On November 1, 1999 the Cato Institute released a paper by Stephen Moore and Richard Vedder titled: "The Case Against a Tennessee Income Tax."⁽¹⁾ In their paper, Moore and Vedder endeavor to demonstrate that the adoption of an income tax in Tennessee would reduce economic growth in the state and result in much greater government spending.⁽²⁾

The economic consequences of Tennessee adopting an income tax is, of course, an important issue. The Cato paper, however, has critical methodological flaws which greatly limit the value of its conclusions. When the methodological errors in Cato's paper are corrected, we find no evidence that the presence of an income tax hurts state economies. Nor is there evidence that state and local spending grows more rapidly in states that have personal income taxes than in states that do not.

High-Tax States and Economic Growth

The Cato paper claims that states with high taxes experience worse economic performance than states with low taxes. As proof, Moore and Vedder compare the economic performance of two groups of ten states selected by level of state taxation. They find that the ten highest tax states in the country have had worse economic performance than the ten lowest tax states. From this evidence, they argue that high taxes hurt state economic performance.

Methodological Flaws

In general, Cato's methodology is flawed in two ways. First, Cato incorrectly selects states for its highest tax and lowest tax groups. Cato is not comparing the economic performance of a group of states that is, in any relevant way, "high tax," with a group of states that is "low tax." Thus, the Cato analysis cannot demonstrate the relative economic performance of high and low tax states.

The second general flaw in Cato's methodology is the way it measures economic performance. Even if the Cato paper was comparing the right states, it is not using the best means for comparing how they have done economically. An analysis of relative economic performance of states that doesn't use reliable measures of how state economies perform is of little value.

Selecting the Comparison Groups of States

For several reasons, the two groups of states selected by Cato for comparison do not represent a relevant group of high and low tax states.

The Wrong Taxes

One problem that pervades the Cato paper is a confounding of two questions: (1) Do higher taxes, in general, hurt a state's economy? And, (2) does a personal income tax, in particular, hurt a state's economy? The Cato analysis compares the economic performance of states with high total state tax burdens with states with low total state tax burdens--counting all state taxes. Thus, their high tax list includes states such as Alaska, Wyoming and Washington that don't have personal income taxes and their low tax list includes states that have significant income taxes. Even if there were no other flaws in Cato's methodology, comparing these two groups of states would, at best, tell us whether high overall state tax burdens hurt economic growth. The impact of higher overall taxes is, however, a different question than whether the use of a personal income tax in a state's revenue mix hurts economic growth. To answer this question--a more relevant one for low-tax Tennessee--Cato should have compared high income tax states to low income tax states.

Cato's selection of states for its high and low tax lists is also distorted by the exclusion of local taxes. Local taxes account for 40 percent of combined state and local tax revenues nationally. The level of local taxes varies significantly among states. If tax burden affects economic performance of states, as Cato claims, the amount of local taxes must matter. Yet in determining whether a state is high tax or low tax, Cato ignores them. Without local taxes, Cato is not correctly identifying which states impose low versus high overall burdens on their taxpayers.

The Wrong Measure of Tax Burden

In addition to problems in selecting the taxes to include in its analysis (all taxes versus the personal income tax only, state versus state and local), Cato's measure of the level of the burden imposed by the taxes selected is flawed. Cato uses per capita taxes--the amount of taxes collected per-person in the state--to categorize states as either high tax or low tax. But the relative level of per capita taxes among states is a very poor measure of the relative tax burden imposed by states.

For example, Connecticut collects $1,040 per capita in personal income tax, while Maine collects $728. Yet Connecticut's personal income tax has much lower tax rates and higher exemptions than Maine's personal income tax. Virtually anyone moving from Maine to Connecticut would, in fact, see their income tax liability go down.

Why does Connecticut collect more income tax per person even thought its income tax rates are lower and exemptions higher? Because Connecticut is a richer state than Maine. Overall, the people of Connecticut have significantly more income per person than the people of Maine so they pay more income tax per person. But if a Connecticut taxpayer and a like Maine taxpayer with identical income are compared, the Connecticut taxpayer will, in the vast majority of cases, pay less in income tax.

It is very hard to see how Connecticut's personal income tax could be a greater economic burden to Connecticut than Maine's income tax is to Maine.

• The overall percentage of income paid in income tax is lower in Connecticut than in Maine (2.8% vs. 3.2%).
• The amount left to taxpayers after their taxes are paid--money they can invest or spend for the good of their state's economy--is more in Connecticut than in Maine.
• If someone were to decide whether to move to Connecticut or Maine based on the amount of income tax they would pay, they'd move to Connecticut.

In other words, notwithstanding its higher per-capita rating, Connecticut's personal income tax burden is lower than Maine's.

By using the per-capita measure of tax level, Cato is putting states on its high and low tax lists because of their relative wealth instead of the tax burdens they impose. A better measure of tax burden is taxes as a percentage of personal income. This measure accounts for differences in the level of income supporting the taxes collected. It is much more likely to accurately reflect how taxpayers' burdens are different between states.

A final criticism of Cato's identification of states as either high or low tax is that it was based on the level of taxes for a single year even though economic growth was measured over seven years. A better approach is to measure average tax level over the entire period over which economic growth is measured.

Measuring Economic Performance

In addition to the problems in Cato's identification of states as either high or low tax, the measures used for economic performance are flawed.

Time Period

Cato's selection of 1990 to 1997 for its analysis is an arbitrary one. Of course, any time period selected could be criticized for being arbitrary. It is important, however, to point out that analyses of economic growth rates in states are extremely sensitive to the time periods selected. Regional economic trends are much more powerful than state-specific economic trends. Thus, choosing years in which there are large differences in regional economic performance can greatly distort results. Differences in state economic conditions can be inappropriately attributed to characteristics of the states (such as taxes) instead of regional factors.

The 1990 to 1997 period is problematic because it covers only part of a single national economic cycle. Different states and regions were affected differently by the recession of the early 1990s and have recovered at different rates. And, among other distorting factors, there have been large differences in the impact of federal defense spending cutbacks in this period. Any economic trend analysis that uses such a short period is likely to confront problems of this nature. Growth rates for the 1990 to 1997 span are particularly difficult to use for an accurate assessment of states' relative economic health.

Growth Statistics

Cato uses growth in population, personal income and employment as its measures of economic performance. Although it is a point on which reasonable minds may differ, we do not believe these are the best measures. One problem is that, for some states, population growth can be the cause of personal income and employment growth. States with high population growth, for example, often have high aggregate income growth simply because people's income moves into the state along with the people themselves.

Where population change is driving income and employment change, either up or down, the question then becomes whether population growth is, in itself, a good measure of economic performance. The argument is sometimes made that population growth is a good measure because people move to where the economic opportunities are. This is, to some degree, true. Of course, people also move to where the weather is good. In addition, if low-income people are flocking to a state and taking poorly paying jobs, while current residents' incomes are stagnant, it would be misleading to characterize the state as a good economic performer.

For these reasons we prefer per-capita income growth as a measure of economic performance. This measure shows whether the average income in the state, for whomever is there, newcomers or long-time residents, is going up or down. Income of a state's residents is, after all, the best measure of how well a state's economy is performing for its citizens.

Summary of Methodological Flaws

Cato's choice of taxes to include and use of per-capita taxes to categorize states as high- or low-tax is highly misleading. Cato puts states with modest tax burdens on the high tax list, independent of whether they have a high personal income tax. Some states on the low tax list have significant income taxes. Moreover, by ignoring local taxes, Cato disregards an important factor in the differing levels of tax encountered by taxpayers in different states. Furthermore, the lists of states are highly dependent on states' relative wealth.

In addition to the problems in identifying states as high tax or low tax Cato uses inadequate measures of economic growth. The time period selected and the growth statistics used make Cato's characterization of states as strong or weak economic performers unreliable.

In sum, Cato's lists of high tax and low tax states are not actually lists of high tax and low tax states by any relevant and significant measure. Nor are Cato's measures of economic performance well chosen. Thus, correlation between high economic performance by the one group and low economic performance by the other group, does not provide us with useful information on whether high taxes, or personal income taxes in particular, adversely affect economic growth.

Revising the Cato Analysis

ITEP has conducted analyses similar to the Cato Institute's but correcting for the methodological flaws identified above. We, as did Cato, take two groups of states--one high tax and one low tax. We then compare the economic performance of the two groups.

The key differences between ITEP's and Cato's analyses are:

• In addition to comparing states with high total taxes to states with low total taxes, ITEP compares states with high personal income tax collections with states with low personal income tax collections. In this way we separately address two questions: do high overall taxes hurt states' economies and do personal income taxes, in particular, hurt states' economies?
• The ITEP analyses include local taxes. If state taxes affect economic performance, local taxes do too. Any rational study looking for a connection between tax burden and state economic performance has to include both.
• ITEP uses taxes as a percentage of personal income instead of per-capita taxes as the measure of tax burden. This is a superior measure of the burden a tax might place on an economy.
• ITEP uses the average tax burden over the entire period of analysis instead of a single year snapshot.
• The ITEP analysis is over a longer period and covers more economic cycles than the Cato analysis. ITEP uses 1980 - 1998 for the personal income tax analysis and 1980 - 1996 for the total taxes analysis. Cato uses 1990 - 1997.
• For the reasons given above, ITEP uses per-capita personal income growth as its primary measure of economic performance.

ITEP Results

State & Local Taxes as a % of Personal Income: The 10 Highest and the 10 Lowest
	State & Local Tax Burden, 1980-96	US RANK	Annual Per Capita Income Growth, 1980-96	US RANK
US Average	10.7%		+1.46%
Alaska	23.6%	1	–0.5%	50
Wyoming	15.1%	2	–0.1%	49
New York	14.7%	3	+2.0%	7
Hawaii	12.6%	4	+1.2%	37
Wisconsin	12.3%	5	+1.2%	33
Minnesota	12.1%	6	+1.7%	22
New Mexico	11.8%	7	+1.1%	40
Maine	11.7%	8	+1.8%	15
Vermont	11.6%	9	+1.9%	11
Montana	11.5%	10	+0.7%	46
Highest 10 Average	13.9%		+1.50%
1996 Per Capita Income (1998 dollars)			$27,158
Colorado	9.8%	41	+1.4%	29
Texas	9.6%	42	+1.0%	42
Indiana	9.6%	43	+1.4%	30
Virginia	9.4%	44	+1.7%	19
Arkansas	9.4%	45	+1.8%	16
Florida	9.1%	46	+1.4%	28
Alabama	8.9%	47	+1.9%	13
Missouri	8.8%	48	+1.5%	24
Tennessee	8.7%	49	+2.2%	5
New Hampshire	8.4%	50	+2.1%	6
Lowest 10 Average	9.4%		+1.30%
1996 Per Capita Income (1998 dollars)			$23,717
NOTE: 1996 is the latest year for which state & local tax data are available. SOURCE: Bureau of the Census, BEA, ITEP's Calculations.

First, we examine the question of whether raising taxes, in general, is likely to hurt Tennessee's economy. Note first that Tennessee, over the 1980 to 1996 period⁽³⁾, ranked 49^th in state and local taxes as a share of personal income. In 1996 Tennessee also ranks 49^th. If Governor Sundquist's tax package were to pass, we estimate that Tennessee's rank would, indeed, go up--to 48^th(4). In other words, Tennessee would continue to be an extremely low tax state. There is nothing in Cato's data, or ours, that indicates that small differences in tax burden between very low tax states make a noticeable difference in economic performance.

Notwithstanding the fact that there is no plan in Tennessee to make it a high tax state, ITEP redid the Cato analysis comparing the economic performance of high- and low-tax states--revising the methodology as described. With the corrected methodology, the ITEP analysis produces very different results than the Cato analysis. As the table shows, ITEP found that the ten highest tax states actually out-performed the ten lowest tax states--with a per-capita income growth rate of 1.5 percent versus 1.3 percent (these, and other growth rates presented here are adjusted for inflation). This is not to say that higher taxes are the linchpin to economic success. We do not claim that simple correlations imply causation. But, clearly, over the long-run, being a high tax state is not the bar to a successful economy that Cato suggests.

Since, however, the net tax increase proposed for Tennessee would have little effect on Tennessee's level of taxation relative to other states, the more important question is whether adopting a personal income tax is particularly harmful to an economy. The next table shows the results of the ITEP analysis of the impact of personal income taxes on economic performance. The results are striking: the 10 states with the highest personal income tax burdens had higher average annual growth (+1.9%) in per capita personal income than the 10 states with the lowest personal income tax burdens (with an average annual growth in per capita personal income of +1.5%). While the 10 high tax states experienced significantly higher growth than the US average, the 10 low tax states grew more slowly than the US average.

• the 10 very highest income tax states have a higher average per-capita income that the 10 lowest income tax states.
• Nine of the top ten states with the highest per capita personal incomes have broad-based income taxes (New Hampshire is the only non-income tax state to crack the top ten).
• Six of the nine states with no broad-based income tax have per capita personal incomes below the national average. Five of those states have been "income tax-free" for over one hundred years and still have per capita income below the national average.⁽⁶⁾
• The average per capita personal income of the 41 states (and the District of Columbia) with personal income taxes is more than four percent higher than states without income taxes.

The Nine States Without Broad-Based Personal Income Taxes
	1998 Per Capita Personal Income	% above/below national average	# of years state has not had an income tax
Alaska	$25,675	–2.8%	19
Florida	$25,852	–2.1%	153
Nevada	$27,200	+3.0%	134
New Hampshire	$29,022	+9.9%	210
South Dakota	$22,114	–16.3%	109
Tennessee	$23,559	–10.8%	202
Texas	$24,957	–5.5%	153
Washington	$27,961	+5.9%	109
Wyoming	$23,167	–12.3%	108
Average for Nine No-Income Tax States	$25,532	–3.3%
Average for All States WITH Income Tax	$26,614	+0.8%
UNITED STATES	$26,412
NOTE: Except for Alaska (which repealed its income tax in 1979), the number of years the state has not had an income tax is the difference between 1998 and the year the state was admitted to the Union. SOURCE: Bureau of Economic Analysis; ITEP's Calculations.

And what of Tennessee? The state ranks 33^rd, with a per capita personal income 11 percent below the national average. Twenty-six states with broad-based personal income taxes have per capita personal incomes higher than Tennessee.

These data hardly suggest that forgoing a personal income tax is a key to long-term economic success. The bottom line is that forgoing an income tax has not delivered superior economic conditions.

How does Tennessee Compare with its Neighbors?

In addition to its nationwide comparison, Cato attempts to show that the lack of a broad-based income tax in Tennessee provides a competitive advantage over its immediate neighbors. To prove this point, Cato compares Tennessee to only one of its eight neighbors, Kentucky--finding that Kentucky's economic growth since 1980 has lagged behind that of Tennessee. Cato justifies the choice of Kentucky for comparison purposes by stating that "Kentucky has the longest border and is probably the state most similar to Tennessee in most respects, including, historically at least, economically. For example, in 1980, per capita personal income in the two states differed by only eight dollars."

Looking beyond border length, Kentucky and Tennessee have two quite different economies. Less than half of Kentucky's population lives in urban areas versus 68 percent of Tennessee's population. In fact, Tennessee has two of the nations most populous 25 cities. These cities, Nashville and Memphis, are the largest in the region. Kentucky does not have a city in the top 50 nationally. Also, Kentucky relies more heavily on manufacturing, mining, and agriculture than Tennessee. Finance and services are more important in Tennessee's economy.⁽⁷⁾

Tennessee's economy is as similar to several of its neighbors' economies as it is to Kentucky's economy. The only obvious factors that make Kentucky uniquely appealing for Cato's comparative analysis are that it has an income tax and has had worse economic performance than Tennessee.⁽⁸⁾

To do a more complete analysis, ITEP compared Tennessee's economic growth to all eight of its bordering states. In addition, trends in the region are compared to national trends. The following table compares the average annual per capita personal income growth rates from 1980 to 1998 for Tennessee, all eight of its bordering states, and the national average, along with the average personal income tax burdens for the time period. The results of this comparison show that all nine states in the region had higher-than-average growth in per capita personal income, and more than half (five of the nine) had personal income tax burdens higher than the national average. The two states with the highest economic growth in the region, Georgia and North Carolina, both have higher personal income tax burdens than the national average. Yet they out-performed all their neighbors, including income-tax-free Tennessee.

ITEP's analysis of state and local direct general expenditure growth from 1990 to 1996⁽⁹⁾ shows Tennessee had the 7^th fastest spending growth rate in the nation. But three of its neighbors (Mississippi, Kentucky, and Arkansas) saw their per capita spending grow even faster than Tennessee's. Seven of the eight states bordering on Tennessee had higher-than-average expenditure growth (only Virginia had slower growth than the national average). As a result, Tennessee's national ranking in per capita expenditures actually fell from 46^th in 1990 to 47^th in 1996. Thus, it appears that many low-spending states (four of the five per capita lowest-spending states are located in the region) have seen their rate of spending increase at higher-than-average rates--probably due to the need to improve neglected public services. In any case, it is clear that Tennessee has been, continues to be, and will continue to be, a very low spending state.

The Income Tax as a Ticket to High Spending?

Cato also cites a "landmark" study that claims that when states introduce an income tax, state governmental expenditures rise dramatically.⁽¹⁰⁾ But a paper recently released by ITEP⁽¹¹⁾ found the conclusions of the National Taxpayers Union paper to be misleading. The NTU report claims that in seven of the nine states that enacted income taxes since 1967, state spending has grown at a faster rate since adoption of those income taxes than before adoption. But this ignores an important factor by looking at only state spending rather than total state and local outlays. Many states adopt income taxes for the express purpose of providing local property tax relief. In such states, an increase in state spending would be expected. But the net impact, state and local, is much smaller, or even non-existent.

Of course, some states have adopted state personal income taxes with the objective of increasing revenues to pay for expanded government services. There is nothing inherently wrong with the citizens of a state choosing to do that--but it does not mean Tennessee will embark on a long term spending binge simply because it adopts an income tax.

This is confirmed by expanding on NTU's comparison of the experience of states that adopted an income tax since 1967 with those that still have no income tax. Of the nine states that adopted an income tax since 1967, five increased per capita state and local spending at a faster rate than the national average from 1991 to 1996. The other four states saw lower growth in per capital state and local spending than the national average.

Likewise, of the nine states without an income tax, five increased per capita state and local spending at a faster rate than the national average from 1991 to 1996, while four saw lower growth.

Thus, in the 1990s, there has been little difference in state and local spending growth rates between the states that most recently adopted income taxes and those that have remained income-tax free. Hence, adoption of a personal income tax does not appear to lead to long-term, consistent, high rates of growth in spending.

Conclusion: No Evidence That the Presence of an Income Tax Harms a State's Economy

Policy makers in Tennessee are currently in the process of making major decisions about the future structure of the state's tax system. It is essential that Tennesseans understand the economic consequences of the tax reforms currently under discussion.

Yet, as the preceding analysis has shown, data recently released by the Cato Institute tends to obscure--rather than clarify--the relationship between tax burdens and economic growth. While the Cato paper's authors find that high tax states have grown more slowly than have low tax states in the 1990s, this finding is dependent on the highly questionable way they classified the states and the way they measured economic growth. Our analysis finds that, when several methodological errors in the Cato study are corrected, the relationship between tax burdens and economic growth is exactly the opposite of Moore and Vedder's claim: high tax states have outperformed low tax states. More importantly, since Tennessee will remain a low-tax state even if it adopts an income tax, states relying most heavily on personal income taxes have experienced greater economic growth than have low (and no) income tax states.

Other evidence, whether comparing no-income-tax states to income-tax states, or focusing purely on Tennessee and her neighbors, confirm this analysis. This doesn't prove that adopting a personal income tax would be beneficial to the Tennessee economy. But forty-one states, with more than 80 percent of the nation's population, have broad-based income taxes. At any given time, some of these states have had strong economic growth and others have performed less well. Likewise, some of the nine states that choose not to impose an income tax have performed well, while some have done poorly. Our conclusion is simply that if the presence of an income tax has any economic impact--positive or negative--it is dwarfed by far more powerful economic forces.

The Cato paper's assertions regarding Tennessee's government spending habits and the probability of a spending binge if an income tax becomes law are similarly misleading. Tennessee will continue being one of the lowest spending states in the country whether an income tax passes or not.

Back To ITEP Home

1. Stephen Moore and Richard Vedder, The Case Against a Tennessee Income Tax, Cato Institute Briefing Paper No.53, November 1, 1999.

2. The Cato paper raises a potpourri of reasons that Tennessee should not adopt a broad-based personal income tax. This analysis does not attempt to address every assertion made by Cato. We confine ourselves to what we see as their most condemning arguments.

3. 1996 is chosen as the latest year for this analysis because it is the latest year for which the US Bureau of the Census makes available local tax collection data for all 50 states.

4. This is estimated by taking the official revenue estimate for the Governor's plan and reducing it to 1996 levels. This amount is then added to what Tennessee state and local governments actually collected in 1996 and the national rankings recalculated.

5. Although we believe the 1980 to 1998 period produces more reliable results, we also did this analysis for the 1990 to 1998 period. This analysis provides further evidence of the lack of any strong and persistent relationship between the level of personal income tax and economic performance. When we examine the 1990 to 1998 period, we find that the ten-lowest income tax states had slightly higher growth (1.5% average annual growth) than the ten highest tax states (1.3% average annual growth). But, interestingly, both had higher growth rates than the national average (1.2%). Thus, states with high personal income tax burdens and states with low personal income tax burdens outperformed states in the middle. Using the mode of analysis endorsed by Cato--equating correlation with causation--this would suggest that states should either have very low or very high income taxes to maximize their economic performance, but should avoid having a middling level of taxes. This is, of course, a bizarre conclusion. In fact, what these data demonstrate is that (i) the level of personal income tax has little to do with the level of economic growth and (ii) one can "prove" almost any causal relationship between taxes and economic growth one wants if the time span of the analysis is chosen carefully.