Concept

Fortress Balance Sheet

conceptbankingfinancerisk-managementstrategyinsurancemulti-source
Teach me this →

Fortress Balance Sheet

Fortress Balance Sheet is Jamie Dimon‘s term for the financial discipline he has practiced since his Primerica days and deployed at every institution he has run: hold more capital than required, maintain a real liquidity buffer, account conservatively even when aggressive accounting is legal, and stress-test to the worst outcomes ever recorded — not to regulatory minimums or recent averages. The fortress is not a static condition; it is a management culture that actively resists the pressures that lead banks toward leverage and fragility.

The five components

1. Conservative accounting. Dimon’s formulation: ‘you can drive a truck through accounting rules.’ The fortress choice is always the conservative path — spread revenues over time rather than front-loading, mark loans to what they are actually worth rather than what held-to-maturity accounting allows. Aggressive accounting produces temporary profits that conceal real losses; the Silicon Valley Bank collapse in 2023 was partly a consequence of held-to-maturity accounting masking interest-rate losses that had halved the bank’s tangible book value.

2. Capital above minimums. After acquiring Washington Mutual in September 2008 — one week after Lehman failed — JP Morgan immediately raised $11 billion of new equity it did not strictly need: ‘I don’t want to be short capital or liquidity.’ Excess capital is not dead weight; it is the precondition for acting in a crisis when competitors cannot.

3. Liquidity buffer. Dimon began stockpiling liquidity in 2006 when he saw early signs of subprime stress, well before anyone declared a crisis. Liquidity is the buffer between a run and a failure; Bear Stearns failed not because it was insolvent but because it ran out of liquidity in a single night.

4. Stress testing to historical worst-ever. When Dimon arrived at JP Morgan and reviewed the risk books, the high-yield stress scenario was a 40% credit-spread widening. The historical worst was 17%. He reset the scenario to worst-ever. In 2008, spreads hit 20% and the bond market stopped functioning. The principle: do not plan for the bad outcome people think will not recur — plan for the worst that has actually happened, then assume something worse is possible.

5. Incentive alignment. The fortress requires removing compensation arrangements that reward short-term risk-taking. At JP Morgan, there are ‘no winks, no nods, no side deals’ — no senior banker is paid on a particular position. The 2006 decision to eliminate profit-pool percentages on leveraged positions was part of what kept JP Morgan at one-third the leverage of peer investment banks going into 2008.

The compounding effect

The fortress imposes a real cost in good years. Dimon acknowledges that JP Morgan was ‘a little bit’ less profitable than peers from 2000 to 2007, when many large banks were earning 30% return on equity through leverage. Most of those banks went bankrupt or required rescues. The compound return over a full cycle, including the good years and the crisis, was dramatically better for the fortress. This is the same logic Morgan Housel describes in Compounding: the prerequisite for compounding is survival; voluntary or forced exits near the trough reset the clock to zero.

The fortress also creates an asset that only emerges in crises: the ability to act while others are frozen. Bear Stearns, Washington Mutual, and First Republic were all available at deep discounts because no one else had the capital, the liquidity, and the credibility to execute. The fortress balance sheet is not merely defensive — in banking, conservatism in good years converts to offensive capability in bad ones.

Barbalat: the insurance fortress is built on the asset side

Dimon’s five components are all liability-side and defensive — how not to blow up. Vlad Barbalat, who invests Liberty Mutual’s roughly $120 billion of capital with no third-party money in it, uses the same term for something built differently and aimed elsewhere.

He names two engines of profitability. Underwriting is one, and it is ‘a thin margin business’. The other is the asset side, and that is where the fortress actually comes from: ‘if you simply take the approach of let me buy a 4% investment grade bond or 5% investment grade bond versus try to achieve a 7, 8, 9, 10% return on the totality of your portfolio, it’s all the difference in the world.’ A bank builds its fortress by holding capital back; an insurer builds its fortress partly by putting capital to work.

The payoff Barbalat names is capacity rather than survival: ‘if you build a fortress balance sheet you’re able to do things that others will not.’ His boundary case is data centres — ‘a totally different scale of the asset and values than of an asset that have existed before’, so large that ‘insurance balance sheets aren’t large enough to just absorb that’, which is why third-party capital is flooding into the category. The fortress determines the opportunity set, on both the liability and the asset side.

This converges with Dimon on the conclusion — conservatism converts into offensive capability — while inverting the mechanism. Dimon’s fortress lets you buy Bear Stearns when nobody else can; Barbalat’s lets you underwrite and own asset classes as they emerge. The common thread is that both treat the balance sheet as the thing that decides what you are allowed to do, not merely what you can survive. See Vlad Barbalat on Permanent Capital, Branded Capital, and AI's Repricing of Risk.

Where mainstream views differ

Short-term shareholders value ROE in the current year. The fortress intentionally sacrifices ROE in good years to survive bad ones; shareholders who hold only in bull markets pay a real cost for the strategy and receive none of the crisis-acquisition upside.

Regulatory optimists argue that better regulation, not voluntary capital excess, is the right solution to banking fragility. Dimon’s implicit position is that regulations will always lag innovation in risk-taking and that management conservatism is the only reliable defence — regulators told banks that rates would stay low forever and contributed to the SVB interest-rate exposure that destroyed the bank.

Private credit represents the current version of the same dynamic Dimon has watched in every cycle: a new product category grows rapidly, contains actors who do not fully understand the risks, and hides leverage in ways that are not immediately visible. Dimon does not consider 2024-era private credit ($2 trillion) to be 2008-scale systemic risk, but he holds the same scepticism he applies to every new financial product that has ‘never been through a cycle.‘