Friday, December 28, 2012

"All I need is a miracle" -- growth in the no-personal-income-tax states

Earlier this week, I analyzed a proposal that the John W. Pope Civitas Institute has made to eliminate the corporate, personal, and business franchise taxes that North Carolinians pay and to replace these with a higher and expanded sales tax, a business license fee, and a real estate conveyance fee (a real estate sales tax). The proposal would shift the responsibility for paying taxes away from North Carolina's wealthiest households and most prosperous corporations and towards its poorest households and smaller businesses.

Civitas acknowledges that the tax proposal would be regressive but claims that it would lead to greater economic growth. As evidence it cites the experiences of states that do not assess personal income taxes and of those that do not assess corporate income taxes. These states experienced higher growth in their Gross Domestic Products (GDPs) than other states.

As I wrote in the earlier post, this contention is correct, but it is hardly evidence of the effect of the states' tax policies. For example, total GDP is influenced by the number of people in a state, and GDP growth is influenced by population growth. Each of the no-personal-income-tax (NPIT) states experienced above-average population growth. Several of the states have other things that make them unique. Below I analyze the growth rates for the states, highlighting some of their characteristics.

Compounded Annual Growth Rates 2002-10

Actual nominal GDP growth
Predicted growth from national price trends
Predicted growth from national output trends
Predicted growth from national trends
Population growth
Predicted growth with population adjustment

US Total

S. Dakota

The first column in the table shows the compounded annual growth rate (CAGR) in U.S. GDP and each state's GDP from 2002-2010. The data are taken from the Bureau of Economic Analysis, and the CAGRs are calculated as
The calculations confirm that GDP growth in the NPIT states was higher than the national average.

To examine what might have contributed to the states' growth rates, I went back to the BEA and grabbed all of the available industry detail for the 2002 GDP figures. These figures show how much of each states' initial GDP came from particular industries. For example, in 2002, 11.6 percent of Alaska's economic output came from oil and gas extraction; 5.2 percent came from construction, 7.0 percent came from pipeline transportation, and so on.

Definitionally, total GDP growth comes from two factors: the change in the value or price of the output and the change in the quantity of the output. If prices of certain goods have changed over time, the value of GDP could increase without any change in quantity (or in tax policy). Energy products are one such good; their price increased at several times the rate of other goods. If a state specializes in producing a good (like energy) that experiences a big price rise, its GDP will grow faster, other things held constant, than other states' GDPs.

To examine how much price changes for each state's specialized mix of produced goods might have contributed to nominal GDP growth, I calculated the growth rates for the price changes for each specialized industry from 2002-10 (e.g., prices of oil and gas extraction output increased just under 13 percent) and applied these price changes to each state's initial mix of outputs. This tells us how much the state's GDP would have increased if the quantity of its output had remained fixed and if prices had followed the national trends. For comparison purposes, I also performed the same calculation for the US as a whole. The results are listed in the second column of the table.

As one might expect, the energy-rich states of Alaska, Texas, and Wyoming enjoyed a tremendous advantage over the other states in terms of the prices that they were able to get for their particular outputs. From these price contributions alone, we would have expected annual GDP growth to be 2.6 percent higher in Wyoming, 2.2 percent higher in Alaska, and 0.7 percent higher in Texas than the national average. Expected GDP growth from price changes for Nevada and Florida were very close to the national average; Washington was slightly disadvantaged, and South Dakota had a 0.4 percent disadvantage. On average, the NPIT states had a 0.7 percent growth advantage in terms of prices for their outputs relative to the U.S. as a whole.

We can also consider how much national trends in output contributed. For example, if a state was specializing in a sector with output that was generally expanding, such as farming which grew at more than three times the rate of other output from 2002-10, it might be expected to have a GDP growth advantage. Conversely, if the state initially specialized in a declining sector, such as automobile manufacturing which declined at a rate of 8 percent per year from 2002-10, it would have a disadvantage.

To examine how much national output trends might have contributed to GDP growth, I performed a similar analysis in which I took the states initial output mix in 2002 and then artificially grew each component by the corresponding national trend in output growth. The results from this exercise are shown in the third column of the table. The figures indicate that South Dakota had a 1.5 percent output advantage; this was due mainly to its specialization in agriculture, which contributed just over a fifth of the state's total GDP growth. Texas also had an output advantage. Florida and Washington had slight disadvantages, and Alaska, Nevada, and Wyoming had sizable disadvantages. On average, the NPIT states had a slight 0.1 percent disadvantage in terms of the national trends in the quantity of their outputs.

If we put the price and output trend figures together, we can calculate how much a state's GDP would have been expected to change as a result of national price and quantity trends. These figures are shown in the fourth column. The figures show that the price trends generally outweighed the quantity trends. On the basis of national trends, we would have expected Alaska to have 1.7 percent higher growth than the national average, Wyoming to have 1.4 percent higher growth, Texas to have 0.7 percent higher growth, and South Dakota to have 0.4 percent higher growth. Expected growth in Florida and Washington would have been close to the national trend, while growth in Nevada would have been 0.3 percent lower than the national trend. On average, the NPIT states had a 0.5 percent advantage in terms of national trends in prices and quantities of their outputs.

Finally, all but one of the NPIT states added population faster than the rest of the country. The annual population growth rates are shown in the fifth column of the table. Nevada, the fastest growing state in the country in terms of population over this period, had a growth rate that was three times the national average. Along with population growth, the big drivers in Nevada's economy over this period were a joint boom in real estate and lending. On average, the NPIT states added population at a rate that was 0.7 percent faster than the national average.

So, how did the NPIT states perform relative to what we might have (admittedly crudely) expected on the basis of price growth, national output trends, and population growth? Wyoming outperformed this measure of expectations by 2.0 percent. Alaska outperformed by 0.4 percent, and Washington outperformed by 0.3 percent. Actual GDP growth in Texas was exactly the amount predicted. Growth in Nevada and South Dakota was slightly worse than expected, and growth in Florida was 0.4 percent worse than expected.

This is hardly resounding evidence of the superiority of NPIT policy. Instead it shows that the NPIT were "lucky ducks" in a number of ways. Three states (Alaska, Texas, and Wyoming) benefited strongly from the run-up in energy prices. One state (South Dakota) caught a tremendous break from the increased demand for agriculture, especially corn used for the ethanol mandate in transportation fuels. A fifth (Nevada) grew because of the housing bubble.

There are miracles in these growth patterns, just not any that are tax-related.

1 comment:

Joe said...

Nice and clear. I like this kind of analysis.

How many significant digits do figures like this have?