UBS: why selling a subsidiary can weaken the parent bank

A bank sells a subsidiary to strengthen its finances. Cash comes in, a business leaves the balance sheet and the bank appears to have reduced its risk. Yet a sale below the stake’s carrying value also forces the parent to recognise a loss. The transaction can improve its liquidity while consuming capital it needs to keep operating.

Switzerland is trying to address this weakness after Credit Suisse. The proposal concerns how a parent bank finances its equity investments in foreign subsidiaries. It does not require every loan made by those subsidiaries to be financed entirely with equity. The Federal Council wants the parent’s foreign participations fully backed by Common Equity Tier 1, or CET1, the highest-quality category of regulatory capital. S02

On 17 September 2026, the Council of States postponed its decision after several hours of debate. Discussion is due to resume on 23 September. A competing proposal would allow up to half the backing to come from Additional Tier 1 instruments; a 90% CET1 alternative has also been advocated. As of 22 September, this is an unresolved legislative debate, not a final new rule. S01

The practical question behind the percentages is how much protection the parent retains when it has to write down, restructure or sell a foreign business. The total amount of capital is only part of the answer. Its quality and its location matter too.

The debt sits one level higher

A banking subsidiary has its own assets, liabilities and equity. The parent owns shares in that company and records the investment as an asset. It may have financed the investment partly with equity and partly with borrowing.

What counts as equity inside the subsidiary may therefore have been financed with debt at the parent. FINMA calls this double leverage. It does not mean a group has conjured genuine consolidated equity into existence through accounting. It describes how an equity participation is funded. S03

Consider an entirely fictional example. A parent invests 100 in a subsidiary, using 45 of its own equity and borrowing 55. The subsidiary really does receive 100 of equity. But the parent still owes its lenders 55, even if the value of that investment falls.

Consolidation eliminates the participation and the corresponding intragroup entries. It provides an economic view of the group without making its separate legal entities interchangeable. A claim against the parent is not an identical claim against every subsidiary. The constraints of the entity that actually owes the debt still matter. FINMA’s explanatory note explicitly distinguishes the parent-bank and consolidated perspectives. S04

In this debate, the relevant parent bank is UBS AG on a standalone basis. That is not the same perimeter as consolidated UBS Group AG, or UBS Switzerland AG alone. Nor should a foreign subsidiary be confused with a branch: the former is a separate company; the latter is not. Blurring these distinctions obscures the source of the risk. S07

A recovery plan can bring the loss to light

The value of a stake depends on what the subsidiary can generate for its owner, subject to the applicable valuation rules. Reducing a business or revising its profit prospects may therefore require a lower carrying value before a buyer has even been found.

FINMA’s retrospective analysis explains that Credit Suisse’s autumn 2022 restructuring lowered expected profits at several subsidiaries. The resulting decline in participation values weakened the parent’s capital and constrained further restructuring options. This identifies a particular vulnerability, not a complete explanation for Credit Suisse’s failure. S04

An actual disposal introduces another test: what an outside buyer is willing to pay. In a fictional example, selling a stake carried at 100 for 70 brings in 70 of cash and creates an accounting loss of 30, provided no impairment has already been booked. If the stake had previously been written down to 70, selling it for that amount would not create a second loss of 30.

The sale need not create the underlying economic loss. It can confirm it or bring forward its recognition. Cash received is not the same thing as capital rebuilt.

CET1 includes ordinary shares and retained earnings, net of regulatory adjustments. It is loss-absorbing funding, not a segregated pile of banknotes. A bank can have capital and run short of liquidity, or have cash available while lacking sufficient protection against losses. S16

Making the deduction before the crisis

The Federal Council’s proposal would deduct the full carrying value of foreign participations from the parent bank’s CET1. The parent would have to cover those investments with ordinary capital in addition to the capital needed for its other activities. S02

A second fictional example shows the mechanism. Start with a capital base of 160 before the specific deduction examined here, and a foreign participation of 100. The deduction leaves 60 of eligible CET1.

Now the participation loses 30 of value. The capital base falls to 130 and the stake is worth 70. Subtract 70 from 130 and eligible CET1 is still 60.

Shareholders have not escaped the loss. Their capital has fallen by 30. But the regulatory deduction, already applied to the entire investment, falls by the same amount. Under these assumptions, the impairment does not reduce the CET1 left to support the bank’s other risks.

This identity describes an impairment while the investment remains on the balance sheet, not the complete outcome of a sale. On disposal, both the participation and its deduction disappear. The broader consequences also depend on how the proceeds are used and how other risks change.

The protection is not a free accounting trick. Before the shock, the bank must be able to fund the larger deduction while maintaining the capital required for the rest of its business. FINMA’s capital-backing notes explain that advance requirement. S05

A loss-making sale can still improve capital headroom

It would nevertheless be wrong to conclude that any sale below book value must weaken a bank’s capital ratios. Removing the participation may also remove the capital requirement attached to it. The realised loss has to be compared with the requirement released.

Take a separate simulation that does not reproduce UBS’s actual requirements. A participation carried at 100 creates an illustrative capital requirement of 50. All other requirements remain fixed. The cash received is assumed to create no additional requirement in this example; taxes and transaction costs are excluded.

A sale for 70 produces a loss of 30 but releases a requirement of 50: capital headroom improves by 20. A sale for 30 produces a loss of 70 and releases the same 50: headroom declines by 20.

These are changes in the amount of capital above a requirement, not profits of 20 or percentage-point moves in a ratio. Under the stated assumptions:

Change in capital headroom = requirement released − disposal loss.

A write-down alone releases only the portion of a value-linked requirement associated with the decline in carrying value. A restructuring, an impairment and an outright sale therefore need not have the same effect. The reform is intended to give a troubled bank more room to act without making recovery depend on a sufficiently favourable sale price.

The compromise changes the mix of protection

The Council of States’ economic affairs committee also proposes a full deduction, but from Tier 1, which combines CET1 and AT1: 50% of the participation would be deducted from CET1, with up to 50% deducted from AT1. The Federal Department of Finance’s comparison, published on 1 September, sets out this architecture. S06

Additional Tier 1 (AT1) instruments are subordinated, perpetual instruments that qualify as regulatory capital when they meet specified criteria. Their distributions can be cancelled. Their principal can absorb losses through conversion or write-down under the relevant terms and triggers. They are neither ordinary senior debt nor a permanent equivalent of ordinary shares. S16

The difference is visible even before a trigger is reached. Assume another fictional loss of 30 on a participation, an exactly equal split of the deduction and enough qualifying AT1 to accommodate it.

The CET1 base falls by 30, while the CET1 deduction falls by only 15. Eligible CET1 therefore declines by 15. At the same time, the deduction from AT1 shrinks by 15. With no new issuance or conversion, 15 of previously deducted AT1 becomes eligible again. Total Tier 1 is unchanged in this model, but the mix has shifted away from CET1.

This is not a simulation of UBS’s balance sheet or all its capital ratios. It isolates one consequence of the proposed deduction rule before any AT1 loss-absorption mechanism is activated. It explains why two approaches both described as full backing need not protect ordinary capital in the same way while the bank is still trying to recover.

When AT1 becomes usable

Supporters of the compromise propose redesigning AT1 to absorb losses earlier, providing protection at a lower cost. UBS supports that direction in its 1 September statement. S07

This is a serious policy option. A Financial Stability Institute paper published by the BIS on 11 June 2026 examines why AT1 has struggled to perform its going-concern role. Its authors consider higher triggers and better-designed equity conversion, while recognising trade-offs for funding costs and the usability of capital buffers. These are research proposals, not a newly adopted Basel Committee standard. S11

But cancelling a coupon and immediately absorbing a principal loss are different operations. Coupon cancellation avoids a cash outflow and preserves earnings. It does not instantly replace the entire value lost on an investment.

Preparatory documents expose the disagreement. In an opinion dated 24 August and published on 1 September, the Swiss National Bank questions whether the arrangement examined would convert AT1 early enough, particularly given the uncertainty of investor acceptance. FINMA’s assessment of a 10 August proposal raises concerns about timing, incentives and legal risk. These opinions address the drafts they identify, not automatically every subsequent version. S12 S13

Homburger’s legal opinion of 23 August, commissioned by the State Secretariat for International Finance, welcomes the objective of improving AT1 but criticises aspects of the proposed automatic mechanisms and Swiss-specific design. It assesses the proposal dated 13 August at SIF’s request. S14

The relevant test is operational. Can a bank recognise losses and carry out a restructuring while markets, creditors and depositors still give it time to act? Loss absorption that depends on an uncertain process does not provide the same room for manoeuvre as ordinary capital already available.

The billion-dollar estimates use different baselines

UBS estimates that the compromise would require around $13 billion of additional Tier 1 at UBS AG, which could be met with AT1. Its calculation uses the 30 June 2026 balance sheet, assumes adoption of the proposed measures and applies an assumed 12.5% CET1 ratio for UBS AG. The bank separately identifies roughly $2 billion of CET1 associated with ordinance-level changes. These are its conditional estimates, not realised losses. S07

The Federal Department of Finance puts the increase in the parent’s CET1 requirements under the government package at roughly $20 billion, including the ordinance measures. Using an end-2025 snapshot, it estimates an actual CET1 capital gap for UBS of about $9 billion, reflecting, among other things, capital already held above requirements at parent and group level. Raising a requirement by 20 does not necessarily force an immediate equity issue of 20. S06

In its Financial Stability Report, published on 2 July 2026, the SNB makes the same distinction. It says including the $9 billion of capital reserves available at UBS AG at end-2025 would already be sufficient, in its calculation, to meet the proposed reform. It also notes the envisaged seven-year transition. This is a pro forma assessment, not a guarantee that the bank’s headroom will remain unchanged. S09

Management choices also affect the comparison. UBS’s 22 April statement, filed with the SEC, uses the bank’s own target-capital assumptions and challenges the government’s international comparison. The two sides do not necessarily mean the same thing by the amount of capital the bank needs. S08

Simply subtracting 13 from 20 to declare a certain saving of $7 billion would mix reporting dates, capital categories and assumptions. Adding together parent-bank and consolidated-group surpluses would be equally misleading: they are not independent pools of spare capital.

Who bears the cost of the safeguard?

More equity funding can constrain shareholder distributions and change returns measured against capital. It does not mean the money has been destroyed or ceases to finance assets. In a simple arithmetic example, retaining 10 of profit rather than distributing it leaves 10 of additional funding inside the bank. The value has not been paid to the state.

The claim that lending must become more expensive needs separate evidence. On 21 September, Reuters reported that Swiss business organisations had urged lawmakers to support the compromise, warning about financing costs. That is an interested-party argument, not a measured outcome. S15

Assessing it requires examining total funding costs, commercial margins, competition and balance-sheet choices. A higher requirement might be met through retained earnings, a different funding mix or smaller operations. The increase alone does not establish which response will prevail.

The SNB also describes a banking sector with substantial capital and liquidity buffers and considers UBS able to meet the proposed backing requirement when reserves are included. The reform addresses the bank’s room to act in a future crisis. S10

Full CET1 deduction directly addresses the transmission of participation losses into ordinary capital available for other risks. The compromise requires confidence that another capital layer can act early enough to provide comparable protection. Neither approach removes the need for sound management, usable liquidity and credible recovery planning. S03 S17

Recovery depends on capital remaining available in the right entity, at the moment when accepting a loss becomes necessary to prevent a worse crisis.

Further reading: reading bank health and reading bank stress tests.

Sources

1. S01 · Le Conseil des Etats gagne du temps sur les fonds propres d’UBSATS / Parlement suisse · 2026-09-17. Opening paragraphs; alternatives and closing paragraph.
2. S02 · Too-big-to-fail regulations: Federal Council adopts dispatch and Capital Adequacy OrdinanceDépartement fédéral des finances · 2026-04-22. Foreign participations, recovery phase and CET1 deduction paragraphs.
3. S03 · FINMA welcomes the Federal Council’s proposed measures on banking stabilityFINMA · 2025-06-06. Foreign-subsidiary funding and double-leverage passage.
4. S04 · Simplified explanatory notes on the financing and valuation of foreign participationsFINMA · 2025-06-06. Pages 1-4: parent/group distinction and Credit Suisse; pages 5-6: stylised examples.
5. S05 · Capital backing for foreign participationsFINMA · 2025-06-06. Pages 1-2: mechanism; pages 4-5: deduction and illustrative shock.
6. S06 · Faktenblatt: Eigenmittelunterlegung ausländischer Tochtergesellschaften beim Schweizer Stammhaus (Massnahme 15)Département fédéral des finances · 2026-09-01. Page 1: CET1/Tier 1 deductions; page 3 and footnotes 1-2: estimates and end-2025 balance sheet.
7. S07 · UBS statement on WAK-S recommendation on banking regulationUBS · 2026-09-01. Paragraphs on $13bn and $2bn; footnote 1: June balance sheet and assumed 12.5% ratio.
8. S08 · UBS statement on Swiss government’s regulatory capital announcementsUBS / SEC EDGAR · 2026-04-22. Notes on the end-2025 balance sheet and the bank’s target-capital assumptions.
9. S09 · Financial Stability Report 2026Banque nationale suisse · 2026-07-02. Printed page 27 (PDF page 29): “Including reserves, UBS already has sufficient capital…”; printed page 24 for perimeters.
10. S10 · Financial Stability Report 2026: selected contentsBanque nationale suisse · 2026-07-02. Summary on capital/liquidity buffers and UBS’s ability to meet the proposal.
11. S11 · Strengthening the going-concern role of AT1: options and trade-offsBRI / Financial Stability Institute · 2026-06-11. FSI Briefs 32, Highlights, reform options and authors’ views disclaimer.
12. S12 · Stellungnahme zum Antrag an die WAK-S zur Kapitalunterlegung ausländischer BeteiligungenBanque nationale suisse · 2026-09-01. PDF pages 1 and 4: document date, voluntary conversion, coupons and recovery capacity.
13. S13 · Einschätzung der FINMA zum Antrag an die WAK-S vom 10.8.2026FINMA · 2026-09-01. Three-page assessment: triggers, timing, incentives and legal consequences.
14. S14 · Vorschlag zur Stärkung der Funktion von AT1-Anleihen im Going ConcernHomburger AG, pour le SIF · 2026-09-01. 23 August opinion, introduction and assessment of automatic mechanisms; 13 August draft.
15. S15 · Swiss business groups pressure parliament over UBS capital rules before voteReuters · 2026-09-21. Report on a letter dated 18 September.
16. S16 · Basel Framework CAP10: Definition of eligible capitalComité de Bâle / BRI · 2020-06-05. CAP10.1, 10.2, 10.6 and 10.11: categories, components and AT1 criteria.
17. S17 · Too Big To FailDépartement fédéral des finances · 2026-04-22. Overview: governance, supervision, liquidity and crisis preparedness.

Limitations

Research cut-off: 22 September 2026, before the planned resumption of parliamentary debate on 23 September. No future vote is treated as settled. The August opinions from the SNB, FINMA and Homburger relate to the proposals they identify, not to an assumed final law.

Actual figures are in US dollars and retain their reporting dates and institutional perimeters. The 45/55 funding example, the 160/100/30 impairment, the disposal cases and the split CET1/AT1 calculation are entirely fictional. They exclude taxes, fees, foreign-exchange effects, other regulatory adjustments, leverage-ratio constraints and AT1 trigger activation. They are not UBS stress tests or exhaustive calculations of its ratios.

The analysis concerns foreign participations at parent-bank level. It does not assess all liquidity risks, intragroup guarantees, local rules or the outcome of Credit Suisse AT1 litigation. Interested sources are identified. No interviews or private responses from UBS or the authorities are claimed.