Insider Trading and Innovation

Insider Trading and Innovation Ross Levine, Chen Lin and Lai Wei Hoover IP 2 Conference Stanford University January 12, 2016 Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 1

Motivation and Question Innovation is a major source of economic growth. Which laws and regulations foster/impede innovation? Considerable research on intellectual property rights. Our research agenda focuses on finance: Which financial laws and regulations foster/impede innovation? This paper examines insider trading laws Does restricting insider trading increase, decrease, or have no effect on technological innovation? Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 2

Differing Views and Predictions Efficiency of Insider Trading View Insider trading is an efficient way to compensate insiders for undertaking risky investments, such as innovation. Insider trading reveals information, improving stock price informativeness and hence the quality of investment. If insider trading is costly for some types of firms, they will design contracts to credibly eliminate it. Restricting insider trading will hurt innovation. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 3

Differing Views and Predictions Inefficiency of Insider Trading View Possibility of insider trading makes outside investors reluctant to finance firms, which raises the cost of capital. Possibility of insider trading makes outside investors reluctant to expend resources researching firms. This will hinder the effective valuation of projects and firms Especially opaque, hard to value endeavors like innovation. Since legal systems are imperfect, insiders cannot credibly contract around fears of insider trading. Restricting insider trading spurs innovation. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 4

This paper Provides the first assessment of whether restricting insider trading increases or decreases innovation. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 5

Empirical Strategy Exploit the staggered enforcement of insider trading across countries from 1976 to the present. Enforcement is defined by the date that a country first implements a prosecution under its insider trading laws. 36 economies of 94 countries have enforced laws. Examine what happens to patent-based measures of innovation. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 6

Data Patent Data We measure innovation based on patent metrics PATSTAT EPO Worldwide Patent Statistical Database It covers over 80 million patent applications originated from more than 100 countries that are filed with more than 80 patent offices worldwide from 1800s to present. We focus on patents filed with and granted by patent offices in OECD countries plus EPO to ensure comparability. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 7

6 Measures of innovation Quantity Patent Count The total number of distinct patents (eventually granted) in each country-industry-year Ln (1 + patent count) c,i,t Scope Number of Patenting Entities The total number of distinct entities filing patent in a particular country-industry-year Ln (1 + # entities) c,i,t Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 8

6 Measures of innovation Impact Patent Citation Number of distinct forward citations made to patents in industry i, filed in year t, by applicants from country c. Adjusted for truncation bias (Hall et al., 2001, 2005) Ln (1 + citations c,i,t ) High impact PC Top10% Number of patents in an industry-country-year whose forward-citation count falls into the top-10% of the citation distribution across all patents in the same technology class (IPC subsection) and application year. Ln (1 + high impact patents c,i,t ) Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 9

6 Measures of innovation Originality Originality Score = 1 HHI(tech-classes of cited patents) higher score indicates the patent cites a diversity of technology classes. Generality Generality Score = 1 HHI(tech-classes citing the patent) higher score indicates the patent is cited by a diversity of technology classes Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 10

Measures of innovation Some details We use first granted patent of each invention Date the innovation by its application year Assign its nationality based on the country of residence of the applicant Unit of analysis: industry-country-year level. Given data availability, we aggregate information to the industry level, not to the firm level. We omit the U.S. because we use it as a benchmark. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 11

Before and After Enforcement Average annual innovation before and after. 4500 Median Annual Citation Pre vs. Post 4000 3870.38 3500 3000 2500 2000 1500 1000 500 0 651.54 Pre [-5,-1] Post [+1,+10] Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 12

Empirical Challenges Although there is a huge increase in innovation after countries start enforcing insider trading laws, we face three broad categories of challenges to drawing causal inferences. 1 Yes, innovation went up after enforcement, but was it already trending up? Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 13

Perhaps, there is pre-existing trend Innovation Enforcement Date We will test for this. Time

Empirical Challenges 1 Pre-existing trends Is there a material change in the time-path of innovation after a country starts enforcing the law? 2 Reverse causality Does innovation trigger a change in insider trading laws? We examine this below. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 15

Empirical Challenges 1 Pre-existing trends Is there a material change in the time-path of innovation after a country starts enforcing the law. 2 Reverse causality Does innovation trigger a change in insider trading laws? 3 Other things? Other changes in the country or industry? Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 16

Pre-trends and dynamics No pre-trend and a notable break Citation Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 17

Reverse causality Neither the level nor the growth rate in innovation predict the timing of the enforcement Citation Δ Citation Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 18

Empirical Strategy Baseline: Difference-in-differences specification Question: Is α 1 positive, negative, and how big is it? Does restricting insider trading accelerate or slow innovation and how big is the effect? Controls: X: GDP, GDP per capita, Stock Market Capitalization/GDP, Credit/GDP, Export activities, more too Country fixed effects control for all time-invariant country traits, e.g., culture, perhaps quality of education, etc. Industry fixed effects control for all time-invariant industry traits, e.g., the natural rate of innovativeness, capital intensity, size, etc. Time fixed effects control for all factors in each year exerting a common effect on countries and industries, e.g., global recessions, global technological changes, etc. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 19

Empirical Results Baseline regression: Dependent variable Patent Count Patent Entities Citation PC Top10% Generality Originality (1) (2) (3) (4) (5) (6) Enforce 0.2594** 0.2061** 0.3666*** 0.1301*** 0.1584*** 0.1809*** (2.19) (2.04) (2.67) (2.76) (2.80) (2.93) Controls Yes Yes Yes Yes Yes Yes Country Fixed Effect Yes Yes Yes Yes Yes Yes Industry Fixed Effect Yes Yes Yes Yes Yes Yes Year Fixed Effect Yes Yes Yes Yes Yes Yes Observations 70,319 70,319 70,319 70,319 65,641 67,014 Adjusted R-squared 0.858 0.873 0.863 0.723 0.781 0.788 Note on magnitude: 13% increase in the number of top-10 patents Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 20

We can push this further We test whether the results are driven by some other law or regulation that just happens to occur at the same time? We control for many laws, regulations, and policies. E.g, Credit allocation laws, interest rate controls, new bank entry limits, the stringency of bank supervision, impediments to the operation of securities markets, restrictions on international capital flows, etc. All of the results hold. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 21

We can push this further Are the results stronger where the Inefficient Insider Trading View suggests that they should be? Theory: Restricting insider trading will induce outside investors to expend more resources researching firms and activities. This is especially important for opaque, hard-to-value activities, such as innovation. Therefore, restricting insider trading should have a particularly big impact on opaque and innovative activities. Do we find this? Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 22

We can push this further To measure the degree to which an industry is naturally innovative or opaque, we use the US as a benchmark. The US has comparatively strong legal and financial systems. We measure the degree to which industries in the US are innovative or opaque. We then use this to characterize industries around the world. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 23

Differentiating by Industry: Strategy Industry-partitioned regression: Question: Do naturally opaque and innovative industries experience a bigger increase in innovation after the enforcement of IT laws? Is β 1 > 0? Controls: country-time and industry-time fixed effects. Controls for all time-varying country traits, such as all national policies, including patenting laws, in each country, etc. Controls for all time-varying industry traits, such as global factors that shape the rate of innovation in particular industries over time. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 24

Empirical Results: Naturally innovative Measures of naturally innovative (US benchmark) High Tech = 1 if greater than the sample median of the average growth rate of R&D expenditures in each industry. Innovation Propensity = 1 if greater than the sample median of the average # of patents filed by public firms in each industry. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 25

Empirical Results: Naturally innovative Growth of Patent Count and Citation in naturally innovative and naturally less innovative groups. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 26

Empirical Results: Naturally innovative Stronger results among naturally innovative industries. Dependent variable Patent Count Patent Entities Citation PC Top10% Generality Originality (1) (2) (3) (4) (5) (6) Enforce High Tech 0.4283*** 0.3729*** 0.4293*** 0.3540*** 0.4240*** 0.4212*** (6.28) (6.73) (6.37) (5.23) (5.37) (5.62) Controls Yes Yes Yes Yes Yes Yes Country-Year Fixed Effect Yes Yes Yes Yes Yes Yes Industry-Year Fixed Effect Yes Yes Yes Yes Yes Yes Observation 73,410 73,410 73,410 73,410 68,010 69,403 Adj. R-squared 0.894 0.905 0.898 0.755 0.811 0.823 Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 27

Empirical Results: Naturally Opaque Measures of opacity (US benchmark) Intangibility = 1 if greater than the sample median intangibility; Intangibility = one minus the average share of tangible assets over total assets in each industry. STD of MTB = 1 if greater than the sample median of the average standard deviation of MTB across public firms in each industry, scaled by average MTB in the industry. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 28

Empirical Results: Naturally Opaque Stronger results among naturally opaque industries Dependent variable Patent Count Patent Entities Citation PC Top10% Generality Originality (1) (2) (3) (4) (5) Enforce Intangibility 0.2961*** 0.2638*** 0.2648*** 0.2560*** 0.2639*** 0.2715*** (6.89) (7.15) (5.75) (5.63) (5.68) (6.03) Controls Yes Yes Yes Yes Yes Yes Country-Year Fixed Effect Yes Yes Yes Yes Yes Yes Industry-Year Fixed Effect Yes Yes Yes Yes Yes Yes Observation 76,321 76,321 76,321 76,321 70,684 72,111 Adj. R-squared 0.892 0.903 0.896 0.749 0.803 0.815 Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 29

We can push a bit further We can evaluate an additional prediction from theory: If restrictions on insider trading spur innovation by improving valuations and therefore making it easier for such firms to raise money, then we should see them raise more money. Let s look at IPOs and SEOs. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 30

Empirical Results Growth of IPO and SEO proceeds in naturally innovative and naturally less innovative groups. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 31

Conclusion The enforcement of insider trading laws is associated with faster innovation. Consistent with Inefficiency of Insider Trading view: Increase is especially pronounced in naturally innovative and opaque industries. Equity issuances rise more these industries too. Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 32

Thank You! Levine, Lin, Wei Insider Trading and Innovation 1/17/2016 33