State policymakers are increasingly using “triggers” to pass big tax cuts while ducking tough decisions on how to pay for them. They are bad tax policy. Naturally, they are incredibly popular.

Here’s how a tax trigger works: A state cuts taxes over a period of years. There may be an initial tax cut that takes effect right away but future reductions are tied to some other benchmark, typically (but not always) achieving an overall revenue target. Advocates argue triggers are a fiscally responsible way to cut taxes. After all, they say, the state only reduces the tax burden if it has enough money to pay for it.

But because the same people who enact the tax change also create the trigger, they are all-but-guaranteeing the future tax cuts. All policymakers really do is cut taxes without busting the budget in the current year. If their budget predictions are wrong, no worries. They’ll just leave it up to future legislatures to figure out how to absorb the cost and explain why the tax cut isn’t happening.

The last few years have highlighted the popularity and problems of triggers. In Oklahoma, despite a revenue shortfall and budget crunch, the Board of Equalization had to certify a tax rate reduction passed several years ago. Why? Because the trigger was based on estimated and not actual revenue.  Kansas, which this year had to freeze income tax cuts and increase the sales tax rate due to lost revenue from prior year tax cuts, still has additional personal and corporate income tax cuts on the books waiting to be triggered in 2019.

While triggers usually drive tax cuts, they sometimes set off tax hikes. In Virginia, policymakers ducked a tax increase to pay for roads by substituting hypothetical revenue from the pending Marketplace Fairness Act. When Congress (predictably) didn’t pass the legislation, the lack of sales tax revenue triggered a gas tax increase. The Virginia pols just blamed Washington.

New Hampshire is the latest state to fall victim to the allure of tax triggers. The Republican-led legislature and the Democratic governor agreed to a budget that cut the Business Profits Tax and the Business Enterprise Tax rates in 2017 and again in 2019—as long as revenues for the 2016-2017 biennium exceed $4.64 million. This amount is already below the level projected in the final budget. Here’s the rub: The revenue projections are conservative, so any bump in revenues triggers the tax cut. It happens even if the revenue increase is caused by a one-time event (for example, if revenues rise because the first presidential primary occurs in your state).

There are several problems with tax cut triggers. Unlike spending programs they typically kick-in outside the state’s one- or two-year budget window and can be passed without bumping against balanced-budget requirements. And any attempt to repeal them is easily labeled a “tax hike,” effectively locking them in regardless of the budget situation.

If tax cuts are triggered by estimated rather than actual revenue, it can lead to interesting results. That’s what happened in Oklahoma, where the trigger was the difference between two projections of revenue and not actual collections. As a result, the personal income tax rate was automatically reduced because projected revenue beat expectations from two years before even though actual revenues were below the budgeted amount.

Next year’s trigger (yes, there’s still another rate cut waiting in the wings) is little better: It is based on a forecast that predicts revenue will offset the estimated cost of reducing the tax rate to 4.85 percent.

California, however, offers a cautionary tale for trigger-happy legislators. In 1998, it used triggers (as well as projected increases in capital gains tax revenue) to cut its vehicle license fee by two-thirds. Its trigger went both ways: If revenues fell below projections, the vehicle fees went back up.

That’s just what happened. And car owners, who barely noticed the initial cut in fees, were outraged when they seemingly tripled without anyone taking responsibility. Cue the recall of Gray Davis, and the election of Arnold Schwartzenegger.

But unlike California and Virginia, all of the other state triggers only go one way. Strong revenues can trigger the cuts even if they are only one time. When revenues fall, the rates don’t go back up in these states.