---
title: "Distressed, Sovereign and Bank-Capital Credit"
book: "Markets II: Rates, FX and Credit"
subject: quant
language: en
chapter: 26
exercises: 8
source: https://one-course.com/books/quant/2/en/chapter/26-distressed-sovereign-and-bank-capital-credit
---

# Chapter 26 — Distressed, Sovereign and Bank-Capital Credit

On Sunday 19 March 2023 the Swiss authorities announced that UBS would take over Credit Suisse. The same day the supervisor, FINMA, said that the extraordinary government support the deal required triggered a complete write-down of all of Credit Suisse’s Additional Tier 1 bonds, about CHF 16 billion, to zero. Credit Suisse’s shareholders, who rank below those bonds in a liquidation, received UBS shares worth CHF 3 billion. The next day the European supervisors and resolution authority stated that in the European Union equity absorbs losses first. In October 2025 a Swiss court held the write-down unlawful; FINMA appealed. This chapter is about credit after things go wrong: [distressed debt](#def-m2-distressed-sovereign-and-bank-capital-credit-distressed) and [exchange offers](#def-m2-distressed-sovereign-and-bank-capital-credit-distressed), sovereigns that cannot be taken to a bankruptcy court, and the bank bonds designed to absorb losses before the bank fails.

## 26.1 Restructuring and exchange offers

**Definition 26.1 (Distressed debt, exchange offer).**

*Distressed debt* is debt of a borrower in or near default, trading at a price that reflects an expected loss rather than a spread over the risk-free rate. An *exchange offer* is a borrower’s offer to its creditors to swap their claims for new instruments, typically worth less, with lower face value, lower coupons or longer maturities, sometimes with cash.

A company in trouble can file for bankruptcy, where a court imposes a plan on all creditors by class; or it can negotiate out of court, and [exchange offers](#def-m2-distressed-sovereign-and-bank-capital-credit-distressed) are how. Out of court can be cheaper and faster, but each creditor decides for itself, and a creditor who does not accept keeps its old claim. The value of an offer is the present value of what it gives, at the yield at which the new instruments will trade after the exchange, the *exit yield*, compared with the old claim: the loss in present value is the offer’s haircut, often much larger than the cut in face value.

**Example 26.2 (An exchange offer).**

A borrower offers, per 100 of old bonds, 50 of new 15-year bonds paying 4% and 5 in cash for those who tender. At an exit yield of 9% the new bonds are worth 29.85 and the package 34.85: a face-value haircut of 50% but a present-value haircut of 65.2%. At an exit yield of 6% the package is worth 45.29, at 12% only 27.76 ([Figure 26.1](#fig-m2-distressed-sovereign-and-bank-capital-credit-offer)).

![Value of the chapter’s exchange offer, per 100 of old face, against the exit yield at which the new 15-year 4% bonds will trade: with the 5 of cash paid to those who tender, and without it, which is what a holder bound by a collective action clause receives. The dots are . Illustrative; data: the chapter’s tutorial.](https://one-course.com/images/onecourse/chapters/quant-2/m2-distressed-sovereign-and-bank-capital-credit/fig-3c3ae9896947.svg)

***Figure 26.1.** Value of the chapter’s [exchange offer](#def-m2-distressed-sovereign-and-bank-capital-credit-distressed), per 100 of old face, against the exit yield at which the new 15-year 4% bonds will trade: with the 5 of cash paid to those who tender, and without it, which is what a holder bound by a [collective action clause](#def-m2-distressed-sovereign-and-bank-capital-credit-cac) receives. The dots are [Example 26.2](#ex-m2-distressed-sovereign-and-bank-capital-credit-offer). Illustrative; data: the chapter’s tutorial.*

## 26.2 Sovereign debt: collective action clauses and holdouts

A sovereign has no bankruptcy court. Its bonds can be restructured only by agreement, and every creditor who does not agree can refuse, keep its claim and sue.

**Definition 26.3 (Collective action clause, holdout creditor).**

A *collective action clause* (CAC) in a bond lets a qualified majority of its holders change its payment terms for all holders, the minority included. A *holdout creditor* is one that does not accept a restructuring and keeps, or litigates on, its original claim.

The IMF describes the traditional clause as series by series: typically 75% of the holders of one bond can amend it, so a creditor who buys a blocking stake in one series can hold that series out. Newer clauses aggregate: in the model clauses of the International Capital Market Association, a single vote of 75% of the principal of all the series affected binds every series, provided all are offered the same terms, or, in a two-limb version, 66 2/3% in aggregate and more than half of each series.

**Definition 26.4 (Pari passu clause).**

A *pari passu clause* states that a bond ranks equally with the issuer’s other unsecured debt of the same kind.

Two cases shaped current practice. Argentina defaulted in 2001 and restructured about 93% of its external debt in exchanges in 2005 and 2010. Holdouts led by NML Capital, with claims of about USD 1.6 billion, sued in New York and won a reading of the [pari passu clause](#def-m2-distressed-sovereign-and-bank-capital-credit-pp) as requiring ratable payment: Argentina could not pay the exchanged bonds without paying the holdouts in full at the same time, and the banks in the payment chain were barred from passing on payments. The Supreme Court declined to hear Argentina’s appeal in June 2014, and at the end of July 2014 the exchanged bondholders, who had settled for about 27 cents on the dollar in present value, went unpaid. Greece in 2012 did the opposite: a law passed in February introduced [collective action clauses](#def-m2-distressed-sovereign-and-bank-capital-credit-cac) into its Greek-law bonds, binding if two thirds of the participating holders agreed. Holders of about 85% of the EUR 206 billion of privately held bonds accepted, the clauses made the exchange binding on the Greek-law bonds, and new bonds of about EUR 100 billion of face value replaced the old. About EUR 6 billion of English-law bonds held out; in May 2012 Greece repaid a maturing EUR 435 million English-law bond in full.

Why hold out? A holdout’s value is a probability tree ([Figure 26.2](#fig-m2-distressed-sovereign-and-bank-capital-credit-tree)): the sovereign may pay the holdouts in full to be rid of them, which is likelier the fewer they are, as Greece did; if not, a lawsuit may win the whole claim years later, as NML won its case; or the bond may stay in default. The more others tender, the better holding out looks, which is the free-rider problem CACs exist to solve.

![The holdout’s probability tree in the chapter’s example, values per 100 of old face discounted at 9%. The chance of being paid in full soon rises steeply with the share of creditors who tendered, since the sovereign can afford to pay a small remainder. Illustrative.](https://one-course.com/images/onecourse/chapters/quant-2/m2-distressed-sovereign-and-bank-capital-credit/fig-bada9186ddfd.svg)

***Figure 26.2.** The holdout’s probability tree in the chapter’s example, values per 100 of old face discounted at 9%. The chance of being paid in full soon rises steeply with the share of creditors who tendered, since the sovereign can afford to pay a small remainder. Illustrative.*

**Proposition 26.5 (The free-rider band).**

Let $T$ be the value of tendering, $H(p)$ the value of holding out when a share $p$ tenders, increasing in $p$, and $p^*$ the participation at which $H(p^*) = T$. Without a [collective action clause](#def-m2-distressed-sovereign-and-bank-capital-credit-cac), holding out pays whenever $p > p^*$. With an aggregated clause of threshold $\tau > p^*$, a holdout is bound to the offer once $p \geq \tau$ and receives only what the offer gives the bound; holding out pays only for $p^* < p < \tau$, and not at all if $\tau \leq p^*$.

**Proof.** Below $\tau$ the holdout keeps its claim, worth $H(p)$, which exceeds $T$ exactly when $p > p^*$. At or above $\tau$ the clause amends its bond to the offer’s terms, which are worth at most $T$ and less by any cash reserved for those who tendered. ∎

**Example 26.6 (Where holding out pays).**

In the tree of [Figure 26.2](#fig-m2-distressed-sovereign-and-bank-capital-credit-tree), holding out is worth 33.75 if no one tenders, 34.73 at 60% participation and 38.97 at 74%, against 34.85 for tendering. It pays from a participation of 60.9%. With a 75% aggregated clause, at 75% or more the holdout is bound and receives the new bonds alone, 29.85: the band in which holding out pays is 60.9% to 75%. Without a clause it pays at any participation above 60.9%, and at 90% it is worth 58.72 ([Figure 26.3](#fig-m2-distressed-sovereign-and-bank-capital-credit-holdout)).

![Value of holding out against the share of creditors who tender, without a collective action clause and with a 75% aggregated clause, and the value of tendering. Holding out pays above 60.9% participation; the clause removes the gain from 75%, where the holdout is bound to the offer and loses the cash paid to those who tendered. Illustrative; data: the chapter’s tutorial.](https://one-course.com/images/onecourse/chapters/quant-2/m2-distressed-sovereign-and-bank-capital-credit/fig-540e0a770b6e.svg)

***Figure 26.3.** Value of holding out against the share of creditors who tender, without a [collective action clause](#def-m2-distressed-sovereign-and-bank-capital-credit-cac) and with a 75% aggregated clause, and the value of tendering. Holding out pays above 60.9% participation; the clause removes the gain from 75%, where the holdout is bound to the offer and loses the cash paid to those who tendered. Illustrative; data: the chapter’s tutorial.*

## 26.3 Bank capital: contingent convertibles

A bank that fails in a crisis may be too important to liquidate. After 2008 the regulators’ answer was to make some of its creditors absorb losses while the bank keeps operating.

**Definition 26.7 (Contingent convertible bond, additional tier 1).**

A *contingent convertible bond* (CoCo) is a bank bond that is converted into shares or written down when a trigger is hit, such as the bank’s common equity ratio falling below a level set in the bond. *Additional tier 1* (AT1) capital consists of perpetual instruments, usually CoCos, whose coupons the bank may cancel and which absorb losses while the bank is a going concern; they rank above common equity and below all other debt in a liquidation.

**Definition 26.8 (Point of non-viability, bail-in).**

The *point of non-viability* is the moment an authority decides that a bank would fail without a write-down of its capital instruments or public support. A *bail-in* is the write-down or conversion into equity of a failing bank’s liabilities by an authority, so that creditors instead of taxpayers bear its losses.

The Basel Committee required in January 2011 that, from 2013, all capital instruments other than common equity issued by internationally active banks can be written off or converted into common equity, at the authority’s option, at the earlier of a decision that a write-off is necessary to keep the bank viable and a decision to inject public capital or equivalent support without which it would not be. An AT1 bond therefore carries two triggers: its own ratio trigger, while the bank is a going concern, and the authority’s at the [point of non-viability](#def-m2-distressed-sovereign-and-bank-capital-credit-pon). Its holders are paid a high coupon for the risk that either is hit, and for the risk that the bank simply stops paying the coupon, which it may do without defaulting.

![The loss-absorbing stack of a bank. The EU authorities stated in March 2023 that common equity absorbs losses first and AT1 only after its full use; in Credit Suisse’s case the AT1 bonds were written down under their contractual viability trigger and an emergency ordinance, while shareholders received UBS shares. Schematic; heights not to scale.](https://one-course.com/images/onecourse/chapters/quant-2/m2-distressed-sovereign-and-bank-capital-credit/fig-0be15405e6eb.svg)

***Figure 26.4.** The loss-absorbing stack of a bank. The EU authorities stated in March 2023 that common equity absorbs losses first and AT1 only after its full use; in Credit Suisse’s case the AT1 bonds were written down under their contractual viability trigger and an emergency ordinance, while shareholders received UBS shares. Schematic; heights not to scale.*

## 26.4 The 2023 write-down

Credit Suisse’s AT1 bonds said that they would be written down completely in a *viability event*, in particular if the bank received extraordinary government support. On 19 March 2023 it received liquidity assistance loans backed by a federal default guarantee, and an emergency ordinance of the Federal Council authorised FINMA to order the write-down of AT1 capital; FINMA ordered it, and the bonds’ nominal value of about CHF 16 billion became zero, raising the bank’s core capital. Tier 2 bonds were not written down. Under the merger terms, Credit Suisse shareholders received one UBS share for every 22.48 of theirs, CHF 3 billion in all ([Figure 26.4](#fig-m2-distressed-sovereign-and-bank-capital-credit-stack)).

To bondholders who read the capital stack as a liquidation order, this was a reversal: the claim senior to equity received nothing while equity received something. The European supervisors’ statement of 20 March, that equity absorbs losses first and AT1 only after its full use, told investors that the European framework would not do the same. The write-down followed the bonds’ contractual terms rather than a liquidation order, and whether those terms and the ordinance allowed it became a matter for the courts: on 1 October 2025 the Federal Administrative Court, in the first of about 360 proceedings, held that the contractual viability event had not occurred and that the write-down lacked a legal basis; the judgment was not final, and FINMA appealed to the Federal Supreme Court.

**As of September 2026 — The Credit Suisse AT1 litigation.**

The Federal Administrative Court’s judgment of 1 October 2025 annulled FINMA’s order of 19 March 2023 writing down about CHF 16.5 billion of AT1 bonds; FINMA announced on 15 October 2025 that it would appeal to the Federal Supreme Court. No final ruling had been found when this chapter was written.

**Example 26.9 (Two orders of loss).**

An illustrative bank has 45 of common equity, 16 of AT1, 12 of Tier 2 and 80 of senior debt eligible for [bail-in](#def-m2-distressed-sovereign-and-bank-capital-credit-pon), in billions. A loss of 20 absorbed in the order of the EU statement leaves equity with 25 and every bond intact. An AT1 write-down under a contractual trigger removes all 16 of the AT1 and leaves the equity with 45.

## 26.5 Tutorial: an exchange offer against holding out

**Goal.** Value an [exchange offer](#def-m2-distressed-sovereign-and-bank-capital-credit-distressed) at exit yields, value holding out with a probability tree, find the participation at which holding out starts to pay, and see what an aggregated [collective action clause](#def-m2-distressed-sovereign-and-bank-capital-credit-cac) changes. **End state:** Figures [26.1](#fig-m2-distressed-sovereign-and-bank-capital-credit-offer) and [26.3](#fig-m2-distressed-sovereign-and-bank-capital-credit-holdout), Examples [26.2](#ex-m2-distressed-sovereign-and-bank-capital-credit-offer), [26.6](#ex-m2-distressed-sovereign-and-bank-capital-credit-band) and [26.9](#ex-m2-distressed-sovereign-and-bank-capital-credit-stack) and the numbers of the weekend problem.

1. **The offer**: bond values at an exit yield, haircuts, the package. `def bond_pv (face: float , coupon: float , years: int , y: float ) -> float : """Present value of an annual-coupon bullet bond at yield y.""" return sum (face * coupon / (1 + y) ** t for t in range (1 , years + 1 )) + face / (1 + y) ** years def npv_haircut (value: float , claim: float = 100.0 ) -> float : return 1.0 - value / claim @dataclass (frozen=True ) class Offer : new_face: float # per 100 of old face coupon: float years: int cash: float = 0.0 # paid only to those who tender def value (self , y: float ) -> float : return bond_pv(self .new_face, self .coupon, self .years, y) + self .cash` **Listing 26.1.** Bond values, haircuts and the exchange package. code/firm/recovery/firm_recovery.py
2. **The holdout**: the probability tree, the clause, the break-even participation and the loss-absorbing stack. `@dataclass (frozen=True ) class Holdout : paid_soon: float # PV of being paid in full soon lawsuit_prob: float lawsuit_pv: float # PV of the whole claim recovered by litigation stuck: float # value of a bond that stays in default power: float = 8.0 # P(paid soon) = participation ** power def value (self , p: float ) -> float : paid = p ** self .power return paid * self .paid_soon + (1 - paid) * (self .lawsuit_prob * self .lawsuit_pv + (1 - self .lawsuit_prob) * self .stuck) def holdout_payoff (p: float , h: Holdout, bound_value: float , cac: float | None ) -> float : """Value of holding out at participation p; above a CAC threshold the holdout is bound to the offer (without the cash reserved for tendering).""" return bound_value if cac is not None and p >= cac else h.value(p) def breakeven_participation (h: Holdout, tender_value: float ) -> float : """Participation at which holding out, not yet bound, is worth as much as tendering.""" lo, hi = 0.0 , 1.0 for _ in range (100 ): mid = 0.5 * (lo + hi) lo, hi = (mid, hi) if h.value(mid) < tender_value else (lo, mid) return 0.5 * (lo + hi) def absorb (stack: list [tuple [str , float ]], loss: float ) -> dict [str , float ]: """Losses taken by each layer, the first-listed absorbing first; returns what each layer keeps.""" kept = {} for name, size in stack: hit = min (size, loss) loss -= hit kept[name] = size - hit return kept` **Listing 26.2.** The holdout’s tree, the CAC, the break-even and loss absorption. code/firm/recovery/firm_recovery.py
3. **Run** `restructuring_demo.tutorial()` , `restructuring_demo.capital()` and `fig_restructuring.py` .

**What to change next.** Replace the aggregated clause by series-by-series clauses on four bonds, and let a holdout buy 26% of one series: find the participation the sovereign now needs. Then make the chance of being paid soon depend on the sovereign’s cash rather than on participation alone.

## 26.6 Build: restructuring-offer valuation

**Purpose.** The miniature firm’s distressed desk decides whether to tender or hold out, and its credit desk values bank capital bonds under both orders of loss.

**Interface.** `bond_pv(face, coupon, years, y)`; `npv_haircut(value, claim)`; `Offer(new_face, coupon, years, cash).value(y)`; `Holdout(paid_soon, lawsuit_prob, lawsuit_pv, stuck, power).value(p)`; `holdout_payoff(p, holdout, bound_value, cac)`; `breakeven_participation(holdout, tender_value)`; `absorb(stack, loss)`.

**Rules.** Annual coupons; values per 100 of old face at one exit yield; the chance of being paid soon is participation to a power; a CAC binds all holders at or above its threshold, without the tender cash; losses absorbed in the order of the stack.

**Acceptance tests.** `code/firm/recovery/tests/`: a bond at its coupon yield is worth par; the offer’s value falls with the exit yield; the holdout’s value rises with participation and the clause binds; the break-even solves $H(p) = T$; the stack absorbs in order.

**Stretch.** Series-by-series clauses and blocking stakes; a game between creditors with a distribution of beliefs about participation; recovery analysis of a company’s capital structure in a bankruptcy plan; AT1 pricing with coupon cancellation and extension risk (One Quant Book 6).

Sources and further reading

- FINMA, press releases of 19 and 23 March 2023 and of 15 October 2025; Swiss Federal Administrative Court, media release of 14 October 2025.
- UBS, media release of 19 March 2023.
- SRB, EBA and ECB Banking Supervision, statement of 20 March 2023.
- Basel Committee on Banking Supervision, press release of 13 January 2011 on loss absorbency at the point of non-viability.
- BIS Quarterly Review, December 2012, box on governing law and the Greek debt restructuring.
- IMF, “Strengthening the contractual framework to address collective action problems in sovereign debt restructuring”, October 2014.

## 26.7 Exercises

**Exercise 26.1 ★.**

Per 100 of old bonds, an offer gives 60 of new 10-year 5% bonds. What is its face haircut, and its present-value haircut at an exit yield of 8%?

**Solution of Exercise 26.1.**

A face haircut of 40%. The new bonds are worth 47.92 at 8%, a present-value haircut of 52.1%.

**Exercise 26.2 ★.**

What is the difference between a series-by-series and a single-limb aggregated [collective action clause](#def-m2-distressed-sovereign-and-bank-capital-credit-cac)?

**Solution of Exercise 26.2.**

A series-by-series clause needs a qualified majority, typically 75%, of each bond separately, so a holdout with a blocking stake in one series can keep that series out. A single-limb aggregated clause counts one vote across all the series affected, 75% of their total principal in the ICMA model, so no single series can block, provided every series is offered the same terms.

**Exercise 26.3 ★.**

Name the two triggers an AT1 bond carries, and what each depends on.

**Solution of Exercise 26.3.**

Its contractual trigger, usually the bank’s common equity ratio falling below a set level, which depends on the bank’s losses; and the [point of non-viability](#def-m2-distressed-sovereign-and-bank-capital-credit-pon), which depends on an authority’s decision that the bank would fail without a write-down or public support.

**Exercise 26.4 ★★.**

With the Greek threshold of two thirds instead of 75%, over what range of participation does holding out pay in the chapter’s example?

**Solution of Exercise 26.4.**

From 60.9% to 66.7%: the clause binds holdouts at two thirds, so the band shrinks by 8.3 points to 5.8.

**Exercise 26.5 ★★.**

Why did the ratable-payment reading of the [pari passu clause](#def-m2-distressed-sovereign-and-bank-capital-credit-pp) make Argentina’s restructuring harder, and how have newer bond contracts responded?

**Solution of Exercise 26.5.**

It meant that paying the creditors who had accepted the exchange required paying the holdouts in full at the same time, and let the holdouts stop the payments through the banks that processed them; accepting creditors lost the certainty of being paid, which made future exchanges harder to sell. Newer contracts state that the clause does not require ratable payment, and use aggregated [collective action clauses](#def-m2-distressed-sovereign-and-bank-capital-credit-cac).

**Exercise 26.6 ★★.**

Why did Greece repay its English-law holdouts in full in 2012 when the Greek-law bondholders took a large loss?

**Solution of Exercise 26.6.**

The Greek law could change only Greek-law bonds; the English-law bonds could not be amended without their holders’ votes, and a default on them would have exposed Greece to lawsuits abroad. The holdouts were small relative to the whole, so paying them in full was cheaper than the fight: the free-rider paid.

**Exercise 26.7 ★★★.**

*Coding.* Raise the lawsuit’s probability of success from 0.3 to 0.5 and find the new break-even participation. What does a stronger legal position do to the free-rider band?

**Solution of Exercise 26.7.**

Holding out is then worth 46.26 even if no one tenders, more than the 34.85 of tendering: it pays at any participation below the clause’s 75%. A stronger legal position widens the band to everything below the threshold, so only the clause stops free-riding.

**Exercise 26.8 ★★★.**

*Find the flaw.* “AT1 bonds rank above shares, so AT1 holders cannot lose everything while shareholders keep something.” Correct it.

**Solution of Exercise 26.8.**

The ranking applies in a liquidation. An AT1 bond is also written down or converted by its own terms and by the authority at the [point of non-viability](#def-m2-distressed-sovereign-and-bank-capital-credit-pon), while the bank is still operating, and nothing in those terms requires the shareholders to be wiped out first; in Credit Suisse’s case the AT1 bonds went to zero while shareholders received CHF 3 billion of UBS shares. The EU authorities have said they would apply the liquidation order; the Swiss case is before the courts.

## 26.8 Problem: Hold Out or Tender

**Problem 26.1.**

Weekend problem — the free-rider band

A sovereign in default offers, per 100 of old bonds, 50 of new 15-year 4% bonds and 5 in cash for those who tender. The new bonds are expected to trade at a 9% yield. A fund holding 100 million of old bonds weighs holding out: the sovereign may pay remaining holdouts in full next year, with a probability equal to the participation raised to the eighth power; if not, a lawsuit has a 30% chance of recovering 130 per 100 in six years; otherwise the bonds stay in default, worth 15. The bonds carry a single-limb aggregated [collective action clause](#def-m2-distressed-sovereign-and-bank-capital-credit-cac) with a 75% threshold.

**Part I — The offer.**

1. What is the offer worth per 100, with and without the cash?
2. What are its face and present-value haircuts?
3. What is it worth at exit yields of 6% and 12%?
4. Why does the exit yield matter so much?
5. What does the fund receive if it tenders its 100 million?

**Part II — Holding out.**

6. Value each branch of the tree.
7. What is holding out worth if no one tenders, and at 60% and 74% participation?
8. At what participation does holding out start to pay?
9. What happens to a holdout at 75% participation or more?
10. What would holding out be worth at 90% participation without a clause?

**Part III — Strategy.**

11. If every creditor reasoned like the fund, could participation settle between 60.9% and 75%?
12. Why does the cash for tendering matter once the clause binds?
13. How could the fund block the clause, and why is that harder with aggregation?
14. What did Greece’s English-law holdouts receive, and why?
15. What did Argentina’s exchanged bondholders suffer from the holdouts’ victory?

**Part IV — Judgement.**

16. Why is the probability of being paid soon hard to estimate?
17. Should a fund that buys defaulted bonds cheaply to litigate be allowed to block a restructuring?
18. How would you hedge the fund’s position while it waits?
19. State the *named result* : the participation band in which holding out pays, and what the clause does above it.
20. In one sentence: what problem do [collective action clauses](#def-m2-distressed-sovereign-and-bank-capital-credit-cac) solve?

**Solution of Problem 26.1.**

**1.** 34.85 with the cash, 29.85 without. **2.** 50% of face; 65.2% in present value. **3.** 45.29 at 6%, 27.76 at 12%. **4.** Because the new bonds pay a low coupon for fifteen years: their value, and so the creditors’ real loss, depends on the market’s view of the sovereign after the exchange. **5.** 34.85 million: 29.85 million of new bonds and 5 million in cash. **6.** Paid next year: $100/1.09 = 91.74$; the lawsuit: $130/1.09^6 = 77.51$; in default: 15; the branch without early payment is worth $0.3 \times 77.51 + 0.7 \times
15 = 33.75$. **7.** 33.75; 34.73; 38.97. **8.** At 60.9%. **9.** The clause binds it to the exchange: it gets the new bonds, 29.85, without the cash. **10.** 58.72. **11.** Not stably: at any participation in that band each creditor prefers to hold out, which lowers participation; creditors who expect 75% to be reached tender to keep the cash, which raises it. Expectations decide which way it goes. **12.** The cash is the difference between tendering and being bound, 5 per 100: it is what a holdout loses by waiting once the clause binds. **13.** By buying more than 25% of the aggregated principal, which is far more than a quarter of one series; aggregation makes a blocking stake much more expensive. **14.** They were paid in full as their bonds matured, such as EUR 435 million in May 2012, because the Greek law could not reach them and they were few. **15.** In July 2014 they were not paid, because the courts barred payment on the exchanged bonds unless the holdouts were paid in full at the same time. **16.** It depends on the sovereign’s cash, politics and legal exposure, and on how many others hold out, all of which change during the restructuring. **17.** A matter of judgement: holdouts enforce contracts and discipline debtors, but they can block relief that most creditors accept; aggregated clauses let a large majority decide. **18.** Little is liquid: CDS on the sovereign, if any settled on the default, are gone; the fund can hedge the exit yield with the sovereign’s other traded bonds or related currencies, but the legal outcome is unhedgeable. **19.** Named result: *the free-rider band*: holding out pays for participation between 60.9% and 75%; at 75% the clause binds the holdout to the offer, worth 29.85 against 34.85 for tendering. **20.** They stop a minority from free-riding on a restructuring that a qualified majority has accepted.

## 26.9 Interview questions

**Interview question 26.1 ★ trader, researcher.**

How is a sovereign debt restructuring different from a corporate bankruptcy?

**Solution of Interview question 26.1.**

A company can file for bankruptcy, where a court imposes a plan on classes of creditors and can liquidate assets; a sovereign cannot, has few seizable assets, and restructures by [exchange offers](#def-m2-distressed-sovereign-and-bank-capital-credit-distressed) and contract terms. Holdouts can sue but struggle to collect, and the outcome depends on governing law, [collective action clauses](#def-m2-distressed-sovereign-and-bank-capital-credit-cac) and politics.

*What the interviewer is looking for: no court, contracts, holdouts.*

**Interview question 26.2 ★ trader, bank.**

What is an AT1 bond, and what are its risks?

**Solution of Interview question 26.2.**

A perpetual bank bond counted as going-concern capital, with discretionary coupons, written down or converted when a capital ratio trigger is hit or when the authority declares the [point of non-viability](#def-m2-distressed-sovereign-and-bank-capital-credit-pon). Risks: coupon cancellation, the bank not calling it at the first call date, write-down or conversion, and the authorities’ choice of loss order, as 2023 showed.

*What the interviewer is looking for: coupons, extension, triggers, authority discretion.*

**Interview question 26.3 ★★ researcher.**

How would you value a distressed [exchange offer](#def-m2-distressed-sovereign-and-bank-capital-credit-distressed)?

**Solution of Interview question 26.3.**

Value the new instruments at an exit yield estimated from comparable restructured credits, add cash and any sweeteners, and compare with the recovery from holding out, a probability tree over being paid, suing and staying in default, discounted consistently; the choice depends on expected participation and the legal terms.

*What the interviewer is looking for: exit yield, holdout alternative, participation.*

**Interview question 26.4 ★★ trader, researcher.**

Explain the holdout problem and how [collective action clauses](#def-m2-distressed-sovereign-and-bank-capital-credit-cac) address it.

**Solution of Interview question 26.4.**

Each creditor prefers others to accept the loss while it keeps its full claim, so participation can stall. [Collective action clauses](#def-m2-distressed-sovereign-and-bank-capital-credit-cac) let a qualified majority bind the minority; aggregated clauses count across all bonds, so a blocking stake in one bond no longer suffices.

*What the interviewer is looking for: free-riding and the binding vote.*

**Interview question 26.5 ★★ risk, bank.**

What did the Credit Suisse AT1 write-down teach investors about bank capital?

**Solution of Interview question 26.5.**

That contractual terms and emergency law, not only the liquidation order, decide who loses; that authorities can differ, as the EU’s statement showed; and that the legal basis can be challenged for years. Investors may now demand a premium for that jurisdictional risk.

*What the interviewer is looking for: contract versus hierarchy, jurisdiction, legal risk.*

**Interview question 26.6 ★★★ developer.**

Design a tool that tracks, for a sovereign’s outstanding bonds, the governing law, the type of [collective action clause](#def-m2-distressed-sovereign-and-bank-capital-credit-cac) and the votes needed to bind each series.

**Solution of Interview question 26.6.**

Parse each bond’s prospectus into a record: governing law, pari passu wording, clause type (series-by-series, two-limb, single-limb), thresholds, quorum and disenfranchised holdings; keep outstanding amounts current from issuance and buyback data; compute the votes needed under each procedure and the stake that would block each series; flag changes when new bonds are issued.

*What the interviewer is looking for: structured contract data and vote arithmetic.*