For full notes, graphs and charts please download the PDF file
When Americans think they can find a better job and higher quality of life somewhere else, they move. Migration between the states is the ultimate expression of “voting with your feet.” Some states have growing populations due to in-migration, while others are losing residents to other states. Connecticut is one of the states that is losing population.
This study looks at Connecticut migration trends and how peoples’ decisions to move out of and into the state affect tax revenue. Key findings include:
- Connecticut lost a net of 325,526 residents to other states between 1991 and 2008, or about one in ten residents.
- The top states that people from Connecticut move to are Florida, North Carolina, Georgia, Virginia, and South Carolina.
- The top states that people move into Connecticut from are New York, New Jersey, Rhode Island, Illinois, and Nebraska.
- The total net income leaving the state was nearly $5 billion between 1995-2006. Had this income stayed in Connecticut, state and local governments would have collected an estimated $566,520,000 in additional tax revenue.
- Of course, when someone leaves, state and local governments don’t just lose income and taxes for one year, but rather for all future years as well. Compounding these figures over the twelve years assessed in this study, the state has lost $31.2 billion in net income and $3.7 billion in state and local tax revenue due to out-migration.
- People move to states where the weather is warmer, taxes are lower, union membership is lower, population density is lower, and the cost of housing is lower.
- The number one destination state for former Connecticut residents is Florida, a state with no income tax and no inheritance tax.
- An August, 2009 poll conducted by The Yankee Institute found that 45 percent of state residents have considered moving out of Connecticut due to high taxes.
MEASURING CONNECTICUT'S OUT-MIGRATION PROBLEM
Population change occurs through births, deaths, migration between the states, and international immigration and emigration. The most comprehensive data available on domestic migration comes from the U.S. Department of Commerce’s Census Bureau. Charts 1 and 2 and Table 1 show Connecticut’s net domestic migration between 1991 and 2008. In every year except one, Connecticut has been an out-migrant state with a net total of approximately 325,536 residents leaving the state during these years.
While the rate of population loss declined from 1992-2002, the trend was still decidedly net-negative in terms of aggregate population loss. The year with the greatest out-migration was 1992 with 40,282 leaving the state. The sole year of in-migration was 2003 by a mere 47 people.
While the Census Bureau data is comprehensive, data from the Internal Revenue Service provides a more detailed picture of migrants. The number of tax returns is a good proxy for the
number of households; the number of exemptions is a good 1993 proxy for the number of people in the household; and reported Adjusted Gross Income (AGI) is a good proxy for household 1995 income.
Table 2 shows the aggregate migration data from the IRS for 1997. Connecticut. In 2006 (the most recent data available), 43,270 taxpayers left the state while 34,293 taxpayers entered the 1999 state—for a net loss of 8,977 taxpayers. Overall, Connecticut lost 14,599 exemptions (people) and $748,731,000 in AGI.
From 1995 to 2006, Connecticut lost 84,963 taxpayers, 124,324 exemptions, and $4,866,235,000 in AGI (nominal dollars). Both taxpayers and exemptions have been negative in every year of this time-period. AGI was positive in only one year, 2002, by a mere $37,975,000.
Of course, when someone moves out, the state doesn’t just lose their income and for one year, but for all future years as well. Compounding the net AGI loss over the twelve years above, the total comes to $31.2 billion over time (not adjusted for inflation or present value).
WHERE THE OUT-MIGRANTS ARE GOING
The IRS data also provides migrant data by state showing where out-migrants are going and where in-migrants are coming from. Tables 3a and 3b rank the net migration totals for the years 1995 to 2006 for exemptions (people) and AGI, respectively.
As shown in Table 3a, the top exemption out-migrant states are Florida (73,560), North Carolina (21,437), Georgia (15,265), Virginia (7,322) and South Carolina (4,358).
The top exemption in-migrant states are New York (81,173), New Jersey (6,001), Rhode Island (1,637), Illinois (944) and Nebraska (298). Overall, Connecticut loses exemptions to 43 states while gaining exemptions from only seven states.
WHY POLICY MAKERS SHOULD WORRY ABOUT OUT-MIGRATION
These out-migrants take their incomes and purchasing power with them. As shown in Table 4, between 1995 and 2006, the total amount of AGI leaving the state was at least $4,866,235,000 (nominal dollars). The greatest out-flow of AGI was in 2006 at $748,731,000. In only one year, 2002, was there in-migration of AGI by a mere $37,975,000.
Overall, had this income stayed in Connecticut, state and local governments would have collected an estimated $566,520,000 in additional revenue over this time-period. This not only includes higher income taxes, but also higher sales taxes and property taxes.
Of course, when someone leaves, the lost revenue to state and local government isn’t limited to the year the person left. It’s lost for every year moving forward, too. Compounding the tax losses over the twelve years considered above, the total tax losses come to roughly $3.7 billion dollars (not adjusted for inflation or present value).
The out-migration of exemptions suggests Connecticut is also losing children. The long term income and tax loss of these future workers to Connecticut’s economy is incalculable, it is not hard to imagine them dwarfing the numbers presented previously.
As shown in Table 3b, the top AGI out-migrant states are Florida ($3,501,002,000), North Carolina ($720,395,000), Massachusetts ($704,996,000), Virginia ($583,698,000) and California ($583,698,000).
The top AGI in-migrant states are New York ($4,109,543,000), New Jersey ($426,730,000), Illinois ($185,511,000), Michigan ($28,313,000) and Ohio ($20,766,000). Overall, Connecticut loses AGI to 42 states while gaining AGI from only eight states.
REVERSING OUT MIGRATION
Reversing Connecticut’s out-migration problem requires an understanding of why residents are leaving. As shown in Table 5, one way to do this is by comparing various characteristics of Connecticut versus the destination states. In economic terms, out-migrants are expressing their “revealed preferences” by moving to another state more in-line with their preferences and values. We compare Connecticut to these destination states via six common variables used in migration studies: state and local tax burdens; income tax burdens; union membership; population density; cost-of- housing; and average temperature.
State and Local Tax Burden: This variable measures total state and local taxes collected as a percent of personal income as averaged over the 1995 to 2006 time-period. Connecticut’s average tax burden was 11.41 percent. Taxpayers left for states where tax burdens were 10.13 percent lower (10.26 percent), while exemptions were 10.56 percent lower (10.21 percent) and AGI was 10.54 percent lower (10.21 percent). Overall, exemptions were most sensitive to state and local tax burdens.
Income Tax Burden: This variable measures total state and local income taxes collected as a percent of personal income as averaged over the 1995 to 2006 time-period. Connecticut’s average income tax burden was 2.81 percent. Taxpayers left for states where income tax burdens were a whopping 37.51 percent lower (1.75 percent), while exemptions were 42.12 percent lower (1.62 percent) and AGI was 44.21 percent lower (1.57 percent). Overall, AGI was the most sensitive to state and local income tax burdens. It is worth noting that the number one destination state for former Connecticut residents is Florida, a state with no income tax and no inheritance tax.
Union Membership: This variable measures the percent of the state’s employed labor force who are members of a union as averaged over the 1995 to 2006 time-period. Connecticut’s average union membership was 16.6 percent. Taxpayers left for states where union membership was 47 percent lower (8.8 percent), while exemptions were 50.63 percent lower (8.2 percent) and AGI was 60.34 percent lower (8.5 percent). Overall, exemptions were most sensitive to union membership.
Population Density: This variable measures total population divided by land area and is as averaged over the 1995 to 2006 time-period. Connecticut’s population density was 704.5 people per square mile. Taxpayers left for states where the population density was 50.81 percent lower (346.6 people per square mile), while exemptions were 60.81 percent lower (276.1 people per square mile) and AGI was 60.34 percent lower (279.4 people per square mile). Overall, exemptions were most sensitive to population density.
Cost of Housing: This variable measures the median cost of housing as reported from the 2000 Census. Connecticut’s median cost of housing was $166,900. Taxpayers left for states where the cost-of-housing was 24.95 percent lower ($125,253), while exemptions were 29.77 percent lower ($117,212) and AGI was 27.01 percent lower ($121,813). Overall, exemptions were most sensitive to cost-of-housing.
Average Temperature: This variable measures the annual average of the daily mean temperature. Connecticut’s temperature by this measure was 50.2 degrees Fahrenheit. Taxpayers left for states where temperatures were 12.1 degrees warmer (62.3 degrees), while exemptions were 12.7 degrees warmer (62.9 degrees) and AGI was 12.4 degrees higher (62.6 degrees). Overall, exemptions were most sensitive to temperature.
People are most inclined to move where it is warmer, taxes are lower (especially income taxes), union membership is lower, population density is lower and the cost of housing is lower. However, there is one notable exception where AGI is the most sensitive to the income tax burden. As such, Connecticut should work toward reducing the state and local tax burden via reductions in the income tax which would encourage both people and income to stay in Connecticut or move into the state.
However, not all of these variables can be changed by policymakers — weather cannot be changed through legislative action. Other variables can be changed by policymakers on an annual basis — tax burdens can be reduced. Most variables can only be influenced by legislation and even then will take years to establish measurable change such as union membership, population density, and cost of housing.
While identifying specific remedies for each of these issues is beyond the scope of this study, without action, out- migration will continue to reduce the ability of both the private and public sector to ensure Connecticut’s economy remains the wealthiest and most dynamic in the country.
The IRS data used in this study is derived from the calendar year (CY) 1995 to 2005 State-to-State Migration Data Set (SSMD) that is published annually by the Statistics of Income Division (SOI) of the Internal Revenue Service (IRS). To qualify for inclusion in the SSMD, the IRS compares address information supplied on the taxpayer’s tax form between two years. If the address is different in Year 2 from Year 1, then the taxpayer is classified as a “migrant;” otherwise, the taxpayer is classified as a “non-migrant.”
The IRS is required by law to ensure that its data products do not reveal the identity of any taxpayer. In the SSMD, the data suppression affects its “data fidelity,” to borrow a musical term. In music, the term “recording fidelity” describes a recording’s ability to capture as much of the total sound as possible, i.e., the lower the recording fidelity, then the lower the recorded sound quality.
Analogous to this is the data fidelity within the SSMD. For example, if only a single taxpayer moved from state A to state B, it would be relatively simple (for those with the know-how) to identify that taxpayer. Therefore, the IRS lumps all such taxpayers into a residual category in order to prevent identification. As a result, the exact movement of all taxpayers is unknown. The percentage that is shown represents the SSMD’s data fidelity which is higher in the state- level migration data than the county-level migration data.
The major strength of the SSMD is that it is based on actual data — not a survey — that is enforced with criminal penalties. This makes the SSMD especially reliable as a data source given people’s incentive to be truthful in their data reporting. In addition, the SSMD includes reported AGI which allows researchers to not only track population flows, but also income flows.
On the other hand, the major weakness of the SSMD is that it excludes certain segments of the population. First, it excludes low-income groups such as students, welfare-recipients and the elderly because the standard deduction and exemptions are greater than their income. Second, it under-represents the very wealthy because they are more likely to request a filing extension and miss the late September cut-off for inclusion in the data-set. Finally, it may miss taxpayers who have changed filing status—especially from “married filing joint” to “married filing separately.”