Re Rev. Proc. 96-41
Dec 17, 1996
Arbitrage and Rebate
Committee Comments on Rev. Proc. 96-41
Editor's Note: The
following comment letter on Rev. Proc. 96-41 and Notice 96-49 was
submitted to the IRS by the Association's Arbitrage and Rebate Committee
on December 17, 1996.
Internal Revenue
Service
CC:DOM:CORP:R (Notice 96-49)
Room 5226
POB 7604
Ben Franklin Station
Washington, D.C. 20044
Gentlemen and
Ladies:
We are writing on
behalf of the Arbitrage and Rebate Committee of the National Association
of Bond Lawyers ("NABL") to
comment on Rev. Proc. 96-41 (the "Rev.
Proc.") and Notice 96-49 (the "Notice") and to request to speak at the public
hearing to be held regarding the Rev. Proc. and the Notice.
The National
Association of Bond Lawyers was incorporated in 1979 for the purpose of
educating its members and others in the law relating to municipal bonds,
providing a forum for exchange of ideas as to law and practice,
improving the understanding and knowledge in this field of law,
providing advice and comment at the federal, state and local levels with
respect to existing or proposed legislation, regulations, rulings and
other action affecting state and municipal obligations, and providing
advice and comment in proceedings before courts and administrative
bodies through briefs and memoranda as a friend of the court or agency.
NABL has over 3,000 members who actively participate in some aspect of
public or private municipal finance practice. They are bond counsel,
underwriter's counsel, municipal attorneys, issuer's counsel, corporate
counsel, trustee's counsel, defense counsel in tax and
securities-related proceedings, and others involved in municipal
finance. The members of NABL who contributed to the preparation of these
comments are: Jeannette M. Bond, David A. Walton, Jeremy A. Spector,
David A. Caprera, Arthur M. Miller, Margaret C. Henry, William H.
Conner, Julianna Ebert, Amy K. Dunbar, and Mitchell H.
Rapaport.
INTRODUCTION
We appreciate the
opportunity to comment on the Rev. Proc. and commend the IRS for the
prompt issuance of the Notice and elimination of the one-year deadline
in the Rev. Proc. as originally issued. However, we continue to believe
that the Rev. Proc. and Notice create a situation which results in the
IRS coercing issuers into paying to the IRS the difference between the
so-called "spot price" for their escrow securities and the price
actually paid. By "offering" issuers a declared maximum penalty of the
difference between "spot price" and the price paid, the IRS' amnesty
forces issuers to evaluate their transactions based upon the conclusions
in the Rev. Proc. regarding pricing and the definition of fair market
value. Issuers who do not take advantage of the Rev. Proc. and pay the
penalty run the risk that an IRS agent will audit their transactions,
resulting in a greater penalty or even the loss of tax-exempt status.
This is a real risk, particularly since we understand that IRS National
Office personnel recently provided training to agents regarding
overcharging for escrow securities. Administrative convenience does not
justify this coercive approach.
As discussed below,
issuers are in an impossible situation as a result of the Rev. Proc.
With at most a few exceptions, issuers have no idea whether they
overpaid for their open market escrows and no practical way to determine
whether or not they did. Bond counsel are in no position to evaluate the
matter, and investment bankers generally shy away from second-guessing
the pricings of others on a retrospective basis, given the subjective
nature of such pricing considerations.
Given the huge volume
of advance refunding transactions during the period from 1991 through
1994, when open market securities were commonly used in escrows, the
Rev. Proc. will affect a tremendous number of issuers. The claim that
the Rev. Proc. would only "affect" about 100 issuers is far from
accurate since every issuer who used open markets must evaluate its
situation under the Rev. Proc. As a result of the Rev. Proc., issuers
who did not know or have reason to know of potential overpayments are
being (or will be) forced to spend time and money attempting to
self-audit their transactions. All of this seems unnecessary and unfair.
The Rev. Proc. should either be revoked or revised so that it only
includes the IRS' offer to resolve concerns about overcharging for
investments through the closing agreement procedure for issuers that
voluntarily come forward. Further, either the Rev. Proc. or another
appropriate document should instruct IRS agents not to assert penalties
against issuers (or the holders of their bonds) who did not know or have
reason to know of overcharging for investments, provided that those
issuers cooperate with any IRS or SEC investigations of the pricing of
open market securities. Alternatively, with the issuer's consent, cases
should be referred to the SEC, an agency that may be more able to police
abuses in pricing matters and which has the power to obtain disgorgement
of overcharges from securities providers. Where, however, an issuer knew
or had reason to know it was being overcharged for its investments, the
IRS should be free to pursue a recovery from the issuer.
GENERAL
COMMENTS
(1)
Retroactivity. The Rev. Proc. is an attempt by the
IRS to impose rules retroactively against issuers where no such rules
previously existed. Under the Rev. Proc., unless an issuer purchasing
open market securities used the bidding process described in the Rev.
Proc. or managed to pay no more than "spot price" (an extremely unlikely
possibility) the Rev. Proc. unmistakably asserts that the issuer
overpaid for those securities and may have caused its bonds to be
arbitrage bonds. We believe that this retroactive approach is
unwarranted and runs counter to the spirit of cooperation that we
believe should exist between state and local governments and the federal
government. Moreover, by imposing these rules retroactively, the IRS has
effectively turned much of the proposed regulations relating to the
purchase of open market securities into regulations that are
retroactively effective.
(2) Disregard for
past comments and industry practice. The Rev.
Proc. and the Notice have the effect of retroactively prohibiting
industry practices that have existed and been relied on by issuers for
many years. These industry practices developed to deal with a SLGS
program that, until recently, was often not a viable investment
alternative. Until relatively recently, it was common for issuers to
purchase open market securities on a negotiated basis and rely on
certifications made by the providers of those securities. The IRS has
every right to look at that practice and conclude that more restrictive
prospective rules are needed. It is unwarranted, however, to use
hindsight to outlaw retroactively the reliance on practices the IRS and
Treasury were long aware of and never objected to until now. Not only
were the IRS and Treasury aware of these practices, they were implicitly
approved by the IRS in 1992 and 1993. In the proposed arbitrage
regulations issued in 1992, the IRS requested comments on whether
specific rules were needed for the purchase of open market securities.
The IRS received a number of comments to the effect that such rules were
not needed. Some of these comments, including those of NABL and the
American Bar Association, described industry practices in detail. After
having received and considered such comments, the final regulations
contained no new rules or safe harbors for purchases of open market
securities.1 Not
surprisingly, this was widely viewed as an IRS approval of these
industry practices. In light of this, it is unwarranted and unfair for
the IRS three years later to in effect retroactively impose bidding
rules on the entire industry. If the IRS believes it made a mistake in
not imposing bidding rules in 1993, it must accept some responsibility
for the situation, not try to shift it to innocent issuers. What it
should not now do is impose a set of rules that have the effect of
forcing every issuer that did not foresee that such rules would apply to
pay--in fact, overpay--significant amounts to the IRS or risk loss of
tax-exemption.
(3) Strong-arm
approach. The Rev. Proc. and the Notice are a
clear attempt by the IRS to coerce issuers into paying the so-called
"differential" amount. This approach unfairly attacks those who did not
benefit from the overpricing. Although the Notice removed the one-year
deadline on the availability of the IRS' settlement offer, issuers are
still faced with a choice: pay the differential amount or run the risk
that an IRS examining agent will audit its transaction. If a transaction
is audited, the IRS agent will come armed with the legal assumptions in
the Rev. Proc. and assert, based on the statements in the "background"
section, that the issuer paid too much (e.g., because it paid for remote
contingencies, etc.), relied on nonconclusive certifications, and,
therefore violated the arbitrage rules. In this situation, the Rev.
Proc. states that the result upon audit "could be different from the
terms of" the Rev. Proc. The clear intent is that those who do not
settle up on their own may have their bonds taxed or, if the agent is
generous, have to pay "lost taxes," an amount likely to be far greater
than the differential. The Rev. Proc. applies this strong-arm approach
to every issuer in the country. We believe this is the wrong way for the
IRS to deal with municipalities with which it believes it has, as
repeatedly stated by IRS staff, a "special relationship."
A fairer approach would
permit an issuer to provide the relevant facts to the IRS so that the
IRS or the SEC could pursue the wrongdoers if it wished to do so. By
this method, issuers can avoid the extensive administrative and
litigation costs that will be involved in any attempt to recover damages
from the investment providers.
(4) Yield restriction and reasonable expectations.
Under Section 148(c) of the Code and '1.148-2 of the regulations, a bond
is an arbitrage bond if the proceeds are reasonably expected to be used
to acquire higher yielding investments (or to replace funds so used) or
if the issuer intentionally uses any proceeds in a prohibited manner.
Thus, compliance with the arbitrage yield restriction rules requires
that the issuer have the proper reasonable expectations on the issue
date and that it thereafter did not take any intentional act to violate
the arbitrage rules.
In the context of the
Rev. Proc., did an issuer that purchased open market securities at a
price that is subsequently determined to have exceeded fair market value
lack the proper reasonable expectation or commit an intentional act to
earn arbitrage? The regulations contain the following definition of
"reasonable expectation:"
An issuer's
expectations or actions are reasonable only if a prudent person in the
same circumstances as the issuer would have those same expectations or
take those same actions, based on all the objective facts and
circumstances. Factors relevant to a determination of reasonableness
include the issuer's history of conduct concerning stated expectations
made in connection with the issuance of obligations, the level of
inquiry by the issuer into factual matters, and the existence of
covenants, enforceable by bondholders, that require implementation of
specific expectations. For a conduit financing issue, factors relevant
to a determination of reasonableness include the reasonable expectations
of the conduit borrower, but only if, under the circumstances, it is
reasonable and prudent for the issuer to rely on those
expectations.
'1.148-1. Thus, in
order to satisfy the reasonable expectations requirement, an issuer's
actions in purchasing open market securities would have to be consistent
with those of a "prudent person in the same circumstances." Issuers
satisfied this requirement in several ways, including the issuer's own
review and diligence, review by a financial advisor, review by another
broker-dealer, and certifications provided by the seller of the
securities. We understand that only the reliance on the certification of
the seller is being questioned by the Rev. Proc. We believe that
reliance on the certification of the seller of the securities satisfies
the reasonable expectations standard. First, reliance on such
certifications was a common practice in the industry for many years and,
therefore, similarly situated persons were "taking those same
actions."
Second, issuers were
generally advised by their bond counsel that reliance on such
certifications was appropriate. Third, the legislative history to the
amendments to section 6700 states that reliance on certifications of
experts is appropriate in connection with tax-exempt
financings:
bond counsel, issuer's
counsel, and underwriters' counsel would be entitled to rely upon a
feasibility study conducted by an engineering firm reputed to be an
expert in the subject matter and area of study, unless such counsel
independently knew or had reason to know information bringing into
question the results of that study. Absent that, counsel would not be
required to question the assumptions underlying, or results reached by,
the study. Similarly, bond counsel would be able to rely, as to matters
of fact or expectation relevant to his or her opinion, on information
provided by other parties (including the issuer) absent actual knowledge
or a reason to know of its inaccuracy or the use of statements not
credible or reasonable on their face.
Omnibus Budget
Reconciliation Act of 1989, House Report, pp. 1398. These statements
strongly support the notion that issuers acted reasonably in relying on
statements of investment providers where they did not know or have
reason to know of information that would cause them to question those
certifications and that issuers need not have questioned the assumptions
or conclusions of those certifications.
Fourth, the IRS
specifically reviewed the long-standing practice of relying on provider
certifications in connection with the finalization of the arbitrage
regulations and provided no indication whatsoever that this was not an
acceptable practice. Finally, the certification itself was provided by a
person with whom the issuer probably had a previous relationship and
whose statements the issuer had no reason to believe might be
false.2 If such a
person overcharged for the securities or lied in the certification, it
would have violated various securities, tax, and criminal laws, and
violated NASD procedures. Given all of these factors, we believe a
prudent person would reasonably have relied on such a certification. We
also note that Black's Law Dictionary defines "prudent" as judicious,
careful, circumspect, sensible; synonymous with cautious. Similarly,
"prudence" is commonly associated with care and diligence and is
contrasted with negligence. We believe that reliance by issuers on
certifications satisfied this standard and that, in doing so, issuers
complied with the requirement that they reasonably expected that they
would make only fair market value purchases.
Similarly, we do not
see how an issuer that did not know or have reason to know that it
overpaid for securities has taken an "intentional act" to violate the
arbitrage rules. With respect to the definition of intentional acts, the
regulations contain the following statement:
The taking of any
deliberate, intentional action by the issuer or person acting on its
behalf after the issue date in order to earn arbitrage causes the bonds
of the issue to be arbitrage bonds if that action, had it been expected
on the issue date, would have caused the bonds to be arbitrage bonds. An
intent to violate the requirements of section 148 is not necessary for
an action to be intentional.
'1.148-2(c). To have
intentionally overpaid for open market securities, an issuer must have
known that it was doing so.
Based on this analysis,
we believe that an issuer who did not know or have reason to know that
it was overpaying for open market securities and who relied on a
certification from the provider of the securities did not violate the
arbitrage yield restriction rules as a result of the purchase of those
securities.
(5) Arbitrage rebate
consequences. Given that there has been no
violation of the yield restriction rules by an innocent issuer, the only
other potential arbitrage violation would be with respect to the
arbitrage rebate rules. Under those rules, rebate must be computed based
on the fair market value of the investments purchased and an issuer who
overpaid for its investments may owe rebate based on the difference
between fair market value and the price actually paid.
Can the bonds of an
issuer who unknowingly overpaid for open market securities but who
relied on a provider's certification be declared taxable as a result of
the failure to pay rebate in a timely manner? Section 1.148-3(h) of the
regulations provides as follows:
(1) In
general. The failure to pay the correct rebate
amount when required will cause the bonds of the issue to be arbitrage
bonds, unless the Commissioner determines that the failure was not
caused by willful neglect and the issuer promptly pays a penalty to the
United States. If no bond of the issue is a private activity bond (other
than a qualified 501(c)(3) bond), the penalty equals 50 percent of the
rebate amount not paid when required to be paid, plus interest on that
amount. Otherwise, the penalty equals 100 percent of the rebate amount
not paid when required to be paid, plus interest on that
amount.
(2) Interest on
underpayments. Interest accrues at the
underpayment rate under section 6621, beginning on the date the correct
rebate amount is due and ending on the date 10 days before it is
paid.
(3) Waivers of the
penalty. The penalty is automatically waived if
the rebate amount that the issuer failed to pay plus interest is paid
within 180 days after discovery of the failure, unless, the Commissioner
determines that the failure was due to willful neglect, or the issue is
under examination by the Commissioner at any time during the period
beginning on the date the failure first occurred and ending on the date
90 days after the receipt of the rebate amount. For purposes of section
1.148-3(h)(3), willful neglect does not include a failure that is
attributable solely to the permissible retroactive selection of a short
first bond year if the rebate amount that the issuer failed to pay is
paid within 60 days of the selection of that bond year. Generally,
extensions of this 180-day period and waivers of the penalty in other
cases will be granted by the Commissioner only in unusual
circumstances.
We believe that issuers
that acted "prudently" could not be guilty of willful neglect and,
therefore, cannot have taxable arbitrage bonds under the rebate rules.
Further, the worst case scenario for these issues is that they owe the
unpaid rebate plus interest plus a 50 percent penalty. Under the
regulations, this penalty is automatically waived if the rebate amount
plus interest is paid within 180 days "after the discovery of the
failure." We do not believe that issuers should be considered to have
discovered a failure merely because the IRS asserts questionable pricing
theories in the "background" in a Rev. Proc. This would force every
issuer to self-audit each of its transactions. Further, by the time the
IRS decides what to do with the Rev. Proc., the 180 days will likely
have run.
Given all this, it is
in fact questionable whether the Rev. Proc. (even ignoring all its
technical problems) provides anything that is not already available to
issuers under the regulations, since the rebate provisions permit
issuers to correct their errors.
(6) Ignoring
available penalties. By issuing the Rev. Proc. and
the Notice and holding issuers accountable under what is in effect a
strict liability standard for failure to purchase at fair market value,
the IRS seems to have deliberately decided not to impose the penalties
under section 6700 that Congress provided to the IRS to pursue the real
wrongdoers in municipal finance abuses. Although we disagree entirely
with the overall approach of the Rev. Proc., the IRS could just as
easily have directed the Rev. Proc. at providers of open market
securities who signed fair market value certifications that were false
or incorrect. Those providers could have been informed that, unless they
paid the differential, they would run the risk of having to pay a $1,000
per bond penalty. Congress intentionally revised section 6700 in 1989 so
that the IRS could impose significant penalties on the wrongdoers in
municipal finance transactions. To ignore or reject this available
penalty and focus on issuers is the wrong approach. The IRS seems to
have chosen to go after the easiest party to attack: issuers of
outstanding bonds with limited resources and a limited ability to fight
back.
In 1989, Congress
conducted a general re-examination of the Internal Revenue Code's
penalty provisions. A group of administration and Congressional staff,
including staff of the Ways and Means Committee, then-Congressman Beryl
Anthony, and the IRS, together with representatives of NABL,
successfully advocated the modification and expansion of section 6700 so
that enforcement efforts could be properly targeted at the "deserving"
participants in a tax-exempt bond transaction (rather than the
bondholders or issuer). We understand that there may be a concern at the
IRS that section 6700 may only provide a penalty of $1,000 per bond
issue. We believe that such a reading of section 6700 in light of the
legislative history is nonsensical. To read the amendments in this
manner leads to the conclusion that Congress acted in a wholly
impractical manner. It makes no sense, and certainly does not deter
abuse, to provide a $1,000 penalty in transactions that average in the
millions of dollars. Congress certainly is not so naive as to enact a
new penalty that would have absolutely no deterrent effect and that
would cost more to enforce than it would produce. Statutes are not to be
interpreted in such an illogical manner.3 Furthermore, the IRS specifically sought to
and did clarify confusion in the Circuits regarding the application of
section 6700 as it had been applied to tax shelters to make clear that
the penalty was not to be applied on a per transaction basis but on a
per share, per plan, or per denomination basis.
More specifically,
section 6700 focuses on a "plan or arrangement," which term was modified
in 1989 to include tax-exempt bonds. See Omnibus Budget Reconciliation
Act of 1989, House Report, pp. 1397-1399. The organization of a plan or
arrangement or the sale of each interest in a plan or arrangement is a
separate activity for purposes of applying the section 6700 penalty; the
penalty for each sale is $1,000. See Omnibus Budget Reconciliation Act
of 1989, Conf. Report, pp. 657-658. In the context of municipal bonds,
the issuance or sale of each bond is a "plan or arrangement." Neither
the statute nor the legislative history used the term "issue of bonds,"
clearly evidencing that it is not the "issue" that is the plan or
arrangement; it is each bond. Further, even if plan or arrangement
refers to the issue, the term "interest" must refer to each bond. The
clear purpose of the changes to section 6700 was to treat bonds like tax
shelters. To interpret "plan or arrangement" or "interest" to mean the
entire bond issue would clearly subvert this intent.
(7) The IRS should
identify and clearly state its goal. We suspect
that no matter how pervasive the IRS believes investment pricing abuses
may have been, it does not believe that issuers knowingly participated
in (or even should have known they were participating in) many of these
abuses.4 As the
Government Finance Officers Association stated in its comments on the
proposed regulations, "yield burning" was predominantly a sales abuse
visited upon unknowing and innocent state and local officials. If the
IRS has a different view of this matter, it should say so publicly since
that will make its policies and positions more
understandable.
If we are correct that
the IRS believes that investment pricing abuse was in most cases
something perpetrated on both the IRS and governmental issuers, we
believe that the IRS should take a fresh look at the Rev. Proc. The
first step in such a fresh look is for the IRS to identify what it is
trying to achieve. The Rev. Proc. seems to have as its goal the recovery
from state and local governments of something approximating the
difference between what issuers paid for their open market securities
and what they would have paid to purchase the same securities for next
day delivery (with no contingencies). We question the purpose of this
approach. It cannot be to punish the wrongdoer, since most issuers
neither knew of nor should have known that they were paying too much (if
that is, in fact, the case). Further, such issuers did not benefit from
the "wrongdoing," since they are out of pocket the amount of any
overpayment. Issuers cannot easily recover these overpayments from their
securities providers; it is likely that they will have to sue and prove
in court that the providers sold the securities at more than fair market
value. If all this is correct, most issuers settling up under the Rev.
Proc., despite having neither known of their "wrongdoing" nor benefitted
from it, will have to pay municipal funds to the IRS. The only goal this
approach fosters is to hold state and local governments absolutely
accountable for any improprieties involving their tax-exempt bonds
regardless of any knowledge or benefit. We know of no other areas in the
tax law that the IRS enforces in this manner or that may effectively
force a taxpayer to turn itself in and pay a penalty for a wrong
committed by another.
We believe that there
is no real benefit to the above-described approach, other than as a
revenue raising function. In the tax-exempt bond area a little
enforcement goes a long way. Further, in our view, the targeting of
issuers who knew (or should have known) of investment pricing abuse and
of the securities providers who benefitted will be effective to limit
the scope of this and other types of abuse in the future. In short, if
the intent of the drafters of the Rev. Proc. was to write a
legalistically perfect cure to overcharging for investments, without
regard to the fact that it may force them to settle by overpaying to the
IRS any possible overpayment for its investments, the Rev. Proc. has
succeeded. On the other hand, if the intent was to eliminate abuses, the
IRS needs to re-think its approach. When the IRS finalized the arbitrage
regulations, it indicated that it would concentrate its efforts on
effective enforcement, not on industry-wide rulemakings. Three years
later, the IRS should recognize that investment pricing abuses by
issuers are more the exception then the rule. The IRS should enforce and
regulate accordingly, not by issuing a Rev. Proc. that essentially
declares on a retroactive basis that all purchases of open market
securities other than on a bidding basis involved so-called "yield
burning." It is appropriate for the IRS to condition any non-enforcement
against issuers who neither knew nor should have known of investment
overcharging on their cooperation in the IRS' enforcement
efforts.
(8) Need to
distinguish wrongdoing. There may well be
situations where issuers overpaid for their open market securities and
knew (or should have known) that they were doing so. We believe that
such issuers do not represent more than a relatively small percentage of
the market. The far greater percentage is made up of issuers who
purchased securities at a price that reflected a mark-up for certain
contingencies and in reliance on certifications from the investment
providers to determine fair market value. These issuers did not know of
any overpricing, did not benefit from such overpricing, yet will bear a
significant financial cost if punished for allegedly overpaying for
securities. We believe that there is no need for the IRS to punish
innocent issuers merely because the IRS can easily do so, particularly
when there are other parties to whom the IRS can appropriately direct
its enforcement activities.
(9) Failure to suspend the Rev. Proc. While the
Notice's elimination of the one-year deadline for entering into
settlement agreements was helpful, we strongly believe that the Rev.
Proc. should be suspended and that IRS agents should be instructed not
to raise fair market value issues or rely on the Rev. Proc. until the
policy and legal questions raised by the Rev. Proc. are finalized.
Despite the elimination of the one-year deadline, issuers remain at
serious risk that agents will examine their transactions and use the
Rev. Proc. to extract penalties significantly in excess of those
provided under the Rev. Proc. In addition, any inference intended by the
Notice that the IRS would hold enforcement actions in abeyance was
eliminated by the IRS' recent training of agents regarding overpayments
for escrow securities. As long as this threat exists, issuers are put in
a dangerous position. Further, as described above, the continued
applicability of the Rev. Proc. threatens to serve as notice to issuers
of a rebate violation for purposes of determining whether a rebate
penalty is owed. NABL continues to favor the withdrawal of the Rev.
Proc. Alternatively, the Rev. Proc. could be revised and reissued
without the substantive provisions contained in the background
section.
(10) The Rev. Proc.
eliminates the value of safe harbors. One of the
many disturbing aspects of the Rev. Proc. is what it means for issuers
who do not follow IRS safe harbors. Even if issuers had perfect
hindsight and knew that the bidding safe harbors for guaranteed
investment contracts applied equally by analogy to open market
securities, still those provisions are only safe harbors. For example,
an issuer who solicited three or more bids but only received two from
disinterested parties might determine that, despite a technical failure
to comply with the safe harbor, its procedure had been sufficient to
establish a market price purchase. The Rev. Proc., however, effectively
turns the safe harbors into explicit rules on a retroactive basis. Any
issuer who deviates from the safe harbors identified by the Rev. Proc.
is considered to have overpaid and must cure this noncompliance by
paying the differential to the IRS or risk loss of tax-exemption. If
this is the IRS' view of how safe harbors work, we suggest that safe
harbors are a trap and unnecessary (and "rebuttable presumptions" even
more so).
(11) Pricing practices and roles of other
regulators. We cannot comment on those portions of the Rev.
Proc. in which the IRS questions the propriety of mark-ups for
contingencies and delayed closings since, as bond counsel, we do not
profess to possess the expertise to judge the correctness of these
points. However, we believe that the IRS will find relatively consistent
mark-ups whether there was significant arbitrage or not. What is
necessary before imposing rules such as those contained in the Rev.
Proc. is that the IRS and the securities providers gather information on
the pricing of open market securities under these various scenarios and
evaluate the real forces involved in pricing decisions. In this context,
it seems totally unfair that the SEC and the NASD could investigate the
pricing of open market securities and determine that there is no
liability (that is, no overcharging for investments) yet issuers would
have been compelled by the Rev. Proc. to make payments to the
IRS.
Regardless of whether
the IRS views a customary market risk as remote and, therefore, not
worthy of compensation, if such compensation has been standard in
competitively bid transactions and transactions in which issuers
incurred significant negative arbitrage, that is a strong indication
that the IRS is wrong in its premises. Some have suggested that, despite
these facts, issuers still paid too much (that is, even where issuers
were in a significant negative arbitrage situation, they overpaid). It
is unwarranted for the IRS to retroactively impose its view as to how
much an issuer (or any person) should pay for a contingent, delayed
delivery of securities. The regulations and the Rev. Proc. define fair
market value as what a willing buyer would pay a willing seller in a
bona fide arm's length transaction. The Rev. Proc. questions valuations
where one party to the purchase lacked a financial incentive to obtain
the best price. Where there is significant negative arbitrage, the
issuer has every incentive to obtain the best price and is a willing
buyer in an arm's length transaction. That the IRS thinks the issuer
paid too much in such a transaction should be irrelevant since the
transaction is, by definition, at "fair market value." If issuers paid
the same prices (and paid for the same risks) in situations where there
is no negative arbitrage, those purchase prices equally satisfy the
definition of fair market value.
It is also clear to us
that the pricing of open market securities is a complex, fact-specific
process. The strongest evidence of this is that the National Association
of Securities Dealers ("NASD"), which requires that sales occur at a "fair
price," has a complex series of factors to take into account in
determining whether pricing is fair. Given this complexity, it is
unlikely that the IRS can appropriately conclude that issuers who paid
more than the so-called spot price paid too much.
(12) The purported
lack of a financial incentive. Both in the Rev.
Proc. and in its subsequent defense of the Rev. Proc., the IRS has made
much of the purported lack of a financial incentive to invest escrows at
market rates. We believe that this point has been greatly exaggerated
and that state and local officials responsible for these investments and
securities providers certifying to the pricing of the securities have
significant incentives in this regard. First, as the Rev. Proc. amply
demonstrates, issuers who overpay for open market securities risk having
their bonds declared taxable or paying significant penalties. Second,
the individuals who sign the issuer's arbitrage certificates and the
securities provider's fair market value certificates risk the section
6700 and other, more serious, penalties for tax and securities fraud.
Finally, state and local officials who pay too much for open market
securities risk violating local laws that prohibit paying more than fair
market value for anything as they carry out their fiduciary
responsibilities. Contrary to the IRS' apparent belief, these are
significant disincentives to intentionally or recklessly "burn"
yield.
(13) Inability to
challenge IRS determinations. Under current law,
when the IRS challenges the tax-exempt status of an outstanding issue of
bonds, as a practical matter, the issuer has few options available if it
disputes the IRS' conclusions. While the issuer can argue with the
agent's findings and request technical advice from the National Office,
once these options have been exhausted, the issuer has little practical
choice but to accede to the IRS' demands. Further, throughout any such
controversy, the issuer must do everything possible to make sure that
the IRS continues to be amenable to a closing agreement. Unlike other
"taxpayers," except in the rarest of circumstances, an issuer of
outstanding tax-exempt bonds has little ability to litigate the tax
issues that the IRS has raised. The reason for this is that, even
assuming that the issuer, rather than a holder, can litigate on its own,
the issuer will have to disclose to the market that its bonds are being
audited, that the issuer has chosen to take the IRS to court over the
issue, and that an adverse court decision could result in a loss of
tax-exempt status. This situation would likely result in an immediate
reduction in the price of the issuer's bonds, lawsuits against the
issuer, and potential difficulties in issuing bonds in the future. In
the case of variable rate bonds, this would result in an immediate
effect on interest rates.
In the context of the
Rev. Proc., an issuer that is audited by the IRS is already out of
realistic options; the Rev. Proc. sets forth the National Office
position and, therefore, the issuer's only choices are to pay up or
litigate. Because of these circumstances, we believe that the IRS needs
to be very careful in implementing its tax-exempt bond audit program.
For the most part, in public discussions of the audit program, the IRS
has generally indicated that it is proceeding cautiously in this area.
The Rev. Proc. and the Notice, however, are wholly inconsistent with a
careful, considered approach. For this reason, we believe the Rev. Proc.
and the approach of the Notice should be reconsidered.
(14) Use of Rev. Proc. to make law. We also believe
that it is inappropriate for the IRS to use the background section of a
Rev. Proc. to attempt to make law, particularly in a retroactive manner.
Outside of the Rev. Proc. and the Notice, the IRS has been justifiably
cautious in issuing retroactive guidance in the tax-exempt bond area.
This makes it all the more inappropriate for the IRS to use a section
labeled as "background" to state its views regarding the law. In fact,
were it not for the background section, it is likely that the municipal
bond industry would have viewed the Rev. Proc. as a positive development
-- an attempt to provide relief to issuers who knew or suspected that
they overpaid for their open market securities. Moreover, much of the
background section of the Rev. Proc. contains inappropriate statements
of factual rather than legal conclusions, such as those relating to the
proper pricing of contingent, forward delivery securities.
For example, Sections
2.15 and 2.16 of the Rev. Proc. claim that the simultaneous purchase of
open market securities and SLGS constitutes the receipt of a call
option, the value of which must be included in any settlement. Leaving
aside the accuracy of the typical factual situation as stated in the
Rev. Proc., this is a clear-cut example of substantive law and rules
being announced and made effective retroactively through a revenue
procedure. There are no case law, regulations, or rulings cited in
support of the IRS' position. The conclusions reached are wholly
inappropriate in the context of a revenue procedure. To the extent that
the IRS believes in the correctness of its approach, it should
promulgate it in the form of a revenue ruling, rather than through a
closing procedure, particularly one with the coercive elements of the
Rev. Proc.
(15) Failure to
obtain industry input. We believe that many of the
problems with the Rev. Proc. could have been avoided if the IRS had
attempted to obtain industry input before releasing a document in "final
form" and not subject to comment (at least until the industry uproar
occurred) and which is in effect to this day. There is only so much that
can be learned about the proper pricing of open market securities
without getting input from the securities industry. Given that so much
of the Rev. Proc. is based on the proper pricing of open market
securities, industry input should have been a vital part of the
process.
(16) Future Rev.
Procs. Although we have many concerns with the
Rev. Proc. as it applies to open market escrows, we are also concerned
that the IRS not view the Rev. Proc. as an available format for future
enforcement actions. In other words, many of the comments included
herein apply equally to any use of the format used in the Rev. Proc. to
"suggest" noncompliance and offer a purportedly reasonable settlement
offer. Prior to the Rev. Proc., the IRS has only used this approach in
rare, unusual, and extremely limited circumstances. This should continue
to be the case.
In the rare situations
where, in the past, the IRS has used the type of approach taken in the
Rev. Proc., it has always stated a clear standard that did not leave
issuers wondering where they stood. For example, both in Rev. Ruls.
80-328 and 82-101, the IRS issued revenue rulings clearly describing the
facts, the law as the IRS saw it, and its legal conclusion. In those
situations, issuers clearly knew if they had a problem and knew the
potential consequences of failing to make use of the IRS' closing
agreement offer.
The Rev. Proc. on the
other hand, provides little certainty. There is no statement in the Rev.
Proc. that tells issuers that they do or do not have a problem. For
example, the Rev. Proc. states that the risk of non-settlement and
provider certifications are "factors" to be considered in determining
whether an issuer paid in excess of fair market value. Issuers must ask
their counsel whether they have a problem and counsel must respond that
there is no problem if a market price was paid. An issuer who, based on
this advice, does not choose to settle under the procedure in the Rev.
Proc. will be uncertain as to what the IRS' position will be if the
issuer's bonds are challenged. It is possible to interpret the Rev.
Proc. as an attempt to force issuers into settling out of IRS concern
that it could not make its views hold up in court and that, in any
event, would tax the IRS' resources if it pursued a full legal
challenge. We believe that if the Rev. Proc. is revised, in fairness it
should provide certainty to issuers rather than the in terrorem rules it
currently provides.
(17) Equal treatment of issuers. The Rev. Proc.
indicates that the penalty under the Rev. Proc. (the differential) is
not necessarily available to issuers who do not voluntarily enter into a
settlement agreement. If this penalty is the right amount, we fail to
see why it should not apply equally to all issuers, including those who
do not self-audit and voluntarily settle with the IRS. We believe that
all "taxpayers" should be treated equally. We question this approach,
both in the context of the Rev. Proc. and future enforcement actions
that might take a similar approach. The IRS should not be establishing
post-closure self-audit rules and penalties especially through revenue
procedures.
(18) Ability to determine the differential. The Rev.
Proc. is also lacking in the practicality of its "remedy." Except for
the most sophisticated issuers, we suspect that issuers will not have
access to the information that would enable them to determine the
differential, assuming that they understood what was required. The
intra-day changes in the market for both spot and forward pricing are
such that we believe issuers will have great difficulty in making a
practical application of the Rev. Proc. As we understand it, experts can
review the same escrows on the same day and conclude that different
prices are nonetheless each "fair market value." So some adjustments
must be made to clarify and simplify the determination of the
"remedy."
(19) Spot price as
fair market value. Despite the clear statements to
the contrary by some at Treasury and the IRS, we continue to be
concerned that in applying the Rev. Proc., IRS auditors will equate spot
price with fair market value (or to state it another way, that issuers,
given the Rev. Proc., will not be able to "prove" fair market value and,
therefore, by default, spot price is the closest thing). We have similar
concerns about the presumption that a delay of more than 30 days between
pricing and settlement is "bad." We know of at least two major
metropolitan areas that, as a result of local law, must have pricing and
settlement separated by more than 30 days.
SPECIFIC
COMMENTS
(1) Two certificate
procedure. As stated throughout these comments, we
advocate the withdrawal of the Rev. Proc. and significant changes in any
new versions of the Rev. Proc. Until recently, some issuers used a two
certificate process in purchasing open market securities. Under this
procedure, both the provider of the securities and another expert (e.g.,
a financial advisor or a primary dealer) certified that the purchase was
made at a price not in excess of fair market value. Even if the IRS
makes no other changes in the Rev. Proc., this procedure should be
effective in establishing fair market value and the Rev. Proc. should be
revised to clarify that, like a bidding process, for purposes of
applying the fair market value rules retroactively, issuers who relied
on this two certificate procedure are presumed to have satisfied the tax
laws. This clarification is needed by issuers both in trying to decide
whether to avail themselves of the offer contained in the Rev. Proc. and
to ensure that zealous auditors will not use the statements in the Rev.
Proc. regarding certifications to challenge purchases using the two
certificate procedure.
(2) Negative
arbitrage. Similar to the two certificate
procedure, many issuers and their counsel relied on the fact that the
yield on the securities that they purchased was well below the bond
yield to determine that a market price was paid for the securities
(since even the IRS ought to concede that an issuer would not willingly
give away its own earnings to a securities provider). The interim final
arbitrage regulations issued in 1992 adopted this approach for
investment contracts. The Rev. Proc. should be revised to clarify that a
material amount of negative arbitrage (that is, at least 12.5 basis
points) establishes a fair market value purchase.
(3) Form
10001. Form 10001, which is required to be filed
to be eligible for a closing agreement, requires issuers to submit a
significant amount of documentation. Much of this documentation seems
designed to enable the IRS to determine whether the issuer's bonds might
have violated any other rules rather than to provide the information
necessary to pay the differential. Given that the form must be signed
under penalty of perjury, copies of the Form 8038 or 8038-G,
confirmations, and a statement explaining the determination of the spot
price and differential should be sufficient. Copies of the offering
document, escrow agreement, and arbitrage certificate are unnecessary
and will only make issuers more suspicious of the IRS' motivations in
issuing the Rev. Proc. and the benefit of a closing agreement
offer.
REQUEST FOR
COMMENTS
In the Notice, the IRS
requested specific comments on several questions. We wish to respond to
certain of these questions:
(i)
Appropriate time period for closing agreement requests. If the Rev. Proc. retains a legal background section that
suggests that a vast majority of open market securities were overpriced,
there is no significant reason to provide a deadline on availability of
the settlement offer, since issuers who do not take advantage of the
settlement offer will always be at risk. If, on the other hand, the Rev.
Proc. is revised to more effectively identify the limited class of
transactions in which issuers knew or should have known that improper
conduct was occurring, then a deadline is appropriate and a one-year
period could be satisfactory.
(ii) Determining spot price. In the
period since the release of the Rev. Proc., we have spent a significant
amount of time discussing "spot price" with issuers, financial advisors,
investment bankers, and others. While we are not expert regarding the
pricing of securities, we would like to pass on our understanding of the
issues relating to "spot price." We understand that it is extremely
difficult and costly to attempt to recreate "spot price" on a historical
basis as of the time that securities were purchased. Investment
providers tell us that here is also great uncertainty as to the meaning
of spot price: is it the interdealer price, the interdealer price plus
markup, the prevailing market price (institutional or retail?), the bid
price or the ask price? Does it matter if the dealer is a primary dealer
or not? What if there are no contemporaneous prices? The only easily
obtainable, published information that we are aware of is the closing
prices in The Wall Street Journal and similar publications. We
understand that, for a fee, there are also other sources of end of the
day prices, as well as sources for real time prices. We are told that
these sources of information have several significant defects: First,
the end of the day prices ignore intra-day price volatility which, we
understand can be significant. Second, these prices do not take into
account any of the peculiarities of contingent, delayed-delivery
purchases or open-ended offers. We are also told that the price quotes
are often based on estimates, and therefore are generally not reflective
of the securities purchased by issuers for escrows. Given these
problems, the published prices need to be adjusted to take these factors
into account. For example, to adjust for volatility and forward
delivery, the higher of the opening or closing prices for a day could be
adjusted upward by an arbitrary number of basis points simply to achieve
rough justice. We note that such a mechanism would still ensure that
many purchases that were significantly overpriced would result in a
payment to the IRS that issuers would have to make out of their own
pockets.
Even assuming that an
appropriate reference price can be determined, we understand that it is
still unclear how the "spot price" is to be determined, creating
further, unacceptable ambiguities. For example, does "spot price" allow
for the "extra" accrued interest to the actual day of settlement or does
it merely reflect the price, including accrued interest, for regular
settlement? A similar problem exists for securities that are traded and
quoted on a yield basis: does "spot price" mean the price as of next day
settlement or is it the price on the actual settlement day determined
based on the yield quoted for next day settlement? Common sense
indicates that these differences can be significant and result in
ambiguities that a "simple" closing agreement procedure should not
contain.
(iii) Other methods
to determine closing agreement amounts. We do not
have any specific suggestion for other closing agreement amounts. We do
wish to emphasize that no fixed amount or formula can be established for
all transactions. The IRS should provide flexibility to determine the
settlement amount for determination on a case-by-case basis. While this
will be more difficult for the IRS and issuers, it is the only fair way
to proceed. Further, a more open-ended Rev. Proc. might get more issuers
to volunteer to settle with the IRS. It does not make sense to have a
program that requires the same amount to be paid regardless of the
facts. Any such voluntary program would still be more effective than
reliance on the examination program.
We would be pleased to
discuss these matters further and provide any additional assistance. Any
questions should be directed to Mitchell H. Rapaport at (202)
457-5305.
Sincerely,
Mitchell H. Rapaport