Much of the debate in Washington concerns policy issues. The system of rules and procedures within which this debate occurs is rarely the focus of attention. The use—some would say abuse—of the filibuster is a rare exception. But when it comes to the federal budget, rules and procedures are critical, shaping everything from estimates of expenditures and revenues to the timetable for making critical decisions.
The House of Representatives’ recent decision to change the way tax and entitlement proposals are evaluated has refocused attention on this seemingly arcane topic. As we will see, questions can be raised concerning both the status quo and proposed reforms—not only about the recent House action, but also about a wide range of other changes that have been discussed from time to time in recent years. Despite the defects of the current system, I will argue, there are good reasons why would-be reformers should proceed cautiously.
In January 2015, the House approved a change to the rules governing the evaluation of major budget proposals. Henceforth, official cost estimates must include an estimate of the economic effects of such legislation.
Democrats have excoriated this change—“dynamic scoring”—as a Trojan horse for tax cut proposals that would otherwise be evaluated as generating unaffordable revenue losses. The matter is more complicated. The Joint Committee on Taxation (JCT), which assesses tax proposals, already evaluates—indeed, is required to evaluate—the effects of these proposals on growth, job creation, and other key economic variables. The Congressional Budget Office (CBO) has conducted macroeconomic analyses on programs such as President Obama’s stimulus package and on proposals such as the Senate’s comprehensive immigration reform bill. In 2013, moreover, CBO published a report assessing the impact of higher and lower long-term budget deficits on economic growth.
The difficulty is that different models produce very different estimates. Often they agree on the direction of change but diverge sharply on its magnitude. For former Rep. Dave Camp’s comprehensive tax reform bill, the high estimate of revenues from added growth was 14 times the low estimate (for a useful summary, see Jonathan Weisman’s New York Times Article). CBO’s estimates of the growth effects of a ten-year, $2 trillion reduction in the budget deficits ranged from 0.5% of GDP to 1.3%.
Sometimes the models even disagree about the direction of change. When the JCT asked nine different modelers to assess the impact of a value added tax on economic growth, their estimates of the short run effects ranged from +16.4% to -4.2%.
The new rule requires the JCT—and for large spending programs, the CBO—to merge these differences into single point estimates and to include those estimates in their official budget scoring. In response to the House’s action, the JCT has stated that “providing dynamic scores is a new and significant challenge” and that it is assessing “the best way to provide Members of Congress information of likely macroeconomic feedback effects that best represents the current state of macroeconomic research.”
It remains to be seen how this can be done. Qualitative differences among models make it difficult to select a single “central case” as the basis for point estimates. Reporting the average of the various models may be arbitrary. However carefully worked out, new procedures are unlikely to resolve fundamental disagreements between the political parties about how the economy works and will respond to policy changes.
Timely and efficient budgeting
Whatever the merits of the dynamic scoring rule, it has helped to reopen discussion on a wider range of budget rules that could alter the substance as well as the process of federal budgeting. For example, both budget experts and the public have been dismayed by Congress’s repeated failure to enact annual budgets prior to the start of each fiscal year. According to the Congressional Budget Act of 1974, the House and Senate are supposed to agree by April on an overall budget framework designed to guide the preparation of the appropriations bills governing the spending of departments and agencies.
During the past generation, Congress has rarely completed these tasks on time. Indeed, 1996 was the last time the federal government had a complete budget prior to the start of the fiscal year. The delay since then has averaged 104 days. (As of February 10, the full-year Department of Homeland Security funding bill for FY2015 is 132 days behind schedule.) Often the government is funded through a series of short-term bills that make it almost impossible for departments and agencies to operate efficiently. On several occasions the government has shut down when the House and Senate could not reach agreement with each other or with the president.
Not surprisingly, these failures have evoked numerous proposals for reform. One would establish a default mechanism: if the House and Senate cannot agree on a budget resolution by April 15, last year’s resolution would automatically go into effect so that the appropriations process could proceed. Another proposal would do the same for full-year appropriations bills not signed into law by September 30.
Another strategy is to enact new incentives for Congress to meet its budget obligations in a timely manner. In recent years, for example, bipartisan proposals offered in both the House and the Senate would make it difficult for them to consider non-budgetary matters after April 15 until they have reached agreement on a budget resolution.
Some believe that even tougher steps are required to break through congressional inertia and gridlock. One proposal would halt members’ pay after April 15 and place the withheld funds in an escrow account until the passage of a budget. Another—known as No Budget, No Pay—would halt pay after September 30 until all appropriations bills are completed and sent to the president. The toughest version of these proposals would not provide back pay; members of Congress would lose—permanently—the paychecks withheld in response to missed budget deadlines.
Some critics have charged that all these proposals would violate the 27th Amendment, which states that “No Law, varying the compensation for the services of the Senators and Representatives, shall take effect, until an election of representatives shall have intervened.” Proponents of such proposals have responded in two ways. The version that passed the House and Senate in 2013 would have placed the withheld funds in an escrow account to be paid out at the end of that Congress. One of the nation’s leading constitutional scholars, Yale’s Akhil Amar, has stated that because the members eventually get paid, this procedure “is not really varying the compensation.” Advocates of the tougher approach—permanent pay losses for missed deadlines—have drafted their proposal to take effect in the Congress after it passes. This version arguably would pass constitutional muster because the members of the new Congress would not have acted to vary their own pay. In the same way, proposals to link congressional pay to the performance of the economy would be constitutional as long as they did not take effect until the next Congress is elected.
Although congressional experts are divided on the merits of linking congressional pay to performance, the American people support this strategy by a margin of 4 to 1. They may be on to something. After Congress reluctantly approved a modest version of No Budget No Pay in January 2013 as part of a package to postpone the ongoing debt ceiling crisis, the Senate passed its budget resolution for the first time since 2009.
Many states budget on a two-year (“biennial”) cycle rather than annually, and in recent decades proposals to extend this practice to the federal government have found numerous advocates. Indeed, every administration from Ronald Reagan to George W. Bush has supported the idea.
As a recent report from the Congressional Research Service notes, proponents of biennial budgeting have offered three main arguments in its favor. First, it would reduce the congressional workload by eliminating the need to consider routine and repetitious matters every year. Second, by requiring two-year appropriations bills to be completed by the end of the first year of the cycle, biennial budgeting would free up the second year for systematic oversight of programs whose efficacy and efficiency Congress seldom reviews. And third, a longer time-horizon would allow departments and agencies to implement their programs and administer their budgets more effectively.
Opponents counter that biennial budgeting would require departments and agencies to put together budgets too far in advance to project conditions with any confidence. The farther in advance budgets must be enacted, the more likely it is that they will have to be adjusted, complicating the process that two-year budgeting was intended to simplify.
Nor, they argue, would this reform have a large impact on congressional operations. Most oversight is conducted by the committees with jurisdiction to authorize programs, not those tasked with funding them. And to the extent that appropriations committees do perform oversight, the move away from annual budgets could actually reduce program evaluation during the budget process. And because Congress seems to have acquired the habit of budgeting only on the brink of government shutdown or default, a two-year budget cycle could lead to longer delays and even steeper fiscal cliffs.
Finally, say the opponents, the states have voted with their feet. In 1940, 44 states had biennial budgets; today, just 19 do. Of the 18 states that have changed their budget cycles since 1968, 15 have moved from biennial to annual. And most of the large states, the complexity of whose affairs comes closest to that of the federal government, do their budgeting each year.
In sum, say the critics, it is at best premature to conclude that the advantages of biennial budgeting outweigh the disadvantages. If federal policymakers want to explore this option, they should test it on a handful of budget accounts and assess its consequences. Against the background of uncertainty sketched out above, this cautionary note seems wise.
When business leaders gather to discuss the federal budget, it rarely takes long for someone to express incredulity that the government does not distinguish more sharply between capital expenditures and operating expenses. They are different qualitatively as well as quantitatively; thus, the argument runs, the government should budget for them differently, finance them differently, and evaluate them differently. And besides, most states have capital budgets, so why shouldn’t the federal government?
Other advocates for capital budgets argue that the current system biases the federal budget against spending that would produce long-term returns. For example, if the federal government had a capital budget, it would be easier to fund levels of infrastructure investment: “Capital outlays would be seen for what they are—net additions to the government’s capital stock, which like the capital assets of a company, would be depreciated over their useful life.”
A report from the Congressional Budget Office offers some cautionary notes. First, the economic benefits of capital spending by the federal government would depend on the government’s ability to select projects on an economic rather than political basis. Second, to the extent that capital budgeting makes outlays in this category easier, it could provide incentives to reclassify ordinary outlays as investments. Third, arriving at a definition of capital expenditures would be a significant challenge. For better or worse, the tax code provides guidelines for the private sector, but there is no agreed template for the public sector. These difficulties help explain why several presidential commissions have rejected the idea of a separate capital budget for the federal government.
In addition, CBO says, the analogy between states and the federal government is imperfect. To a greater extent than the federal government, states face credit-market constraints; most have requirements for balanced operating budgets and restrictions on debt; and the competition among states for taxpayers is more limiting than the competition among countries. And crucially, the federal budget serves macroeconomic objectives in ways that state budget cannot.
As Brookings economist Charles Schultze pointed out many years ago, capital budget advocates typically propose a general budgetary rule: revenues should cover current operating expenses while net investment outlays should be financed through borrowing. Although this rule may make sense for the private sector, it can be counterproductive in circumstances of high employment when the economy is operating at or near full capacity. The federal government needs the flexibility to finance capital expenditures through either revenues or borrowing, depending on circumstances. In addition, Schultze echoed the worries of other analysts: capital budgets would invite advocates to game the system—for example, by ensuring accommodative depreciation schedules for their favorite categories of investments.
This brief discussion offers only a sample of possible budget rules drawn from a much longer list. There is, for example, an arcane but consequential debate over the “baseline” from which CBO measures additions to or cuts in the federal budget. But whatever the specific proposal, one generalization seems safe: new rules change incentives, often in unanticipated ways. Measures designed to solve one problem can end up creating or exacerbating others. In times of intense political polarization, moreover, partisan objectives will often shape changes in budget rules.
Granted, there may be no such thing as neutral procedures. Nonetheless, some are more partisan than others, more dependent on controversial economic theories. Good budget rules reflect the limits of our knowledge as well as the changing circumstances that federal policymakers will always face.