public policy, entrepreneurship, and economic freedom

PUBLIC POLICY, ENTREPRENEURSHIP, ANDECONOMIC FREEDOM

Steven F. Kreft and Russell S. Sobel

The “entrepreneurial spirit” is something that has long been asso-ciated with the driving force behind economic progress and growth.Joseph Schumpeter (1942) stated that the key to the success of mar-kets lies in the spirits of entrepreneurs who persist in developing newproducts and technologies, through a process he termed as “creativedestruction.” Kaiser (1990) modeled the entrepreneur on the basis ofmany historical characterizations, including the Schumpeterian inno-vator, and concluded that the major characteristics of the entrepre-neur—innovator, risk taker, and resource allocator—are complemen-tary and inseparable facets of entrepreneurship. Kirzner (1997) ar-gues that the entrepreneurial discovery process is vital to theeffectiveness of markets, where discovery entails entrepreneurs dis-covering profit opportunities by trial and error. In this same respect,Jenner (1998) models the Schumpeterian entrepreneurial process asa dynamic process in which entrepreneurs search for new combina-tions of products and production techniques that will lead to in-creased productivity and economic growth. Knight (1921) views theentrepreneur as the bearer of the uninsurable uncertainty present inthe marketplace, with the profit earned being the compensation forbearing this uncertainty.

Recently, the conceptual link between entrepreneurship and eco-nomic growth has received renewed interest by economists. As ar-gued by Minniti (1999), entrepreneurs are the catalysts for economicgrowth because they create a networking externality that promotesthe creation of new ideas and new market formations. The finding

Cato Journal, Vol. 25, No. 3 (Fall 2005). Copyright © Cato Institute. All rightsreserved.

Steven F. Kreft is the Clinical Assistant Professor in the Department of Business Economicsand Public Policy at the Indiana University Kelley School of Business. Russell S. Sobel isProfessor of Economics and Director of the Entrepreneurship Center at West Virginia Uni-versity. The authors would like to thank participants at the 2005 meetings of the Association forPrivate Enterprise Education and the Southern Economic Association, as well as Don Bruceand an anonymous referee for helpful comments and suggestions.

595

that increased entrepreneurial activity leads to greater economicgrowth has been well-established at both the national and local levels.For example, Reynolds, Hay, and Camp (1999) show that one-third ofthe differences in national economic growth rates can be attributed tothe level of entrepreneurship in each country. Supporting these find-ings, Zacharakis, Bygrave, and Sheperd (2000) study 16 developedeconomies and find that entrepreneurial activity explains approxi-mately one-half of the differences in GDP growth between countries.More recently, Henderson (2002) shows that entrepreneurs signifi-cantly impact economic activity at a more local level through fosteringlocalized job creation, increasing wealth and local incomes, and con-necting local economies to the larger global economy.

Based on the increasing awareness of the role of entrepreneurs indriving economic growth, state and local economic development ef-forts have been more heavily directed toward promoting entrepre-neurship. These development efforts have mainly focused on reduc-ing the financial constraints that entrepreneurs face—either throughpreferential loans to new businesses, as those supported by the SmallBusiness Administration, or preferential tax treatment for new orsmall businesses. One such policy that has recently gained popularityaims to devote public resources toward attracting and building alarger amount of venture capital to encourage entrepreneurial activ-ity. This development strategy is largely based on casual observationthat areas with larger amounts of entrepreneurial activity generallytend to also have a larger amount of venture capital.

A recent controversial policy alternative has been popularized byRichard Florida (2002) in his book The Rise of the Creative Class. Theauthor proposes that instead of focusing on developing capital inputs,development efforts should be focused toward making areas moreattractive to bring in and nourish creative, entrepreneurial individu-als. In addition, recent work by Gwartney and Lawson (2002), Farr,Lord, and Wolfenbarger (1998), Gwartney, Lawson, and Holcombe(1999), Cole (2003), and Powell (2003) highlight the role of economicfreedom in promoting economic prosperity and growth. The results ofthis research suggest that policies consistent with expanding the eco-nomic freedom of individuals are the cornerstone of successful eco-nomic development policy.

In this article, we propose that the main difference between thesecompeting development strategies is a question of the direction ofcausation between entrepreneurial activity and the quantity of ven-ture capital. We then proceed to answer this question with an em-pirical test to determine whether it is more venture capital that causesmore entrepreneurial activity in an area, or whether the presence of

CATO JOURNAL

596

more entrepreneurial activity simply, and automatically, causes moreventure capital to flow into an area. Not only is this an interestingacademic question, it also has significant implications for how to bestdirect the limited resources available for state and local economicdevelopment efforts. The basic question is whether it is better todevote development efforts toward bringing in venture funds or toencourage more entrepreneurial activity among individuals in an area(or alternatively, to attract entrepreneurs to the area) by enactingpolicy reform that expands economic freedom. Even more interestingis the possibility that there is causation running simultaneously inboth directions between venture capital and entrepreneurial activity.If these two phenomenon have this type of relationship, developmentefforts will only be successful if resources are devoted simultaneouslyto promoting both larger venture funds and encouraging entrepre-neurial activity among individuals.

The next section of this article proceeds to uncover the direction ofcausality between venture capital and entrepreneurial activity. On thebasis of these results, we then consider the issue of which governmentpolicies best stimulate the underlying causal factors that promoteentrepreneurship. Finally, we present concluding remarks.

Direction of Causality between Venture Capital andEntrepreneurial Activity

One variable that has been widely supported in the literature as amajor determinant of entrepreneurial activity is the amount of ven-ture capital investment that is available to entrepreneurs. The Cor-poration for Enterprise Development (2001) lists eight core elementsof an infrastructure necessary for supporting entrepreneurship. Six ofthose elements revolved around the financing that was available forpotential entrepreneurs. Also, highlighting the importance of financ-ing, Henderson (2002) states that the availability of financial re-sources in an area, especially venture capital investment, is vital todeveloping entrepreneurs. However, one important idea that has gen-erally been overlooked by previous authors is the notion that venturecapital investment may be endogenous to the model of entrepreneur-ial activity. More specifically, it is hard to determine if the venturecapital investment is creating entrepreneurship, or if the investmentis simply flowing to the states that already have significant levels ofentrepreneurial activity.

We perform state-level panel causality tests on venture capital in-vestment and two measures of entrepreneurial activity (sole propri-etorships and patent activity). The measure of venture capitalinvestment is from the 2002 Venture Capital Profiles published by

ENTREPRENEURSHIP AND ECONOMIC FREEDOM

597

PricewaterhouseCoopers/Thomson Venture Economics/NVCA Mon-eytree, and includes cash-for-equity investments by professional ven-ture capital firms in private emerging companies in the United States,where the venture capital firm can be based, or based abroad.1 Thefirst measure of entrepreneurial activity, sole proprietorships, hasbeen widely supported in the literature as a good proxy for the levelof entrepreneurship.2 The second measure of entrepreneurship, pat-ent activity, is new to this article, and is measured as the number ofutility patents (those received for general inventions or innovations)granted annually in each state. The logic behind patent activity as ameasure of entrepreneurship rests in the notion that the most directand visible outcome of the entrepreneurial process is innovation,which should be reflected in the quantity of patents.3

The causality test procedure used here builds on the Granger(1969) and Sims (1972) causality framework by modifying the test toincorporate the pooled time-series properties of all the 50 states. Oneproblem that may arise in using the pooled state data is that thedifferences across states may be significant enough to bias the truetime-series information that is available in the data. Following theapproach of Blomstrom, Lipsey, and Zejan (1996) and Farr, Lord,and Wolfenbarger (1998), state intercept dummies were included ineach regression specification to avoid the possible bias by controllingfor any state-specific influences.4 Specifically, the effect of the stateintercept dummies is to remove the cross-sectional differences of thestates, while leaving only the time-series variations to be analyzed.5

The general Granger-Sims causality test of two variables X and Y,modified for state panel data, can be seen in the following equations,where equation (1) tests causality running from X to Y, and equation(2) tests causality running from Y to X.

1Professional venture capital firms include the following types of firms: Small BusinessInvestment Companies (SBICs), venture arms of corporations, institutions, investmentbanks, and similar entities whose primary activity is venture capital investing.2The Bureau of Economic Analysis reports the number of sole proprietors on the basis offederal income tax forms filed by individuals of each state.3Ideally, a more exhaustive indicator of innovative activities would be to measure thenumber of patent applications per state, but those data do not appear to exist at the statelevel. Griliches (1990) found that there has been a 65 percent patent application grantingrate in the United States, and from 1880 to 1989, patent grants have followed closely withthe trend of patent applications (see p. 1664, Figure 1). Therefore, the patent grants shouldbe a reasonable measure of innovation in the states.4The state intercept dummy parameter estimates are not reported with the causality re-gression results but are available on request.5The Granger causality framework and the testing procedures involved are still somewhatcontroversial in economics. Obvious limitations to the methodology (like Christmas cardsales causing Christmas) are discussed and highlighted in Bishop (1979).

CATO JOURNAL

598

�1� Yt,i = �i + �m=1

M

�m Yt−m,i + �n=1

N

�n Xt−n,i + �t,i

�2� Xt,i = �i + �v=1

V

�vXt−v,i + �w=1

W

�wYt−w,i + �t,i

Note that the subscript i refers to the corresponding state obser-vation; the error terms �t,i and �t,i are assumed to be white noise; and,the number of lagged values (M and N or V and W) of the indepen-dent variables are chosen to adequately capture the relationship be-tween X and Y.

To check for a one-way causal relationship, both directions of cau-sality have to be investigated. In order to test if X Granger causes Y,equation (1) is estimated with and without the lagged X variables, andthen an F-test is performed to test the null hypothesis that �n = 0 forn = 1,. . .,N. Rejecting the null hypothesis would show that X Grangercauses Y. In order to test if Y Granger causes X, equation (2) isestimated with and without the lagged Y variables, and then an F-testis performed to test the null hypothesis that �w = 0 for w = 1,. . .,W.Rejecting the null hypothesis would show that Y Granger causes X.

This modified Granger-Sims causality framework is used to runcausality tests between venture capital investment, sole proprietor-ships, and patent activity in the United States during the 10-yearperiod 1992–2001.6 Descriptions of all variables used in this article,along with the sources of these data, are given in Table 1. The resultsof the causality tests results are presented in Table 2, and show thata one-way causal relationship exists between entrepreneurship andventure capital investment.7 Specifically, the level of sole proprietorswas found to Granger cause venture capital (specifications 1 and 3 in

6Due to data limitations on venture capital investment, Alaska, Hawaii, North Dakota,South Dakota, Vermont, West Virginia, and Wyoming were not included in the causalitytests between entrepreneurship and venture capital investment. Also, the California ob-servations were suppressed because standard outlier tests revealed that California is astatistical outlier in venture capital investment. Specifically, California’s observations hadstandardized residuals that were greater than 2.5 standard deviations from the mean inabsolute value.7The test structure reported in Table 2 includes only one lag of the independent variables,in part, because of the limited number of observations and also to conserve on degrees offreedom. However, this one-lag relationship seems to best reflect the highly mobile char-acteristics of venture capital investment. It should be noted that the causality tests were runusing two-and three- lags and the results were not substantially different. The only differ-ence is that there is a weak dual-causality relation found between patent activity andventure capital investment at the 10 percent confidence level. Also, it should be noted thata more simplified t-test can be run in the causality tests that incorporate only one lag;however, the more general F-test is also acceptable.


599

TA

BL

E1

DA

TA

DE

SCR

IPT

ION

AN

DSO

UR

CE

S

Var

iabl

eN

ame

(Sou

rce)

Des

crip

tion

Mea

n(S

t.D

ev.)

Cau

salit

yT

est

Var

iabl

esSo

lepr

opri

etor

ship

aA

nnua

lnon

farm

prop

riet

ors

empl

oym

ent

asre

veal

edth

roug

hin

com

eta

xda

ta38

9,60

8(4

66,8

30)

Pate

ntac

tivity

bN

umbe

rof

annu

alut

ility

pate

nts

gran

ted

inth

eU

nite

dSt

ates

,whi

char

ere

ceiv

edfo

ral

lge

nera

lU.S

.inv

entio

ns

1,02

8(1

,589

)

Ven

ture

capi

tali

nves

tmen

tcV

entu

reca

pita

linv

estm

ent

toU

.S.c

ompa

nies

(fro

mal

lsou

rces

,inc

ludi

ngno

n-U

.S.)

inm

illio

nsof

curr

ent

dolla

rs

426

(1,0

12)

Ent

repr

eneu

rial

Gro

wth

Var

iabl

esSo

lepr

opri

etor

grow

tha

Perc

ent

chan

gein

nonf

arm

prop

riet

ors

empl

oym

ent

(NPE

),ca

lcul

ated

as(N

PE20

00−

NPE

1996

)/N

PE19

96

10.9

8(4

.77)

All-

gove

rnm

ent

econ

omic

free

dom

inde

xdC

ompo

site

inde

xm

easu

reof

fede

ral,

stat

e,an

dlo

calp

olic

ies

that

affe

ctin

divi

dual

econ

omic

free

dom

6.68

(0.5

0)

All-

gove

rnm

ent

com

pone

nt1d

Indi

cato

rsof

the

size

ofgo

vern

men

t7.

24(0

.69)

All-

gove

rnm

ent

com

pone

nt2d

Indi

cato

rsof

disc

rim

inat

ory

gove

rnm

ent

taxa

tion

5.77

(0.5

6)A

ll-go

vern

men

tco

mpo

nent

3dIn

dica

tors

ofla

bor

mar

ket

regu

latio

n6.

86(0

.68)

Stat

ego

vern

men

tec

onom

icfr

eedo

min

dexd

Com

posi

tein

dex

mea

sure

ofst

ate,

and

loca

lpo

licie

sth

ataf

fect

indi

vidu

alec

onom

icfr

eedo

m

6.96

(0.6

7)

CATO JOURNAL

600

Stat

ego

vern

men

tco

mpo

nent

1dIn

dica

tors

ofth

esi

zeof

gove

rnm

ent

7.21

(1.0

5)St

ate

gove

rnm

ent

com

pone

nt2d

Indi

cato

rsof

disc

rim

inat

ory

gove

rnm

ent

taxa

tion

6.85

(0.6

5)St

ate

gove

rnm

ent

com

pone

nt3d

Indi

cato

rsof

labo

rm

arke

tre

gula

tion

6.83

(0.6

6)In

heri

tanc

eta

xla

we

Dum

my

=1

ifth

est

ate

levi

esan

esta

te,i

nher

i-ta

nce,

orgi

ftta

xbe

yond

the

fede

ralr

ate

0.32

(0.4

7)

Perc

ent

with

high

scho

olde

gree

fPe

rcen

tof

popu

latio

nre

ceiv

ing

ahi

ghsc

hool

degr

eeas

itshi

ghes

tle

velo

fed

ucat

ion

82.8

8(4

.90)

Perc

ent

with

colle

gede

gree

fPe

rcen

tof

popu

latio

nre

ceiv

ing

a4-

yr.c

olle

gede

gree

asits

high

est

leve

lof

educ

atio

n22

.99

(4.3

5)

Perc

ent

mal

esg

Perc

ent

ofth

ela

bor

forc

eth

atis

mal

e53

.38

(1.1

9)Pe

rcen

tw

hite

gPe

rcen

tof

the

labo

rfo

rce

that

isw

hite

84.9

8(1

1.78

)M

edia

nag

efM

edia

nag

eof

the

popu

latio

n34

.15

(1.7

0)U

nem

ploy

men

tra

teg

Civ

ilian

unem

ploy

men

tra

te5.

15(1

.16)

Perc

ent

serv

ice

empl

oym

entg

Perc

ent

ofth

ela

bor

forc

eem

ploy

edin

serv

ice

indu

stri

es21

.48

(3.4

8)

Prop

erty

crim

era

teh

Num

ber

ofpr

oper

tycr

imes

per

100,

000

popu

latio

n4,

313

(1,0

40)

a U.S

.Dep

artm

ent

ofC

omm

erce

,Bur

eau

ofE

cono

mic

Ana

lysi

s,St

ate

and

Loc

alA

rea

Dat

a,W

ashi

ngto

n,D

.C.

bU

.S.P

aten

tan

dT

rade

mar

kO

ffic

e,U

tilit

yPa

tent

Cou

nts

byC

ount

ry/S

tate

and

Yea

r,W

ashi

ngto

n,D

.C.,

2001

.c Pr

icew

ater

hous

eCoo

pers

/T

hom

son

Ven

ture

Eco

nom

ics

/N

VC

AM

oney

tree

,Ven

ture

Cap

ital

Prof

ile,2

002.

dK

arab

egov

ic,A

mel

a;F

red

McM

ahon

,and

Dex

ter

Sam

ida.

Eco

nom

icF

reed

omof

Nor

thA

mer

ica,

2002

.eC

CH

Inco

rpor

ated

,Sta

teE

stat

eT

axes

,200

2.f U

.S.D

epar

tmen

tof

Com

mer

ce,C

ensu

sB

urea

u,St

atis

tica

lA

bstr

acts

ofth

eU

nite

dSt

ates

,Was

hing

ton,

D.C

.gU

.S.D

epar

tmen

tof

Lab

or,B

urea

uof

Lab

orSt

atis

tics,

Stat

ean

dA

rea

Lab

orF

orce

Stat

isti

cs,W

ashi

ngto

n,D

.C.

hU

.S.D

epar

tmen

tof

Just

ice,

Bur

eau

ofJu

stic

eSt

atis

tics,

FB

I’s

Uni

form

Cri

me

Rep

orts

,Was

hing

ton,

D.C

.


601

TA

BL

E2

CA

USA

LIT

YT

EST

RE

SUL

TS

BE

TW

EE

NST

AT

EE

NT

RE

PRE

NE

UR

IAL

AC

TIV

ITY

AN

DV

EN

TU

RE

CA

PIT

AL

INV

EST

ME

NT

Ven

ture

Cap

ital

Inve

stm

ent

Sole

Prop

riet

ors

Pate

ntA

ctiv

ity

(1)

(2)

(3)

(4)

(5)

(6)

Ven

ture

Cap

italI

nves

tmen

t(t

-1)

0.42

8***

0.21

4***

−0.4

410.

002

(mill

ions

ofdo

llars

)(8

.42)

(4.3

9)(0

.55)

(0.1

4)So

lePr

opri

etor

s(t

-1)

4.55

8***

969.

335*

**92

6.33

0***

1.72

8***

(tho

usan

dsof

prop

riet

ors)

(5.3

9)(7

2.18

)(5

3.61

)(6

.15)

Pate

ntA

ctiv

ity(t

-1)

146.

664*

**79

1.40

4***

90.1

24**

*69

.845

***

(hun

dred

sof

pate

nts)

(11.

60)

(2.9

4)(2

3.32

)(1

5.98

)R

esul

t/Fin

ding

sole

prop

riet

ors

caus

eve

ntur

eca

pita

l

pate

ntac

tivity

caus

esve

ntur

eca

pita

l

vent

ure

capi

tal

does

not

caus

eso

lepr

opri

etor

s

pate

ntac

tivity

caus

esso

lepr

opri

etor

s

vent

ure

capi

tal

does

not

caus

epa

tent

activ

ity

sole

prop

riet

ors

caus

epa

tent

activ

ityF

-sta

tistic

[1,3

34]

29.1

1***

134.

66**

*0.

308.

65**

*0.

0237

.87*

**N

umbe

rof

Obs

erva

tions

378

378

378

378

378

378

R-s

quar

ed0.

610.

700.

990.

990.

980.

98N

OT

ES:

Stat

edu

mm

yva

riab

les

wer

ein

clud

edin

each

regr

essi

onsp

ecifi

catio

n,an

dth

ees

timat

edco

effic

ient

sar

eav

aila

ble

upon

requ

est

toth

eau

thor

s.C

alifo

rnia

was

omitt

edas

anou

tlier

.Sig

nific

ance

leve

lsar

ere

pres

ente

dby

the

follo

win

g:**

*1pe

rcen

t,**

5pe

rcen

t,*1

0pe

rcen

t;ab

solu

tet-

stat

istic

sar

ein

pare

nthe

ses.

CATO JOURNAL

602

Table 2), and the level of patent activity was shown to Granger causeventure capital (specifications 2 and 5 in Table 2). Also, tests wereperformed to determine the direction of causality between the twomeasures of state entrepreneurship, and the tests revealed that dualcausality exists between sole proprietors and patent activity (specifi-cations 4 and 6 in Table 2). The dual causality result is not surprising,considering that sole proprietors and patent activity are intended tomeasure the same thing—the level of entrepreneurial activity.

The causality results, showing a one-way causal relationship run-ning from entrepreneurship to venture capital investment, show thatventure capital investment funds are simply flowing to states withalready well-established entrepreneurial activity. One explanation ofour finding is that venture capital investment is inherently more mo-bile than labor, which would imply that, as the level of entrepreneur-ial activity rises in a particular geographic region, new venture capitaltends to automatically, and freely, flow from all parts of the UnitedStates, and also from areas abroad, to fund the entrepreneurial ac-tivity in that region.

It is important to note that our results do not contradict the ideathat venture capital is important in the entrepreneurial process. Infact, our results are most consistent with the literature on entrepre-neurial survival, which suggests that once an entrepreneurial ventureis started, venture funding will significantly increase the chance ofsurvival.8 What our results do say, however, is that focusing develop-ment efforts on attracting more venture funding will not be an effec-tive method of encouraging the higher levels of entrepreneurial ac-tivity necessary for economic growth. Rather, attracting and promot-ing underlying entrepreneurial activity must be the focus ofdevelopment efforts and venture funding will automatically, and natu-rally, flow into the area to support this activity.

State Policies that Promote EntrepreneurshipOur empirical results from the previous section suggest that entre-

preneurial activity (measured by patents and sole proprietorships)tends to be the underlying factor that attracts more venture capital toan area. The remaining question is then how to structure governmentpolicy to encourage more entrepreneurial activity among individualsin an area (either by making current residents more entrepreneurial,or by attracting new entrepreneurs to the area).

8Holtz-Eakin, Joulfaian, and Rosen (1994) and Blanchflower and Oswald (1998) presentevidence that financing is key to the survival of entrepreneurial ventures.


603

One such structure for government policy is suggested by anotherstrand of literature attempting to explain economic growth differen-tials across countries by differences in a well-constructed index ofeconomic freedom. Generally, these indexes attempt to condense intoa single number the degree of economic freedom individuals have ina geographic area in several key categories, such as low taxes, lowregulations, and secure property rights. Studies using these indexessuch as Gwartney and Lawson (2002), Farr, Lord, and Wolfenbarger(1998), Gwartney, Lawson, and Holcombe (1999), Cole (2003), andPowell (2003) have generally found that countries with a higher eco-nomic freedom score not only have larger per capita incomes, but alsotend to have higher rates of economic growth.9

In this article, we propose that the “missing link” that has yet to bedemonstrated between economic freedom and economic growth isentrepreneurial activity. That is, underlying economic freedoms gen-erate growth primarily because they promote underlying entrepre-neurial activity, which is then the source of economic growth. Thishypothesis is consistent with the recent works of Baumol (1990, 2002)and Boettke (2001, 2003), who propose that the efforts of creative,entrepreneurial individuals in different countries, or areas, are di-rected in different ways depending on the prevailing economic andpolitical institutions. In areas with institutions providing secure prop-erty rights, a fair and balanced judicial system, contract enforcement,and effective limits on government’s ability to transfer wealth throughtaxation and regulation, individuals are more likely to engage in thecreation of new wealth through productive market entrepreneurship.In areas without these institutions, creative individuals are more likelyto engage in attempts to capture transfers of existing wealth throughunproductive political entrepreneurship.

The freedom index that we use is developed by Karabegovic,McMahon, and Samida (2002), and is a composite index measure ofmany public policies that affect the economic freedom of individualsin that state. The authors construct two different freedom indexes.First, an index is created that reflects the policies that are put in placeby federal, state, and local governments (what we refer to as the“all-government economic freedom index”). Second, an index is cre-ated that reflects the policies that are put in place by state and localgovernments only (what we refer to as the state and local government

9The idea that free-market institutions, such as secure private property rights, are vital toeconomic growth was also stressed in the works of Peter Bauer, see Dorn (2002) for a gooddiscussion of Bauer’s contributions in this area. Economic freedom has also been shown tobe correlated with lower levels of violent conflict by Tures (2003).

CATO JOURNAL

604

economic freedom index). Furthermore, each freedom index can bebroken down into three major components: (1) size of government,(2) government taxation, and (3) labor market flexibility. First, thesize of government is based on general government purchases, trans-fer payments, and subsidies. Second, government taxation incorpo-rates total government revenue, income tax rates and thresholds, in-direct taxes, and sales taxes. Third, the labor market flexibility is basedon minimum-wage earnings, government employment, occupationallicensing, and union density.

To illustrate the general relationship between economic freedomand productive private-sector entrepreneurial activity, Figure 1 showsthe relationship between state economic freedom scores (using theall-government economic freedom index) and the annualized growthrate of sole proprietors between 1996 and 2000. The scatter plot ofthe raw data is supported by a simple regression line fit between thetwo variables. The positive correlation can be seen clearly in thefigure, which shows that the states with more economic freedom in1996 experienced higher growth rates in entrepreneurship over thefollowing five years. This view is highlighted by Lee (1991: 50) whowrites:

No matter how fertile the seeds of entrepreneurship, they witherwithout the proper economic soil. In order for entrepreneurship togerminate, take root, and yield the fruit of economic progress it hasto be nourished by the right mixture of freedom and accountability,a mixture that can only be provided by a free market economy.

To verify our hypothesis, we have gathered state-level data on thegrowth of entrepreneurial activity, on other key factors that have beenshown to be correlated with entrepreneurial activity, and on the de-gree of economic freedom.

In modeling entrepreneurship, the existing literature has mainlyfocused on the question of what characterizes an individual entrepre-neur. Our analysis is the first attempt to model environments forentrepreneurship at the state level. Also, in an attempt to stay con-sistent with the previous literature and to directly test whether theexisting model can be applied to state-level aggregated data, ouranalysis will focus attention on modeling sole proprietors.

Furthermore, in order to be consistent with the earlier causalityanalysis, we take the values of the state explanatory variables in theinitial year and see which variables significantly impact each state’sgrowth rate of entrepreneurship over the following five years. Thistechnique tries to get at the heart of causality by taking the existing


605

state characteristics at one point in time and asking what character-istics lead to entrepreneurial growth in the next period.

Generally, the formal estimated regression takes the followingfunctional form:

�3� SPGRi = �i + �x=1

X

�x DEMx,i + �y=1

Y

�y POLy,i + �z FREEi,

where SPGRi is the annualized growth rate of sole proprietors be-tween 1996 and 2000 in state I, DEMi is a set of demographic andeconomic control variables for state I, POLi is a set of political vari-ables for state I, and FREEi is the specified economic freedom index,or index component for state i. In an attempt to uncover the causalnature of this relationship, we use the beginning-year (1996) valuesfor the independent variables. In other words, we attempt to see whatinfluences were in place in the initial year (1996) that led to the rateof entrepreneurial growth that occurred over the following five years(1996 to 2000).

In determining the state control variables, we relied heavily onthose variables proposed in the literature, and our underlying controlsfollow the approach of Bruce (2000, 2002). The variables included to

FIGURE 1RELATIONSHIP BETWEEN STATE ECONOMIC FREEDOM AND

GROWTH OF ENTREPRENEURIAL ACTIVITY

CATO JOURNAL

606

capture the demographic characteristics of entrepreneurs are popu-lation statistics on the median age, percent males, percent white,percent receiving a high school education, and percent receiving acollege education. The variables included to capture the socio-economic characteristics of the states are the unemployment rate,percent employed in service industries, and the property crime rate.

With regard to the policy variable, which is separate from thefreedom index measures, we include a variable that captures thosestates that enacted inheritance taxes that went over and beyond thefederal taxes. Holtz-Eakin, Joulfaian, and Rosen (1994) and Blanch-flower and Oswald (1998) find evidence that an individual’s inheri-tance increases the probability of entering and succeeding in an en-trepreneurial venture. The reason is that inheritance often providesthe seed funding necessary to develop and finance a new venture upuntil the point at which it becomes possible to secure outside debt orventure funding. Furthermore, Holtz-Eakin (1999), in a survey of theliterature on estate taxes, concludes that entrepreneurs are morelikely to bare the burden of estate taxes because they are inherentlymore exposed to the taxation of wealth accumulation. Thus, inheri-tance taxes, which directly reduce the ability of successful entrepre-neurs to pass on their wealth to fund future generations of entrepre-neurs, should lead to less entrepreneurial activity in states with oner-ous estate laws.

Also relevant to the focus of our research is the strand of literaturefocusing on income and payroll tax policy.10 For example, Bruce(2000) examines income and payroll taxes of the self-employed andwage-and-salary workers to see if tax differentials affect the choice tobe self-employed. The author finds that the differential tax treatmentsignificantly affects the probability of leaving self-employment for awage-and-salary job. Bruce (2002) extends his original work to allowfor the endogeneity of individual tax rates and finds that taxes havemixed effects on the level of entrepreneurial activity. His resultshighlight the overall findings of the previous literature—namely, thatthere is no conclusive evidence on the relationship between incometax rates and entrepreneurial activity. However, one would expectthat the tax environment would influence the states attractiveness toentrepreneurship.

As noted earlier, one of the propositions of this article is that astate’s underlying economic freedom is an essential determinant of

10For a good review of the literature on the relationship between income and payroll taxesand entrepreneurship, see Bruce (2002).


607

the state’s ability to create and attract entrepreneurial activity. Putsimply, an environment of low taxes, low regulations, and secureprivate property rights (as measured by the economic freedom index)is what is necessary to encourage growth in entrepreneurial activity.Therefore, in all cases, the economic freedom indexes, and theircomponents, are expected to have a positive sign showing that moreeconomic freedom will lead to more entrepreneurial activity.

The estimated determinants of entrepreneurial growth are pre-sented in Tables 3 and 4, which differ in that Table 3 reports theresults of the estimations that incorporate the all-government eco-nomic freedom index and components, while Table 4 only reports theresults of the estimations that incorporate the state and local govern-ment economic freedom index and components. Examining Tables 3and 4, the all-government economic freedom index and the state andlocal index, included in specifications (1) and (2), are significant at the1 percent level or better. Thus, the states with the most economicfreedom in 1996 had the highest subsequent growth of entrepreneur-ial activity over the next five years, regardless of which economicfreedom index we use. This result shows that state policymakers needto ensure that economic freedom exists in their state in order topromote entrepreneurial growth, which in turn naturally attracts thenecessary venture capital.

Specifications (3) and (4) in both tables present the results ofincluding the freedom index components separately. This is an inter-esting exercise to see which components exert stronger relative in-fluences on the growth of entrepreneurs. At the all-government level,low government taxes are the only component that exerts a significantinfluence on the growth of entrepreneurship. Focusing solely on thestate and local level, labor market flexibility is the only componentthat exerts a significant influence on the growth of entrepreneurship.It is important to point out, however, that economic freedom consistsof an environment of low taxes, low regulations, and secure privateproperty rights—and those factors may only work jointly, as seen bythe overall high significance of the composite indexes and the re-duced significance of the component measures. This lack of signifi-cance when all separate categories are included is also the result ofthe high degree of correlation among the components.

In the four specifications in which the inheritance tax is included,it is found to be significant in explaining the growth of entrepreneur-ial activity. Specifically, the presence of state inheritance taxes beyondthe federal level exerts a negative influence on the growth of stateentrepreneurial activity. These results lend support to the findings ofHoltz-Eakin, Joulfaian, and Rosen (1994), Blanchflower and Oswald

CATO JOURNAL

608

(1998), and Holtz-Eakin (1999) that high inheritance taxes directlyreduce the likelihood of individuals becoming entrepreneurs and alsolower the reward from entrepreneurship.

The demographic control variables that were consistently signifi-cant were the median age and percent with college degree. First, theestimated influence of the median age shows that states with youngerpopulations experienced higher entrepreneurial growth. The influ-ence of age has had mixed results in the literature, and often thenegative relation is explained by the view that older individuals maybe more risk averse with regard to income. Moreover, entrepreneur-ial ventures are characterized by high risk and are more likely to beundertaken by younger, more risk-loving individuals. Second, the per-cent of the population receiving a college education is shown to exerta negative and significant influence on state entrepreneurial growth.The influence of education has also had mixed results in the litera-ture. However, this somewhat counterintuitive result is often ex-plained in the literature by the notion that a high school educationgives an individual the basic training and understanding needed tostart his or her own business without specifying a certain way ofthinking or performing tasks (positive influence), while a four-yearcollege education trains an individual to think in a more specializedfield, which may be better suited for a wage-and-salary job (negativeinfluence). The other three demographic variables (percent of popu-lation with high school degree, percent males, and percent white) areinsignificant in all the regression specifications.

The economic control variables that were consistently significantwere the percent service-sector employment and the property crimerate. Specifically, our results show that states with larger service sec-tors in their economies experience more growth in entrepreneurialactivity. As noted by Blau (1987), the industries in which entrepre-neurship is more common are typically the service and retail tradesectors, so the positive influence is in line with past findings. Withregard to the crime rate, we find a negative influence on the growthof entrepreneurial activity. Thus, the protection of property rights isimportant for stimulating entrepreneurial activity. The final economiccontrol variable, the unemployment rate, was statistically insignificantin all the regression specifications.

Taken as a whole, our results have significant policy implicationsfor state and local development agencies. To encourage economicgrowth, localities must encourage entrepreneurial activity; and to doso, they must focus on creating an environment consistent with eco-nomic freedom rather than on bringing in more venture capital to thearea. Again, a state’s economic freedom consists of an environment of


609

TA

BL

E3

EST

IMA

TE

DD

ET

ER

MIN

AN

TS

OF

STA

TE

EN

TR

EPR

EN

EU

RIA

LG

RO

WT

H,1

996–

2000

Dep

ende

ntV

aria

ble:

Stat

eSo

lePr

opri

etor

Gro

wth

Rat

es

(1)

(2)

(3)

(4)

Stat

eP

olic

yV

aria

ble

sA

ll-go

vern

men

tec

onom

icfr

eedo

m3.

784*

**3.

668*

**(3

.33)

(3.1

4)C

ompo

nent

1:Sm

alls

ize

ofgo

vern

men

t0.

253

0.08

3(0

.24)

(0.0

8)C

ompo

nent

2:L

owgo

vern

men

tta

xatio

n3.

484*

**2.

974*

*(2

.65)

(2.1

8)C

ompo

nent

3:L

abor

mar

ket

flexi

bilit

y1.

406

1.52

3(1

.54)

(1.5

9)In

heri

tanc

eta

xla

w−2

.055

*−2

.467

**(1

.79)

(2.2

5)St

ate

Con

trol

Var

iab

les

Con

stan

t16

.332

21.0

11−3

1.85

7−2

0.86

9(0

.42)

(0.5

2)(0

.78)

(0.4

9)Pe

rcen

tof

popu

latio

nw

ithhi

ghsc

hool

dipl

oma

−0.0

010.

033

0.12

70.

159

(0.0

1)(0

.21)

(0.8

2)(0

.98)

CATO JOURNAL

610

Perc

ent

ofpo

pula

tion

with

colle

gede

gree

−0.3

96**

−0.4

22**

−0.3

64**

−0.4

01**

(2.2

3)(2

.32)

(2.1

1)(2

.22)

Perc

ent

ofpo

pula

tion

mal

e−0

.379

−0.5

080.

054

−0.1

50(0

.59)

(0.7

8)(0

.09)

(0.2

3)Pe

rcen

tof

popu

latio

nw

hite

0.08

30.

103

0.06

20.

087

(1.2

5)(1

.52)

(0.9

8)(1

.34)

Med

ian

age

ofpo

pula

tion

−0.8

25**

−0.9

49**

−0.4

55−0

.665

(1.9

9)(2

.27)

(1.0

7)(1

.52)

Une

mpl

oym

ent

rate

0.54

70.

731

0.36

20.

617

(0.9

2)(1

.22)

(0.6

1)(1

.01)

Perc

ent

serv

ice

empl

oym

ent

0.55

1***

0.59

5***

0.67

1***

0.72

0***

(3.0

4)(3

.22)

(3.8

5)(3

.95)

Prop

erty

crim

era

te−0

.001

**−0

.002

**−0

.001

*−0

.001

*(2

.05)

(2.2

7)(1

.70)

(1.9

4)N

umbe

rof

obse

rvat

ions

5050

5050

R-s

quar

ed0.

560.

520.

630.

58N

OT

ES:

Sign

ifica

nce

leve

lsar

ere

pres

ente

dby

the

follo

win

g:**

*1pe

rcen

t,**

5pe

rcen

t,*1

0pe

rcen

t;ab

solu

tet-

stat

istic

sin

pare

nthe

ses.


611

TA

BL

E4

EST

IMA

TE

DD

ET

ER

MIN

AN

TS

OF

STA

TE

EN

TR

EPR

EN

EU

RIA

LG

RO

WT

H,1

996–

2000

Dep

ende

ntV

aria

ble:

Stat

eSo

lePr

opri

etor

Gro

wth

Rat

es

(1)

(2)

(3)

(4)

Stat

eP

olic

yV

aria

ble

sSt

ate

and

loca

lgov

ernm

ent

econ

omic

free

dom

3.24

7***

2.78

9***

(3.2

1)(2

.69)

Com

pone

nt1:

Smal

lsiz

eof

gove

rnm

ent

0.41

70.

159

(0.4

7)(0

.17)

Com

pone

nt2:

Low

gove

rnm

ent

taxa

tion

0.97

40.

671

(0.8

7)(0

.58)

Com

pone

nt3:

Lab

orm

arke

tfle

xibi

lity

2.13

5**

2.38

8**

(2.0

4)(2

.22)

Inhe

rita

nce

tax

law

−2.6

15**

−2.3

44*

(2.2

1)(1

.93)

Stat

eC

ontr

olV

aria

ble

sC

onst

ant

−14.

349

−2.2

38−1

3.89

7−3

.335

(0.3

4)(0

.05)

(0.3

3)(0

.08)

Perc

ent

ofpo

pula

tion

with

high

scho

oldi

plom

a0.

145

0.16

00.

137

0.15

2(0

.85)

(0.9

0)(0

.79)

(0.8

5)

CATO JOURNAL

612

Perc

ent

ofpo

pula

tion

with

colle

gede

gree

−0.3

83**

−0.4

08**

−0.3

88**

−0.4

11**

(2.1

5)(2

.18)

(2.1

5)(2

.21)

Perc

ent

ofpo

pula

tion

mal

e−0

.028

−0.2

06−0

.027

−0.1

89(0

.04)

(0.3

1)(0

.04)

(0.2

9)Pe

rcen

tof

popu

latio

nw

hite

0.05

80.

085

0.06

50.

091

(0.8

6)(1

.22)

(0.9

4)(1

.30)

Med

ian

age

ofpo

pula

tion

−0.7

22*

−0.9

00**

−0.7

77*

−0.9

53**

(1.7

2)(2

.08)

(1.8

1)(2

.20)

Une

mpl

oym

ent

rate

0.82

30.

933

0.58

40.

642

(1.2

9)(1

.40)

(0.8

1)(0

.86)

Perc

ent

serv

ice

empl

oym

ent

0.61

4***

0.62

2***

0.65

5***

0.71

4***

(3.3

6)(3

.49)

(3.4

6)(3

.69)

Prop

erty

crim

era

te−0

.001

−0.0

01*

−0.0

01−0

.001

*(1

.38)

(1.7

7)(1

.41)

(1.7

5)N

umbe

rof

obse

rvat

ions

5050

5050

R-s

quar

ed0.

550.

490.

570.

52N

OT

ES:

Sign

ifica

nce

leve

lsar

ere

pres

ente

dby

the

follo

win

g:**

*1pe

rcen

t,**

5pe

rcen

t,*1

0pe

rcen

t;ab

solu

tet-

stat

istic

sin

pare

nthe

ses.


613

low taxes, low regulations, and secure private property rights, wherethese factors jointly work to produce economic freedom. Therefore,one component by itself will not necessarily encourage entrepreneur-ial activity without the other factors in place.

ConclusionWe began this article by reviewing the well-documented link be-

tween entrepreneurial activity and economic growth. Local economicdevelopment efforts have recently recognized this link and have be-gun to enact policies targeted at increasing entrepreneurial activity.Many localities have focused these efforts toward attracting new ven-ture capital investment funds. The underlying, but unsubstantiated,assumption is that more venture capital will cause an increase insuccessful entrepreneurial activity. Recently, however, some criticshave questioned whether the limited resources available for develop-ment efforts would be better directed toward attracting and nurturingindividual entrepreneurs.

The state panel causality tests we perform in this article concludethat entrepreneurial activity causes an inflow of venture funding, andnot vice versa. Because entrepreneurial activity tends to be the un-derlying factor that automatically and naturally attracts more venturecapital to an area, economic development policies should focus oncreating an environment attractive to individual entrepreneurs, ratherthan on attracting venture capital.

We gathered data across U.S. states on the growth of entrepre-neurial activity in each state, other key factors that have previouslybeen shown to be correlated with entrepreneurial activity for thatstate, and the degree of economic freedom in the state. Our resultsshow that an area’s degree of economic freedom significantly impactsthe underlying level of entrepreneurial activity. Put simply, an envi-ronment of low taxes, low regulations, and secure private propertyrights is what is necessary to encourage the entrepreneurial activitythat is vital to produce economic growth.

In addition to the clear implications our results have for economicdevelopment efforts, we also provide a significant contribution to thegrowing literature on the relationship between economic freedomand economic growth. More specifically, our results fill in the “miss-ing link” in this well-documented relationship by showing that theconduit between economic freedom and economic growth is entre-preneurial activity. Economic freedoms generate growth primarilybecause they promote underlying productive private-sector entrepre-neurial activity.

CATO JOURNAL

614

ReferencesBaumol, W. J. (1990) “Entrepreneurship: Productive, Unproductive and De-

structive.” Journal of Political Economy 98 (5): 893–921.(2002) The Free-Market Innovation Machine: Analyzing the

Growth Miracle of Capitalism. Princeton, N.J.: Princeton University Press.Bishop, R. V. (1979) “The Construction and Use of Casualty Tests.” Agri-

cultural Economic Research 31: 1–6.Blanchflower, D. G., and Oswald, A. J. (1998) “What Makes an Entrepre-

neur?” Journal of Labor Economics 16 (1): 26–60.Blau, D. M. (1987) “A Time-Series Analysis of Self-Employment in the

United States.” Journal of Political Economy 95 (3): 445–67.Blomstrom, M.; Lipsey, R. E., and Zejan, M. (1996) “Is Fixed Investment the

Key to Economic Growth?” Quarterly Journal of Economics 111 (1): 269–76.

Boettke, P. J. (2001) Calculation and Coordination: Essays on Socialism andTransitional Political Economy. New York: Routledge.

Boettke, P. J., and Coyne, C. J. (2003) “Entrepreneurship and Development:Cause or Consequence?” Advances in Austrian Economics 6: 67–87.

Bruce, D. (2000) “Effects of the United States Tax System on Transitionsinto Self-Employment.” Labour Economics 7 (5): 545–74.

(2002) “Taxes and Entrepreneurial Endurance: Evidence from theSelf-Employed.” National Tax Journal 55 (1): 5–24.

Corporation for Enterprise Development (2001) Infrastructure for Entre-preneurial Development: A Scan of Ten Southern States. Battle Creek,Mich.: Mary Reynolds Babcock Foundation (April).

Cole, J. H. (2003) “The Contribution of Economic Freedom to World Eco-nomic Growth, 1980–99.” Cato Journal 23 (2): 189–98.

Dorn, J. A. (2002) “Economic Development and Freedom: The Legacy ofPeter Bauer.” Cato Journal 22 (2): 355–71.

Farr, W. K.; Lord, R. A.; and Wolfenbarger, J. L. (1998) “Economic Free-dom, Political Freedom, and Economic Well-Being: A Causality Analysis.”Cato Journal 18 (2): 247–62.

Florida, R. (2002) The Rise of the Creative Class and How It’s TransformingWork Leisure, Community and Everyday Life. New York: Basic Books.

Granger, C. W. J. (1969) “Investigating Causal Relations by EconometricModels and Cross-Spectral Methods.” Econometrica 37 (3): 424–38.

Griliches, Z. (1990) “Patent Statistics as Economic Indicators: A Survey.”Journal of Economic Literature 28 (4): 1661–1707.

Gwartney, J. D.; Lawson, R. A.; and Holcombe, R. G. (1999) “EconomicFreedom and the Environment for Economic Growth.” Journal of Insti-tutional and Theoretical Economics 155 (4): 1–21.

Gwartney, J. D., and Lawson, R. A. (2002) Economic Freedom of the World:2002 Annual Report. Vancouver, B.C.: Fraser Institute.

Henderson, J. (2002) “Building the Rural Economy with High-Growth En-trepreneurs.” Federal Reserve Bank of Kansas City Economic Review 87(3): 45–70.

Holtz-Eakin, D. (1999) “The Death Tax: Investments, Employment, andEntrepreneurs.” Tax Notes 84 (5): 782–91.


615

Holtz-Eakin, D.; Joulfaian, D.; and Rosen, H. S. (1994) “Sticking It Out:Entrepreneurial Survival and Liquidity Constraints.” Journal of PoliticalEconomy 102 (1): 53–75.

Jenner, R. A. (1998) “Endogenous Schumpeterian Growth, The ProductivitySlowdown, and Entrepreneurial Dynamics.” Small Business Economics 11(4): 343–51.

Kaiser, C. P. (1990) “Entrepreneurship and Resource Allocation.” EasternEconomic Journal 16 (1): 9–20.

Karabegovic, A.; McMahon, F.; and Samida, D. (2002) Economic Freedom ofNorth America. Vancouver, B.C. Fraser Institute and the National Centerfor Policy Analysis.

Kirzner, I. M. (1997) “Entrepreneurial Discovery and the Competitive Mar-ket Process: An Austrian Approach.” Journal of Economic Literature 35(1): 60–85.

Knight, F. H. (1921) Risk, Uncertainty, and Profit. Boston: Houghton Mif-flin.

Lee, D. R. (1991) “The Seeds of Entrepreneurship.” Journal of PrivateEnterprise 7 (1): 20–35.

Minniti, M. (1999) “Entrepreneurial Activity and Economic Growth.” GlobalBusiness and Economics Review 1 (1): 31–42.

Powell, B. (2003) “Economic Freedom and Growth: The Case of the CelticTiger.” Cato Journal 22 (3): 431–48.

Reynolds, P. D.; Hay, M., and Camp, S. M. (1999) Global EntrepreneurshipMonitor. Kansas City, Mo.: Kauffman Center for Entrepreneurial Lead-ership.

Schumpeter, J. A. (1942) Capitalism, Socialism, and Democracy. New York:Harper.

Sims, C. A. (1972) “Money, Income, and Causality.” American EconomicReview 62 (4): 540–552.

Tures, J. A. (2003) “Economic Freedom and Conflict Reduction: Evidencefrom the 1970s, 1980s, and 1990s.” Cato Journal 22 (3): 533–42.

Zacharakis, A. L.; Bygrave, W. D.; and Shepherd, D. A. (2000) Global En-trepreneurship Monitor: National Entrepreneurship Assessment: UnitedStates of America. Kansas City, Mo.: Kauffman Center for EntrepreneurialLeadership.

CATO JOURNAL

616

public policy, entrepreneurship, and economic freedom

Documents