TB: Fragile by Design (Calomiris & Haber)
Core Thesis
Banking crises are not random natural disasters — they are political outcomes. A country gets the banking system consistent with the institutions that govern its distribution of political power. Banking is a partnership between government and bankers, and the terms of that partnership are set by political bargaining, not by technical efficiency criteria. The US has had 14 banking crises in 180 years; Canada has had none. This is not an accident.
Key Takeaways
Banking as politics
- Banks are an institutional mirror of the political system — a product of a society's deep history.
- A country does not "choose" its banking system; it gets a banking system consistent with its institutions governing the distribution of political power.
- Modern banking is best thought of as a partnership between government and a group of bankers. There are no fully private banking systems; the idea is a libertarian fairy tale.
- Government has a conflict of interest: it simultaneously regulates banks, borrows from them, and taxes them.
Why banking is inherently unstable
- A bank is inherently unstable: its inputs (deposits = IOUs to depositors) and outputs (loans = IOUs from borrowers) are both promises to repay.
- Any enterprise whose inputs and outputs consist primarily of debt contracts is inherently unstable and risky.
- Deposit insurance destabilizes the banking system: depositors no longer discipline bank insiders because their funds are no longer at risk. Insiders are then free to take riskier bets.
The US vs. Canada comparison
- Since 1840 the US has had 12 major banking crises; Canada has had none, including during the Great Depression.
- The difference: Canadian central government monopolized bank chartering (to manage French-English political tensions); US had unit banking laws (to protect small politically powerful bankers from competition).
- The FDIC was not created to protect depositors — it was created by small unit bankers to stifle branch banking competition.
- 1994 ended barriers to interstate banking; the US was finally positioned to have a stable banking structure. Then the subprime political bargain undid it.
The subprime crisis as political bargain
- The subprime crisis was the outcome of a political bargain: activists used congressional influence to impose HUD mandates on Fannie/Freddie; Fannie/Freddie accepted on condition of taxpayer guarantees.
- The US addressed inequality and urban poverty through bank regulation ("let them eat credit") while Western Europe expanded welfare programs.
- The problem was not lack of regulation — it was piles and piles of ineffectual regulation.
Democracy, populism, and banking
- Liberal democracy constrains majority rule to protect minorities; populist democracy gives majorities unlimited power. Banking crises cluster in populist political systems.
- The populist notion that unconstrained majority rule empowers the common man is based on a misunderstanding — populist governments consistently disadvantage the majority.
- Societies do not choose their banking systems in any meaningful sense. They get what their political institutions and dominant coalitions allow.
Historical patterns
- Chartered banks and nation-states co-evolved: banks aligned incentives of rulers, merchants, and financiers simultaneously.
- Paper currency was invented in Amsterdam to fight debasement of coins; governments quickly turned it into an inflation tax on the poor.
- The Hamiltonian formula: find the right moment to push measures that wouldn't pass in normal times, hoping the complexity of the legislative process prevents reversal later.
Mental Models
- Incentives Matter — the book is an extended study of how political incentives shape every banking outcome
- Complex Systems: Features from Path, Not Design — US unit banking persisted 150 years because of path-dependent political coalitions, not because it was efficient
- Second-Order Thinking — deposit insurance was designed to stabilize banks; the second-order effect was moral hazard that destabilized them