In the Nation's Interest
The Discretionary Wild Card
By Joseph Minarik
Two weeks ago, we talked about the onset of the sequester – which applies almost exclusively to “discretionary” spending, that is, annual appropriations for federal agencies – and what it implies for both the macroeconomy and the performance of government functions.
Next week, a further development will pull that issue back on stage: House Budget Committee Chairman Paul Ryan (R-WI) will release his budget resolution for next year. As part of the deal struck within the House Republican caucus to pass the fiscal cliff bill at the turn of the calendar year, that resolution will claim to achieve a balanced budget – a deficit of zero – in the tenth year from now (that is, in fiscal year 2023). To reach budget balance that quickly – last year’s House resolution was estimated to achieve balance in 2040 – will require greater savings, and some of those additional savings likely will come from a lower path for discretionary spending than is set under current law – that is, even with the sequester.
Then, to avoid a government shutdown (that is, a lapse of appropriations for federal agencies), the Congress will need to replace the current continuing resolution before its expiration on March 27. That will highlight once more the implications of the level of annual appropriations (although its effect will be restricted to the current fiscal year – fiscal year 2013, which ends on September 30, 2013 – more on that point later).
And at some time in March or April, the President will submit his fiscal year 2014 budget. Many people will rush to see how the budget proposals compare with the House budget resolution, and in particular its FY 2023 budget balance. The President’s proposed future-year levels for discretionary spending will be an important element in determining how that comparison will look. So for yet a third time in just a few weeks, proposals for levels of discretionary spending will get at least a little air time – even though we do not yet have specialty cable channels that focus on the federal budget.
Background: The right way to construct a budget, as any public finance textbook will tell you, is to weigh the benefits of each individual program with its costs, and to fund each program only to the point where it provides a marginal net benefit. In practice, that is an awfully high bar. And standards of practice have fallen markedly in recent years. But even assuming the very best practice, given that appropriations are funded one year at a time, any longer-term budget plan requires some extrapolation of current levels into the future. Hence this discussion.
Assessing claims of budget savings out of future-year discretionary spending is both impossible and difficult.
Why is it impossible? Consider federal entitlement programs, which are written into permanent law and continue unchanged unless the law is changed. Any proposal to change any such entitlement program in any future year can be compared precisely with the law that otherwise would apply. Winners and losers can be identified clearly – and the losers can protest in explicit terms.
Future appropriations cuts cannot be so scrutinized. Appropriations laws extend for one year and one year only, and provide very little information about future years (except to the extent that they bind the federal government to long-term contracts). So Chairman Ryan and the President can make their own proposals for discretionary spending in fiscal year 2023, and no one can challenge them. We don’t yet have appropriations laws for fiscal year 2023 to which we can compare the proposed changes. We don’t yet have appropriations laws for fiscal year 2022. Heck – right now, we don’t even have an appropriations law for fiscal year 2013.
So someone can propose an outlandishly low appropriations total for fiscal year 2023, and if someone else were to complain that some important program would be cut to ineffectiveness, the advocate could answer that, no, we won’t cut that program – we will cut something else. And no one can prove, or even argue on the basis of any tangible evidence, that such a cut in a specific program would in fact occur. This happens year after year, in Congressional budget resolutions and in Presidents’ budgets. Both parties do it. Out-year appropriations spending cuts are the tool of choice for achieving future deficit-reduction targets – because they are utterly resistant to fingerprints.
In the late 1980s, former Representative Lee Hamilton (D-IN) was asked why the Congress was making so little progress in reducing the budget deficit despite all of the passionate declarations on the topic. Rep. Hamilton thought for a moment, and then said that “Reducing the federal budget deficit is the second-highest priority of every Member of Congress.” The audience got the idea very quickly: Every Member of Congress is bound and determined to cut “something else.” But after all of those Members protect their highest priorities, there isn’t much “something else” left. Remember that no expenditure is appropriated by accident; every expenditure is in the law because someone wants it there. And that Member (or those Members) will resist if their priorities are slated for repeal or reduction.
So that is why assessing claims of future-year discretionary spending cuts is impossible. But people continue to make those claims, and so other people continue to try to assess them. And there are important reasons why evaluating those claims, beyond being impossible, is also difficult.
Unlike the entitlement part of the budget, which is numerically dominated by Social Security, Medicare, and Medicaid, the discretionary budget (especially the non-defense part) is a mass of a large number of relatively smaller programs. Enacted appropriations laws contain line upon line of detailed budget accounts funded with specific numbers of dollars. But future year appropriations levels are only highly aggregated numbers, with virtually no underlying detail. “Defense” and “nondefense,” or “security” and “non-security,” are about as specific as these numbers get. So again, any outlandishly low proposed number for a future year can be said in one conversation to be sufficient to fund programs A, B and C, in a second conversation to cover programs L, M and N, and in a third to fund programs X, Y and Z, even if the total amount would fall far short of covering all of those programs. And again, no one can disprove such mutually inconsistent claims.
So instead of searching in an infinitely large haystack for a very small needle, analysts tend instead to look for aggregate rules of thumb: How much must total discretionary spending increase to continue doing what government is doing now? That generalization helps a little, but it falls far short of answering any specific questions.
One possible rule is a “freeze:” the same number of dollars this year as last year. Some justify a “freeze” rule by saying that the private sector does not expect to get an automatic raise each year, so government should not expect an automatic raise, either. Such a rule would be fine, provided that its advocates are prepared to tell private vendors who sell to the federal government that their prices will be frozen, too. So, for example, the supplier of asphalt should be ready to sell for the same numbers of dollars even if the price of the oil in the asphalt rises. And those who argue that federal government employee salaries should be frozen, also should be prepared to freeze the salaries of private-sector vendors that sell to the government.
Because those conditions generally are not acceptable or realistic, analysts typically move on to alternative rules with more flexibility. There is the “baseline,” under which the federal government’s costs for salaries increase at the same rate as salaries in the private sector, and its costs for purchased goods and services rise at the same rate as similar prices in the private sector. Generally speaking, if appropriations increase at the “baseline” rate, then the federal government can continue doing next year exactly what it is doing this year.
But for some, doing exactly the same thing next year is not enough. They see the federal government as providing services to the population, and believe that therefore discretionary spending needs to increase at the rate of growth of the population, in addition to the rate of growth of prices. Others go still further, and say that at least some activities of government provide the “overhead” for the workings of the economy, and therefore believe that discretionary spending needs to keep pace with the size of the GDP – and therefore, that the ratio of discretionary spending to the GDP needs to be about constant. To put all of this into perspective, the rate of growth of the GDP is very roughly equal to the rate of growth of prices, plus the rate of growth of the population, plus the rate of growth of productivity; and so those alternative rules of thumb for the size of appropriated spending – the baseline, inflated spending per person, and spending per dollar of GDP – are ordered essentially systematically along those lines.
Apart from its prescriptive value, the measure of discretionary spending as a percentage of GDP has a descriptive use. Analysts can compare levels of spending as a percentage of GDP over time in a meaningful and consistent way, and in particular can look into the future and obtain a reasonable understanding of the orders of magnitude of policy change that will be needed to reduce the deficit and control the public debt. So discussion of the ratio of discretionary spending to the GDP is common, even among those who differ about whether such spending should bear any consistent relationship to the GDP.
But back to its appropriate size: House Budget Chairman Ryan, in his breakfast talk to a Wall Street Journal audience in January, argued strongly against a “floor” for discretionary spending as a percentage of the GDP ( click here for part one and part two of the talk). Chairman Ryan pointed out that if spending growth is constrained successfully, the growth of the economy will naturally lead the measure of spending as a percentage of the GDP to progressively lower levels. If such restraint is a live possibility, then setting a spending floor as a percentage of the GDP clearly would be inappropriate.
Discretionary spending growth could be restrained through productivity growth within government. If government is behind the curve in process improvement, then there could be a one-time catch-up in productivity. Beyond that, continuing process improvement could yield further ongoing productivity growth.
However, the base of discretionary spending to which productivity improvements might apply is less than the total. There might even be components of the total where one might expect growing costs, whatever the growth of productivity. For that reason, one might accept that there should not be a floor on the ratio of discretionary spending to the GDP, but it also might not be appropriate that there be a ceiling, either. (One might go one step further and ask whether, if a floor on discretionary spending as a percentage of the GDP is inappropriate, should a ceiling on revenues as a percentage of the GDP be considered inappropriate, too?)
But what about productivity growth in the federal government? And what about significant components of the federal government’s discretionary spending where needs might be growing faster than the GDP?
On the productivity front, there are parts of the government’s activity that are uniquely resistant to process improvement, identified by the so-called “Baumol effect,” named after noted economist William Baumol. The quintessential example of the Baumol effect is grants to symphony orchestras. One wag pointed out that the orchestra cannot play Beethoven’s fifth symphony 3.2 percent faster each year. Baumol argued that government is uniquely stuck with the responsibility to finance activities such as this, which entail personal services in which the potential for productivity growth is constrained.
But many would react today that the severity of the “Baumol effect” is overstated. Specifically, some of the activities that Baumol considered inherently resistant to productivity growth no longer are seen in that light. We observe process improvement in many personal services – for example, the cost of hair-cutting services has been the object of process improvement and vigorous competition in the private sector. We demand process improvement in health care; in fact, if we don’t get it, we will give new meaning to the admonition spoken by John Maynard Keynes that “in the long run, we are all dead.” (See CED’s prescription here) The cost of higher education, a significant part of which is borne by the federal government, is running up the exponential curve even faster than health care, and must be controlled through process improvement (as CED has explained here). And we even could choose to eliminate grants to those quintessential symphony orchestras; some have argued that if the symphonies cannot compete in the marketplace, they do not deserve to survive, which would close forever that branch office of the Baumol problem. (I have views, but I will spare you.)
However, at least the symphony case does raise one additional problem area for productivity growth in government. A non-trivial part of discretionary spending is grants to state and local governments, and to private institutions. It will be more difficult for the federal government to reach through those transactions to those entities and to engineer productivity improvement in their operations, than it will be for the federal government to improve its own processes. Other parts of federal discretionary spending are purchases from the profit-making private sector, which is assumed by some to be inherently competitive and effective in process improvement. It would at least seem heavy-handed and optimistic that the federal government could somehow enforce greater process improvement and productivity growth on its contractors than is generated in the rest of the private sector. Those portions of the base of discretionary spending might not be thought to be subject to any quantum productivity improvement beyond what already is incorporated into the “baseline.”
Furthermore, there are significant parts of the discretionary budget where growth in need might far overwhelm not only the rate of growth of productivity, but also the full growth of the GDP.
One example is the cost of the Veterans healthcare system – which is a classic and touching good-news bad-news story. Improvements in battlefield medicine have been enormous. Many of our wounded servicepersons have made it home because of those improvements. The bad news is that almost by definition, those who are saved because of those improvements are the most complex cases. Those servicepersons will remain in need of our care, and this care will remain very costly. Meanwhile, veterans with less serious conditions will age, and will become newly eligible and more in need of care. (Cash disability and pension obligations also will rise, but those costs are legislated as entitlements and are included in that part of the budget. Veterans health care is funded in annual appropriations.) Many do not realize, but the Veterans medical system is quite large. The VA is the largest employer in the federal government after the Department of Defense, and constitutes almost one-quarter of all non-DoD federal employment.
The problem is not a lack of productivity improvement in the VA system. As was mentioned last week in the discussion of Medicare, VA health care is by any reasonable standard a success story, though productivity always can be improved. And policymakers could choose in whatever fashion to means test eligibility for the program. But such changes would merely be nibbling around the edges. The caseload, and the intensity of need in the beneficiary population, are growing massively – and will overwhelm potential additional productivity growth and any cost sharing on the part of those who would by intention and design be the least-needy cases.
A second area of out-of-scale need is infrastructure – and in particular, deferred maintenance, even apart from the needs imposed by ongoing population and economic growth. The American Society of Civil Engineers estimated in 2009 that the nation faced a backlog of deferred maintenance on transportation, water and sewer systems, and other components of infrastructure that would require $2.2 trillion to liquidate over five years. (These estimates are updated every four years, and the next release is due later this month.)
Now, to be sure, the ASCE is not exactly a disinterested party on this issue; and one can argue that individual infrastructure projects, even for maintenance, must be evaluated on cost-benefit grounds. So not every engineering deficiency needs to be corrected, and not every deficiency that needs to be corrected must be addressed within five years. And some of those infrastructure components are state- or local-government obligations – though the federal government typically shares the financial responsibility through grants. But more-demonstrably-objective observers have said for some time that other nations around the world have achieved advantages over the United States in terms of the capacity, speed and reliability of infrastructure. And as maintenance continues to be deferred under the ongoing fiscal pressures, the backlog will only continue to grow.
And the key point in this connection is that the growth of the backlog of infrastructure maintenance need not bear any relationship to the rate of growth of the GDP; it can exceed it, substantially – as the $2.2 trillion estimate suggests. And many would argue that the sum of the need for new infrastructure plus the nation’s maintenance backlog are a high priority that will challenge attempts at achieving substantial savings in the discretionary-spending budget.
The conclusion of all of this is not that annual appropriations must increase, or even that they cannot be cut. It is, rather, that the nation’s budget policymakers are taking a high-stakes leap into the void. Three years ago, without the resources of either the executive or the legislative branches behind them, both the Bowles-Simpson and the Domenici-Rivlin panels recommended substantial reductions of appropriations from their then-projected paths. The most knowledgeable experts said that those recommendations were highly adventurous, but possibly doable. In August of 2011, the Congress and the President put what amounted to those recommendations into law, with no meaningful analysis of the changes in policy that would result. Then, one week ago, Washington policymakers accepted stalemate and allowed further across-the-board spending cuts – the “sequester” – on top of that unanalyzed leap of faith.
Now, it appears that the Congress will triple-down on those two rounds of speculative, unanalyzed future spending cuts with still more assumed but unspecified reductions. Ironically, these optimistic assumptions about future discretionary spending cuts are becoming the most-effective Washington argument against healthcare reform. House Budget Committee Chairman Ryan reportedly proposed bringing forward the effective date of his market-based Medicare system, which bears significant similarity to CED’s proposal, to help to achieve his tenth-year budget balance. By all reports, his Republican colleagues refused to accept that, on the ground that Medicare reform was too politically risky. They told Chairman Ryan to go back to the spreadsheet and find the necessary savings somewhere else. So healthcare reform, reportedly, will no longer be a part of Chairman Ryan’s ten-year budget proposal. And where in the budget did Chairman Ryan go for those extra billions of dollars to keep his tenth-year numbers in balance? I don’t know for sure, but I can guess.
And soon, the President, and Senate Budget Committee Chair Patty Murray (D-WA), will present their budget visions. Neither will attempt to balance the budget in 2023; but both recognize that for many Americans, Chairman Ryan’s claim that his budget balances while the Democrats’ do not ipso facto will demonstrate that his is better. So the President and Senator Murray will want to get as close as possible to that goal. If Chairman Ryan claims highly speculative but non-disprovable (sorry for the double negative) future-year discretionary spending cuts, they will do the same.
And there we will be, with both parties assuming the solution to the problem through policies that have not in any way been tested, and that by any reasonable judgment actually substitute for an attack on our real budget woes. If those assumptions prove unrealistic, then the nation will be that much further behind the curve, with that much more debt on board, and with that much less time to address the real problems – including what will be a highly contentious reform of health care. It is so much easier to assume that you have a wild card than actually to face up to the issues.