what was in both the Senate version (discussed here) and the House version (discussed here), it is further expected that the draft released by the Conference Committee will eliminate tax-exempt advance refunding bonds.[1] In a misconception akin to an economic “epiphany” scribed on the back of a cocktail napkin, the House indicated that the elimination of tax-exempt advance refunding bonds will raise revenue because the federal government will no longer be extending a subsidy to two issues of tax-exempt bonds that are outstanding simultaneously to finance the same project. According to the Joint Committee on Taxation, eliminating tax-exempt advance refunding bonds would save $16.8 billion over the next 10 years. Aside from the savings identified by the Joint Committee on Taxation, the Senate could not be bothered to provide an explanation for the elimination. We discuss the proposed elimination of advance refunding bonds and scrutinize the purported savings here.
If prevailing market interest rates decline after an issuer issues fixed-rate tax-exempt bonds, an issuer can capture the benefits of the decline in interest rates by using an advance refunding structure to refinance those tax-exempt bonds because most fixed-rate tax-exempt bonds have “call protection” of around 10 years from the issue date, meaning that the issuer cannot redeem the bonds during that period. Call protections are important to investors because they guarantee a particular return for a particular period of time.
Usually an issuer will embark on an advance refunding when market interest rates have dropped sufficiently since the bonds were issued. This is referred to as a “high to low” advance refunding because the high interest rate the issuer paid on the refunded bonds is reduced to a lower interest rate on the refunding bonds. (There are other reasons for refunding bonds, such as defeasing bonds to eliminate the yoke of burdensome financial covenants, but usually refundings are done for savings.)
When deciding whether the drop in interest rates in the market is “sufficient” to warrant a refunding, the issuer takes into account the period prior to the call date of the refunded bonds. During this period both the refunded bonds and the advance refunding bonds are outstanding. Thus, the issuer is paying interest on two sets of bonds at the same time, offset only by the interest earnings on the refunding escrow, which for many years now have been well below the interest cost of the refunding bonds. In other words, the investment of the advance refunding bond proceeds in the defeasance escrow results in negative arbitrage. To proceed with a tax-exempt advance refunding, interest rates must have dropped sufficiently so that, even taking into account the fact that for a period of time the issuer will be paying interest on two sets of tax-exempt bonds (at least to the extent of the negative arbitrage), the issuer will still save a sufficient amount of money on a present value basis.
After the refunded bonds are redeemed, the net effect of a high-to-low advance refunding to the U.S. Treasury is positive. This is because the lower yield on the tax-exempt advance refunding bonds means that less tax-exempt interest is paid and the U.S. Treasury loses less money in foregone income tax on the interest. To illustrate the point further, suppose a person (we’ll call him “Wesley”) owned a home and his lender suggested that he refinance the mortgage secured by the home because the prevailing interest rates are lower than the rate that Wesley currently pays on his home mortgage. He would undoubtedly say “as you wish” and go forward with the refinancing. This is a fairly common scenario with which any homeowner is familiar.
However, what if instead the lender told Wesley that he couldn’t pay off the prior mortgage loan until the 10th anniversary of the mortgage loan (“Inconceivable!”)? Wesley might be tempted to draw his sword and do battle, but it would still make sense first to compare the interest paid over the life of his current mortgage loan and the refinancing loan that his lender proposed to see whether the “double interest” that he would pay on both mortgages during the overlap period, net of his paltry escrow investment earnings, would on a present-value basis eat up the savings from a lower interest rate on the refinancing loan
The refunding makes sense if the borrower pays less in interest on a present value basis over the life of the loan. The same is true of high to low advance refunding bonds. Importantly, the interests that motivate State and local issuers to issue advance refunding bonds are aligned with those of the Federal government – to minimize interest expense!
A high-to-low tax-exempt refunding, and the attendant benefit to the federal treasury, would arise only if the decline in prevailing market interest rates continues to the point in time at which the tax-exempt bonds can be currently refunded. In the end, the inability to lock in lower market rate of interest through a tax-exempt advance refunding should cast doubt on the proposed savings projected by the Joint Committee on Taxation that would result from the elimination of tax-exempt advance refunding bonds.
[1] For those of you who may be newer to our blog, an “advance refunding bond” is a refunding bond that is issued more than 90 days before the date that the bond that it will refund can be redeemed. Only governmental units and 501(c)(3) organizations can benefit from tax-exempt advance refunding bonds.