Finance theory has a clear answer to the question of which risks deserve a premium. Lenders do not follow it. Valuation experts often do not either. And bankruptcy courts, which must set cramdown rates and resolve valuation disputes, have developed their own methods for pricing risk. The gap between theory and practice is not academic. It determines whether a secured creditor receives prime plus 1% or something closer to a market rate, whether a business is worth enough to support a plan, and why a lender insists on a personal guarantee rather than simply charging more. Counsel who understand where the answers diverge are better positioned in loan negotiations, workouts, and valuation disputes

 

 

1.The Theory: Only Systematic Risk Is Priced

Modern portfolio theory divides risk into two categories. Systematic risk moves with the economy as a whole: recessions, interest rate shocks, credit cycles. Idiosyncratic risk belongs to a single firm: the loss of a key customer, a fire at the plant, a founder who retires without a successor. The distinction matters because idiosyncratic risk can be diversified. An investor holding enough unrelated companies will see one firm’s bad luck offset by another’s good luck. Systematic risk cannot be diversified, because the whole portfolio falls together when the economy does.

 

The Capital Asset Pricing Model (CAPM) is built on that distinction. Because investors can eliminate idiosyncratic risk at no cost, the model assumes the market will not compensate them for bearing it. Under CAPM, expected return depends only on an asset’s exposure to systematic risk, measured by beta. Beta is simply a measure of how much an asset’s value tends to rise or fall when the market as a whole does. Nothing in the model asks whether the company has one customer or a thousand. CAPM has its empirical critics, but it remains the standard framework for how diversified investors price public equities. The trouble starts when the investor is not diversified, the instrument is not equity, or the one setting the price is a judge.

 

 

Why Debt Pricing Departs From the Model

 

Lenders price borrower-specific risk as a matter of course. That is not a failure to understand finance. Three features of lending explain it.

The payoff is asymmetric. An equity holder captures the upside if the company triples in value. A lender’s best case is repayment with interest, and its downside runs to zero. The lender’s return turns almost entirely on two questions: will this borrower default, and how much will be lost if it does.

 

 

Diversification reduces the variance of losses, not their expected amount. This point is easy to miss. A bank with 10,000 loans can forecast its aggregate losses with reasonable accuracy, but it still has to charge for them. How much it expects to lose turns on facts specific to each borrower and each deal: how likely the borrower is to default, how much the lender would lose if it did, and how much would be outstanding at the time. A loan spread therefore bundles several components: expected loss, compensation for systematic risk, liquidity, servicing costs, and the cost of the capital the lender must hold against the loan. The first two are easy to confuse. Expected loss is the lender’s estimate of what it will lose on average across loans like this one, much as an insurer estimates claims. Charging for it does not reward the lender for taking risk. It only keeps the lender whole. A risk premium is the additional return the lender demands because actual losses are uncertain and tend to arrive together in a downturn, when many borrowers struggle at once. Only the risk premium is compensation for bearing risk in the CAPM sense.

 

Many lenders are not diversified. A community bank with a regional footprint and a concentrated commercial real estate portfolio cannot diversify away any single borrower’s risk. For that lender, idiosyncratic risk is simply risk.

 

Structure as a Substitute for Price

The obvious question is why a lender facing a riskier borrower does not just charge more. The answer is that a higher rate can make the problem worse. Raising rates can attract riskier applicants (adverse selection) and push existing borrowers toward riskier projects (moral hazard). Past a certain point, a higher rate can lower the lender’s expected return rather than raise it.

Lenders respond with non-price terms. Each targets a specific source of loss or the incentive problem behind it:

 

 

Lender tool	What it targets
Financial covenants (leverage, coverage, liquidity)	Early warning that the probability of default is rising; control rights before a payment default
Collateral and borrowing base	Reduces loss given default
Personal guarantee	Reduces loss given default and moral hazard by putting the owner’s own assets at stake
Reporting requirements	Reduces information asymmetry, which narrows uncertainty about the probability of default and limits adverse selection
Amortization and shorter maturity	Reduces exposure at default over time and shortens the window for deterioration

Rate, covenants, collateral, and guarantees are therefore not independent negotiating points. They are different ways of pricing the same risk, and a concession on one tends to reappear in another.

 

Where Courts Price Risk

Courts price risk directly in two recurring settings: setting a cramdown interest rate and valuing a business. In both, the court decides which risks deserve a premium, whether or not the opinion frames it that way.

 

Cramdown Rates

In Till v. SCS Credit Corp., 541 U.S. 465 (2004), the lender in a Chapter 13 case sought its 21% contract rate. The debtors proposed 9.5%, a prime rate of approximately 8% plus 1.5% for default risk, and the bankruptcy court confirmed the plan over the lender’s objection. Id. at 471–72. A plurality of the Supreme Court endorsed that formula approach. The plurality declined to set a specific adjustment but noted that courts had generally approved adjustments of 1% to 3% over prime. Id. at 480.

 

Viewed through the lens of risk theory, the formula is a hybrid. The prime rate is what banks charge creditworthy commercial borrowers, so it already reflects baseline credit risk and the economic conditions behind it. The adjustment on top is expressly debtor-specific, turning on the circumstances of the estate, the nature of the security, and the duration and feasibility of the plan. Id. at 479. Till, in other words, prices exactly the idiosyncratic risk that CAPM says should earn nothing, and in practice the premium stays small: if a plan would need an “eye-popping” rate to compensate the creditor, it probably should not be confirmed. Id. at 480–81. Despite criticism that the size of the adjustment is arbitrary, the Till formula has become standard practice in cramdown. See Wells Fargo Bank, N.A. v. Tex. Grand Prairie Hotel Realty, L.L.C. (In re Tex. Grand Prairie Hotel Realty, L.L.C.), 710 F.3d 324, 333, 336 (5th Cir. 2013).

 

The case for the formula is stronger than its critics allow. A cramdown loan is imposed over the creditor’s objection, so there is no voluntary market to observe. The formula is inexpensive to apply, starts from an objective benchmark, and keeps every confirmation hearing from becoming a contest between competing financing experts. See Till, 541 U.S. at 476 n.14, 479.

 

The case against it is just as direct. A rate no lender would voluntarily offer does not give the creditor the present value of its claim, which is what the statute requires. Even the Fifth Circuit, while affirming a formula rate, conceded that no willing lender would have extended credit on the terms the secured creditor was forced to accept, and described that result as the natural consequence of a method that trades market realities for simple, feasible reorganizations. Tex. Grand Prairie, 710 F.3d at 336. The plurality left a door open. In a footnote, it suggested that in Chapter 11, where lenders actively market financing to debtors, a court might ask what rate an efficient market would produce. Till, 541 U.S. at 476 n.14.

 

That footnote produced a circuit split. The Sixth and Second Circuits apply a two-step test: if an efficient market exists for comparable financing, the court uses the market rate, and if not, it falls back on the Till formula. Bank of Montreal v. Official Comm. of Unsecured Creditors (In re Am. HomePatient, Inc.), 420 F.3d 559, 568 (6th Cir. 2005), cert. denied, 127 S. Ct. 55 (2006); Apollo Glob. Mgmt., LLC v. BOKF, NA (In re MPM Silicones, L.L.C.), 874 F.3d 787, 800 (2d Cir. 2017), cert. denied sub nom. BOKF, N.A. v. Momentive Performance Materials, Inc., 2018 U.S. LEXIS 3822 (U.S. June 18, 2018). The Second Circuit reasoned that where an efficient market exists, its rate is preferable to a formula improvised by a court. MPM Silicones, 874 F.3d at 800–01. The Fifth Circuit, by contrast, declines to require any particular method and reviews the rate only for clear error. Tex. Grand Prairie, 710 F.3d at 331.

 

The difference is not trivial. In Texas Grand Prairie, the debtor’s expert proposed prime (3.25%) plus 1.75%, for a 5% rate. The lender’s expert proposed 8.8%, built from a blended market rate of 9.3%. Id. at 327, 334–35. The bankruptcy court confirmed at 5%, and the Fifth Circuit affirmed. Id. at 335, 337. The two approaches produce different numbers because they answer different questions. A market rate captures the full spread a real lender would demand: expected loss, systematic risk, liquidity, and capital cost. The formula rate captures what a creditworthy borrower would pay, plus a judicial estimate of debtor-specific risk, typically 1% to 3%. In circuits applying the two-step test, the contest has largely shifted to whether an efficient market exists at all, and courts almost invariably conclude that one does not. Id. at 333.

 

Valuation and the Company-Specific Risk Premium

The same problem appears in reverse in discounted cash flow valuations. Experts valuing a company may build a discount rate from CAPM and then add a company-specific risk premium for risks beta does not capture, such as customer concentration, key-person dependence, or a thin management bench. The argument for the premium is straightforward. CAPM’s diversification logic assumes an investor who holds many companies. A buyer acquiring a single company holds one, bears its idiosyncratic risk in full, and has every reason to price it. If the purpose of the valuation is to estimate what a real buyer would pay, ignoring that risk understates the discount rate.

 

The argument against it is practical rather than theoretical. Unlike beta, which is estimated from market data, the premium has no observable input to anchor it. It is a judgment call layered on top of a size premium that may already capture much of the same risk, and it can duplicate risks already reflected in the cash flow projections. An input that is unobservable, difficult to verify, and capable of moving value more than any operating assumption invites reverse-engineering.

 

Bankruptcy courts have not settled the question. Some build company-specific risk into the discount rate. See, e.g., In re Zenith Elecs. Corp., 241 B.R. 92, 104 (Bankr. D. Del. 1999) (rejecting a 17% CAPM-based discount rate in favor of a 25% rate drawn from venture fund returns because the debtor’s new technology faced regulatory and market-adoption risks and the debtor “is no Microsoft”); CNB Int’l, Inc. v. Kelleher (In re CNB Int’l, Inc.), 393 B.R. 306, 320–22 (Bankr. W.D.N.Y. 2008) (adding a 3% to 4% premium for customer concentration, an untested business plan, and divided management, and rejecting the argument that a small-company premium already captured those risks), aff’d in part, vacated in part on other grounds, 440 B.R. 31 (W.D.N.Y. 2010). Others credit experts who leave it out. See, e.g., Am. Classic Voyages Co. v. JP Morgan Chase Bank (In re Am. Classic Voyages Co.), 367 B.R. 500, 513–14 & n.20 (Bankr. D. Del. 2007) (crediting an expert’s CAPM-based discount rate, which excluded an unsystematic risk premium the expert testified was inappropriate, based on his experience and adherence to generally accepted procedures).

 

In a bankruptcy valuation dispute, the consequences are direct. A higher discount rate lowers enterprise value. Lower enterprise value can mean the debtor is insolvent or that a junior creditor class is out of the money. Whether that saves or sinks a plan depends on the plan. A plan that hands the reorganized company to senior lenders needs a low valuation, while a plan that promises recoveries to junior classes needs a high one. For a company whose cash flows grow 5% a year, raising the discount rate from 10% to 11% cuts its value by about 17%. A few points of unexplained company-specific risk can decide which plan survives. The two settings sit in tension. In cramdown, courts add a debtor-specific premium deliberately but keep it small. In valuation, courts are divided over whether to add one at all. Both settings reflect the same concern: idiosyncratic risk is real, but it is difficult to measure, and an unmeasured premium is easy to manipulate.

 

Where the Disagreement Shows Up in Practice

 

Loan negotiations. Rate, covenants, collateral, and guarantees function as substitutes. A lender that agrees to a lower rate without a guarantee has to account for the risk somewhere else, whether in tighter covenants or a larger collateral cushion. It follows that a borrower who reduces the lender’s uncertainty, through better reporting or cleaner collateral, has something to trade for room on the other terms. A borrower seeking concessions on all of them at once is asking the lender to absorb idiosyncratic risk without compensation, and a lender pricing that risk has no reason to agree.

 

Workouts. Because expected loss turns on both the probability of default and the loss given default, a workout proposal can address either. New collateral, a paydown from the sale of a non-core asset, or an expanded reporting package reduces the lender’s exposure or its uncertainty without adding to the debtor’s debt service. A higher rate does the opposite: for the adverse-selection and moral-hazard reasons discussed earlier, it may increase the very default risk it is meant to compensate.

 

Cramdown litigation. The governing framework determines what evidence matters. Where the two-step test applies, the central dispute is whether a comparable market exists for financing of similar term, size, and collateral. See Tex. Grand Prairie, 710 F.3d at 337. Where the formula governs, the dispute centers on the Till factors, and an adjustment untethered to specific facts about the estate, the collateral, and plan feasibility is vulnerable to the criticism that it is arbitrary.

 

Valuation disputes. A discount rate can be broken into its components. Beta captures some risks, and the size premium captures others. A company-specific premium is defensible to the extent its components can be identified, quantified, and shown not to duplicate risks already built into the projections. A risk counted once in the cash flows and again in the discount rate is a risk counted twice, and value is understated accordingly.

 

Conclusion

CAPM’s answer is elegant, but for lenders and courts it is mostly beside the point. Lenders price borrower-specific risk because their payoff is asymmetric, because diversification does not eliminate expected loss, and because many of their portfolios are concentrated. Courts price it in cramdown because Till ties the adjustment to this debtor and this plan, and divide over it in valuation because it is so easily inflated. The more useful question is not which theory is correct. It is where each risk has been priced in the number at issue, who priced it, and whether it has been priced twice.