Would you make an effort to find ways to reduce your tax burden if your tax rate was suddenly raised 50 percent? The higher one’s income, the more incentive a person has to find ways to minimize his tax burden — which is why very high tax rates on the rich always fail to produce the projected revenue.

The new Republican Congress is going to make a couple of critical personnel decisions within the next few weeks. The Congressional Budget Office has the responsibility for scoring the cost and economic effects of spending bills, and the Joint Committee on Taxation (JTC) has the responsibility for scoring and gauging the effects of tax bills. The staff directors of these committees are of great importance because conclusions they make regarding the impact of tax and spending bills greatly influence decisions of Congress. Historically, these number-crunchers relied on “static” models, which largely ignore changes in behavior as a result of changes in tax rates or government spending. The alternative approach is to use what are commonly referred to as “dynamic” models, which incorporate expected changes in behavior. Critics of the dynamic model approach argue that it is difficult to project changes in behavior, and it is easier to calculate changes in expected tax revenues if no change in behavior is assumed. In other words, the advocates of the static approach believe it is better to be precisely incorrect than approximately accurate.

Reliance on static models leads to an overestimation of tax revenues from any tax rate increase and an overestimation of the tax revenue loss from any tax rate reduction. This bias leads to higher levels of taxation and spending than is best for economic growth and job creation. The current JTC tax model now includes some dynamic estimates, but the committee has not revealed the assumptions and interactions within the model, thus keeping outside economists from being able to review it.

The errors in the tax and spending models in the past have been considerable and destructive for policymakers. For example, there have been a number of changes to the capital gains tax rates over the years. In 1978, 1981, 1986 and 1996, the Joint Committee on Taxation not only produced grossly inaccurate numbers but did not even get the direction of the sign correct. That is, it scored what turned out to be a revenue gain as loss and vice versa because it largely ignored the “unlocking effects” (the greater willingness to sell and buy assets with lower rates), and the increase in investment and job growth resulting from the lower rates.

The current tax and spending models do not adequately incorporate behavioral changes. Thus, they are biased toward the shorter rather than the very long run. For instance, if there is a sudden and unexpected increase in tax rates, most people will pay the higher rate in the immediate weeks and months after the rate increase. Over time, though, people will find more and more legal or even illegal ways to avoid the higher rates. It was this type of analysis (not apparently included in the JTC models) that caused the Scottish economist and 1996 Nobel prize recipient James Mirrlees to conclude that over the long run, tax rates of more than 20 percent on the richest people reduce both tax revenue and economic growth.

Republican congressional leaders have pledged to undertake tax and spending reform. To do so, they need accurate projections of the impact on tax revenues, job creation and economic growth resulting from their possible alternative reforms. This requires realistic models of what is likely to happen, so they need to find the best possible experts to manage the Joint Committee on Taxation and the Congressional Budget Office.

These are some of the questions I would pose to the candidates for these positions: First, what do you think the revenue and growth-maximizing rates are over the long run for the corporate income tax, individual income tax and capital gains tax?

Second, what level of government spending do you think maximizes economic growth and job creation over the long run?

Respondents would likely give ranges rather than point estimates as answers to each of the questions and refer to specific studies to back up their answers.

Congress should also obtain commitments from whomever it selects to publish the models and open them up for peer review and comment. Finally, Congress should put on retainer organizations that have credible models (such as the Tax Foundation) as a check on the Joint Committee on Taxation and Congressional Budget Office — and Congress should feel free to accept the results of the alternative private models when making tax and budget decisions.