The Hutchins Roundup brings the latest thinking in fiscal and monetary policy to your inbox. Have something you'd like us to include in the next Roundup? Email us and we'll take a look.
This edition was written by Chesapeake Dowdy, Chase Parry, Jack Spira, and Louise Sheiner
Danila Smirnov of the IMF examines how unexpected increases in the supply of U.S. Treasuries affect foreign interest rates and economic activity. To identify these surprises, he measures changes in Treasury yields over the 30-minute windows around scheduled announcements of upcoming Treasury auctions. Using data from 1998 to 2026, he estimates that a debt supply shock that raises U.S. 10-year Treasury yields by 10 basis points raises comparable foreign government bond yields by about 7 basis points within three business days. Despite the smaller rise in yields abroad, industrial production declines by roughly the same amount in the U.S. and foreign economies in response to U.S. debt supply shocks, with the effects abroad lasting longer. The international spillovers appear to operate primarily through higher global borrowing costs and reduced U.S. demand for foreign exports. Smirnov finds that declines in industrial production are larger in countries with more integrated financial markets, as measured by fewer restrictions on cross-border capital flows and greater foreign ownership of government debt. Both the declines in production and the increases in yields are greater in countries with greater risks to government debt sustainability and are smaller in countries with more fiscal space and stronger current accounts.
Paid subscriptions are an important source of revenue for AI providers, but little is known about how consumers choose among competing services. Mark L. Egan of Harvard and co-authors use debit and credit card transaction data from January 2021 through April 2026 to analyze demand for paid AI services. They find that only 2.9% of consumers were paying for an AI service by the end of their sample, with nearly all choosing the least expensive plans. Customer retention is also low. Only about 45% of OpenAI subscribers were still paying after 12 months, compared with about 70% of Netflix and Spotify subscribers. Consumers also appear willing to switch to or add competing services when providers release improved models, suggesting that switching costs are low. Using a structural model of the market for AI services, the authors find that consumers are very price sensitive and that AI services have become closer substitutes over time as providers have expanded their capabilities. These findings suggest that the market for AI services may be highly competitive with limited pricing power for providers.
Credit rating agencies say sovereign ratings reflect the likelihood that a government will default on its debt. Olivier J. Blanchard of MIT, Daniel Leigh of the IMF, and Prachi Mishra of Ashoka University develop a simple model to assess how debt, expected primary balances (budget balances excluding interest payments), and the gap between interest rates and economic growth affect default risk. They compare the model's predictions with ratings assigned by three major credit rating agencies to advanced and developing economies. They find that ratings put much more weight on current debt relative to forecasts of primary balances than their model does. Ratings are also much less sensitive than their model to the difference between interest rates and growth rates. Ratings also differ substantially across countries even after accounting for debt and expected primary balances. To illustrate the size of these country differences, the authors calculate the debt level at which each country would have an even chance of an S&P rating of AA or better, holding projected primary balances at zero.The estimated debt levels are 272% of GDP for Singapore and 248% for Germany, compared with 161% for the U.S. and 89% for Italy.The Baltic states wouldn't have an even chance of receiving an AA or better rating even if their debt were zero. The authors conclude that either their model misses something important about what drives default or the ratings agencies should revisit their methods.
“Since the ECB only has domestic counterparties in its refinancing operations, it is essential to complement the domestic liquidity framework with international liquidity lines. These lines, which consist of both repo and swap lines, mitigate adverse spillovers in the event of increased demand for euro liquidity abroad and/or impairment in cross-border euro liquidity flows. In the absence of international liquidity provision, such shocks might exert upward pressure on euro money market rates or disrupt the functioning of securities markets and the banking system, and thereby impair monetary policy transmission," says Philip R. Lane, member of the Executive Board of the European Central Bank.
"In recent years, the world economy has undergone profound structural shifts related to geopolitics and changes in the international financial system. These shifts have altered the dynamics of global trade and financial integration and suggest that the macroeconomic and financial environment will remain uncertain and potentially more volatile. More frequent financial disruptions and possible knock-on effects on euro area financial markets have the potential to hamper the smooth transmission of monetary policy.
"Given their backstop pricing, international liquidity lines are designed to be drawn down only in response to shocks and in crisis situations. However, their mere existence – even without being drawn upon – can help to prevent adverse market dynamics from arising in the first place. Empirical evidence suggests that the ECB’s liquidity lines have been effective in reducing financial pressures on euro funding markets."
