Patricio Goldstein

We interpret the 1971–1973 collapse of Bretton Woods as a sequence of reverse speculative attacks driven by the incompatibility of asymmetric monetary policies, capital mobility, and fixed dollar parities. In a two-country model with faster US domestic credit growth and no US foreign-exchange intervention, the peg survives only while foreign central banks absorb the excess supply of dollars as international reserves. A ceiling on their dollar holdings triggers the attack: reserves accumulate abroad rather than run down in the US. The model matches the crises’ timing and reserve and inflation dynamics; archival evidence supports this interpretation, which extends to the 1992–1993 EMS crises.

Governments have long used sovereign lending as a tool of economic statecraft, conditioning finance on foreign policy concessions. This paper studies how debt contracts are designed when lending serves geopolitical objectives, and what this implies for the allocation, terms, and timing of official credit. We develop a model in which an official lender offers take-it-or-leave-it contracts to a cross-section of borrowing countries, with concessional terms and pro-poor allocation of credit emerging endogenously from the interaction between economic and geopolitical motives. Empirically, we use large language models to identify foreign policy concessions, events that benefit one state and impose a cost or policy shift on another, in a corpus of bilateral treaties drawn from national and UN registries and in the published record of American diplomacy. We show that concessions increase with official lending and use the model to price them from observed loan terms.

This paper studies the cognitive costs of inflation. Households pre-commit to expenditure shares across products before making purchases (‘mental budgeting’) and choose those shares using imperfect memories of previously paid prices. When purchases are infrequent, inflation between purchase occasions distorts perceived relative prices. This generates persistent expenditure misallocation even when prices are fully flexible and inflation is steady. We derive a closed-form second-order approximation for the consumption-equivalent welfare loss and calibrate the key sufficient statistic using household purchase microdata and evidence on perceived inflation. The implied welfare costs of 10 percent steady inflation can be up to 4.5 percent of consumption expenditures, significantly above standard estimates from money-demand and nominal-rigidity channels.