NCPA - National Center for Policy Analysis

Are Government Revenue Projections Biased?

March 5, 2014

All budget revenue projections require forecasts of economic activity -- will the economy grow or shrink in future years? These agency forecasts then determine whether politicians can, or should, make spending cuts or tax increases. But how open to influence and bias are these forecasts? Politicians may want to avoid program cuts, for example, and they may reward agencies for issuing rosy forecasts that would seemingly allow for such spending to continue, says Robert Krol, a professor of economics at California State University, Northridge.

Krol analyzed how accurate the Congressional Budget Office (CBO), Office of Management and Budget (OMB), and Blue Chip Consensus (a combination of private sector projections) forecasts of gross domestic product (GDP) growth actually were, as public choice models would indicate that government agencies are likely prone to bias.

  • For the OMB, Krol found a significant bias, indicating that political pressure from the executive branch largely influenced the agency's forecasts. OMB forecasters overcast GDP growth at two- and five-year horizons by 5 percent and 14 percent, respectively.
  • Moreover, Krol's OMB calculations indicate that a low GDP forecast was more costly to an administration than a rosy GDP forecast. With an upward bias, a cheerier forecast can help the executive branch avoid spending cuts or tax increases that could be costly politically.
  • Krol did not find a bias for the CBO forecasts, finding that they were conservative and consistent with those in the private sector. The CBO reports to Congress rather than a single party, and it is more institutionally independent than the OMB. Controlled directly by the president, the OMB is more likely to face significant amounts of pressure.
  • Rudolph Penner, former director of both the OMB and the CBO, has said that the CBO does not want to deviate from the private sector consensus forecast, because large differences would make the agency appear partisan.
  • The Blue Chip Consensus has a downward bias, similar to the CBO, in its forecasts of real GDP two to five years out.

While agency actors undoubtedly would like to protect their own reputations for accuracy, there is likely to be political pressure on them to produce favorable forecasts. Politicians may control agency salaries, budget and promotions, and agency forecasters may rely on politicians for job referrals down the road. Ignoring political pressures to bias a forecast could also carry other costs, as those promoting an unwelcome view may be ignored and unable to exercise any influence over the budget.

Source: Robert Krol, "Forecast Bias of Government Agencies," Cato Institute, Winter 2014.


Browse more articles on Economic Issues