IMAGINE the most indispensable institution in human civilization.
Not hospitals. (though vital)
Not governments.
The answer is simpler and stranger: the bank. When the economy booms, banks fund the ambition. When it crashes, banks prevent the collapse. They are equally needed by the entrepreneur signing a term sheet and by the finance minister scrambling for a bailout. There is no “modern economy” without them.
So if demand for banking is essentially permanent, in booms, in busts, in panics, in prosperity then why don’t banks simply compound upward, like a well-behaved S&P 500 chart drawn by someone who has never experienced a bad day in their life?
Why, instead, do they surge, stagnate, implode, rebuild, and repeat, like a soap opera with leverage!
The Illusion: “Permanent Demand Means Permanent Growth”
Banks are needed in booms for credit expansion, M&A financing, and investment. They are needed in recessions for liquidity provision and corporate restructuring. And they are needed in full-blown crises as the lender-of-last-resort, coordinating the kind of desperate plumbing that keeps ATMs from going dark.
This feels like a growth investor’s dream that an industry with a structural moat, serving customers who literally cannot function without it. And yet, banking profits are famously cyclical, and hundreds of banks collapse every decade. From 2008 to 2011 alone, 411 U.S. banks failed, collectively holding over $668 billion in assets.
The reason for this paradox is: banks do not sell products, they intermediate risk. And risk, unlike a product, is not “linearly scalable.” The more of it you accumulate, the more non-linear and explosive its behavior becomes. This is the first crack in the illusion of permanent growth.
The Balance Sheet:
Banks grow through three primary mechanisms: expanding their loan book, increasing leverage, and engaging in maturity transformation (borrowing short, lending long). Each of these carries a hard ceiling.
- Capital Adequacy Is a Structural Brake:
Under Basel III, the global regulatory framework introduced after the 2008 crisis — banks must hold minimum capital ratios against their risk-weighted assets. Every new loan requires equity capital as a backstop. Capital is expensive; deposits are cheap. But regulators, sensibly, care only about the former.
Research by Torres-García, Ballesteros-Ruiz, and Villca-Condori (2020) in the Journal of Banking Regulation found that capital requirements “act as a significant financial accelerator in the presence of productivity and monetary shocks.” Loose translation: when the economy turns, the capital constraint doesn’t just slow growth, it amplifies the slowdown. As Repullo and Suarez (2013) showed, each percentage point increase in the capital ratio translates into roughly a 0.09 percent loss in steady-state output. Small numbers; enormous consequences at scale.
2. Risk Is Non-Linear by Nature:
Loan growth looks safe, right up until it doesn’t. In 2006, subprime mortgage defaults looked “diversified” across geography, borrower profiles, and credit tranches. CDO and CDS markets had ballooned to over $10 trillion, injecting massive liquidity on the assumption that defaults were uncorrelated. They weren’t.
By 2008, correlation went to 1. As economist A. Michael Spence observed: “when formerly uncorrelated risks shift and become highly correlated… diversification models fail.” The financial system’s losses reached trillions of dollars globally. Twelve of the thirteen largest U.S. financial institutions, according to Fed Chair Ben Bernanke, were at risk of failure during 2008.
Interest Rate Cycles:
Bank profitability depends critically on one number: Net Interest Margin (NIM) — the spread between what banks earn on loans and what they pay on deposits. When central banks tighten monetary policy, deposit costs rise quickly while loan yields adjust slowly. Margins compress. And when the duration mismatch is severe, the result is not compression but catastrophe.
Silicon Valley Bank is the case study for the ages. SVB had accumulated $191 billion in deposits by 2021, largely from venture-backed tech firms. It invested heavily in long-duration U.S. Treasuries and mortgage-backed securities, betting on persistently low rates. When the Federal Reserve raised the federal funds rate from 0%–0.25% in March 2022 to 4.75%–5% by March 2023, the value of those long-duration assets collapsed. On March 8, 2023, SVB disclosed a $1.8 billion loss on the sale of $21 billion in securities. By March 9, clients withdrew over $40 billion in a single day. On March 10, regulators seized the bank. The 16th largest bank in the United States was gone in 48 hours.
More than 85% of SVB’s securities portfolio was invested in assets with maturities exceeding five years — compared to a 20% average for peer U.S. banks. The Federal Reserve’s own review later noted that “contagion from the firm’s failure posed systemic consequences.” Duration mismatch is not a quirk of poorly run banks. It is an inherent feature of banking itself. SVB merely ran it to the extreme.
Banks Price Risk:
This is perhaps the most underappreciated truth in finance: banks are fundamentally different from technology companies in the mathematics of their growth.
Apple Inc. scales intellectual property, a piece of software, once built, costs virtually nothing to distribute. JPMorgan Chase scales exposure, every additional dollar lent requires new capital, new risk assessment, and new counterparty relationships. Technology growth compounds innovation. Bank growth compounds leverage. These are very different mathematical functions, and only one of them has a natural ceiling.
The data bears this out painfully. Pre-crisis, global banking sector Return on Equity (ROE) averaged around 13%. Post-crisis, following the tightening of Basel III capital requirements and regulatory reforms, it fell to approximately 9% globally, and in Europe, even lower. McKinsey’s Global Banking Annual Review (2021) found that the industry’s average ROE fell from 8% in 2011 to 6% in 2020. More brutally, the cost of equity for most banks exceeds their return on equity, meaning they are, technically, destroying value even as they “grow.”
The Political Paradox:
Banks are essential to every government, backstopped by every government in crisis, and despised by most governments in between. After each financial shock, the political response is entirely predictable: regulations tighten, capital requirements rise, executive compensation faces scrutiny, M&A activity gets blocked, and the industry’s ability to generate excess returns shrinks by design.
The Basel III countercyclical capital buffers, introduced after the 2008 crisis, are a direct expression of this dynamic. Research from the Cleveland Fed (2018) found that while these buffers theoretically reduce procyclicality, in practice their countercyclical impact may be small, because banks are deeply averse to constraints on earnings distributions, and regulators are reluctant to raise buffers frequently during expansions. The result is a system designed to be stable but structurally capped.
Regulation is not the enemy of banks. It is the inevitable consequence of their systemic importance. And that importance has a price: exponential growth is off the table. Banks are too big to fail and, precisely because of that, too regulated to compound.
Credit Has a Ceiling:
There is a hard economic truth that no amount of financial engineering can override: banks cannot sustainably lend faster than the underlying economy grows. Credit, ultimately, must be repaid from productive activity. When credit outpaces productivity, you do not get growth, you get a bubble, followed by deleveraging, followed by a generation-long hangover.
Schularick and Taylor (2012) documented that bank lending in advanced economies quadrupled from about 30% of GDP in 1945 to 120% of GDP right before the 2008 global financial crisis. That trajectory was not growth. It was accumulation of a debt stock that ultimately required a violent correction.
Japan provides the most sobering case study in modern economic history. After the asset bubble burst in 1990, when stock prices had tripled and land prices had surged to absurd multiples during the late 1980s, the banking system entered a condition of prolonged, almost theatrical paralysis. From 1991 to 2003, Japan’s GDP grew just 1.14% annually. Banks wrote off a cumulative $318 billion (¥37.2 trillion) in non-performing loans, but new ones appeared so fast the total barely moved. By the mid-1990s, zombie firms (companies surviving only on subsidized bank credit) exceeded 25% of all listed companies.
As the ADB Working Paper (2015) on Japan’s lost decades noted, the cumulative output lost in the three years following the 1997 banking crisis alone could be as large as 18% of GDP. Japan’s banks survived. They simply did not grow — for three decades. Banks that lend beyond productive capacity don’t just stagnate. They slowly fossilize.
The Deeper Truth:
All of this converges on a thesis that sounds almost paradoxical when stated plainly that banks don’t fail because they are unnecessary. They stagnate because they are too necessary.
Their systemic importance forces regulation. Regulation forces capital buffers. Capital buffers constrain leverage. Constrained leverage limits returns. Lower returns discourage equity investment. Less equity capital slows loan book expansion. And slower growth, paradoxically, is exactly the design intent because the alternative is the cycle described above: surge, crisis, collapse, rebuild.
Every banking cycle through history has followed the same arc: credit expansion, asset inflation, shock, deleveraging, regulation, slow rebuild. BIS research by Claudio Borio on the financial cycle has documented this rhythm across decades and geographies. It is not a bug. It is a feature of what it means to intermediate risk at the heart of a capitalist economy. Banks oscillate around macro cycles precisely because they are embedded in them.
Conclusion:
Because growth in banking is not like growth in technology, or consumer goods, or pharmaceuticals. It is not driven by “innovation compounding on innovation.” It is driven by risk accumulation and risk, unlike software, eventually bites back.
Banks cannot outgrow the economy, because their loans must be repaid from the economy’s output. They cannot defy interest rate cycles, because their very business model is built on the spread between short and long rates. They cannot escape regulation, because their failure is a systemic event, not a private one. And they cannot ignore correlation risk, because when a crisis hits, diversification — the great financial comfort blanket — tends to vanish precisely when it is most needed.
“Banking is very good business if you don’t do anything dumb.”
— Warren Buffett
Why Banks Don’t Just ALWAYS Grow? was originally published in DataDrivenInvestor on Medium, where people are continuing the conversation by highlighting and responding to this story.
