SAN FRANCISCO, March 5 (Reuters) - Toward the end of her stint as chair of the Federal Reserve, Janet Yellen assured markets that a flat U.S. Treasury yield curve - noted with worry by some of her colleagues - was simply a benign reflection of a new normal in the U.S. economy.
But two recent papers, one from researchers at the Fed's Board of Governors in Washington, and one from economists at the Federal Reserve Bank of San Francisco, show the debate still simmers at the central bank, now led by Yellen's successor Jerome Powell.
Historically, an inverted yield curve - where long-term Treasury yields are lower than short-term ones -- has in nearly all instances been followed by an economic recession. (For a graphic please see http://reut.rs/2zVhQyJ)
Last year, as the yield curve flattened despite the Fed raising interest rates, some policymakers including St Louis Fed President James Bullard worried publicly that the flat curve could be prelude to an inversion, and thus a recession.
But Yellen and others argued that because long-term growth prospects are now lower for the United States than before because of the aging of the population, and because interest rates are already low, yield curves will necessarily be flatter than in the past.
In a paper released on Friday, the Washington-based economists offered a view that bolsters that point, saying that while taken by itself the flat yield curve suggests a higher probability of recession, incorporating information from corporate bond spreads and surveys of economists among other data points to a "much lower" probability of recession ahead.
Indeed the New York Fed's most recent survey of Wall Street banks showed they saw just a 10 percent chance of a recession in the next six months, unchanged from recent surveys.
But research published Monday from the San Francisco Fed argues that concerns about the yield curve shouldn't be dismissed.
"An extensive analysis of various models leads us to conclude that the term spread is by far the most reliable predictor of recessions, and its predictive power is largely unaffected by including additional variables," wrote Michael Bauer and Thomas Mertens in the San Francisco Fed's latest Economic Letter.
(Reporting by Ann Saphir, with additional reporting by Richard Leong Editing by Phil Berlowitz )