Transcript PowerPoint

The Impact of Patent Protection on
Outsourcing Decisions by U.S.
Manufacturing Firms
Mike Palmedo
Presented at
URPE Panel on Financialization, R&D, Patents and Development
ASSA Annual Meeting | Chicago, IL | January 2017
Many Agree: Stronger Intellectual Property
Protection = More Jobs!
“...if China protected intellectual property as
the U.S. does, there would be approximately
923,000 new U.S. jobs”
- Richard Trumka, AFL-CIO
"Sound IP policies and enforcement of IP rights
abroad are essential to advancing U.S. economic
recovery, driving America’s competitiveness and
export growth, and creating high-quality, high-paying
American jobs"
- David Hirschmann, U.S. Chamber of Commerce
Employment in IP-intensive and Non-IPIntensive Industries
Source: U.S. Department of Commerce. Intellectual Property in the U.S. Economy: Industries in Focus. March, 2012.
Hypothesis:
U.S. firms between 1997 and 2010 were more likely to
offshore production of patent-intensive intermediate goods to
countries with stronger patent laws
Continuum of Inputs with Different Levels of
IP Intensity
Inputs (z)
→ → IP Intensity → →
Expected Cost of Production Depends Partially
on Probability of IP Theft
Home:
E[c(z)] = βX + Pr(THEFTz) ● V(IPz)+ ε
Foreign:
E[c(z)]* = βX* + Pr(THEFTz*) ● V(IPz*)+ ε
Foreign Versus Home Production of Inputs
E[c(z)]* /
E[c(z)]
Relative Cost
1
Foreign
Production
Home
Production
z’
→ → IP Intensity → →
Inputs (z)
Foreign Versus Home Production of Inputs
E[c(z)]* /
E[c(z)]
Relative Cost
Relative Cost ‘
1
Home
Production
Foreign
Production
z’
z’’
→ → IP Intensity → →
Inputs (z)
Empirical Test: Data Overview
Dependent Variable:
• Estimated U.S. imports of intermediate goods
Independent Variables:
• Ginarte-Park Patent Index
• Wages
• Cost of capital
• GDP
• Distance
• World Institutional Quality Index
0
.5
Test Variable: Ginarte-Park Patent Index
0
1
2
3
4
5
Patent Index
kdensity patent_index_1995
kdensity patent_index_2005
kdensity patent_index_2000
kdensity patent_index_2010
Dependent Variable
Estimated value of imported patent-intensive commodities that are used
as intermediate goods from various trading partners
• “Patent intensive commodities” are those produced by industries with
high ratios of patents to employees (Dept. of Commerce, 2012)
• BEA reports percentage of imports from each industry that are used as
intermediate inputs. ITC reports total imports from each trading partner
• Proportionality assumption: “within each sector imports from each
source country are split between final and intermediate in proportion to
the overall split of imports between final and intermediate use.”
(Johnson and Noguera, 2012).
Total U.S. Employment in NAICS 333, 334, 335
1,800,000
NAICS 333: Machinery
Manufacturing
1,600,000
1,400,000
NAICS 334: Computer &
Electronics Mfg.
1,200,000
1,000,000
NAICS 335: Elec. Equip.,
Appliance, & Components
800,000
600,000
400,000
200,000
Source: U.S. Census Bureau, Statistics of U.S. Businesses
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
0
Descriptive Statistics for Logged Dependent
Variable
Industry
Obs.
Mean
St. Dev.
Machinery (333)
429
7.81
4.27
Computer and electronic products (334)
456
8.13
4.46
Electrical equipment, appliances, and components (335)
393
7.83
4.08
Results: OLS Regressions with Fixed Effects
Dep. Variable:
Estimated
imports of
intermediate
goods
VARIABLES
patent
wage
rent
gdp
distance
(1)
(2)
(3)
(4)
0.677**
(0.275)
0.264**
(0.116)
-1.227***
(0.373)
1.272***
(0.0699)
-0.773***
(0.212)
0.430**
(0.174)
0.333
(0.368)
0.334
(0.590)
0.341
(0.420)
-0.730
(0.797)
-17.94***
(2.922)
7.354
(7.391)
0.458*
(0.245)
0.0622
(0.176)
-0.869**
(0.405)
1.335***
(0.0734)
-0.731***
(0.208)
0.600**
(0.270)
-18.44***
(2.772)
0.426**
(0.173)
0.334
(0.372)
0.326
(0.598)
0.355
(0.422)
-0.778
(0.857)
-0.0482
(0.394)
7.406
(7.591)
672
0.702
Yes
No
672
0.937
Yes
Yes
669
0.706
Yes
No
669
0.937
Yes
Yes
institution
Constant
Observations
R-squared
Time and industry F.E.
Country F.E.
Results: Separate Regressions for Each Industry
Dep. Variable:
Estimated
imports of
intermediate
goods
VARIABLES
patent
wage
rent
gdp
distance
institution
Constant
Observations
R-squared
Time F.E.
Country F.E.
(1)
NAICS 333
(2)
NAICS 334
(3)
NAICS 335
0.488**
(0.197)
-0.539
(0.343)
-0.351
(0.643)
0.824**
(0.398)
-2.359***
(0.817)
0.138
(0.408)
14.74**
(6.563)
0.404**
(0.179)
0.837**
(0.337)
0.557
(0.732)
0.432
(0.384)
0.0979
(0.822)
-0.0166
(0.373)
-5.988
(9.867)
0.644***
(0.176)
0.892
(0.615)
0.542
(1.078)
-0.538
(0.613)
0.774
(1.163)
-0.138
(0.484)
13.68
(8.991)
224
0.983
Yes
Yes
227
0.985
Yes
Yes
218
0.973
Yes
Yes
Results: Panel Regressions
Dep. variable =
estimated imports of
intermediate goods
Panel variable =
Industry/country
groups
VARIABLES
Patent
wage
rent
gdp
(1)
Wage
Rent
Gravity
(2)
Wage
Rent
Gravity
Institution
0.620***
(0.109)
0.360
(0.283)
0.159
(0.485)
0.0590
(0.254)
4.844
(4.227)
0.617***
(0.110)
0.369
(0.294)
0.146
(0.478)
0.0667
(0.253)
-0.0789
(0.225)
4.606
(4.201)
672
0.212
215
669
0.212
215
institution
Constant
Observations
Within-Entity R-squared
# Industry/country groups
Conclusion
• Controlling for variables related to factor costs, gravity model
determinants, and institutional quality, as well as time, industry, and
country fixed effects, an increase of one standard deviation in a
country’s score on the Ginarte-Park Patent Index was associated with
a 43% increase in U.S. imports of these intermediate goods from it.
• The degree to which patent protection affected these trade flows
differed by industry, as is predicted by previous literature.
• As individual countries adjusted the level of patent protection they
provided during this time, firms in those countries found themselves
shipping more intermediate goods from these industries to the U.S.
Thank you!
Mike Palmedo
[email protected]