Dynamic Scoring: Caveats and Controversy

Dynamic Scoring and Georgia

While the debate over dynamic scoring has been focused on the national level, the Fiscal Research Center recently released “A Multiregional Model of Growth Oriented State Tax Reforms: An Application to Georgia and Five Comparison States,” which applies a form of dynamic modeling to tax reform in Georgia. This report was highly controversial internally, even among the authors of the report, not least because the model finds a relatively significant positive effect of fundamental tax reform on a variety of metrics, but most notably, in one scenario, personal income shows a 0.67 percentage point increase in the compound growth rate over 10 years over a baseline of 3.37 percent compound growth rate. (Yes, that is a 20 percent increase!)

First a recap of the report for the layperson:

The analysis uses a Computable General Equilibrium (CGE) model, a series of economic equations based on economic theory interacted with input-output tables. The input-output tables are matrices that mathematically describe the interrelationships between different industries throughout a state’s economy. This particular model looks at Georgia and five neighboring states: Alabama, Florida, North Carolina, South Carolina and Tennessee. The analysis establishes a baseline growth rate in Georgia and the other states for personal income, consumption, labor supply and demand for capital. Perhaps of most interest is the projected 3.37 percent compound rate of growth over the next 10 years for Georgia. This growth rate is greater than Georgia realized over the past decade (so it may raise some eyebrows), but the main point is to establish a baseline over which the growth from tax reform can be measured.

The model then examines three revenue-neutral tax reform scenarios over 10 periods and finds the following:

  1. Expanding the sales tax base to include services and a revenue-neutral reduction of the personal income tax (PIT) increases the 10-year compound growth rate of personal income by 0.51 percentage points over the baseline.
  2. Eliminating the PIT with a revenue-neutral increase in the sales tax rate on the expanded sales tax base increases the 10-year compound growth rate of personal income by 0.67 percentage points over the baseline.
  3. Eliminating PIT and the corporate income tax (CIT) with a revenue-neutral increase in the sales tax rate on the expanded sales tax base increases the 10-year compound growth rate of personal income by 0.62 percentage points over the baseline.

Caveats and points of controversy:

A CGE model is a very fancy mousetrap, but an important piece to keep in mind is that it will largely produce the results that theory predicts because the equations themselves are derived from theory. Given that economic theory predicts that reducing the price of capital (by cutting income and corporate income taxes) will cause capital to flow into a state and will make labor more productive relative to capital, we would be surprised if there were not a positive effect. But bear in mind, if the theory is wrong or the underlying equations are wrong then the results are wrong.

A point of internal dissention revolved around the difference between theory and empirical findings – namely most research finds small to no effect from shifting from income to sales taxes (and the relative shifts in capital flow) or are not able to appropriately measure the impact (i.e., see Sjoquist, FRC Policy Memorandum Replacing the Income Tax with Increased Sales Taxes: What Do We Know?, http://cslf.gsu.edu/files/2014/06/Memo_16FIN.pdf and Bluestone and Bourdeaux, CSLF Report No. 12 Dynamic Revenue Analysis: Experience of the States, http://cslf.gsu.edu/files/2015/04/Dynamic-Revenue-Analysis_April2015.pdf). Notably, in another FRC report examining other state estimates of dynamic effects on taxes, the effects appear to be small relative to a state’s overall revenues and will often be within the error rate of the revenue estimate.

The FRC analysis does not directly calculate the revenue effects, but obviously, if there is a 20 percent (0.67/3.37) increase in personal income, it would follow that consumption also would go up and feed back into tax revenues in a positive way.

Some other important points:

  • The growth in the model is not so much because Georgia simply becomes a stronger economy but because it “steals” capital from neighboring states. Other state economies have declining personal income because capital is shifting into Georgia. The model also assumes that the other states do not respond to Georgia’s changing tax structure.
  • The model does not capture the pyramiding effects of the retail sales tax but assumes that the state is able to effectively tax only the final products sold in the state (i.e., a very well-designed retail sales tax is imposed).
  • The model assumes that the state would be able to expand the sales tax base — something that the state failed to do in its previous efforts at fundamental tax reform.

So, consider this a very interesting and provocative analysis, but also recognize that there are some important assumptions in it.

Carolyn Bourdeaux