By
Moshe Adler
Dept. of Urban Planning
Columbia University
212 873 6803
When deciding how to
close states' and cities' budget gaps one of the questions that must be
answered is whether a moderate tax increase by a city or a state--when the
alternative is cuts in government services--would cause a gain or a loss of
jobs.
The evidence from
empirical studies subjected to peer review is that when tax increases prevent
service cuts they also save jobs. But
the Beacon Hill Institute at Suffolk University in Boston has created its own
unpublished model of the relationship between employment and taxes and its results
differ sharply from those of the peer-reviewed studies. The Manhattan Institute
relies on the Beacon Hill model to support its claim that increasing taxes will
cost jobs, no matter what the circumstances are. Below is a brief review of the
literature and of the Beacon Hill model.
The research on the
effect of taxes on the economy experienced a great leap forward with the publication in 1985 of the article “The
Effect of State and Local Taxes on
Economic Growth: A Time Series-Cross Section Approach” by University of California
economist Jay Helms.[1] Instead of
treating government as a black hole into which funds disappear,
Helms included in his study both the
taxes paid and the services
financed by these taxes. Helms discovered that expenditures on health,
highways, schools and higher education cause growth in a state's
personal income. The only expenditure that caused a decrease
in a state's personal income in Helms'
study was welfare; interestingly, this is also the only result that did not withstand the test of time, because it
was unique to a particular
time-period. Carroll and
Wasylenko (1994) discovered that the effect of
welfare payments on employment is mostly limited to employment in manufacturing and that the effect on
manufacturing was present in the 1970s
but not in the 1980s.[2],
[3]
(Helms' data are for the years 1965-1979).
Helms' study remains
the last word in estimating the empirical effect that taxes have on economic performance overall.[4] Other researchers examined more limited issues that are relevant to the effect of
taxes.
--Bartik (1989) found
that increased spending on local schools and fire protection increases the
creation of small businesses.[5]
--Munnell (1990) found that tax increases
that finance improvements to highways,
sewage systems and other infrastructure increase the growth rate of private employment.[6]
--Bartik (1996) found that increased spending
on higher education and health,
financed by property tax increases, stimulate state manufacturing output.[7] On the other hand, increases in spending
on roads, financed by non-property tax
increase, reduce manufacturing
employment in the long run. According to Bartik's findings all other tax
increases had little impact on economic
performance.
The Beacon Hill
Institute applied its model, called State Tax Analysis Modeling Program, or STAMP, to New York
State and City. Economists Andrew
Haughwout, Robert Inman, Steven Craig
and Thomas Luce, researchers at the National Bureau of Economic Research, also constructed a model of
employment and taxes in New York
City. This model is strikingly similar
to the Beacon Hill study and achieves a similar result.
As we show in an
article in a State Tax Notes article[8]
and in an as yet unpublished response
to a rebuttal by the Beacon Hill Institute,[9]
the Beacon Hill studies and the study
by the NBER researchers are fatally
flawed. Briefly, the main problems with
these studies are:
1. The
studies account for taxes in only one jurisdiction. Thus, in the case of New York City, the studies contain
information only about the city itself. As the data that these studies use show, between 1989 and 1992 private sector employment in NYC fell
by 10% at the same time that the top
personal income tax rate increased by 31%.
But these studies completely
ignore the fact that New Jersey raised its top rate by a whopping 90% and Connecticut instituted its
first state income tax during the same period. New Jersey and Connecticut are often
described as competing with New York
City for jobs, and their inclusion in the
study is therefore mandatory. It
should be noted that Helms' study contains tax data for all the relevant jurisdictions, states in his case, and this
is one reason why his results are so
different.
2.
The Beacon Hill Institute and the Haughwout et al. studies do not control for employment, wages or rates of unionization across
competing jurisdictions. Thus it is impossible to know whether employment
is determined by regional economic conditions (such as the loss of jobs that
affected the whole urban Northeast in the 1970s)[10]
and/or by labor market conditions or by taxes.
3. In the Beacon Hill Institute and the
Haughwout et al. studies state and city
services play no role. In the Beacon
Hill studies the only thing the
government does is pay welfare and unemployment, and these are the
only expenditures that are
included. In the study by Haughwout et
al. only transfer payments from the
federal government and the state are
included. Expenditures on
schools, police and roads are missing in
all these studies. Thus the Beacon
Hill and the Haughwout et al. studies
do not tell us whether taxes have an
effect on employment at all, and if they do what this effect is.
Conclusion
The empirical evidence is that cuts in government services in order to avoid tax increases will result in a reduction in incomes and therefore jobs. Claims to the contrary are based on research that is deliberately shoddy.
[1] L. Jay Helms, “The Effect of State and Local Taxes on Economic Growth: A Time Series-Cross Section Approach,” The Review of Economics and Statistics, 1985, pp. 574-582.
[2] Carroll, Robert, and Michael Wasylenko, “Do State Business Climates Still Matter: Evidence of a Structural Change.” National Tax Journal, Vol. 47 (March), 1994, 19-37.
[3] Helms’ result regarding welfare payments probably confuses cause and effect. The 1970s saw a restructuring of the U.S. economy, with employment migrating from cities to suburbs and from the sun belt to the snow belt. As a result, personal incomes in urban Northeastern states declined during that period and because of this decline in incomes welfare payments increased. See Moshe Adler, Oliver Cooke and James Parrott, “Do Tax Increases In New York Cause a Loss of Jobs? A Review of the Evidence,” State Tax Notes, Feb. 4 2002. (www.columbia.edu/~ma820/State.Tax.Notes.pdf)
[4] See Robert G. Lynch, Do State & Local Tax Incentives Work? Economic Policy Institute, 1996, and Michael Wasylenko, “Taxation and Economic Development: The State of Economic Literature,” New England Economic Review, March/April 1997, 37-52.
[5] Bartik, Timothy, “Small Business Start-Ups in the
United States: Estimates of the Effects
of Characteristics of States.” Southern
Economic Journal, Vol. 55, No. 4, pp.
1004-18.
[6] Munnell, Alicia, 1990. “How Does Public Infrastructure
Affect Retinjal Economi Performance?” New England Economic Review (September/October),
pp. 11-13.
[7] Bartik, Timothy, Growing State Economies: How Taxes and Public Services Affect Private Sector Performance, The Economic Policy Institue, 1996.
[8]Moshe Adler, Oliver Cooke and James Parrott, “Do Tax Increases In New York Cause a Loss of Jobs? A Review of the Evidence,” State Tax Notes, Feb. 4 2002.
[9] Moshe Adler, “Response to the Confusers,” unpublished working paper.
[10] See footnote 3 above.