The Deficit Estimating Cycle


Both President Bush and Sen. John Kerry promise to cut the budget deficit in half in four years. At first glance, that does not appear to offer any reason to prefer one candidate over another.

The first difference is Mr. Kerry hopes to get to the same point with higher tax rates on dividends, capital gains and two-earner professional couples. The second big difference is his most optimistic dreams of sums to be collected from higher taxes are dwarfed by just one of his big spending plans -- $950 billion for additional subsidies to health insurers (supposedly reduced to $650 billion by forcing the sickest into HMOs.)

Even the New York Times could not fail to notice additional pricey medical subsidies "could endanger deficit cutting" -- a politically polite way of saying Mr. Kerry has to add a lot more tax increases to make it all add up. But that article nonetheless made a heroic effort to put Mr. Bush and Mr. Kerry in the same boat, by saying, "Most of the deficit they promise is dependent on the accuracy of a forecast from the Congressional Budget Office (CBO)."

Specifically, the CBO's March estimate already assumed the deficit would fall $200 billion by 2006 (from 4.2 percent to 2.1 percent of GDP). That is what normally happens in cyclical recoveries. The New York Times implied, however, the CBO is likelier to be underestimating rather than overestimating future deficits. Yet the usual pattern at this stage of the cycle is for the budget estimates, particularly for the distant future, to be too gloomy.

Even with short-term budget estimates -- those presented every January after three months of the fiscal year are already over -- the estimated deficit is often off by $100 billion or more. By June, for example, this year's deficit was reportedly expected to be $100 billion smaller than the Office of Management and Budget estimated in January. When peering five years into the future, such errors can become 5 times larger. Stretching the estimate to 10 years can be wilder than a wild guess.

In January 1995 -- after tax rates had been increased in 1991 and 1993 in response to previous dismal budget forecasts -- the director of the Congressional Budget Office testified before Congress that the baseline budget deficit would reach $284 billion by the year 2000, or 3.1 percent of gross domestic product. By 2000, however, that budget was actually in surplus by nearly $237 billion, or 2.4 percent of GDP. Within just five years, the 1995 deficit estimate turned out to be exaggerated by $521 billion -- for a single year.

In recent years, Congress has become obsessed with a ritual of combining 10 such future estimates in a single total, which can easily produce numbers that are off by trillions. It would be prudent to keep the awesome magnitude of past errors in mind whenever someone advocates a big, pre-emptive tax increase on the basis of some tenuous estimate of what the deficit might or might not be five or 10 years from now.

That cautionary note is amplified for the self-styled "independent" estimates, which are too cleverly spun with the too obvious purpose of influencing policy.

It is often said ridiculous exaggerations of future deficits, as in 1993-95, are offset by big underestimates of future deficits, as in 2000-2001, so we supposedly don't know whether today's deficit estimates are too small or too large. Not true.

Budget forecasting errors follow an obvious cyclical pattern -- much too optimistic near cyclical peaks (such as 2000) and much too pessimistic in the early stages of recovery (such as 1984, 1994 and 2004). The error at cycle peaks is to project good times too far ahead. When that proves wildly incorrect (as in early 2001), there is a natural bureaucratic inclination to overcompensate by marking down the future "as far as the eye can see" (which is not far, actually).

One reason deficit projections are always exaggerated during the first few years after a recession is the CBO's method of forecasting long-term potential growth of the economy. It more or less assumes the average rate of economic growth -- including recessions -- will also be the average growth rate between recessions. Since recessions have lately been a decade apart, economic growth is therefore persistently underestimated for about 10 years. And deficit estimates have to be repeatedly scaled down as the economy continually outperforms the unduly dismal estimates.

As the economic expansion takes hold, reality repeatedly compels the CBO to raise its GDP growth forecast in the short run. But sticking to their dismal long-term outlook requires that faster economic growth today must, as a mathematical necessity, be offset by assuming slower economic growth tomorrow. Stronger economic growth therefore has the paradoxical effect of making future budget deficits look worse rather than better: "Changes in the economic forecast had a slightly positive effect on revenues in 2004 and 2005," the CBO reported earlier this year, "but an increasingly negative impact on revenues during the remainder of the projection period." That is so typical -- actual good news in the short run is always purchased at the price of assuming worse news ahead.

In this instance, that familiar "negative impact" was made even more negative because the revised CBO forecast "lowered inflation, which tended to reduce receipts from all sources." Why lower estimated inflation? Probably "to reduce receipts." No change in the news over a few months could ever legitimately change estimated average inflation over 10 years. But such assumptions may frequently need to be adjusted, to keep the future looking as grim as possible.

In just five months between last August and this January, the CBO increased its forecast of economic growth for the current year by a full percentage point, from 3.8 percent to 4.8 percent. Whenever it does that, however, it feels compelled to lower estimated growth in the future to make it all add up to an appropriately (in their view) low number. As a result, the CBO then estimated the economy could not possibly grow faster than 2.8 percent a year from 2006 to 2009, nor faster than 2.5 percent from 2010 to 2014.

To put such figures in perspective, real GDP grew by 2 percent a year from 1930 to 1940. So, the CBO expects future economic growth after 2009 to be slightly better than during the Great Depression.

Many professional deficit hawks will spend the next year or two telling us the sky is falling, as they did in 1985 and 1995. Private researchers may be "independent," but not disinterested. Those of us who are eager to roll back the runaway federal gravy train have an incentive to exaggerate future budget deficits, even though we know very well it is the spending itself rather how it is financed that matters most.

Those who want to increase taxes (at all times and for any conceivable reason) have even more incentive to exaggerate future budget deficits, because spending cuts take booty from their most bountiful constituents, which explains why such cuts are termed "unrealistic."

Just remember that this budget forecasting ritual is repeated during the first few years after every recession, roughly every 10 years. Be patient, and the latest scary deficit estimates for the distant future are likely to be repeatedly adjusted downward, grudgingly, once again.