PAYGO: Improving State Budget Discipline While Retaining Flexibility
September 22, 2011
States can help policymakers and the public understand the consequences of budget decisions and properly weigh the long-term impact of alternative proposals by adopting important budget management tools that also promote fiscal responsibility. The cornerstone of these tools is "pay-as-you-go" (PAYGO), a requirement that the governor and the legislature fully offset over a 5-year period the cost of proposed and enacted increases in spending or reductions in revenues through spending cuts or revenue increases.
States can facilitate implementation of PAYGO by adopting two other tools: multi-year projections, or projecting revenues and expenditures over several years as part of the normal budget process, and current services budgeting, or basing expenditure projections on the cost of continuing the current level of services, given expected changes in the number of recipients, per-recipient costs, and other relevant factors.
Today in most states, it is impossible for policymakers to know whether proposed program increases or tax cuts are affordable over the longer term. Nor do most states have appropriate mechanisms for considering and implementing tradeoffs among fiscal policy options. These problems impede decision-making and leave states vulnerable to serious long-term budget problems.
PAYGO, together with the other two budget management tools mentioned above, could help assure that states do not use temporary spikes in revenues to enact spending increases or tax cuts that they cannot afford over the long term. It could also minimize the need for deep budget cuts and large tax increases when the economy is weak, particularly if the state deposits revenues that cannot be spent under PAYGO rules in a well-designed stabilization fund.
Today, some states make multi-year projections and some states use current services budgeting — although few states use both. No state, however, now uses PAYGO as a budget control mechanism. The PAYGO rules to which the federal government adhered from 1991 through 2000 helped lead to federal budget surpluses during 1998 – 2001.
A PAYGO approach is superior to more draconian, non-flexible limitations on revenues or expenditures because it imposes budget discipline while recognizing the need to adapt to changing circumstances in a responsible manner. And, unlike many rigid tax or expenditure limitations, a PAYGO budget process maintains legislative prerogatives over the budget and allows for the possibility of program expansions or tax reductions when they are affordable. While PAYGO is not necessarily the best path for every state, many states could benefit from it.
If the revenue estimates a state made before the beginning of the year turn out to be too pessimistic, or if an improving economy lessens the need for services more than anticipated, a state is said to end the fiscal year with a "surplus." Such a surplus, however, may not be indicative of a state's fiscal health. The state might have guessed wrong about the course the economy would take during the year, for example, or revenue collections might be particularly strong because a rising stock market led to high capital gains realizations. Nevertheless, such surpluses typically spark demands to expand programs and/or enact new tax cuts for the subsequent year, even though the surplus may not represent a permanent increase in revenue or permanent reduction in needs.
In most states, it is impossible for policymakers to weigh whether proposed program increases or tax cuts are affordable over the longer term.
- Some states only consider the upcoming fiscal year or biennium as part of their budget process, and few states consider as many as five future years. Yet it is common for states to enact policies that begin in the middle of the year or are phased in over a number of years, sometimes with costs that grow substantially over time. Without credible multi-year budget projections, policymakers and the public have no way of knowing if the policies being contemplated will be affordable over time.
- Some states practice multi-year budgeting but do not credibly project future-year costs because they do not take into account inflation, population, or anticipated caseload growth resulting from demographic or economic changes when making their projections. In other words, they fail to project "baseline" or "current services" costs — even though those costs are highly likely to materialize, absent action to change the provisions of a program. Unless factors such as these are considered, projections of future-year costs will not be accurate.
- Few, if any, states have a PAYGO requirement or other flexible mechanism that allows policymakers to make reasonable changes in expenditures and revenues while assuring that those changes are affordable over the long term. States that restrict revenues or expenditures according to a rigid formula do not have that flexibility.
- Even states with statutory or constitutional tax or expenditure limitations may enact phased-in programs or tax cuts that are unaffordable in the long term, leaving it to future policymakers to determine how to fit those enhanced expenditures within their caps or how to maintain expenditures with lower revenues.
States can address these problems by adopting a baseline, multi-year budgeting process and adhering to PAYGO, as explained below.
Policymakers routinely design changes in spending programs or the tax code so that they have a modest budget impact in the initial year or biennium but a much larger one in subsequent years. This enables policymakers to squeeze their initiatives into annual or biennial balanced-budget requirements, leaving to subsequent governors and legislatures the problem of how to balance future budgets.
If states provide budget data only for the immediate budget period, and/or limit their fiscal-impact estimates to that period, it is difficult to detect these tactics or to gauge whether the out-year impacts are affordable. Thus, policymakers considering policy changes typically give little consideration to their longer-term implications.
Multi-year budgeting can help address this problem. About a quarter of the states provide detailed budget data that extend four years beyond the current budget cycle. (Some of those, however, fail to base those budgets on meaningful out-year forecasts, as described below.) At the other extreme, more than a third of the states consider only the current budget cycle or one additional year. Ideally, states should include a five-year period in their budgets.
It is important to provide multi-year data for the total budget, for summary program areas or departments, and for individual programs. Needless to say, summary data can hide large changes in individual programs or areas. Moreover, different types of programs grow at different rates; detailed data is needed to understand what programs are — or are not — driving budget trends.
Policymakers or technical staff members that work in budget offices or legislative fiscal analyst offices are sometimes reluctant to provide longer-term projections. They may argue that even short-term estimates often prove to be wrong and that it is even more difficult to foresee economic conditions five years out. While that surely is true, a best estimate is better than no estimate at all; even an estimate for a fifth year that is 10 percent off, for example, would give policymakers a better understanding of the future impact of policies enacted today than they otherwise would have in most states.
Baseline (Current Services) Budgeting
Multi-year budgeting will not accomplish its purpose if projections are not done properly. Revenue projections should reflect credible economic forecasts and analysis of the behavior of the revenue system relative to the economy (a procedure that most states already use). Expenditure projections should reflect "current services" or "baseline" analysis — a procedure few states use.
A baseline analysis determines how much it will cost to continue existing policies and programs in future years. Such an analysis projects current-year spending (as reflected in the current adopted budget) forward based on anticipated increases in the population receiving the service or using the program and on expected changes in the per-person cost, as well as the impact of any previously enacted rule changes that have not yet phased in. Per-person costs typically are adjusted by some measure of inflation, preferably with special indices used to project increases in health-care programs or other programs for which general inflation is not the appropriate measure.
A baseline analysis does not commit the state to continue all programs and benefits; it just provides an accurate base on which policymakers can consider potential changes.
| TABLE 1:
Connecticut Makes Detailed Comparisons between the Amount Required to Continue the Current Activities, Given Changes in Population Served and Cost, and the Governor's Recommendation
|Agency Programs by Total Funds
(net of reimbursements)
|Care and Custody||525,416,967||580,125,667||558,724,128|
|Health and Addiction services||100,263,997||104,775,491||102,553,239|
|Parole and Community Services||55,293,177||58,714,055||58,226,602|
|Correctional Enterprises of CT||20,940,815||20,750,000||20,750,000|
|Considerations of Pardons and Parole||6,773,118||6,540,176||6,280,668|
|Agency Management Services||10,126,811||11,870,548||10,872,286|
|Total- All Funds Gross||718,814,885||782,755,937||757,406,923|
|Total – All Funds Net||718,814,885||764,955,630||739,606,616|
|Difference between Current Services and Recommended||-25,349,014|
A current services estimate should be done for each program, or the lowest feasible budget level, and the results then summed to determine a baseline for total expenditures. Without this level of detail, policymakers will lack the information needed to make informed budget choices.
The Connecticut governor's proposed budget provides an example of an effective use of a current services budget. It summarizes proposed increases or decreases from the baseline at the beginning of each program area section and then gives details. Table 1 shows a portion of the state's corrections spending; note that the current services estimate is higher than the prior year in some cases and lower than the prior year in others, evidence that these estimates do not reflect the application of a one-size-fits-all-programs formula but rather careful consideration of the circumstances.
The federal PAYGO rule, first in effect between 1991 and 2000, proved to be an effective part of the federal budget process. It helped lead to four consecutive years of budget surpluses (1998- 2001), which had last occurred in the 1920s.
Congress enacted PAYGO in the Budget Enforcement Act of 1990. It required policymakers to offset new legislation that would increase entitlement spending or lower taxes over the subsequent five years with reductions in entitlement spending for other programs or increased revenue from other sources to avoid adding to the deficit. A 60-vote supermajority in the Senate and a majority in the House were required to override PAYGO.*
Before PAYGO's adoption, the federal deficit was $220 billion in 1990; by 2000, the federal budget had a surplus of $236 billion. PAYGO expired in 2002, but already by 2001 — the year Congress enacted President George W. Bush's tax cuts — Congress had stopped adhering to it. In 2001, the federal budget fell back into deficit.
In 2010, at President Obama's request, Congress reenacted the PAYGO rule under the Statutory Pay-As-You-Go-Act of 2010 as a permanent feature of federal budgeting. The new law measures costs as five- and ten-year averages rather than year by year, but the principles are the same as the original PAYGO law.
*Federal PAYGO applied only to the budget category known as "mandatory" spending, more colloquially called entitlements. The federal category of "discretionary" spending had a separate budget control mechanism. States do not have such categories in their budgets; rather, they appropriate all general fund expenditures annually or biennially. Thus, this paper recommends that states apply PAYGO to the entirety of their general fund budget.
Nine states plus the District of Columbia prepare a current services baseline for the upcoming fiscal year and also project beyond that year. They are Alaska, Arizona, Connecticut, Florida, Maryland, Minnesota, New York, South Carolina, and West Virginia.
An effective PAYGO system would require policymakers to offset, for each of the next five years, the cost of any legislated reductions in future revenues below baseline projections and any program expansions that raise future expenditures above baseline levels. Policymakers can offset a tax cut or program increase by increasing revenues — either directly by raising a tax or fee, or by eliminating a tax expenditure — or by reducing on-budget expenditures for a program. Thus, expenditures and revenues would remain approximately at baseline levels over time, unless revenues were explicitly increased or programs deliberately decreased. This approach has a number of advantages:
- It allows regular adjustments in current programs to keep up with changes in costs or populations served, because the starting point for the budget process is a baseline that accommodates these changes rather than nominal prior-year spending.
- It makes explicit the cost of cutting taxes, forcing policymakers to declare the specific programs they will cut to offset the loss of revenue.
- It encourages regular scrutiny of both on-budget expenditures and tax expenditures to determine whether any programs have lost their usefulness or could be provided more efficiently, as policymakers search for offsets for proposed program increases or tax reductions.
- It makes surprise deficits in future years much less likely, except during economic downturns.
- If used in conjunction with a well-designed stabilization or "rainy day" fund, into which the state deposits revenues that exceed baseline expenditure levels, it evens out the highs and lows of expenditures over the business cycle.
PAYGO works best if the state engages in high-quality multi-year baseline budgeting with reasonably good revenue and expenditure forecasts.
Getting Started on a Baseline
The easiest way to implement PAYGO is to construct an expenditure baseline based on current spending adjusted for prior legislation that has yet to take effect, caseload changes (where appropriate), and inflation. In states that have imposed severe program reductions in recent years because of the recession, however, it could be problematic to establish a PAYGO system based on maintaining the current, reduced level of spending, since current expenditures in some program areas are well below the level required to adequately serve residents. A more appropriate approach in these states would be to construct a baseline that allows the state to reverse spending cuts enacted during the recession without offsetting the cost. The Appendix shows how a state could construct such a baseline and how to allow the restoration of recession-era cuts while still using PAYGO as a strong budget control mechanism.
It would be desirable to develop baselines for detailed subcategories of expenditures (as in the Connecticut example above) to allow annual calculations of the current services level based on utilization and cost changes. (See, however, the box on page 8.) By contrast, it should not be necessary to develop a baseline for each type of revenue, since general fund revenues for the support of general fund expenditures are fungible — that is, it is the total amount of revenues, rather than their specific types, that matters. PAYGO requirements would apply to net changes in total revenue. For this reason, changes in revenue sources would not trigger PAYGO requirements if they were revenue-neutral over the five-year period that PAYGO would cover.
The precise way in which a state would implement PAYGO would depend on the state's budget processes. Most likely, the executive budget office would be responsible for creating the current services budget for expenditures, although the legislative fiscal office might create a baseline in addition to or instead of the one created by the executive, depending on the state's budget traditions. PAYGO would arguably work best in states that have non-partisan fiscal offices or commissions that could calculate baselines and produce revenue forecasts to which the executive and legislature would adhere. But it could be used in states that do not have such institutions as well.
Some states may have to change the way they enact budgets. A state operating under PAYGO must establish some point in the budget process where the legislature approves a framework for enacting the budget (similar to the federal budget resolution) that conforms with PAYGO and/or brings together the various pieces of the budget that committees have tentatively approved (including revenue changes) in one package to determine if the package conforms with PAYGO. In many states, both changes are likely to be necessary.
PAYGO would apply solely to states' general funds. Most states do not include funds received from the federal government within their general fund, but those that do should make a rule to exclude federal funds from PAYGO. Conversely, states that do not include some programs such as education or Medicaid in their general funds should make a rule to consolidate the outlying funds with the general funds for PAYGO purposes. Dedicated funds with their own revenue sources, such as the transportation fund in most states, would not fall under PAYGO.
Even though funds that states borrow by issuing bonds would be outside the general fund and thus not subject to PAYGO, any interest payments the state would have to make from general funds on those bonds would come under the PAYGO restrictions and would have to be offset, just as any new expenditure would. Transferring funding streams out of general funds, such as dedicating a portion of the sales tax to a particular purpose, would be treated as a tax reduction under PAYGO and would have to be offset.
Some states would also have to decide how to handle under PAYGO any revenue or expenditure changes adopted through initiative or referendum votes. In states that require legislative action to implement the results of a ballot measure, it may be possible to subject the results to PAYGO. PAYGO definitely could be required for any measure referred by the legislature, and it might be possible in some states to amend initiative rules to require PAYGO for initiatives that seek to change revenues or expenditures. States in which the results cannot be subject to PAYGO should adjust the baselines to accommodate the results of ballot measures.
State Variations on PAYGO
This paper presents an outline of how PAYGO could work in a state budgeting process. Every state is different, so it is important to realize that PAYGO could work for a state in a somewhat different form than presented here.
For example, a state may not want to take the full step of preparing and/or publishing current services or baseline budgets for five years into the future. It certainly would be important to have the capability to make such projections and to calculate and present baseline budgets for the upcoming fiscal year or biennium. But for the purposes of PAYGO, a state could make current services estimates just for the program increase being proposed and for its proposed offsets. Or, in the case of a proposed tax cut, a state could make current services estimates just for any program reductions being proposed as offsets. (Revenue offsets, or the costs of tax cuts, would be projected based on economic data and models, as most states currently do.)
The potential disadvantage of such a partial use of current services budgeting is that it would be harder for policymakers and the public to consider alternative policies to those being proposed — particularly alternative offsets — because projections of the costs of all programs would not be available. There may be ways to overcome that problem, though. One example is the publication of a strong tax expenditure budget that includes five-year estimates and some sort of options document that provides five-year current services estimates for some programs.
At the other end of the spectrum, a state may want to take the opportunity afforded by implementing PAYGO to require that budgets be balanced on a current services basis for each year of the five-year PAYGO window.
Not every state needs PAYGO. Some states can be fiscally responsible without such a mechanism; in others, PAYGO would not fit well with existing budget control mechanisms. But there are many states in which adoption of PAYGO could improve long-term fiscal health.
Finally, even though PAYGO would require program increases or tax cuts to be offset for five years, states should take care to ensure that costs do not explode just outside of that window. The same would be necessary for tax changes that are revenue-neutral for five years. Thus, a state may want to institute a rule that the full cost of a policy change — as if it were fully implemented — be counted for PAYGO purposes. This could be done for the final year covered by PAYGO, which could discourage overly long phase-in periods.
Dealing With State Revenue Surpluses
Forecasted Surpluses That Exceed the Current Services Budget
In times of strong economic growth, revenue forecasts for the coming year sometimes exceed the estimated amount necessary to fund a current services budget. Under PAYGO, the state would not be permitted to use the "surplus" revenues to increase program spending; any increase would still require a statutory increase in revenues or a reduction in other spending.
This requirement makes sense for a number of reasons. If a state budgets its expenditures based on the level of revenues collected when economic growth is strongest, it will not be able to maintain those expenditures when economic growth slows. In addition, some types of revenues are more volatile than others. Income tax revenues from capital gains realizations are particularly volatile across the business cycle, and a state cannot rely on high capital gains revenues in a strong-growth year to fund a recurring expenditure.
A PAYGO process should specify the use of revenues that are forecast to exceed the current services budget for the upcoming fiscal year or biennium, and for the five-year period. There are several options:
- PAYGO would work best if a state required such revenues (or a specified portion thereof) to be placed in a "rainy day" fund or other type of reserve fund. The state could then use those funds in years in which revenues are forecast to fall below the current services budget, thereby stabilizing its budgets across the business cycle. The rules regarding the rainy day or stabilization fund would have to allow the state to build up sufficient funds when the economy is strong and use those funds without undue barriers when the economy is weak. 
- The state could use such revenues (or a specified portion thereof) for non-recurring expenses, such as a capital expenditure that otherwise would be funded by bonds or the non-recurring portion of the cost of upgrading a computer system. Similarly, the state could use the revenues to retire previously issued bonds if that is financially advantageous.
- The state could use the revenues to pay down a portion of any unfunded liability in the state pension fund that has developed because of investment losses or prior actions relating to the fund. The state should not use the revenues to pay the normal cost of pensions (i.e., the cost of the pensions earned in the current year) because that is a recurring cost.
- The state could specify circumstances under which it would return such revenues to taxpayers as a one-time tax rebate. This might be appropriate, for example, if the reserve fund had reached an adequate specified percentage of the annual budget. It would not be appropriate to use the revenues for a permanent tax cut, just as it would not be appropriate to use them for a permanent program expansion.
Of course, a forecast is not actual revenue, so a temporary allocation of the "surplus" expected revenues should be made at the time the budget is enacted, with the allocation confirmed to the extent the revenue materializes.
It is common for a state to have "unexpected" revenues — i.e., revenues that were not forecast but materialize during the fiscal year — in the aftermath of a recession, because forecasters tend to be cautious in predicting the strength of an economic recovery. PAYGO rules regarding unexpected revenues could vary depending on the situation in which they occurred.
Sometimes unexpected revenues occur when the state has budgeted expenditures below baseline (usually because the revenues forecast were insufficient to allow spending up to the normal baseline level). For such situations, a state could create a rule allowing some unpaid-for expenditure increases to restore total expenditures up to baseline levels in light of the unexpected additional revenues. Those potential expenditure increases could follow a normal legislative process in legislatures that are in session for most or all of the year or have a tradition of special sessions, or could take place automatically. The District of Columbia, for example, sometimes creates a prioritized list of contingent expenditures, to be funded if revenues come in higher than forecast.
If unexpected revenues occur in years in which expenditures had been enacted at baseline, the state should treat the unexpected revenues as one-time revenues that cannot support ongoing expenditures or tax reductions. The acceptable uses would be similar to those listed above for forecasted surplus revenues.
PAYGO in Future Recessions
It is common in recessions for states to reduce expenditures below baseline because revenues are not available to support baseline spending. But if PAYGO baselines for future years are based on those depressed expenditures, a "ratchet" effect would occur, under which it could be difficult or impossible to restore service levels after the economy improved. Special PAYGO rules are necessary to prevent a ratchet.
For example, in any year in which revenues are forecast to grow by less than a specified percentage — for example, 3 percent, 4 percent, or 5 percent, depending on the history of revenue growth in the state — the expenditure baseline could be "frozen" at the prior year level. For any future year, the expenditure baseline would be the higher of the frozen baseline and the normally calculated baseline.
Alternatively, the state could freeze the expenditure baseline but with some adjustments. A baseline that is frozen for a number of years during a prolonged economic downturn would arguably be inadequate by the time a normally calculated baseline would exceed it because the population will have grown and the need for services may have increased. The current downturn provides an example of this phenomenon. The U.S. Education Department projects that there will be about 260,000 more K-12 students in the 2011-12 school year than in 2007-08, before the recession began, and 1.5 million more public college and university students. And an estimated 5.6 million more people will be eligible for Medicaid in 2012 than were enrolled in 2008, due to the loss of jobs, wages, and job-based health coverage. To at least partially account for this problem, the state could adjust the expenditure baseline using the type of procedure outlined in the Appendix.
Enforcement of PAYGO
The PAYGO enforcement mechanisms a state chooses will likely depend on its current budget process rules and traditional budget enforcement methods. One approach would be to allow a point of order to be raised against any piece of legislation that violates PAYGO. The legislation could not move forward unless it was amended to include sufficient offsets. PAYGO could not be waived.
Another option would be to require a separate, explicit vote of the legislature to override PAYGO for a given legislative proposal. Depending on the parliamentary procedures in use in a state, this could occur by allowing a point of order to be raised against any legislation that violated the PAYGO rules, but in this case the point of order could be overcome though a vote. This would force legislators to go on record that they favor breaking the PAYGO rules, and would alert the media and the public to that fact.
A state could require a supermajority vote to override the PAYGO rules. Unless a state already has a tradition of requiring supermajority votes for budget actions, however, it makes sense to avoid using a supermajority requirement; such requirements in California and elsewhere have often allowed a minority to hold up budget changes that are widely deemed necessary, resulting in prolonged gridlock.
While a state could adopt PAYGO as a legislative rule or enact it as a statute, it should not enshrine PAYGO in its constitution. PAYGO rules can help a state maintain an appropriate level of fiscal discipline under current conditions, but it is impossible to predict what circumstances will be like 25 or more years from now. Once fiscal policy is embedded in a constitution, it becomes difficult or impossible to change, even if it no longer useful or appropriate — or even if it is harmful.
State Experience With PAYGO-Like Rules
No state currently uses PAYGO. But a few states have budget process elements that help stabilize their budgets and mitigate excessive program expansions or tax reductions resulting from temporary revenue spikes.
Massachusetts enacted legislation in 2010 that seeks to compensate for the volatility of capital gains realizations by requiring the state to transfer all capital gains revenues that exceed $1 billion in any fiscal year to the Commonwealth Stabilization Fund.  The state would then draw down the funds when the economy is weak and capital gains realizations are low in order to stabilize the budget.
A Virginia statute requires a PAYGO-like assessment process for corrections spending; as an official explained:
In Virginia, all sentencing changes proposed by lawmakers are evaluated by the Virginia Criminal Sentencing Commission that projects the effect on correctional resources and any additional costs. All proposed laws are given a price tag based on the commission's analysis. When sentencing laws pass the public safety committee, Virginia lawmakers must go before the appropriations committee to identify cuts in other government services or increases in revenue to pay for the new law.
However, enactment of corrections changes is not contingent upon enactment of the offsets, which are handled in the overall budget process.
In the District of Columbia, legislators must pay for all tax cuts and program expansions by showing that they will not push the budget out of balance over the next four years. That certainly helps budget discipline, but it does not necessarily prevent tax cuts or budget growth based on temporary revenue increases that are built into the revenue forecasts. Projected increases from revenues currently on the books can be used to increase programs well beyond baseline or to cut taxes without triggering a need for offsets.
In 2011, Colorado legislators considered a proposal to implement a PAYGO system in the budget process for the first time.  It would have required all measures that reduce revenue to pay for themselves through an offsetting revenue increase or spending cut.
PAYGO Superior to Rigid Tax and Expenditure Limits
There are many types of state tax and expenditure limitations around the country. None of them tie their limit on taxes or expenditures to any measure of the level of need for expenditures; instead, most use some arbitrary measure such as inflation, inflation plus population growth, or personal income growth. Moreover, none of them prevent a state from responding to a temporary increase in revenues by cutting taxes to a level that is insufficient to maintain the current level of services, and most require either a supermajority vote or a citizen vote to raise revenues, so it becomes extremely difficult or impossible to recover from an overly-generous tax cut made on the basis of temporarily high revenues. Finally, most tax and expenditure limitations do not have features that help states weather recessions.
PAYGO is far more fitted to the challenges that states face in providing needed services, and it is more flexible. Current services projections for expenditures, if done properly, will reflect the need for expenditures at different points in the business cycle — such as the higher Medicaid funding needed when the economy is weak and more people qualify for the program, and the lesser amounts necessary when the economy is stronger.
Moreover, PAYGO evens out the highs and the lows of revenues and expenditures by preventing states from expanding programs or cutting taxes based on temporarily higher revenues. This can help states meet their balanced budget requirements without making massive cuts in services, large tax increases, or resorting to more harmful forms of budget gimmicks, particularly if states deposit "surplus" or "windfall" revenues in a well-designed stabilization fund.
Finally, PAYGO does not prevent program expansions or tax cuts; it just requires that they be enacted in a responsible manner.
PAYGO, if used in combination with current services budgeting and multi-year projections, could improve the budget processes of many states by stabilizing revenues and expenditures. It imposes budget discipline but lets the state retain the flexibility to adapt to changing circumstances. While no state currently uses PAYGO, it is a far better choice for a budget control mechanism than rigid tax and expenditure limits.
Constructing a Starting Point for PAYGO
In states that imposed severe program reductions during the recent recession and its aftermath, it would be appropriate to begin PAYGO with a constructed baseline rather than an actual current services level. The best way to construct such a starting point for PAYGO would be to extend fiscal year 2008 expenditures to the current year. For example, one could assume that the 2008 expenditures would have grown by the state's historical annual rate of growth of expenditures or by some fraction thereof, such as two-thirds or one-half.
Consider a hypothetical state with general fund expenditures of $10 billion in fiscal year 2008. If it followed the average expenditure pattern of states in the recent recession and its aftermath, it would be spending about $9.848 billion in 2012 — a substantial reduction, given the growth over that period in both the populations to be served (such as schoolchildren and the elderly) and the per-person cost of providing services.
Table 1 shows the results. We assume that expenditures would have grown at 5 percent a year from 2008 to 2012 if there had not been a recession and a sharp drop in revenues — a rate typical of long-term annual expenditure growth across states. The state's 2012 baseline would therefore be $12.155 billion rather than $9.848 billion.
| Example Method of Creating a Baseline Starting Point for PAYGO Hypothetical State
(Dollars in Millions)
|Fiscal Years||2008||2009||2010||2011 Est.||2012 Est.|
|Adjusting Using Historical Growth of Expenditures*||$10,000||$10,500||$11,025||$11,576||$12,155|
|Adjusting Using 2/3 of Historical Growth in Expenditures||$10,000||$10,333||$10,850||$11,393||$11,962|
|* Assumed to be 5% per year. NASBO shows the average historical growth across all states from 1979 to 2011 to be 5.7%.|
This example creates an overall expenditure baseline for the state general fund. In reality, a state would also have to set initial expenditure baselines for subcategories of expenditures, because once the initial expenditure level is established, it should be adjusted each year in the normal way current services are calculated, based on utilization and cost changes. This can only be calculated at a program level or at the level of a category of programs. The program or subcategory beginning baselines could be established using a procedure similar to establishing the overall baseline.
Constraining Restorations to Maintain Budget Discipline
One purpose of a constructed baseline is to allow a state to restore programs to pre-recession levels — or to some agreed-upon proportion of pre-recession levels — without having to count those restorations as new expenditures, while creating a benchmark level above which program increases must be offset. Of course, revenue would have to be available to support the restorations. Presumably, as revenues recover along with the economy, revenues would be available to support some restorations.
There are various other ways that a state could allow program restorations without losing the budget discipline of PAYGO. One way would be to create a specific list of restorations that would not trigger PAYGO. Examples might include changes in school aid formulas that reduced per-pupil aid, elimination of medical services covered under Medicaid (or elimination of coverage under Medicaid for certain categories of residents), cuts to child care or pre-kindergarten programs that reduced the number of children that could participate, and reductions in aid or services for disabled residents. A state could make some or all such cuts eligible for restoration, as determined through a committee or commission process.  PAYGO would continue to be in force for the enactment of any other programs, any new programs and for any enacted reduction in revenues.
Transitioning to an Actual Baseline
A state could choose to allow the constructed baseline to be in effect for a specific number of years. Alternatively, the constructed baseline could remain in effect until actual expenditures projected for the upcoming fiscal year equal or exceed the constructed baseline, or until unemployment in the state has shrunk to its normal non-recession level.
 A described below, there are a number of options of ways to use "surplus" or "windfall" revenues that can strengthen, rather than jeopardize, the long-term fiscal health of the state. Moreover, the potential use of revenues that exceed forecasts could depend on whether expenditures were enacted at baseline levels or whether revenue constraints had resulted in below-baseline expenditure levels.
 There are many examples of phased-in tax reductions. The most recent one is from Arizona. In the 2011 legislative session, Arizona enacted a tax package that reduces the corporate income tax rate to 4.9 percent from 6.98 percent and reduces commercial property taxes by 10 percent. The package will cost the state $38 million in fiscal year 2012, but by fiscal year 2018 the cost will balloon to $538 million, half of which will result from the corporate tax rate cut.
 A state might also consider other factors, such as provisions of collective bargaining agreements or anticipated sharp increases in employee health insurance costs. In addition, it might have to adjust baselines to accommodate the cost of a new federal mandate or requirement or to reflect the transfer of a program from the general fund to special funds — to avoid the potential of manipulation for PAYGO purposes.
 Connecticut has a biennial budget process. The budget gives current services levels and proposed spending for each year of the biennium, but Table 1 shows only one year. Connecticut only projects at this level of detail for the two years of the biennium; the state publishes a second document (along with the budget) that makes current services projections for the subsequent three years at some level of detail, but the categories do not necessarily match those in the budget.
 Revenue forecasts typically specify revenue amounts by type of tax, which is useful for other purposes.
 The PAYGO requirement of projecting into the future should help assure (if the state's revenue projections for the five-year period are accurate) that changes that proponents claim are revenue neutral really are revenue neutral over time. This is important because different types of revenues grow at different rates relative to the economy (technically, have different elasticities), so a change that is revenue neutral in the first year may not be over five years.
 Many states' rainy day funds do not meet these criteria. Thirty-three states and the District of Columbia cap fund accumulations at inadequate levels; 10 states either require a supermajority vote to release the funds or place an arbitrary limit on how much of the fund can be released at one time; 12 states and the District of Columbia require the state to replenish the funds quickly after they are used, even if economic conditions have not improved. See Elizabeth McNichol and Kwame Boadi, Why and How States Should Strengthen their Rainy Day Funds, Center on Budget and Policy Priorities, February 3, 2011.
 If revenues are below a constructed baseline, the suggested rules allowing program restorations would apply.
 A state also could enact contingent tax increases to allow expenditures to rise back up to the baseline, with explicit provisions for rolling back the tax increase without PAYGO offsets if total revenues reach a specified target level.
 The state then allocates 5 percent of the transferred amount to the State Retiree Benefits Trust Fund.
 Richard P. Kern, Director, Virginia Sentencing Commission, testimony to the California "Little Hoover Commission," August 24, 2006.
 House Bill 11-1052, PAYGO, sponsored by Rep. Hullinghorst and Sen. Morse, was defeated in the House Finance Committee.
 Based on National Association of State Budget Officers data. The 2012 data is for governors' proposed budgets.
 Inevitably, the question would arise of whether to allow an increase in a program that does not restore a previous cut but has the same cost and accomplishes a similar purpose An example might be if legislators wanted to enact a refundable credit for child care expenses rather than restore funding in a direct child care subsidy program. In creating a PAYGO process that includes a constructed baseline, a state would have to answer that question.