The Arkansas Tax Reform and Relief Legislative Task Force met today (August 6) to hear the results of its contracted dynamic revenue modeling. The results generally indicated that the state could expect that the net costs of its proposed tax cuts and reforms should be somewhat less than the static revenue impacts projected by the Department of Finance and Administration ("DFA"), but not by much: up to 6% depending on the specific proposals and modeling assumptions. There is concern that the dynamic scoring failed to fairly consider the competitiveness and long-term growth benefits from tax reforms that would encourage business investment in Arkansas.
Dynamic modeling is the use of economic modeling to take into account the secondary effects of a tax policy change and their impact on bottom-line revenue. In general, policies that grow the economy should cost somewhat less using a dynamic revenue model. The use of dynamic scoring could allow the state to include more pro-growth reforms in the hoped-for 2019 tax reform bill.
The Task Force had engaged Regional Economic Models, Inc. (REMI) to conduct the modeling, and it had submitted the modeling questions at the beginning of July. Now a month later, Dr. Peter Evangelakis from REMI presented a detailed slide deck and fiscal impact report to the Task Force. It seems that Task Force members did not receive the materials ahead of the meeting. As reflected in the materials, REMI presented a somewhat bewildering array of scenarios under different assumptions, without giving much guidance to the Task Force what impact was most likely. There seemed to be some confusion among Task Force members as to what REMI's modeling really means at the end of the day.
To summarize the results of the modeling:
- The Governor's proposed tax cut from 6.9% to 6.0% for top marginal rate has a static revenue loss of $180 million. Dynamic scoring puts this number between $170 and $181 million, depending on assumptions.
- The Task Force's individual income tax cuts Option A, which basically gives everyone a slightly-modified version of the low-income tax bracket schedule, now has a static impact of $276 million. (Recall the DFA modeling adjustment from June.) Dynamic scoring put this in the range of $260 to $279 million.
- The Task Force's individual income tax cuts Option B, which consolidates all individual taxpayers onto a modified version of the middle-class bracket schedule, has a static impact of $126 million. It was analyzed in combination with an $80 million refundable earned income tax credit (EITC). Dynamic scoring gave a range between $195 and $206 million.
- Repeal of the throwback rule, which has a static loss of $25 million, is reduced by 2% to $24.5 million.
- Single sales factor, which has a static revenue score of a $9 million increase, was given a dynamic score of $8.6 million.
- Repeal of the franchise tax was scored as further reducing revenue by 1%, from a static score of $29 million to a dynamic score of a $29.3 million.
- The inventory tax repeal was modeled to grow state revenue by $2.8 million due to economic growth. The estimated $70 million static revenue cost would be to local tax revenues.
Net operating loss (NOL) reforms were not modeled.
Nicole Kaeding from the Tax Foundation critiqued REMI's dynamic scoring. While she was not present at today's meeting, Professor Jeremy Horpedahl from the Arkansas Center for Research in Economics testified to the points laid out in the Tax Foundation response. It seems that the Tax-PI model REMI used is more typically for modeling the multiplier effects of increased spending rather than for positioning states for long-term economic growth. "In short, REMI’s model is driven by the demand-side of the economy. As governments or individuals spend more money, the economy grows." More neoclassical modeling would have favored policies that increased investment by private businesses. So, for example, REMI's franchise tax repeal model had reduced government expenditure impacts that actually caused the state's economy to shrink despite the competitiveness benefits of repealing the franchise tax. Overall, the modeling seemed to overvalue government spending and individual consumption and undervalue business investment decisions.
From my perspective, the most obviously problematic modeling was for single sales factor. Instead of modeling this as a tax reduction for some in-state businesses and an increase on some out-of-state businesses, it treated the $9 million revenue gain as a tax increase on in-state production that resulted in reduced employment in the state. There did not appear to be any modeling of more favorable competitiveness in business location decisions resulting from the lower in-state tax burden from single sales factor apportionment.
Another Tax Foundation critique that seems particularly well founded is that each proposal was modeled separately. For example, the apportionment reforms of throwback repeal and single sales factor (SSF) apportionment were modeled separately rather than together. From a competitiveness standpoint, it would be surprising to go to single sales factor without also repealing the throwback rule.
DFA testified at the end of the meeting on this very point, albeit more from a revenue modeling concern: One cannot simply add up the revenue scores of single sales factor ($9 million gain) and throwback repeal ($25 million loss). Going to single sales factor would double the weight of sales factor, and thus the impact of the throwback rule. While DFA had not released an official combined static score, presumably it is around $41 million ($25 million doubled, net $9 million single sales factor gain). Other linkages in proposed income tax policy changes may emerge as well.
The Task Force will meet again tomorrow to vote on proposals for inclusion in its final recommendations.