2026-09-15 Dienstag

The Swiss National Bank runs its policy rate "clean", aimed at price stability alone, and handles financial-stability vulnerabilities with macroprudential tools — of which the sectoral countercyclical capital buffer is the only one with a Swiss legal mandate. That buffer sits at its legal maximum of 2.5%, reactivated in 2022, so there is no macroprudential room left to lean against a further build-up of risk in credit and real estate markets. The SNB's answer is institutional rather than immediate: it wants a separate releasable buffer for shocks unrelated to property, and a higher legal cap on the sectoral CCyB, so that room exists before it is needed. The speaker is explicit that the buffer should not be raised now — the argument is about optionality, not about the current level of risk.

The case rests on the cost of the alternative. Leaning against the wind with the policy rate is expensive, and the Tinbergen separation principle argues for assigning each instrument to one target; the SNB therefore keeps the rate for prices and the buffer for the credit cycle. The historical record it cites is Swiss: activation at 1% in 2013, an increase to 2% the following year, and the 2022 move to the cap. The cost of getting this wrong is quantified from crisis experience — a peak loss of GDP per capita of about 1.5%, and GDP around 4% lower five years after a financial crisis. The tightening cycle itself is the counter-example on the rate side: the SNB raised the policy rate by 250bp within a year from mid-2022, a pace that would have been hard to justify on financial-stability grounds alone.

The structural backdrop has changed. UBS's share of the domestic deposit and loan market rose to around a quarter in 2024, against 16% for deposits and roughly half that level for loans in 2022, so the banking sector is both more concentrated and larger relative to GDP. The Credit Suisse crisis exposed weaknesses in capital requirements and in collateral preparation, and the SNB's preferred remedy is full CET1 backing of foreign participations as a targeted, proportionate too-big-to-fail measure. On the borrower side, the existing toolkit is a set of minimum standards rather than a buffer: a minimum down payment of 10% for owner-occupied properties with linear amortisation over 15 years, 25% for investment properties with amortisation over 10 years, and linear amortisation applying to the portion of a loan above a 66% loan-to-value ratio.

The counterargument the speaker concedes is that the sectoral CCyB has been doing two jobs — building resilience against mortgage and real estate vulnerabilities, and holding releasable capital for shocks with nothing to do with property — and that this double role is problematic. Switzerland is also unusual in having a strict legal cap, and it has not adopted a positive cycle-neutral CCyB the way many countries did after the pandemic. The market mechanism follows from the cap: with the buffer pinned at 2.5% since January 2022, any renewed vulnerability build-up has to be met either by the policy rate or by the Federal Council's too-big-to-fail capital package, which shifts the adjustment onto UBS's balance sheet and, through it, onto Swiss credit supply.

Isabel Schnabel's message to the BMWE Symposium on European Competitiveness is that macroeconomic, fiscal and financial stability in the euro area are not separate problems but one balance sheet: the same sovereign that must absorb defence, interest, ageing and climate costs also stands behind the banks. The good news in her account is that decisive monetary policy brought inflation back to target without a recession, and that sound regulation left euro area banks with a median core capital ratio of 21% and a weighted-average return on equity of 16% in 2026-Q2, so they can absorb shocks rather than transmit them. The bad news is the fiscal arithmetic. Public debt sits at 160% of GDP in 2025, government interest expenditure at 3% of GDP, and defence spending is on a path toward 5% of GDP, while the maturity structure of outstanding debt is shortening and government bill issuance reached 70% of gross borrowing in 2025. Schnabel's conclusion is that debt stabilisation therefore requires large fiscal adjustments, not because markets are forcing them today but because the adjustment need compounds the longer it is deferred.

The flexibility in the EU fiscal framework is real but bounded, and Schnabel quantifies both sides. The national escape clause allows extra defence spending of 1.5% of GDP, and the total allowed deviation from the Stability and Growth Pact requirements for a country such as Germany is 0.6% of GDP. The German fiscal package is the clearest live experiment: ECB-BASE simulations put its impact on German real GDP growth at 0.6pp in 2026, but the sign of the medium-term effect depends on composition — an investment-led package behaves differently from one where half the spending is government consumption, and a scenario with no additional consolidation behaves differently again. On the structural side, the EU's minimum renewable energy target of 42.5% by 2030 implies a linear investment trajectory that competes for the same fiscal space, while venture capital investment of 1.6% of GDP and roughly 30,000 patent applications in 2021 illustrate why Europe's innovation gap is not closing on its own.

The method matters for how much weight the projections carry. The debt paths to 2031 assume the IMF Spring 2026 Forecast, and the interest expenditure figures are a no-policy-change scenario in which the structural primary balance stays at its 2025 level except for expected ageing costs; the defence line assumes every euro area country activates its national escape clause in 2026 and raises spending linearly to 2028. Change any of those assumptions and the required adjustment changes with them. The counterpoint Schnabel herself raises is that potential growth is slowing, which makes any given debt ratio harder to stabilise, and that the shortening maturity structure means more of the stock reprices at market rates sooner — the channel through which a higher realised or implied deposit facility rate of 4% reaches sovereign budgets. For holders of euro area sovereign debt, the mechanism is straightforward: a larger structural primary balance adjustment, a higher bill share in gross borrowing and a shorter average maturity together raise rollover frequency and the sensitivity of debt-service costs to the level of rates, which is why the composition of the German package and the activation of escape clauses across member states are the variables to watch rather than the headline deficit alone.

Peter Kažimír, an ECB Governing Council member, called the September rate increase an appropriate response to enormous uncertainty and a persistent risk of higher inflation, saying it leaves room to react flexibly and that decisions stay data-driven and meeting-by-meeting. That is the same framing the Council used when it raised the three key rates, so the new element is his warning that the energy shock is lasting longer than expected and has not fully passed through, with the pressure rotating from oil and fuel into gas and electricity and food prices accelerating. For bond markets the channel is the wage and inflation-expectations leg: if food and power costs keep feeding pay demands, the room to react flexibly that he cites is room to hold or hike again rather than to cut, and the front end of the curve prices that optionality.

Euro area households hold about €440bn in US technology firms, ECB President Christine Lagarde said in Vienna on 14 September 2026, while AI-related borrowing accounted for about a quarter of the growth in credit to firms in the first quarter of 2026. The exposure sits on the household side of the euro area's balance sheet, not in its own capital stock, and the ECB puts the cost of closing Europe's data centre gap over the next decade as high as €600bn, with more than €100bn a year of the public share of those investment needs uncovered by existing budgets and EU instruments. Lagarde's answer is to route the financing through capital markets rather than banks, which points at the savings and investments union and at the pool of household savings Europe accumulates each year as the money to be redirected; the counterpoint she acknowledges is that a correction in the US AI buildout would reach Europe through exactly the holdings she is describing.

A Bank of Italy paper by Onofrio Panzarino and Antonio Perrella, using trade-level MTS data, documents that auto-hedging — algorithmically driven offsetting trades executed by market makers after their quotes are consumed — reached peaks of roughly two fifths of total market transactions at end-2024, against a share in the low single digits of overall market activity between 2016 and 2019. The strategy is overwhelmingly internal: the large majority of identified AH transactions are executed in the same bond, and a substantial minority incur zero execution cost. The paper's baseline logit (January 2023–April 2026) attributes variation mainly to time-invariant market-maker characteristics — dropping market-maker fixed effects collapses the R² from a moderate fit to almost none — while liquidity conditions matter modestly: an extra billion euros of quoted depth raises the predicted probability of auto-hedging by a fraction of a percentage point, and a sell-initiated trade raises it by a bit under two percentage points. That is the channel to the price of money: the marginal buyer of BTPs is increasingly hedging in the same instrument rather than absorbing inventory, so the depth that shows up in MTS quotes is partly reflexive, and the trade-size threshold above which AH probability falls significantly marks where that breaks. The finding sits in a Banca d'Italia working paper, not in the MTS market notices dealers actually read; the test is whether the AH share holds near its recent peak through the next BTP auction cycle or retreats toward the pre-2020 band, and whether the zero-cost share stays where it is once bid-ask spreads widen from their current level.

Malaysia's $1.5bn Global Sukuk, issued last month, drew more than $9.5bn in orders — a book several times oversubscribed — and priced at record-low spreads. The pricing channel runs through the sovereign's own curve: a heavily oversubscribed book at record-low spreads compresses the premium Malaysia pays relative to conventional dollar paper, and with Islamic financing at almost half of the domestic banking system, that spread sets the funding benchmark for the banks and corporates that issue into the same investor base. The number sits in a BIS note on Islamic finance rather than in any auction calendar or syndication league table, and the next Malaysian sovereign sukuk — a scheduled issuance whose pricing will show whether the record-low spread held or whether the heavily oversubscribed book was a one-off bid for scarcity.

An ECB working paper by Andrea Barbon, Jean Barthélemy and Benoit Nguyen, covering January 2021 to January 2026, finds that USD-pegged stablecoin deposit rates in DeFi lending protocols averaged above the Fed funds target rate, with occasional spikes far higher, but averaged below the policy rate from mid-2022 to end-2023. The decoupling is stark in the tightening phase: between March 2022 and July 2023 the Fed raised rates eleven times in a cumulative tightening of several percentage points, yet DeFi rates fell from a low single-digit percentage in March 2022 to near zero in September 2022. Around a median-sized policy hike, DeFi rates rose only a sliver after one week — a pass-through of roughly a third. The loosening phase ran the other way: between September and December 2024 the Fed cut while DeFi rates rose about six times as much, tripling over the same three months. The channel to the price of money is the one Danmarks Nationalbank flagged — a stablecoin complex worth hundreds of billions of dollars and tens of billions of dollars of total value locked in DeFi lending protocols as of September 2025 that prices dollar liquidity on its own schedule, so the Fed's marginal move reaches crypto collateral and, through it, dollar funding outside the banking perimeter with a lag and a sign of its own. The evidence is buried in an ECB working paper series number rather than in the FOMC's own materials; the settling test is the next policy step in either direction, where a pass-through still near a third of the policy move would confirm the decoupling rather than a sample artefact.

In a Handelsblatt guest contribution, Bundesbank board member Michael Theurer argues Europe must mobilise its savings to finance the climate transition, digitalisation and defence, and that its financial system must become more independent. Deregulating banks would be a mistake, he writes, because the resilience built on post-2008 reforms is what allows the sector to carry high and growing investment needs. He wants the EU bank rulebook simplified — overlaps removed, supervision streamlined — ahead of Commission legislative proposals expected in early 2027, following the Commission's 17 July bank-competitiveness communication. The channel is the rulebook itself: if the early-2027 proposals tilt toward the simplification Theurer describes rather than toward loosening capital requirements, the cost of bank funding and the capacity of European banks to hold sovereign and infrastructure paper stay where post-2008 rules put them. The argument sits in a German-language op-ed by a Bundesbank board member, not in the consultation documents, and the Commission's early-2027 legislative proposals are what will show which reading of "simplification" won.

The Bank of Russia's payment-system monitoring shows incoming payments in August 2026 running 2.6% above the average daily level of the second quarter of 2026, but the aggregate hides a sharp split. Excluding mining, refined petroleum products and public administration, receipts rose 5.6%, so the headline is dragged down by exactly the sectors the central bank singles out. Consumer-demand industries led with 6.4%, intermediate demand followed at 5.3%, and external-demand industries added 4.9%. Against that, public-demand receipts fell −8.5% and investment-demand receipts −6.8%.

The public-demand decline is not a broad phenomenon. Public administration alone fell −32.1% on a seasonally adjusted basis, and the central bank attributes the bulk of the group's contraction to that single activity. The group's year-on-year reading is −20%, and the July seasonally adjusted figure was already deeply negative before August. Investment demand is the second drag: −12% year-on-year, with the August seasonally adjusted change at −6.8%. Consumer demand, by contrast, is accelerating on both bases — 12.5% year-on-year against a much smaller July reading, and 6.4% seasonally adjusted after a modest July gain.

The method matters for how these numbers should be read. The data cover payments through the Bank of Russia payment system, classified by the recipient's main OKVED code as recorded in the EGRUL/EGRIP at month-end, with a sample running from January 2016 through 31 August 2026. Seasonal adjustment uses sparse Bayesian learning ridge regression, series are winsorised at the extreme upper tail of the distribution of non-zero payments, and the winsorisation boundary is re-estimated each half-year. Use-category weights come from the 2021 input-output tables, and an industry counts as belonging to a category if more than roughly a tenth of its output falls there, so industries can appear in several groups. The comparison base is the average daily level of the previous calendar quarter, with flows indexed to 2019.

The central bank itself flags that the perimeter covers less than half of all payments once LORO/NOSTRO accounts, intra-bank transfers and other payment systems are included, and that the methodology is still being revised, so retrospective estimates may change. Growth was also uneven within the positive categories: refined petroleum products and metallurgy held back external demand, agriculture held back both consumer and intermediate demand, and finished metal products weighed on intermediate demand. Excluding mining and public administration, the Central Federal District rose 5.2%, but the Southern, Siberian, South Caucasian, Ural and Volga districts declined. The market mechanism runs through the budget channel: public administration carries the dominant weight in the public-demand group, so a sustained −32.1% fall in its receipts implies fewer budgetary payments flowing through the central bank's system, while the 6.4% rise in consumer-demand receipts points to firmer household-sector cash turnover.