Millionaires Pay to Stay: The Millionaire Tax Migration Myth

Summary

• Despite anecdotal claims and reference to flawed data, there is no defensible proof of high-income residents leaving Connecticut due to taxes.
• Recent research shows that the 1991 income tax did not affect migration into or out of Connecticut.
• National studies show tax increases minimally impact migration, with increased revenue outweighing losses from a handful of departures.
• High-income individuals stay in areas offering resources, including public goods, quality schools, proximity to major cities, and an attractive environment.

Introduction

The wealthy and those with high incomes are more capable of moving than other populations; nonetheless, they remain rooted and concentrated in high-cost-of-living enclaves. Connecticut is a case in point. Connecticut continues to rank at, or near the top, of all states across most metrics of wealth and income:

• One percent of all adults earn more than $819,630. Only DC has a higher threshold for being in the top 1%.¹
• 3^rd highest share of households with at least $1,000,000 in investable assets.²
• 2^nd in per capita income – behind only Massachusetts.³
• The number of Connecticut residents on the 2022 Forbes 400 list of wealthiest Americans (6) is disproportionately large relative to the state’s population and larger than it has been several times in the previous decades (e.g., 4 in 1982 and just 2 in 2010).

Nonetheless, there is a powerful narrative, built on faulty data and anecdotes, that wealthy and high-income residents are fleeing the state due to the cost of living, taxes, and public finances.

The Misuse of Migration Data

Questionable data and methods drive the “millionaire with a suitcase” political discourse. For example, moving van data frequently show a net outflow of population from Connecticut, yet this data does not represent the population or movers.  Furthermore, these sorts of data are contradicted by representative data gathered by the US Census, which shows a small net out-migration from Connecticut that is, statistically speaking, effectively zero and very similar to nearby states.

The media also frequently site surveys that report that a large portion of the population is considering a move out of state, but the connection between stated migration intentions and migration is not robust. For example, the following figure shows that while 13% of adults are “definitely” planning to move, only 14% of this 13% move! The disconnect between moving intentions and actual migration also applies to second-hand reports that the handful of billionaires in the state would move if marginal tax rates were increased. In this regard, “talk is cheap”; people across the income spectrum frequently consider moving, but less than 2% of Americans move from one state to another – including billionaires. As discussed in greater detail below, the fact remains that billionaires are geographically rooted in what are already high-tax and high-cost-of-living states. 

The narrative of “income migration” relies heavily on flawed Survey of Income (SOI) data from the IRS.⁴The SOI details the number of tax returns, dependents, and total adjusted gross income (AGI) by state with further detail for those who moved into or out of state over the previous year. This data is further broken down by income level (the top level being $200,000) and age of the primary taxpayer. Estimates of state “income migration” are then made by comparing the income earned over the previous year for migrants relative to non-migrants.

The concept of “income migration” itself is suspect.⁵ First, when discussing income migration, it is essential to note that there is no such thing as income migration for most workers. While remote work is increasing, most workers remain tied to a physical place of work or must live near a home office. They must change employers or jobs within the same organization when they move. Since their job does not move with them, the income associated with their previous job also does not. For example, an individual who lived in Connecticut for the first six months of a year and earned $50,000 during that time, then moved to Florida to retire and had no earnings after moving. The IRS SOI will incorrectly treat the $50,000 earned in Connecticut as earned in Florida. Consequently, it will appear as if $50,000 of income “migrated” from Connecticut to Florida when, in fact, this retired person earned nothing in Florida. If they had stayed in Connecticut, their income would have been zero. Second, while the IRS SOI covers 97% of the US population, it does not represent that population. Most significantly, the SOI excludes recent immigrants. Immigrants are a significant source of income for Connecticut since immigrants to Connecticut are, on average, relatively well-educated and concentrated in high-wage occupations.⁶

These two errors – counting the income of recent retiree migrants as if they had been working in their new state the whole year and excluding the income of recent immigrants – understate income in-migration and overstate income out-migration for Connecticut. There is no doubt that IRS SOI estimates of a net outflow of income from Connecticut due to migration are inflated. Unfortunately, there is no simple way to address these flaws, nor can solutions be found in other publicly available data, such as the US Census American Community Survey, which shares the same limitations as the IRS SOI data.

Our recommendation is to ignore IRS SOI data, non-representative data such as moving van data, and anecdotes such as billionaire threats to move for public policy in Connecticut, to rely on peer-reviewed studies that have used better data and methods, and to develop the capacity in the executive branch to analyze individual-level administrative data on migration by income levels.

What is the Actual Effect of Income Taxes on the Migration of High Earners?

Peer-reviewed research concludes that the migration response among high-income earners to an increase in marginal tax rates is minimal, so small that the gain in total tax revenues far outweighs the loss of what would be a handful of residents. The most reliable studies take a longitudinal approach by comparing migration into and out of a state before and after a change in tax policy. A good example is a recent study comparing migration into and out of Connecticut before and after the start of Connecticut’s income tax in 1991.⁷ In this case, the implementation of the income tax caused a decline in net migration by just 580 people out of Connecticut’s total population of about 3.3 million people in the early 1990s. The 1991 income tax effectively did not affect migration into or out of Connecticut.

Cornell Sociologist Cristobal Young, along with colleagues at the US Department of Treasury and Stanford University, has taken a similar approach but focused explicitly on high-income individuals (~$1m of income a year) and how migration into and out of states was affected by increases in top marginal tax rates in California which added a 1% surcharge on income over $1m in 2005⁸, New Jersey which increased the maximum marginal tax rate from 6.37% to 8.97% in 2004 on income over $500,000⁹, and nationally using individual-level administrative data.¹⁰

Young and colleagues analyzed tax return data from every million-dollar income earner in the United States. The data includes 3.7 million top-earning individuals who collectively filed more than 45 million tax returns over more than a dozen years – showing where millionaires live and where they move to:¹¹

• Income taxes have only a tiny impact on millionaires’ moves. If the average state raised its income tax rate on millionaires by one percentage point and all other states kept their rates constant, 12 fewer millionaires would move into that state, and 11 more would move out of that state. That’s 23 moves compared to 9,000 households earning more than $1m in the average state. The study estimates that only 2.2 percent of all millionaire interstate moves have any income tax motivations whatsoever. This migration response to an increase in marginal tax rates is so tiny that the gain in total tax revenues far outweighs the loss of what would be less than two dozen residents.
• Florida’s the only state for which there’s any evidence that low taxes even slightly attract millionaires.  There’s no effect for other states without income taxes, such as Texas, Tennessee, and New Hampshire. That this effect is absent from other states without an income tax suggests that the Florida effect may be more about climate and other factors than the lack of an income tax.
• Millionaire business owners are even less likely than other millionaires to move in response to interstate differences in income taxes.  Presumably, their incomes are even more closely tied to the networks of customers, employees, investors, and suppliers they’ve built over the years. 
• Millionaires also tend to move to states with high residential land prices. This is a significant result, as it shows that the cost of living is not a significant factor in the location choices of millionaires. 

Young concludes that “. . .elites are embedded in the regions where they achieve success, and they have limited interest in moving to procure tax advantages” because “. . . most millionaires are at their peak years of earnings and are drawing on long personal investments in a career or business line from which they cannot easily migrate away. Income-earning capacity derives not just from individual talent and human capital (which is movable) but also from place-based social capital— social and business connections to colleagues, collaborators, funders, and co-founders.”¹²

Migration and the Life Course

That taxes and other economic factors are not connected to the migration of wealthy or high-income populations – or any other people, for that matter – may be counterintuitive. While economics certainly matter in moving – people both need a source of income and are likewise affected by the cost of living wherever they live – the modern view is that migration is a complex decision focused on the tradeoffs between stage in the life course, lifestyle, quality of life, and economic factors:

. . . US internal migrants find themselves in the wrong places at the wrong times in their lives. At some ages, large agglomerations hold key advantages; at others, sparsely settled environments are preferable. Increasingly, people adjust their location reflective of their changing circumstances as they experience and anticipate the next stages of life.¹³

The role of the life course and quality of life in shaping migration into and out of Connecticut is seen in US Census estimates of reasons for moving into and out of Connecticut.

Migration is never one way; the same thing that may cause some people to move out may cause others to move in. The following figure shows 29,095 people left Connecticut annually between 2015 and 2019 for “economic reasons.”¹⁴At the same time, 22,807 people moved to Connecticut for the same reasons. Hence, “economic reasons” only caused an annual net loss of 6,289 people. It contradicts the narrative of an economic exodus from Connecticut and shows that the economic situation in Connecticut is a positive attribute for many people.

The net loss of 6,289 people per year for economic reasons is more than compensated by the net gain of 11,628 people for family¹⁵ (9,962) and housing¹⁶ (1,666) reasons. This housing effect contradicts rhetoric about Connecticut’s rising cost of living and declining quality of life. Access to mountains and sea, four seasons, proximity to New York and Boston, the New England landscape, and high-quality schools and public services matter to many people. This fact should not be discounted in making public policy. Many benefits of living attracting people to Connecticut are public goods and services supported by tax revenue.

The net loss of population for college and retirement are important examples of how migration is a tradeoff between stages in the life course, lifestyle, quality of life, and cost of living. Connecticut is an attractive destination at various times in the life course; at other periods, it is not. Positive net migration for family and housing reasons reflects the attraction of Connecticut for working-age adults with families. Negative net migration for retirees and college students reflects that Connecticut is not a desirable state to retire in and that it has relatively few college opportunities relative to its size and is near education opportunities in other states.¹⁷

Summary and Recommendations

Despite frequent anecdotal statements to the contrary, there is no evidence of a tax-induced flight of high-income residents from Connecticut. Narratives in this regard are built on deeply flawed IRS data that should be disregarded in forming public policy. A study of the impact of the 1991 income tax showed that it may have only reduced net migration to Connecticut by less than 600 people. Several well-regarded studies have also demonstrated that taxes do not play much of a role in the location decisions of people with high incomes. Instead, the wealthy are rooted in areas that provide resources – many of which are public goods and services supported by public finance – which enhances their ability to earn and preserve their wealth. Connecticut is well positioned by quality schools and housing stock, proximity to New York and Boston, and an attractive environment for working-age adults and their families. Research indicates that even moderately significant increases in the top marginal tax rates may only induce a dozen high-income residents to leave and that the gain in tax revenue far outweighs the loss due to the migration of a few people.

Notes

¹ Source: Internal Revenue Service, Statistics of Income Division, December 2022.

² https://graphics.wsj.com/table/Millionaires_0207

³ https://www.bea.gov/data/income-saving/personal-income-by-state

⁴ For more detail see Pierce K. SOI migration data: A new approach: Methodological improvements for SOI’s United States population migration data, calendar years 2011-2012. Statistics of Income. SOI Bulletin. 2015 Jun 22;35(1).

⁵ For a more detailed discussion of the flaws with IRS SOI data for income migration research see Mazerov M. State Income Migration Claims are Deeply Flawed. Center on Budget and Policy Priorities. 2014 Oct 20.

⁶ Cooke TJ. What Exodus? High-Skill Workers Drawn to CT’s Quality of Life. Hartford Courant. 2016.

⁷ Afonso WB. The effect of a state income tax on migration: the example of Connecticut. Journal of Public Policy. 2018 Mar;38(1):113-40.

⁸ Varner C, Young C. Millionaire migration in California: The impact of top tax rates. National Tax Journal. 2012;64(2):255-84.

⁹ Young C, Varner C. Millionaire migration and state taxation of top incomes: Evidence from a natural experiment. National Tax Journal. 2011 Jun 1;64(2):255-83.

^10 Young C, Varner C, Lurie IZ, Prisinzano R. Millionaire migration and taxation of the elite: Evidence from administrative data. American Sociological Review. 2016 Jun;81(3):421-46. & Young C. The myth of millionaire tax flight: How place still matters for the rich. Stanford University Press; 2017 Oct 31.

¹¹ Mazerov, M. Millionaire Tax Flight Myth Debunked — Again. Center on Budget and Policy Priorities. 2016 May 26.

¹² Young C, Varner C, Lurie IZ, Prisinzano R. Millionaire migration and taxation of the elite: Evidence from administrative data. American Sociological Review. 2016 Jun;81(3):421-46.

¹³ Plane DA, Henrie CJ, Perry MJ. Migration up and down the urban hierarchy and across the life course. Proceedings of the National Academy of Sciences. 2005 Oct 25;102(43):15313-8.

¹⁴ “Economic Reasons” include new job, job transfer, to look for work or lost job, for easier commute.

¹⁵ “Family Reasons” include change in marital status, to establish own household, and other family reasons.

¹⁶ “Housing Reasons” include want to own home, not rent, wanted new or better housing, wanted better neighborhood, housing cost, and other housing reason.

¹⁷ Cooke TJ, Boyle P. The migration of high school graduates to college. Educational Evaluation and Policy Analysis. 2011 Jun;33(2):202-13.