How the Rainy Day formula requires us to make mistakes

Last Monday, State Finance Director Preston Doerflinger announced that the state would be depositing $219 million in the Constitutional Reserve Fund (commonly referred to as the “Rainy Day Fund”) this year. Seen alongside news that lawmakers had to overcome a $500 million budget shortfall, with resulting severe cuts to child care subsidies and teacher training, among others, it may seem strange that we are setting aside so much money that could otherwise be used to protect these important programs.

The reason is an artifact of how the Rainy Day Fund is implemented. The trigger for making deposits into the fund depends not on budget needs, but on how good we are at forecasting revenues. Each year, the State Board of Equalization estimates how much tax revenue the state will receive in the coming fiscal year. If General Revenue (GR) collections come in above projection, the Rainy Day Fund gets the surplus, until the fund is at 15 percent of the previous year’s GR certification.

Spending out of the Rainy Day Fund is also partially tied to the accuracy of our estimates, since up to 3/8ths of it can be appropriated when revenue comes in under the projected amount.  Another 3/8ths can be used when revenue is projected to be below the previous year. The remaining 1/4th can be used when an emergency is declared by the Governor and 2/3rds of the House and Senate or, without the Governor, 3/4ths of the House and Senate.

The most obvious problem with this system is that it depends on the Board of Equalization to make inaccurate revenue estimates in a way that underestimates swings in both directions. Luckily, that has usually been the case, as the graph below shows.

Because of a natural tendency to underestimate growth for years in which revenues are increasing and underestimate drops in years when it is decreasing, the rainy day fund has generally worked as it is supposed to. When times are good, revenues come in above projections, so we deposit money in the fund. When times are bad, our projections are too high, so we are able to withdraw funds.

However, it doesn’t always work that way, particularly in years when we are coming out of a recession. As can be seen in FY ’03-FY ’04 and in FY ’10-FY ’11, revenues actually increased and came in above projections when they were expected to decrease. In this situation, the Rainy Day Fund will absorb a big portion of new revenues, despite the fact that they represent recovering normality, not excess prosperity. That creates a situation like we have this year, when there are “surplus” funds even though most agencies are still undergoing budget cuts. Perversely, this could prolong the pain of the recession, which is the opposite of what the Rainy Day Fund is intended to do.

Another drawback is that if our revenue projections ever become more accurate, we will stop making any deposits into the Rainy Day Fund, and it will be harder to make withdrawals.

A more rational method would be to make deposits based on revenue increases above a multi-year average from previous years, so that it both reflects real surplus revenues and is not distorted by one very bad year. For example, Virginia makes deposits when tax revenues are at least 8 percent above the previous year and 1.5 percent above the six-year average.

Deposits and withdrawals could also be tied to measures of economic growth besides tax revenues. For example, Indiana allows deposits to its Rainy Day Fund only when adjusted personal income shows annual growth of more than 2 percent and withdrawals only when it is less than negative 2 percent. An advantage of this type of metric is that it won’t be distorted by changes in tax policy.

We can debate which specific triggers to use, but almost any trigger that reflects real changes in state revenues or the economy would be an improvement over a system that only works when the State Board of Equalization makes the right mistake.