Do budget deficits affect long-term interest rates? U.S. federal budget deficits are back big time. What will be their long-term consequences? Seven big thinkers enter the ring, gloves up

International Economy, The, Summer, 2003

In a January 9 letter to the Wall Street Journal, Gale and Orszag said, "[S]tandard economic reasoning, present in almost all macroeconomic textbooks, implies that budget deficits reduce national savings." But an appeal to standard reasoning and textbooks is evidence of nothing. The authors' Brookings Institution paper quotes "the views of numerous leading academics, policy makers and government agencies." Yet those same apostles of standardized reasoning were once equally self-assured about "twin deficits." If orthodoxy was proof of reality, the world would still be flat.

From 1998 to 2001, while the budget was in surplus, national savings amounted to 18 percent of GDP. From 1981 to 1989, when deficits averaged 3.8 percent of GDP, national savings was 18.2 percent of GDP. When other countries moved from deficit to surplus, their national savings also did not rise. Taking a dollar from a taxpayer's budget and adding it to the government's budget does not improve both budgets. And governments do not borrow from the flow of national savings, as the authors theorize, but from the world's stock of assets.

The letter goes on to say, "Recent evidence supports the view that deficits affect interest rates. This evidence comes from almost all major macroeconomic models ... and in particular from studies that properly focus on the relation between interest rates and expected future deficits." But those models are built on assumptions, not facts, and cannot predict interest rates or anything else.

The "recent" evidence shows "a strong positive relationship between deficits projected by the Congressional Budget Office and the spread between long-term interest rates and short-term interest rates." But the spread between long and short rates is actually widest when bond yields are low and the fed funds rate even lower. Gale and Orszag simply showed that short-term interest rates fall after recessions, when the CBO forecast turns gloomy. The CBO's August 1994 projected deficit for 2000 was exaggerated by 5.3 percent of GDP, an annual error of $520 billion. The Gale-Orszag theory that such projections are more potent than actual deficits is nonsensical. And their own evidence about the yield curve refutes their hypotheses about the level of interest rates.

According to the Gale-Orszag hypothesis, expected future deficits have fallen dramatically since 2000 because mortgage rates fell by two percentage points. With standardized textbook theory so obviously at odds with the facts, it is the facts and not the theory that Gale and Orszag find expendable.