38E00100 Economics and Management of Intellectual Property Lecture 13, Part IX “Innovation and IP...

38E00100Economics and Management of Intellectual Property

Lecture 13, Part IX

“Innovation and IP in Financial Service Sector continues”

Tuomas Takalo, 26.2.www.takalo.net

Part I. Basic IP Law (V)

Part II. Use of IPRs (V)

Part III. Basic Economics of IP (T)

Part IV. Breadth and Duration of IP and Their Optimal Design (T)

Part V. Cumulative Innovation and IP (T)

Part VI. Competition Policy and IP (T+V)

Part VII. Litigation (Cases) (V)

Part VIII. Basic IP Corporate Strategy (T)

Part IX. Industry Studies 1. Financial Services (T)

Outline of Core Lectures

Financial services

Growth

Innovation in other sectors

Innovation in financial service sector

Recapping

Welfare effects of financial innovation

• Appropriating strategies in financial services traditionally:

• Patents were not available (in such a way as today)

• Secrecy does not work for product innovations

• Imitation of product innovations rapid

Lead time & marketing the best means of appropriation

• However, the things have changed in the U.S & Asia in the aftermath of the court decision in 1998 on State Street Bank vs. Sigature Financial Group that allowed for financial method patents

intense debate of pros and cons of patents

deep concern for deteriorating patent quality

• Europe?

– Note: patents have always been available for technological inventions in financial services sector, too

e.g. a USPTO patent, March 19, 1799, on an invention to detect counterfeit notes

– But now they are available (US, Asia) for abstract methods, mathematical formulas (or software that implements them)

• The entry of patents into Wall Street is a part of a broader story of how software/ business methods became patentable.

• The US Patent Law §101 says on patentability:

“Who ever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent thereof…”

• Prior to 1980 this was generally interpreted to mean that software and abstract ideas are not patentable

The US Case

• 1981, the Supreme Court decision (Diamond v. Diehr) upheld the validity of a patent on an improved process for making rubber goods that was assisted by computer software program

patent lawyers started to write patent claims so that software patent applications did not appear to be patents on software

• Attention to tangible innovations such as new industrial processes, not software

• Patent claims were also written in terms of hardware that included software

• This became standard and accepted practice but it also became increasingly difficult to hide software in the patent claims

• By the mid 1990 there was little uncertainty about the patentability of software but still there were some ambiguity over which fields software patents were valid

• 1998, the Court of Appeals for the Federal Circuit (CAFC) decision in State Street Bank and Trust v. Signature Financial Group

• The case originates from a software program used to determine the value of mutual funds, on which Signature had obtained a patent in 1993.

• State Street sued to have the patent invalidated on the grounds that it covered a business method and was hence not patentable

• District court considered the patent invalid but the CAFC did not. The Supreme court declined to hear State Street’s appeal

• The CAFC deliberately wanted to clarify the situation. §101 is unambiguous: Who ever invents or discovers any new and useful

Business methods should be patentable, a finance patent just happened to one the table then...

Surge of business methods and finance patents in the US (Figure)

US PTO Class 705

0

2000

4000

6000

8000

10000

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

Years Applications

Issues

• The patentability of software is part of broader “pro-patent” shift of the US policy

• CAFC was created in 1982. Its purpose was to unify the appellation procedure. A by product it has

• increased the probability that a patent is held valid

• reduced the patent scope?

• reduced non-obviousness standard

• increased preliminary injunctions

• increased damage awards and the success rates of patent holders in infringement cases

patents have become more cost effective surge in patenting in general (Figure)

USPTO Patents

050

100150200250300350400450

19

63

19

67

19

71

19

75

19

79

19

83

19

87

19

91

19

95

19

99

20

03

Years

Th

ou

sa

nd

s

Applications

Years

Where does State Street lead? (Lerner 2002)

• Lerner (2002) is a pioneering study on finance patents

• An empirical study of IPRs.

• Theory and empirics are complements.

• How to identify finance patents? In general two methods to assign patents to an industry, an easier and a more difficult

• “Automatized”, i.e., using the patent classification system developed by the POs

• “Hand-picking”, i.e., reading the patents and classifying them independently

• Lerner (2002) uses both approaches but hinges more on the easier (automatized) way

• How to identify financial innovations? cf. Lerner (2006)

• Patent classification systems developed to facilitate the search of prior art by the POs

• WIPO’s IPC classification and EPO’s ECLA

• USPTO’s UCLA

Patent examiners mainly search patents for prior art

problems in new patent subject matter areas where the prior art is not in patent files

• Lerner’s analysis limited to five subclasses of the USPTO class 705 “Data processing, financial, business practice, management, or cost/price determination” ≈ “Business methods”

• 705/35: Finance • 705/36: Portfolio selection, planning, or analysis• 705/35: Trading, matching, or bidding• 705/35: Credit or loan processing

• In addition, 705/4: Insurance, from which the finance patens are hand-picked

• Problems with the easier approach:

• patent classes are made so as to aid prior art search, not to aid social scientist they do not correspond economic concepts of industry, markets or technology

• the classification system changes over time

• patens are reclassified

• Strategic patent drafting to be classified in a favorable way

• because of the wide-spread concern for patent quality, the USPTO introduced a second pair of eyes review of patent applications in class 705

finance patents drafted to be classified in 902 “Electronic Funds Transfer”

• this happened 2000 so does not concern this study

• But because the easier way is, well, easier, it has been widely used and is of course a natural starting point

• According to this quite narrow definition of finance patens, 445 finance patents have been awarded 1971-2000

• Focus on finance innovations that are business methods: • a part of finance patents

• excludes many banking patents, payment media patents etc

• A small part of business method patents (figure)

• Some of the main patentees are Wall Street investment banks such as Merrill Lynch, Citigroup (Table 1)

• What explains the patenting behavior of the investment banks?

• A “patent production function”: Patents = f(R&D, controls)

• Scotchmer’s critique: ideas model vs. production function

• Patents/applications are taken as a proxy for innovation output and R&D as an input

• However, distribution of patent values highly skewed

• only a small fraction of patents is valuable,

• most are useless and are hence not kept in force for a long time (less than 50% are kept in force 10 years)

• a way to address this problem is to employ forward citations

• citations to the patent by subsequent patents

• patent applicants must cite the relevant prior art they are aware, if not the patent can be held invalid due to “fraud on the patent office”

• strong correlation between patent value and the number of citations

• each patent weighted by citations, i.e., one plus number of citations it has received as of 2000

• new problem: the patents had different periods to be cited. Ceteris paribus, older ones should receive more citations than new ones

“normalized” citation-based weight (W) for each patent

• where A = is the year of award, Ct is the actual number of citations received year t

• Pt is the predicted number of citations in year t = the average number citations of similar patents (same classification and same age) in the past

• so W should be a more robust dependent variable (than patents/applications) in the LHS of the regression

2000

2000

1

1

Att

Att

P

CW

• Independent/explanatory variables in the RHS, 30 largest investment banks

• The volume of equity securities underwritten by the investment bank

• The volume of debt securities underwritten by the investment bank

• The investment bank’s reputation among its peers from “tombstone announcements”

• The bank’s R&D effort

• Usually R&D expenditures or R&D staff used as a proxy, for banks difficult to find this information

• the average share of bank’s employees on the editorial/advisory boards of academic-practioneer journals used as a proxy

• Patent awards used from 1980-1997 in the LHS and characteristics of 30 largest investment banks in the RHS

• Both time (T)and cross-section dimensions (N) where N>T panel data

• The problem with the independent variables is that the there has been a lot of merger activity in the banking industry

investment banks measured as if they were in the current configuration (i.e., if bank k consisted previously from banks h and j, the observations from h and j were combined)

a (fixed effect) panel data regression:

itititiit uXXY ...2211

Estimation techniques

• Most straightforward is a linear model, estimated by OLS

• Good starting point but

• can predict negative patent numbers

• if log transformation used, what about the cases when the banks had no patents?

the Poisson regression

• pooled, fixed effects, and between estimation

• pooled: neither time nor cross section dimension

• fixed effects: time variation within each cross section

• arbitrary correlation between unobservables

• between/random effects: variation in cross-section dimension

• no correlation between unobservables

• Heteroscedasticity-robust standard errors

• Results: Figure 3

• Measure for academic ties/R&D effort significant except for the fixed effects regressions

• if the academic ties of investment banks vary little over time, difficult to separate from the fixed effects

a better proxy needed to assess of how important is the bank’s knowledge of frontier academic research for the innovativeness of the bank

a better proxy for banks R&D needed

• Other typical problems in these kind of regressions

• Endogeneity, i.e., “the chicken-and-egg” problem

• Omitted variables

• The finding also suggests that academic finance research is patentable

• This notion supported by the interviews of patent lawyers and the share of finance patents that have roots in the top academic journal publications

• However, most academic research has not been

patented…Lerner argues that this is because lack of awareness and interest

• Book authorship and professional technology transfer offices of universities explain finance faculty patenting

• University faculty faces trade-offs: make commercial/patent applications vs. invest in frontier research/teaching?

• more generally, how well financial institutions manage their IP?

• Cf. the difference in his 2006 article between financial innovators and financial patent holders

• Europe?

Where will the State Street lead?

• Countries that allow business method patenting:• US• Australia• Korea• Japan

• Countries that do not allow (in principle)• EPO• UK• Canada

• In Europe the law explicitly rules out business method and software patents as such

European Patent Convention Article 52:

(1) European patents shall be granted for any inventions which are susceptible of industrial application, which are new and which involve an inventive step

(2) The following in particular shall not be regarded as inventions within the meaning of paragraph 1:

• (a) discoveries, scientific theories and mathematical methods;

• (b) aesthetic creations; • (c) schemes, rules and methods for performing

mental acts, playing games or doing business, and programs for computers

• Instead, software is protected as a literary work under copyright protection

• In practice, EPO has become increasingly tolerant on software & business method patents, situation roughly speaking similar as in the US after Diamond v. Diehr

• Computer programs can be patented if they produce technical effects that go beyond the normal physical interactions between software and hardware

• EPO & EC wanted to clarify the situation similar manner as CAFC

the Software Patent Directive (allowed software patents but not business method patents)

one of the most heated lobbies ever & and the surprise NO vote in EP

no drastic change in the situation

The case against finance patents

• Patents are uninvited quests, finance/banking industry did not ask for them

• e.g. railroads in the 1900th century

• Finance/banking industry has been highly innovative without patents

• Theory predicts patents should spur further innovation but the evidence on software is weak

• Patenting will change the nature of innovation and industry organization, nothing else

• concern for bad patent quality

Financial services

Growth

Innovation in other sectors

Innovation in financial service sector

The case for finance patents

Stronger IP protection

• Vertical disintegration possible?

• e.g., eSpeed spinoff of Canton Fitzgerald

• Canton a leading player in the bond market

• eSpeed develops and seels pricing and trading software for various securities markets, an IPR oriented finance firm

• If finance patents are not allowed, how European financial service sector firms will manage IPs?

• Meridea• A joint venture between Nokia, 3i, Sampo &

Accenture

• Patents needed to stimulate disclosure?

• Patents needed to create a functioning market for technology?

• Tacit knowledge vs. codified knowledge

• Patents needed to finance start-ups?• Without patents new start-ups would have no

collateral, no way to tie down human capital of their founders

• Are patents good for innovation?

• Inverted U shape again?