Financial Regulation in a System Context

Morris and Shin argue that financial regulation built around ensuring the solvency of each institution individually – the traditional, risk-based-capital logic of the Basel accords – cannot by itself ensure the stability of the financial system, because actions that make one institution safer can make the system as a whole more fragile. Using the March 2008 collapse of Bear Stearns, which met its regulatory capital requirements yet still failed, as their organizing example, they show that what actually destroyed the firm was a run on its liabilities (its cash fell from $18 billion to $2 billion in three days), not a shortfall of capital against risky assets. Using a global-games coordination framework (following Morris and Shin 1998, 2002), they formalize how a creditor’s incentive to withdraw funding depends on both the proportion of other creditors needed to stay in for the debtor to survive and the cost of guessing wrong, and show that holding more liquid assets reduces fragility through two separate channels: it makes debtor banks more able to meet withdrawals, and it makes creditor banks themselves less prone to pull funding pre-emptively. They further argue that an asset’s own credit risk and its systemic importance are distinct: a fully collateralized, short-maturity asset can look safe in isolation under Basel’s risk-weighting logic while still being highly systemically important, because its liquidation transmits a funding shock to whoever is on the other side of the repo chain. On this basis they argue for liquidity requirements that constrain the composition (not just the size) of a bank’s balance sheet, and for a maximum leverage ratio whose rationale is not the traditional “capital absorbs losses” logic but the goal of preventing the system-wide buildup of leverage during good times that makes a later increase in haircuts so destructive.

Unique Equilibrium in a Model of Self-Fulfilling Currency Attacks

This 1998 American Economic Review paper by Stephen Morris and Hyun Song Shin resolves the multiple-equilibrium indeterminacy of standard self-fulfilling currency-attack models by replacing common knowledge of economic fundamentals with a small amount of private, noisy information. In a model where a government defends an exchange-rate peg against a continuum of speculators who may each pay a transaction cost to short-sell the currency, perfect-information versions of the model have a “ripe for attack” range of fundamentals in which both “everyone attacks” and “no one attacks” are self-fulfilling equilibria, so the theory cannot say whether or when a crisis occurs. The authors show that once each speculator observes only a noisy private signal of the true fundamental state (rather than the state itself), the game has a unique equilibrium: there is a single threshold fundamental state below which the government always abandons the peg and above which it never does, pinned down (in the limit of vanishing noise) by the condition that the marginal speculator is exactly indifferent between attacking and not attacking. Because the equilibrium is unique, the model supports genuine comparative statics and policy analysis – transaction costs, the quantity of speculative (‘hot’) money, and above all the transparency of public information about fundamentals – which indeterminate multiple-equilibrium models cannot provide.

Stephen Morris | Macro Paper Warehouse

Financial Regulation in a System Context

Unique Equilibrium in a Model of Self-Fulfilling Currency Attacks