Do Lower Taxes Stimulate
Small Business Economics
19: 135–145, 2002.
2002 Kluwer Academic Publishers. Printed in the Netherlands.
ABSTRACT. This paper investigates the relation between
taxation and the share of self-employed in two data sets. First,
in a panel of OECD countries a strong negative correlation is
found between the tax burden and the share of self-employed
in total employment. Unfortunately the OECD data contains
a number of measurement problems, and too few observations
to allow a reliable detection of reverse causation. Therefore
the analysis is complemented with regressions on the relation
between the income tax and the share of self-employed using
a panel of Swedish counties over the previous two decades.
Here again there is a strong negative correlation between the
tax burden and the share of self-employment. This survives
simultaneous equations estimates and tests of Granger
causality. Both samples indicate that reducing the tax burden
by 10 percentage points (of GDP) increases the share of self-
employed by about 3 percent of total employment.
Should self-employment be encouraged with lower
taxes? Yes, is the answer often given in the public
policy debate. One line of argument is that some
people who become self-employed are entrepre-
neurs who are creators of jobs and new products.
Another line of argument is that self-employment
is a safety valve for unemployed people or people
that are discriminated in the labour market (e.g.
Moore, 1983). In fact, a number of studies find
that overall employment tends to be higher if a
greater share of employed are self-employed
(Thurik, 1999; Fölster, 1998).
But do lower taxes stimulate self-employment?
Some policy proposals focus on asymmetric tax
cuts, making the tax system more favorable for
self-employed relative to employed people. For
example, in many countries capital gains tax or
wealth tax is reduced for investment in small
businesses. But the argument is also made that
general tax cuts for both self-employed and
employed strengthens incentives to invest and
work, thus making it more likely that potential
entrepreneurs start firms and that unemployed
people move into self-employment.
This paper investigates the link between the
general tax level and the frequency of self-
employment. Regressions based on a sample of
OECD countries indicate a rather strong relation-
ship between the general tax burden and the share
of self-employed. This sample does not, however,
permit a fuller investigation of a number of econo-
metric issues, in particular of simultaneity issues.
Therefore we present a more sofisticated analysis
using a 20-year panel of Swedish counties. This
second sample appears to confirm the relation
between the tax burden and the share of self-
Why taxes might affect self-employment
Theoretical models of the determinants of self-
employment usually draw on Knight’s (1921)
notion that the individual responds to the risk-
adjusted relative earnings opportunities associated
with employment and self-employment. More
recent theoretical models in this tradition consider
explicitly how stochastic shocks in demand or cost
functions affect the decision to become self-
employed (Appelbaum and Katz, 1986; Kihlstrom
and Laffont, 1979). In this type of model it is
obvious that an asymmetric tax reduction, favoring
the self-employed relative to others, increases self-
employment. It is less obvious, however, why a
general tax reduction might increase self-employ-
ment. In the models that build on Knight’s tradi-
tion this only occurs if a general tax reduction
increases willingness to accept a more uncertain
Final version accepted on January 5, 2001
Confederation of Swedish Enterprise
114 82 Stockholm