Economic Models of Trade Secrets

Một phần của tài liệu The Economics of Trade Secrets: Evidence from the Economic Espionage Act (Trang 53 - 66)

Further  papers  attempt  to  provide  a  stronger  theoretical  framework  to  the   discussion  of  trade  secrets  by  modelling  either  the  decision  not  to  patent  and/or   the  decision  to  use  trade  secrets.    I  highlight  the  distinction  because  the  former   approach  treats  trade  secrets  as  a  catch-­‐all  strategy  for  actions  other  than   patents,  as  discussed  in  Arundel  (2001.)    The  implication  is  that  using  no  IP   protection  at  all  fits  into  the  same  category  as  Trade  Secrets;  using  Trade  Secrets   is  a  passive  action.    The  second  approach  of  modelling  the  decision  to  use  Trade   Secrets  treats  Trade  Secrets  as  an  active  decision.    This  approach  takes  into   account  the  “reasonable  steps”  required  to  protect  Trade  Secrets.    However,  this   approach,  while  excluding  using  no  IP  from  using  Trade  Secrets,  fails  to  offer   using  no  IP  as  a  strategy.      Arundel  (2001)  points  out  that  a  number  of  policy   discussions  model  the  decisions  to  use  patents  with  the  assumption  that  patents   are  an  obvious  decision  or  mutually  exclusive  from  trade  secrets;  both  

assumptions,  he  argues,  that  are  in  conflict  with  the  empirical  evidence.    This   section  of  the  chapter  examines  the  literature’s  analysis  of  the  firm’s  IP  decisions   with  respect  to  the  decision  not  to  patent  and  the  decision  to  use  trade  secrets.  

Decision  Not  to  Patent    

The  decision  not  to  patent,  as  Arundel  (2001)  notes,  is  sometimes  modelled  as  a   default  decision  to  use  trade  secrets.    One  example  is  that  of  Erkal  (2005),   discussed  in  Section  2.4.1,  in  which  the  author  examines  the  impact  of  policy   regimes  on  the  behaviour  of  firms  in  two  sequential  R&D  races.    In  the  first   round,  depending  on  the  regime,  the  winner  may  decide  to  maintain  the  first   innovation  secret  in  order  to  delay  the  progress  of  its  competitors  in  the  second   round.    Erkal  suggests  that  strong  trade  secret  protection  should  be  accompanied  

by  broad  patent  protection  and  that  allowing  collusion  to  encourage  disclosure   may  be  optimal.    However,  Erkal  assumes  reverse  engineering  to  be  easy  and   that  commercialization  results  in  the  same  amount  of  information  disclosure  as   patenting.    However,  reverse  engineering  can  be  difficult  and/or  costly  and,   consequently,  commercialization  can  result  in  only  limited  disclosure.    Thus,   these  two  assumptions  are  fairly  strict  and  reduce  the  model’s  ability  to  make   policy  recommendations.  

Decision  to  Use  Trade  Secrets    

The  active  decision  to  use  trade  secrets  presents  a  more  complete  analysis  of  the   firm’s  strategic  protection  of  innovation,  as  noted  in  Arundel  (2001.)  An  example   is  that  of  Bessen  (2004),  where  the  author  develops  a  three-­‐stage  model  with   two  competing  firms.    The  Bessen  model  also  incorporates  issues  of  disclosure   and  argues  that  diffusion  of  knowledge  is  not  necessarily  more  likely  with  a   patent  system.  

 

In  the  first  stage  of  the  Bessen  (2004)  model,  the  innovator  decides  whether  to   patent  or  use  trade  secrets;  in  the  second,  the  follower  decides  whether  to   develop  innovation  independently  or  not  (potentially  by  inventing  around);  and,   in  the  final  stage,  the  firms  produce  and  compete  in  the  market.    The  author   assumes  that  patenting  costs  more  than  using  trade  secrecy  and  concludes  that   firms  use  patents  when  they  reduce  or  eliminate  imitation.      

 

If  the  follower  chooses  not  to  imitate,  then  the  innovator  will  receive  monopoly   profits  under  trade  secrecy,  or,  under  patent  protection,  monopoly  profits   followed  by  duopoly  profits  for  both  firms  once  the  patent  expires.    (For   simplicity,  Bessen  chooses  not  to  incorporate  discounting.)    If  the  follower   chooses  to  imitate  or  invent  around,  they  will  incur  R&D  costs  and  their  efforts   may  not  result  in  a  successful  imitation.    If  the  follower  is  successful,  then  both   firms  will  receive  duopoly  profits  under  both  IP  regimes.    Unsuccessful  imitation   results  in  the  innovator  retaining  monopoly  profits.  The  decisions  of  each  player   are  determined  by  the  probability  of  diffusion.    Bessen  argues  that  firms  use  

patents  when  they  serve  to  reduce  or  prevent  imitation  and  diffusion  from   imitation  is,  therefore,  reduced  under  patents.  

 

Bessen’s  analysis  focuses  on  the  decision  to  use  secrecy  and  the  subsequent   diffusion  of  technical  information.    The  author  argues  that  technology  is  more   diffused  when  it  is  profitable  for  the  follower  to  imitate  and  that  patent  regimes   slow  down  this  diffusion.    Bessen  notes  that  firms  can  choose  whether  to  protect   inventions  by  patents  or  by  trade  secrecy  and  predicts  that  the  diffusion  of   technical  information  of  inventions  is  not  improved  by  the  patent  system  and   may  be  delayed.    

Trade  Secret  Models  and  Licensing    

The  licensing  of  trade  secrets  allows  for  extensions  of  simpler  models  and  can   affect  the  social  surplus  effects  of  trade  secrecy.      

 

Bhattacharya  and  Guriev  (2006)  develop  a  two-­‐stage  model  in  which  a  research   unit  develops  an  innovative  idea,  which  is  then  licensed  to  a  development  unit  to   potentially  develop  into  an  innovation.    The  focus  of  their  paper  is  how  the  units   can  use  licensing  strategically.    In  their  model,  licensing  occurs  either  through  an   open  sale,  in  which  the  knowledge  is  protected  through  patenting,  or  a  closed   sale,  in  which  trade  secret  protection  is  used.    The  authors  conclude  that  the   closed  licensing  with  trade  secrecy  is  most  often  used  if  the  knowledge  is  highly   valuable,  if  intellectual  property  rights  are  not  well  protected  and  if  negotiations   involve  substantial  knowledge  leakage.  

 

The  Bhattacharya  and  Guriev  approach  highlights  an  important  aspect  of  using   trade  secrecy  in  that  the  leakage  during  negotiations  is  paramount.    The  authors   also  note  that  valuable,  trade  secret  protected  knowledge  is  more  likely  to  result   in  an  exclusive  license  that  minimizes  leakage.    If  the  monopoly  rents  created  by   an  exclusive  license  are  high  enough,  the  licensor  has  no  incentive  to  sell  the   knowledge  to  a  third  party.    As  the  licensing  literature  acknowledges,  self-­‐

reinforcing  mechanisms  are  crucial  to  successful  licensing  of  knowledge.  

 

The  Bessen  model  is  extended  to  include  licensing  and  concludes:  

The  extent  of  the  market  for  licenses  may  actually  be  greater  without   patents.    The  intuition  is  simple:  licensing  occurs  where  there  is  a  credible   threat  of  imitation.    Because  imitation  occurs  in  more  restricted  

circumstances  with  patents  than  without  patents,  the  extent  of  licensing  is   less  with  patents.  

 

Cugno  and  Ottoz  (2006)  also  develop  a  model  detailing  how  the  innovator’s   choice  to  use  patents  or  trade  secrets  as  protection  affects  social  surplus,  which   they  term  social  welfare.    In  their  view,  patents  represent  a  temporary  monopoly   followed  by  perfect  competition,  whereas  secrecy  represents  a  long-­‐term  

oligopoly.    The  introduction  of  licensing  under  the  secrecy  regimes  allows  for   this  oligopoly.    Cugno  and  Ottoz  conclude  that  social  welfare  is  greater,  ex  post,   when  the  innovator  chooses  secrecy  given  that  transactions  costs  do  not  get  too   costly.    Licensing  avoids  welfare  wasting  duplication  costs  (the  cost  of  

reinventing  an  existing  innovation)  as  these  are  instead  appropriated  by  the   inventor  via  licenses.      

 

Applying  a  ratio  test,  also  used  in  Scotchmer  (2005),  Cugno  and  Ottoz  (2006)   examine  the  conditions  of  the  deadweight  losses  and  profits  under  different  IP   regimes.    The  ratio  test  provides  an  overview  of  the  per-­‐period  deadweight  loss   with  respect  to  the  per-­‐period  profit.    As  Scotchmer  notes,  this  test  is  used  in   Gilbert  and  Shapiro  (1990)  to  argue  that  substituting  lower  prices  for  longer   protection  is  socially  beneficial.”37    The  ratio  examines  the  ratio  of  deadweight   loss  and  profits;  where  the  ratio  of  deadweight  loss  to  profits  is  lower  indicates   the  better  policy.      

 

In  Cugno  and  Ottoz  (2006),  the  authors  evaluate  this  ratio  under  secrecy  and   patents.    Primarily,  they  are  interested  in  cases  in  which  the  deadweight  loss  to   profit  under  patents  (WP)  is  greater  than  the  deadweight  loss  to  profit  under   secrecy  (WS):    

                                                                                                               

37  Scotchmer  (2005),  p.  109.  

[2-­1]  

   

 

Assuming  that  patent  duration  and  secrecy  are  equally  profitable  for  the  

innovator,  the  authors  conclude  that  social  welfare  is  greater  under  secrecy.    The   authors  draw  this  conclusion  because  duplication  costs  under  trade  secrecy  are   converted  into  licensing  fees.    Thus,  these  duplication  costs  are  not  put  towards   the  wasteful  task  of  reinventing  the  wheel.    The  introduction  of  licensing  makes   trade  secrecy  a  socially  beneficial  regime  by  increasing  social  surplus  (which  the   authors  refer  to  as  social  welfare.)  However,  this  analysis  ignores  enforcement   costs  and  the  incentive  to  innovate  goal  of  IP  systems.  

 

The  licensing  of  trade  secrets  does,  however,  present  some  practical  difficulties.    

Primarily,  given  the  intangible  nature  of  trade  secrets,  it  is  difficult  to  determine   if  the  licensee  has  truly  received  the  trade  secret  knowledge  and  difficult  to   determine  if  the  licensee  ceases  to  use  this  knowledge  upon  the  license’s  

expiration.    Nonetheless,  as  numerous  authors  examine  (Scotchmer,  1991;  Cugno   and  Ottoz,  1991;  Choi,  2004;  and  Anand  and  Kahnna,  2000),  the  introduction  of   licenses  to  trade  secret  models  can  allow  for  increased  social  surplus  and   diffusion  of  knowledge.  

The  Anton  and  Yao  Model    

However,  one  weakness  of  many  of  the  previously  discussed  models  is  that  they   treat  trade  secrets  and  patents  as  mutually  exclusive,  which,  as  noted  by  Arundel   (2001),  is  not  necessarily  true.    An  example  of  an  exception  to  these  models  is   the  Anton  &  Yao  (2004)  paper.    In  “Little  Patents  and  Big  Secrets,”  Anton  and  Yao   (2004)  argue,  as  the  title  suggests,  that  firms  should  patent  small  innovations   and  use  trade  secrets  to  protect  larger  innovations.    Furthermore,  the  authors  

allow  for  a  mixing  of  trade  secrecy  and  patent  for  medium-­‐sized  innovations.    

This  widely  cited  paper  provides  a  relatively  comprehensive  model  in  addressing   the  decision  to  use  patents  or  trade  secrets.    Their  argument  is  based  on  three   fundamental  assumptions  that  “innovation  creates  asymmetric  information,   innovation  often  has  only  limited  legal  protection  and  disclosure  facilitates   imitation.”    The  authors  view  the  choices  of  IP  protection,  and  its  subsequent   disclosure,  as  an  important  signalling  mechanism.    Based  on  this  disclosure  and   limited  legal  protection,  others  determine  whether  to  imitate  or  not.  

Model    

The  Anton  and  Yao  (2004)  model  begins  with  a  cost  reducing  process  innovation   by  an  innovator.    The  innovation  reduces  the  marginal  cost  (MC)  of  production  to   c.    In  the  first  stage,  the  protection  and  disclosure  stage,  the  innovator  chooses   the  form  of  IP  protection  (either  secrecy  or  patent)  and,  as  a  consequence,  the   level  of  disclosure,  which  acts  as  a  signal.    Disclosure  occurs  under  both  Secrecy   (S)  and  Patent  (P),  but  only  patent  provides  protection  for  disclosure.    The   disclosure  allows  the  second  player,  the  follower,  the  option  of  reducing  his   costs.  

Table  2-­2:  Key  Variables  in  Anton  and  Yao  (2004)  model  

i=Innovator j=Follower

c =Marginal cost of prior technology c=Marginal cost of new technology

s=Follower's marginal cost with disclosure where

sc

 

 

The  Innovator  must  decide  between  Secrecy  (S)  and  Patenting  (P)  the  

innovation.    In  the  second  stage,  if  the  Innovator  has  chosen  Patenting,  then  this   stage  is  the  infringement  risking  imitation  stage.    In  this  case,  the  follower   chooses  to  not  imitate  (N)  or  to  imitate  (I),  which  is  actually  a  decision  to  risk   losing  an  infringement  lawsuit.    If  the  Innovator  has  chosen  Secrecy,  the  Follower   is  assumed  to  Imitate.    The  Innovator  and  Follower  have  the  following  decisions:  

 

Table  2-­3:  Key  Decisions  in  Anton  &  Yao  (2004)  model    

 

For  the  follower,  choosing  to  imitate  allows  the  follower  to  operate  at  marginal   cost  s.    However,  if  the  innovator  has  chosen  Patent,  the  probability  of  being   found  to  infringe  is  γ  and  damages  are  calculated  based  on  the  principle  of   reasonably  royalty  at  τ.      

In  the  final,  competition  stages,  the  firms  compete  in  a  Cournot  duopoly  where:  

   

The  model  can  be  summarized  in  the  following  game  tree:  

   

 

The  best  response  (BR)  functions  for  each  firm  are:  

 [2-­2]  

   

 [2-­3]  

     

Where  g  =  damages  rate  =  γ  τ,  which  is  the  probability  of  the  Follower  being   found  to  infringe,  multiplied  by  the  royalty  rate  that  the  court  would  require  the   Follower  to  pay  the  Innovator.    

Innovator’s  Strategy    

Using  separating  perfect  Bayesian  equilibrium  and  backwards  induction,  the   authors  determine  that  the  best  strategies  for  the  innovator  are  based  on  their   innovation  size.    The  Innovator  chooses  to  signal  its  innovation  either  partially   Figure  2-­2:  Anton  and  Yao  (2004)  Game  Tree  

(via  secrecy)  or  more  fully  (via  patenting.)    Anton  and  Yao  define  the  size  of  the   innovation  by  the  cost  reduction  it  creates;  the  results  are  summarized  here:  

Figure  2-­3:  Anton  and  Yao  (2004)  Marginal  Cost  and  Innovation  Size  

 

Table  2-­4:  Anton  and  Yao  (2004)  Model  Conclusions   Anton  and  Yao  model  conclusions  

Innovation  Size   Large   Medium   Small  

Effect   Waiver  Effect   Licensing  Effect   No  Imitation  

IP   secrecy   patents   patents  

Disclosure   partial   partial   full  

Follower’s  Action   Produces  at  s   Imitates,  risks  

damages   Does  not  imitate   (s  ≥  c  ≥  c*)   Small  innovations  –  No  Imitation  Effect  

 

Small  innovations  are  always  patented  because  the  innovator  knows  the  follower   will  not  imitate.    To  the  follower,  because  the  cost  advantage  is  so  small,  the  risk   of  paying  damages  outweighs  the  benefits  of  the  lower  marginal  cost.    Therefore,   the  disclosure  associated  with  patents  is  acceptable  for  the  innovator.  

 

More  specifically,  a  small  innovation  is  defined  as  one  that  the  marginal  cost   remains  above  c*.    When  c  ≥  c*,  disclosure  is  full.    The  follower  could  imitate,  but   the  risk  of  damages  outweighs  the  cost  reduction  benefits  thus  the  follower   remains  at  s  where  s  ≥  c  ≥  c*.    The  follower  then  earns  (1/9  β)(α  –  2c  +   )2      

MC  and  Innovation  Size    

 

Large                         Medium                Small  

 

          0                                               cL                  c*                        

 

c*  is  cost  above  which  j  chooses  N,  below  I  

 

which  is  dependent  on  the  cost  of  the  prior  technology  (where  the  follower   remains)  and  the  cost  with  the  innovation  (where  the  innovator  operates.)    The   disclosure  of  c  ≥  c*  only  affects  the  follower’s  behaviour  in  that  they  do  not   imitate,  knowing  that  damages  would  outweigh  the  cost  benefit.  

Medium  innovations  –  Licensing  Effect    

For  medium  innovations,  the  follower  is  enticed  to  imitate,  as  expected  damages   no  longer  outweigh  the  benefit  from  infringing  and  the  innovator  patents.    The   follower  imitates  incurring  the  expected  damages  knowingly.    These  expected   damages  are  transferred  to  the  innovator  through  litigation  and  the  damages   function  as  a  license.        

 

Anton  and  Yao  define  the  medium  innovation  as  larger,  where  c<c*.    This  cost   reduction  is  enough  to  trigger  imitation  by  the  follower.    Knowing  that  types   above  c*  (small  innovations)  disclose  fully,  disclosure  to  signal  cost  below  c*  has   to  be  less  than  c*.    Thus,  the  innovator  will  patent  and  disclose  partially  (e.g.  not   patent  the  entire  innovation.)    It  is  here  that  Anton  and  Yao  add  the  crucial   observation  that  patents  and  secrecy  are  not  mutually  exclusive.    The  decision  to   disclose  partially  and  entice  innovation  means  that  the  innovator  will  lose  in   terms  of  market  revenue,  but,  as  the  follower  is  infringing,  gain  in  terms  of   expected  damages.    Thus  the  authors  define  this  area  as  the  “licensing  affect”,   because  the  follower  chooses  to  use  the  innovation  in  exchange  for  expected   damages  payments.  

Large  innovations  -­  Waiver  Effect    

As  the  innovation  becomes  larger,  c≤cL,  the  innovator  faces  a  trade-­‐off,  when   signalling  low  cost,  between  operating  revenues  and  damages  (licensing.)    Lower   costs  will  reduce  the  market  price  and,  therefore,  reduce  expected  damages   (which  are  calculated  based  on  reasonable  royalty.)    Thus,  the  innovator  utilizes   the  cost  advantage  in  the  market  as  the  main  source  of  revenue  and  elects  not  to   use  damages  by  choosing  secrecy.  

 

Thus,  the  larger  innovation  uses  trade  secrecy  to  minimize  disclosure  and  limit   the  follower’s  ability  to  imitate.    However,  the  authors  highlight  some  of  the   practical  difficulties  with  this  conclusion.    The  decision  not  to  patent  may  be  due   to  the  non-­‐patentability  of  the  innovation.    They  also  suggest  that  a  patent  with  a   nominal  lump  sum  may  be  a  viable  alternative  but  leave  this  option  for  later   research.  

 

Anton  and  Yao  make  one  statement  that  contradicts  the  technical  definition  of  a   trade  secret.    They  call  the  decision  not  to  patent  a  “non-­‐action.”    The  implication   is  that  the  decision  to  use  trade  secrets  is  thereby  a  non-­‐action.    This  is  not   necessarily  true,  as  discussed  in  Arundel  (2001),  as  the  decision  to  use  trade   secrets  is  an  active  decision.    The  decision  to  use  trade  secrets  is  a  decision  to   utilize  confidentiality  agreements,  secrecy  and  other  legal  measures  for  the   protection  of  the  innovation.  

Furthering  Anton  and  Yao    

Encaoua  and  Lefouili  (2006)  develop  a  model  similar  to  Anton  and  Yao’s.    

However,  Encaoua  and  Lefouili  focus  on  the  probabilistic  nature  of  patents,   rather  than  the  signalling  aspect  of  choosing  an  IP  regime.    They  focus  on  three   parameters  that  affect  the  firm’s  IP  choice:  patent  strength  (likelihood  patent  will   be  held  up  in  court),  cost  of  imitating  (the  cost  to  the  follower  of  imitating  a   patented  innovation  relative  to  a  secret  innovation)  and  the  innovation  size  (the   extent  of  the  cost  reduction.)  

 

Both  patent  strength  and  innovation  size  are  referred  to  explicitly  in  the  Anton   and  Yao  model.    However,  the  cost  of  imitating,  in  Anton  and  Yao,  is  not  explicit   and  is  only  captured  partially  in  a  consequential  cost  of  imitation  –  that  of  

expected  damages  under  the  patent  regime.    Instead,  the  Encaoua  model  focuses   on  the  costs  to  reverse  engineer  or  develop  the  innovation  independently,  which   is  typically  higher  under  secrecy  than  under  patent  protection.    The  Encaoua   model  treats  expected  damages  as  a  separate  cost.    

 

A  higher  cost  of  imitation  is  likely  to  dissuade  a  follower  from  imitating.    At  the   same  time,  a  strong  patent  regime  (patent  strength)  is  also  likely  to  dissuade   imitation  due  to  expected  damages  payments.    Lower  costs  or  weaker  patent   strength,  on  the  other  hand,  make  imitating  more  attractive.    Thus,  Encaoua  and   Lefouili  view  patent  strength  and  the  cost  of  imitation  as  strategic  substitutes.  

The  Encaoua  and  Lefouili  (2006)  paper  also  departs  significantly  from  Anton  and   Yao’s  focus  on  signalling  by  permitting  the  innovation  (cost  reduction)  to  be   directly  observable.  This  is  a  fairly  limiting  assumption,  as  it  would  require  firms   to  know  marginal  costs  of  the  innovator  directly,  before  and  after  the  innovation.  

 

The  Encaoua  and  Lefoulli  paper  focuses  on  two  competing  effects:  the  damage   effect  and  the  competition  effect.    The  damage  effect  is  the  advantage  that   patents  have  over  secrecy.    Patents  allow  for  the  possibility  of  damages   payments,  whereas  secrecy  does  not38;  thus,  ceteris  paribus,  the  firm  would   choose  patents.    The  competition  effect  allows  for  imitation  levels  to  differ   between  patents  and  secrecy.    The  regime  that  has  less  imitation  will  be   preferred.    The  Encaoua  and  Lefoulli  paper  highlights  the  interaction  between   these  two  effects.  

 

The  Encaoua  and  Lefoulli  (2006)  model  is  very  similar  to  the  Anton  and  Yao   model.    An  innovator  creates  a  cost  reducing  process  innovation  and  then   proceeds  to  compete  with  a  follower  in  a  Cournot  duopoly  in  three  stages:  the   protection  stage,  the  imitation  stage,  and,  finally,  the  market  competition  stage.    

However,  the  new  model  introduces  one  important  new  variable:  I  (cost  of   imitating),  where  imitating  a  patented  innovation  costs,  at  most,  as  much  as   under  imitating  an  innovation  protected  by  secrecy.  

 

The  results  of  the  Encaoua  and  Lefoulli  model  are  similar  to  that  of  Anton  and   Yao.    They  argue  that  small  innovations  are  always  patented.    This,  however,  is   due  to  the  low  cost  of  imitating  a  small  innovation  coupled  with  the  lack  of   damages  under  secrecy.    Also  like  Anton  and  Yao,  the  Encaoua  model  predicts                                                                                                                  

38  In  the  case  of  reverse-­‐engineering,  which  is  permitted  under  trade  secrets  law.  

that  large  innovations  are  kept  secret  because  any  disclosure  reduces  I  (the   imitation  cost)  and  invites  imitation.  

 

For  medium  sized  innovations,  the  two  models  differ.    In  the  Encaoua  and   Lefoulli  model,  medium  sized  innovations  are  either  patented  or  kept  secret,  as   opposed  to  Anton  and  Yao,  where  they  are  patented  and  partially  disclosed.    

Imitation  occurs  partially  under  secrecy  and  may  occur  under  patenting.    The   difference  between  the  two  models’  conclusions  stems  from  the  fact  that  patent   strength  and  imitation  costs  interact  to  influence  the  follower’s  behaviour  in  the   Encaoua  model,  whereas  signalling  by  the  innovator  plays  a  role  in  the  Anton   and  Yao  model.  

 

Encaoua  and  Lefoulli  (2006)  also  develop  the  licensing  option  hinted  at  in  Anton   and  Yao.    They  develop  a  fixed  fee  plus  royalty  licensing  scheme  which  allows  for   the  same  equilibrium  outcomes  as  in  patenting  under  the  shadow  of  

infringement.    This  is  akin  to  Anton  and  Yao’s  “licensing  effect”  which  arises   through  litigation.    Encaoua  and  Lefoulli  frame  licensing  deliberately  as  an   alternative  to  litigation.    Jointly,  these  two  papers  embrace  Shankerman  and   Scotchmer’s  (2001)  observations  on  the  circularity  of  damages,  in  that   reasonable  royalty  is  determined  by  a  combination  of  willing  licensing,  and   coerced  licensing  through  litigation.    Encaoua  and  Lefoulli  argue  that  licensing  in   lieu  of  litigation  may  lead  to  poor  quality  patents  being  licensed  and  decrease   social  surplus.      

 

As  the  Encaoua  and  Lefoulli  paper  notes,  the  introduction  of  probabilistic  patents   to  theoretical  models  allows  for  more  thorough  analysis  of  IP  policy  by  opening   up  new  research  avenues.    The  authors  claim  that  the  one-­‐size-­‐fits-­‐all  approach   of  patents  does  not  take  into  account  the  heterogeneous  nature  of  innovations   and  patents.    Instead,  they  suggest  that:  

 

…some  flexibility  could  be  introduced  by  allowing  each  innovator  to   choose  a  patent  inside  a  menu  of  characteristics.  For  instance  an   innovator  may  have  to  choose  between  a  patent  with  strong  property   rights  and  high  disclosure  requirements  and  a  patent  with  weak  property  

Một phần của tài liệu The Economics of Trade Secrets: Evidence from the Economic Espionage Act (Trang 53 - 66)

Tải bản đầy đủ (PDF)

(328 trang)