Global Edition · Version 6 · 10 Pages · 31 Sources

Banking Fundamentals

Plain language · Everyday analogies · Global examples · Mercatus Center research · Verified sources

Page 1 of 10What is a bank?

What a bank actually does

A bank borrows money from depositors at a low interest rate, lends it to borrowers at a higher rate, and keeps the difference as profit. You — the depositor — are the real lender. The bank is the trusted middleman managing thousands of people's money simultaneously.

Everyday analogy

A trusted neighbour collects grain from 100 families, stores it safely, and lends it to those who need it — returning it with a little extra. Most families never need their grain back at the same time. The system works — until fear changes that.

Bank run

A bank run occurs when too many depositors try to withdraw simultaneously — triggered by fear or rumour. Banks only hold a fraction of deposits in cash. They cannot pay everyone at once. The run can collapse a bank — but research shows it almost always starts because fundamentals were already weak long before.

The Big Analogy — The Rotten Tree

🌳

A bank run is the wind that knocks the tree over. Everyone blames the wind. But the roots were already rotten — for years. The wind just finished what the rot started.

Healthy years: Deposits grow, loans are made, profits look good. Nobody checks the roots.

Rot builds: Bad loans accumulate. Risk is hidden. Earnings managed. Roots weaken — silently, over 3 to 5 years.

Wind arrives: A rumour. A rate hike. Depositors run. The tree falls. Everyone calls it sudden.

It was never sudden. The rot decided the outcome. The wind just picked the date.

Historical bank run — crowds outside a bank with signs reading We want our deposits back

A historical bank run — depositors queue outside a bank demanding their money back. The sign reads "We want our deposits back! We're ruined!" — the panic that MIT research shows was almost always preceded by years of deteriorating fundamentals.

The Digital Bank Run — SVB collapsed in 48 hours infographic

The digital bank run: Silicon Valley Bank collapsed in 48 hours in 2023. In the digital age, a bank run takes hours, not days — making sound fundamentals more critical than ever.

Three global examples

🇺🇸 Silicon Valley Bank, 2023 $42 billion withdrawn in 48 hours. The rot: interest rate risk ignored for years, no Chief Risk Officer for 8 months.

🇬🇧 Northern Rock, 2007 First British bank run in 150 years. The rot: dangerously fragile short-term wholesale funding built deliberately into its business model.

🇦🇷 Argentina, 2001 Bank runs swept the nation. The rot: a decade of unsustainable sovereign debt — visible for years before depositors panicked.

Mercatus Center — Verner (2024)

Runs on otherwise healthy banks are not a common cause of failure. The root cause is almost always problems on the asset side — losses that lead depositors to run because they are responding to those losses, not causing them.

Sources

[2] FDIC. SVB Material Loss Review. 2023. · [3] Bank of England. Northern Rock. 2008. · [4] IMF. Argentina FSSA. 2001. · [5] Verner, E. Mercatus Center Macro Musings. Nov 2024.

NPA — Non-Performing Asset (NPL globally)

A loan where the borrower has not paid interest or principal for more than 90 days. It sits on the bank's books as an asset — but it is not one in reality. Banks can carry thousands of silent loans for years before anyone outside notices.

Everyday analogy

You lent $1,000 to a friend three months ago. He has gone silent. That money still appears as your asset — but it isn't coming back. Now imagine a bank holding millions of such silent debts, quietly, for years.

Mercatus Center — Verner (2024)

Rising NPAs are the earliest detectable signal of a bank heading toward failure — visible 3 to 5 years before collapse. Accumulating bad loans is not a bad quarter. It is the beginning of the rot.

Provisioning — the rainy-day fund for bad loans

Provisioning is when a bank sets aside money from its profits to cover loans that might go bad. Every time a bank gives a loan, it also puts a small percentage into a separate provision account. If the borrower defaults, the bank already has money to absorb the loss — before the loss arrives.

Everyday analogy

You lend $1,000 to a friend. You know there is a 10% chance he might not pay back. So you put $100 in a separate envelope labelled "just in case." That $100 is your provision. If he pays back, you keep it. If he defaults, you use it. The bank does exactly this — across millions of loans simultaneously. The question is whether it is setting aside enough.

Global examples — provisioning in practice

🇺🇸 Citigroup, 2008–09 Forced to set aside over $40 billion in provisions — taken directly from profits, pushing the bank into reported losses. Painful and public. But honest. The provisioning was the mechanism of recovery. The bank survived.

🇺🇸 JPMorgan, 2020 Set aside $15.7 billion in provisions in a single quarter during COVID-19 — the largest in its history. Strong coverage ratios maintained in good years meant it absorbed the shock and released provisions back to profit by 2021.

🇮🇹 Monte dei Paschi, 2013–16 Carried NPLs exceeding 35% of its loan book — but provisions consistently fell short. The gap was visible to analysts for years. The crisis was not a surprise. It was a disclosure. Required repeated EU bailouts.

Types of provisions

Type	Definition	When made
Generic	Set aside at loan approval, before any default	At loan origination
Specific	Set aside after a loan is identified as impaired	After 90 days of non-payment
Collective	Set aside for unidentified losses across a portfolio	Based on historical loss experience

IFRS 9 — the global provisioning standard

Before 2018, banks only provisioned after a loan had gone bad. After the 2008 crisis, regulators required banks to provision for expected losses — not just confirmed ones.

Stage 1Loan performing normally. Provision for 12-month expected credit losses only.12-month ECL

Stage 2Credit risk has increased significantly — but not yet in default. Provision for full lifetime expected losses.Lifetime ECL

Stage 3Loan is credit-impaired — borrower has defaulted or is close. Full lifetime expected losses provisioned immediately.Full lifetime ECL

Why IFRS 9 matters

Forces banks to recognise deterioration before default — making balance sheets more honest and crises more predictable. Adopted across EU, UK, Australia, Canada, and most major economies. The US uses CECL (Current Expected Credit Losses), adopted from 2020.

Key metrics & global PCR comparison

Provision Coverage Ratio

Provisions ÷ Gross NPLs × 100

Net NPL

Gross NPL − Provisions

Credit Cost

Provisions ÷ Avg Loans

BankPCRWhat it means

🇺🇸 JPMorgan Chase~87%

Very strong buffer

🇬🇧 HSBC~77%

Good coverage

🇩🇪 Deutsche Bank~72%

Adequate — improved post-restructuring

🇮🇹 Monte dei Paschi~55%

Weak — required EU bailouts

How provisions kill bank profits — the math

Bank lends $100 million to a corporate borrower.

Loan shows stress (Stage 2 under IFRS 9). Bank must provision lifetime expected losses → $20 million set aside.

That $20 million is recorded as an expense — taken directly from the profit and loss account.

If the bank earned $25 million that quarter, reported profit falls to just $5 million.

If provisions exceed profit — the bank reports a loss. It is not failing. It is being honest.

This is exactly what happened to Citigroup, Deutsche Bank, and many European banks in 2008–09. Their underlying businesses were often viable. Their provisions were massive — because their loan books had genuine problems that could no longer be hidden.

Credit rating vs provisioning — a summary

Credit Rating

What it measuresLikelihood of default by external assessment

Who does itIndependent agencies — Moody's, S&P, Fitch

Warning signRating downgrade — public and immediate

Global exampleCredit Suisse — downgrades accelerated collapse

Provisioning

What it measuresMoney set aside internally for expected losses

Who does itThe bank itself — an accounting requirement

Warning signSudden spike in provisions — internal and early

Global exampleMonte dei Paschi — under-provisioning before bailout

MIT Research Lesson

A bank's provisioning level is one of the earliest warning signs of trouble — visible years before collapse. When provisions rise sharply, loans are going bad faster than expected. When provisions stay suspiciously low despite rising NPLs, the bank is hiding its problems. Investors who watch provisioning ratios can see the rot 3 to 5 years before the bank fails. The bank run is not the cause of death. It is the moment the body stops pretending to be alive.

Sources

[6] IMF. Global Financial Stability Report: NPLs. 2022. · [7] EBA. NPL Guidelines. 2018. · [8] ECB. Monte dei Paschi Data. 2016. · [9] IMF. Greece FSAP. 2015. · [25] Correia, Luck, Verner. Failing Banks. QJE 2026. · [30] IFRS Foundation. IFRS 9. 2018. · [31] FASB. CECL. 2020.

The five things that determine bank health

Banking Fundamentals are not complicated — the five pillars of bank health

Banking Fundamentals are not complicated. There are only five things that determine bank health: Asset Quality (NPAs), Capital Adequacy (CAR), Profitability, Liquidity, and Underwriting. MIT research confirms all five are measurable — and all five deteriorate years before a bank fails.

MIT Research Connection

Verner's research confirms that all five of these fundamentals are detectable in publicly available financial statements — 3 to 5 years before a bank fails. None of them hide invisibly. They are simply ignored.

Liquidity

A bank's ability to meet immediate obligations. A bank can be profitable on paper and still collapse overnight if it cannot convert assets to cash fast enough.

Everyday analogy

You own a house worth $500,000. But if you cannot pay today's $200 grocery bill, you are illiquid. Wealth on paper and cash in hand are fundamentally different.

Mercatus Center — Verner (2024)

If a bank is already insolvent due to bad assets, liquidity support merely delays the inevitable and transfers the cost to taxpayers. The distinction between liquidity and solvency crises has enormous policy consequences.

Global examples

🇺🇸 Lehman Brothers, 2008 $639 billion in paper assets. When overnight lenders stopped, Lehman had no liquid cash. Filed for bankruptcy within 24 hours.

🇺🇸 Bear Stearns, 2008 Acquired by JPMorgan at $2 per share against a recent price of $170 — because no counterparty would lend overnight.

Sources

[11] Financial Crisis Inquiry Commission. Final Report. 2011. · [12] FRBNY. Bear Stearns Overview. 2008. · [5] Verner, E. Mercatus Center. 2024.

Market capitalisation

Total value of a bank's shares as priced by investors. A collapsing market cap is often the earliest public warning that trust is eroding — usually before regulators have noticed.

Everyday analogy

Your business is valued at $1 million on paper. But buyers only offer $400,000. The market is telling you something your accountant isn't. Markets are often right — and almost always earlier.

Global example

🇨🇭 Credit Suisse, 2021–23 Market cap fell from $30 billion to under $8 billion in two years — pricing in governance failures long before regulators acted. The market saw the rot first.

Sources

[13] UBS. Acquisition of Credit Suisse. March 2023. · [14] Bloomberg. Credit Suisse market cap history, 2021–2023.

Basel norms are international banking regulations set by the Bank for International Settlements in Basel, Switzerland — telling banks globally how much capital to hold, how much risk to take, and how to manage liquidity.

Basel I — 1988: The first global safety net

For every $100 of risky loans, hold at least $8 of your own capital. Simple — but it treated all loans as equally risky.

Everyday analogy

A building code: walls at least 8 inches thick — whether it is a garden shed or a 40-floor skyscraper. A start, but clearly insufficient.

Source

[15] Basel Committee. International Convergence of Capital Measurement (Basel I). BIS, 1988.

Basel II — 2004: Risk-sensitive rules

Introduced risk-weighting — riskier loans require more capital. The critical flaw: banks could use their own internal models to assess risk, which many manipulated to minimise capital requirements.

Everyday analogy

Code now requires thicker walls for taller buildings — but architects get to decide how tall their own building is. Some declared 10-floor buildings as 3-floor cottages.

Global example

🇺🇸 Lehman Brothers, 2008 Used internal Basel II models to classify toxic mortgage securities as low risk. The models were catastrophically wrong.

Sources

[16] Basel Committee. Basel II Revised Framework. BIS, 2004. · [11] FCIC Final Report. 2011.

Basel III — 2010 onwards: Post-crisis overhaul

Raised capital requirements, introduced liquidity ratios ensuring cash for a 30-day stress period, and capped leverage. Most major economies adopted it from 2013.

Everyday analogy

After a city-wide collapse, the code is rewritten. Every building must have an emergency generator, a 30-day water reserve, and cannot be more than 10 times as tall as its foundations are wide.

Global example

🇺🇸 SVB, 2023 Mid-size US banks exempted from full Basel III stress testing after 2019 rule relaxation. Had stricter liquidity rules applied, SVB's risk would have been caught years earlier.

Mercatus Center — Verner (2024)

The Basel III debate — more capital or more liquidity — depends on whether you believe the solvency or liquidity view of bank failures. Verner's research supports focusing on capital adequacy and solvency as the primary preventive tool.

Sources

[17] Basel Committee. Basel III Global Framework. BIS, 2010. · [18] Federal Reserve. SVB Supervisory Review. April 2023. · [5] Verner, E. Mercatus Center. 2024.

Risk management

The system a bank uses to identify, measure, and control how much risk it carries — across loans, investments, currencies, interest rates, and operations. A bank without functioning risk controls is flying blind at altitude.

Everyday analogy

A responsible household tracks expenses, maintains an emergency fund, and never borrows beyond its repayment capacity. Risk management is that discipline — scaled to billions and governed by regulation.

Global example

🇺🇸 SVB, 2023 Had no Chief Risk Officer for 8 consecutive months before collapse — during which it massively increased exposure to long-duration bonds with zero hedging. The vacant seat was the warning sign nobody acted on.

Sources

[18] Federal Reserve. SVB Supervisory Review. 2023. · [19] FDIC OIG. SVB Material Loss Review. 2023.

Earnings management

When a bank manipulates how it reports profits — deferring recognition of bad loans or inflating income — to appear healthier than it is. MIT and Mercatus research identify it as one of the earliest signals of failure, often 3 to 5 years before collapse.

Everyday analogy

A family counts an unpaid loan from a relative as monthly income — so the budget looks healthy. Nothing has been falsified exactly. But the truth is being obscured from everyone, including themselves.

Global examples

🇩🇪 Deutsche Bank, 2008–16 Repeatedly delayed recognising losses on derivatives — reporting profits while regulators warned of hidden risks. Billions in fines followed.

🇺🇸 Washington Mutual, 2008 Recognised loan losses far later than prudence demanded — masking mortgage book deterioration until collapse was unavoidable.

Sources

[20] US DOJ. Deutsche Bank Settlement. 2017. · [21] US Senate. Wall Street and the Financial Crisis: WaMu. 2011.

Underwriting

The evaluation of whether a borrower genuinely deserves a loan — assessing income, assets, employment, and repayment capacity. Weak underwriting is the first domino in every major banking crisis. It is where the rot begins.

Everyday analogy

Before lending money to someone, a sensible person asks: do they earn enough? Have they repaid before? Can they afford this? Skipping these questions — for speed or volume — is poor underwriting. Someone always pays the price.

Global example

🇺🇸 US Mortgage Crisis, 2008 Millions of NINJA loans — No Income, No Job, No Assets — approved and sold globally. When borrowers defaulted, the damage spread across every continent.

Source

[22] Financial Crisis Inquiry Commission. Final Report: Mortgage Underwriting. 2011. pp.84–106.

Simple meaning

A grade given to a bank, company, or country telling you how likely they are to pay back their debts. Think of it as a report card for borrowing. AAA means "very safe — will definitely pay back." D means "already failed to pay."

Everyday analogy

You want to lend $1,000 to a friend. You check their credit rating by asking three questions: Have they paid back before? Do they have a stable job? Do they own anything valuable? Banks face the same scrutiny from global investors, every single day.

Two global examples — two different failure modes

Credit Suisse — downgrade as accelerant

For years, Credit Suisse reported profits while quietly absorbing trading losses and governance scandals. Rating agencies downgraded it repeatedly through 2022 and 2023. Each downgrade made it more expensive to borrow, drained client confidence, and tightened liquidity — a death spiral. By March 2023, UBS acquired it in a forced weekend merger at a fraction of book value.

Lehman Brothers — downgrade arriving too late

In the months before its 2008 collapse, Lehman's rating still showed investment grade. But the market had already stopped trusting it — overnight lenders refused to roll over credit. When the downgrade came, the bank had hours left. A credit rating tells you where a bank has been — not always where it is going.

Rating agencies & scale

Moody's

Founded 1909 · New York

S&P Global

Founded 1860 · New York

Fitch Ratings

Founded 1913 · NY & London

AAAInvestment GradeHighest quality — extremely strong capacity to pay

AAInvestment GradeVery strong capacity to pay

A / BBBInvestment GradeStrong to adequate — susceptible to adverse conditions

BB / BSpeculativeSignificant risk — vulnerable to default (Junk)

CCC–CSpeculativeHigh default risk — currently very vulnerable

DDefaultAlready in default — payments have been missed

How a rating downgrade affects a bank

Cost of borrowing

Lower rating → higher interest rates when raising money from markets. Each notch costs millions.

Depositor confidence

A public downgrade can trigger panic withdrawals — the very bank run the bank was trying to avoid.

Regulatory capital

Many regulations use ratings to determine how much capital banks must hold against certain assets.

Market access

Below investment grade (BB+ or lower) can shut a bank out of entire funding markets entirely.

Why ratings failed in 2008

The conflict of interest problem

Rating agencies were paid by the same companies whose bonds they rated — the issuer-pays model. Agencies gave AAA ratings to mortgage-backed securities full of bad loans. When housing collapsed, these were downgraded by 14 notches at once — from AAA to junk in months. The lesson: credit ratings are opinions, not guarantees.

MIT Research Lesson

A bank's credit rating does not suddenly collapse overnight. It deteriorates slowly as fundamentals weaken — rising NPLs, falling profitability, shrinking capital. The downgrade is not the cause of failure. It is the warning sign that everyone saw and nobody acted on.

Sources

[23] Moody's. Credit Suisse Rating History. 2023. · [24] US Treasury. TARP: Citigroup. 2009. · [25] Correia, Luck, Verner. Failing Banks. QJE, Feb 2026.

The core finding — 160 years, 5,000+ bank failures

After studying over 5,000 bank failures across 160 years and 46 countries, MIT economists reached one clear conclusion: bank failures are almost always predictable — detectable in declining fundamentals 3 to 5 years before collapse. These are not accidents. They are choices.

Early warning

Rising bad loans

Middle stage

Falling profits

Final stage

Run or forced exit

Three banks. Three continents. One story.

🇺🇸 Lehman Brothers Toxic asset buildup, zero underwriting → 2008 collapse. Triggered the global financial crisis.

🇺🇸 SVB Interest rate risk ignored, no risk officer → 2023 collapse in 48 hours.

🇨🇭 Credit Suisse Governance failures, hidden losses → 2023 forced merger with UBS.

In every case: the rot came first. The crisis came second. The headlines came last.

Primary MIT Source

[25] Correia, S., Luck, S., Verner, E. Failing Banks. Quarterly Journal of Economics, Vol.141, Issue 1, Feb 2026, pp.147–204. doi.org/10.1093/qje/qjaf044

Not all credit booms are equal

Verner's research across 117 countries since 1940 found that what matters is not the size of the boom but where the credit flows.

Bad boom — non-tradable sector

Real estate, construction, retail

Predicts growth slowdowns

Higher crisis risk

Lower future productivity

Good boom — tradable sector

Manufacturing, technology, exports

Predicts sustained growth

Lower crisis risk

Higher future productivity

Mercatus Center — Verner (2024)

"Ex ante, you want to focus on times of rapid lending growth and focus on what credit is financing." The warning signal is not the size of the boom. It is the sector the boom is feeding.

Source

[26] Müller, K. and Verner, E. Credit Allocation and Macroeconomic Fluctuations. Review of Economic Studies, Vol.91(6), Nov 2024.

Banking crises without panics

A banking crisis can devastate an economy even without a dramatic bank run. The quiet accumulation of bad loans and reduced lending does its damage slowly — long before any panic is visible.

Global example

🇯🇵 Japan, 1990s No dramatic runs or single catastrophic collapse. Banks quietly carried massive bad loans for years, suppressing lending — costing Japan nearly a full decade of economic growth.

Source

[27] Baron, M., Verner, E., Xiong, W. Banking Crises Without Panics. QJE, February 2021.

Financial crises and the rise of populism

Financial crises do not just damage economies — they reshape politics. When households face debt distress, they become receptive to populist parties that promise debt relief and frame the crisis as ordinary people versus banks and elites.

Case study — Hungary, 2008–10

By 2008, two-thirds of Hungarian household mortgages were denominated in Swiss francs. When the forint depreciated 23% in six months, household debt surged by 4% of national GDP overnight.

Research found approximately one-fifth of the total political shift rightward in Hungary can be attributed directly to foreign-currency debt exposure. The financial crisis did not just shrink Hungary's economy. It permanently reshaped its politics.

Mercatus Center — Verner (2024)

"The financial crisis was really an accelerant. Populist parties exploit the cleavage between debtors and banks — and conflict between debtors and banks has been a particularly fruitful way for them to do that."

Source

[28] Gyöngyösi, G. and Verner, E. Financial Crisis, Creditor-Debtor Conflict, and Populism. Journal of Finance, Vol.77(4), Aug 2022.

Bank run

When too many depositors withdraw simultaneously — usually triggered by fear. Almost always happens at banks already weakened by poor fundamentals.

NPA / NPL

Non-Performing Asset or Non-Performing Loan — a loan where the borrower has not paid for 90+ days. The earliest and most reliable warning sign of bank weakness.

Provisioning

Money set aside from profits to cover loans that may go bad. Higher provisioning = more honest and better prepared. Under-provisioning = hiding the problem.

Provision Coverage Ratio (PCR)

Total provisions ÷ Gross NPLs × 100. Tells you how much of the bad loan book is already covered. A PCR above 80% is generally considered healthy.

IFRS 9

The global accounting standard that requires banks to provision for expected credit losses — before the loan goes bad — adopted across the EU, UK, Australia, Canada, and most major economies from 2018.

Liquidity

The ability to pay obligations in cash today. A bank can be asset-rich but cash-poor — and that mismatch can trigger collapse overnight.

Market capitalisation

Total value of a bank's shares as priced by investors. A falling market cap often signals deteriorating fundamentals before regulators or management acknowledge it publicly.

Credit rating

An independent grade (AAA to D) of how likely a bank is to repay its debts. A downgrade raises borrowing costs, drains confidence, and can accelerate the very collapse it is measuring.

Basel Norms

International banking safety rules set by the Bank for International Settlements — covering minimum capital, liquidity, and leverage requirements. Basel III is the current standard.

Underwriting

The process of evaluating whether a borrower deserves a loan. Weak underwriting — lending without proper checks — is where every major banking crisis begins.

Earnings management

When a bank manipulates reported profits by deferring recognition of bad loans. MIT research identifies this as one of the earliest detectable signals of a bank heading toward failure.

Credit boom (good vs bad)

Not all credit booms are dangerous. Lending to the tradable sector (manufacturing, exports) builds productivity. Lending to the non-tradable sector (real estate, consumption) predicts future crises. The sector, not the size, is the warning signal.

[1]Bank for International Settlements. The role of banks in the economy. BIS Working Papers, 2021. bis.org BIS

[2]FDIC. Silicon Valley Bank: Material Loss Review. April 2023. fdic.gov SVB

[3]Bank of England. Northern Rock: Financial Stability and the Banking System. Quarterly Bulletin, 2008. bankofengland.co.uk

[4]IMF. Argentina: Financial System Stability Assessment. 2001. imf.org

[5]Verner, E. Beckworth, D. (Host). Emil Verner on Banking Crises, Credit Booms, and the Rise of Populism. Macro Musings, Mercatus Center. November 18, 2024. mercatus.org Mercatus — Primary

[6]IMF. Global Financial Stability Report: Non-Performing Loans. 2022. imf.org

[7]European Banking Authority. EBA Guidelines on the Management of NPLs. 2018. eba.europa.eu

[8]European Central Bank. Supervisory Banking Statistics: Monte dei Paschi di Siena. 2016. bankingsupervision.europa.eu MPS Italy

[9]IMF. Greece: Financial Sector Assessment Program. 2015. imf.org

[10]FINMA. FINMA's Supervisory Activity Report: Credit Suisse. 2023. finma.ch Credit Suisse

[11]Financial Crisis Inquiry Commission. The Financial Crisis Inquiry Report. US Government Publishing Office, 2011. fcic.law.stanford.edu Lehman / 2008

[12]Federal Reserve Bank of New York. Bear Stearns: Overview of the Rescue. 2008. newyorkfed.org

[13]UBS Group AG. Acquisition of Credit Suisse: Transaction Details. March 2023. ubs.com

[14]Bloomberg Terminal. Credit Suisse Group AG market capitalisation data, 2021–2023.

[15]Basel Committee on Banking Supervision. International Convergence of Capital Measurement (Basel I). BIS, July 1988. bis.org

[16]Basel Committee on Banking Supervision. Basel II: Revised Framework. BIS, June 2004. bis.org

[17]Basel Committee on Banking Supervision. Basel III: A Global Regulatory Framework. BIS, December 2010. bis.org

[18]Board of Governors of the Federal Reserve System. Review of SVB's Supervision and Regulation. April 2023. federalreserve.gov SVB

[19]FDIC Office of Inspector General. SVB Material Loss Review. September 2023. fdicig.gov

[20]US Department of Justice. Deutsche Bank AG Agrees to Pay $7.2 Billion. January 2017. justice.gov Deutsche Bank

[21]US Senate Permanent Subcommittee. Wall Street and the Financial Crisis: Washington Mutual. April 2011. hsgac.senate.gov WaMu

[22]Financial Crisis Inquiry Commission. Final Report: Mortgage Underwriting. 2011. pp.84–106.

[23]Moody's Investors Service. Credit Suisse Group AG Rating History. 2023. moodys.com

[24]US Department of the Treasury. TARP: Citigroup Transaction Report. 2009. treasury.gov Citigroup

[25]Correia, S., Luck, S., Verner, E. Failing Banks. Quarterly Journal of Economics, Vol.141, Issue 1, Feb 2026, pp.147–204. doi.org/10.1093/qje/qjaf044 MIT — Primary

[26]Müller, K. and Verner, E. Credit Allocation and Macroeconomic Fluctuations. Review of Economic Studies, Vol.91(6), November 2024, pp.3645–3676. Mercatus — Credit Booms

[27]Baron, M., Verner, E., Xiong, W. Banking Crises Without Panics. Quarterly Journal of Economics, February 2021. doi.org/10.1093/qje/qjaa034 Mercatus — Crises

[28]Gyöngyösi, G. and Verner, E. Financial Crisis, Creditor-Debtor Conflict, and Populism. Journal of Finance, Vol.77(4), August 2022, pp.2471–2523. doi.org/10.1111/jofi.13138 Mercatus — Populism

[29]Correia, S., Luck, S., Verner, E. Why Do Banks Fail? Three Facts About Failing Banks. Liberty Street Economics, FRBNY. November 2024. libertystreeteconomics.newyorkfed.org

[30]IFRS Foundation. IFRS 9: Financial Instruments. 2014, effective 2018. ifrs.org IFRS 9

[31]Financial Accounting Standards Board (FASB). ASU 2016-13: Measurement of Credit Losses (CECL). 2016, effective 2020. fasb.org CECL / USA