Knowledge Transmission by Mobile Employees:

1
Knowledge Transmission by Mobile Employees

• Do We Know Enough to Make Policy on Trade Secrets
  and Covenants Not to Compete?

2
By Alan Hyde

• Conference, IP_at_Work
• Radzyner School of Law, Interdisciplinary Center
  (IDC), Herzliya, Israel
• 28-29 December 2005

3
Eight conjectures about trade secrets in Silicon
Valley

• From Alan Hyde. Working in Silicon Valley
  Economic and Legal Analysis of a High-Velocity
  Labor Market (M.E. Sharpe 2003).
• Basic thesis The public interest in
  technological and economic growth requires
  lifting impediments to mobility of knowledge
  workers, such as covenants not to compete and
  trade secrets law.
• Obstacles to recognizing thesis (stronger ten
  years ago)
• Economic analysis that assumes that
  information/knowledge can have value only if it
  is property
• Legal obsession with implicit employment
  contracts for lifetime employment.

4
1.

• Some networks of firms may attain unusual
  technological and economic growth when
  information/knowledge is shared among firms.

5
2.

• There is a public interest in not impeding this
  growth.

6
3.

• Employee mobility is one valuable institution of
  information transmission

7
4.

• Silicon Valley, California, illustrates both high
  employee mobility and long-term technological and
  economic growth.

8
5.

• Some of the valuable information transmitted by
  mobile employees fits the legal definition of
  trade secrets. It is not possible to form a
  start-up, or even hire an experienced employee,
  without simultaneously acquiring information that
  is, legally, the trade secret of a former employer

9
6.

• Employers in Silicon Valley do not enforce their
  rights under trade secret law.

10
7.

• This refusal is rational and reflects many
  factors
• A. reputation for suing employees will hurt both
  recruiting, stock price, and reputation among
  peers
• B. trade secrets suits are hard to win and rarely
  accomplish anything
• C. the firms best investment is in employee
  information that it cannot retain
• D. firms hire as well as lose information.
  Shared information reduces RD costs
• E. an implicit promise not to interfere with
  future employment is the price of the employees
  willingness to accept precarious employment

11
8.

• Legal enforcement of trade secrets should, (a) at
  least, reflect the underlying implicit employment
  contract. (b) consider reducing the scope of
  trade secret, to advance the public interest in
  technological and economic growth.

12
What was the evidence?

• 1. growth through sharing anecdotal comparisons
  of Silicon Valley with Boston (Saxenian 1994)
  let alone Germany New Jersey. Franco Filson
  (2000) Knowledge Diffusion Through Employee
  Mobility hard disk drives.
• 2. public interest normative
• 3. employee mobility no rigorous comparisons
  with other information spillover
• 4. Silicon Valley hard to untangle relative
  contributions of HR practice, venture capital,
  Stanford, history, sunshine.
• 5. employees take trade secrets many anecdotes
  (Saxenian, Sitkin, etc.)
• 6. Employers dont enforce rights anecdotes
  interviews
• 7. Legal change normative suggestions very
  cautious

13
What have we learned 2003-05?

• 1. High turnover of employees is positive for
  productivity of computer firms.
• 2. How employees transmit information, documented
  through patent citations.
• 3. Silicon Valley really does have more job
  hopping.
• 4. Failure of German attempt to encourage venture
  and equity investment in tech, without changing
  employment market.
• 5. Employee incentives on open-source software
  projects illustrate rationality of open
  collaboration followed by employee mobility

14
Data on turnover and productivity

• Fredrik Andersson, Clair Brown, Benjamin
  Campbell, Hyowook Chiang Yooki Park. The Effect
  of HRM Practices and RD Investment on Worker
  Productivity http//web.mit.edu/ipc/sloan05/HRM_R
  D_Andersson_et_al.pdf
• Study relates data for computer firms employee
  earnings and employment from state unemployment
  insurance Census and Bureau of Labor Statistics
  data on RD and HRM practices (unpublished,
  preliminary)
• For firms with high levels of RD, high hiring
  rate for high- and low-educated workers is
  positive for productivity. So are multiple
  ports of entry and payment by incentives.
• For low RD firms, internal job ladders for
  low-educated workers are positive for
  productivity

15
How employees transmit information

• Patent citation literature, e.g. Adam B. Jaffe
  Manuel Trajtenberg, Patents, Citations, and
  Innovations A Window on the Knowledge Economy
  (MIT Press 2002).
• Only one reference to labor market, at 138 n.59
  full employment equilibrium a nuisance
  parameter they assume inelasticity of labor

16
New understandings of knowledge transmission from
patent citations

• Almeida Kogut Silicon Valley the only region
  where firm patent applications cite patents from
  other firms.
• Lee Fleming, Charles King III Adam Juda. Small
  Worlds and Regional Innovative Advantage.
  Harvard Business School Working Paper 04-008
  (2004) map of all collaborations in 2.8 million
  US patents 1975-2002. Silicon Valley many more
  collaborations. Agglomeration of paths of
  collaboration.
• Ajay Agrawal et al. Gone But Not Forgotten Labor
  Flows, Knowledge Spillovers, and Enduring Social
  Capital. NBER WP 9950 (2003). Disproportionate
  patent citations from regions where inventor was
  formerly employed. Demonstrates social capital.
  Even (especially?) mobile employees invest in
  social capital that survives geographic
  separation.

17
Data on Silicon Valley job-hopping

• Bruce Fallick, Charles A. Fleischman James B.
  Rebitzer. Job Hopping in Silicon Valley Some
  Evidence Concerning the Micro-Foundations of a
  High Technology Center. NBER WP 11710 (Oct.
  2005). College-educated men in computer industry
  in Silicon Valley change jobs 40-90 more
  frequently than mean (depending on definition of
  computer industry). Other California computer
  clusters similar to Valley. No general tendency
  among Californians to change jobs frequently.

18
You cant do it without labor market reforms

• Sigurt Vitols Lutz Engelhardt. National
  Institutions and High Tech Industries A
  Varieties of Capitalism Perspective on the
  Failure of Germanys Neuer Markt. Discussion
  Paper SP II 2005-03, Wissenschaftszentrum Berlin
  für Sozialforschung. German attempt 1997-2003 to
  encourage venture capital and equity investment
  in high tech startups though more frequent and
  stricter reporting requirements, ease of listing
  new companies, and protection for minority
  shareholders. Failed for several reasons,
  including low rates of investment by German
  households and pensions. However the lack of
  an open market for mid-career scientists and
  entrepreneurs makes it very difficult to attract
  the type of labor needed to create successful
  high-growth entrepreneurial companies.. (at 2).
  Only Italy (6.9) has lower percentage of high
  tech scientific and technological personnel who
  took a new job in the last year than does Germany
  (10.1)(Eurostat). Contrast Spain (23.8),
  Ireland (22.6), UK (18.3).

19
New Jersey Germany of the New World?

• Among US states, New Jersey is like Germany
  educated workforce many scientists and engineers
  working for large companies (ATT, Lucent, drug
  companies) highest per capita income but
  essentially no start-ups, venture capital, legal
  or financial infrastructure. Not coincidentally,
  it is the only state where an appellate court has
  adopted the doctrine of inevitable disclosure of
  trade secrets, and NJ vigorously enforces
  covenants not to compete.

20
Open Source Projects and Employee Incentives

• Open source software (Apache, Linux, other ).
  Extreme example of shared information with no
  property rights. What is employee incentive?
  Just love?
• Il-Horn Hann et al. An Empirical Analysis of
  Economic Returns to Open Source Participation
  (unpub. 2004). http//www.rcf.usc.edu/hann/publi
  cations_files/economic_returns_to_open_source_part
  icipation.pdf . Among programmers who
  contributed to Apache, 36 changed jobs in 1998.
  Those who had attained high rank within the
  Apache community enjoyed wages 14-29 higher
  (depending on rank). Rank signals desirable
  skills that cannot otherwise be easily observed.
  Confirms rationality of software employee
  accepting precarious employment (or volunteering)
  that pays off on next, paid job.

21
Legal Conclusions Reinforced

• Suit against departing employees to be evaluated
  against implicit employment contract. Did
  employer implicitly promise stability, or only
  that employee would be gone in 18 months? If the
  latter, did the employer implicitly promise not
  to interfere with employee mobility and
  transmission of some information about what he
  worked on? If the evidence is unclear, would
  such an agreement have been ex ante efficient for
  the employer (in light of studies reviewed here)?
• Public interest favors finding most or all
  covenants not to compete to be unreasonable,
  hence unenforceable. Concept of trade secret
  should conform to actual usage of employers and
  employees. Strong burden should be placed on
  employer to show a social advantage in
  recognizing trade secret.