St. Paul Pioneer Press (MN) – Sunday, May 19, 2013
Gov. Mark Dayton wisely urged Minnesotans to use facts in the budget debate in a speech to the Chamber of Commerce. Despite this sage advice, he continued to rely on a flawed analysis as an important rationale for his budget’s higher taxes.
“If the lowest taxes equaled the highest job growth and the highest per capita income, public policy would be clear and simple,” Dayton told the Chamber. “In fact, however, the opposite is true. If you look at the last chart, two pages of your handout, people living in the states with the lowest taxes generally have among the lowest per capita incomes. And those states have inferior job growth,” he stated.
Gov. Dayton’s data analysis has multiple statistical mistakes. One critical error is selection bias. He looks at a small subsample of the available evidence that was carefully cherry-picked, instead of drawing a conclusion from all of the available evidence.
Dayton showed data from only six states in one table and from only eight states in a second table before drawing the specious conclusion that these cherry-picked numbers prove that lower taxes equal lower job growth and lower per capita income.
New Hampshire, for instance, was clearly picked for one table to “prove” that New Hampshire’s low taxes lead to New Hampshire being 46th in job growth in 2012, while New Hampshire’s high per capita income failed to appear in the other table. Similarly, Dayton’s tables suggest that North Dakota’s job growth comes from its high tax collections, instead of the Bakken oil boom that has led to both North Dakota’s job growth and a gusher of energy-related tax revenue.
Dayton’s analysis also suffers from a tiny sample, a lack of statistical hypothesis testing, and the omission of any control variables. These serious statistical flaws mean that Dayton’s inference that higher taxes equal higher incomes and job growth is not supported by his methodology.
Most peer-reviewed research examining the aggregate effect of state taxes finds “a negative relationship between taxation and growth,” while a few papers “do not detect any significant negative impact of taxes on growth” according to a careful literature review by economists Ergete Ferede and Bev Dahlby in last September’s National Tax Journal. This literature review does not find a robust positive relationship between aggregate state taxes and income growth.
Research on the aggregate effects of taxes is complicated by several technical issues of data quality. Aggregate measures of taxes mix more- and less-distortionary types of taxes. Aggregation also conflates productive public expenditures with wasteful public expenditures.
Even careful analysis of the aggregate effects of typical state tax and spending patterns doesn’t provide the best guidance about the economic consequences of specific increases in taxes and spending.
A better way to understand the effects of higher taxes and spending is to consider each tax hike and spending program on a case-by-case basis.
Each dollar of government spending costs the private sector both the dollar taken from people in the private sector plus the economic damage from all the ways taxes distort decision-making.
People change their behavior in response to changes in taxes. Higher income tax rates are a disincentive to work, work overtime, or work multiple jobs. Higher income tax rates penalize education and on-the-job-training by lowering the after-tax rewards for investment in productive human capital.
At low tax rates, this economic damage from taxes is small. But at higher tax rates, these negative effects of taxes are larger.
The tax bills in St. Paul would increase effective marginal income tax rates above 50 percent for some primary earners and above 60 percent for affected secondary earners. In other words, these taxpayers would gain less than the government from additional hard work to earn another dollar of pre-tax wages.
At the proposed federal plus Minnesota income-tax rates, each extra dollar of government spending is likely to cost the private sector around $1.25 to $3. I applied standard economic formulas and the proposed tax rates to a range of estimated behavioral responses from other economists’ research to produce this imprecise estimate.
The logical case for higher taxes and spending requires explaining why the private sector should shrink by more than a dollar for each additional dollar spent by the public sector.
Gov. Dayton should lead elected officials from both parties to eschew flawed analysis in favor of careful enumeration and consideration of the real costs of additional taxes and the benefits of each spending program based on the best possible analysis.
John Spry of St. Paul is an associate professor in the Department of Finance at the University of St. Thomas. He’s on Twitter at @JohnASpry .