The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
April 1, 2000 Federal Reserve Bank of Cleveland Fiscal Policy and Fickle Fortunes: What’s Luck Got to Do With It? by David Altig A mong the more amazing elements of the amazing U.S. economic landscape of the latter 1990s was the transition from federal budget deficits as far back as memory can recall to budget surpluses as far forward as the eye can see. This development was as sudden and unexpected as it was welcome, and virtually overnight the political debate shifted from how to restrain government spending and enhance government revenues to how best to allocate the spoils of the seemingly successful federal budget battles. Not everyone, however, is convinced that the war is over. In his testimony to Congress on February 17, 2000, Federal Reserve Board Chairman Alan Greenspan commented that “it might … be prudent to eschew new longer-term, potentially irreversible commitments until we are assured that the on-budget surplus projections are less conjectural than they are, of necessity, today.” This comment reflects a concern, shared by many, that there is an unquestioned presumption that projected surpluses are as good as achieved, and that such a presumption is a shaky foundation from which to launch major new budget initiatives. In this Economic Commentary, I briefly review the fiscal fortunes of the 1990s and attempt to answer the question posed in this article’s title. What’s luck got to do with it? An awful lot, I contend, and with that conclusion comes a strong cautionary tale about counting our future fiscal surpluses before they hatch. ISSN 0428-1276 ■ Birds of a Feather In November 1990, President Bush signed into law the Omnibus Budget Reconciliation Act of 1990 (OBRA90). This legislation provided for deficit reductions of $482 billion over the fiveyear period 1991–95, to be accomplished with a combination of tax hikes, restraints on mandatory spending programs, and caps on discretionary expenditures. Over that five-year period, actual deficits rose $6 billion relative to the cumulative levels projected prior to the bill’s passage. In August 1993, President Clinton signed into law the Omnibus Budget Reconciliation Act of 1993 (OBRA93). This legislation provided for deficit reductions of $433 billion over the five-year period 1994–98, to be accomplished with a combination of tax hikes, restraints on mandatory spending programs, and caps on discretionary expenditures. Over that five-year period, cumulative deficits fell more than $1 trillion relative to the cumulative levels projected prior to the bill’s passage. As OBRA93 began to take form, the general perception was that OBRA90 had been tried and had failed. In the summer of 1993, before OBRA93 was passed, the Congressional Budget Office (CBO) foresaw deficits rising to $360 billion by 1998. By fiscal year 1998, however, the federal budget was $69 billion in the black, making a strong argument for the case that the later legislation succeeded where the earlier could not. But two outcomes so differenct hardly seem to belong to the same domain and century, let alone two nearly identical pieces of legislation spaced a mere three If you buy a lottery ticket, you usually wait to see if you’ve won before going on a spending spree. This Economic Commentary explains why we ought to be just as careful about spending projected federal budget surpluses. years apart. Furthermore, the overachievement of OBRA93 was as unexpected as the underachievement of OBRA90. Even after the passage of OBRA93, the CBO was projecting that the federal budget for fiscal year 1998 would be $200 billion in deficit. ■ Details, Details The first obvious question is whether the specifics of OBRA90 and OBRA93 were sufficiently different to explain why one would succeed and the other would not. The answer is no. In terms of the sheer magnitude of their goals, the two pieces of legislation were nearly identical. OBRA90 had five-year deficit-reduction provisions accumulating to $482 billion, and OBRA93 had provisions accumulating to $433 billion. Nor were the broad details of the packages dissimilar, the primary distinction being that OBRA90 was more heavily weighted toward discretionary expenditure cuts and OBRA93 more toward revenue increases. (Discretionary spending items are subject to annual approval by Congress.) About 39 percent of the planned deficit reduction in the 1990 legislation came from nonentitlement spending cuts, and 33 percent from revenue increases. The corresponding figures for the 1993 bill were 16 percent and 55 percent. Important to both of these pieces of legislation were institutional procedures to restrain federal expenditure and tax policy changes. These procedures were introduced in the 1990 Budget Enforcement Act (BEA), the companion legislation to OBRA90. The potential for OBRA90 and OBRA93 to achieve their deficit goals clearly depended on subsequent Congresses adhering to the provisions of the BEA. So, did OBRA93 succeed where OBRA90 failed because Congress and the Administration became more disciplined and more faithfully hewed to the BEA’s procedural limitations? Again, the answer is no. A fair reading of the record clearly shows that the failure of OBRA90 to meet its stated deficit reduction goals did not occur because Congress and the Administration significantly reneged on their legislative promises. ■ It’s the Economy, Stupid? One very plausible explanation for the dramatically different records of the two budget packages lies in the fact that the post–OBRA93 economic landscape was much different than that of the short post–OBRA90 era. For most of the period after the 1990 legislation, the American economy has been in the upswing phase of the business cycle. But the record-breaking episode of economic growth that commenced in April 1991 seems like a tale of two expansions: Uncharacteristically sluggish at its inception, astonishingly robust well past the prime of most other business cycles. While higher-than-average economic growth has been the hallmark of the early phases of most postwar U.S. expansions, economic performance in the first years of this expansion was surprisingly soft. In the first three years of the four previous sustained expansions, the growth rate of real gross domestic product (GDP) averaged nearly 17 percent.1 This represents yearly growth rates of about 5.1 percent, well above the postwar (1958–99) average of about 3.4 percent. In contrast, during the three years following the 1990–91 recession— roughly the period from the implementation of OBRA90 through the passage of OBRA93—GDP grew only 10 percent, just less than the annual postwar average. This performance, along with other coincident macroeconomic developments— such as the paths of interest rates and inflation—took its toll on the federal budget. Over the three years from 1991 through 1993, economic projection errors accounted for actual federal deficits that were some $66 billion in excess of initial budget resolution estimates. Initial budget resolutions are the annual “first drafts” of the federal budget prepared by Congress to guide spending and taxation for the subsequent fiscal year. Thus, this $66 billion miscalculation derives from projections made at the beginning of each fiscal year, not from five-year estimates. Although one might expect imprecision in longer-term prognostications, the substantial differences between budget resolution projections and actual outcomes drive home the enormous difficulties inherent in budget forecasting, even at horizons as short as one year. The economy over the course of the current expansion has, of course, steadily improved. Annual GDP growth averaged about 3.4 percent per year during 1994– 96 and about 4.2 percent in 1997–99. The sustained acceleration of growth so long into this cycle was as surprisingly favorable to the budget picture as the “delayed response” of the economy early in the cycle was unfavorable. In budgetary terms, economic influences over the immediate post–OBRA93 period (1994– 96) were the mirror image of the post– OBRA90 episode. In contrast to the $66 billion dollar increase associated with economic developments in 1991–93, over the 1994–96 period macroeconomic forecast mistakes reduced actual deficits by $67 billion relative to initial budget resolution projections. And the hits just kept coming. In 1997 and 1998, the economy contributed $108 billion more than expected toward higher revenues and lower outlays, enough in 1998 to help shove the federal budget from the red-ink territory originally anticipated to the first surplus in almost 30 years. Over the entire 1997–99 period, the U.S. economy contributed $203 billion more to the government’s coffers than the CBO had expected in January 1997. But what caused this striking economic turnaround? It is implausible that policy differences explain the differential economic performance in the episodes immediately following OBRA90 and OBRA93. A more convincing case might be made that the economy needed the double barrels of OBRA90 and OBRA93 to slay the deficit beast and unleash the potential that was finally realized in the latter half of the 1990s. Possibly, but few are the informed observers who would make such dramatic claims for fiscal legislation in isolation. Writing in a recent issue of the Wall Street Journal, economist (and former Federal Reserve Governor) Lawrence Lindsey points to a collection of regulatory reforms, changes in monetary policy, tax reform, and financial-market innovation dating back to 1982 for an explanation of recent economic prosperity.2 Even these —very significant—influences may ultimately be viewed as minor when juxtaposed with the building momentum of productivity advances unleashed by the information-technology revolution, developments that the University of Rochester’s Jeremy Greenwood dates back to the mid-1970s.3 But even if we unrealistically attribute all of the positive economic surprises to the combined effects of OBRA90 and OBRA93, we would still fall far short of an explanation for the dramatic deficit reductions that we have actually enjoyed. Consider, again, some numbers. Over the five years from 1994 through 1998, budget deficits were cumulatively some $415 billion lower than projected in the year-by-year budget resolutions. Of this total, only about 42 percent can be attributed, after the fact, to imperfect foresight about macroeconomic developments. What, then, accounts for the remaining 58 percent in cumulative annual misestimates from 1994 through 1998? The official answer: “Technical factors.” ■ The Measure of Our Ignorance The distinction between economic and technical factors is a rather fine one, in that technical factors are typically economic in nature. In formulating its budget projections, the CBO uses a methodology that partially relies on forecasts of such things as GDP, inflation, and interest rates. Also key are projections about the components of income. How fast will wages and salaries grow? Corporate profits? Based on these projections, the CBO uses statistical models to estimate the likely magnitudes of spending and revenues, and then combines these estimates with projections for other key elements of the macroeconomic landscape (like capital-gains realizations) to develop an overall budget outlook. Over time, and in particular circumstances, economic developments will FIGURE 1 SOURCES OF DIFFERENCES BETWEEN ACTUAL BUDGET TOTALS AND BUDGET RESOLUTION ESTIMATES Billions of dollars 120 100 Technical 80 60 40 20 0 –20 Policy –40 –60 Economic –80 –100 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 SOURCE: Congressional Budget Office. conspire to undermine budget projections, not just because the broad macroeconomic forecasts of GDP, interest rates, and inflation are wrong, but also because the methods used to connect the forecasts to revenue and spending outcomes fall short. The behavior of the stock market and revenues from capital-gains realizations, for example, are notoriously difficult to predict, even with perfect knowledge of GDP growth, interest rates, and inflation. Although capital-gains realizations clearly represent an economic decision by equity owners, forecast errors from this source are classified as “technical,” rather than “economic,” in the framework of the CBO’s official reports. How problematic is the observation that technical factors account for a good chunk of unexpected surprises over the past decade? Very. Technical explanations of discrepancies between the budget outlook and actual outcomes are by nature made after the fact. Otherwise, they would be part of the outlook in the first place. This fact suggests the nearinevitability that the complicated interactions critical to budget forecasts will remain a blight on the best-laid plans. This inevitability is readily acknowledged by budget analysts. And some discrepancies there can be. Figure 1 illustrates the sources of differences between estimates based on initial budget resolutions and the actual deficit outcomes for 1980–98. The most striking element of this figure is the very large and persistent role played by technical factors in the history of budgetresolution/actual-budget discrepancies since 1993. In fact, although subsequent revisions were not always as dramatic as that year’s, it is now clear that 1993 was just an early entry in an unbroken string of technical developments that have substantially altered the federal budgetary landscape. Over the entire period from OBRA90’s birth through fiscal year 1998, technical misses were over one and a half times as large, in absolute value, as the misses attributable to simple macroeconomic outcomes. ■ The Roseanne Roseannadanna Factor The sources of this string of technical misses are found in familiar budgetary issues, both chronic and acute. In 1991 and 1992, lower-than-expected expenditures associated with the cost of the savings and loan crisis were helping to disguise misses in Medicare and Medicaid spending and capital-gains-tax collections that were expanding the size of the federal deficit. Unlike the thrift-related developments, many of which might reasonably have been associated with the policy-change category, these items were more truly technical in nature, arising from incomplete information and our imperfect ability to estimate their budgetary impact.4 What changed after OBRA93 was that the mistakes in guessing future federal revenue receipts and Medicare/Medicaid outlays were happy ones, made happier, perhaps, by being so unexpected. As an illustrative case, consider the history of post–OBRA93 technical revisions to the fiscal year 1998 (FY98) budget.5 In September 1993, the CBO released its first regular budget outlook report subsequent to the passage of OBRA93, projecting a deficit of $200 billion for FY98. As noted earlier, the FY98 federal budget actually registered a surplus of $69 billion. One-half of that turnaround resulted from technical revisions to revenue estimates and Medicare/Medicaid expenditure projections. One of the more striking aspects of this history is how late in the game it became apparent that revenues would be much higher, and social health insurance outlays much lower, than had been expected. More than 90 percent of the total technical revisions made subsequent to the initial 1993 projection for the FY98 surplus came in 1997 and 1998. Nearly half of that total revision was made in 1998. And these numbers capture only the forecast errors that can’t be explained by unforeseen technical forces. Discrepancies from technical and economic sources combined account for more than the $269 billion swing from the FY98 deficit projected in September 1993 to the surplus that was actually realized. Almost all of these discrepancies were unexpected prior to 1997. About 40 percent were unforeseen prior to 1998. We can, of course, construct reasonable explanations after the fact. The stronggrowth, low-inflation U.S. economy of the latter 1990s brought with it a booming stock market, an acceleration of capital-gains realizations, and burgeoning revenues from this source. Changing methods of delivering health care services, many of which were responses to earlier policy decisions, no doubt helped to moderate Medicare and Medicaid expenditure growth. But these are just examples of the “Roseanne Roseannadanna factor”: It’s always something. Ex post rationalization of extreme miscalculations hardly makes the case for confident extrapolation. ■ Fiscal Policy and Fickle Fortunes The record clearly shows that budget projections retain their conjectural nature for a very long time—sometimes nearly to the point where they pass from projections into history. Faced with this reality, prudence requires a lot of patience. Consider this. In 1995, the Republican Congress and the Clinton Administration became embroiled in a dispute over budgetary priorities that resulted in two partial shutdowns of the federal government, a dispute with political ramifications that still persist today. Among the major items of contention was the Congressional intention to cut outlays on Medicare and Medicaid by $452 billion over the period from 1996 through 2002. As of the latest CBO projections, cumulative changes in actual and projected outlays in this category have yielded 55 percent of this total from technical adjustments alone—those that are not directly related to macroeconomic developments or the projected impact of policy changes. In this case, it was definitely better to be lucky than good. But luck has a history —and a recent one, at that—of turning south. That is a lesson worth remembering as we contemplate how we might exploit projected federal surpluses as far as the eye can see. ■ Footnotes 1. The real GDP growth rates for each of these episodes was 17.5 percent (1961:IIQ– 1964:IIQ), 12.4 percent (1971:IQ–1974:IQ), 17.8 percent (1975:IIQ–1978:IIQ), and 17.8 percent (1983:IQ–1986:IQ). These, and other GDP figures reported in the text reflect recent revisions in the U.S. national income and product accounts. 2. Lawrence Lindsey, “America’s 17-Year Boom,” Wall Street Journal, January 27, 2000. 3. Jeremy Greenwood, “The Third Industrial Revolution: Technology, Productivity, and Income Equality,” Federal Reserve Bank of Cleveland, Economic Review, vol. 35, no. 2 (1999, Quarter 2). 4. In both 1992 and 1993, anticipated Congressional funding of the Resolution Trust Corporation—the organization created to implement the thrift-industry reorganization— failed to materialize, and in the end total outlays associated with financial sector distress fell significantly short of initial guesses. The big S&L-related budget-projection errors in these years are attributed to technical rather than policy sources only because the CBO classified them as such. 5. The figures in this section are taken from various issues of the CBO’s beginning-of-year publication, The Economic and Budget Outlook, and its midyear updates. The revisions reported across these publications may not sum to the actual cumulative change in budget estimates between any two publication dates due to interim changes in the baseline. David E. Altig is a vice president and economist at the Federal Reserve Bank of Cleveland. The views stated here are those of the author and not necessarily those of the Federal Reserve Bank of Cleveland or of the Board of Governors of the Federal Reserve System. Economic Commentary is published by the Research Department of the Federal Reserve Bank of Cleveland. To receive copies or to be placed on the mailing list, e-mail your request to 4d.subscriptions@ clev.frb.org or fax it to 216-579-3050. Economic Commentary is also available at the Cleveland Fed’s site on the World Wide Web: www.clev.frb.org/ research, where glossaries of terms are also provided. We invite comments, questions, and suggestions. E-mail us at editor@clev.frb.org. Federal Reserve Bank of Cleveland Research Department P.O. Box 6387 Cleveland, OH 44101 Return Service Requested: Please send corrected mailing label to the above address. Material may be reprinted if the source is credited. Please send copies of reprinted material to the editor. BULK RATE U.S. Postage Paid Cleveland, OH Permit No. 385