The value of software-related patents in the European Patent System Salvatore Torrisi Department of...
-
Upload
kory-watkins -
Category
Documents
-
view
216 -
download
0
Transcript of The value of software-related patents in the European Patent System Salvatore Torrisi Department of...
The value of software-related The value of software-related patents in the European Patent patents in the European Patent SystemSystem
Salvatore Torrisi Salvatore Torrisi
Department of Management, Università di Bologna and CESPRI-Department of Management, Università di Bologna and CESPRI-Bocconi Bocconi [email protected]@unibo.it
EUPACO, What Future for the European Patent System? 15&15 May EUPACO, What Future for the European Patent System? 15&15 May 20072007
May 2007May 2007 s. torrisis. torrisi 22
goalsgoals
• What is the value of EPO patents?What is the value of EPO patents?
• How to find “software-related” How to find “software-related” patents in the EPO dataset?patents in the EPO dataset?
• what is the value of “software-what is the value of “software-related” patents in Europe? related” patents in Europe?
May 2007May 2007 s. torrisis. torrisi 33
ResearchResearch
• This presentation is based on Thoma and This presentation is based on Thoma and Torrisi ( 2006) and Hall-Thoma-Torrisi (2006)Torrisi ( 2006) and Hall-Thoma-Torrisi (2006)
• See See www.cespri.unibocconi.itwww.cespri.unibocconi.it/working/working paperspapers
• Research framework: “Study of the effects of Research framework: “Study of the effects of allowing patent claims for computer-allowing patent claims for computer-implemented inventions” sponsored by the implemented inventions” sponsored by the European CommissionEuropean Commission
• DisclaimersDisclaimers:: The opinions expressed are those of the speaker The opinions expressed are those of the speaker and do not necessarily reflect in any way opinions of the European and do not necessarily reflect in any way opinions of the European Commission or any of the partners Commission or any of the partners
May 2007May 2007 s. torrisis. torrisi 44
MotivationsMotivations• Traditionally, patents are a weak Traditionally, patents are a weak
instrument to protect innovation in instrument to protect innovation in many sectors (many sectors (Cohen, Nelson and Cohen, Nelson and Walsh, 2000)Walsh, 2000)
• … … why, then, so many patents why, then, so many patents around? around?
• Increased tech opportunities? Stronger Increased tech opportunities? Stronger enforceability? Lower barriers to enforceability? Lower barriers to patent?patent?
• What is their value to the owners? What is their value to the owners?
May 2007May 2007 s. torrisis. torrisi 55
MotivationsMotivations
• What about “software” patents?What about “software” patents?
• EPC (art. 52): computer programs “as EPC (art. 52): computer programs “as such” not patentable but …EC has such” not patentable but …EC has been considering standards for been considering standards for “computer implemented inventions” “computer implemented inventions”
May 2007May 2007 s. torrisis. torrisi 66
1. How to find “software” 1. How to find “software” patents?patents?1.1. Keywords method (Bessen-Hunt 2004)Keywords method (Bessen-Hunt 2004)
– Search for Search for ((software) OR (computer AND ((software) OR (computer AND program)) AND NOT (chip OR semiconductor program)) AND NOT (chip OR semiconductor OR bus OR circuit OR circuitry <in> TI) AND OR bus OR circuit OR circuitry <in> TI) AND NOT (antigen OR antigenic OR chromatogra NOT (antigen OR antigenic OR chromatogra phy)phy) in the patent document (title and full text) in the patent document (title and full text)
2.2. Patent class method (Graham-Mowery, Hall-Patent class method (Graham-Mowery, Hall-MacGarvie)MacGarvie)
– IPC classes in the patent portfolios of the IPC classes in the patent portfolios of the world’s largest specialized software firms world’s largest specialized software firms
– 3,518 classes-subclasses (117 if only the 3,518 classes-subclasses (117 if only the main IPC codes in each patent are considered). main IPC codes in each patent are considered).
3.3. Hall,Thoma and Torrisi (2006): a combination of Hall,Thoma and Torrisi (2006): a combination of these two methods these two methods
May 2007May 2007 s. torrisis. torrisi 77
Cumulative software patents granted by EPO
0
3000
6000
9000
12000
15000
18000
1978 1982 1986 1990 1994 1998 2002
Year of application
Num
ber
IPC method BH method
Cumulative number of “Cumulative number of “software”software” patents granted by the EPO by year of patents granted by the EPO by year of applicationapplication
May 2007May 2007 s. torrisis. torrisi 88
Concentration of “software” Concentration of “software” patentspatents
0%
20%
40%
60%
80%
100%
0% 20% 40% 60% 80% 100%
cumulative share of assignees
cu
mu
lative
sh
are
of p
ate
nts
keywordsmethod
IPCmethod
May 2007May 2007 s. torrisis. torrisi 99
““software” patents as a software” patents as a share of all patents: USPTO share of all patents: USPTO vs. EPOvs. EPO
0%
2%
4%
6%
8%
10%
12%
14%
16%
1981 1984 1987 1990 1993 1996 1999 2002 Year
% o
f all
pate
nts
EPO - keyword method
EPO - IPC method
USPTO – keyword method
USPTO - IPC method
USPTO - HM method
May 2007May 2007 s. torrisis. torrisi 1010
2. What is the value of ‘software-2. What is the value of ‘software-related’ patents in Europe?related’ patents in Europe?
• Publicly-traded firms from 17 Publicly-traded firms from 17 European countries (1984-2002)European countries (1984-2002)
• Data sources: Data sources: – EPO PATSTAT, DelphionEPO PATSTAT, Delphion– Bureau van Djik’s Amadeus, Hoovers, Bureau van Djik’s Amadeus, Hoovers,
D&B’s WOW, Thomson Financial’s D&B’s WOW, Thomson Financial’s Datastream Datastream
May 2007May 2007 s. torrisis. torrisi 1111
R&D reporting firms by sector R&D reporting firms by sector (N=2,197)(N=2,197)
2,5 digit industry class Companies % Companies % Companies %
01 Food & tobacco 87 3.96 51 4.11 13 3.27
02 Textiles, apparel & footwear 45 2.05 25 2.01 5 1.26
03 Lumber & wood products 10 0.46 4 0.32 1 0.25
04 Furniture 21 0.96 15 1.21 2 0.5
05 Paper & paper products 30 1.37 20 1.61 8 2.02
06 Printing & publishing 27 1.23 11 0.89 6 1.51
07 Chemical products 92 4.19 66 5.31 22 5.54
08 Petroleum refining & prods 38 1.73 23 1.85 14 3.53
09 Plastics & rubber prods 38 1.73 27 2.17 4 1.01
10 Stone, clay & glass 47 2.14 35 2.82 8 2.02
11 Primary metal products 55 2.5 36 2.9 12 3.02
12 Fabricated metal products 60 2.73 39 3.14 8 2.02
13 Machinery & engines 171 7.78 135 10.87 37 9.32
14 Computers & comp, equip, 50 2.28 28 2.25 14 3.53
15 Electrical machinery 78 3.55 58 4.67 13 3.27
16 Electronic inst, & comm, eq, 224 10.2 128 10.31 48 12.09
17 Transportation equipment 18 0.82 12 0.97 8 2.02
18 Motor vehicles 53 2.41 36 2.9 16 4.03
19 Optical & medical instruments 75 3.41 56 4.51 19 4.79
20 Pharmaceuticals 131 5.96 94 7.57 29 7.3
21 Misc, manufacturing 37 1.68 19 1.53 2 0.5
22 Soap & toiletries 17 0.77 12 0.97 3 0.76
24 Computing software 326 14.84 81 6.52 34 8.56
25 Telecommunications 48 2.18 20 1.61 15 3.78
26 Wholesale trade 53 2.41 24 1.93 5 1.26
27 Business services 50 2.28 13 1.05 5 1.26
28 Agriculture 3 0.14 0 0 0 0
29 Mining 29 1.32 17 1.37 2 0.5
30 Construction 42 1.91 27 2.17 5 1.26
31 Transportation services 17 0.77 11 0.89 5 1.26
32 Utilities 58 2.64 35 2.82 11 2.77
33 Trade 23 1.05 8 0.64 1 0.25
34 Fire, Insurance, Real Estate 27 1.23 10 0.81 2 0.5
35 Health services 9 0.41 4 0.32 2 0.5
36 Engineering services 85 3.87 51 4.11 13 3.27
37 Other services 23 1.05 11 0.89 5 1.26
Overall 2197 100 1.242 100 397 100
with R&D with EPO patents with "software" pats
May 2007May 2007 s. torrisis. torrisi 1212
Estimating the economic value of Estimating the economic value of patentspatents
• The sample: 1,046 firms, 5,444 obs The sample: 1,046 firms, 5,444 obs
• Our dependent variable: Our dependent variable: – Tobin’s q = Market value (M)/tangible assets (A)Tobin’s q = Market value (M)/tangible assets (A)
• main explanatory variables:main explanatory variables:– R&D stock/AR&D stock/A– Pat stock/R&DPat stock/R&D– SWPat stock/R&D stockSWPat stock/R&D stock– Quality: cites and Lanjouw & Schankerman Quality: cites and Lanjouw & Schankerman
‘composite’ index: forward cites, family size, ‘composite’ index: forward cites, family size, No. 8-digit IPC classesNo. 8-digit IPC classes
May 2007May 2007 s. torrisis. torrisi 1313
Variable model 5 model 6 model 7 model 8R&D stock/assets 0.762 (0.079) *** 0.785 (0.094) *** 0.669 (0.066) *** 0.634 (0.068) ***Pat stock/R&D stock 0.153 (0.034) *** 0.152 (0.038) *** 0.131 (0.029) *** 0.116 (0.029) ***D (no patents) 0.174 (0.054) *** 0.171 (0.056) *** -0.013 (0.038) -0.010 (0.036)SW pat stock/R&D stock 0.169 (0.499) 0.221 (0.522)D (no sw patents) 0.051 (0.055) -0.045 (0.032)Cit stock/Pat stock 0.041 (0.008) *** 0.039 (0.008) ***SW cit stock/ SW pat stock 0.013 (0.006) **Index stock/Pat stock 0.313 (0.062) *** 0.310 (0.061) ***SW index stock/ SW pat stock -0.072 (0.058)log (sales) 0.022 (0.010) ** 0.021 (0.012) * 0.020 (0.008) ** 0.016 (0.009) ***D (sales missing) 5.873 (1.579) *** 6.138 (1.703) *** 5.169 (1.313) *** 4.907 (1.268) ***
Adjusted r-squared 0.268 0.264 0.264 0.265Standard error 0.719 0.738 0.721 0.721
R&D stock/assets 0.425 (0.054) 0.438 (0.062) 0.323 (0.040) 0.314 (0.042)Pat stock/R&D stock 0.085 (0.019) 0.085 (0.022) 0.063 (0.014) 0.057 (0.015)SW pat stock/R&D stock 0.095 (0.279) 0.109 (0.259)Cit stock/Pat stock 0.022 (0.004) 0.022 (0.005)SW cit stock/ SW pat stock 0.007 (0.003)Index stock/Pat stock 0.169 (0.035) 0.173 (0.037)SW index stock/ SW pat stock -0.040 (0.033)
All equations include country dummies, year dummies, and sector dummies.
Elasticities from nonlinear least squares
TableLog Q regressions 1984-2002
(5444 observations, 1,046 firms, 17 countries)
Nonlinear least squares (robust s.e.s)
May 2007May 2007 s. torrisis. torrisi 1414
Main resultsMain resultsvariablevariable Effect on M/AEffect on M/A commentcomment
R&D stock/AR&D stock/A ++ +++ + < than in USPTO< than in USPTO
Pat stock/RDstkPat stock/RDstk + ++ + > than in USPTO> than in USPTO
Cites/Pat stockCites/Pat stock ++ from USfrom US
Swpatstk/RDstkSwpatstk/RDstk Not significantNot significant H-M (2006):+/signH-M (2006):+/sign
Cites/SwpatstkCites/Swpatstk Not significantNot significant H-M (2006): sign. x H-M (2006): sign. x specialised sw firmsspecialised sw firms
LS composite LS composite index (software)index (software)
++ (software: ++ (software: not sign)not sign)
Total patents: similar Total patents: similar to LS(2004) on US datato LS(2004) on US data
May 2007May 2007 s. torrisis. torrisi 1515
Conclusions (1) Conclusions (1) • EPO patents worth more to European firms EPO patents worth more to European firms
than US patents to US firmsthan US patents to US firms• But citation-weighted EPO patents worth But citation-weighted EPO patents worth
less than citation-weighted US patents. less than citation-weighted US patents. • Why?Why?
– Are EPO patents different? Are European firms Are EPO patents different? Are European firms different?different?
– in the EPO system cites are assigned by in the EPO system cites are assigned by examiners … upper bound to the number of examiners … upper bound to the number of citationscitations
• Future research: US patents and EPO Future research: US patents and EPO patents for the same sample of firms patents for the same sample of firms
May 2007May 2007 s. torrisis. torrisi 1616
Conclusions (2)Conclusions (2)
• LS index: Investors have information LS index: Investors have information to distinguish patents with a to distinguish patents with a consistent set of characteristics consistent set of characteristics (forward cites, family size and tech (forward cites, family size and tech fields)fields)
May 2007May 2007 s. torrisis. torrisi 1717
Conclusions (3)Conclusions (3)
• Why are software patents not valued by Why are software patents not valued by the market? (the private value)the market? (the private value)
• few observations and high concentration few observations and high concentration of EPO software patsof EPO software pats
• Limited enforceability (normative Limited enforceability (normative framework)framework)
• Investors see these as ‘strategic’ patents Investors see these as ‘strategic’ patents in a field with highly cumulative, in a field with highly cumulative, sequential change sequential change litigations, litigations, transaction costs … transaction costs …
May 2007May 2007 s. torrisis. torrisi 1818
What about the social value of What about the social value of software patents?software patents?• Sequential, cumulative changeSequential, cumulative change• Relatively large number of claims/patentRelatively large number of claims/patent• Lack of documented prior artLack of documented prior art• IP protection reinforces increasing returns IP protection reinforces increasing returns
(network externalities)(network externalities)• The low (private and social?) value of sw The low (private and social?) value of sw
patents suggests: patents suggests: – reduce the scope of patents (n. of claims)reduce the scope of patents (n. of claims)– Allow a limited right to reverse engineering Allow a limited right to reverse engineering – More aggressive compulsory licensingMore aggressive compulsory licensing– …………..
May 2007May 2007 s. torrisis. torrisi 1919
thanksthanks
May 2007May 2007 s. torrisis. torrisi 2020
Appendix: main variables and Appendix: main variables and resultsresults
• Market value (Market value (VV): market cap (equity) plus current ): market cap (equity) plus current and non current liabilities less current assets plus and non current liabilities less current assets plus inventoriesinventories
• Tangible assets (Tangible assets (AA): gross fixed assets and ): gross fixed assets and inventories less [depreciation, depletion, and inventories less [depreciation, depletion, and amortization (accumulated), and other amortization (accumulated), and other deductions]deductions]
• R&D and patent stocks:R&D and patent stocks:
KKt t = I= Itt+(1-+(1-) I) It-1 t-1 KK0 0 = I= I00/(/(+g) +g)
= depreciation rate = depreciation rate gg = constant growth rate= constant growth rate
May 2007May 2007 s. torrisis. torrisi 2121
Variable N valid Mean s.d. Median 1Q 3Q Min MaxSales* 5437 3804.19 11932.92 321.63 68.6 2070.62 0.01 190000Tobin's q 5444 2.95 3.51 1.69 1.13 3.11 0.10 24.85
R&D expenditure* 5206 139.07 502.11 9.22 2.13 42.91 0.00 6786.63R&D stock* 5206 684.54 2473.86 39.23 9.37 209.95 0.03 33127.46R&D stock/assets 5444 0.50 0.72 0.25 0.09 0.58 0.00 4.99
Patents this year 4070 20.51 79.66 2.00 0.00 9.00 0.00 1448.00Patent stock 4070 115.97 431.67 10.60 2.29 55.46 0.05 7056.78Pat stock/R&D stock* 4070 0.44 1.19 0.16 0.06 0.39 0.00 19.82
SW patents this year 1469 2.10 8.34 0.00 0.00 1.00 0.00 112.00SW patent stock 1469 10.57 38.44 1.41 0.61 4.84 0.07 473.48SW pat stock/R&D stock* 1469 0.03 0.23 0.00 0.00 0.02 0.00 8.19
Citations this year 3925 47.09 148.65 6.36 0.00 36.00 0.00 2086.00Citation stock 3925 310.25 918.59 54.02 11.49 236.81 0.04 10350.85Cit stock/Pat stock 3925 4.61 3.73 3.68 2.54 5.43 0.11 49.01
SW citations this year 1352 6.79 20.76 0.00 0.00 4.38 0.00 227.00SW citation stock 1352 38.08 96.51 9.38 3.13 29.98 0.10 856.32SW cit stock/SW pat stock 1352 6.22 4.87 4.68 2.90 8.30 0.23 36.02
Index this year 1933 2.16 6.46 0.02 0.00 1.30 -7.78 65.79Index stock 1933 11.67 30.58 1.50 0.35 6.79 0.00 272.81Index stock/Pat stock 1933 0.21 0.21 0.15 0.07 0.28 0.00 2.50
SW index this year 794 0.28 1.23 0.00 0.00 0.03 -5.26 15.85SW index stock 794 1.38 3.67 0.44 0.15 1.15 0.00 42.26SW index stock/SW pat stock 794 0.35 0.29 0.31 0.13 0.46 0.00 2.41*In millions of current euros
Sample statistics(5,444 observations, 1,046 firms, 17 countries)
May 2007May 2007 s. torrisis. torrisi 2222
Distribution of “Distribution of “software”software” patents patents by sectorby sector
2 digit2 digitNACE NACE CodesCodes
2-digit 2-digit USUSSICSIC
DescriptionDescriptionNumber Number
of of assigneesassignees
Number of Number of softwaresoftwarepatentspatents
Share of Share of all software all software
patentspatents
5151 3535Industrial machinery and Industrial machinery and
equipment equipment 267267 25442544 20.4%20.4%
5151 3636Electrical and electronic Electrical and electronic
equipment equipment 185185 24312431 19.5%19.5%
7272 7373Business services incl. Business services incl.
software and IT software and IT 271271 11901190 9.5%9.5%
3333 3838Instruments and related Instruments and related
products products 174174 795795 6.4%6.4%
6464 4848Telecommunication Telecommunication
services services 4646 790790 6.3%6.3%
Other sectorsOther sectors 887887 47244724 37.9%37.9%
TotalTotal 18301830 1247412474 100%100%
Share overall datasetShare overall dataset 38%38% 54%54%
Assignees are classified according to their primary industry code (e.g., SIC code 35 for IBM)
Source: Our computations using Delphion and CESPRI data.
May 2007May 2007 s. torrisis. torrisi 2323
Cumulative software patents granted by EPO
0
3000
6000
9000
12000
15000
18000
1980 1984 1988 1992 1996 2000 2004
Year of publication
Nu
mb
er
IPC method BH method