Commentary

Don’t Trust the CBO’s Numbers

This article was first published in the Wall Street Journal, August 30, 2001.
The Congressional Budget Office is right where it likes to be — in the news.

First reports of its latest budget outlook came out of the same spin machine. The Washington Post headline was “Budget Will Tap Into Social Security,” and the New York Times’ was “Report Says Lower Surplus Will Affect Social Security.” The CBO report said no such thing, of course. What it said was, “the distinction between on- and off-budget surpluses is unimportant from an overall economic perspective” and “CBO is projecting small on-budget surpluses or deficits in 2001 through 2005.” The Social Security trust fund cannot possibly be “raided” or “tapped,” but has to grow by the amount that payroll taxes exceed benefits. The CBO thinks an already large surplus will grow to 3.7% of GDP by 2011, noting, “all debt available for redemption will be retired by 2010.” Dick Gephardt calls that a “fiscal crisis.”

The compelling part of CBO estimates has nothing to do with Social Security or the odd priority accorded to retiring debt. The real story is the way in which each forecast is treated as indisputable fact, despite the CBO’s record of exaggerating deficits and underestimating surpluses.

In 1993, the CBO predicted that the deficit would soar to $653 billion in 2003. This week, they said that same budget will be in surplus by $172 billion. Little of that $825 billion revision can be explained by legislation or luck. Nearly all of it reflects the magnitude of past forecasting blunders.

Errors are unavoidable, but perpetual bias is another matter. CBO errors always tilt in a specific direction. Aside from the first year of recessions, the CBO always exaggerates future budget deficits and underestimates surpluses.

Past forecasts often overstated deficits by huge amounts even for the current year — by $78 billion in 1992 and $102 billion in 1997. In early 1998, the CBO thought the next year’s surplus would be $2 billion, but it turned out to be $125 billion. Looking further ahead, CBO errors have been staggering. Next year’s budget, now estimated to be in surplus by $176 billion, had once been expected to show deficits of $579 billion (per the CBO’s 1993 forecast), $349 billion (1995 forecast), and $188 billion (1997 forecast).

There are good reasons to expect the CBO now underestimates future surpluses in the same way it used to overestimate future deficits. There is a cyclical pattern to CBO budget forecasts. They always turn overly pessimistic during economic slumps, and move closer to reality only after years of economic expansion.

>From the fall of 1990 until early 1993, the CBO began to overcompensate for its understandable failure to forecast recession, just as it had in 1982-85, by fabricating increasingly bleak deficit forecasts “as far as the eye could see.” In October 1990, an unsuspecting Congress capitulated by enacting an ill-timed tax increase. Yet the CBO’s estimates soon turned even gloomier. Tax receipts for 1991-95 were predicted to wind up smaller by $537 billion than had been predicted before the tax increase.

By March 1992, with half of that fiscal year behind them, the CBO thought the 1992 deficit would hit $368 billion — an estimate $78 billion too large. Undaunted, the CBO turned even more pessimistic about future years.

In early 1993, the CBO warned the Clinton administration that the deficit would remain stuck near $300 billion until 1997, and then soar to $653 billion by 2003. The assumptions were remarkable. The CBO wrote that “real economic growth is posited to continue at about 2% a year.” Yet the only time the economy has done nearly that badly for an entire decade was 1930-40, when annual growth averaged 2.2%. Rather than question the CBO’s assumption that the economy was about to underperform the Great Depression, the administration pushed through another destructive tax hike in 1993. But all of Mr. Clinton’s new taxes were merely worth $48 billion a year, said the CBO. A number that small could easily be erased by reducing estimated economic growth a bit more.

In August 1995, with most of the year behind them, the CBO predicted 1.3% growth for that year. But actual growth was 2.7%. In August 1996, the CBO predicted 2.1% growth for 1996 and 1.9% for 1997. Actual growth was 3.6% in 1996 and 4.4% in 1997. But feeble GDP estimates served their purpose, keeping deficit frenzy alive and the CBO in the spotlight.

By January 1998, the CBO’s pattern of overestimating future deficits had been transformed into one of underestimating future surpluses, which they kept doing until recently when the economy stalled.

So we, in a slump, are asked to worry about whether projected surpluses will be large enough. The question is so inapt it barely merits response. The CBO’s history reveals that if the economy performs nearly as well as usual between slumps, then future surpluses will be larger than the CBO predicts. To its credit, the CBO has finally raised its estimate of “potential” growth to 3.3%, up from 2.1% in 1996. That first figure is simply the 1980-2000 average. But that average included recessions in 1980-82 and 1990-91. Economic growth between recessions has been much faster — 4% in 1983-89 and 3.7% in 1992-2000.

The CBO’s unique method of averaging is to disregard below-average growth in 2001-02, and restart the clock in 2003. That means that the CBO’s “projected” growth never exceeds 3.2%. That is, the CBO assumes that we never make up for below-average growth during the downturn. That has the effect of turning a temporary cyclical problem into the illusion of permanent and unabating losses of tax revenue.

The CBO is predicting that not one future year will ever match the mere average of the past two expansions. That 3.2% ceiling on growth is bad math. We can never get back to a 3.3% average by adding a bunch of 3.2% figures to 1.7% this year and 2.6% in the next. It is also bad economics. The conventional concept of potential growth assumes “full employment” of men and machines — not, as now, when starting from a point where nearly 30% of industrial capacity lies idle. In reality, the economy will grow much faster than 3.2% for at least a few years, sooner or later, by simply putting idle plants back to work.

Unless the next expansion is much weaker than the past two, the CBO is still underestimating the economy and future surpluses. In truth, the budget surplus is already larger than it should be in hard times. In any case, Washington spends too much time worrying about the government’s budget, and not enough worrying about the budgets of hapless taxpayers.

Alan Reynolds is a senior fellow with the Cato Institute.