Revenue “Triggers” for State Tax Cuts Provide Illusion of Fiscal Responsibility
Ten states (including the District of Columbia) have enacted laws in recent years that cut personal or corporate income taxes at a future date, contingent on state revenues first reaching a certain level or growth rate. Though these “triggered” tax cuts — like other tax cuts enacted with a significantly delayed effective date — are sometimes portrayed as fiscally responsible, they are nothing of the sort. Lawmakers enacting them typically have no idea if they will be affordable or desirable when they take effect, and they can cause deep and lasting damage to a state’s ability to invest in its people and communities.
Triggered tax cuts have caused serious financial problems in states that have adopted them.Triggered tax cuts have caused serious financial problems in some states that have adopted them (see the Appendix) and will likely continue causing problems as they kick in in additional states, because they:
Are based on inadequate information about projected revenues and spending. When lawmakers approve them, they often lack up-to-date, multi-year forecasts of what the tax cuts will cost when they take effect, and they almost never know what state services will likely cost at that time. Both kinds of forecasts could be done in most cases, but multi-year estimates of revenue losses are the exception rather than the rule, and long-term estimates of the cost of providing “current services” (taking into account inflation, population growth, and caseload growth) are almost never done. Without that information, policymakers cannot responsibly evaluate the tax cuts’ impact on state services.
For example, none of the ten states that have enacted triggered income tax cuts in recent years estimated the cost of providing existing services over the full period over which tax cuts might take effect.
Moreover, even the best forecast cannot predict how state needs and citizen preferences might change between the time the tax cuts are enacted and the time they take effect, and whether policymakers and the public will still see them as a priority.
- Can take effect even during economic downturns or at other times when revenues are particularly needed. For example, a tax cut triggered in Oklahoma last year amidst a severe state economic downturn caused by falling oil prices. Several triggers are based on achieving modest revenue growth over a single year, meaning that new tax cuts can take effect as soon as a state’s economy starts recovering from a recession even though revenues remain well below pre-recession levels. A triggered tax cut could easily take effect just as a natural disaster or a cut in federal assistance hit a state treasury.
- Typically fail to account for state fiscal needs. Tax cuts can trigger in most states even if recent revenue growth has not been enough to offset inflation, population growth, and other factors that affect the cost of current services. And they almost never account for every state’s need to maintain adequate financial reserves for fiscal and other emergencies.
- Offer no meaningful benefits compared with deferring action on tax cuts until closer to the implementation date, when policymakers will know more about whether they are affordable. At any future point when lawmakers decide that tax cuts are affordable and/or a priority, they can be put into effect quickly. While knowing future tax rates may help individuals and businesses decide whether to make certain long-term investments, these benefits are likely quite limited when it comes to state taxes. State personal and corporate income tax rates are relatively low to begin with and thus unlikely to tip investment decisions. Also, most of the triggered tax cuts enacted in recent years involve personal income taxes, and only a small minority of income tax payers are business owners making significant capital investment decisions.
Enable policymakers to claim credit for cutting taxes while avoiding accountability for the consequences. Enacting an income tax cut with a future effective date — whether or not a trigger is attached — effectively acknowledges that the cut is not affordable now: given state balanced-budget requirements, lawmakers would have to enact politically unpopular cuts to state services or offsetting tax increases. The delayed effective date offers the best of all worlds. Policymakers can garner political credit for approving the tax cuts, even if taxpayers won’t actually benefit for years, and gamble that the cuts will turn out not to harm public services or the state’s financial stability down the road.
Most important for many policymakers, enactment establishes the tax cuts as a top fiscal policy priority, forcing future lawmakers to take a highly visible vote to rescind a tax cut that has already been promised. Not surprisingly, triggered tax cuts are rarely modified, let alone repealed: many state lawmakers have signed public pledges not to vote for tax increases and would fear being accused of violating that pledge if they voted to postpone or cancel a triggered tax cut — even one that proved unaffordable.
No simple formula can take the place of careful deliberation by lawmakers with the information they need to make good decisions, yet that is what triggers typically are — a highly inadequate formula for estimating whether tax cuts might be affordable built on very limited information about the future. Lawmakers who agree to cut state revenues without knowing whether the cuts will be affordable abdicate their responsibility to prudently manage state finances, often at significant cost to the state’s future.
Which States Have Enacted Triggered Tax Cuts?
Since 2002, lawmakers in ten states (including the District of Columbia) have enacted what were intended to be permanent tax cuts to which triggers were attached (see Table 1 at the end of this report). Most involved personal income taxes, with corporate income taxes the second most common.
Triggered tax cuts have fully taken effect in North Carolina and West Virginia. Multi-year triggered tax cuts have taken effect for some years in the District of Columbia, Massachusetts, and Oklahoma. Triggered tax cuts in Kansas, Maine, Michigan, Missouri, and New Hampshire are likely to take effect in future years.
Maine enacted triggered tax cuts as free-standing legislation. In Massachusetts and Michigan the triggered tax cuts were included in bills that increased taxes. In the other seven states, the triggered cuts were in larger bills that also included tax cuts that took effect more or less immediately or automatically in the future.
The trigger mechanisms in these laws vary significantly by state. See the Appendix for a discussion of the triggered tax cuts that have already been implemented in several states.
The Political Attractiveness of Triggered Tax Cuts
States do not legally lock in the path of future spending; they budget two years into the future at most. Enacting tax cuts that will begin and/or continue beyond the budget period increases the odds they will take precedence over spending needs and priorities at the time they take effect. The tax cuts present a fait accompli to any future policymakers or citizens who might prefer to allocate the higher revenue that would otherwise be available to different priorities: for example, making new investments in education or infrastructure, expanding the state’s “rainy day” fund, addressing a long-range fiscal problem (like an underfunded pension fund), or substituting a different type of tax cut.
Enacting triggered tax cuts also enables policymakers to avoid accountability for the consequences. They can immediately claim credit for voting for tax cuts that — given state balanced-budget requirements — might force politically unpopular service cuts if they took effect right away, and gamble that the cuts will not harm public services or the state’s financial stability down the road. Disavowing responsibility for harmful tax cuts is even easier if officials can point out that the cuts took effect only after a supposedly “fiscally responsible” trigger mechanism allowed them to.
Proponents respond to criticisms of delayed and triggered tax cuts by noting that future policymakers can always reduce, postpone, or cancel the tax cuts — and can fix adverse consequences of badly designed triggers. For example, a New Hampshire lawmaker defended his state’s trigger by asserting: “A future legislature could come in and say ‘Even though we have a trigger met, we have a [revenue] problem . . . and we need to fix it. . . . It’s not like it’s constitutionally protected.”
It is true that a future effective date for a tax cut constitutes a political obstacle to setting different priorities, not a legal one. But at a time when a large number of lawmakers have signed public pledges never to vote for a tax increase — and many who haven’t are concerned that voting to scale back or suspend a scheduled tax cut would open them to attack for raising taxes — it is a significant obstacle. Also, several states have supermajority requirements to increase taxes.  Simply stated, “once a tax cut is set in law, it is politically hard to reverse.”
Policymakers Typically Have No Idea How Triggered Tax Cuts Will Affect Future Services
When enacting triggered tax cuts (or other tax cuts with a significantly delayed effective date), policymakers often lack up-to-date, multi-year forecasts of what they will cost in forgone revenue when they take effect, and they almost never know what state services will likely cost at that time. Both kinds of forecasts could be done in most cases, but multi-year estimates of revenue losses are the exception rather than the rule and long-term estimates of the cost of providing “current services” are almost never done. And without that information, policymakers cannot evaluate the tax cuts’ impact on state services. (See box.)
Even if such a forecast is available, it is unlikely to predict the occurrence or magnitude of an economic downturn before the tax cut takes effect. Nor can an economic forecast predict random events that can also drastically affect state finances, such as natural disasters or cuts in federal financial aid to states.
In nearly all of the ten states that enacted triggered tax cuts since 2002, policymakers did so without information needed to assess their likely impact on the state’s ability to provide services:
- None of the ten states that have enacted triggered income tax cuts in recent years estimated the cost of providing existing services over the full period over which tax cuts might take effect, leaving policymakers in the dark as to whether the tax cuts will force cuts in state services.
- In all ten states the tax cuts could take full effect after the period for which the state forecasts revenue.
- Only three of the nine states with multiple triggered tax cuts in successive years estimated the combined revenue loss in the first fiscal year in which all of them take full effect.
- In seven of the eight states in which the tax cuts are not yet fully in effect there are no time limits on when they can trigger, so it is impossible to estimate their cost in relation to the cost of providing services in the year they take effect.
Missouri Example Highlights Perils of Inadequate Informationa
In May 2014, Missouri enacted a tax cut that phases in over five years, beginning in tax year 2017. The legislation lowers income tax rates, increases the personal exemption for low-income taxpayers, and establishes an income tax deduction of up to 25 percent for business income. The tax cuts can phase in in any year in which the revenue collected in the previous fiscal year was at least $150 million more than the highest amount collected in any of the three years before that. According to the official fiscal note, the changes will reduce revenue by $620 million in 2022, when they were expected to be fully phased in.b
While the fiscal note was available before enactment, policymakers and the public considering the legislation did not have critical information against which to compare this revenue loss or to judge the potential impact of the trigger. There was no baseline for revenues or current services spending that covered the period in which the tax cut would phase in. Missouri prepares a consensus revenue forecast, but it provides information only for the current year and the upcoming (budget) fiscal year. Similarly, the governor’s budget provides spending information only for the previous, current, and upcoming fiscal year. So when the tax cut was enacted, the most forward-looking revenue and spending estimates were for fiscal year 2015 — and therefore useless for evaluating a tax cut that would not even begin taking effect until tax year 2017.
It is worth noting that the $150 million revenue growth needed to trigger the phase-in is just 1.7 percent of forecasted fiscal year 2015 revenue. That is highly unlikely to be sufficient to maintain current services. The fiscal note assumed inflation alone would be 2.3 percent per year during the phase-in period.
After the legislation’s enactment, the independent Missouri Budget Project prepared longer-term revenue and current services estimates. These showed that by 2023, available revenues following full implementation of the tax cuts would fall $1.95 billion short of the amount necessary to fund current services expenditures (the same services provided in fiscal year 2014 plus funding needed under the state’s K-12 school funding formula).c Eliminating that gap would require a spending cut of more than 16 percent below a current services level.
It is always difficult to think about what any situation will be eight years from now. But, as the Missouri example shows, acting without fully understanding the longer-term implications can have a huge cost.
a For more on these issues, see the CBPP report from which this box was taken: Elizabeth McNichol, Iris Lav, and Michael Leachman, “Better State Budget Planning Can Help Build Healthier Economies,” October 15, 2015, https://www.cbpp.org/research/state-budget-and-tax/better-state-budget-planning-can-help-build-healthier-economies.
b In fact, revenue growth in fiscal year 2016 (July 1, 2015 through June 30, 2016) was too slow to trigger a January 2017 tax cut, meaning that full implementation of the five possible tax cuts in the Missouri legislation will be delayed at least a year beyond 2022.
c Tom Kruckemeyer and Amy Blouin, “Tax Cuts Fling Missouri Toward Massive Budget Cliff,” November 12, 2014. The report also points out that fiscal year 2014 expenditures were still well below pre-recession levels.
Tax Cuts Can Trigger When Revenues Are Particularly Needed, as in Recessions
Some state trigger mechanisms are particularly ill conceived and can cause tax cuts to occur during recessions or before state economies have recovered:
- In Oklahoma, tax cuts triggered twice in the middle of major state budget shortfalls or before revenues recovered enough from a recession to permit the restoration of services. This happened primarily because the tax cuts were triggered on the basis of revenue forecasts that were largely obsolete by the time the tax cuts took effect.
- The triggers in four states (the District of Columbia, Kansas, Massachusetts, and Oklahoma) depend only on one-year growth in state revenue. Under such a mechanism, modest revenue growth could trigger tax cuts even if revenues remained deeply depressed following a nationwide or state-specific economic downturn (for example, one occurring in an energy-producing state because of a decline in oil prices). These kinds of triggers can lead to a ratcheting down of state spending over time as successive recessions occur.
- The triggers in six states (Kansas, Maine, Massachusetts, Michigan, Missouri, and North Carolina) are written in such a way that if the revenue targets are met, the tax cuts could cause revenues to fall below the targets in the very first year they take effect.
- In only one of the ten states (West Virginia) does the trigger require revenues to remain above a threshold amount for more than the year in which the tax cuts take effect.
Triggered Tax Cuts Typically Fail to Account for State Fiscal Needs
The tests of tax-cut “affordability” embedded in trigger mechanisms are inadequate in several ways. First, tax cuts can trigger in most states even if recent revenue growth has not been enough to finance current state services:
- Inflation raises the cost of the goods and services that states buy. But in only two of the ten states with triggers (Massachusetts and Michigan) does the trigger explicitly allow tax revenues to grow with general inflation before cutting them is considered affordable. Moreover, both of these states use the Consumer Price Index (CPI) to measure inflation, even though an index of prices covering the specific goods and services that states and localities buy rose 40 percent faster than the CPI on average from 1990 to 2015.
- State spending also has to grow to accommodate such factors as growth in the overall population and in groups that tend to consume more state services (such as students and the elderly). For example, as the number of cars on the road rises, so should the number of state troopers. And to attract good employees, states need to increase their salaries at close to the same rate as private-sector salaries increase. Yet only in Massachusetts and Michigan does the trigger explicitly allow revenues to grow faster than general inflation before cutting them can be considered affordable.
- In at least five of the seven of the states with triggers based on attaining a certain revenue level or revenue gain, the targets are almost completely arbitrary; they were not based on any systematic analysis of the revenue needed to allow for recent inflation and growth in population and caseloads.
Second, the trigger mechanisms fail to consider whether the tax cuts are affordable over the long term:
- In only one of the ten states (North Carolina) does the trigger require revenues to hit a target for multiple years before the tax cuts take effect.
- In Missouri and Oklahoma, tax rate cuts with no time limit on when they can take effect are based on reaching a revenue level measured in dollars or achieving a one-year revenue increase measured in dollars. Since inflation erodes the value of those dollar thresholds, it will get steadily easier over time for the tax cuts to take effect, even though the share of revenue that the tax cuts will strip away remains constant or even increases.
- In Kansas and Michigan, triggers remain in place until personal income taxes are completely phased out, which would force some combination of extremely deep cuts in critical services like K-12 schools, higher education, and health care, substantial depletion of state reserves, and sharp increases in other taxes or fees.
Third, triggers almost never account for every state’s need to maintain adequate financial reserves for fiscal and other emergencies (recessions, natural disasters, federal aid cuts, or other unforeseen circumstances) and the readily apparent long-term fiscal problems in many states:
- Only in West Virginia does the trigger require the state to accumulate a particular level of reserves in a formal “rainy day fund” before the tax cuts can take effect.
- In only one of the ten states is there any evidence that policymakers, in considering triggered tax cuts, gave meaningful consideration to their potential impact on the state’s ability to address long-run needs such as underfunded public pension plans or inadequate infrastructure investment.
More broadly, the strong push for tax cuts in recent years — including the growing number of tax cuts with triggers attached — ignores the long-term trends that are causing state revenues to fall behind the costs of critical programs and investments. On the revenue side, these trends include the shift in household purchases toward services (which are largely untaxed) and online sales (many of which are effectively untaxed), declining alcohol and tobacco taxes as consumption falls, and states’ failure to adapt their corporate income taxes to the growing sophistication of corporate tax-avoidance strategies. On the spending side, these trends include the tendency of health-care costs (which represent a large share of state budgets) to grow faster than other expenses, the growing demand for a college education, and the need to make long-deferred investments in schools, roads, and other critical infrastructure.
These long-term trends imply that — regardless of the level of services they choose to provide — states will eventually need to generate additional revenue, by closing loopholes, scaling back ineffective economic development giveaways, and/or increasing tax rates.
No Meaningful Benefits to Enacting Tax Cuts Far in Advance
There are no practical benefits to enacting tax cuts, triggered or un-triggered, far in advance of when they will take effect. Whenever elected officials decide that taxes can be cut without requiring cutting services or are a priority even if they might negatively affect services, the cuts can be implemented almost immediately. New withholding schedules for individual income taxes can be published by state tax departments and implemented by payroll processors in a few weeks’ time. Point-of-sale systems that calculate sales taxes can be reprogrammed quickly. Even an immediate corporate income tax cut only requires action two to three months down the road when companies calculate their quarterly estimated payments.
Indeed, most “immediate” cuts in individual and corporate income taxes do not actually take effect until the beginning of the next calendar year, which is adequate time to ensure proper compliance.
With regard to the potential economic benefits, knowing future tax rates may help individuals and businesses decide whether to make certain long-term investments, but these benefits are likely quite limited when it comes to state taxes. Most of the triggered tax cuts enacted in recent years involve personal income taxes, and only a small minority of income tax payers are business owners making significant capital investment decisions. Also, state personal and corporate income tax rates are relatively low to begin with — especially in comparison to federal rates — and thus unlikely to tip investment decisions. And because elected officials may cancel or postpone previously enacted tax cuts, investors can never be 100 percent certain of future tax rates. This is especially true of triggered tax cuts, which depend on the state hitting revenue targets.
In short, any theoretical economic benefits of delayed tax cuts are relatively minor, especially when compared to the very real economic damage and disruption if the tax cuts turn out to be unaffordable. The lack of concrete benefits suggests that the fundamental objective of delayed tax cuts (whether or not they have triggers) is to avoid the political backlash that could arise if state services had to be cut immediately while erecting a political obstacle to making different fiscal policy choices down the road.
Triggers help perpetuate the widespread misconception that states can cut taxes without cutting services by devoting “new” revenue they collect in future years to tax reduction. This idea ignores the fact that the cost of maintaining services rises over time due to general inflation, a growing and aging population, and a host of other factors. Triggers offer the appearance of taking recent revenue growth trends into account in determining if and when new tax cuts will take effect and how large they will be when they do, but this creates only an illusion of the cuts’ affordability.
Enacting delayed tax cuts with revenue triggers is unnecessary because legislatures can rapidly put tax cuts into effect any time they wish. It is unwise because the balance between revenues and spending can change drastically before the tax cuts take effect. And it is irresponsible because it enables lawmakers and governors to impose policies on future voters that they are unwilling to impose on themselves and be held accountable for, leaving it to future policymakers to deal with the consequences. For these reasons, state policymakers should reject triggered tax cuts.
|Key Provisions of Recent State Tax Cuts Subject to Revenue Triggers|
|State||Year Enacted||Taxes Subject to Triggered Cuts||Potential Triggered Cuts||Specific Date by Which All Cuts Take Effect?||First Date on Which First Tax Cut Could Take Effect||First Date on Which First Tax Took Effect||All Tax Cuts Subject to Triggers in Effect?||Summary of Trigger Mechanism|
|DC||2014; amended 2015, 2016||
|17 different tax cuts: PIT and CIT rate cuts; increase in SDs and PEs; increase in estate tax exemption||No||01/2016||01/2016||No||If updated forecasted revenue exceeds revenue upon which budget was based, “excess” revenue is dedicated to tax cuts whose first-year cost is below “excess” revenue.|
|KS||2013; amended 2015||PIT
|Phase-out of PIT with rate cuts, then CIT||No||01/2021||No||Rate cuts trigger when year-over-year revenue growth exceeds 2.5%.|
|ME||2012||PIT||Rate cuts to convert multiple-rate PIT with top rate of 7.95% to flat-rate tax at 4.0%||No||01/2015||No||Revenue received in excess of state spending cap and/or budgeted revenue deposited to fund. Rate cut triggers when fund accumulates sufficient revenue to pay first-year cost.|
|MA||2002||PIT||Cut of flat-rate PIT from 5.3% to 5.0% in six 0.05% increments||No||01/2009||01/2012||No||Rate cuts trigger when year-over-year revenue growth is 2.5% above inflation.|
|MI||2015||PIT||Elimination of PIT via rate cuts||No||01/2023||No||Annual revenue cap calculated by inflating FY21 revenue by roughly 1.5 times cumulative inflation after that year. Revenue in excess of cap triggers rate cuts.|
|MO||2014||PIT||Cut in top PIT rate from 6.0% to 5.5% in five 0.1% increments. Exclusion of up to 25% of business income from PIT in five 5% increments||No||01/2017||No||Tax cuts trigger when actual revenue in previous fiscal year is at least $150 million higher than highest amount collected in three prior years|
|Cut in CIT rate from 8.2% to 7.9%; cut in VAT rate from 0.72% to 0.675%||Yes||01/2018||No||Tax cuts trigger if combine General and Education Fund revenues collected in FY2016-FY2017 biennium exceed $4.64 billion.|
|NC||2013||CIT||Cut in CIT rate from 5% to 3% in two 1% increments||Yes||01/2016||01/2016||Yes||First rate cut triggered when actual FY15 revenue hit $20.2 billion. Second cut triggered when FY 2016 revenue hit $20.975B.|
|OK||2014||PIT||Cut in top PIT rate from 5.25% to 5.0%, then to 4.85%||No||01/2016||01/2016||No||First rate cut triggered because FY16 revenue predicted in 12/14 exceeded FY14 revenue predicted in 02/13. Second rate cut did not trigger because growth in FY18 revenue (in dollars) predicted by 12/16 forecast was less than estimated tax year 2018 cost of rate cut; can still trigger in future based on similar test to be done each December|
|WV||2008||CIT||Cut in CIT rate from 8.5% to 7.75%, then to 7.0%, then to 6.5%||No||01/2012||01/2012||Yes||Rate cuts took effect in three consecutive years because rainy day fund exceeded 10% of general revenue fund budget.|
Appendix: Case Studies of States in Which Tax Cuts Have Triggered
District of Columbia
The District of Columbia’s tax trigger was designed to dedicate all “unanticipated” revenue growth above budgeted amounts to 17 different tax cuts until they all go into effect — whenever that may occur. The tax cuts trigger regardless of whether the cost of providing the existing level of services has also increased unexpectedly.
The City Council enacted the trigger law in 2014 after the city’s Chief Financial Officer (CFO) was unable to certify the affordability of the Council’s proposed fiscal year (FY) 2015 budget, as required by federal law governing the District’s finances. A major reason for the lack of affordability was that the budget included virtually all of the tax cuts that the DC Tax Revision Commission had recommended in 2013 but omitted several proposed revenue-raising offsets.
To obtain the CFO’s certification, the Council amended the budget to implement only a few of the tax cuts during FY 2015 (October 2014 through September 2015) and establish revenue triggers and a priority ranking for 17 additional tax cuts in later years. The amended budget stipulated that if the revenue forecast to be conducted in the middle of FY 2015 showed revenues for FY 2015 or FY 2016 above levels assumed in the enacted FY 2015 budget, the first $181 million of these “unanticipated” revenues would go toward filling the expected gap between predicted revenues and predicted spending for FY 2016. Any unanticipated revenues above that $181 million would go to implementing as many of the tax cuts as those revenues could pay for in the first year the tax cuts would take effect. This provision effectively assumed that the extra revenue would materialize in all future years and pay for the tax cuts in perpetuity.
More significantly, the Council mandated that in all future fiscal years, all nominal (that is, non-inflation-adjusted) increases in revenue predicted in a quarterly revenue forecast conducted after the budget’s adoption would be dedicated to implementing the 17 tax cuts, in order of listed priority, until all were in place. Proponents claimed this trigger structure would permit the Mayor and City Council to spend as much of the available revenue shown in each year’s February revenue forecast as they wished, but any subsequent additional revenue would have to go for tax cuts. Any tax cuts that were triggered would take effect on January 1, just three months into the fiscal year. This mechanism created the possibility that a one-time upward revision in a revenue forecast would lead almost immediately to a reduction in revenue even if the forecast were later revised downward.
In both 2015 and 2016, the Council amended the trigger to accelerate implementation of some of the tax cuts by a full year, nullifying the claim that the trigger would allow the Mayor to propose a budget that would spend the available revenue shown in the February forecast. The first seven tax cuts have since taken effect.
Those seven tax cuts plus those in the original FY 2015 budget reduce revenues by approximately $100 million annually, according to the DC Fiscal Policy Institute. This revenue loss has contributed to very slow revenue growth: 2.6 percent in FY 2016, an estimated 1.1 percent in FY 2017, and only around 3 percent predicted for the subsequent three years. Given that prices are predicted to rise more than 2 percent annually over this period, the school-age population is rising, and Medicaid caseloads continue to increase, revenue will likely grow more slowly than the cost of providing current services. Current services funding levels rose an average of 6.4 percent annually between FY 2012 and FY 2016.
In short, the already implemented tax cuts will likely compel some cuts in current services. And the remaining ten tax cuts, if triggered, would reduce yearly revenues by an additional $128 million, likely leading to additional service reductions. Regardless of whether the tax cuts lead to cuts in existing services, some portion of the $100 million in annual revenue they have already cost could have been devoted to addressing critical problems, such as the District’s growing homeless population and shortfall in affordable housing.
In 1999, responding to strong revenue growth from the dot-com bubble of the late 1990s, Massachusetts lawmakers approved a three-year, phased-in cut in the state’s flat-rate income tax from 5.95 percent to 5.75 percent. Dissatisfied with that reduction, then-Governor Paul Cellucci sponsored a ballot initiative to cut the rate over three years to 5.0 percent. The voters approved Question 4 in November 2000, and the first rate cut, to 5.6 percent, took effect in January 2001.
A national recession began in March 2001, and Massachusetts’ revenues fell. Nonetheless, the second rate cut (to 5.3 percent) took effect as scheduled in January 2002. But, confronting a combined $1 billion revenue shortfall in FY 2002 and FY 2003, lawmakers overrode a gubernatorial veto and suspended the final rate cut (to 5.0 percent) scheduled for January 2003. They substituted a provision that would make the final 0.3 percentage-point cut in six annual increments, each subject to two triggers.
The first trigger provided that a tax cut would take effect in any year in which tax collections rose at least 2.5 percent over the prior year, after factoring out inflation and any growth attributable to other tax-law changes. The second trigger provided that a tax cut couldn’t take effect on January 1 if revenue growth had slowed sharply since the end of the prior fiscal year on June 30. The bill also provided that the triggered rate cuts could not take effect until a reduction in personal exemptions included in the bill were reversed.
Triggered rate cuts occurred in January of 2012, 2013, 2015, and 2016. The rate now stands at 5.1 percent. The most recent revenue forecast projects that the reduction to 5.05 percent will occur in January 2018. In short, even 16 years after their enactment, the triggered cuts will not be fully in effect; the 2002 legislature will continue to impose its fiscal priorities on a younger generation of lawmakers for at least another year beyond that.
The tax cuts that have triggered thus far will cost an estimated $585 million this fiscal year (2017); the two rate cuts yet to trigger will cost at least $300 million more each year. Although the revenue loss from these cuts is smaller than that resulting from the cuts approved in Measure 4, they have compounded the damage. The state has deeply cut services since the 2001 and 2002 tax cuts took effect.
As part of a major tax restructuring plan enacted in 2013 (House Bill 998), North Carolina reduced the corporate income tax rate for tax year 2014 to 6.0 percent from 6.9 percent, and to 5.0 percent for 2015. The plan also included additional, triggered cuts — to 4.0 percent for 2016 and 3.0 percent for 2017— if revenues reached $20.2 billion in FY 2015 and $20.975 billion in FY 2016. Both targets were hit and both tax cuts have now taken effect.
The triggers, added in a House-Senate conference committee to mollify House members concerned about the corporate tax cuts’ cost, were largely symbolic. Around the time HB 998 was enacted, the state was predicting FY 2015 revenue of $20.95 billion, or $750 million more than needed to trigger the first tax cut and only $25 million less than needed to trigger the second one. A post-enactment analysis indicated that average inflation and population growth alone would have put FY 2016 revenue $668 million over the level needed to trigger the second cut.
Lawmakers approved the tax-cut bill despite grossly insufficient information with which to assess the potential impact of the triggered tax cuts on the state’s ability to finance future services. At the time of enactment, no fiscal note was available showing the year-by-year impact of the corporate tax cuts in the final bill. A fiscal impact analysis of the bill published after enactment only estimated the revenue loss for the two corporate rate cuts not subject to triggers, even though the analysis extended through FY 2018, when both cuts subject to triggers were scheduled to take effect. North Carolina does no revenue forecasting beyond the upcoming biennium and no current-service spending estimates at all.
An analysis after the 2013 legislative session pegged the FY 2018 revenue loss from HB 998 at $728 million (again, not including the potential impact of the two triggered cuts). In mid- 2015, the state estimated that the triggered cuts would cost an additional $503 million in FY 2018. In other words, the two triggered tax cuts raised the ultimate cost of HB 998 by more than two-thirds.
In the four years since FY 2013 (the last year before HB 998 started reducing revenues), North Carolina’s General Fund budget has increased an average of only 2.6 percent per year. Between FY 2008 and FY 2014, North Carolina cut K-12 spending per student by 14.1 percent, after adjusting for inflation. Between FY 2008 and FY 2016, it cut higher education spending per student by 20.1 percent. Moreover, the full revenue losses from the two triggered tax cuts have not yet been felt; many corporations won’t file their tax year 2016 corporate tax returns until the fall of 2017, and their 2017 tax returns a year beyond that. Even greater austerity may well be in store for North Carolina’s students, elderly, and other citizens who rely on the critical services the state provides. For the foreseeable future, the revenue reductions caused by HB 998 will make it all but impossible for the state to achieve a reasonable level of investment in education and other critical services cut during the Great Recession.
Oklahoma has enacted two ill-conceived tax triggers since 2006. Both have resulted in tax cuts occurring at times when the state already faced sharp revenue losses due to economic downturns and deep, phased-in tax cuts not subject to triggers. Although the triggered tax cuts were smaller than the mandated tax cuts, they have compounded the financial damage that the latter created.
Enjoying strong revenue growth from a sharp run-up in oil prices that started in 2002, Oklahoma began a decade-long tax-cutting spree in 2004. In 2004 and 2005, lawmakers cut the top income tax rate from 6.65 percent to 6.25 percent. In 2006 they voted to phase out the state estate tax, increase the income tax standard deduction, and cut the top income tax rate to 5.25 percent in four steps over four years. The fully-phased-in cost of these tax cuts was more than $550 million annually. Oklahoma does no revenue or expenditure forecasting beyond the next budget year, so lawmakers had no information available to assess the potential impact on the state’s ability to finance services.
In a concession to lawmakers concerned about the long-range impact, a revenue trigger for the last of the four rate cuts — which could take effect as early as January 2010, midway through FY 2010 — was added in a House-Senate conference committee. But the trigger mechanism was seriously flawed. First, it made the rate cut contingent on reaching an arbitrary rate of revenue growth: lawmakers had no information to assess how much revenue growth in FY 2010 would be needed to accommodate the normal growth in the cost of services, so they arbitrarily assumed 4 percent. Second, the trigger’s measure of revenue growth in FY 2010 was not based on the most up-to-date information available before the rate cut would take effect. It relied on a forecast of FY 2010 revenue made in February 2009 rather than the regular December 2009 forecast.
The final rate cut did not, in fact, trigger in January 2010. Oklahoma’s revenues — like those of virtually all states — plunged due to the Great Recession, though Oklahoma’s revenue decline was much more severe than the average state because the rate cuts for 2008 and 2009 enacted in 2006 took effect despite the economic crisis. The revenue crisis did, however, delay the final rate cut until 2012. At that point the state still faced a $500 million revenue shortfall for FY 2013 and the rate cut dug the hole at least $75 million deeper.
In 2014, lawmakers approved a bill cutting the top tax rate further, to 5.0 percent in 2016 and to 4.85 percent in 2018. Both cuts had triggers attached. The first tax cut would take effect if the December 2014 forecast predicted that FY 2016 revenues would be at least as high as a forecast of FY 2014 revenues made in in February 2013. That was a very low bar. First, given inflation and a growing school-age and elderly population, it clearly would cost somewhat more in FY 2016 to provide the level of state services provided in FY 2014. Worse, the tax cut — whose estimated FY 2016 cost was $57 million — would take effect even if this brought FY 2016 revenue below the FY 2014 amount. Finally, if the conditions for the January 2016 tax cut were not satisfied, they would remain in effect for all future years until the rate cut triggered.
Like the trigger for the first tax cut, the trigger for the second was based on a comparison of revenue forecasts for two fiscal years two years apart and ignored the need for some nominal revenue growth to accommodate the growing cost of services, though it did require sufficient revenue growth between the two years to avert a decline in nominal revenue after the tax cut took effect. The Oklahoma Tax Commission estimated that both rate cuts would cost at least $267 million in the first full year in which both took effect.
Because the triggers were based on revenue forecasts made months or years before the tax cut would take effect, they created the possibility that the tax cut could take effect even if actual revenues fell sharply after those forecasts were made. This is precisely what happened with the triggered tax cuts in the 2014 legislation. The 2016 tax cut took effect as scheduled, even though the last three forecasts of FY 2016 revenue made before the January 2016 implementation date were too low to have triggered it due to the impact of falling oil prices. FY 2016 revenues ultimately fell $800 million (13.3 percent) short of the amount predicted at the time the tax cut triggered. The triggered tax cut was responsible for $49 million of that shortfall.
Because of the plunge in state revenues, the second tax cut did not trigger in December 2016. But it is still on the books to trigger in any future year in which there is even minimal growth in nominal revenue.
Since 2004, Oklahoma’s top income tax rate has fallen by 1.65 percentage points (25 percent); the two triggered cuts comprise 0.5 percentage points of that. The annual revenue loss from these rate reductions is over $1 billion. They have devastated state services, leading to “acute teacher shortages, college tuition and fee hikes, critically understaffed correctional facilities, longer waiting lists for services, and lower reimbursement rates for medical and social service providers.” Between 2008 and 2014, Oklahoma cut per-pupil spending for K-12 education by more than 14 percent, after adjusting for inflation. It cut per-pupil funding for higher education by almost 22 percent from 2008 to 2016. As of December 2016, the gap between available revenues and current spending levels for FY 2018 stands at almost $900 million.
In February 2008, West Virginia Governor Joe Manchin proposed additional cuts in the state’s (flat) corporate income tax rate, which had been cut the previous year from 9.0 percent to 8.75 percent. Manchin proposed four cuts beginning on January 1, 2009, ending at a 6.5 percent rate for tax years beginning January 1, 2014. The enacted bill also included a trigger conditioning the rate cuts on the maintenance of a rainy day fund equal to at least 10 percent of annual appropriations in the General Revenue Fund.
The official analysis of the bill’s fiscal impact noted that when the rate cuts had their full effect in FY 2016, they would cost the state “roughly $70 million” annually. The governor’s FY 2009 budget, released a month before he proposed the new tax cuts, showed shortfalls of $120 million in FY 2012 and $140 million in FY 2013 between projected revenues and the cost of maintaining current services. It would be reasonable to assume that the shortfall would grow in subsequent years, but even assuming it would have stayed at $140 million through FY 2016, the tax cut would have been responsible for fully half of the predicted gap in FY 2016.
In short, the trigger based on maintaining a rainy day fund target provided political cover for a tax cut that was not “affordable” in any meaningful sense when it was enacted. West Virginia cut higher education spending per student by almost 24 percent between 2008 (when the tax cut was enacted) and 2016 (when it was fully in effect). That the income tax cut enacted in 2008 — and several other large business tax cuts in 2006 and 2007 — did not lead to even deeper cuts in critical services was due to the severance tax revenues generated by the growth in West Virginia’s natural gas output.
 Those states (other than the District of Columbia) are Kansas, Maine, Massachusetts, Michigan, Missouri, New Hampshire, North Carolina, Oklahoma, and West Virginia.
 All changes in tax laws, both those that increase tax payments and those that reduce them, allow some amount of time for taxpayers and tax administrators to get ready. For example, most changes in income taxes take effect on January 1 of the upcoming year.
 Reliable forecasts may not be doable at all if the effective date of a scheduled tax cut is more than four or five years in the future or if a triggered tax cut contains no time limit on when the cut can trigger.
 As former Oklahoma Senate President Cal Hobson has candidly acknowledged: “It’s irresistible. You get a tax reduction bill out there, and we don’t think past the end of our noses.” Warren Vieth, “Study Finds a Decade of Income Tax Cuts Deprived State of $1B This Year,” Oklahoma Watch, January 12, 2016.
 The Massachusetts legislation substituted triggered income tax cuts for cuts that had originally been scheduled to phase in automatically.
 For a CBPP analysis of recent tax cuts that were or are being phased in over time, see Eric Figueroa, Michael Leachman, and Michael Mazerov, “Phasing in State Tax Cuts Doesn’t Make Them Fiscally Responsible,” February 2, 2017.
 A 2015 Michigan law takes lack of political accountability to an extreme. It delays implementation of personal income tax cuts subject to triggers until 2023 at the earliest. Michigan House members cannot serve for more than six years and senators are limited to eight years, so the vast majority of lawmakers who voted for these tax cuts did so knowing that they could not be held accountable for any adverse consequences when the cuts took effect.
 Quoted in Elaine S. Povich, “‘Triggers’ Cut State Taxes; But Are They Good Policy?” Governing, November 16, 2015.
 Legislative supermajorities are required to increase taxes in 16 states. (See Figueroa.) Two of those (Missouri and Oklahoma) are among the ten states that have enacted triggered tax cuts. In both states it takes only a legislative majority to approve tax cuts but a supermajority to approve tax increases. Whether reducing, postponing, or repealing a triggered tax cut would require such a supermajority vote is unclear in these states, but any such step would very likely lead to litigation that would at least temporarily prevent such a modification. In other words, in these two states, triggers may create a legal obstacle to altering future state fiscal priorities in addition to a political one.
 Povich, Note 8.
 Of course, unforeseen events can occur at any time, including right after a tax cut that took effect immediately. But given the lack of meaningful benefits from enacting delayed tax cuts (as discussed below), states should not run the risk of these events occurring in the interval between when a tax cut is enacted and when it takes effect.
 See Elizabeth McNichol, Vincent Palacios, and Nicholas Johnson, “Budgeting for the Future: Fiscal Planning Tools Can Show the Way,” CBPP, February 4, 2014, Table 5, Column 2, https://www.cbpp.org/research/state-budget-and-tax/budgeting-for-the-future-fiscal-planning-tools-can-show-the-way?fa=view&id=4085.
 The three states are the District of Columbia, Oklahoma, and West Virginia. New Hampshire is the only one of the ten states in which all triggered tax cuts would occur in a single year.
 Tax cuts subject to triggers have not yet fully taken effect in the District of Columbia, Kansas, Maine, Massachusetts, Michigan, Missouri, New Hampshire, and Oklahoma. New Hampshire is the only one of these states in which the cut can only occur in a specific year.
 See the Appendix.
 A tax cut did, in fact, trigger in Massachusetts in January 2012 because revenue growth between fiscal years 2010 and 2011 exceeded the requisite 2.5 percent rate (after adjusting for inflation). Yet fiscal year 2011 revenue was more than $300 million below the pre-Great Recession peak, in fiscal year 2008. See Massachusetts Department of Revenue, “Briefing Book, FY 2018 Consensus Revenue Estimate Hearing,” December 5, 2016, p. 26.
 The trigger in Missouri compares revenue collected in the most recent fiscal year with the highest annual revenue collected in the three fiscal years prior to that one. This mechanism prevents a revenue rebound after a brief recession from triggering a tax cut, but an extended, deep recession could still trigger a tax cut before revenue reached pre-recession levels.
 In Missouri, for example, one of five possible cuts in the top personal income tax rate takes effect in January following any fiscal year (concluding June 30) in which revenue increases by at least $150 million from the prior fiscal year. The first-year cost of the tax cut was estimated at $69 million (Missouri Committee on Legislative Research, “Fiscal Note for Senate Committee Substitute for Senate Bills Nos. 509 and 496,” April 29, 2014). If the target were just barely met and revenue growth remained low in the fiscal year in which the tax cut took effect, this trigger could cause revenue in the latter year to drop below the revenue level that triggered the cut.
 The West Virginia trigger required the state’s rainy day fund to equal at least 10 percent of the General Fund budget for the three years in which the tax cuts took effect, but not thereafter.
 Even if a trigger took account of recent inflation, the tax cut would not be meaningfully “affordable” if it did not allow future revenues to grow with inflation as well.
 Even if a trigger took account of recent population and caseload growth, the tax cut would not be meaningfully “affordable” if it did not allow future revenues to grow to in line with good faith forecasts of these expenditure drivers as well.
 In Kansas, Massachusetts, Missouri, New Hampshire, and Oklahoma, the revenue targets were essentially arbitrary and not based on any analysis of recent annual growth in the cost of providing services. It is unclear if any analysis was done in Michigan to arrive at the trigger provision that allows revenue to grow about 50 percent faster than inflation before tax cuts trigger. Only in North Carolina was the revenue-growth target based on a rough analysis of recent inflation and population growth. (But see Notes 19 and 20 above.)
 In North Carolina, only the second of two triggered cuts in the corporate income tax rate required revenue targets to be hit two years in a row.
 In Kansas, the trigger provides for the complete phase-out of the corporate income tax as well.
 However, having a certain level of reserves on hand was the only requirement that had to be met for the West Virginia tax cuts to take effect. That is, the tax cuts were enacted with no consideration of their potential impact on service provision. Moreover, the reserves threshold only had to be met for the three years in which corporate income tax rate cuts were going into effect, not for any years thereafter.
 The Tax Foundation has in the past sounded the alarm about the underfunding of state pension plans. (See: Joseph Henchman, “Monday Map: Funded Ratio of State Public Pension Plans,” September 16, 2013.) Its recent report touting triggered tax cuts observes that states could dedicate part of revenue growth to underfunded pensions or rainy day funds. However, only one of the states that have enacted triggered tax cuts (Kansas) has actually done so. Jared Walczak, “Designing Tax Triggers: Lessons from the States,” Tax Foundation, September 2016, pp. 15, 18. In fact, Maine reduced an existing mandatory allocation of surplus revenue to closing a funding gap in the state’s pension fund and dedicated the money instead to financing the income tax cuts. (See Section 2 of LD 849, enacted May 25, 2012.)
 See Iris J. Lav, Elizabeth McNichol, and Robert Zahradnik, “Faulty Foundations: State Structural Budget Problems and How to Fix Them,” Center on Budget and Policy Priorities, May 2005, https://www.cbpp.org/research/faulty-foundations-state-structural-budget-problems-and-how-to-fix-them.
 See Michael Mazerov, “Cutting State Personal Income Taxes Won’t Help Small Businesses Create Jobs and May Harm State Economies,” CBPP, February 19, 2013, https://www.cbpp.org/research/cutting-state-personal-income-taxes-wont-help-small-businesses-create-jobs-and-may-harm. The vast majority of small business owners whose profits are subject to taxation under personal income taxes are self-employed professionals or tradesmen whose capital equipment needs are likely relatively modest.
 Curiously, policy analyst Jared Walczak at the Tax Foundation, an organization that promotes tax triggers, argued recently that they “eliminate some of the uncertainty surrounding changes to the tax code.” (Quoted in Bloomberg BNA SALT Talk Blog, September 12, 2016.) In reality, triggers create more uncertainty by making future tax rates contingent on future revenues.
 Memo from City Council Chair Phil Mendelson to the Council, June 23, 2014.
 “What Revenue Changes Are in the Final FY 2017 Budget?,” June 30, 2016.
 District of Columbia revised revenue estimate; September 30, 2016.
 Ilana Boivie, “Tax Cut ‘Triggers’ Are Limiting Our Ability to Fund Necessary Services and Should Be Modified,” DC Fiscal Policy Institute, May 11, 2016; District of Columbia revised revenue estimate; September 30, 2016.
 See letter from CFO DeWitt to Mayor Gray, November 21, 2014.
 H. 5250 approved July 2002.
 For an example of the calculation of this somewhat complex trigger, see the letter from Tax Commissioner Mark E. Nunnelly to Kristen Lepore, Secretary of the Executive Office for Administration and Finance, December 15, 2015.
 See Note 16.
 “Lepore, Spilka, Dempsey Announce Consensus Revenue Forecast of $27.072B for Fiscal year 2018,” press release, January 12, 2017, http://www.mass.gov/anf/press-releases/fy2017/consensus-revenue-forecast-for-fy18-announced-.html.
 Commonwealth of Massachusetts, investor “Information Statement,” November 1, 2016, p. A14, http://www.massbondholder.com/sites/default/files/downloads/Commonwealth%20Information%20Statement%20dated%20November%201%2C%202016%20%28with%20exhibits%29.pdf.
 Kurt Wise and Marie-Frances Rivera, “Income Tax Cuts Have Reduced Funding for People and Communities,” Massachusetts Budget and Policy Center, January 22, 2015.
 The trigger language was written in such a way that both targets had to be hit for both rate cuts to go into effect. If only the second-year target had been hit, the tax rate would have gone down only one percentage point rather than two.
 The Senate version of the bill would have phased out the corporate income tax and eliminated it effective January 1, 2018. The final House version of the bill would have cut the rate only from 6.9 percent to 5.4 percent.
 North Carolina Fiscal Research Division, North Carolina General Assembly, “2013 Annotated Conference Committee Report on the Continuation, Expansion, and Capital Budgets, Revised November 6, 2013, p. 11, http://www.ncleg.net/FiscalResearch/budget_summaries/budget_summaries_PDFs/2013_Annotated_Committee_Report_Final_revised_13-11-06.pdf.
 Three fiscal notes for the bill are available on the North Carolina General Assembly’s website. The first was based on a version of the bill with corporate income tax rate cuts that were less drastic than those ultimately enacted, and the latter two were based on versions of the bill with rate cuts more drastic than those ultimately enacted. But none estimated the exact pattern of rate cuts actually enacted, including the triggered cuts.
 See the source cited in Note 43.
 “Revised General Fund Revenue Consensus Forecast,” North Carolina Fiscal Research Division memorandum, May 6, 2015, http://www.ncleg.net/fiscalresearch/generalfund_outlook_pdfs/15-16_Outlooks/Consensus%20Revenue%20Report%20May%202015.pdf.
 Michael Leachman, Kathleen Masterson, and Marlana Wallace, “After Nearly a Decade, School Investments Still Way Down in Some States,” CBPP, October 20, 2016, https://www.cbpp.org/research/state-budget-and-tax/after-nearly-a-decade-school-investments-still-way-down-in-some-states.
 Michael Mitchell, Michael Leachman, and Kathleen Masterson, “Funding Down, Tuition Up: State Cuts to Higher Education Threaten Quality and Affordability at Public Colleges,” CBPP, August 15, 2016, https://www.cbpp.org/research/state-budget-and-tax/funding-down-tuition-up.
 Oil prices had more than doubled between 2003 and 2006. See historical data for the Cushing, Oklahoma West Texas Intermediate crude oil spot market at http://www.eia.gov/dnav/pet/pet_pri_spt_s1_d.htm.
 Oklahoma Tax Commission, Fiscal Impact Statement for HB 1772, June 26, 2006.
 Elizabeth McNichol, Vincent Palacios, and Nicholas Johnson, “Budgeting for the Future: Fiscal Planning Tools Can Show, the Way,” CBPP, February 4, 2014, p. 39, https://www.cbpp.org/research/state-budget-and-tax/budgeting-for-the-future-fiscal-planning-tools-can-show-the-way?fa=view&id=4085.
 The Senate President had said, “It would be irresponsible not to look beyond the next fiscal year when considering a [tax cut] program of this magnitude.” Kenneth L. Hunt, “Oklahoma House Approves Cut in Top Individual Income Tax Rate,” State Tax Notes, April 18, 2006.
 It might be objected that December 2009 would be too late to make a final determination about a January 2010 rate cut, but recall that 2010 taxes aren’t filed until February 2011 at the earliest. If the tax cut had triggered on January 1, 2010, the only consequence would have been that wage earners would have had a very small amount of extra tax unnecessarily withheld from their paychecks for a few weeks until the Tax Commission could publish updated withholding tables taking into account the lower tax rate.
 When the rate cut did finally trigger on January 1, 2012, its FY 2012 cost was estimated at $37.6 million. (See State Board of Equalization, “Proposed FY-2012 Revenue Certification,” February 22, 2011, pp. 2 and 13; https://www.ok.gov/OSF/documents/boe02222011.pdf.) Note that because the tax cut went into effect midway through the fiscal year, its full-year cost was actually at least twice as large, that is, at least $75 million. See also: Ted Streuli, “Oklahoma Senate Approves Income Tax Cut with Revenue Trigger,” State Tax Notes, March 3, 2014.
 It is unlikely the bill would have been approved without them; it passed with only three votes to spare in the House.
 Oklahoma Tax Commission analysis of SB 1246, February 26, 2014.
 See State Board of Equalization, “Proposed FY 2016 Revenue Certification,” December 18, 2014, pp. 1, 10, https://www.ok.gov/OSF/documents/boe12182014.pdf; and “Proposed FY 2018 Revenue Certification,” December 21 2016, p. 3, https://www.ok.gov/OSF/documents/boe12212016.pdf. (Again, that was a half-year cost.)
 Legislation has been introduced for the 2017 session to require year-over-year revenue growth to equal $500 million plus the cost of the rate cut before it could trigger. See David Blatt, “Trigger Warning: Legislature Sets Itself Up for Another Ill-Timed Income Tax Cut,” Oklahoma Policy Institute, December 8, 2016.
 David Blatt, “The Cost of Tax Cuts in Oklahoma,” Oklahoma Policy Institute, January 2016.
 Blatt, “Cost of Tax Cuts.”
 Leachman, Masterson, and Wallace.
 Mitchell, Leachman, and Masterson.
 Rick Green, “Oklahoma Budget Hole Nearly $900 Million,” The Oklahoman, December 20, 2016.
 Mitchell, Leachman, and Masterson.
 See Sean O’Leary, “How the Marcellus Boom Masked West Virginia’s Tax Cut Problems,” West Virginia Center on Budget and Policy Blog, November 7, 2014.