low innovationone
TRANSCRIPT
-
8/12/2019 Low Innovationone
1/16
Innovation and Entrepreneurship : Framingthe Questions in Rural Space
Sarah A. Low
Regional Economics and Public Policy
University of Illinois
The Importance of Innovation andEntrepreneurship
Innovation drives the economy, the
nnova or s e en repreneur c umpe er
1911)
Ample work on entrepreneurships
contribution to national growth
2
-
8/12/2019 Low Innovationone
2/16
The Importance of Innovation and
Entrepreneurship Entrepreneurs are innovative users of
reg ona asse s
What is the link between innovation and
entrepreneurship in rural America?
3
Innovation + EshipProsperity?Agglomeration + LuckProsperity???
Relevant Literature
Schumpeter (1911) predicts entrepreneurs
w ea o econom c grow
Larger firms replacing smaller, has
negative effect on innovation and
entrepreneurship (Schumpeter 1942)
4
urban areas-thicker markets, lower fixed
costs (Shane 2003)
-
8/12/2019 Low Innovationone
3/16
Relevant Literature
Acs: 2 competing theories Diversity; density not necessary
,
innovation and eship
Knowledge Spillover Theory of Entrepreneurship:
eship provides mechanism for knowledge spillovers
5
Focus on supply side: aligns with Schumpeter,regions, infrastructure, and knowledge
Relevant Literature Eship Supply (knowledge, opportunity) and
Demand (backward linkages)
2005)
Which determines rate of eship will vary regionally
Rural eship and innovation not studied as much
6
Strumsky 2006)
Our measure of eship has positive impact on factors
of production in the rural US (Thompsen et al. 2006)
-
8/12/2019 Low Innovationone
4/16
Framing the Question
If we adopt the agglomeration idea Eship and innov are higher in urban areas
controls for population
Supply side factors needed for entrepreneurship
to create growth
Market opportunities-thick markets, people
7
-
Next best occupation -opportunity cost of self-
employment
Infrastructure, amenities-is this a nice place to be and
do business?
Research Questions
Does supply side reasoning hold in rural
p aces
Is agglomeration necessary for eship and
innovation to drive growth?
How do eship and innovation relate to
8
-
8/12/2019 Low Innovationone
5/16
Measures of Entrepreneurship
Breadth calculated as:
nonfarm proprietors nonfarm employment
Proprietor Value calculated as:
nonfarm proprietor income nonfarm total
income
9
Entrepreneurship Breadth
10
< 20% (Mean)
20-27.5% (Mean Mean + 1 St Dev)
>27.5%
-
8/12/2019 Low Innovationone
6/16
Proprietor Value
11
< 13% (Mean)
13-17.5% (Mean Mean + 1 St Dev)
>17.5%
Measure of Innovation
Innovation calculated as:
Patents (1990) per capita
Not a good measure of innovation
Wei ht number of atents for their im ortance?
12
ie- incremental patents v. radical patents, #
citations
-
8/12/2019 Low Innovationone
7/16
Innovation
13
Average
Average + 1 StDev
> 1 StDev above Ave
< .00036 (Mean)
.00063-.0036 (Mean Mean + 1 St Dev)
>.0036
Hypotheses
H1: Innovation and entrepreneurship drive
rura emp oymen grow
H2: Combining innovation and eship is the
best way to produce rural employment
growth
14
-
8/12/2019 Low Innovationone
8/16
Hypotheses
H1: Innovation and Entrepreneurship driverural employment growth
s agg omerat on necessary or nnov an
eship to generate growth?
15
Entrepreneurship and Population.4
Eship by
Agglomeration
Eship
.1
.2
. necess ty
1.5 million
16
0
0 1.00e+07 2.00e+07 3.00e+07 4.00e+07pop05
Predicted fit for Breadth05 based on County Population
-
8/12/2019 Low Innovationone
9/16
Entrepreneurship and Population
.2
5
.1
.15
.2
Fitted
values
3.5 Million
17Predicted Fit for PropValue05 based on Pop05
.0
5
0 2000000 4000000 6000000 8000000 1.00e+07pop05
Hypotheses H2: Combining innovation and eship is the
best way to produce rural employment
Should rural policy focus on growing the
Entrepreneur
Innovator
Or both?
18
-
8/12/2019 Low Innovationone
10/16
Patents per Capita and Population
5
.002
.0005
.001
.001
Fitted
values
Patents have pos itive
relationship with
population
19Predicted Fit for PatPC99 based on Pop05
0
0 2000000 4000000 6000000 8000000 1.00e+07pop05
Eship and Patents per Capita
5
.0006
.0003
.0004
.000
Fitted
values
Interesting because
Eship Breadth is highest
in rural areas
20Predicted Fit for PatPC based on Breadth05, zero patents dropped
.0002
0 .2 .4 .6nf proprietor employment over nf employment, reis, 2005
-
8/12/2019 Low Innovationone
11/16
Interaction Variables & Growth
Breadth*Innov:
Proprietor Value*Innov: Each is normalized to one, then multiplied
Growth measured as employment growth
over 1991-2001 business cycle
21
Expected Results H1 : Growth = (+) innov or (+) eship
H2: Eship/Innov interaction (+)
relationship with growth
22
-
8/12/2019 Low Innovationone
12/16
Control Variables
Agglomeration Population
Recast Creative Class, McGranahan and Wojan 2007
Natural amenities Distance to National Park
Financial capital
23
Bank Deposits per Capita
Infrastructure None, difficulty with 1990 measure
Empirical Model Hausman test detects simultaneity inmodel
I.V. choice is problematic
Also: how to control for spatial dependency
within 2SLS framework
24
-
8/12/2019 Low Innovationone
13/16
Spatial Dependency
Controlled for in all counties model
Queen contiguity weights matrix
LM error and LM lag tests significant
Robust LM-lag test significant for all models
We use spatial lag model
No ood solution for rural counties
25
H 1: Growth=Eship or Innov
Rural All-OLS All-LAG
Breadth + + +
PropValue - +
PatPC
26
R-square .20 to .14 .20 to .14 .22 to .16
N 2009 3105 3050
-
8/12/2019 Low Innovationone
14/16
H2:Growth=Breadth, PatPC, Interaction
Rural
Rural-
2SLS
All-
OLS
All-
2SLS
All-
LAG
PatPC + + +
Breadth + + + + +
Breadth
* + + +
27
R-square .20 .19 0.21 0.22
N 2006 2006 3046 3046 3105
H2:Growth=PropVal, PatPC, Interaction
Rural- All- All- All-
PatPC + +
PropValue + + +
PropValue
28
*PatPC - - -
R-square 0.14 0.15 0.16
N 2006 2006 3046 3046 3105
-
8/12/2019 Low Innovationone
15/16
Conclusion
Research Question: Do rural places with
g er es p an nnova on ave g er
growth?
Positive relationship exits
Eship Growth
Results of interaction inconclusive
29
Results dependent upon definition of eship
and innovation
Conclusion
Little difference b/w all & rural counties
Agglomeration not necessary to create
employment growth from entrepreneurship
Human Capital is strongest (+) control variable
Suggests that knowledge is important, not
necessarily agglomeration
30
-
8/12/2019 Low Innovationone
16/16
Weaknesses of Analysis
Methodological difficulties
Endogeneity
Instrumental Variables
Measures of Innovation
Not capturing what Id hope it would
31