The “reasonable expectations” approach to determining the issue price of a tax-advantaged bond[1] has been the law since 1989. On June 7, it is scheduled to join Betamax tapes and parachute pants as another relic of that bygone decade. Barring intervention (either Divine or as part of the President’s executive order to undo recent regulations that “add undue complexity to the Federal tax laws”), the new issue price regulations will take effect for tax-advantaged bonds sold on or after June 7. Though we don’t often have to rely on reasonable expectations because underwriters usually actually sell at least 10% of each bond maturity at the initial offering price to the public on the sale date, the reasonable expectations rule has been a useful tool and a dear friend. As it prepares to ride off into the sunset,[2] a eulogy is in order. And bittersweet that eulogy shall be, for the death of the reasonable expectations standard seems senseless.

What began as a regulatory effort that, for all its faults, was at least focused around a goal, has devolved through the comment process into regular old horse-trading.

The relevance of issue price for arbitrage purposes

 The main reason – maybe the only reason – that we care about the issue price of a tax-exempt bond issue for arbitrage purposes is that it sets the target for the bond yield calculation, which in turn sets the limit on the investment return that an issuer can earn on its investment of unspent bond proceeds.

We should pause to realize that this is a policy choice – of course it is – and not the inescapable result of mathematical logic. The arbitrage rules intend to prevent an issuer from earning more on its investment of the money that it borrows than it pays out in interest to the bondholder. Conceptually, then, the restriction on an issuer’s investment of bond proceeds should be a comparison of its cost of borrowing the money to the return that it earns on the money that it borrowed before it spends that borrowed money. But Congress took the opposite approach legislatively in the Tax Reform Act of 1986, avenging the IRS’s bitter defeat in the State of Washington case.[3] Congress reasoned that because this approach to calculating yield would allow issuers to recover their costs of issuance through arbitrage profit (by subtracting them from the target amount for the arbitrage yield calculation),[4] the issuer’s investment limit should be set with reference to the issue price, unadjusted for costs of issuance. So Congress gave us Section 148(h) of the Code, which says that yield is determined “on the basis of the issue price (within the meaning of sections 1273 and 1274).”

The history of the reasonable expectations rule for setting issue price

The rules in sections 1273 and 1274 of the Code tell us how to calculate the amount of original issue discount for all debt instruments. But that topic is most relevant to debt that bears taxable interest, because the OID must be included in gross income as it accrues in each year that the lender holds the debt. These general OID rules say nothing at all about reasonable expectations – the issue price of a debt instrument under these rules is based on actual sales.

Yet when the IRS released the first set of arbitrage regulations after the 1986 Act,[5] they allowed an issuer to use reasonable expectations to set the issue price of a publicly offered bond. These regulations didn’t explain why we got this helpful rule, but a later amendment[6] explained that Treasury would allow an issuer to use reasonable expectations “because, on the date of issue, the exact price at which the bonds subsequently will be sold to the general public may not be known.” Left unsaid is that an issuer needs to know the issue price prior to the issue date, so that it can size its refunding escrows and make sure that it has complied with all of the tax-exempt bond rules that depend on the issue price of a bond issue, and more fundamentally, so that bond counsel can give its unqualified opinion. These concerns are not present in a taxable debt issuance, which is the main topic of concern of the OID regulations. This accommodation is consistent with the approach taken generally in the arbitrage regulations, where an issuer can establish its compliance based on its reasonable expectations.

In other words, the unique circumstances of the tax-exempt bond world justified a difference between these regulations and the general OID rules. These regulations finally settled into place in 1993, and there haven’t been any changes to the issue price rules since then.

What problems are the new issue price regulations trying to solve?

A few years ago, Treasury and IRS officials began making public comments that the issue price rules “weren’t working,” and needed to be changed. Treasury and the IRS expressed concern in the Preamble to the 2013 proposed issue price regulations “that certain aspects of the Existing Regulations for determining the issue price of tax-exempt bonds are no longer appropriate in light of market developments since those regulations were published.” The problem was “that the ten-percent test does not always produce a representative price for the bonds.” The Preamble went on to say that the IRS had observed underwriters deliberately selling the first 10% of a maturity at an artificially low price, which set the issue price for the issuer, and then selling the remainder at a higher price.

If 10% isn’t enough to produce a “representative price,” the 2013 proposed regulations reasoned, then let’s use 25% instead. This is a solution that would have at least responded to the perceived problem. But after all of the tomato-throwing about the 2013 proposed regulations, the IRS changed this aspect of the rule. So we’re now back to 10% again, even though the concern was that sales – actual sales – of 10% are not enough to establish issue price.

In connection with this rationale, Treasury attempted to lump in the reasonable expectations test as another of the existing rules that was enabling bad behavior by underwriters. The Preamble to the 2013 proposed regulations stated that “increasing transparency about pricing information” (i.e., EMMA) “has led to heightened scrutiny of issue price standards.” (This is an odd, somewhat circular phrasing – the words “led to” attempt to transform a decision by Treasury and the IRS to change the issue price rules into some inexorable factual result occurring out in the wild. They were the only ones doing the “heightened scrutiny.”) The Preamble went on to say that “[t]he reported data has [sic] shown, in certain instances, actual sales to the public at prices that differed significantly from the issue price used by the issuer. These price differences have raised questions about the ability of the reasonable expectations standard to produce a representative issue price.”

The reasonable expectations test only matters, though, in cases where less than 10% of a particular maturity is actually sold on the sale date. In that situation, we can typically infer that the initial offering price of the bonds was too high, not too low, meaning that the initial offering price reflected a bond yield (and permitted investment yield for the issuer) that was too low, so that the representative price/yield and permitted investment yield for the issuer would be higher than the pricing reflected. This is the opposite of the concern that seemed to sit at the heart of the 2013 proposed regulations.

The most charitable explanation for dropping the reasonable expectations test is that it brings the issue price rules for tax-exempt bonds back into consistency with the issue price rules for all debt in sections 1273 and 1274 of the Code. But it does so notwithstanding the fact that the inconsistency seemed fine when the regulations were put out in 1989 and no one has seemed concerned since, and the fact that there are very good reasons – reasons that Treasury and the IRS have acknowledged in published guidance – for allowing this small departure from the general OID rules to provide certainty to tax-exempt bond issuers. This is particularly so when the arbitrage rules generally allow an issuer to rely in the first instance on its reasonable expectations as of the issuance date.

One cannot help but wince at the irony, then, that the new issue price regulations have justified themselves as providing greater “certainty.” And one also cannot help but compare the sense of administrative restraint and the desire to hew closely to Congress’s statutory text shown in these final regulations with the opposite approach in the other regulatory project that appeared side-by-side with the issue price regulations – the seemingly doomed[7] attempt to upend the definition of a “political subdivision” for tax-advantaged bond purposes.

In addition to the concerns about manipulations of the issue price, Treasury and the IRS were also concerned under the prior regulations that underwriters were treating sales of bonds to other “underwriters” as sales to the public. In other words, either the existing definition of underwriter was too narrow to catch all of the participants in the sale of bonds that might properly be considered underwriters, or it was being abused by underwriters. And yet now we find ourselves on the eve of these new issue price regulations with a rule that we are told by Treasury and the IRS should be celebrated because it has further narrowed the definition of an underwriter to those entities that are in the contractual chain with the issuer in connection with the initial sale of the bonds to the public.

In a case where a rogue underwriter engaged in the bad behavior described above, whether it be selling the first 10% of an issue of bonds at an artificially low price or misrepresenting the status of the party to whom it sold the bonds, the underwriter would nevertheless provide the issuer and bond counsel with a certificate stating or implying that it did neither of these things. The problem under the old regulations, then, was that the behavior of the parties conflicted with the statements in the certificate. We would expect that a regulatory project intended to fix this would firm up the certificate practice. And indeed the 2013 proposed regulations did that. But then came more tomato-throwing, and now we’re back to final regulations that continue to allow issuers and bond counsel to rely on the same sorts of certifications without more (other than a requirement to attach the pricing wire in certain cases).

The thread running through all of the concerns above is the perception by the IRS that underwriters were behaving badly and the potential unreliability of the underwriter’s certifications to the issuer. The preamble to the final issue price regulations notes that the IRS may impose penalties under Section 6700 on underwriters whose certificates contain false information. As Brother Cardall once described the state of play on the other aforementioned regulatory project on political subdivisions, aren’t these issue price regulations also the right idea, but the wrong regulation? A glance at the Treasury Regulations shows that there isn’t yet a Reg. § 1.6700-1. Maybe there should’ve been.

*           *           *

To be sure, these final regulations are light-years to the good side of the 2013 proposal. Perhaps we should all be thankful.

But the final regulations do not seem to respond to the concerns that started all of this. It was said that sales – even actual sales – of 10% of a maturity did not produce a representative price. But yet the 10% threshold remains. It was said that the definition of underwriter was either too narrow or it was being manipulated. But yet the definition of underwriter has been narrowed even more. It was said that the manipulations of underwriters were causing their certifications to be unreliable. But yet issuers can still rely on those certifications, with the same due diligence standard that has always applied to the arbitrage rules.

What began as a regulatory effort that, for all its faults, was at least focused around a goal, has devolved through the comment process into regular old horse-trading. Maybe there’s something to be said for bringing these rules into line with the OID rules, but since the reasonable expectations approach appeared in 1989, no one (certainly not Congress itself) has seemed too bothered by it.

There may still be some chance that the Administration makes a last-minute decision to stall and review these regulations as part of the recent executive order. (Anyone who has seen the flurry of model documents and webinars and teleconferences among our industry groups would tell you that it looked an awful lot like the “complicated forms and frustration,” mentioned in the order. And the repeated exhortations by these groups that the new rules will require significantly more tax lawyer involvement from the very beginning of a transaction cannot be described as anything other than an “undue burden.”) But, nevertheless, a delay looks less likely as each new day passes.

And if no delay comes, then the reasonable expectations method for setting issue price is toast. What began as a reasonable accommodation to reflect differences between the tax-exempt and taxable debt markets now represents an inconsistency and an invitation to mischief that can no longer be tolerated.

Good-night, sweet prince, and flights of angels sing thee to thy rest.




*Thanks as always to our retired partner Jack Browning for his thoughts and contributions to this post.

[1] For arbitrage purposes only, of course; the final regulations leave us all just as much in the dark as we have always been about whether the rule applies for any other tax-exempt bond purposes.

[2] Mercifully not the sort of agonizing “sunset” that applies to the $150 million nonhospital bond limitation that applies to qualified 501(c)(3) bonds.

[3] There, the Court of Appeals for the D.C. Circuit held that “yield” should be calculated with reference to the issuer’s cost of borrowing.

[4] E.g., H.R. Rept. 99-426, p. 517-18.

[5] T.D. 8252 (May 15, 1989). These regulations were largely concerned with implementing – poorly – the new rebate regulations that were part of the new rebate requirement applicable to all tax-exempt bonds in the 1986 Act.

[6] T.D. 8345 (Apr. 19, 1991).

[7] This is what § 150 of the Code defines as “whistling past the graveyard.”