My last comparison of U.S. states and Canadian provinces with respect to their federal transfer revenue shares got me thinking about the other revenue sources and whether any relationship could be found between economic growth and revenue composition. Income taxation is supposed to have incentive and distortion effects on saving, risk taking and labor supply and therefore one might expect to see a relationship between reliance on income taxation and economic growth rates.
Figure 1 ranks and plots real GDP growth rates in 2012 for these 60 jurisdictions. Average real GDP growth across these 50 U.S. states in 2012 was 2.1 percent whereas it averaged only 0.8 percent across the ten Canadian provinces mainly because of Newfoundland. Removing Newfoundland would raise average real GDP growth to 1.4 percent. Either way, average growth was higher across U.S. states than it was across Canadian provinces in 2012. When ranked together, the top three jurisdictions were all American – North Dakota, Texas and Oregon (with Alberta a close fourth) and the bottom three were all Canadian – Nova Scotia, New Brunswick and Newfoundland.
Figure 2 ranks and plots personal income tax revenues as a share of total government revenues for provincial and state governments in 2012. They range from a high of 28.5 percent for Quebec to a low of zero for seven U.S. states – Alaska, Florida, Nevada, South Dakota, Texas, Washington and Wyoming. Unlike the real GDP growth ranking, which saw Canadian provinces spread out from the highs to the lows, when it comes to reliance on personal income taxation, Canadian provinces are pretty much clustered in the top half. Some American states also have high revenue shares from PIT such as Connecticut and Massachusetts.
So, bringing the two variables together in Figure 3 shows a negative linear relationship between the two variables. A greater reliance on personal income taxation as a revenue source does seem to be weakly correlated with lower real GDP growth. However, this is not too definitive a result for several reasons. After all, it is only one year of observations and there are some definite outliers in the data. If you get rid of the two biggest outliers – North Dakota and Newfoundland – you get Figure 4. There is a negative relationship still but not much. Getting some more solid results definitely requires more years of data as well as data on other tax revenue shares (corporate income tax, general sales taxes) and other variables that are determinants of growth. Seems like a good project for a graduate student thesis.