Parsing the Difference Between Conventional and Dynamic Scoring | Commentary
By Robert G. Lynch The new rule adopted by the House of Representatives earlier this year requiring the use of dynamic scoring for certain types of legislation seems, at first glance, like an elegant solution to a complex budgetary issue: how to take into account the effects of proposed legislation on economic growth and the feedback effects of that growth on the budget. Yet upon close inspection, the uncertainty and biases inherent in the assumptions of the macro models that underpin dynamic scoring render the exercise largely unworkable.
Factoring in the consequences of future economic growth due to tax reform enacted in major legislation — the essence of dynamic scoring as envisioned by the House — makes sense in theory. So, too, does its theoretical use on the spending side of the ledger for big legislative initiatives such as public infrastructure investment, or investment in early childhood education. After all, if dynamic scoring were done across the board using agreed upon, accurate, non-partisan macroeconomic models then we would presumably have a better bead on the future costs and benefits of current legislation.
But the first problem with dynamic scoring is this: The economics profession today simply does not have the capabilities and tools it needs to do dynamic scoring well. We do not know how to accurately measure the growth effects of many policies. Current macroeconomic modeling, upon which dynamic scoring depends, is unsophisticated and inaccurate. The theoretical models are built upon less than rigorous, evidence-based assumptions. And the political biases that can be built into these models leaves them subject to easy manipulation.
Here’s just one of many cases in point. Back in 2003, the Joint Committee on Taxation experimented with dynamic scoring, using three different macroeconomic models with multiple sets of assumptions to come up with five different predictions, as well as the non-dynamic conventional score, about the revenue impacts of the House version of tax cut legislation that year. Twelve years later, we know the most accurate prediction was the conventional scoring method used by both the JCT and Congressional Budget Office.
The reasons? All the macro models failed to predict the Great Recession of 2007-2009 and their dynamically scored rosy revenue estimates missed by miles the actual revenue outcomes.
Conventional scoring assumes that legislation will have no effect on economic growth, calculating only how much revenue will be lost or gained by a tax change or spending proposal, although this method does take into account many changes in individual economic behavior. The evidence required for conventional scoring is well understood by economists and is rigorously empirical. The method lacks predictive power with respect to economic growth, but such prowess is only impressive if done well — an impossible task given the state of the prediction art..
Indeed, macro model-based dynamic scoring as proposed by the House ignores well-documented facts about economic growth- and revenue-inducing spending proposals grounded in decades of data-driven empirical analysis. Investing more in policies to raise academic achievement and narrow education-achievement gaps between children of wealthy families and other children could boost economic growth and revenue significantly, as I detail in my recent paper, “The Economic and Fiscal Consequences of Improving U.S. Educational Outcomes.” The same has been documented about much infrastructure spending or investment in early childhood education.
The new House rule for dynamic scoring would miss these benefits outright because it only applies, essentially, to tax legislation. And even if dynamic scoring were applied to major spending proposals the level of precision would be poor because the macro models required for dynamic scoring do not accurately incorporate the evidence of the growth effects of public investment. And then there’s the non-trivial cost of doing dynamic scoring. It’s expensive, takes time and would require regularly updated baseline economic assumptions — three things CBO and JCT neither can afford to do nor has the time to do.
The bottom line: economists at the two scoring arms of Congress just aren’t able to produce robust dynamic scores of proposed legislation because the macro models aren’t good enough to accurately estimate economic growth and the revenue consequences of that growth. It’s a fact and we need to deal with it.
Robert G. Lynch is a visiting fellow at the Washington Center for Equitable Growth and the Everett E. Nuttle Professor of Economics at Washington College. His areas of specialization include human capital, public policy, public finance and income inequality.
The 114th: CQ Roll Call’s Guide to the New Congress
Get breaking news alerts and more from Roll Call in your inbox or on your iPhone.