State “Income Migration” Claims Are Deeply Flawed
October 20, 2014
Some proponents of state income tax cuts are making highly inaccurate claims about the impact of interstate migration patterns on states with relatively high income taxes based on a misleading reading of Internal Revenue Service data.
Those making these arguments claim that many of the people who leave states with relatively robust income taxes do so largely in order to pay little or no income tax in another state, and that they take their incomes with them when they move, harming the economies of the states they left. As a consequence, these “income migration” proponents claim, states with relatively high income taxes are suffering severe damage from the loss of income as “money walks” out of their states to lower-tax states.
The first part of this argument — that interstate differences in tax levels are a major explanation for interstate migration patterns — is not supported by the evidence, as we documented in an earlier paper. People rarely move to lower their state income taxes. Other factors, such as job opportunities, family considerations, climate, and housing costs, are much more decisive.
The second part of the argument — that states with relatively high income taxes are suffering severe economic damage because they are losing the incomes of people who migrate to other states — is also deeply flawed.
- Income migration analyses ignore that the vast majority of people can’t take their income with them to a new state because they work for someone else. When people leave a state, they usually also leave their job. The income they made in that job then typically goes to the person who gets that job next; it does not leave the state.
For example, consider a California sales representative who is transferred to Nevada. According to “income migration” proponents, California’s economy is weakened because the sales representative moved away and took her income with her. In reality, her income stayed with her employer and was then transferred to her replacement. California’s economy was not harmed.
- These analyses also ignore the income gains that accrue to other in-state small businesses when business owners move away. For example, if a New York doctor in private practice retires and moves to Florida, his or her patients ? and their payments ? will go to some other New York provider, increasing that provider’s income. Also, the owner of a successful business who leaves will often sell it to someone who will continue to operate it. “Income migration” analyses miss these realities because they focus only on people who move from one state to another, ignoring what happens to the incomes of people who don’t move.
- These analyses also do not trace what happens to the income of a person after he or she leaves a state. Income migration proponents effectively assume that the income of a person who leaves a state will stay the same after the move, even if the person doesn’t find a job in the new location or is moving there to retire. That assumption further skews their results. For example, when someone from New Jersey retires to Florida, income migration analyses claim that New Jersey’s economy lost income equal to the person’s pre-retirement salary, even though their income probably would have declined even if they had stayed in New Jersey.
- Other shortcomings in income migration analyses further exaggerate the size of interstate income flows. For example, some people leave a state but continue to work there. These people usually continue to contribute to the economy and tax revenue base of the state where they still work, even though their home is now elsewhere. Income migration analyses exaggerate the income lost to the worker’s old home state by effectively claiming that all of their income is lost to the economies of the states from which they moved.
To be sure, some income does automatically follow a person when he or she leaves a state — pensions, Social Security, and investment earnings, for example. But they represent a relatively small share of total taxable income — under one-fifth in most states. And, as with other forms of income, much of such income that is “lost” to a state when people move out is replaced by income “gained” when others move in.
Policymakers should focus their attention on the policy choices most likely to grow the incomes of their current and future residents, and not be distracted by misleading claims about income migration. The chief policy prescription that the income migration concept is used to justify — deep cuts in (or outright repeal of) state income taxes — would likely prove self-defeating, leading to deteriorating K-12 education, state universities, parks, roads, public safety, and other services that make states places where businesses want to invest and where the engineers, managers, and other personnel they need to hire want to live.
State Taxes Have Negligible Impact on Interstate Moves
Proponents of the income migration concept cite it — and the IRS data that allegedly measure it — in service of a broader claim: that interstate differences in tax levels drive large numbers of people to consciously “vote with their feet” and move from higher-tax states to lower-tax ones. Studies, however, don’t support this claim, as a recent CBPP papera explained:
- Relatively few Americans relocate from state to state, and a miniscule share of them report that they moved because of taxes. Only about 1.5 to 2 percent of U.S. residents relocate across state lines each year, and the rate seems to be declining. And of that 1.5 to 2 percent, the vast majority cite new, transferred, or lost jobs or family-related reasons (like needing to care for an ailing relative).
- People who move are nearly as likely to move from low-tax states to high-tax states as the reverse — in some cases, more likely. Between 1993 and 2011, for example, no-income-tax Florida lost households to 15 states, 11 of them with income taxes. Net in-migration of households to North Carolina, which had the highest income taxes of any Sunbelt state throughout this period, was more than double that of its no-income-tax neighbor Tennessee. Other migration patterns also confound the taxes-drive-migration thesis; for example, almost as many people moved to Arizona (which has an income tax) as to Texas (which doesn’t), even though Texas is a much larger state with many more jobs for a potential in-mover to fill.
- Migrants to no-income-tax states aren’t disproportionately high-income. For example, more than three times as many people moving from New York to Florida between 2008 and 2012 had incomes below $50,000 as above $100,000 — a ratio roughly in proportion to their shares of the population. If income taxes were a major reason that more people move from New York to Florida than vice versa, one would expect the people moving to Florida to be disproportionately high income.
- Climate is a major driver of interstate migration; people — especially retirees — continue to move from cold, snowy states to Sunbelt states regardless of the tax levels in either the origin or destination state. No-income-tax Florida and Texas had the highest net in-migration of any states from 1993-2011, but income-tax-levying Arizona and relatively high-income-tax North Carolina were close behind.
- Reductions in housing costs, not taxes, are what save families the most money when they move from states like New York and California to states like Texas and Florida. Taxes are much less consequential than housing costs for most families making many of the specific state-to-state moves often attributed to taxes. For example, a typical family with a $75,000 income selling its home in Los Angeles in 2010 and buying one in Houston would have saved more than two and a half times as much in mortgage payments as in state and local taxes. The same family moving from New York City to Miami would have saved more than three times as much in housing costs as in state and local taxes.
aMichael Mazerov, “State Taxes Have a Negligible Impact on Americans’ Interstate Moves,” Center on Budget and Policy Priorities, revised May 21, 2014, http://www.cbpp.org/cms/?fa=view&id=4141.
IRS Data on Interstate Migration
Proponents of the income migration thesis base their findings on a misleading reading of IRS data. Those data define an interstate move as having occurred when a tax return filed under a particular Social Security number in a specific state in one year is filed under the same Social Security number but in a different state the following year. These data are available for all possible combinations of state-to-state moves from the late 1980s to 2011.
The IRS data confirm what common sense would suggest: every year, every state sees some households moving in and others moving out. “Net migration” is the difference between the two and can be positive or negative. Many states, primarily northern “Frostbelt” states, have experienced fairly consistent net out-migration for several decades, while many “Sunbelt” states have experienced fairly steady net in-migration.
Beginning with interstate moves that occurred between 1992 and 1993, the IRS included in its migration data the total household income (adjusted gross income, or AGI) of the tax returns that moved between states. Proponents of the income migration concept simply aggregate the AGIs reported on all the migrating household tax returns — those moving into a given state and those moving out. Accordingly, a state experiencing net out-migration of households will generally show what proponents characterize as a net “loss of income due to migration.”
For example, the IRS data show that the 3.3 million households moving out of New York between 1993 and 2011 reported aggregate AGIs of approximately $170 billion, while the 2.3 million households moving into New York reported aggregate AGIs of approximately $103 billion. Income migration proponents cite these numbers to argue that New York “lost” approximately $67 billion in income as a result of the net out-migration of households and suffered lost jobs and other economic damage as a result, since that income was no longer available for consumer spending in the New York economy. These claims are deeply flawed, for a number of reasons discussed below.
Vast Majority of Income “Lost” Due to Out-Migration Never Leaves the State
The IRS data provide an incomplete — and thus seriously misleading — picture of whether states experiencing out-migration actually “lose income.” In reality, the income allegedly lost when people leave a state generally goes to two groups of people: those moving into the state and those already in the state who are entering the state’s labor force.
- People moving into a state replace the vast majority of the income of people moving out of it. The IRS migration data show that in 20 of the 26 states that experienced net out-migration of households between 1993 and 2011, at least 80 percent of the income supposedly lost through out-migration was replaced by income gained through in-migration.
- People entering a state’s labor force receive most of the rest of the income previously earned by people who left the state. If the jobs of people who leave a state are filled by people already in the state who have just reached working age, graduated from high school or college, were previously unemployed, or are otherwise re-entering the labor force, their income will not show up in the IRS migration data because they haven’t migrated. Likewise, if immigrants from foreign countries fill the jobs of out-movers, their incomes usually will not be counted either.
As Table 1 indicates, every state that experienced net out-migration in 1993-2011 except Michigan nevertheless saw an increase in the total number of people employed. For example, even though California lost 768,000 households due to migration, it both filled the jobs of those who left and generated almost 2.3 million new jobs. Since these states did not suffer a net loss of jobs, it is highly misleading to treat the wages of people who left the state as income “lost” to the state economy, as income migration proponents do. This particular misinterpretation of the IRS data has a substantial impact on the overall income migration figures often cited, because wages and salaries account for more than two-thirds of total AGI reported on federal tax returns.
All States Except Michigan Gained Jobs and Income from 1993-2011 Regardless of the Change in Households from Interstate Migration
|Net Change in Households
Due to Interstate Migration
(in thousands) 1993-2011
|New Non-Farm Jobs
|Average Annual Growth in Adjusted Gross Income of State Residents 1993-2011|
|District of Columbia||-25||56||5.1%|
|Source: Internal Revenue Service interstate migration and Statistics of Income databases, and Bureau of Labor Statistics|
Similarly, most of the business income of self-employed people and other small business owners isn’t lost when they leave a state. Departing small business owners often sell their businesses to people who will keep operating them. Even if they shut them down, different businesses already in the state will often fill the demand for the goods and services they provided. If a doctor or a plumber leaves a state, for example, some other doctor or plumber will pick up the leaver’s patient or client. Under scenarios like these, business owners or self-employed individuals remaining in the state will see a corresponding increase in their AGIs, but that increase will not show up in the IRS migration data.
As Table 1 shows, every state saw total AGI reported by state residents grow significantly between 1993 and 2011. In other words, none of the states that experienced net out-migration during that period actually “lost income.”
Other Limitations of IRS Migration Data Further Exaggerate Income Migration
Other significant limitations in the IRS migration data cause them to substantially exaggerate the amount of income that can reasonably be characterized as “lost due to (out) migration.” Those limitations concern four groups of people:
- People who are retiring. A large share of the people leaving relatively high-income-tax states like New York and New Jersey for no-income-tax states like Florida do so upon retirement. Their AGIs in the IRS migration data likely include, in part, salary or wages received before they quit their job. (The IRS data cannot determine the month in which the person moved, nor does the IRS trace movers to determine how much income they report in the first full year of residence in their new state.) Counting all of the income of retiring workers as “income lost due to migration” ignores the fact that much of it would have been “lost” even if they hadn’t left, since people’s AGI usually drops sharply when they retire.
- People who are laid off. When people leave a state after being laid off, a substantial share of their AGI reflected in the IRS migration data was likely earned before the layoff and would have been lost to the state economy even if they hadn’t left. In other words, the state economy did not “lose income” because they migrated; they migrated because they — and the state economy — lost income.
- People who leave the state but continue to work there. The IRS migration data treat people who change their state of residence as having migrated even if they continue to work in their old state. That income legally can be (and usually is) taxed by the state in which the work is performed. In turn, that tax revenue will be spent — and thereby provide employment — in that state’s economy. In addition, people who move away but still work in their old state likely patronize that state’s restaurants and stores to some extent, injecting additional demand (and sales tax revenue) into the state’s economy.
This shortcoming of the IRS data may exaggerate the “income loss due to migration” of New York and Massachusetts in particular, two states that income migration proponents often criticize for their income tax rates. Between 1993 and 2011, 492,000 households left New York for New Jersey, 164,000 left New York for Connecticut, and 151,000 left Massachusetts for New Hampshire. Many workers commute into New York City from New Jersey and Connecticut, as well as from southern New Hampshire to the Boston metro area. These commuters undoubtedly include many people who migrated into these three states, and the amount of income migration out of New York and Massachusetts is exaggerated because both states continue to tax the salaries of these non-residents and benefit from their workday spending.
- People who haven’t completely left their home state. There is no requirement that the address on a federal tax return be that of the taxpayer’s legal state of residence under state law. Accordingly, if a legal resident of New York (for example) who spends December through April at a second home in Florida decides to start listing the Florida address on his or her tax return, the IRS will deem this person — and all of his or her income — to have migrated to Florida in that year, even if the person continued to spend the same amount of the year (and the same share of his or her income) in New York.
Florida, which has no income tax, has an enormous “snowbird” population, estimated at 800,000 people in a 2007 study. Accordingly, a significant share of the alleged income migration to this state actually reflects the IRS’s equation of a change in tax filing address with a permanent move. The same is likely true to a lesser extent of some other areas in no-income tax states that are popular locations for second homes or winter rentals — for example, Jackson Hole, Wyoming, southern Nevada, and the south Texas coast.
Only a Small Share of Total Income Actually “Migrates”
Some types of income can reasonably be described as “migrating” when their recipients move, but they account for a relatively small share of total taxable income — under one-fifth in most states. These include things like pensions, Social Security benefits, and interest, dividends, and capital gains from bank accounts and other passive investments in stocks and bonds. In general, these types of income can only be taxed by the state in which the person receiving them resides. Accordingly, if a recipient of this income permanently leaves a state, the state’s economy and income tax base generally lose that income.
Taxable interest comprised just 1.2 percent of total AGI reported on federal tax returns in 2012, dividends 5.1 percent, and capital gains 6.8 percent. Taxable Social Security benefits comprised 2.5 percent of federal AGI and pensions and annuities an additional 6.7 percent, but a majority of states substantially or completely exclude these types of income from taxation.
Out-migration to Other States Remains a Concern
The fact that proponents of the income migration concept wildly exaggerate the economic harm to states from out-migration does not mean that policymakers should be indifferent to people leaving their states. Net out-migration can be a signal that a state’s economy is not providing sufficient opportunities for some residents to use their skills fully and to improve their standard of living. Out-migration can damage a state’s long-term growth prospects if the people leaving tend to be working-age adults with critical job skills, high levels of education, and strong entrepreneurial drive.
While some claim that high state taxes are driving these types of people to move out of state, the evidence doesn’t support this claim. Moreover, the incorrect assumption that state and local taxes are driving entrepreneurs, engineers, venture capitalists, and other skilled professionals to low-tax states leads some conservatives to advocate policy choices that ultimately could be self-defeating with respect to stemming out-migration and encouraging long-term growth.
Because income taxes provide such a large share of revenue in the states in which they are levied, even modest cuts would likely impair a state’s ability to provide high-quality education programs that train the engineers and business executives of tomorrow. Declining quality of local schools, state universities, roads, parks, mass transit, and other critical services also would likely discourage in-migration by the same kinds of highly skilled people that states need to attract. State policymakers thus would be wise to reject both the income migration concept and the misguided tax policy prescriptions that flow from it.
 For examples of this sort of claim see Travis H. Brown’s 2013 book How Money Walks and Arthur Laffer, et. al.’s 2014 book An Inquiry into the Nature and Causes of the Wealth of States. See also Richard Borean, “Monday Map: Migration of Personal Income,” Tax Foundation, August 19, 2013, http://taxfoundation.org/blog/monday-map-migration-personal-income, and Wendell Cox and E.J. McMahon, “Empire State Exodus: The Mass Migration of New Yorkers to Other States,” Empire Center for New York State Policy, October 2009, http://www.empirecenter.org/publications/empire-state-exodus/.
 Michael Mazerov, “State Taxes Have a Negligible Impact on Americans’ Interstate Moves,” Center on Budget and Policy Priorities, revised May 21, 2014; http://www.cbpp.org/cms/index.cfm?fa=view&id=4141. See the text box on page 3 for a summary of this report.
 Such a “migrated” tax return is a reasonable proxy for the interstate move of a single household, since most married couples file joint returns.
 Tax year 2009 and 2010 tax returns were usually filed in calendar years 2010 and 2011, respectively. Thus, a move that occurred between the filing of the 2009 and 2010 tax year returns appears in what the IRS refers to as its “2010-11 migration file.” The Adjusted Gross Income (AGI) data in the file is that shown on the return filed in the second year, tax year 2010 in this example.
 For moves that occurred between 1992 and 1993, 1993 and 1994, and 1994 and 1995, the income data included in the IRS migration database was actually “total money income” rather than adjusted gross income. The amounts generally should be of the same order of magnitude.
 The measured “migration of income” is also affected by the relative AGIs of the households moving in and out. It is possible for a state to experience net out-migration of households during a particular time interval but experience net in-migration of income during that period if the net loss of households is relatively small and the households moving into the state report higher average incomes than the households moving out. Nonetheless, whether a state shows “net in-migration of AGI” or “net out-migration of AGI” in the IRS data is determined in most cases by whether it is a net in- or out-migration state with respect to total households.
 As discussed below, it is impossible to determine the extent to which the incomes that appear in the IRS migration database were earned in the origin state, the destination state, or a combination of the two. The discussion in this paragraph assumes for the sake of argument that people who move to a different state manage to find jobs in their new location that pay comparable salaries.
 The other six states — Illinois, Louisiana, Michigan, New Jersey, New York, and Ohio — replaced between 60 and 80 percent of the income “lost” to out-migration through in-migration alone.
 The IRS migration database includes a line for the aggregate AGIs of returns that moved from a foreign location to the United States. However, that line only includes people who moved to the United States from a U.S. possession (such as Puerto Rico) in which they previously were required to file a U.S. tax return or U.S. citizens (such as military personnel) returning from abroad. It does not include someone who immigrates into the United States from a foreign country and takes a job here but who did not previously file a U.S. return because he or she had no U.S.-source income.
 Of course, the statement that California “filled the jobs of those who left” is not meant to be taken literally. State economies are dynamic, and specific, existing jobs are eliminated all the time. The point is that California created a sufficient number of jobs over this period to effectively replace the jobs of all employed people who left the state and to employ an additional 2.3 million persons.
 This is not to say that no income is ever lost to a state’s economy when the job of someone who leaves the state is filled by someone already in the state. It is possible, for example, for the job to be filled by someone with less experience than the incumbent, who accordingly will be paid a lower salary. On the other hand, it is also possible for the job to be filled by someone who is promoted into the job and receives the incumbent’s salary, with a chain of upward promotions being triggered and resulting in the same total salary payments within the business or organization.
 Calculated from Internal Revenue Service data for tax year 2012 (the most recent available) at http://www.irs.gov/file_source/pub/irs-soi/12in14ar.xls.
 The departure of a business can result in a loss of income to a state in certain circumstances. As an example, consider the owner of a small manufacturing company who decides to move to another state and closes the business. Assume that an out-of-state business manages to grab the customers of the liquidating firm. In that case, the state in which the closing business was located would likely lose the salary and taxable profit of the former owner from its income tax base because the owner of the out-of-state business probably lives out-of-state as well. But this scenario would likely be the exception rather than the rule because most small businesses supply goods and services to a fairly localized market. See Erik Hurst and Benjamin Wild Pugsley, “What Do Small Businesses Do?” Brookings Papers on Economic Activity, August 2011, http://www.brookings.edu/~/media/Files/Programs/ES/BPEA/2011_fall_bpea_papers/2011_fall_bpea_conference_hurst.pdf.
 The discussion in this paragraph assumes that the person who retired from a job and then left the state is not replaced in that job by someone already in the state or by an immigrant from another country. If he or she is replaced by such a person, then the discussion of the preceding section in this report is applicable and there may be no “loss of income” to that state whatsoever.
 To see this, consider someone who lived and worked in New York in 2009, filed his 2009 tax return reporting an annual salary of $50,000 on April 15, 2010, retired on December 31, 2010 and saw his 2011 AGI drop to $25,000, moved to Florida on February 15, 2011, and filed his 2010 return on April 15, 2011. This person would show up in the 2010-2011 IRS migration file (the most recent the IRS has made available) with an AGI of $50,000. “Income migration” proponents would deem all $50,000 to be “income lost due to migration,” despite the fact that if the person had chosen to stay in New York, New York would have lost $25,000 in AGI anyway because the person’s income dropped.
 According to a 2012 Social Security Administration study, the total household income of the median retiree household falls by almost one-fourth in the first or second year after retirement. The total household income of a retiree household in the bottom quarter of the income distribution falls by more than half in the first or second year. (Patrick J. Purcell, “Income Replacement Ratios in the Health and Retirement Study,” Social Security Bulletin, 2012, Table 1.) It seems likely that adjusted gross income reported on tax returns falls even further, because a considerable share of the Social Security income that a person usually begins receiving immediately upon retirement is tax-exempt.
 There are two exceptions. The District of Columbia is barred by a federal law from taxing the salaries of people who work there but reside elsewhere. Some states have income tax “reciprocity agreements” with other states, under which they mutually agree not to tax the salaries of residents of those states — ceding to the state of residence the sole authority to tax the income.
 Stanley K. Smith and Mark House, “Temporary Migration: A Case Study of Florida,” Population Research and Policy Review, 2007.
 Calculated from Internal Revenue Service data for tax year 2012 (the most recent available) at http://www.irs.gov/file_source/pub/irs-soi/12in14ar.xls.
 Mazerov, 2014.