Taxes and Interstate Migration
New Jersey Department of the Treasury
Office of the Chief Economist
Office of Revenue and Economic Analysis
The New Jersey Tax Flight Controversy
• In 2004 New Jersey sharply increased its highest tax
– Marginal rate on income over $500,000 was boosted from
6.37% to 8.97%.
• Hughes and Seneca (2007) first documented increased
outmigration from New Jersey.
• Havens (2010) documented large unfavorable shift in
wealth movements from New Jersey than from
• Young and Varner (2011) document outmigration rise at
upper end, but
– Regard loss of tax base as very small relative to revenue gain
– Argue that tax increase had little or nothing to do with
• Assert high home prices in New Jersey may have spurred
outmigration of the well-to-do.
Net Migration from New Jersey
Thousands of Returns (Exemptions)
1000 Net income
1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009
Taking another look at the issue
• Young and Varner
▫ Only considered taxes and outmigration, not taxes and
inmigration (did not consider individuals who did not
move to New Jersey as a result of the tax change)
▫ Only looked at New Jersey
▫ Assumed that 2004 tax hike had no implications on
migration patterns of individuals not immediately
Essentially looked at direct effect on individual migration
decisions, and did not take into account expectational
and other indirect effects on the rest of the population.
▫ Did not systematically look at the influence of housing
Study of Interstate Migration
• Started with the IRS migration dataset.
• This series presents annual state-to-state
movements of taxpayers and adjusted gross income.
• Assumed migrations are influenced by
• Differences in unemployment rates (people likely
move from high-unemployment rate to low-
unemployment rate states)
• Differences in home prices (people likely move from
high-home price to low home-price states)
• Differences in state income taxes (people likely move
from high-tax to low-tax states, especially to those
with no state income tax)
• Statistically significant effects of income tax rate
differentials on annual migration.
• Housing price differentials also significant.
• People and income are attracted to zero income-
▫ Hard to distinguish from tax differential effect.
Implication of Results for New Jersey
• Suppose NJ boosted its income tax schedule by one
percentage point across the board (and no other state changes
– This would be a very large tax hike
• More than twice as big as the 2004 increase.
• Static revenue increase would be nearly $2 ½ billion a year.
– One model (without zero tax attractor or state fixed effects)
suggests NJ would see increased annual net outflows of
• About 4,200 taxpayers (likely about twice as many persons)
• About $530 million in AGI
• Average income of affected taxpayers equal to about $125,000.
• Suggests higher-income taxpayers are most sensitive to changes
associated with tax increase.
• Model with state fixed effects has comparable income loss, but higher
loss of taxpayers.
• Similar calculations can be made for any other state.
Calculated Effects of 2004
• If NJ effective rates had remained at 2003 levels (movement since
reflects bracket creep as well as changes in the law), then by 2009
▫ NJ would have had roughly 20,000 more taxpayers.
▫ Adjusted gross income would be approximately $2.4 billion
higher, generating more than $125 million in state income tax.
Suggests that revenue boost from 2004 increase has been
partly reversed by induced out-migration of the base (as well as
Some Limitations of the Results
• Model does not estimate migration effects on
▫ Business tax revenue
▫ Property tax revenue
▫ Sales tax revenue
• The cumulative tax losses to the state could
potentially be larger than the model’s predictions
• Result not necessarily causal
▫ A weak state economy could spur both outmigration
and tax increases.
▫ State fixed effects, unemployment rate would help
control for this.
▫ Still suggests that widened tax differential is at least a
symptom to be concerned with.
Effect of House Prices
• A $10,000 increase in New Jersey home prices
would be associated with a loss of 1200
taxpayers a year and $66 million in adjusted
▫ Average income of lost taxpayers is in one
specification smaller than those affected by a rate
increase (about $50,000 a year vs. about
$125,000 a year).
Preliminary Follow-Up Results
• Estimation of model for New Jersey alone shows
▫ Very similar tax effects as that derived from
▫ Stronger home price effects.
• Little evidence for effect of longer lags on tax
▫ Suggests that the effect works through the level of
interstate differential, not its change.
▫ Suggests that the effect does not materially grow
or fade over time.
• IRS data does not allow separation of “high-
income” and “low-income” taxpayers.
▫ Perhaps county migration data can help capture
▫ In any event, effects of tax changes on migration
can permeate throughout population.
• Effect of top marginal rate (purer policy
variable) is similar to that of weighted-average
• In July 2011 subscribers to the NJ Division of Taxation’s Tax E-
Notes were asked about migration patterns.
▫ Most likely, these are accountants or other financial advisers.
• Approximately 200 responded.
▫ More than half said that clients had recently left, or expressed
interest in leaving the state.
▫ Given the option to pick 5 out of a list of 11 potential reasons for
outmigration, the top ones were
State income taxes (85.4%)
State property taxes (77.7%)
State estate taxes (67.0%)
Housing costs (43.7%)
• Analysis of IRS data suggests some connection between interstate
tax differentials and migration patterns.
▫ High taxes seem to be associated with loss of higher-income
▫ High home prices seem to be associated with loss of middle-
• Application to New Jersey suggests cumulative outmigration may
have offset some portion of the 2004 tax increase.
▫ Given New York’s 2012 reduction in income taxes, sensible to be
aware of the potential impact of New Jersey taxes on migration.
Survey responses suggest that New Jersey’s taxes may be on
many people’s minds when making location decisions.