| Syllabus | Opinion [ Stevens ] | Concurrence [ Thomas ] | Dissent [ Scalia ] |
|---|---|---|---|
| HTML version PDF version | HTML version PDF version | HTML version PDF version | HTML version PDF version |
[May 17, 2004]
Justice Scalia, with whom The Chief
Justice, Justice O’Connor, and Justice Kennedy join,
dissenting.
My areas of agreement with the
plurality are substantial. We agree that, although all
confirmed Chapter 13 plans have been deemed feasible by a
bankruptcy judge, some nevertheless fail. See ante, at
14. We agree that any deferred payments to a secured creditor
must fully compensate it for the risk that such a failure will
occur. See ante, at 7. Finally, we agree that adequate
compensation may sometimes require an “
Our only disagreement is over what
procedure will more often produce accurate estimates of the
appropriate interest rate. The plurality would use the prime
lending rate–
a rate we know is too low–and
require the judge in every case to determine an amount by which
to increase it. I believe that, in practice, this approach
will systematically undercompensate secured creditors for the
true risks of default. I would instead adopt the contract
rate–i.e., the rate at which the creditor actually
loaned funds to the debtor–as a presumption that the
bankruptcy judge could revise on motion of either party. Since
that rate is generally a good indicator of actual risk,
disputes should be infrequent, and it will provide a quick and
reasonably accurate standard.
I
The contract-rate approach makes two assumptions, both of which are reasonable. First, it assumes that subprime lending markets are competitive and therefore largely efficient. If so, the high interest rates lenders charge reflect not extortionate profits or excessive costs, but the actual risks of default that subprime borrowers present. Lenders with excessive rates would be undercut by their competitors, and inefficient ones would be priced out of the market. We have implicitly assumed market competitiveness in other bankruptcy contexts. See Bank of America Nat. Trust and Sav. Assn. v. 203 North LaSalle Street Partnership, 526 U.S. 434, 456—458 (1999). Here the assumption is borne out by empirical evidence: One study reports that subprime lenders are nearly twice as likely to be unprofitable as banks, suggesting a fiercely competitive environment. See J. Lane, Associate Director, Division of Supervision, Federal Deposit Insurance Corporation, A Regulator’s View of Subprime Lending: Address at the National Automotive Finance Association Non-Prime Auto Lending Conference 6 (June 18—19, 2002) (available in Clerk of Court’s case file). By relying on the prime rate, the plurality implicitly assumes that the prime lending market is efficient, see ante, at 12; I see no reason not to make a similar assumption about the subprime lending market.
The second assumption is that the expected costs of default in Chapter 13 are normally no less than those at the time of lending. This assumption is also reasonable. Chapter 13 plans often fail. I agree with petitioners that the relevant statistic is the percentage of confirmed plans that fail, but even resolving that issue in their favor, the risk is still substantial. The failure rate they offer–which we may take to be a conservative estimate, as it is doubtless the lowest one they could find–is 37%. See Girth, The Role of Empirical Data in Developing Bankruptcy Legislation for Individuals, 65 Ind. L. J. 17, 40—42 (1989) (reporting a 63.1% success rate).1 In every one of the failed plans making up that 37%, a bankruptcy judge had found that “the debtor will be able to make all payments under the plan,” 11 U.S.C. § 1325(a)(6), and a trustee had supervised the debtor’s compliance, §1302. That so many nonetheless failed proves that bankruptcy judges are not oracles and that trustees cannot draw blood from a stone.
While court and trustee oversight may provide some marginal benefit to the creditor, it seems obviously outweighed by the fact that (1) an already-bankrupt borrower has demonstrated a financial instability and a proclivity to seek legal protection that other subprime borrowers have not, and (2) the costs of foreclosure are substantially higher in bankruptcy because the automatic stay bars repossession without judicial permission. See §362. It does not strike me as plausible that creditors would prefer to lend to individuals already in bankruptcy than to those for whom bankruptcy is merely a possibility–as if Chapter 13 were widely viewed by secured creditors as some sort of godsend. Cf. Dunagan, Enforcement of Security Interests in Motor Vehicles in Bankruptcy, 52 Consumer Fin. L. Q. Rep. 191, 197 (1998). Certainly the record in this case contradicts that implausible proposition. See App. 48 (testimony of Craig Cook, sales manager of Instant Auto Finance) (“Q. Are you aware of how other lenders similar to Instant Auto Finance view credit applicants who appear to be candidates for Chapter 13 bankruptcy?” “A. Negative[ly] as well”). The better assumption is that bankrupt debtors are riskier than other subprime debtors–or, at the very least, not systematically less risky.
The first of the two assumptions means that the contract rate reasonably reflects actual risk at the time of borrowing. The second means that this risk persists when the debtor files for Chapter 13. It follows that the contract rate is a decent estimate, or at least the lower bound, for the appropriate interest rate in cramdown.2
The plurality disputes these two assumptions. It argues that subprime lending markets are not competitive because “vehicles are regularly sold by means of tie-in transactions, in which the price of the vehicle is the subject of negotiation, while the terms of the financing are dictated by the seller.” Ante, at 15.3 Tie-ins do not alone make financing markets noncompetitive; they only cause prices and interest rates to be considered in tandem rather than separately. The force of the plurality’s argument depends entirely on its claim that “the terms of the financing are dictated by the seller.” Ibid. This unsubstantiated assertion is contrary to common experience. Car sellers routinely advertise their interest rates, offer promotions like “zero-percent financing,” and engage in other behavior that plainly assumes customers are sensitive to interest rates and not just price.4
The plurality also points to state and federal regulation of lending markets. Ante, at 15—16. It claims that state usury laws evince a belief that subprime lending markets are noncompetitive. While that is one conceivable explanation for such laws, there are countless others. One statistical and historical study suggests that usury laws are a “primitive means of social insurance” meant to ensure “low interest rates” for those who suffer financial adversity. Glaeser & Scheinkman, Neither a Borrower Nor a Lender Be: An Economic Analysis of Interest Restrictions and Usury Laws, 41 J. Law & Econ. 1, 26 (1998). Such a rationale does not reflect a belief that lending markets are inefficient, any more than rent controls reflect a belief that real estate markets are inefficient. Other historical rationales likewise shed no light on the point at issue here. See id., at 27. The mere existence of usury laws is therefore weak support for any position.
The federal Truth in Lending Act, 15 U.S.C. § 1601 et seq., not only fails to support the plurality’s position; it positively refutes it. The plurality claims the Act reflects a belief that full disclosure promotes competition, see ante, at 16, and n. 24; the Act itself says as much, see 15 U.S.C. § 1601(a). But that belief obviously presumes markets are competitive (or, at least, that they were noncompetitive only because of the absence of the disclosures the Act now requires). If lending markets were not competitive–if the terms of financing were indeed “dictated by the seller,” ante, at 15–disclosure requirements would be pointless, since consumers would have no use for the information.5
As to the second assumption (that the expected costs of default in Chapter 13 are normally no less than those at the time of lending), the plurality responds, not that Chapter 13 as currently administered is less risky than subprime lending generally, but that it would be less risky, if only bankruptcy courts would confirm fewer risky plans. Ante, at 16—17. Of course, it is often quite difficult to predict which plans will fail. See Norberg, Consumer Bankruptcy’s New Clothes: An Empirical Study of Discharge and Debt Collection in Chapter 13, 7 Am. Bankr. Inst. L. Rev. 415, 462 (1999). But even assuming the high failure rate primarily reflects judicial dereliction rather than unavoidable uncertainty, the plurality’s argument fails for want of any reason to believe the dereliction will abate. While full compensation can be attained either by low-risk plans and low interest rates, or by high-risk plans and high interest rates, it cannot be attained by high-risk plans and low interest rates, which, absent cause to anticipate a change in confirmation practices, is precisely what the formula approach would yield.
The plurality also claims that the contract rate overcompensates creditors because it includes “transaction costs and overall profits.” Ante, at 11. But the same is true of the rate the plurality prescribes: The prime lending rate includes banks’ overhead and profits. These are necessary components of any commercial lending rate, since creditors will not lend money if they cannot cover their costs and return a level of profit sufficient to prevent their investors from going elsewhere. See Koopmans v. Farm Credit Services of Mid-America, ACA, 102 F.3d 874, 876 (CA7 1996). The plurality’s criticism might have force if there were reason to believe subprime lenders made exorbitant profits while banks did not–but, again, the data suggest otherwise. See Lane, Regulator’s View of Subprime Lending, at 6.6
Finally, the plurality objects that
similarly situated creditors might not be treated alike.
Ante, at 11, and n. 17. But the contract rate is
only a presumption. If a judge thinks it necessary to modify
the rate to avoid unjustified disparity, he can do so. For
example, if two creditors charged different rates solely
because they lent to the debtor at different times, the judge
could average the rates or use the more recent one. The
plurality’s argument might be valid against an approach
that irrebuttably presumes the contract rate, but that
is not what I
propose.7
II
The defects of the formula approach far outweigh those of the contract-rate approach. The formula approach starts with the prime lending rate–a number that, while objective and easily ascertainable, is indisputably too low. It then adjusts by adding a risk premium that, unlike the prime rate, is neither objective nor easily ascertainable. If the risk premium is typically small relative to the prime rate–as the 1.5% premium added to the 8% prime rate by the court below would lead one to believe–then this subjective element of the computation might be forgiven. But in fact risk premiums, if properly computed, would typically be substantial. For example, if the 21% contract rate is an accurate reflection of risk in this case, the risk premium would be 13%–nearly two-thirds of the total interest rate. When the risk premium is the greater part of the overall rate, the formula approach no longer depends on objective and easily ascertainable numbers. The prime rate becomes the objective tail wagging a dog of unknown size.
As I explain below, the most relevant factors bearing on risk premium are (1) the probability of plan failure; (2) the rate of collateral depreciation; (3) the liquidity of the collateral market; and (4) the administrative expenses of enforcement. Under the formula approach, a risk premium must be computed in every case, so judges will invariably grapple with these imponderables. Under the contract-rate approach, by contrast, the task of assessing all these risk factors is entrusted to the entity most capable of undertaking it: the market. See Bank of America, 526 U.S., at 457 (“[T]he best way to determine value is exposure to a market”). All the risk factors are reflected (assuming market efficiency) in the debtor’s contract rate–a number readily found in the loan document. If neither party disputes it, the bankruptcy judge’s task is at an end. There are straightforward ways a debtor could dispute it–for example, by showing that the creditor is now substantially oversecured, or that some other lender is willing to extend credit at a lower rate. But unlike the formula approach, which requires difficult estimation in every case, the contract-rate approach requires it only when the parties choose to contest the issue.
The plurality defends the formula approach on the ground that creditors have better access to the relevant information. Ante, at 18—19. But this is not a case where we must choose between one initial estimate that is too low and another that is too high. Rather, the choice is between one that is far too low and another that is generally reasonably accurate (or, if anything, a bit too low). In these circumstances, consciously choosing the less accurate estimate merely because creditors have better information smacks more of policymaking than of faithful adherence to the statutory command that the secured creditor receive property worth “not less than the allowed amount” of its claim, 11 U.S.C. § 1325(a)(5)(B)(ii) (emphasis added). Moreover, the plurality’s argument assumes it is plausible–and desirable–that the issue will be litigated in most cases. But the costs of conducting a detailed risk analysis and defending it in court are prohibitively high in relation to the amount at stake in most consumer loan cases. Whatever approach we prescribe, the norm should be–and undoubtedly will be–that the issue is not litigated because it is not worth litigating. Given this reality, it is far more important that the initial estimate be accurate than that the burden of proving inaccuracy fall on the better informed party.
There is no better demonstration of the inadequacies of the formula approach than the proceedings in this case. Petitioners’ economics expert testified that the 1.5% risk premium was “very reasonable” because Chapter 13 plans are “supposed to be financially feasible” and “the borrowers are under the supervision of the court.” App. 43. Nothing in the record shows how these two platitudes were somehow manipulated to arrive at a figure of 1.5%. It bears repeating that feasibility determinations and trustee oversight do not prevent at least 37% of confirmed Chapter 13 plans from failing. On cross-examination, the expert admitted that he had only limited familiarity with the subprime auto lending market and that he was not familiar with the default rates or the costs of collection in that market. Id., at 44—45. In light of these devastating concessions, it is impossible to view the 1.5% figure as anything other than a smallish number picked out of a hat.
Based on even a rudimentary financial analysis of the facts of this case, the 1.5% figure is obviously wrong–not just off by a couple percent, but probably by roughly an order of magnitude. For a risk premium to be adequate, a hypothetical, rational creditor must be indifferent between accepting (1) the proposed risky stream of payments over time and (2) immediate payment of its present value in a lump sum. Whether he is indifferent–i.e., whether the risk premium added to the prime rate is adequate–can be gauged by comparing benefits and costs: on the one hand, the expected value of the extra interest, and on the other, the expected costs of default.
Respondent was offered a risk premium of 1.5% on top of the prime rate of 8%. If that premium were fully paid as the plan contemplated, it would yield about $60.8 If the debtor defaulted, all or part of that interest would not be paid, so the expected value is only about $50.9 The prime rate itself already includes some compensation for risk; as it turns out, about the same amount, yielding another $50.10 Given the 1.5% risk premium, then, the total expected benefit to respondent was about $100. Against this we must weigh the expected costs of default. While precise calculations are impossible, rough estimates convey a sense of their scale.
The first cost of default involves depreciation. If the debtor defaults, the creditor can eventually repossess and sell the collateral, but by then it may be substantially less valuable than the remaining balance due–and the debtor may stop paying long before the creditor receives permission to repossess. When petitioners purchased their truck in this case, its value was almost equal to the principal balance on the loan.11 By the time the plan was confirmed, however, the truck was worth only $4,000, while the balance on the loan was $4,895. If petitioners were to default on their Chapter 13 payments and if respondent suffered the same relative loss from depreciation, it would amount to about $550.12
The second cost of default involves liquidation. The $4,000 to which respondent would be entitled if paid in a lump sum reflects the replacement value of the vehicle, i.e., the amount it would cost the debtor to purchase a similar used truck. See Associates Commercial Corp. v. Rash, 520 U.S. 953, 965 (1997). If the debtor defaults, the creditor cannot sell the truck for that amount; it receives only a lesser foreclosure value because collateral markets are not perfectly liquid and there is thus a spread between what a buyer will pay and what a seller will demand. The foreclosure value of petitioners’ truck is not in the record, but, using the relative liquidity figures in Rash as a rough guide, respondent would suffer a further loss of about $450.13
The third cost of default consists of the administrative expenses of foreclosure. While a Chapter 13 plan is in effect, the automatic stay prevents secured creditors from repossessing their collateral, even if the debtor fails to pay. See 11 U.S.C. § 362. The creditor’s attorney must move the bankruptcy court to lift the stay. §362(d). In the District where this case arose, the filing fee for such motions is now $150. See United States Bankruptcy Court for the Southern District of Indiana, Schedule of Bankruptcy Fees (Nov. 1, 2003) (available in Clerk of Court’s case file). And the standard attorney’s fee for such motions, according to one survey, is $350 in Indiana and as high as $875 in other States. See J. Cossitt, Chapter 13 Attorney Fee Survey, American Bankruptcy Institute Annual Spring Meeting (Apr. 10—13, 2003) (available in Clerk of Court’s case file). Moreover, bankruptcy judges will often excuse first offenses, so foreclosure may require multiple trips to court. The total expected administrative expenses in the event of default could reasonably be estimated at $600 or more.
I have omitted several other costs of default, but the point is already adequately made. The three figures above total $1,600. Even accepting petitioners’ low estimate of the plan failure rate, a creditor choosing the stream of future payments instead of the immediate lump sum would be selecting an alternative with an expected cost of about $590 ($1,600 multiplied by 37%, the chance of failure) and an expected benefit of about $100 (as computed above). No rational creditor would make such a choice. The risk premium over prime necessary to make these costs and benefits equal is in the neighborhood of 16%, for a total interest rate of 24%.14
Of course, many of the estimates I have made can be disputed. Perhaps the truck will depreciate more slowly now than at first, perhaps the collateral market is more liquid than the one in Rash, perhaps respondent can economize on attorney’s fees, and perhaps there is some reason (other than judicial optimism) to think the Tills were unlikely to default. I have made some liberal assumptions,15 but also some conservative ones.16 When a risk premium is off by an order of magnitude, one’s estimates need not be very precise to show that it cannot possibly be correct.
In sum, the 1.5% premium adopted in this case is far below anything approaching fair compensation. That result is not unusual, see, e.g., In re Valenti, 105 F.3d 55, 64 (CA2 1997) (recommending a 1%—3% premium over the treasury rate–i.e., approximately a 0% premium over prime); it is the entirely predictable consequence of a methodology that tells bankruptcy judges to set interest rates based on highly imponderable factors. Given the inherent uncertainty of the enterprise, what heartless bankruptcy judge can be expected to demand that the unfortunate debtor pay triple the prime rate as a condition of keeping his sole means of transportation? It challenges human nature.
III
Justice Thomas rejects both the formula approach and the contract-rate approach. He reads the statutory phrase “property to be distributed under the plan,” 11 U.S.C. § 1325(a)(5)(B)(ii), to mean the proposed payments if made as the plan contemplates, so that the plan need only pay the risk-free rate of interest. Ante, at 3 (opinion concurring in judgment). I would instead read this phrase to mean the right to receive payments that the plan vests in the creditor upon confirmation. Because there is no guarantee that the promised payments will in fact be made, the value of this property right must account for the risk of nonpayment.
Viewed in isolation, the phrase is
susceptible of either meaning. Both the promise to make
payments and the proposed payments themselves are property
rights, the former “to be distributed under the plan”
immediately upon confirmation, and the latter over the life of
the plan. Context, however, supports my reading. The cramdown
option which the debtors employed here is only one of three
routes to confirmation. The other two–creditor acceptance
and collateral surrender, §§1325(a)(5)(A),
(C)–
are both creditor protective, leaving the secured
creditor roughly as well off as he would have been had the
debtor not sought bankruptcy protection. Given this, it is
unlikely the third option was meant to be substantially
underprotective; that would render it so much more
favorable to debtors that few would ever choose one of the
alternatives.
The risk-free approach also leads to anomalous results. Justice Thomas admits that, if a plan distributes a note rather than cash, the value of the “property to be distributed” must reflect the risk of default on the note. Ante, at 4—5. But there is no practical difference between obligating the debtor to make deferred payments under a plan and obligating the debtor to sign a note that requires those same payments. There is no conceivable reason why Congress would give secured creditors risk compensation in one case but not the other.
Circuit authority uniformly rejects the risk-free approach. While Circuits addressing the issue are divided over how to calculate risk, to my knowledge all of them require some compensation for risk, either explicitly or implicitly. See In re Valenti, supra, at 64 (treasury rate plus 1%—3% risk premium); GMAC v. Jones, 999 F.2d 63, 71 (CA3 1993) (contract rate); United Carolina Bank v. Hall, 993 F.2d 1126, 1131 (CA4 1993) (creditor’s rate for similar loans, but not higher than contract rate); In re Smithwick, 121 F.3d 211, 214 (CA5 1997) (contract rate); In re Kidd, 315 F.3d 671, 678 (CA6 2003) (market rate for similar loans); 301 F.3d 583, 592—593 (CA7 2002) (case below) (contract rate); In re Fisher, 930 F.2d 1361, 1364 (CA8 1991) (market rate for similar loans) (interpreting parallel Chapter 12 provision); In re Fowler, 903 F.2d 694, 698 (CA9 1990) (prime rate plus risk premium); In re Hardzog, 901 F.2d 858, 860 (CA10 1990) (market rate for similar loans, but not higher than contract rate) (Chapter 12); In re Southern States Motor Inns, Inc., 709 F.2d 647, 652—653 (CA11 1983) (market rate for similar loans) (interpreting similar Chapter 11 provision); see also 8 Collier on Bankruptcy, ¶1325.06[3][b], p. 1325—37 (15th ed. rev. 2004). Justice Thomas identifies no decision adopting his view.
Nor does our decision in Rash,
520 U.S. 953,
support the risk-free approach. There we considered whether a
secured creditor’s claim should be valued at what the
debtor would pay to replace the collateral or at the lower
price the creditor would receive from a foreclosure sale.
Justice Thomas contends that Rash selected the former in
order to compensate creditors for the risk of plan failure, and
that, having compensated them once in that context, we need not
do so again here. Ante, at 5. I disagree with this
reading of Rash. The Bankruptcy Code provides that
“value shall be determined in light of the purpose of the
valuation and of the proposed disposition or use of [the]
property.” 11
U.S.C. § 506(a). Rash held that the
foreclosure-value approach failed to give effect to this
language, because it assigned the same value whether the debtor
surrendered the collateral or was allowed to retain it in
exchange for promised payments. 520 U.S., at 962. “From
the creditor’s perspective as well as the debtor’s,
surrender and retention are not equivalent acts.”
Ibid. We did point out that retention entails risks for
the creditor that surrender does not. Id., at
962—963. But we made no effort to correlate that
increased risk with the difference between replacement and
foreclosure value. And we also pointed out that retention
benefits the debtor by allowing him to continue to use the
property–a factor we considered “[o]f prime
significance.” Id., at 963. Rash stands
for the proposition that surrender and retention are
fundamentally different sorts of “disposition or
use,” calling for different valuations. Nothing in the
opinion suggests that we thought the valuation difference
reflected the degree of increased risk, or that we adopted the
replacement-value standard in order to compensate for
increased risk. To the contrary, we said that the
debtor’s “actual use … is the proper guide under
a prescription hinged to the property’s ‘disposition
or use.’
If Congress wanted to compensate secured creditors for the risk of plan failure, it would not have done so by prescribing a particular method of valuing collateral. A plan may pose little risk even though the difference between foreclosure and replacement values is substantial, or great risk even though the valuation difference is small. For example, if a plan proposes immediate cash payment to the secured creditor, he is entitled to the higher replacement value under Rash even though he faces no risk at all. If the plan calls for deferred payments but the collateral consists of listed securities, the valuation difference may be trivial, but the creditor still faces substantial risks. And a creditor oversecured in even the slightest degree at the time of bankruptcy derives no benefit at all from Rash, but still faces some risk of collateral depreciation.17
There are very good reasons for Congress to prescribe full risk compensation for creditors. Every action in the free market has a reaction somewhere. If subprime lenders are systematically undercompensated in bankruptcy, they will charge higher rates or, if they already charge the legal maximum under state law, lend to fewer of the riskiest borrowers. As a result, some marginal but deserving borrowers will be denied vehicle loans in the first place. Congress evidently concluded that widespread access to credit is worth preserving, even if it means being ungenerous to sympathetic debtors.
* *
Today’s judgment is unlikely to burnish the Court’s reputation for reasoned decisionmaking. Eight Justices are in agreement that the rate of interest set forth in the debtor’s approved plan must include a premium for risk. Of those eight, four are of the view that beginning with the contract rate would most accurately reflect the actual risk, and four are of the view that beginning with the prime lending rate would do so. The ninth Justice takes no position on the latter point, since he disagrees with the eight on the former point; he would reverse because the rate proposed here, being above the risk-free rate, gave respondent no cause for complaint. Because I read the statute to require full risk compensation, and because I would adopt a valuation method that has a realistic prospect of enforcing that directive, I respectfully dissent.
1. The true rate of plan failure is almost certainly much higher. The Girth study that yielded the 37% figure was based on data for a single division (Buffalo, New York) from over 20 years ago (1980—1982). See 65 Ind. L. J., at 41. A later study concluded that “the Buffalo division ha[d] achieved extraordinary results, far from typical for the country as a whole.” Whitford, The Ideal of Individualized Justice: Consumer Bankruptcy as Consumer Protection, and Consumer Protection in Consumer Bankruptcy, 68 Am. Bankr. L. J. 397, 411, n. 50 (1994). Although most of respondent’s figures are based on studies that do not clearly exclude unconfirmed plans, one study includes enough detail to make the necessary correction: It finds 32% of filings successful, 18% dismissed without confirmation of a plan, and 49% dismissed after confirmation, for a postconfirmation failure rate of 60% (i.e., 49%/(32% + 49%)). See Norberg, Consumer Bankruptcy’s New Clothes: An Empirical Study of Discharge and Debt Collection in Chapter 13, 7 Am. Bankr. Inst. L. Rev. 415, 440—441 (1999). This 60% failure rate is far higher than the 37% reported by Girth.
2. The contract rate is only a presumption, however, and either party remains free to prove that a higher or lower rate is appropriate in a particular case. For example, if market interest rates generally have risen or fallen since the contract was executed, the contract rate could be adjusted by the same amount in cases where the difference was substantial enough that a party chose to make an issue of it.
3. To the extent the plurality argues that subprime lending markets are not “perfectly competitive,” ante, at 15 (emphasis added), I agree. But there is no reason to doubt they are reasonably competitive, so that pricing in those markets is reasonably efficient.
4. I confess that this is “nonresponsive” to the argument made in the plurality’s footnote (that the contract interest rate may not accurately reflect risk when set jointly with a car’s sale price), see ante, at 15, n. 20; it is in response to the quite different argument made in the plurality’s text (that joint pricing shows that the subprime lending market is not competitive), see ante, at 15. As to the former issue, the plurality’s footnote makes a fair point. When the seller provides financing itself, there is a possibility that the contract interest rate might not reflect actual risk because a higher contract interest rate can be traded off for a lower sale price and vice versa. Nonetheless, this fact is not likely to bias the contract-rate approach in favor of creditors to any significant degree. If a creditor offers a promotional interest rate–such as “zero-percent financing”–in return for a higher sale price, the creditor bears the burden of showing that the true interest rate is higher than the contract rate. The opposite tactic–inflating the interest rate and decreasing the sale price–is constrained at some level by the buyer’s option to finance through a third party, thus taking advantage of the lower price while avoiding the higher interest rate. (If a seller were to condition a price discount on providing the financing itself, the debtor should be entitled to rely on that condition to rebut the presumption that the contract rate reflects actual risk.) Finally, the debtor remains free to rebut the contract rate with any other probative evidence. While joint pricing may introduce some inaccuracy, the contract rate is still a far better initial estimate than the prime rate.
5. The plurality also argues that regulatory context is relevant because it “distorts the market.” Ante, at 16. Federal disclosure requirements do not distort the market in any meaningful sense. And while state usury laws do, that distortion works only to the benefit of debtors under the contract-rate approach, since it keeps contract rates artificially low.
6. Some transaction costs are avoided by the
creditor in bankruptcy–
for example, loan-origination
costs such as advertising. But these are likely only a minor
component of the interest rate. According to the record in
this case, for example, the average interest rate on new-car
loans was roughly 8.5%–only about 0.5% higher than the
prime rate and 2.5% higher than the risk-free treasury rate.
App. 43 (testimony of Professor Steve Russell). And the 2%
difference between prime and treasury rates represented
“mostly … risk [and] to some extent transaction
costs.” Id., at 42. These figures suggest that
loan-origination costs included in the new-car loan and prime
rates but not in the treasury rate are likely only a fraction
of a percent. There is no reason to think they are
substantially higher in the subprime auto lending market. Any
transaction costs the creditor avoids in bankruptcy are thus
far less than the additional ones he incurs.
7. The plurality’s other, miscellaneous criticisms do not survive scrutiny either. That the cramdown provision applies to prime as well as subprime loans, ante, at 18, proves nothing. Nor is there any substance to the argument that the formula approach will perform better where “national or local economic conditions drastically improved or declined after the original loan was issued.” Ibid. To the extent such economic changes are reflected by changes in the prime rate, the contract rate can be adjusted by the same amount. See n. 2, supra. And to the extent they are not, they present the same problem under either approach: When a party disputes the presumption, the court must gauge the significance of the economic change and adjust accordingly. The difference, again, is that the contract-rate approach starts with a number that (but for the economic change) is reasonably accurate, while the formula approach starts with a number that (with or without the economic change) is not even close.
8. Given its priority, and in light of the amended plan’s reduced debtor contributions, the $4,000 secured claim would be fully repaid by about the end of the second year of the plan. The average balance over that period would be about $2,000, i.e., half the initial balance. The total interest premium would therefore be 1.5% x 2 x $2,000 = $60. In this and all following calculations, I do not adjust for time value, as timing effects have no substantial effect on the conclusion.
9. Assuming a 37% rate of default that results on average in only half the interest’s being paid, the expected value is $60 x (1 — 37%/2), or about $50.
10. According to the record in this case,
the prime rate at the time of filing was 2% higher than the
risk-free treasury rate, and the difference represented
“mostly … risk [and] to some extent transaction
costs.” App. 42 (testimony of Professor Steve Russell);
see also Federal Reserve Board, Selected Interest Rates,
http://www.federalreserve.gov/releases/
h15/data.htm (as
visited Apr. 19, 2004) (available in Clerk of Court’s case
file) (historical data showing prime rate typically exceeding
3-month constant-maturity treasury rate by 2%—3.5%). If
“mostly” means about three-quarters of 2%, then the
risk compensation included in the prime rate is 1.5%. Because
this figure happens to be the same as the risk premium over
prime, the expected value is similarly $50. See nn. 8—9,
supra.
11. The truck was initially worth $6,395; the principal balance on the loan was about $6,426.
12. On the original loan, depreciation ($6,395 — $4,000, or $2,395) exceeded loan repayment ($6,426 — $4,895, or $1,531) by $864, i.e., 14% of the original truck value of $6,395. Applying the same percentage to the new $4,000 truck value yields approximately $550.
13. The truck in Rash had a replacement value of $41,000 and a foreclosure value of $31,875, i.e., 22% less. 520 U.S., at 957. If the market in this case had similar liquidity and the truck were repossessed after losing half its remaining value, the loss would be 22% of $2,000, or about $450.
14. A 1.5% risk premium plus a 1.5% risk component in the prime rate yielded an expected benefit of about $100, see supra, at 11—12, so, to yield $590, the total risk compensation would have to be 5.9 times as high, i.e., almost 18%, or a 16.5% risk premium over prime.
15. For example, by ignoring the possibility that the creditor might recover some of its undersecurity as an unsecured claimant, that the plan might fail only after full repayment of secured claims, or that an oversecured creditor might recover some of its expenses under 11 U.S.C. § 506(b).
16. For example, by assuming a failure rate of 37%, cf. n. 1, supra, and by ignoring all costs of default other than the three mentioned.
17. It is true that, if the debtor defaults, one of the costs the creditor suffers is the cost of liquidating the collateral. See supra, at 13. But it is illogical to “compensate” for this risk by requiring all plans to pay the full cost of liquidation (replacement value minus foreclosure value), rather than an amount that reflects the possibility that liquidation will actually be necessary and that full payments will not be made.