In his recent budget proposal, Gov. Josh Stein recommended halting the scheduled reductions to personal and corporate income tax rates. Stein’s proposal stated that the recommendation was predicated on the February consensus revenue forecast, which projected the state would experience a budget shortfall in fiscal year (FY) 2026-2027. However, since being established in 2011, the consensus revenue forecast has a history of underestimating the state’s ability to collect revenue.

Consensus revenue forecast

The consensus revenue forecast is released by the Fiscal Research Division and the Office of the State Budget and Management every other February before the start of a new biennium. The forecast estimates the General Fund revenue available for the budget, with projections for the remaining months of the current FY and the two in the upcoming biennium. For example, the following are the estimates made in the February 2025 consensus revenue forecast:

  • FY 2025 = $34.7 billion
  • FY 2026 = $34.9 billion
  • FY 2027 = $34.1 billion

Track record

Forecasts made more than two years out, like the one Stein is basing his tax increase on, can be particularly misleading because they are based on numerous uncertain variables.

The table below depicts the consensus revenue forecast estimates made more than two years in advance and the actual revenue that the state collected in those years. Six out of seven forecasts were underestimates. On average, the projections underestimated revenue by nearly $1.7 billion per year or an average of 6.2 percent.

Suppose we assume that the February 2025 consensus revenue forecast underestimates revenue for FY 2027 by even half the average error rate (or 3.1 percent). In that case, revenue will exceed $35.1 billion, and there will not be a budget shortfall.

Moving forward, legislators should regard the consensus revenue forecast estimates released more than two years in advance with a healthy dose of skepticism.

The cause of underestimation

The consensus revenue forecast utilizes static instead of dynamic scoring. Dynamic scoring models consider how alterations to tax policies change key economic variables, such as employment, wages, and investment. Meanwhile, static scoring models assume that changes to tax policies do not affect these variables. This means that static scoring fails to account for the economic growth tax cuts generate, which causes revenue to be underestimated.

However, it is worth noting that while dynamic scoring models offer improved accuracy over static models by accounting for economic feedback, both models become increasingly unreliable when projecting too far into the future.

Closing thoughts

Moving forward, legislators should regard the consensus revenue forecast estimates released more than two years in advance with a healthy dose of skepticism. These long-term estimates tend to have more political utility than economic predictive power and should not be used for the justification of policymaking. Nevertheless, budget writers must remain vigilant and limit spending growth, as the possibility of an upcoming recession could strain the state budget.

Reductions to the personal and corporate income tax rates that began in 2014 have fostered budget surpluses, increased revenues, bolstered reserve accounts, facilitated population growth, and decreased the poverty rate. As a result, it would be prudent for policymakers to ignore Stein’s recommendation and maintain the scheduled reductions in the personal and corporate income tax rates.