The Role of US Monetary Policy in Banking Crises Across the World

The historical record suggests US monetary tightening is dangerous for banks elsewhere – the early-1980s international debt crises followed the 1980-82 Volcker tightening, the Mexican Peso Crisis followed the 1994 Greenspan tightening – and this paper asks whether that danger is uniform or conditional. Its answer is that it is conditional, and on a specific thing: whether a country’s exposure to the United States is direct. Using an annual unbalanced panel of 69 countries (24 developed, 45 emerging) spanning 1870 to 2010, with systemic banking crises taken from Reinhart and Rogoff (2009) and yielding 239 distinct crisis starts, the authors estimate panel logit regressions in which the change in US short-term interest rates enters only through interactions with exposure measures – necessarily so, since a US variable that does not vary by country would be collinear with the time fixed effects. Direct exposure is measured by a country’s bilateral trade intensity with the United States, and for the post-1990 sample by its dollar-denominated debt liabilities net of assets; indirect exposure by overall trade openness and the Chinn-Ito capital account openness index. Both trade measures are also instrumented with gravity estimates built from distance, population, common language, borders, land area, landlocked status and colonial history, following Frankel and Romer (1999). For directly exposed countries the interaction is positive and significant, and economically meaningful: “a 1 % tightening in monetary policy increases the probability of a crisis by 1.0-6.8% for a given level of direct exposure to the United States.” For countries that are merely globally integrated the effect is ambiguous – the contemporaneous trade-openness interaction is actually negative and significant, which the authors read as openness providing diversification, those countries receiving funds flowing out of the directly exposed ones, or a more orderly reversal helping to correct accumulated imbalances – and “the impact diminishes when we correct for the endogeneity.” The channel is capital flows, and the paper adjudicates between two opposing possibilities: a tightening might lean against the wind and curb credit booms that would otherwise end in crisis, or it might trigger a sudden reversal of flows. “We find evidence that the latter effect dominates”: tightening significantly reduces portfolio flows to countries with direct US exposure but not to merely open ones, and where the adjustment is disorderly the crisis probability rises. Splitting the sample, the effect “is mainly an emerging market phenomena” – the interaction stays positive and significant in every subsample except the developed-country one. Results survive replacing the rate change with Gertler-Karadi, Rogers-Scotti-Wright and Romer-Romer shock series on the 1990-2010 window, OLS and probit estimation, adding local monetary policy and exchange rate changes, an alternative merged crisis database, and dropping the winsorization. The scope conditions are substantial: the exposure measures are themselves slow-moving country characteristics interacted with a common time-series shock, crises are rare events, and the baseline omits local monetary policy and exchange rates because including them shrinks the sample by three-quarters.

US or Domestic Monetary Policy: Which Matters More for Financial Stability?

When inflation is too low or unemployment too high, central banks cut rates; easier financial conditions improve balance sheets and encourage borrowing, but “this inevitably results in higher debt, which brings with it the risk of financial instability.” This paper asks whether prolonged easing raises financial vulnerability, and whether prolonged easing in the United States does so abroad. The design is deliberately not a shock-identification exercise: the key variable is the “duration” of easing, the number of consecutive quarters in which the eight-quarter moving average of a country’s nominal 2-year sovereign yield declines, and the authors state plainly that “we do not explicitly distinguish between systematic and unexpected monetary policy,” because firm leverage is a slow-moving variable responding more to policy expectations than to small surprises. Vulnerability is measured as market-value leverage – the market value of equity plus the book value of liabilities, over the market value of equity – for 988 publicly listed financial firms in 21 countries (18 advanced economies plus Brazil, Mexico and South Africa) from 1998Q1 to 2014Q4, split by the GICS classification into banks, insurance, real estate, asset management, investment banks and a residual category. In panel regressions with firm fixed effects, lagged macroeconomic controls and Driscoll-Kraay standard errors, one additional quarter of domestic easing raises banking-system leverage by 0.191 at the median, and eight consecutive quarters take a representative banking system from 10.5 to 12.0. The paper’s headline is what happens when US easing duration is added and US firms are dropped: the US coefficient is significant at the 1 percent level for every sector except investment banks, and at two years lifts non-US banking leverage from about 16.7 to 19.6, insurance from 8.2 to 9.0 and investment banks from 5.8 to 6.6 – effects “either equal to those of domestic monetary policy easing (for investment banks and asset managers), greater (for banks), or substantially greater (for insurance, real estate and other financial firms).” Repeating the exercise with the 2-year German Bund yield finds euro area spillovers “both economically and statistically very close to zero,” which the authors attribute to the euro’s far smaller role in global trade and finance – non-US banks issue about $15 trillion of dollar liabilities against only about €4 trillion of euro liabilities issued by non-euro-area banks. Cross-country variation lines up with three characteristics: spillovers are larger where financial development is higher, and smaller where trade openness and gross dollar liabilities are larger, because a weaker dollar mechanically shrinks dollar debt and lowers leverage. That dampening is real but partial – push factors dominate throughout, and firms mostly borrow further against the windfall, so “leverage appears to be pro-cyclical.” What the paper cannot claim is causal identification off exogenous policy surprises; it controls for what might have prompted the easing rather than instrumenting it, and notes that if policymakers ease in response to lower leverage the bias runs toward understating the effect.

IMF Economic Review | Macro Paper Warehouse

The Role of US Monetary Policy in Banking Crises Across the World

US or Domestic Monetary Policy: Which Matters More for Financial Stability?