Note: The following members of the Corporate Tax Committee Advisory Board participated in the preparation of these comments: Jasper L. Cummings Jr. and Mark R. Hoffenberg (primary drafters), Andrew M. Eisenberg (Chair), Lisa M. Zarlenga (Vice Chair), William Galanis, Stephen I. Flanagan, Karen G. Gilbreath, Bryan P. Collins, Joseph M. Pari, and Robert H. Wellen. The Corporate Tax Committee would like to thank Christine W. Booth for her assistance in the preparation of these comments. The views expressed in these comments are those of the Corporate Tax Committee of the Tax Section of the District of Columbia Bar Association and not those of the District of Columbia Bar or its Board of Governors.
Comments on Advance Notice of Proposed Rulemaking Concerning Assumption of Liabilities
- Introduction
- Default Rule
- Agreements
- Issues Involving Both Recourse and Nonrecourse Liabilities
- Coordination With Other Code Sections
I. Introduction
On May 6, 2003, the IRS published an advance notice of proposed rulemaking,
REG-100818-01 (the "Notice"), regarding "the amount of a liability
a transferee of property is treated as assuming in connection with a
transfer of property and certain tax consequences that result from the
transferee’s assumption of such a liability."1
The IRS issued the Notice for the purpose of inviting comments regarding
these issues. These Comments respond to that request on behalf of the
Corporate Tax Committee of the Section of Taxation of the District of
Columbia Bar Association.
The Notice considers three section 357(d) rules
governing nonrecourse liabilities and proposes modifications to each
of these. In simplified form, the three general rules the Notice proposes
to modify are: (1) a presumption that a transferee who receives assets
subject to a nonrecourse liability assumes the entire amount of that
liability (the "default rule"),2 (2)
the fact that the above presumption will apply only if the transferee
actually receives assets subject to the nonrecourse liability,3
and (3) a statutory reduction in the amount of the nonrecourse liability
presumed to have been assumed by the transferee (pursuant to the default
rule) equal to the amount of the liability that the transferor (or another
transferee) who has sufficient collateral has agreed and is expected
to pay.4
The modifications the Notice proposes are: (1) changing the presumption
from an assumption of all of the nonrecourse liability to an
assumption of a pro rata amount of the nonrecourse liability,
(2) eliminating the requirement that the transferee actually receive
assets that are subject to the nonrecourse liability the transferee
is treated as assuming, and (3) changing the rule governing the reduction
in the amount of the nonrecourse liability presumed to have been assumed
by the transferee (pursuant to the default rule), to allow the reduction
even if the transferor (or another transferee) who agrees to satisfy
all or part of the liability lacks sufficient (or indeed any) collateral
securing the liability (i.e., eliminating the limitation set forth in
section 357(d)(2)(B)).
As an initial matter, we believe that the statute’s
grant of regulatory authority (“except as provided in regulations”)
and the legislative history to section 357(d) support broad regulatory
action to modify the statutory rules. The legislative history to section
357(d) evinces a plain Congressional intention to conform the determination
of the amount of liability assumed as closely as possible to the underlying
economics of the transaction, and in our view, grants broad regulatory
authority to further that end: “The provision is intended to…better
reflect the economics of these corporate transfers.” *** “The
Treasury Department has authority to prescribe such regulations as may
be necessary to carry out the purposes of the provision.”5
Moreover, in the legislative history to a later, related provision,
Congress describes the statutory framework of section 357(d), and then
states, “Also, under section 357, regulations, if issued, may
provide for different results.”6
We therefore believe that the modifications proposed by the Notice (and
the changes to those modifications we propose below) are within the
authority that has been granted to the Treasury and the IRS, and are
appropriate.
The Comments that follow will address the questions
raised by the Notice by first considering the Notice’s underlying
question about the propriety of a default rule and the optimal formulation
of such a rule. That discussion will first conclude that a default rule
is appropriate and will then evaluate six alternative default rules,
each from three perspectives: (1) conformance with economic reality;
(2) consistency with the principles of Crane v. Commissioner,
331 U.S. 1 (1947), and Tufts v. Commissioner, 461 U.S. 300 (1983);
and (3) impact on the transferee. That discussion will conclude that
the Notice’s proposal to adopt a pro rata assumption approach
is the optimal formulation.
Our Comments will next consider whether, and
to what extent, agreements among the parties regarding satisfaction
of a nonrecourse liability should be respected. This discussion will
begin with a brief overview in which we conclude that agreements should
supplant a default rule to the greatest extent possible. Our Comments
will then address the second of the Notice’s proposed modifications
(eliminating the requirement that the transferee actually receive assets
subject to the nonrecourse loan) and conclude that this modification
is warranted. The discussion will then focus on the Notice’s inquiry
into the application of section 357(d)(2)(the reduction in the amount
presumed to have been assumed under the default rule). Because we believe
that agreements should supplant the default rule without regard to whether
the agreeing person owns collateral securing the liability, we do not
think that a reduction in the default amount for selected agreements
is necessary. Rather, we will evaluate various possible agreements (including
agreements by parties who may have reasons not to perform), and discuss
whether each such agreement should supplant an unreduced pro rata default
rule. This part of our discussion will conclude with a consideration
of the impact of agreements on the proper characterization of the debt
for purposes of subsequent transactions.
Finally, because our views pertaining to various
section 357(d) issues affecting both recourse and nonrecourse loans
and the coordination of the section 357(d) rules with other Code sections
derive from the analysis set forth in detail in the first part of our
Comments, we will set forth only briefly our recommendations with respect
to those matters.
Throughout these Comments, we have assumed that
all transfers of assets satisfy the requirements of section 351. More
specifically, we have assumed that the assets transferred in each case
have positive net value, even after taking account of the liabilities
assumed.
For example, if an asset worth $100 and encumbered
by a $200 nonrecourse liability is transferred by a shareholder to a
corporation that agrees to satisfy the liability (and no other assets
are transferred to the corporation), no net value would be transferred
to the corporation, and thus there arguably would be no room to construct
an issuance of stock, as would be required in a section 351 exchange.7
Similarly, if an asset worth $100 is transferred to a corporation that
assumes a $200 recourse liability, the transaction may be better described
as a net distribution by the corporation, rather than as a section 351
exchange.
Therefore, when our Comments describe a transfer
of an “underwater asset,” we have assumed that the transferor
also transfers additional property in the exchange. For example, if
an asset worth $100 and encumbered by a $200 nonrecourse liability is
transferred to a corporation, we have assumed that an additional asset
not encumbered by a liability and thus having net value also is transferred
as part of the exchange. If an unencumbered asset worth $30 is transferred
along with the encumbered asset described in the previous sentence,
there would be room for a $30 issuance of stock in the exchange, such
that the entire transaction would qualify as a section 351 transaction.
As noted below, the authors of these Comments
believe that principles similar to those described herein should also
apply to determine the amount of liabilities assumed in transactions
other than section 351 exchanges. However, the examples contained in
the Comments focus only on the specific operation of section 357.
II. Default Rule
A. Assuming there is no agreement as to the allocation of a nonrecourse
liability and a default rule is necessary, should a pro rata default
rule be adopted in place of the default rule requiring assumption by
the transferee of all nonrecourse debt collateralized by any transferred
property?
As explained in detail below, we recommend that
a default rule should be retained. In the absence of an agreement, we
recommend a pro rata default rule based on the relative value of the
retained and transferred collateral.
As in the case of a recourse liability, the
determination of the extent of assumption of a nonrecourse liability
would most closely track the real economics of a transaction if it were
based solely on the agreements and expectations of the parties. Unfortunately,
however, it would not appear possible always to rely solely on such
factual indicia, because it is our belief that there will be many circumstances
in which the parties to the transaction will not manifest (and will
not have) any agreements or expectations as to who will satisfy a nonrecourse
liability that is collateralized by assets held by more than one person.
Thus, we do not think it is advisable to dispense with a default rule
governing the determination of whether and to what extent nonrecourse
liabilities are treated as assumed. Nonetheless, as discussed below,
we recommend that the existence of “agreements and expectations”
should be broadly construed to include circumstances in which the relatively
contemporaneous actions of the parties to the transaction (including
payment of interest or principal on the debt) are sufficient to manifest
the bilateral intentions of the parties. The greater the extent to which
“agreements and expectations” do govern the determination
of the assumption of nonrecourse liabilities, the more likely the tax
treatment is to track the real economics of the transaction, and the
less pressure will rest on a default rule to attempt to accomplish that
objective.
Assuming, then, that there will remain a set
of transactions requiring a default rule, the government faces certain
choices, none of which will assure consistency between tax treatment
and the underlying economics of the transaction, and all of which have
certain collateral advantages and disadvantages in particular fact patterns.
We will consider six alternatives available to the government. These
are: (1) all nonrecourse liabilities retained by the transferor (the
“All to Transferor Rule”); (2) all nonrecourse liabilities
assumed by the transferee (the “All to Transferee Rule,”
which is the current statutory default rule); (3) nonrecourse liabilities
assumed to the extent of the fair market value of the transferred collateral
(the “Transferred FMV Rule”); (4) nonrecourse liabilities
assumed to the extent they exceed the fair market value of the retained
collateral (the “Retained FMV Rule”); (5) nonrecourse liabilities
assumed in proportion to the relative fair market values of the retained
and transferred collateral (the “Pro Rata FMV Rule”); or
(6) nonrecourse liabilities assumed in proportion to the relative initial
tax bases of the retained and transferred collateral (the “Pro
Rata Basis Rule). For purposes of the preceding sentence, and for reasons
that will be discussed more fully below, the term “initial”
tax basis (as used in the description of the Pro Rata Basis Rule) assumes
that the collateral was acquired in the transaction in which the debt
was issued and refers to the basis of such collateral in the hands of
the borrower at such time.
We will evaluate these alternatives from three
perspectives. First is a determination of which alternative is most
likely to match the assumption or retention of a liability to the party
that will ultimately satisfy the debt (conformance with economic reality).
Second, is a determination of which of the alternatives best preserves
the principles Crane and Tufts. Third is an evaluation
of the impact of each alternative on the transferee. We will conclude
that, although there may be circumstances in which the Pro Rata Basis
Rule would best serve the third of these goals, the Pro Rata FMV Rule
overall is the optimal rule. Although no default rule can reliably conform
to the actual subsequent repayments of liability, we believe that the
Pro Rata FMV Rule is more likely than any other to conform with such
repayments. Thus, we believe that the Pro Rata FMV Rule best conforms
with economic reality, and most appropriately times the recognition
of gain under Crane and Tufts.
(1) Conformance with economic reality
In the absence of agreements and expectations, any default rule is to
a certain extent arbitrary, in that there can be no assurance that any
alternative ultimately will succeed in matching the assumption or retention
to the ultimate payment of the liability. In other words, there can
be no assurance that any alternative will conform perfectly with economic
reality. We believe, however, that the Pro Rata FMV Rule has the greatest
potential to achieve or approximate that match in the greatest number
of cases. As illustrated below, in the case of an overcollateralized
liability, the All to Transferee Rule (the current statutory presumption)
and the Transferred FMV Rule would always allocate a disproportionately
large amount of the liability to the holder of the transferred collateral,
and the All to Transferor Rule and the Retained FMV Rule would always
allocate a disproportionately large amount of the liability to the holder
of the retained collateral. In addition, the Pro Rata Basis Rule might
allocate a disproportionately large amount to the holder of either the
transferred or the retained collateral. In fact, there may be many cases
in which the actual repayment of liabilities is disproportionately borne
by either the transferor or the transferee, but because the incidence
of those cases cannot be predicted, we see no compelling reason to adopt
a rule that presumes such disproportionate repayment. Although we acknowledge
that any rule that attempts to predict the course of future events will
be imperfect, we nevertheless believe that a prediction that liabilities
will be repaid in proportion to the ownership of collateral is more
likely than any other alternative to mirror economic reality. If this
assumption is correct, then although misallocations may arise in any
event, we believe that alternatives other than the Pro Rata FMV Rule
are likely to increase the incidence of such misallocations.
The same principle appears to hold true in the
case of an undercollateralized liability, if one believes that the likelihood
of appreciation in the value of the collateral also mirrors the relative
values of the collateral. (Again, we can think of no compelling reason
why the likelihood of appreciation should be allocated disproportionately
to the assets of the transferor or transferee.) If that assumption is
correct, then the All to Transferor Rule and the Transferred FMV Rule
would allocate a disproportionate amount of liability to the holder
of the retained collateral; and the All to Transferee Rule (the current
statutory presumption) and the Retained FMV Rule would allocate a disproportionate
amount to the holder of the transferred collateral. Again, the Pro Rata
Basis Rule might allocate a disproportionately large amount to the holder
of either the retained or transferred collateral. It would seem that
the Pro Rata FMV Rule would best approximate the actual likelihood of
appreciation (and thus of repayment of some or all of the underwater
portion of the liability) in the greatest number of instances.
These principles can be illustrated using two
examples:
Example One: Shareholder A owns property worth $200, which is collateral for a $100 nonrecourse liability. A transfers $60 worth of collateral with an initial basis of $40 to Sub as part of a section 351 transaction. The retained collateral has a value of $140 and an initial basis of $60. No agreements or expectations regarding repayment of the liability are manifested.In this example involving an overcollateralized liability, Sub would be treated as assuming (1) $0 under the All to Transferor Rule; (2) $100 under the All to Transferee Rule; (3) $60 under the Transferred FMV Rule; (4) $0 under the Retained FMV Rule; (5) $30 under the Pro Rata FMV Rule; or (6) $40 under the Pro Rata Basis Rule. As discussed above, only the Pro Rata FMV rule would match the liability to the relative amounts of collateral. Although there is no assurance that such a match ultimately will reflect the repayment of the liability, it seems that the other alternatives are even less likely to do so. (It may be the case, for example, that A will simply continue to treat the debt as its own, or that Sub will be expected to use all of its collateral to repay $60 of the debt. Such facts might mean that the determination of assumption is based on the agreements and expectations test rather than the default rule. If not, as noted above, there does not seem to be any one-size-fits-all default rule that can assure consistency of tax treatment and economics.)
Example Two: Shareholder A owns property worth $40, which is collateral for a $100 nonrecourse liability. A transfers $30 worth of collateral with an initial basis of $80 to Sub as part of a section 351 transaction. The retained collateral has a value of $10 and an initial basis of $20. No agreements or expectations regarding repayment of the liability are manifested.In this example involving an undercollateralized liability, Sub would be treated as assuming (1) $0 under the All to Transferor Rule; (2) $100 under the All to Transferee Rule; (3) $30 under the Transferred FMV Rule; (4) $90 under the Retained FMV Rule; (5) $75 under the Pro Rata FMV Rule; or (6) $80 under the Pro Rata Basis Rule. Under the assumptions (1) that the portion of the liability that is adequately secured most likely will be satisfied proportionate to the relative values of the retained and transferred collateral, and (2) that, if there is any appreciation that requires repayment of the currently underwater portion of the liability, such appreciation also most likely would track such relative values, the Pro Rata FMV Rule would have the highest likelihood of tracking the likely economic results. Of course, particularly in such a case where the chance of repayment of the underwater component of the liability is low and where there is little current indication of where appreciation that would increase the value of the liability might arise, there is a substantial likelihood that any choice of allocation will prove arbitrary and incorrect. It still appears reasonable to assume, however, that no other alternative is a better predictor of possible future appreciation and repayment. Moreover, using the same default rule to govern both the overcollateralized and undercollateralized case would reduce burdening an already complex test with additional complexity.
(2) Consistency with Crane and Tufts
In the discussion below, we will evaluate the
alternatives in terms of their consistency with the principles of Crane
and Tufts. In general, Crane and Tufts stand for
the proposition that the amount realized on the disposition of property
must include the full amount of any nonrecourse liability secured by
the property (at least when all the property collateralizing the nonrecourse
debt is conveyed). Crane is premised on the policy that the borrower
should be treated as benefiting for tax purposes from the entire amount
of basis financed by the debt.8 In
other words, according to Crane, the borrower should be required
to recapture as taxable income the depreciation deductions on the property
that were financed by the debt proceeds, and should not be permitted
to receive a portion of that basis tax-free. See also Treas. Reg. Section
1.1001-2(b). In Tufts, in which the liability was underwater
at the time the property was conveyed, the Court reached a similar conclusion
by determining that nonrecourse liabilities should be treated in the
same manner as recourse liabilities.9
The Notice states that the existing statutory
presumption of Section 357(d)(1)(B) (i.e., the All to Transferee Rule)
is “consistent with the principles” of the Crane
and Tufts decisions. To the extent Tufts is premised on
treating recourse and nonrecourse liabilities in the same manner, this
statement is questionable, because absent an agreement, the current
statutory framework leaves a recourse liability with the transferor
and shifts a nonrecourse liability to the transferee. To be consistent
with Tufts, the proper statutory rule for nonrecourse liabilities would
be one based on agreements and not one based on a default rule. As discussed
in greater detail below, we believe that a rule for nonrecourse liabilities
based on agreements is best in certain cases, but as discussed above,
we agree that such a rule is not always workable. Therefore, despite
any possible inconsistency with Tufts, we continue to believe
that a default rule for nonrecourse liabilities is necessary. Moreover,
we believe that Tufts can also be read as merely an alternative
route to the same rationale applied in Crane. In other words,
despite Tufts’ reliance on equating recourse and nonrecourse
liabilities, Tufts, like Crane, has the effect of preventing
a borrower from benefiting from the entire amount of debt-financed basis,
without recapturing as taxable income the depreciation deductions on
that property. Because we are assuming that a default rule for nonrecourse
liabilities will sometimes be necessary, the analysis that follows treats
both Crane and Tufts as decisions predicated on this recapture
theory.
In any event, the salient question is which
of the alternatives under consideration is the most consistent with
Crane and Tufts—the current default rule (the All
to Transferee Rule) or one of the other five alternatives. If one were
to view Crane and Tufts as dictating a rule that accelerates
any recapture of Crane/Tufts income, one would presumably seek
to maximize the amount of liabilities deemed assumed under the default
rule. In the undercollateralized case, where Craneand Tufts
are relevant, the All to Transferee Rule or the Retained FMV Rule would
best advance such an objective. As we read the Crane and Tufts
decisions, however, we see no mandate to accelerate the recapture of
debt-financed deductions as long as the borrower remains liable on the
debt. Certainly, the cases do not stand for the proposition that a borrower
is required to recognize Crane/Tufts income until such time as
the borrower is discharged of the liability, even if it is evident at
an earlier time that the liability is highly unlikely to be repaid.
Accordingly, it would seem that any rule would be consistent with the
principles of Crane and Tufts, provided the borrower is
ultimately required to recapture the benefits of debt-financed basis,
and is not able artificially to defer such recapture beyond such time
as it is effectively relieved of some or all of the liability.
Viewed in this light, it appears to us that
the Pro Rata FMV Rule is consistent with Crane and Tufts,
and reduces the risk of timing distortions that arguably are not mandated
by those decisions. Turning back to the facts of Example Two,
if one were to apply either the All to Transferee Rule or the Retained
FMV Rule, A would recapture immediately all Crane/Tuftsgain associated
with the underwater portion of the liability (i.e., $60), even though
A is still likely to repay a quarter of the debt that is adequately
secured, and, as discussed above, arguably also a quarter of any additional
repayment resulting from subsequent appreciation in collateral. This
result would appear to go beyond what is mandated by Crane and
Tufts, and would have the potential of taxing the use of basis
for which the borrower will ultimately pay (if the retained collateral
appreciates and the amount of repayment increases). This questionable
result is easier to see if one changes the facts of Example Two
to involve a transfer of only $1 (instead of $30) worth of the $40 of
collateral. There would be no reason to accelerate Crane and
Tufts recapture if A retains 39/40 of the collateral, and thus
arguably remains obligated to satisfy 97.5 percent of the liability.
Similarly, the All to Transferor Rule and Transferred
FMV Rule would seem to invite distortions. Applying those rules to the
facts of Example Two would defer any recapture so long as A retains
any portion of the collateral. Again, this result would appear particularly
egregious if the facts of Example Two involved a retention of
only $1 worth of the $40 of collateral. Although A’s use of basis
could still eventually be recaptured, there does not seem to be a compelling
policy reason to defer recapture even after A has effectively been discharged
of what likely will be 97.5 percent of the liability.
Applying the Pro Rata FMV Rule would reduce
these distortions. Under the facts of Example Two, A would realize
an amount equal to 75 percent of the debt at the time that it transfers
75 percent of the collateral. This generally would appear neither to
artificially accelerate—nor to allow the taxpayer to artificially
defer—its obligations to recapture the use of debt-financed basis.
One might argue that the Pro Rata Basis Rule
would be a more appropriate rule governing the timing of Crane
and Tufts recapture. Because the Pro Rata Basis Rule allocates
the liability according to the initial basis of the collateral, that
rule has the ability to recapture precisely the depreciation deductions
taken with respect to the basis of the transferred asset. Accordingly,
in Example Two, if A has taken $60 of depreciation deductions
on the transferred collateral, such that its basis at the time of the
transfer is $20, the amount of assumption of liability under the Pro
Rata Basis Rule would be $80. Thus, the amount of section 357(c) gain
would be $60 (absent the transfer of other assets and setting aside
questions as to whether a transaction not involving a net transfer of
value would be subject to section 351, as discussed above). This result
would precisely recapture the depreciation deductions taken on the asset
in question.
On balance, however, we do not believe that
this feature of the Pro Rata Basis Rule makes it more attractive than
the Pro Rata FMV Rule. First, it is not clear that it is appropriate
to hinge the timing of recapture (under Crane and Tufts)
on the disposition of the depreciated asset, rather than on the time
at which and extent to which the borrower is economically relieved of
the liability in question. Returning to Example Two, if, at the
time of the transfer, the retained collateral had in fact appreciated
in value to $70 (rather than declining in value to $10), the expectation
would be that the nonrecourse liability would be fully repaid (because
it would still be secured by aggregate collateral worth $100, $30 owned
by Sub and $70 owned by A). It would seem inappropriate in such a case
to apply the Pro Rata Basis Rule to treat Sub as assuming $80 of liability
(giving rise to $60 of section 357(c) gain), given that Sub is economically
more likely—under the assumptions discussed above—to repay
only 30 percent of the liability. Thus, there is a good argument that
the timing of recapture should focus on the likely discharge of the
liability, rather than on the disposition of the depreciated asset.
Moreover, there may be circumstances in which
collateral that secures a nonrecourse debt was not actually acquired
in connection with the borrowing. In such cases, the initial bases of
the collateral would not necessarily reflect the amount of the debt,
and the Pro Rata Basis Rule would become meaningless.
Taking these points into account and assuming,
as we do, that a default rule for nonrecourse liabilities will sometimes
be necessary, we believe that the best default method for determining
liability assumption is the Pro Rata FMV Rule. In our opinion, that
alternative offers the best reflection of the principles of Crane
and Tufts, even though in some cases, the Pro Rata Basis Rule
will cause the transferor to recognize an amount of section 357(c) gain
equal to the depreciation benefit the transferor has enjoyed prior to
the transfer.
(3) Impact on the transferee
Our final consideration is whether the determination of the amount of
liability assumed is appropriate from the perspective of the transferee.
Particularly in the case of an underwater liability, the question arises
as to whether it is fair to the transferee to treat the amount of liability
assumption as exceeding the value of the received collateral. As demonstrated
below, we believe that the transferee should not be inappropriately
disadvantaged by the amount of liability allocated to it. The more important
objective is to ensure that sufficient liability is allocated to the
transferor to prevent or minimize a separation of the use of basis and
corresponding income recapture.
Example Three: The facts are the same as in Example Two; however, the initial basis of the transferred collateral is $75, and the initial basis of the retained collateral is $25. At the time of the transfer, the basis of the assets remains unchanged.Focusing on these examples, the question is whether it is inappropriate from the perspective of Sub to treat Sub as assuming an amount of liability in excess of $30, the fair market value of the transferred collateral.
Example Four: The facts are the same as in Example Three, except that the basis of the transferred collateral at the time of transfer has been depreciated to $30, and the retained collateral has been depreciated to $10.
In Example Three, even if Sub were treated as assuming only $30 of the $100 liability (as would be the case under the Transferred FMV Rule), and then each of A and Sub were to transfer their assets in satisfaction of the debt at a time when the value of the assets remain unchanged, the amount realized by A under Crane and Tufts would be $70 and the amount realized by Sub would be $30. However, under section 362(a), Sub would have taken a $75 basis in the transferred collateral. As a result, A would recognize $45 in gain, and Sub would have a $45 loss. In effect, the Transferred FMV Rule would artificially separate the party that uses the debt-financed basis from the party that is required under Crane and Tufts to recapture that basis as income. Even more inappropriate results would obtain if Sub were treated as assuming less than $30 of the liability (as would be the case under the All to Transferor Rule).
In Example Four, assuming the same repayments at some point after the transfer (and assuming for simplicity no change in asset value and no further depreciation deductions taken), if Sub is treated as assuming $30 of liability, Sub would have no gain at the time of repayment ($30 amount realized, less $30 basis), and A would recognize gain of $60 ($70 amount realized, less $10 basis). This result would not suffer from the infirmity of divorcing the recapture of income from the party that enjoyed the benefits of the debt-financed basis, but would effectively permit A to defer any recapture even beyond the point where A has effectively been relieved of 75 percent of the liability.
In this situation, the Pro Rata FMV Rule would be more appropriate. In Example Three, using the Pro Rata FMV Rule, Sub would be treated as assuming $75 of the $100 liability, such that neither party would recognize gain or loss upon the repayment (for Sub, $75 amount realized less $75 basis, and for A, $25 amount realized less $25 basis). Similarly, if Sub has depreciated the $75 in basis between the time of the transfer and the repayment, Sub would recognize the $75 in recapture income.
In Example Four, using the Pro Rata FMV Rule, A would recognize $45 in section 357(c) gain at the time of the transfer ($75 liability assumed less $30 basis asset transferred). Then, at the time of the repayment, A would recognize the last $15 in recapture gain ($25 amount realized less $10 basis). Although Sub would be treated as being discharged of a $75 liability at the time of the repayment, and would have only $30 in basis under section 362(d)(1) (which limits basis to FMV where gain under section 357(c) would otherwise allow basis in excess of FMV), it would seem that the principles of Crane and Tufts should not cause Sub to recognize $45 in gain. In any event, any regulation the government issues should clarify that there should be no recapture to the transferee of basis to the extent that the basis was denied under section 362(d)(1). In other words, there should be recapture to the transferee only if the transferee enjoys the benefit of the debt-financed basis, either because the basis of the collateral remains intact at the time of the section 351 transfer, or because section 357(c) gain to the transferor gives rise to an increase in the basis of the transferred assets (including assets other than the collateral) in the hands of the transferee.
The Pro Rata Basis Rule, in contrast to the other alternatives, would arguably best achieve the objective of marrying recapture under Crane and Tufts to the party that uses the basis. Consider, for example, the facts of Example Two, in the event that the initial basis has not changed by the time of the section 351 transfer. Under the Pro Rata FMV Rule, as discussed above, Sub would assume $75 of liability, even though Sub would get the benefit of $80 in basis. Ultimately, therefore, A would recapture $5 of income with respect to basis that was in fact used by Sub, a result which would not obtain under the Pro Rata Basis Rule, which would allocate $80 of liability to Sub.
Nonetheless, this advantage of the Pro Rata Basis rule would not appear to outweigh the other advantages of the Pro Rata FMV Rule. As discussed above, (i) initial basis will not always have reflected the amount of the debt, and (ii) the Pro Rata Basis Rule will in certain circumstances (particularly where the high basis collateral depreciates, and the low basis collateral appreciates) inappropriately trigger recapture of income even where the debt ultimately will be fully repaid. In sum, taking all of the various considerations into account, and recognizing the imperfection of any particular alternative, we continue to believe that a rule which apportions liabilities in accordance with value is most often predictive of repayment and consistent with the underlying economics.
In conclusion, we believe that notwithstanding any advantages the Pro Rata Basis Rule might have in reflecting the principle of Crane and Tufts and minimizing inappropriate treatment to the transferee, the best alternative remains the Pro Rata FMV Rule, because it provides the best reflection of economic reality, and most appropriately times the recognition of gain consistent with the policies of Crane and Tufts.
III. Agreements
A. Should an agreement to allocate a nonrecourse liability be respected?
As discussed above, a rule that follows the
parties’ agreements and expectations is more likely to reflect
economic reality than is the default rule. Thus, as in the recourse
case, agreements and expectations, as broadly construed, should govern
whenever possible. In fact, the regulations should clarify that therequirement
that the transferee “agree” to satisfy a liability can be
met even if the “agreement” is manifested simply by (i)
the receipt of collateral securing either a recourse or nonrecourse
debt (which, by itself, puts the transferee on the legal hook for repaying
some or all of the debt) and (ii) other relatively contemporaneous actions
(such as payment of principal and interest on the debt). Thus, the default
rule need not apply, provided either the transferor or the transferee
is legally bound to satisfy (or indemnify for the payment of) some or
all of the debt, the actions of the parties indicate a formal or informal
mutual agreement, and there is evidence that the expectations of the
parties are consistent with such agreement.
B. Can a nonrecourse liability be treated as assumed by a party who
agrees to pay but who does not receive a transfer of any of the collateral?
This question addresses the second of the three
modifications to the current statute that the Notice proposes. For the
same reasons discussed in the preceding paragraph, we believe this question
should be answered in the affirmative. The Notice correctly states that
the current statutory scheme only imposes a presumed assumption of a
nonrecourse liability in cases in which a transferee receives property
subject to that liability. In other words, pursuant to the current statutory
regime, if a transferee does not receive property subject to a nonrecourse
liability, there will be no “assumption” within the meaning
of section 357, regardless of whether the transferee has actually agreed
to pay the liability. This result would not comport with the underlying
economics of the agreement, and thus this rule should be altered.
C. Should section 357(d)(2) provide for a reduction in the amount
presumed to have been assumed under the default rule to reflect the
role of an agreement, or does the presence of an agreement (including
agreements not described by section 357(d)(2)(A)) make such an alteration
unnecessary?
Section 357(d)(2) provides for a reduction in
the amount of a liability that is treated as assumed under the current
statutory default rule (the All to Transferee Rule). The reduction is
designed to reduce the amount of the assumption established by the default
rule so as to reflect agreements of the parties, but only in specified
circumstances. The reduction called for by the current statutory scheme
is a reduction equal to the lesser of (A) the amount of the liability
an owner of other collateral not transferred to the transferee has agreed
and is expected to satisfy, and (B) the fair market value of that untransferred
collateral owned by that person. In other words, applying section 357(d)(2),
if the transferor has agreed and is expected to pay a portion of the
nonrecourse liability, but holds assets subject to that liability with
a fair market value lower than that agreed portion, then the amount
presumed to have been assumed by the transferee pursuant to the default
rule will only be reduced by the fair market value of those assets.
The Notice proposes (in the third of the three modifications described
above) that this lesser-of rule be replaced with a rule that would simply
reduce the default amount by the amount of the liability that another
party has agreed and is expected to satisfy (regardless of the value
of any untransferred property). In light of the discussion of the role
of agreements, above, we do not believe that a reduction in the default
rule to reflect the role of an agreement is necessary. Rather, we believe
that certain agreements should simply supplant the (unreduced) default
rule of section 357(d)(1)(B) (as amended to reflect the Pro Rata FMV
Rule rather than the All to Transferee Rule). Our responses to questions
regarding various types of agreements are set forth below.
Provided the agreement is real and enforceable (whether by the creditor or the transferee), such that the transferor is obligated to provide funds to satisfy a portion of the nonrecourse liability attributable to the value of collateral held by the transferee, we believe the nonrecourse liability should not be treated as assumed by the transferee to that extent.
This question is a subset of the broader question of how to determine the amount assumed under all of the facts and circumstances, which we consider to be fundamentally a section 1001 issue, as discussed below. The particular issue here is how to interpret a unilateral promise by the transferee, without a bilateral agreement from the transferor that the transferee’s liability shall be so capped, and without an agreement from the creditor. The amount the transferee agrees to pay certainly sets a floor on the assumption, but the question is whether a higher amount assumed, determined under the default rule, should govern. We suggest that the regulations include an example illustrating the relevant circumstances, such as whether the transferor has sufficient collateral to satisfy the remaining portion and whether there are facts to suggest that the parties have different expectations. Such facts might be sufficient to reflect a mutual arrangement that the transferee’s liability will be limited to the amount it expressly agrees to pay. Otherwise, the “agreement” would seem to have little force, in which case the transferee should be deemed to assume the greater of the amount it explicitly agrees to pay or the amount determined under the default rule.
In general, we believe that agreements that can reasonably be expected to be honored should be given effect. As in all cases, however, an agreement to pay a liability overrides the default rule only to the extent that the parties also expect one another to adhere to the terms of the agreement. An example might discuss the case where several transferees nominally agree with the transferor to pay an amount in excess of their collateral, but only one of the transferees is expected to satisfy that excess liability. In that case, only the one transferee should be treated as assuming the liability. In the absence of any clear expectation, however, the nominal agreement should not control, and the Pro Rata FMV Rule should apply.
We believe that the agreement and expectation rule should apply equally in a case where the liability is undercollateralized, and where the agreement addresses the potential satisfaction of the underwater portion of the liability in the event that collateral appreciates. This would apply whether or not the collateral with the potential to appreciate is held by the party that agrees to satisfy such amount. The most extreme example would be a case where a $100 liability is secured by a single asset, with a basis and value of only $10, the transferor transfers the asset to the transferee, and the transferor agrees to indemnify the transferee for any amount in excess of $10 that the transferee pays to the creditor in satisfaction of the debt. But even the results under that example do not seem offensive. For the same reasons discussed above in the analysis of the interaction of the principles of Craneand Tufts with the choices of a default rule, there would not seem to be a reason to tax the transferor immediately on the recapture of basis, because the transferor remains potentially liable to repay creditor for the full $90 of debt-financed basis that it used prior to the section 351 transfer. If, as discussed above, the central determinant of the timing of Crane/Tufts recapture involves relief from the debt, rather than disposition of the asset, it would seem appropriate to defer recapture in such an instance. Of course, as soon as the liability is ultimately discharged, the transferor would recognize gain in the amount of its relief from liability.
D. If assumption of nonrecourse liabilities based on agreements is adopted, how does that affect the character of the debt (recourse versus nonrecourse) for purposes of subsequent transfers of the collateral? Does it make a difference whether the assumption was by agreement or an assumption that is deemed by a default rule?
If the amount of a nonrecourse liability deemed assumed is determined under a default rule, the liability retains its nonrecourse nature in all respects, and thus should be treated as nonrecourse if the related collateral is subsequently transferred. If, however, a party to the transaction agrees with another party to satisfy an amount of such liability in excess of the collateral held by that person, that person’s liability attains certain features of a recourse obligation. Nonetheless, we believe that a subsequent transfer of the collateral should be governed by the rules applicable to nonrecourse liabilities. The nonrecourse liability rule is appropriate because, as noted above, special rules are needed in the event that no discernible agreements are reached with respect to the allocation of liability following the transfer of collateral securing a nonrecourse liability. Of course, in the event that agreements were reached regarding payment of the liability at the time of the first transfer, and those agreements continue to have force, the agreements may once again alter the application of the default rule on the subsequent transfer.
An example in the Notice posits a situation in which P transfers an asset (with a value that represents one-third of the collateral for a $100 nonrecourse liability) to S1, and P agrees to indemnify S1 for any liability in excess of $20. If the agreement is consistent with the parties’ expectations, S1 would be deemed to have assumed $20 of liability. In the event that S1 subsequently transfers the asset to S2, and no further agreement is reached, the notice asks whether the rules for nonrecourse or recourse liabilities should apply. In our view, the second transfer should be subject to the nonrecourse rules. Thus, if the benefit of P’s indemnity runs to S2, S2 should be deemed to assume $20 of liability. If P’s indemnity is extinguished, S2 would be deemed to assume $33 of liability under the Pro Rata FMV default rule. In that case, S2 may be deemed to assume $13 of the liability directly from P.
The regulations might include some clarification of other collateral consequences resulting from the fact that an agreement of a party to satisfy a portion of a nonrecourse liability in excess of that person’s collateral accords certain recourse features to that person’s liability. In general, we believe that the obligation should continue to be treated as fully nonrecourse from the creditor’s perspective, because the creditor never has a right to more than the value of the overall collateral. From the perspective of the various debtors, it might be appropriate to treat the obligation as nonrecourse to any party who is potentially subject to Crane/Tufts recapture, but as recourse to other parties who assume liabilities in excess of collateral and on a recourse basis.
IV. Issues Involving Both Recourse and Nonrecourse Liabilities
A. Once a liability is treated as assumed under section 357(d), should
the transferor no longer be treated as owing the liability for purposes
of further application of section 357(d), and should that diminution
of liability prevent a subsequent transferee from assuming a liability
that the transferor does not "owe," for purposes of further application
of section 357(d)?
The liability should be treated for all purposes
as a liability of the person who is treated as retaining or assuming
it under section 357(d). If the liability is allocated to a particular
party because that party indemnifies the other party therefor, the indemnifying
party should be viewed as the obligor on the third party obligation.
B. Should rules prescribe how an agreement should be evidenced?
The rules should allow taxpayers to demonstrate
the existence and substance of agreements and expectations with reference
to the facts and circumstances, and should not prescribe a particular
form of agreement. An optional form might be provided for parties to
a transfer to reflect their agreements and expectations, but the existence
of such a form should not relieve a taxpayer from the requirement of
demonstrating an actual legal obligation to satisfy an obligation that
has been allocated to it, or from demonstrating that the expectations
of the parties are consistent with the allocation of the form.
C. How is "expected to satisfy" determined? Should interpretation
of that phrase include amounts the person is expected to pay to the
creditor and to indemnities?
Interpretation of “expected to satisfy”
should take into account indemnities, the reality of any agreements,
and the economic viability of the assuming party.
D. Should subsequent satisfaction inconsistent with the section 357(d)
treatment of the assumption be characterized under general federal income
tax principles, for example as a contribution to capital or distribution?
We believe that subsequent satisfaction of a
liability that is inconsistent with the section 357(d) treatment of
the assumption should be characterized under general federal income
tax principles.10 For example, if a
liability (or portion of a liability) is paid by a transferee who was
not deemed to assume that liability (or portion thereof), the payment
generally would be characterized as a distribution to the transferor.
Conversely, if a liability (or portion thereof) is paid by the transferor
after having been deemed assumed by the transferee, such payment would
be characterized as a contribution to capital.
We recognize that arguments can be made supporting
either the treatment advocated in the foregoing paragraph, or characterization
of the transaction under the principles of Arrowsmith v. Commissioner,
344 U.S. 6 (1952). Under an Arrowsmithanalysis, payment by a
transferee of a liability not treated as assumed at the time of the
prior asset transfer would be taxed under the “relation-back”
principles of sections 357(a), (b) and (c), as if the liability were
assumed at the time of the transfer, rather than as a later, separate
distribution under section 301. For example, if the transferred assets
had a basis of $30, and no liabilities were treated as assumed at the
time of the original transfer, payment of a $50 liability by the transferee
would be deemed to result in $20 of section 357(c) gain, rather than
in a $50 dividend. Payment by a transferor of a liability treated as
assumed by the transferee at the time of the original asset transfer
would be treated as if the liability had not been assumed by the transferee
at the time of the original transfer, instead of being treated as a
separate, later contribution to capital. Thus, if at the time of the
original transfer, a transferor had recognized $20 of section 357(c)
gain upon the transfer of an asset with a basis of $30 and a deemed
assumption of a $50 liability, payment of the liability by the transferor
presumably would result in a $20 loss (reversing the section 357(c)
gain), and in a restoration of $30 of basis to the stock of the transferee.
At the outset, we note that adoption of an Arrowsmith
approach seems fundamentally inconsistent with the adoption of any
default rule. In other words, if the proper approach is simply to defer
initial characterization in favor of a “wait and see” approach,
then a default rule would be superfluous. Other arguments in favor of
applying general tax principles rather than the principles of Arrowsmith
focus primarily on the view that subsequent determinations (or changes
in determination) as to whether the transferor or transferee will satisfy
a liability do not factually relate back to the asset transfer, but
do in substance reflect subsequent decisions effectively to transfer
value either upstream or downstream, as in the case of any contribution
or distribution. For example, if the transferor agrees to retain a liability
at the time of the asset transfer, but later—given the profitability
of the transferee—the parties decide that the transferee will
pay the debt, that decision would appear to be best characterized as
a decision to effectuate a distribution, and not as a part of the initial
transaction involving the asset transfer. Even if the determination
of the amount of liability assumed was made under the default rule,
a subsequent decision that the transferor or transferee will pay the
liability does not necessarily relate back to the earlier transfer—and
instead seems better characterized as a separate transaction.
If one were to treat subsequent payments of
liability as relating back to the initial asset transfer, one would
invite the opportunity for taxpayers to choose among different forms
of economically identical transactions, based on the desired tax consequences.
Focusing again on the case in which a transferee was not treated as
assuming a liability at the time of the asset transfer, but is now in
a better position to repay the liability, adoption of the Arrowsmith
“relation back” approach would permit the transferee to
elect to repay the liability (thereby invoking the rules of section
357) or simply to distribute cash to the transferor, who could itself
repay the liability (thereby invoking the rules of section 301). Adoption
of the “general tax principles” approach, by contrast, would
ensure that the transferee would be treated as making a distribution
under either form, and thus would preclude inappropriate electivity.
Arguments in favor of adopting the Arrowsmith
approach, by contrast, focus largely on a perception that a “general
principles approach” could be used affirmatively by taxpayers
to structure around the rules of section 357. For example, if a corporate
transferor wants to transfer assets with a basis of $30, and wants the
transferee to assume a $50 liability, the transferor might attempt to
circumvent the potential $20 of section 357(c) gain by “agreeing”
to retain the liability itself. Later, the transferee could in fact
satisfy the liability, and treat the satisfaction as a distribution,
resulting in no gain, and a dividend that is offset by a dividends-received
deduction.
On balance, as noted above, we find the arguments
in favor of a “general tax principles” approach more compelling,
because such an approach appears more consistent with a separate subsequent
determination of repayment of liability, and affords less electivity.
We acknowledge that the approach itself would not police against taxpayer
efforts to circumvent section 357(c), as described in the previous paragraph.
We note, however, that subsequent actions by the parties could in certain
circumstances be strong evidence of the existence of an agreement (or
of an agreement inconsistent with one asserted by the parties), which
could call into question the taxpayer’s treatment of the initial
asset transfer. For example, if the taxpayer argues that no section
357(c) gain was incurred in the transfer described above, because the
transferor had agreed to retain the liability, subsequent satisfaction
by the transferee would call into question the bona fides of that original
“agreement,” and may require the taxpayer to present evidence
as to whether the original purported agreement was consistent with the
parties’ expectations, and why the arrangement was revised. If
the taxpayer failed to establish that the original retention of liability
was agreed and expected, section 357(c) could be applied to the initial
transfer.
Finally, we acknowledge the need to address
the manner in which a “general tax principles” approach
would be applied in the event a transferor is no longer a shareholder
of the transferee when the repayment occurs. Regardless of whether the
rules adopt a “general tax principles” or an Arrowsmith
approach to the subsequent repayment of liabilities, we believe
that agreements governing transactions that cause the transferor and
the transferee to cease to be related generally will address the allocation
of liability repayment responsibilities. If that allocation is inconsistent
with previous presumptions about the identity of the party who was expected
to pay the liability at the time of the original transfer, it should
constitute a separate taxable event to be taken into account at the
time of the separation. Thus, repayment of a liability that occurs when
the transferor and transferee are no longer related will either be consistent
with the positions taken by the parities to that separation transaction
or it will be subject to an adjustment between the parties to that separation
transaction. Even if the repayment were characterized by applying a
“relation back” approach, that approach should relate back
to the time of the separation; in no event should a later inconsistent
repayment be an occasion to revisit the tax treatment of the original
transfer.M/p>
V. Coordination with Other Code Sections
A. Should the section 357(d) rules be the same for purposes of section
1001?
The section 357(d) rules should be the same
for purposes of section 1001. We believe the economic analysis set forth
in these comments is equally relevant for taxable transactions falling
outside of section 351.
B. Should section 7701(g) be consistent with section 357(d)?
Section 357(d) (and the regulations proposed
above) remain consistent with the Crane/Tufts objective of ensuring
that debt-financed basis/proceeds are taxed to the extent that the debt
is not fully repaid. Section 7701(g) is intended to advance the same
objective and thus remains applicable.
C. Should the section 357(d) rules apply for purposes of other code
rules that invoke section 357(d) specifically, and those that do not?
The section 357(d) rules should apply for purposes
of other code rules that invoke section 357(d) specifically, and for
those that do not. Although we have not yet analyzed the issue in detail,
we are unaware of reasons why the more economic approach contained in
these rules would be inconsistent with the policy objectives of any
other provisions of the Code. We acknowledge, however, that adoption
of these rules in other areas may require special rules to address issues
raised in specific contexts, and would change existing law, as it has
been construed by the IRS and the courts. See, e.g., Maher v. Commissioner,
469 F.2d 225 (8th Cir. 1972).
D. Should a distinction be retained between liabilities assumed and
liabilities to which property is subject, for purposes of sections 304
and 336 and other rules?
Uniform treatment should be obtained, and thus
the distinction should be eliminated.
E. Should basis rules affecting property transferees be changed?
Section 362(d)(1) would be unaffected by the
changes in these rules. However, it would appear that the IRS could
exercise its regulatory authority to override section 362(d)(2). Under
the revised pro rata default rule, transferees will receive under the
general rules the same amount of basis credit as they would under the
special provision of section 362(d)(2). Any additional amount of basis
credit would be available only if the transferee actually agrees and
is expected to satisfy the assumed liabilities, which appears appropriate.
In any event, the fair market value cap would continue to apply.
-
* The following members of the Corporate Tax Committee Advisory Board
participated in the preparation of these comments: Jasper L. Cummings
Jr. and Mark R. Hoffenberg (primary drafters), Andrew M. Eisenberg (Chair),
Lisa M. Zarlenga (Vice Chair), William Galanis, Stephen I. Flanagan,
Karen G. Gilbreath, Bryan P. Collins, Joseph M. Pari, and Robert H.
Wellen. The Corporate Tax Committee would like to thank Christine W.
Booth for her assistance in the preparation of these comments. The views
expressed in these comments are those of the Corporate Tax Committee
of the Tax Section of the District of Columbia Bar Association and not
those of the District of Columbia Bar or its Board of Governors.
- 68 Fed. Reg. 87.
- I.R.C. § 357(d)(1)(B).
- Id. (“[A] nonrecourse liability shall be treated as having been assumed by the transferee of any asset subject to such liability.”) (emphasis added).
- I.R.C. § 357(d)(2). The statute provides that this reduction is equal to the lesser of: (A) the amount of the liability that is collateralized by assets retained by a party other than the transferee and that the other party has agreed and is expected to satisfy, and (B) the fair market value of such other assets.
- S. Rep. No. 2, 106th Cong. 1st Sess. 75-76.
- H.R. Conf. Rep. No. 1033, 106th Cong., 2d Sess. 1019, n. 53.
- We understand that the Service currently is separately considering the treatment of certain transactions not involving a transfer of net value. 2003-2003 Priority Guidance Plan, reprinted in 2003 TNT 75-10 (LEXIS) (July 25, 2003).
- Crane, 331 U.S. at 6-11 (concluding that “property” does not mean “equity,” but instead means the full amount of the assets (an apartment building) undiminished by a mortgage).
- Tufts, 461 U.S. at 307 (“[W]e read Crane to have approved the Commissioner’s decision to treat a nonrecourse liability in this context as a true loan.”); and 461 U.S. at 313 (“Crane also stands for the broader proposition, however, that a nonrecourse loan should be treated as a true loan.”).
- The discussion in the text below focuses on circumstances in which a party other than the party retaining or assuming the liability under section 357(d) satisfies the liability, and is not entitled to reimbursement by the other party. If, by contrast, the transferee assumes a liability and agrees to indemnify the transferor if the liability is repaid by the transferor, payment of the liability by the transferor would result in no tax consequences, provided the transferor retains the right to indemnity. Only if the transferor were to waive its indemnification rights, would the results described in the text be invoked.
