WHY DO STARTUPS USE TRADE SECRETS?

• Author: Levine, David S.

INTRODUCTION 753 I. EXISTING THEORIES OF WHY STARTUPS USE (AND DON'T 755 USE) TRADE SECRETS A. Reasons Why Startups Use Trade Secrecy 756 1. To Maintain Lead-Time Advantage and Prevent 757 Competition 2. Patent Protection Is Unavailable 758 3. Patent Protection Is Too Costly, Weak, or Difficult 761 to Enforce 4. To Provide Protection Prior to and 763 Complementary with Patenting 5. To Assist in Financing and Financial Exit 766 6. To Prevent Employees from Working at 767 Competitors 7. To Generate Licensing Revenues 768 8. To Provide Strategic Bargaining Leverage 770 B. Reasons for Startups Not to Use Trade Secrecy 770 1. Ease of Reverse Engineering and Advantages to 770 Patenting 2. Fear of Others Independently Patenting the 771 Invention 3. Inability to Market or Even Explain to Investors 773 4. Tension with Open Source, Open Innovation, and 774 Other "Sharing" Models 5. Reliance on Contracts Like Covenants Not to 775 Compete and Nondisclosure Agreements II. THE PAUCITY OF EMPIRICAL STUDIES ON TRADE SECRECY 776 A. Previous Empirical Studies of SMEs and Trade Secrecy 779 B. Speculating on the Results of More General Studies 781 1. The Benefits of "Informal" vs. "Formal" IP 781 2. Products vs. Processes 782 3. Lead-Time and First-Mover Advantage 783 4. Industry Analysis: Software and Biotechnology 784 5. Collaboration 785 6. Knowledge Spillovers 786 7. Research and Development (R&D) 787 8. Hybrid Strategies 788 9. Open Questions Remaining from Previous Studies 789 III. AN EMPIRICAL ASSESSMENT OF STARTUPS AND TRADE SECRECY 789 A. The Data: The Berkeley Patent Survey 790 B. Response Rates and Respondent Characteristics 792 1. Survey Response Rates 792 2. Characteristics of Respondents 793 C. The Results: The Use of Trade Secrets by Startups Highly 794 1. Descriptive Results 795 a. Importance of Trade Secrecy Relative to Other 795 Appropnabihty Mechanisms b. Trade Secrets and Patents as Substitutes or 797 Complements? c. "New" vs. "Old" Startups 800 d. Important of Secrecy by Business Model 801 2. Regression Models 803 a. Explanatory Variables 803 3. Regression Results 805 a. General Trends 805 b. Secrecy vs. Patenting 807 IV. IMPLICATIONS OF OUR RESULTS: THE VARIED USES OF TRADE 808 SECRETS A. The Role of Trade Secrecy in Promoting First-Mover 809 Advantage B. Trade Secrets and Patents as Substitutes and Complements 810 C. Trade Secrets as Strategic IP Assets 811 CONCLUSION 812 APPENDIX 813 INTRODUCTION

Magic Leap is a startup company that is active in the exploding field of augmented reality, head-mounted displays that "superimpose[ ] 3D computer-generated imagery over real world objects." (1) In 2016, it was one of the first companies to sue under the recently passed Defend Trade Secrets Act (DTSA), the first federal law to provide broad protection for private litigants over trade secrets. (2) Suits like Magic Leap's are becoming common, with the "amount and significance" of trade secret litigation "exploding" over the past thirty years. (3)

Yet, despite this massive increase in the number of suits--and the increasing characterization of trade secrecy as part of the panoply of intellectual property rights protecting innovations--there has been little empirical study on what drives the use and effectiveness of trade secrets, unlike other forms of intellectual property, such as patents. (4) Of this small body of trade secret research, very little has even mentioned technology startups. (5) This is perhaps even more surprising given that trade secrets are often viewed as a vital mode of protection for a startup's innovations that may "not otherwise [be] protected by copyright or patent law." (6)

In 2008, a team of researchers (including one of us) at UC Berkeley performed a comprehensive study to better understand the role of patenting among startups in the software, medical device, biotechnology, and information technology hardware industries. (7) Indeed, one of the leading innovation economists, Bronwyn Hall, refers to it as the "most comprehensive evidence of start-up patenting." (8) Important for this Article, as part of that study, several questions compared patenting to secrecy. (9) Although some limited results regarding trade secrecy were published using this dataset, (10) much of the data as it relates to secrecy has been either underexplored or wholly unexplored, including the economic relationship between patents and trade secrets, (11) the role of trade secrets in promoting first-mover advantage, and the use of trade secrets as strategic intellectual property (IP) assets. (12) Indeed, after the publication of the Berkeley Patent Survey articles, leading trade secret researcher Ivan Png, as a result of empirical study, called for a "new research agenda--the impact of trade secrets law on (i) entrepreneurship and venture capital, (ii) collaboration, (iii) business and marketing innovation, and (iv) international trade and investment." (13) This Article addresses the first three of those items.

As such, this Article makes a significant contribution toward understanding these complex relationships, and thereby helps fill a meaningful void in the literature. Specifically, we make three major theoretical and empirical findings. First, trade secrecy may serve important aims other than extending first-mover advantage. (14) Second, trade secrets may act as both economic complements and substitutes to patenting. (15) Third, trade secrets may serve as important strategic assets, functioning much in the same manner as patents in terms of licensing and setting the boundaries of the firm. (16)

Part I of this Article assesses the current theoretical views on the use of trade secrets, applying them in the context of technology startups. Part II then explores the existing empirical research regarding trade secrets, noting the relative paucity of work focused on startups. Part III begins by describing the Berkeley Patent Survey, including its methods and dataset. It then presents the major results as they concern trade secrecy, first descriptively, and then in the form of several regression models. Part IV assesses the results, describing the study's major implications for trade secrecy theory and empirical research.

1. EXISTING THEORIES OF WHY STARTUPS USE (AND DON'T USE) TRADE SECRETS

   As noted earlier, the literature contains a variety of theories attempting to explain why companies choose to use trade secrets (or not). (17) Less prevalent are theories focused on startups. (18) Here, in order to provide context for our empirical study, we draw from the smaller literature on startups and adapt from the larger, more general literature to present a theory as to why startups use (and do not use) trade secrecy. (19) We focus on a startup's decision to choose trade secrecy as an appropriation mechanism in addition to (i.e., as an economic "complement") or in place of (i.e., as an economic "substitute") to other forms of intellectual property and barriers to competition, such as patents, copyrights, and lead-time advantages. (20)

   1. Reasons Why Startups Use Trade Secrecy

      While patents and copyrights are mandated by the U.S. Constitution, (21) the origins of trade secrecy trace to the common law. (22) There are two core elements for trade secret infringement under state law: " (1) the information qualifies as a 'trade secret,' and (2) the defendant acquired, used, or disclosed the information in breach of confidence or by other improper means." (23) The Restatement (Third) of Unfair Competition has articulated a similar framework for trade secrecy. (24) Trade secrecy applies to a range of information, including recipes, (25) software code, (26) customer lists, (27) algorithms, (28) and other technical information. (29) Perhaps the most well-known example of the "best-kept trade secret[ ] in the world" is the recipe for Coca-Cola. (30)

      1. To Maintain Lead-Time Advantage and Prevent Competition

         "Lead-time" or "first-mover" advantage occurs when a company enters the market earlier than competitors, which often allows the company to acquire a dominant market share for a particular product or service. (31) Although startups can maintain a lead-time advantage simply because of the inherent failure of competitors to innovate, (32) a primary reason for choosing trade secrecy is to extend a lead-time advantage by preventing the disclosure of specific information that provides the advantage. (33) For instance, to a large degree, Google--an early pioneer in online search engines--has been able to maintain its dominance by keeping key details of its search algorithms a trade secret. (34)

         Theoretical modeling supports this conclusion. For instance, Alexandra Zaby has constructed an economic model that finds when a technological first-mover advantage is large, inventors will often use secrecy in an effort to extend the lead. (35) Yet, the model also shows that when reverse engineering is straightforward--such as in the pharmaceutical industry--companies will tend to rely upon patents to protect a first-mover advantage. (36)

         Beyond first-mover advantage for a single company, sometimes a small number of incumbent companies may use trade secrets to maintain an oligopoly for a product or service merely by excluding potential competition. (37) Here, the mechanism is similar to that of maintaining a first-mover advantage--by preventing third parties from obtaining critical information, the incumbents enjoy supracompetitive profits flowing from the lack of competition. (38)

      2. Patent Protection Is Unavailable

         As just noted, sometimes patents are a better mechanism to maintain a lead-time advantage or to prevent competition than trade secrets, particularly when reverse engineering or independent invention is likely. (39) However, in many instances, patent protection is unavailable for an invention--for instance, because it is too abstract or too obvious of an idea to be patentable. (40) Thus, another important reason for using trade...