Financial regulation is not a math problem that stays solved; a regulatory regime elicits responses from private firms who adapt to maximize returns within constraints by amending rules and finding practices within the letter but contrary to the spirit of the rules, placing any regime under continual assault so that over time its ability to prevent crises degrades.

Joseph Stiglitz's argument that market failure in the mortgage market justifies Fannie and Freddie ignores that the GSEs would have their own incentives to push the regulatory envelope; moreover, the failure to supply 30-year fixed-rate mortgages is not a true market failure like overfishing a commons, but merely a failure to produce the kind of mortgages Stiglitz prefers.

Because short-term inflation is low, it is not clear that real interest rates are actually low even though nominal rates are very low, undercutting the claim that the Fed has produced an enormous surge of liquidity.

Higher leverage (a higher proportion of the loan using other people's money) means a smaller change in the asset price puts the borrower underwater, which increases the value of the default option and should make that option expensive for lenders, raising interest rates if properly priced.

The decline in enforcement of recourse loans in the U.S. has caused interest-rate differentials between recourse and non-recourse states to disappear, because effective recourse requires courts willing to let banks garnish wages or seize assets, which has eroded.

Fannie Mae and Freddie Mac sustained the 30-year fixed-rate mortgage by being able to borrow long (issuing long-dated bonds to match their mortgage portfolios) far more cheaply than ordinary banks, thanks to their reputations and implicit government guarantee, while the taxpayers bore the tail risk of the default option (and likely prepayment option) becoming extremely valuable—which is what ultimately destroyed them.

The non-financial sector (households and entrepreneurs) wants to hold riskless liquid assets like checking accounts while issuing risky illiquid liabilities like consumer debt or long-term investments; the financial sector exists to take the opposite side, holding risky illiquid assets and issuing riskless liquid liabilities.

A bank that is insolvent will have negative net worth even if all assets are allowed to mature, whereas an illiquid bank would have positive net worth if it could continue but negative net worth if forced to sell assets immediately; this distinction is central to interpreting the financial crisis but is genuinely hard to determine because bank assets are difficult to value.

There is no clean controlled experiment on whether crisis banks were truly solvent because the Fed gave away profits to banks by paying interest on reserves, keeping short-term rates low, and buying assets to raise their market values, making their recovery uninformative about underlying solvency.

U.S. mortgages are peculiar because of valuable embedded options: a non-recourse default option (you can turn in the keys and walk away, and in many states the bank cannot pursue your other assets) that is rooted in state law and custom, and which is most valuable on low-down-payment loans because higher leverage makes it more likely the house value falls below the mortgage.

Because government policies subsidize the 30-year, low-down-payment, non-recourse, no-prepayment-penalty mortgage, its embedded options are mispriced; if the subsidy were removed, the interest rate on such mortgages would be high relative to a Canadian-style 5-year recourse rollover mortgage, and the market would greatly reduce or eliminate the 30-year fixed-rate mortgage.

The government can only sustain the valuable borrower options by providing subsidies to borrowers who exploit them; even if recourse loans were banned, the rate on non-recourse loans would be higher in a working market, and low-down-payment purchases (under 10% down) would carry a high interest-rate penalty that would discourage them.

The 30-year amortizing fixed-rate mortgage originated as a Great Depression policy response: borrowers were defaulting on five-year balloon mortgages with no principal amortization, so the government created the FHA and the savings-and-loan system (under the Federal Home Loan Bank Board) to promote amortizing, fixed-rate, 30-year loans that removed interest-rate risk from borrowers.

Savings and loans went bankrupt in the 1970s and 80s because they took on excessive interest-rate risk by funding 30-year fixed-rate mortgages with short-term deposits—a fundamental duration mismatch where rising rates make the institution lose money on both sides.

A demonstrably solvent bank can rarely have a genuine liquidity crisis because it can borrow against its assets from investors, the interbank lending market, or the Fed's discount window; the 2008 crisis is described as an interbank lending breakdown driven by distrust of mortgage-related asset values.

Banks provide economies of scale by spreading the cost of a vault across many customers, which is cheaper than each person building their own vault.

Banks provide three things individuals cannot replicate themselves: diversification (across both loans and depositors' withdrawal timing yielding a better risk-return tradeoff), specialized expertise in underwriting and servicing loans, and a reputation that induces depositors to trust them with money.

Reputation creates the volatility in finance because reputations can rise and fall; when reputations rise inappropriately, banks can take on more and more risk, which is roughly what happened in the crisis.

As time has passed it has become increasingly clear that there was far more insolvency in the crisis than was claimed at the time, with Fannie Mae and Freddie Mac shifting decisively toward the insolvency interpretation while Citigroup and Bank of America remain ambiguous.

The prepayment/refinance option is a one-way option that goes entirely to the borrower: if rates fall the borrower can refinance without penalty, but if rates rise the bank cannot force the borrower into a higher rate; this option is especially valuable on a 30-year loan because little principal is paid early and longer-dated options are worth more, and properly priced it would imply higher interest rates.