What will the next financial crisis likely look like?

The short answer

The honest answer to “what will the next financial crisis look like?” is that nobody knows the form, and the empirical record suggests that the form is precisely what is hardest to anticipate. What is more anticipatable is the mechanism. Reinhart-Rogoff (2009), Schularick-Taylor (2012), and the BIS literature converge on excessive credit growth — particularly in newly liberalized or under-regulated channels — as the most consistent precursor.

The intuition that “the next crisis will look like 2008” is comfortable but historically misleading. Each crisis tends to occur in the area where regulation has been adapted least to current dynamics. Post-2008 regulation tightened the bank channel, which is why much of the post-2008 risk accumulation has occurred outside it.

The complication is that mechanism-based prediction does not produce timing. Excessive credit conditions can persist for years, and a marginal event may or may not arrive to trigger the reversal. The signals provide elevated probability, not deterministic forecasts.

→ New to systemic risk frameworks? Macro-financial regimes pillar

What the data shows

The empirical record on post-2008 systemic risk accumulation has been compiled by central banks, the BIS, and the FSB.

The figures (FSB, BIS, IMF, 2008-2024):

• The FSB’s Non-Bank Financial Intermediation (NBFI) sector has grown to roughly half of total global financial assets (around $239 trillion at end-2022 according to FSB monitoring), up materially since 2008
• Global government debt to GDP rose from approximately 65% in 2007 to over 90% in 2024 (IMF), with several advanced economies above 100%
• Private credit (direct lending and other non-bank corporate lending) grew from approximately $250bn in 2008 to over $1.5tn in 2023 (Preqin estimates)
• The BIS credit-to-GDP gap indicator was elevated in several emerging markets in 2023-2024, particularly in Asia, signaling potential build-up of credit fragility

The 2023 SVB and Credit Suisse failures are an interesting interim signal. Both occurred in regulated banks but reflected interactions between rapid digital deposit flight, balance-sheet duration mismatches, and confidence loss — a partial preview of how speed and connectivity may shape future crises even within the regulated bank channel.

The exception worth flagging: cryptocurrency markets experienced their own 2022 crisis (TerraLuna collapse, FTX bankruptcy, Three Arrows Capital) that reached approximately $2tn in market value destruction. The crisis stayed largely contained from the regulated financial system, suggesting that “next crisis in crypto” may be self-contained rather than systemic — though regulators worldwide are reassessing as TradFi-crypto connections deepen.

→ Dataset: Credit spreads and recession risk dataset

Why it happens — the macro mechanism

Three structural channels explain where the post-2008 risk has migrated.

Channel 1 — Regulatory arbitrage to non-bank channels. Basel III and Dodd-Frank tightened bank capital, liquidity, and trading-book regulation. Activity that had previously sat on bank balance sheets — leveraged lending, certain mortgage origination, market-making in less liquid securities — migrated to non-bank channels (hedge funds, private credit funds, business development companies, money market funds). The FSB’s NBFI categorization captures the growth; the regulatory architecture for these channels is less developed than for banks.

Channel 2 — Sovereign debt and fiscal-monetary interaction. This is the angle most overlooked in classical bank-centric analysis. Government debt to GDP rose materially in advanced economies post-2008 and again post-2020. Central bank holdings of sovereign debt grew through QE programs. The interaction creates a fiscal-monetary nexus where central banks may face conflicts between price stability mandates and the cost of servicing government debt at higher rates. The 2022 UK gilt market episode (LDI pension-fund stress) showed how this nexus can produce stress even in advanced economies.

The corollary is that contrary to the narrative that the next crisis must look like 2008, several post-2008 mechanisms are structurally distinct and would produce different transmission paths.

Channel 3 — Dollar-funding channels in non-US economies. Eurodollar markets — dollar-denominated lending and funding outside the US — have continued to grow. Non-US banks, corporates and sovereigns hold material dollar liabilities while their assets and revenues are typically in local currencies. Cross-border dollar funding stress (a recurring feature of 2008, 2020 March, and several emerging market episodes) is a documented channel where regulatory remedies in the US do not solve the underlying mismatch.

Synthesis by regime: in pre-2008, bank-balance-sheet leverage was the dominant fragility, with the financial sector’s exposures to subprime structured products as the proximate trigger; in post-2008 to 2020, monetary expansion supported asset prices but accumulated risk in non-bank channels and sovereign balance sheets; in post-2022, the unwinding of monetary expansion has tested duration-mismatched institutions (SVB, LDI pension funds) and revealed vulnerabilities the regulatory framework had not anticipated. Three regimes, three sets of fragilities.

Crises do not repeat in form because regulation adapts to the previous one — they recur in mechanism, with credit growth in newly liberalized channels as the most reliable shared feature across centuries.

→ Framework: Systemic fragilities pillar

What it means for different economic actors

Long-term investors face the practical implication that “stress-testing on 2008” may miss the actual fragility patterns. Stress tests calibrated to the next crisis looking different from 2008 — including non-bank intermediary stress, sovereign debt repricing, and dollar-funding stress in non-US holdings — provide more conservative coverage than 2008-shaped scenarios alone.

Risk managers have a structural challenge that the regulatory framework does not fully solve. NBFI activity, private credit, and sovereign duration risk all sit partly outside bank stress-test perimeters. The 2022 LDI episode showed how interconnected these channels are: stress in pension-fund leverage transmitted to government bond markets and to the broader financial system within hours.

Policymakers and central banks face the question of whether the macroprudential framework built post-2008 covers the relevant fragilities. The FSB’s NBFI monitoring is a step; whether it is matched by tools to act remains contested. The 2022 BoE intervention in gilt markets and the 2023 BTFP (Bank Term Funding Program) were ad-hoc responses suggesting that the framework remains under construction.

A common error is to assume that because the post-2008 regulation tightened bank capital, banks are no longer the relevant fragility. The 2023 SVB and Credit Suisse episodes showed that even regulated banks can fail through interactions (deposit flight, duration mismatch, confidence) that the framework did not fully anticipate. Banks remain part of the systemic risk picture; they are not the entire picture.

Practical observation

Go deeper

Related questions

• Why is financial history the best teacher?
• What is financialization and why has it grown?
• What causes a Minsky moment?
• What is shadow banking and why does it matter?

Frequently asked questions

If we cannot predict the form, how is this analysis useful?

Predicting the mechanism does not give timing precision but does give probability assessment. The Schularick-Taylor 14-country, 140-year empirical work documents that excessive credit-to-GDP gap predicts subsequent banking crisis with statistical significance. This does not say “crisis arrives in 2026” but does say “elevated probability of stress within 5-7 years.” For long-term investors and risk managers, probability assessment with 5-7 year horizons is operationally useful even without specific timing.

Is the post-2008 regulatory framework sufficient?

The framework substantially reduced bank-balance-sheet vulnerabilities — Tier 1 capital ratios at major banks roughly doubled between 2007 and 2023. It did not prevent risk migration to non-bank channels. The 2023 SVB and Credit Suisse failures and the 2022 LDI episode showed that even within the regulated perimeter, interactions between liquidity, duration mismatch, and confidence can produce rapid stress. The framework is better than pre-2008 but is not complete, and the FSB’s continued work on NBFI suggests regulators recognize this.

What single indicator would be most worth monitoring?

The credit-to-GDP gap published by the BIS is the most empirically validated single indicator. Schularick-Taylor’s quantitative work demonstrated that this gap, when materially elevated above its trend for an extended period, has been the most consistent precursor of subsequent banking crisis across the long historical record. It does not provide timing precision, but it does provide a defensible probability signal that regulators, central banks, and investors can integrate into longer-horizon risk frameworks.

Last updated — 30 July 2026