Patent Trolls and Small Business Employment Ian Appel† , Joan Farre-Mensa‡ and Elena Simintzi∗ February 2, 2017 Abstract We analyze how frivolous patent-infringement claims made by “patent trolls” affect small firms’ ability to create jobs, raise capital, and survive. Our identification strategy exploits the staggered passage of anti-patent-troll laws at the state level. We find that the passage of this legislation leads to a 2% increase in employment at small firms in high-tech industries, which are a frequent target of patent trolls. By contrast, the laws have no significant impact on employment at larger or non-high-tech firms. Anti-troll legislation is also associated with fewer business bankruptcies. Financing appears to be a key channel driving our findings: in states with an already established VC presence, the passage of anti-troll laws leads to a 19% increase in the number of firms receiving VC funding. Our findings suggest that measures aimed at curbing the litigation threat posed by patent trolls may play an important role in reducing both the real and financing frictions faced by small businesses.

Keywords: employment, patent trolls, venture capital. Affiliations: † Carroll School of Management, Boston College; ‡ Harvard Business School, Harvard University; ∗ Sauder School of Business, University of British Columbia. e-mails: [email protected], [email protected], [email protected].

I

Introduction

Small businesses are a key engine of employment and economic growth, but they are hindered by a number of frictions that limit their ability to grow, create jobs, and innovate (Himmelberg and Petersen 1994; Carpenter and Petersen 2002; Kerr and Nanda 2011). One friction that has received considerable attention in recent years are “patent trolls”, organizations that own patents but do not make or use the patented technology directly. Patent litigation has increased tenfold since 2000, and in 2015 trolls accounted for 69% of these lawsuits (RPX 2015). Perhaps even more worryingly, survey evidence suggests that those cases that end up in court may just be a small fraction of the instances in which trolls target firms, particularly small ones (Chien 2014). Indeed, one of the most notorious patent trolls, MPHJ, sent patent-infringement “demand letters” to over 16,000 small firms between 2012 and 2013 threatening them with litigation if they did not pay license fees on its scanner patents, but never filed a single lawsuit (FTC 2015). Examples abound of the negative consequences of patent trolls on firm growth. A case in point is Kunin, a Vermont technology company, whose CFO was cited in the Washington Post in 2013 as saying: “We had two projects that we were close to doing. Then our clients contacted us saying, ‘We’re being threatened by patent trolls’.” As a result, the Post reports, “both projects were cancelled, nixing the firm’s plans to hire several new employees.”1 Chien’s (2014) survey provides further evidence of the potential costs associated with patent trolls: 41% of the survey firms that had been targeted by a patent troll reported that the troll had had a significant impact on their operations, including delays in hiring and even having to close the business, with these impacts concentrated among the smaller firms. Motivated in part by these concerns, several bills have been introduced in Congress since 2012 to limit the activities of patent trolls, and in particular their ability to send abusive demand letters. However, as of today, none of them have become law. As Cohen, Gurun, and Kominers (2015) note, a common concern among those opposing restrictions on 1

“How Vermont could save the nation from patent trolls” (Aug. 1, 2013).

–1–

patent trolls is that trolls “serve a key financial intermediary role, policing infringement by well-funded firms that could otherwise infringe upon small inventors’ intellectual property at will” (p. 1). In response to the lack of congressional action, 31 state legislatures, beginning with Vermont in 2013, have passed state laws to limit patent trolls’ ability to target local businesses with bad-faith patent-infringement claims. In this paper, we use a difference-in-differences approach to provide the first analysis of the effect that these statelevel anti-patent-troll laws have had on small firms’ ability to create jobs, raise capital, and survive. We find that the passage of anti-patent-troll legislation in a state leads to an average 2% increase in employment at high-tech small firms located in the state, which are precisely the firms that tend to be the target of patent trolls (AIPLA 2013; Chien 2014). This increase is most significant for college-educated workers; that said, we also find somewhat weaker evidence of an employment increase for workers without a college degree, which highlights the complementary nature of their jobs. While the fact that many of the anti-troll laws have been passed in the last three years limits our ability to capture their long-term effects, we find no evidence that their effect on employment is short-lived—if anything, it appears to increase over time. By contrast, we find no significant effect on the employment at larger high-tech firms (those with more than 20 employees) or at non-high-tech firms regardless of their size. Our results thus indicate that the positive effect of anti-troll laws in protecting small high-tech firms from frivolous legal threats dominates on average any negative effect that the laws may have in limiting trolls’ ability to act as financial intermediaries for these firms. Empirically identifying the effects of policy changes such as the passage of anti-patenttroll legislation is plagued with endogeneity concerns. At the heart of these concerns is the possibility that there could be confounding variables, such as macroeconomic shocks, that may affect both a state’s decision to pass anti-troll legislation and the hiring decisions of firms in the state. Our diff-in-diff identification strategy helps overcome these endogeneity challenges by exploiting the fact that not all U.S. states have adopted anti-patent troll

–2–

laws (as of today, 19 have yet to do so, including some states with a large concentration of high-tech firms like California); in addition, among those states that have adopted them, not all of them did so at the same time. This staggered passage of the anti-patent troll laws allows us to construct a time-varying control group that provides a plausible counterfactual for how employment would have evolved in the treated states had they not adopted antipatent-troll legislation. For our control group to be a valid counterfactual, the employment evolutions at treated and control states need to share parallel trends. Consistent with this assumption, we find that employment trends at treated and control states were indistinguishable pre-treatment, both economically and statistically. While the parallel-trends assumption is ultimately untestable, the following four facts are also consistent with it. First, our results are robust to controlling for standard state-level macroeconomic variables: state aggregate and percapita income, unemployment, and patenting volume. These macroeconomic variables are in addition to the state and year-quarter fixed effects that we include in all specifications, which account for time-invariant employment differences across states and aggregate shocks affecting employment in all states, respectively. Second, if we estimate a placebo diff-in-diff on pre-event years, the estimated treatment effect is indistinguishable from zero. Third, the fact that the effect of the anti-troll laws is concentrated among small high-tech firms, which tend to be the focus of patent trolls’ frivolous demand letters, is consistent with our estimates capturing the effect of the laws and not some other state-level shock. Fourth, small firms are unlikely to be able to spend much on lobbying, thus alleviating the concern that our results may be driven by reverse causality or by some unobservable shock affecting both the growth opportunities and lobbying efforts of high-tech firms. What are the likely mechanisms driving the positive effect that state-level anti-troll laws have on small high-tech firms’ ability to create jobs and grow? By limiting the instances in which trolls target small firms with frivolous patent-infringement claims, anti-troll laws allow firms to avoid having to devote time and money to defending themselves from these claims. Consistent with this direct effect of the laws on improving small firms’ viability by

–3–

allowing them to conserve their often scarce resources, we find (somewhat noisy) evidence that the passage of anti-troll legislation in a state leads to a decrease in the number of business bankruptcies in the state.2 Perhaps even more importantly, by reducing the risk that any small, high-tech firm in a state faces of being targeted by a patent troll, anti-troll laws can increase the net present value (NPV) of investing in such firms in that state. This higher NPV may not only affect the investment decisions of the firms’ founders and managers, but also of professional investors such as venture capitalists (VCs). In line with this hypothesis, we find that, in states with above-median VC presence at the beginning of our sample period, anti-troll laws lead to a 19% increase in the number of firms raising VC funding and to a 29% increase in the capital invested by VCs in the state. By contrast, we find that the laws have no significant effect on VC investments in those states with low VC activity to begin with, which suggests that the passage of anti-troll laws is, on its own, not sufficient to attract VCs to states with historically scarce VC presence. Further reinforcing the importance of this financing channel, we find that the positive effect of anti-patent-troll laws on the employment of small high-tech firms is driven by states with above-median VC presence. In these states, we also find a positive effect of the laws on the number of high-tech establishments, thus suggesting that the growth in employment we observe is driven both by the growth of existing establishments and by the creation of new ones. Our study joins a small but growing literature that analyzes the economic consequences of patent trolls. Cohen, Gurun, and Kominers (2015) show that, when suing public firms, patent trolls tend to target those that are cash rich, leading to a substantial decrease in their innovative activities. Tucker (2014) and Smeets (2015) also find a decline in innovation at firms sued by patent trolls. Our study complements these papers by differing from them 2 By contrast, as expected, we find no effect of anti-troll laws in the number of non-business bankruptcies. This placebo test further alleviates concerns that the paper’s findings are driven by unobservable economic shocks that coincide, both in time and geography, with the passage of the laws.

–4–

in two distinct ways. First, prior work has focused on analyzing the effects of being sued by a patent troll. Lawsuits are indeed the main way how trolls can attract the attention of large firms like the public firms in Cohen et al.’s and Smeets’s samples.3 Crucially, though, our findings indicate that the simple threat of litigation expressed via the frivolous demand letters that state anti-troll laws aim to curtail can disrupt the operations of small high-tech firms. Second, prior studies focus on firms that have already been targeted by patent trolls. By contrast, our empirical setting captures both the realized and expected effects of being targeted by a troll by analyzing the effects of legal changes that reduce the risk that any firm in a state faces of being targeted by trolls. Our paper provides the first analysis of the effects of state anti-patent-troll laws. The adoption of anti-troll legislation has been surrounded by an intense debate on whether limiting the activities of patent trolls may do more harm than good by hindering patent trolls’ ability to help small innovative firms monetize their inventions.4 Also, a number of commentators and scholars have expressed doubts on whether states should legislate in an area, intellectual property protection, that has long been considered a federal matter—and, relatedly, whether such state-level laws would be effective in limiting the activities of patent trolls (e.g., DeSisto 2015). Our results indicate that state anti-troll laws have had a net positive effect for small firms in high-tech industries, helping them create jobs and making them more attractive to VC investors, and no significant effect for larger or non-high-tech firms. These findings suggest that anti-patent-troll laws can have a multiplier effect, by not only decreasing the resources that small firms’ need to spend on litigation expenses, but by also facilitating their access to the funding, monitoring, advice, and networks provided by VCs (Hellmann and Puri 2002; Bernstein, Giroud, and Townsend 2015; Hochberg, 3

Cohen et al. quote the response of one company executive to trolls that send demand letters as being: “If you have a truly viable case you will sue; otherwise don’t waste my time with this letter(!).” 4 See, e.g., “In Defense of the Dreaded ‘Patent Troll’ ”(Business Insider, Nov. 27, 2012), or “Not So Scary, After All: In Defense Of Patent Trolls” (Forbes, Feb. 1, 2013). As we explain in the next section, this debate is also a key reason why some big states like California have not passed a state anti-troll law.

–5–

Ljungqvist, and Lu 2007). Our paper also contributes to the literature in finance studying the effects of frictions that hinder firms’ ability to create jobs. Several papers have shown that frictions in financial markets can impede firm employment growth (e.g., Benmelech, Bergman, and Seru 2011; Chodorow-Reich 2014; Giroud and Mueller 2015, 2016; Benmelech, Frydman, and Papanikolaou 2016). Our study focuses on a different type of friction, showing that frivolous patent-infringement claims can limit small innovative firms’ ability to grow and create jobs. The remainder of the paper is organized as follows. Section II discusses the institutional background of the state anti-patent-troll laws we study. Section III summarizes the data and describes our methodology. Section IV presents the baseline results. Section V discusses the mechanisms that may explain our findings. Section VI concludes.

II

Institutional Background

The number of patent lawsuits has increased tenfold since 2000 (Cohen, Gurun, and Kominers 2015). Of late, much of this litigation has been brought by non-practicing entities or “patent trolls,” organizations that own patents but do not create or use the patented technology directly: The share of patent cases brought by patent trolls almost doubled between 2010 and 2015, increasing from 35% to 69% (RPX 2015; CEA 2016). Importantly, though, litigation is not the only way how allegations of patent infringement can be made. Before suing an alleged infringer, a patent owner often sends a “demand letter” to the alleged infringer. While such letters can take a variety of forms, common purposes are to inquire whether a product uses a particular patent, to make a licensing offer, and/or to threaten litigation if the alleged infringement is not stopped or a royalty is paid (AIPLA 2013). Patent demand letters are “essential enforcement tools for patent owners, and they allow many patent disputes to be resolved long before court intervention is necessary;” however, in recent years a number of patent trolls appear to have abused this tool by “sending thousands of demand letters to unknowing end users of allegedly patented prod-

–6–

ucts, using the recipient’s lack of experience with the patent system to coerce them into paying settlements” (AIPLA 2013; p. 18). One notorious troll is MPHJ, which between September 2012 and May 2013 sent demand letters to 16,465 small businesses in an attempt to collect license fees on the scanner patents it had assembled by threatening the firms with litigation that it did not actually intend to pursue (FTC 2015). Indeed, MPHJ never filed any lawsuits, probably because it knew that its patent-infringement claims were unlikely to withstand a legal challenge.5 Another example is Lodsys, which between 2011 and 2013 sued or threatened to sue thousands of iOS application developers. While Lodsys sued some large firms, “most of its targets [were] tiny app developers who lack[ed] the resources to defend patent litigation” (Nazer 2013). Systematic data on the number of businesses targeted by patent demand letters are not available.6 However, survey evidence indicates that, particularly among small firms, the number of targeted firms is substantially larger than the number of firms that end up being sued in court. In a survey of tech startups, Chien (2014) found that lawsuits represented only 31% of patent-infringement demands received by companies with under $10 million in revenue, while they represented 67% of demands received by companies with over $10 million in revenue.7 Accordingly, Scott Morton and Shapiro (2014) write that patent lawsuits “are just the tip of the iceberg. Surely, there are far more patent assertions than actual litigations, and these assertions impose various costs on targets, including legal expenses, design-around costs, and settlement costs” (p. 469). 5 See Figure A1 in the Internet Appendix for a sample demand letter sent by MPHJ. The website https://trollingeffects.org/letters contains examples of dozens of other demand letters sent by patent trolls. 6

Daniel Nazer, Staff Attorney at the Electronic Frontier Foundation (one of the organizations behind the website https://trollingeffects.org/) wrote to us in an email: “We’ve been disappointed with the low number of submissions of letters to Trolling Effects. We learned that companies are generally quite nervous about sharing information about demand letters.” 7

This evidence suggests that studies that use litigation data to analyze the effects of patent trolls on the behavior of publicly-listed firms (e.g., Cohen, Gurn, and Kominers 2015) are capturing the vast majority of instances in which trolls target a firm of the size of a public firm. However, litigation data provide a far less complete picture of the extent to which small and medium-sized firms are affected by patent trolls.

–7–

Chien’s (2014) survey shows that these costs can indeed be substantial, particularly for small firms. Among the 79 startups in her sample that had been targeted by a patent troll, the most common response was to fight it out of court. This strategy, which was followed by 24% of the targeted firms, resulted in average expenses of $168,000, or 5% of the firms’ annual revenue. For the 18% of firms that ended up entering into a settlement, the expenses increased to $340,000 (13% of annual revenue), while the 11% of firms that fought the demand in court faced average expenses of $857,000 (24% of annual revenue).8 In addition to these monetary costs, 73% of startups reported that dealing with the trolls’ demands had required founder time, and 89% acknowledged that doing so had been a distraction from their core business. The end result of these monetary and time costs was that 41% of targeted firms reported that the troll demand had had a significant impact on their operations, including delays in hiring or in meeting operational milestones, changes in business strategies or products, loss of firm value, and even having to exit a business line or the business altogether. Not surprisingly, these operational impacts were concentrated among the smaller firms.9 The costs associated with patents trolls have led a growing chorus of scholars and public commentators to ask policymakers to pass legislation curtailing trolls’ activities (e.g., Asay et al. 2015). In the US, patent law generally falls under the purview of the federal government. Article One of the Constitution states that the US Congress shall have power to “promote the progress of science and useful arts” by enacting legislation “securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.” The latest patent reform law passed by Congress, the America Invents Act (AIA) of 2011, includes a provision intended to curb abusive patent litigation by making it more difficult to sue multiple defendants in the same patent infringement suit (Bryant 8

The remaining firms did nothing (22% of targeted firms); conducted a product or business change (9%); or followed some other response (17%). 9

Of those firms with revenue below $100,000, 62% reported having suffered a significant operational impact, while for firms with revenue in the $100,000-$1 million range, the fraction was 55%. By contrast, no firm with revenue over $100 million reported that its operations were significantly impacted by the troll demand.

–8–

2012). However, the AIA appears to have had a limited impact in curbing the activities of patent trolls (CEA 2016); in particular, the AIA does not address trolls’ ability to use extrajudicial actions such as extortive demand letters to make frivolous patent demands. Since the passage of the AIA, Congress has considered several pieces of legislation aimed at further curtailing the activities of patent trolls, and in particular their ability to send abusive demand letters.10 Yet, as of this writing, none of them has become law. In response to concerns that the AIA does not sufficiently protect firms from bad-faith patent infringement claims, a number of state legislatures have decided to take action at the state level. Beginning with Vermont in 2013, state legislators have adopted patent reforms that protect local businesses from frivolous patent-infringement claims. These anti-patenttroll laws have been framed as consumer protection laws, thereby sidestepping the fact that patent legislation is considered to be a federal matter.11 The stated goal of these state-level anti-troll laws is to reduce the effect of abusive patent claims on the state economies. For example, the Vermont statute (Act 44) notes the following: Abusive patent litigation, and especially the assertion of bad faith infringement claims, can harm Vermont companies. A business that receives a letter asserting such claims faces the threat of expensive and protracted litigation and may feel that it has no choice but to settle and to pay a licensing fee, even if the claim is meritless. This is especially so for small and medium sized companies and nonprofits that lack the resources to investigate 10

These include the Targeting Rogue and Opaque Letters (TROL) Act (H.R. 2045), the Patent Transparency and Improvements Act (S. 1720), the Saving High-tech Innovators from Egregious Legal Disputes (SHIELD) Act (H.R. 845), the Innovation Act (H.R. 3309), the Stopping the Offensive Use of Patents (STOP) Act (H.R. 2766), the Transparency in Assertion of Patents Act (S. 2049), and the Demand Letter Transparency Act (H.R. 1896). 11 DeSisto (2015) writes about the Vermont law that “the Act attempts to apply consumer protection law to the methods of patent trolls and to avoid an evaluation of the underlying patents by focusing on one of the patent trolls’ favorite and most effective weapons: demand letters. By avoiding the evaluation of patents completely, and instead simply laying out a number of factors courts may consider when deciding whether a demand letter to an alleged patent infringer was sent in bad faith, the Act successfully dodges intrusion upon Congress’s exclusive [patent] jurisdiction” (p. 126).

–9–

and defend themselves against infringement claims. Not only do bad faith patent infringement claims impose a significant burden on individual Vermont businesses, they also undermine Vermont’s efforts to attract and nurture small and medium sized IT and other knowledge based companies. Funds used to avoid the threat of bad faith litigation are no longer available to invest, produce new products, expand, or hire new workers, thereby harming Vermont’s economy.12 The Vermont anti-troll law “seeks to change the calculations of patent trolls in Vermont by increasing the potential costs of sending out mass demand letters. [...] The Act attempts to tackle patent-troll demand letters by allowing a court to consider whether the letter had the required information, requested an unreasonable license fee, or demanded payment in an unreasonably short period of time as factors in determining whether the demand letter was sent in bad faith. If a [targeted firm] is able to establish a reasonable likelihood of a bad-faith assertion, then the asserting party is required to post a bond equal to a ‘good faith estimate of the [target’s] costs to litigate’ ” (DeSisto 2015; pp. 124-125). In addition, the law establishes that if a court finds that a Vermont firm has been the target13 of badfaith patent infringement assertions, then the court may award it the following remedies: “(1) equitable relief; (2) damages; (3) costs and fees, including reasonable attorney’s fees; and (4) exemplary damages in an amount equal to $50,000.00 or three times the total of damages, costs, and fees, whichever is greater.” In Vermont, the push for patent reform was led by a tech firm with around 100 employees 12 The full text of Vermont’s Act 44 is available here: http://www.leg.state.vt.us/docs/2014/Acts/Act044.PDF. A summary of the legislation can be found here: http://apps.americanbar.org/litigation/committees/intellectual/articles/winter20141213-vermont-enacts-legislation-against-patent-trolls.html. 13

The law defines a “target” as a Vermont (physical or legal) person: “(A) who has received a demand letter or against whom an assertion or allegation of patent infringement has been made; (B) who has been threatened with litigation or against whom a lawsuit has been filed alleging patent infringement; or (C) whose customers have received a demand letter asserting that the person’s product, service, or technology has infringed a patent.”

– 10 –

that eventually started a grassroots coalition with other local firms.14 The Vermont law has served as a model for other states, and as of this writing a total of 31 states have adopted anti-patent-troll laws similar to Vermont’s.15 All the statutes passed by these 31 states share two critical components: (1) they aim to curtail bad-faith patent-infringement assertions made via demand letters by allowing courts to impose penalties on the senders of such letters; and (2) they cover any target firm located within that state, regardless of where the firm is incorporated or where the sender of the letter is located. While the anti-patent-troll laws have spread quickly, the states that have yet to pass anti-troll legislation include some of the largest and most innovation-active states, most notably California. In fact, an anti-troll law was introduced in the California State Senate in February 2015 (S.B. 681),16 but it was not passed. The office of State Senator Hill, who introduced S.B. 681, wrote to us in an email that the challenge in California was that legislators could not agree on a text that a sufficient majority thought would deter “fraudulent patent infringement claim letters” without having “a chilling effect on legitimate communications between small innovators trying to engage in coordination, development and licensing communications with larger businesses.” In fact, similar disagreements have been at the root of the U.S. Congress’s inability to pass analogous anti-troll legislation at the federal level.

III III.1

Data and Methodology Data

We obtain state-level employment data from the Census’ Quarterly Workforce Indicators (QWI) database. Employment is measured at the end of each quarter (EmpEnd). Hires 14

“How Vermont could save the nation from patent trolls,” (The Washington Post, Aug. 1, 2013).

15

Table A1 in the Internet Appendix provides a list of the states along with the dates the legislation was signed into law in each state. 16 The proposed text is available here: http://www.leginfo.ca.gov/pub/15-16/bill/sen/sb_06510700/sb_681_bill_20150227_introduced.html.

– 11 –

(HirAEnd) include workers that begin a new job in the quarter that continues into the next quarter. Separations (Sep) include workers whose job with a given employer ended in the specified quarter. Payroll (Payroll) is firms’ total payroll in a given quarter. For each state-quarter, the QWI reports employment data aggregated by firm size and 4-digit NAICS industry. The QWI reports data for five size categories (in employees): 0-19, 20-46, 50-249, 250-499, and 500+. In the interest of space, in our analyses we collapse the three largest size categories into a single group, and report results for small (<20 employees), medium (20-50), and large (50+) firms.17 In much of the paper, we focus on high-tech employment, which we classify using 4digit NAICS codes following Kile and Phillips (2009).18 In addition, we use demographic information from the QWI to analyze how the effect of anti-patent-troll laws on employment depends on the workers’ education level. Our employment analysis sample begins in 2012Q1, five quarters before the passage of the first anti-patent-troll law in Vermont, and ends in 2015Q4, the last quarter for which QWI are available. We obtain establishment count data from the Quarterly Census of Employment and Wages (QCEW) database. The data are available at the industry-state-year-quarter level, and we again focus on high-tech industries. Venture capital (VC) data are from Thomson’s VentureXpert, and we focus on early-stage VC rounds raised by high-tech firms. Statewide bankruptcy data are from the Administrative Office of the United States Courts. Bankruptcies are classified as business bankruptcies if the debtor is a corporation or partnership, or if debt related to the operation of a business predominates; non-business bankruptcies include consumer debt and other debt that is not consumer or business debt. Establishment, VC, and bankruptcy data are available through 2016Q2. We present all our results with and without the following macroeconomic controls: statelevel quarterly gross state product (GSP) and income per capita, both log-transformed, 17

Our results remain qualitatively the same if we instead use the original five categories.

18

Specifically, the following 4-digit NAICS industries are classified as high-tech: 3254, 3341, 3342, 3344, 3345, 3346, 3353, 3391, 5112, 5133, 5141, 5172, 5173, 5179, 5181, 5182, 5413, 5415, 5416, 5417, 5161, 5171, 5179, 5191.

– 12 –

from the Bureau of Economic Analysis; state-level quarterly unemployment rates, from the Bureau of Labor Statistics; and state-level patenting activity, measured by the log number of patents granted in a given state-year as reported by the United States Patent and Trademark Office. Table 1, Panel A reports summary statistics for our entire sample. The average state has approximately 2.1 million people employed. Just over 400,000 of them are employed in establishments with fewer than 20 employees, and 166,000 are in industries that we classify as high-tech. On average, there are 256,000 new hires each quarter and 354,000 separations. Aggregate payroll in the mean state is $28 billion. The number of business bankruptcies averages 150 per state-quarter, while the number of non-business bankruptcies averages 4,772. On average, there are 15 unique firms that receive VC funding in each state-quarter and they raise an aggregate dollar amount of $138 million. As expected, there is substantial variability for both VC variables—their standard deviation is over three times their mean value—reflecting the high degree of concentration of VC activity in a few states (most notably, California). Summary statistics for our macroeconomic controls are as follows: The mean state-level quarterly GSP is $309 billion, income per capita averages $44,800, the number of granted patents in the average state-quarter is nearly 3,000, and the average unemployment rate during our sample period is 6.2%. Panel B compares the mean values of all our outcome and control variables for states with and without patent troll laws, measured in all cases during the four quarters prior to the passage of the first anti-patent-troll law in Vermont. Specifically, Column 1 reports the mean of each variable over 2012Q2 – 2013Q1 for those states that pass an anti-patenttroll law at some point during our sample period, while Column 2 reports the mean for those states that pass no anti-troll legislation; the difference between these means and the corresponding p-values (accounting for clustering at the state-level) are reported in Columns 3 and 4, respectively. While the treated states tend to be smaller prior to the passage of anti-troll legislation than the never-treated states, this difference is not statistically significant. The never-

– 13 –

treated states also concentrate an insignificantly higher level of VC activity, a finding that is largely driven by the presence of California in the never-treated group. To ensure that California is not driving our conclusions, Table A4 in the Internet Appendix shows that our main findings are robust to excluding California from the sample.

III.2

Methodology

To identify the effect of the passage of anti-patent-troll laws on state-level outcomes such as employment and VC investment, we estimate the following difference-in-differences specification at the state-year-quarter level:

yst = δ · P atentLawst + β · Xst−1 + λs + αt + st ,

(1)

where s denotes a state and t denotes a year-quarter. P atentLaw is an indicator set equal to one if state s has passed anti-patent-troll legislation on or before quarter t, and zero otherwise; Xst−1 is the vector of state-level control variables (GSP, income per capita, unemployment rate, and patent activity); λs is a state fixed effect (to control for timeinvariant state characteristics); and αt is a year-quarter fixed effect (to control for aggregate shocks affecting all states). In all specifications, we report robust standard errors clustered at the state level. An important feature of this empirical setting is that the control group allows the same state to be part of the treated and control groups at different points in time. Specifically, at any year-quarter t, the control group includes both states that passed anti-troll legislation after year-quarter t (but before the end of our sample period) and so that will be treated eventually, and states that are never treated (either because they have not yet passed antitroll legislation or did so after the end of our sample period). The staggered nature of the laws that we study thus helps reduce the identification problems that may result in settings where one analyzes a single event, where the estimates are more likely to be confounded by the presence of another contemporaneous shock (Roberts and Whited 2012).

– 14 –

IV

Results: Patent Troll Laws and Employment

We begin by examining how the passage of anti-patent-troll laws in a state affects employment at firms of different sizes and industries in that state. Table 2 presents the results. Column 1 shows that the passage of anti-patent-troll legislation in a state leads to a 2% increase in employment at small (0-19 employees), high-tech firms in the state (this effect is significant at the 5% level). Column 2 shows that this finding is robust—and, if anything, the estimated treatment effect becomes slightly larger—to including our state-level macroeconomic controls. This robustness helps alleviate the concern that our results may reflect the possibility that a state’s economic conditions could affect both the passage of anti-troll legislation and employment. Further alleviating this concern, and consistent with the parallel-trends assumption necessary for identification in our setting, Table A2 in the Internet Appendix shows that a state’s economic conditions are not a significant predictor of whether the state will pass anti-troll legislation in the next quarter or year. Columns 3 and 4 of Table 2 report the effect of anti-troll laws on employment at medium (20-49 employees) and large (50+ employees) firms in high-tech industries. In contrast to the results in Columns 1 and 2, we find no evidence that the laws are associated with economically or statistically significant changes in employment at medium or large hightech firms. Nor do we find any effect for firms in non-high-tech industries, regardless of their size (Columns 5 through 7). Taken together, the results in Table 2 indicate that the effects of state anti-patenttroll laws are primarily confined to small businesses in high-tech industries. This finding is consistent with the notion that the frivolous patent demand letters that are the focus of these laws are particularly costly for small firms, which tend to have little experience with the patent system and are more likely to be coerced into settlements to avoid the threat of litigation (AIPLA 2013; Chien 2014). The fact that we only observe an increase in employment in small high-tech firms also helps mitigate the concern that our estimates may be confounded by an omitted variable correlated with the adoption of anti-patent-troll

– 15 –

laws. To explain our findings, such an omitted variable would not only need to coincide in time and space with the staggered adoption of the laws, but it would also need to differentially affect small high-tech firms. Figure 2 helps visualize how employment in small high-tech firms changes around the passage of anti-patent-troll legislation. Specifically, the figure plots the estimated average quarterly employment change in treated states relative to control states in quarters t– 5 through t+6, where quarter t is the quarter when the anti-patent-troll law is signed into law, alongside the 95% confidence interval of these estimates.19 Consistent with the parallel-trends assumption, we find no significant difference in the evolution of employment at treated and control states prior to the passage of anti-troll legislation. As expected given that employment adjustments are not immediate, we also find no significant difference in employment the quarter that the anti-troll law is signed. However, we begin to observe a significant increase in employment in the treated states relative to the control states the quarter after the legislation is passed. The size of this increase in employment grows steadily over the following five quarters, becoming a highly significant 6.4% increase (p<0.001) six quarters after the anti-troll law is passed. To further examine the internal validity of our identification strategy, Table A3 in the Internet Appendix presents the results of a falsification (or placebo) test. Specifically, we falsely assume that each state that passed an anti-patent-troll law did so three years before the actual law was passed in the state. For example, we assume that Vermont passed its anti-troll law in May 2010 instead of in (the actual date of) May 2013. The placebo sample goes from 2009 to 2012 to be consistent with the panel duration of our baseline analysis. As expected, and contrary to our baseline results in Table 2, the placebo treatment effect is not statistically or economically significant, which supports the validity of our identification strategy. 19

As in Column 2 of Table 2, the underlying analysis includes macroeconomic controls as well as state and year-quarter fixed effects. For treated states, the sample is limited to quarters t–6 through t+6 (the presence of state fixed effects implies that the t–6 indicator is not identified). We focus on the six quarters before and after the adoption of the laws because employment data for most states that adopt anti-troll laws are not available past this window.

– 16 –

Next, we examine whether the increase in employment captured in Table 2 is driven by firms hiring more new employees or firing fewer existing employees. The results of this analysis are presented in Table 3, where, following our findings in Table 2, we focus on small firms in high-tech industries. The dependent variable in Columns 1 and 2 is the natural logarithm of the number of new hires, defined as the number of workers who started a new job in the specified quarter that continued into the next quarter; the dependent variable in Columns 3 and 4 is the logarithm of the number of workers whose job with a given employer ended in the specified quarter. Columns 1 and 2 show that the adoption of anti-troll legislation leads to a 3.6% increase in new hires in treated states relative to control states (in Column 1, the effect is statistically significant at the 10% level, while in Column 2 the inclusion of macroeconomic controls increases the significance to the 5% level). By contrast, we do not find a significant effect of anti-troll laws on employee separations.20 Columns 5 and 6 in Table 3 investigate the effect of anti-patent-troll laws on the aggregate payroll of small high-tech firms. Consistent with the fact that the laws lead to an increase in employment at these firms, we find that their aggregate payroll also goes up: Total wages increase by 3%, an increase that is statistically significant at the 5% level and is robust to the inclusion of macroeconomic controls. This increase in total wages is, if anything, larger than the increase in employment estimated in Table 2, thus suggesting that the increase in employment is not driven by hires of lower paid workers. Table 4 sheds further light on what type of employment increases in states with antipatent-troll laws by comparing the employment change of workers with and without a college degree. We maintain our focus on small high-tech firms. Consistent with the high concentration of college-educated workers in high-tech industries, we find that the increase 20 Our proxy for firings does not only capture involuntary separations, but also voluntary separations of workers who change firms. The positive (albeit insignificant) coefficient estimate we find in Columns 3 and 4 of Table 3 may therefore reflect the fact that the increase in the supply of jobs at small high-tech firms following the passage of anti-troll legislation increases the likelihood that workers voluntarily switch jobs.

– 17 –

in employment is particularly significant among such workers. Specifically, anti-troll laws are associated with a 1.9%-2% increase in the employment of college-educated workers, an effect that is significant at the 5% level when we do not include state-level macroeconomic controls in Column 1 and at the 1% level when we include them in Column 2.21 Interestingly, we also find evidence of an increase in employment for workers without college degrees: When we focus on workers without a college degree in Columns 3 and 4, we find that the point estimates of the effect of anti-patent-troll laws on employment are only slightly smaller than those for college-educated workers, though the non-college estimates are statistically noisier. This finding may reflect the complementary nature of the jobs of high- and low-skilled employees, suggesting that the benefits of protecting small high-tech firms from being targeted by patent trolls can be felt by workers across the wage and skill distributions.

V

Mechanisms: Patent Troll Laws, Firm Survival, and Firm Financing

Section IV shows that the passage of state-level anti-patent-troll legislation leads to an increase in employment at small high-tech firms. In this section, we aim to shed light on the potential mechanisms driving this effect. We begin by examining whether anti-troll laws have a direct effect in improving the viability of small firms by testing whether the laws are associated with a decrease in business bankruptcies. We then turn to studying the laws’ effect on the financing environment faced by small innovative firms by analyzing how the passage of anti-troll legislation in a state impacts venture capital investment in that state. 21

As expected, Table A5 in the Internet Appendix shows that this finding is exclusive to firms in high-tech industries.

– 18 –

V.1

Patent Troll Laws and Business Bankruptcies

Small, privately-held firms face important financing constraints (e.g., Kerr and Nanda 2011; Farre-Mensa and Ljungqvist 2016). Those that receive a demand letter from a patent troll often have to devote their limited financial and time resources to hiring legal counsel to counter the troll’s patent-infringement allegations, even when the allegations are baseless. As noted in Section II, Chien’s (2014) survey finds that the costs incurred by firms that have to defend themselves from these allegations can be substantial, particularly for small firms. These costs can not only limit the firms’ ability to grow and create jobs, as documented in Section IV, but they could even force some of the firms into bankruptcy. Table 5 examines the effect that the passage of anti-troll legislation in a state has on the number of business bankruptcies in that state. Column 1 shows that anti-patent-troll laws are associated with a 3.8% decrease in the number of business bankruptcies in a state, although this estimate is not statistically significant (p=0.199). When we include state-level macroeconomic controls in Column 2, the estimated reduction in the number of bankruptcies increases to 4.9% and becomes significant at the 10% level. The evidence in Table 5 is thus consistent with the notion that by protecting small firms from being frivolously targeted by patent trolls, anti-troll laws have a positive effect on firms’ probability of survival. As expected, Columns 3 and 4 in Table 5 show that anti-troll laws have no economically or statistically significant effect on the number of non-business bankruptcies in a state. This placebo test helps alleviate concerns that our previous estimates of the effects of anti-patenttroll laws may be confounded by unobservable economic improvements that coincide, both in time and geography, with the passage of the laws—indeed, such economic improvements would likely decrease both the number of business and non-business bankruptcies.

V.2

Patent Troll Laws and Venture Capital Investment

In addition to reducing the instances of bad-faith patent-infringement demands and thus the direct costs that these demands have on the targeted firms, anti-troll legislation also

– 19 –

reduces the risk faced by all the firms in a state of at some point being targeted by trolls. By reducing the risk posed by patent trolls, state anti-troll laws can increase the net present value (NPV) of investing in the high-tech small firms that trolls often target, thereby making the firms more attractive to professional investors such as venture capitalists (VCs). In fact, in a survey of 114 VCs, Chien (2013) finds that 75% of surveyed VCs had been impacted by a demand from a patent troll to one (or more) of their portfolio companies. Of the 42 VCs that responded to the question of whether patent trolls were harming innovation in their industry, 67% strongly agreed that they were, and an additional 17% agreed. By contrast, 78% of these 42 VCs strongly disagreed with the notion that the ability of their portfolio companies to enforce or monetize their patents through trolls helped innovation, and 17% disagreed. Similarly, in a separate survey of 200 VCs, Feldman (2014) finds that 48% of respondents said that they would refrain from investing in a new company that had an existing patent demand, as they saw patent demands as a “major deterrent” to investing;22 52% of respondents answered that they may refrain from investing in the company, depending on the situation; and no VCs answered that the patent demand would not be a factor in their investment decision. Table 6 investigates whether the risk posed by patent trolls affects the investment decisions of venture capitalists by examining how the passage of anti-troll legislation in a state affects VC investment in the state.23 Column 1, which is estimated using all 50 U.S. states, shows that anti-troll laws are associated with a 6.8% average increase in the number of firms raising venture capital in treated states relative to control states, though this effect is not statistically significant (p=0.256). We continue to find positive but 22

In Feldman’s survey, patent demands “include letters indicating that the recipient may be infringing a patent and demanding a license fee, threats of litigation, or lawsuits” (p. 238). 23

Venture capital is not the only source of funding for startups, but prior work has shown that it is particularly helpful to the kinds of innovative startups that tend to be targeted by patent trolls (e.g., Hellmann and Puri 2000; Kortum and Lerner 2000; Gompers and Lerner 2001). In addition to offering funding, VCs provide monitoring and advice (Hellmann and Puri 2002; Bernstein, Giroud, and Townsend 2015), access to networks of potential customers, suppliers, and strategic partners (Hochberg, Ljungqvist, and Lu 2007), and help recruiting talented individuals (Gorman and Sahlman 1989).

– 20 –

insignificant estimates in Column 2 (where we include state-level macroeconomic controls) and in Columns 3 and 4, where the dependent variable is the dollar amount of capital invested by VCs in the state (also log-transformed). Recall that Table 1 showed that, in the median state-quarter, only three firms raise venture capital, and so the analysis in Columns 1-4 includes many states that have historically had very low levels of VC activity. When, in Columns 5-8, we exclude states with belowmedian VC activity in 2012 (the year before the first anti-patent-troll law was passed), we find that the effect of anti-troll laws on VC investment becomes stronger, both economically and statistically. Indeed, Column 5 shows that the number of firms raising venture capital increases by 18.7% with the passage of anti-troll legislation (an effect significant at the 5% level). This growth in the number of firms raising venture capital translates into a 28.9% increase in the amount of capital invested by VCs in the state (significant at the 10% level; Column 7). These estimated effects are only slightly smaller when we include state-level macroeconomic controls in Columns 6 and 8. The evidence in Table 6 is consistent with the hypothesis that by reducing the threat posed by bad-faith patent-infringement claims to small innovative firms in a state, antitroll laws make firms in that state more attractive to venture capitalists—although this effect does not appear to be strong enough to draw VCs to states with historically low VC presence. Our results also support the notion that the negative association between the instances of patent litigation by frequent litigators (a proxy for patent trolls) in a federal court district and VC investment in that district found by Kiebzak, Rafert, and Tucker (2016) has a causal interpretation.24 Further reinforcing the importance of the VC financing channel, Table 7 shows that the positive effect of anti-patent-troll laws on the employment of small high-tech firms is driven by states with a level of VC activity in 2012 at or above the median of all 50 24 As we have discussed in Section II and Kiebzak et al. themselves note , the cases that are part of Kiebzak et al.’s litigation sample do not include the “many occasions where patent litigation is threatened through a demand letter and a company chooses to settle rather than face the risks and expense of litigation” (p. 221).

– 21 –

states. Column 1 shows that in high-VC-presence states, anti-patent-troll laws lead to a 2.2% increase in the employment of small firms in high-tech industries (significant at the 5% level), and a 2.4% increase when we include macroeconomic controls in Column 2 (significant at the 1% level).25 By contrast, the magnitude of the employment increase in states with below-median VC presence is 1.0%-1.1% and insignificant (Columns 3 and 4). Similarly, Table 8 shows that in high-VC-presence states, anti-troll laws lead to an increase in the number of high-tech establishments. In Column 1, the estimated increase is 2.4% (significant at the 5% level), and it increases to 2.9% in Column 2 when we include state-level macroeconomic controls (significant at the 1% level). By contrast, Columns 3 and 4 again show no significant effect in states with below-median VC presence. The results in Table 8 thus suggest that the growth in employment at small high-tech firms after the passage of anti-troll legislation found above is driven both by the growth of existing establishments and by the creation of new ones.

VI

Conclusion

We investigate the impact of the staggered passage of anti-patent-troll laws in 31 U.S. states on small-firm employment, survival, and financing. The laws aim to limit patent trolls’ ability to target small firms with bad-faith patent-infringement claims that try to coerce the firms into quick settlements to avoid being exposed to the monetary and time costs of litigation. We show that, following the passage of anti-troll legislation, employment at small high-tech firms in a state increases by 2%; by contrast, the laws appear to have no significant effect on employment at larger high-tech firms or at non-high-tech firms of any size. We also find that, in states with an established VC presence, anti-troll laws lead to a 19% increase in the number of firms raising VC funding and a 29% increase in the dollar amount invested by VCs in the state. Thus, by reducing the risk posed by frivolous patent demands, anti-troll laws appear to increase the NPV of investing in small innovative firms 25

As expected, Table A6 in the Internet Appendix shows that this finding is exclusive to firms in high-tech industries.

– 22 –

in a state, thereby improving the financing environment faced by firms in the state. Chien (2014) writes of the America Invents Act of 2011, the latest patent reform law passed by Congress: “Reforms to reduce the cost of litigation defense are laudable, and likely deter some suits from being brought in the first place, but do not reach small companies against whom litigation is threatened, but not brought” (p. 462). The results in our paper indicate that measures aimed at curbing trolls’ ability to target small firms with bad-faith patent-infringement claims can play an important role in reducing both the real and financing frictions faced by small innovative firms. To the extent that these firms are key engines of productivity and economic growth (Aghion and Howitt 1992; Acemoglu and Akcigit 2012; Akcigit and Kerr 2016), the benefits of such measures could be felt economy-wide.

– 23 –

References [1] Acemoglu, D., and U. Akcigit, 2012, “Intellectual Property Rights Policy, Competition and Innovation,” Journal of the European Economic Association 10, 1-42. [2] Akcigit U., and W.R. Kerr, 2016, “Growth through Heterogeneous Innovations,” Journal of Political Economy, forthcoming. [3] American Intellectual Property Law Association (AIPLA), 2013, Testimony of Todd Dickinson before the U.S. Senate Judiciary Committee. Hearing on “Protecting Small Businesses and Promoting Innovation by Limiting Patent Troll Abuse.” [4] Aghion, P., and P. Howitt, 1992, “A Model of Growth through Creative Destruction,” Econometrica 60, 323-351. [5] Asay, C. D., et al., 2015, Letter to Congress (March 3). Available at: http://www.utdallas.edu/ ugg041000/IPScholarsLettertoCongress_March_2_2015.pdf. [6] Benmelech E., N. Bergman, and A. Seru, 2011, “Financing Labor,” Working Paper. [7] Benmelech E., C. Frydman, and D. Papanikolaou, 2016, “Credit Market Disruptions and Employment during the Great Depression: Evidence from Firm-level Data,” Working Paper. [8] Bernstein, S., X. Giroud, and R. R. Townsend, 2015, “The Impact of Venture Capital Monitoring,” Journal of Finance 71, 1591-1622. [9] Bryant, T., L., 2012, “The America Invents Act: Slaying Trolls, Limited Joinder,” Harvard Journal of Law & Technology 25, 674-695. [10] Carpenter R. E. and B. C. Petersen, 2002, “Is the Growth of Small Firms Constrained by Internal Finance?,” Review of Economics and Statistics 84, 298-309. [11] Chien, C. V., 2013, “Patent Assertion and Startup Innovation,” Working Paper. [12] Chien, C. V., 2014, “Startups and Patent Trolls,” Stanford Technology Law Review 17, 461-506. [13] Chodorow-Reich, G., 2014, “The Employment Effects of Credit Market Disruptions: Firm-level Evidence from the 2008-09 Financial Crisis,” Quarterly Journal of Economics 129, 1-59. [14] Cohen, L., U. G. Gurun, and S. D. Kominers, 2015, “Patent Trolls: Evidence from Targeted Firms,” Working Paper. [15] Council of Economic Advisers, 2016, “The Patent Litigation Landscape: Recent Research and Developments.”

– 24 –

[16] DeSisto, R., 2015, “Vermont vs. the Patent Troll: Is State Action a Bridge Too Far?,” Suffolk University Law Review 109, 109-130. [17] Farre-Mensa, J., and A. Ljungqvist, 2016, “Do Measures of Financial Constraints Measure Financial Constraints?” Review of Financial Studies 29, 271-308. [18] Federal Trade Commission (FTC), 2015, Complaint Against MPHJ Technology Investments. Docket No. C-4513. [19] Feldman, R., 2014, “Patent Demands and Startup Companies: The View from the Venture Capital Community,” Yale Journal of Law and Technology 16, 236-284. [20] Gompers, P., and J. Lerner 2001, “The Venture Capital Revolution,” Journal of Economic Perspectives 15, 145-168. [21] Gorman, M., and W. A. Sahlman 1989, “What Do Venture Capitalists Do?” Journal of Business Venturing 4, 231-248. [22] Giroud, X., and H. Mueller, 2015, “Firm Leverage, Consumer Demand, and Employment Losses during the Great Recession,” Quarterly Journal of Economics, forthcoming. [23] Giroud, X., and H. Mueller, 2016, “Redistribution of Local Labor Market Shocks through Firms’ Internal Networks,” Working Paper. [24] Hellman, T., and M. Puri, 2000, “The Interaction Between Product Market and Financing Strategy: The Role of Venture Capital,” Review of Financial Studies 134, 959-984. [25] Hellman, T., and M. Puri, 2002, “Venture Capital and the Professionalization of StartUp Firms: Empirical Evidence,” Journal of Finance 57, 169-197. [26] Himmelberg, C. P. and B. C. Petersen, 1994, “R&D and Internal Finance: A Panel Study of Small Firms in High-Tech Industries,” The Review of Economics and Statistics 76, 38-51. [27] Hochberg, Y. V., A. Ljungqvist, and Y. Lu, 2007, “Whom You Know Matters: Venture Capital Networks and Investment Performance,” Journal of Finance 62, 251-301. [28] Kerr, W. R., and R. Nanda, 2011, “Financing Constraints and Entrepreneurship,”in David Audretsch, Oliver Falck, and Stephan Heblich, eds. Handbook of Research on Innovation and Entrepreneurship (Cheltenham, U.K.: Edward Elgar Publishing). [29] Kiebzak, S., G. Rafert, and C. E. Tucker, 2016, “The Effect of Patent Litigation and Patent Assertion Entities on Entrepreneurial Activity,” Research Policy 45, 218-231. [30] Kile, C. O., and M. E. Phillips, 2009, “Using Industry Classification Codes to Sample High-Technology Firms: Analysis and Recommendations,” Journal of Accounting, Auditing & Finance 24, 35-58.

– 25 –

[31] Kortum, S., and J. Lerner, 2000, “Assessing the Contribution of Venture Capital to Innovation,” RAND Journal of Economics 31, 674-692. [32] Roberts, M. R., and T. M. Whited, 2012, “Endogeneity in Empirical Corporate Finance,”in George Constantinides, Milton Harris, and Rene Stulz, eds. Handbook of the Economics of Finance 2, 493-572. [33] RPX, 2015, 2015 Report: Patent Litigation, Patent Marketplace, and NPE Cost. [34] Scott Morton, F., and C. Shapiro, 2014, “Strategic Patent Acquisitions,” Antitrust Law Journal 79, 463-499. [35] Smeets, R., 2015, “Does Patent Litigation Reduce Corporate R&D? An Analysis of US Public Firms,” Working Paper. [36] Tucker, C. E., 2014, “Patent Trolls and Technology Diffusion: The Case of Medical Imaging,” Working Paper.

– 26 –

Figure 1: Map of Patent Troll Laws This figure shows states that have adopted a patent troll law. Signing dates are provided in Table A1.

– 27 –

Figure 2: Evolution of Employment Around the Passage of Anti-Troll Laws This figure shows the evolution of employment at small high-tech firms in states that pass an antipatent-troll law relative to those states that do not pass an anti-troll law. We estimate our baseline employment regression as in Table 2, with state and year-quarter fixed effects and macroeconomic controls, except that we replace Patent Law with indicators that identify quarters t–5, t–4, ..., t, t+1, ..., t+6 for states that pass an anti-troll law, where quarter t is the quarter the anti-troll law is signed. Accordingly, for treated states, the sample is limited to quarters t–6 through t+6 (the presence of state fixed effects implies that the t–6 indicator is not identified).The figure plots the point estimates associated with each of these indicators, alongside their 95% confidence interval. Robust standard errors are clustered by state.

0.1 0.08 0.06 0.04 0.02 0 -0.02 -0.04 t-5

t-4

t-3

t-2

t-1

t

– 28 –

t+1

t+2

t+3

t+4

t+5

t+6

Table 1: Summary Statistics Panel A reports summary statistics for the full sample. Panel B compares the mean values of all variables for states with and without patent troll laws, measured in all cases during the four quarters prior to the passage of the first anti-patent-troll law in Vermont. Specifically, Column 1 reports the mean of each variable over 2012Q2 – 2013Q1 for those states that pass an anti-patent troll law at some point during our sample period, while Column 2 reports the mean for those states that pass no anti-troll legislation; the difference between these means and the corresponding p-values (accounting for clustering at the state-level) are reported in Columns 3 and 4, respectively.

Panel A N

Mean

Median

SD

Total Employment, Thousands

788

2086.1

1348.4

2284.2

Employment (<20), Thousands

788

408.1

257.9

481.1

Employment (Tech), Thousands

788

166.4

92.4

222.8

Hires, Thousands

788

255.7

175.3

284.3

Separations, Thousands

788

353.8

236.0

401.9

Payroll, Billions

788

28.2

15.8

34.9

Bankruptcies (Business)

900

150.1

92.0

196.9

Bankruptcies (Non-Business)

900

4772.0

3247.5

5562.6

Establishments (Tech), Thousands

900

14.7

8.7

16.6

VC (# Unique Firms)

900

15.1

3.0

47.7

VC ($ Amount), Millions

900

137.8

9.0

649.2

GSP, Billions

788

308.7

188.8

377.5

Income Per Capita, Thousands

788

44.8

43.7

7.3

Patents

788

2985.9

1100.0

5905.9

Unemployment Rate

788

6.2

6.2

1.7

– 29 –

Panel B Never Treated

Eventually Treated

Difference

P-value

Ln(Total Employment)

14.15

13.97

0.18

0.55

Ln(Employment (<20))

12.52

12.41

0.11

0.70

Ln(Employment (Tech))

11.43

11.26

0.17

0.41

Ln(Hires)

12.00

11.87

0.13

0.66

Ln(Separations)

12.32

12.19

0.13

0.66

Ln(Payroll)

23.60

23.34

0.26

0.43

Ln(Bankruptcies (Business))

4.73

4.59

0.14

0.67

Ln(Bankruptcies (Non-Business))

7.96

7.97

-0.01

0.97

Ln(Establishments (Tech))

9.07

9.04

0.03

0.92

Ln(VC (# Unique Firms))

1.70

1.31

0.40

0.35

Ln(VC ($ Amount))

2.35

2.02

0.34

0.57

Ln(GSP)

12.26

12.00

0.26

0.39

Ln(Income Per Capita)

10.71

10.65

0.06

0.20

Ln(Patents)

7.13

6.91

0.22

0.62

Unemployment Rate

7.47

7.06

0.42

0.40

– 30 –

Table 2: Employment by Firm Size This table reports the effect of state anti-patent-troll laws on employment at firms of different sizes and industries. The dependent variable in Columns 1-4 is state employment in high-tech industries, and the dependent variable in Columns 5-7 is state employment in non-high-tech industries. Patent Law is an indicator for whether a state has passed anti-patent-troll legislation on or before that quarter. The list of states and corresponding signing dates is provided in Table A1. The control variables are defined in the text. Each specification includes state and year-quarter fixed effects. Robust standard errors are clustered by state. ***p<1%. ** p<5%, * p<10%.

Tech Firms Firm Size=

Patent Law

Non-Tech Firms

0-19

0-19

20-50

51-500+

0-19

20-50

51-500+

(1)

(2)

(3)

(4)

(5)

(6)

(7)

0.0197**

0.0203**

-0.0051

-0.0002

-0.0021

-0.0022

0.0015

(0.00985) (0.00812) (0.0202) (0.0106) Ln(GSP)

(0.00522) (0.00582) (0.00453)

0.643*** (0.201)

Ln(Income per Capita)

0.124 (0.248)

Unemployment Rate

-0.0070 (0.0043)

Ln(Patents)

0.0149 (0.0375)

State FE

yes

yes

yes

yes

yes

yes

yes

Year-Quarter FE

yes

yes

yes

yes

yes

yes

yes

Observations

788

788

788

788

788

788

788

0.026

0.176

0.001

0.000

0.000

0.001

0.001

Within R-squared

– 31 –

Table 3: Small Business Hires, Separations, and Payroll – High-Tech This table reports the effect of state anti-patent-troll laws on hires, separations, and payroll at small firms (<20 employees) in high-tech industries. The dependent variables are the natural logarithm of new hires (Columns 1 and 2), separations (Columns 3 and 4), and total payroll (Columns 5 and 6). Patent Law is an indicator for whether a state has passed anti-patent-troll legislation on or before that quarter. The list of states and corresponding signing dates is provided in Table A1. The control variables are defined in the text. Each specification includes state and year-quarter fixed effects. Robust standard errors are clustered by state. ***p<1%. ** p<5%, * p<10%.

Ln(Hires) (1)

Patent Law

Ln(Separations)

(2)

0.0357* 0.0358** (0.0198) (0.0179)

Ln(GSP)

Ln(Income per Capita)

Unemployment Rate

Ln(Patents)

(3)

(4)

0.0205

0.0202

(0.0155) (0.0149)

Ln(Payroll) (5)

(6)

0.0307** 0.0307*** (0.0133)

(0.0107)

0.741

1.310***

0.955***

(0.616)

(0.401)

(0.287)

0.704

-0.220

-0.009

(0.573)

(0.454)

(0.344)

0.0062

-0.0049

-0.0087

(0.0105)

(0.0088)

(0.0053)

0.0867

-0.0165

0.0389

(0.0801)

(0.106)

(0.0554)

State FE

yes

yes

yes

yes

yes

yes

Year-quarter FE

yes

yes

yes

yes

yes

yes

Observations

788

788

788

788

788

788

0.010

0.043

0.004

0.057

0.026

0.152

Within R-squared

– 32 –

Table 4: Small Business Employment by Education Level – High-Tech This table reports the effect of state anti-patent-troll laws on employment at small firms (<20 employees) in high-tech industries for workers of different education levels. The dependent variables are the natural logarithm of employment of college graduates (Columns 1 and 2) and employment of non-college graduates (Columns 3 and 4). College Emp. includes QWI groups “Some college or Associate degree” and “Bachelor’s degree or advanced degree;” Non-College Emp. includes QWI groups “Less than High school” and “High school or equivalent, no college”. Patent Law is an indicator for whether a state has passed anti-patent-troll legislation on or before that quarter. The list of states and corresponding signing dates is provided in Table A1. The control variables are defined in the text. Each specification includes state and year-quarter fixed effects. Robust standard errors are clustered by state. ***p<1%. ** p<5%, * p<10%.

Ln(College Emp.)

Patent Law

Ln(Non-College Emp.)

(1)

(2)

(3)

(4)

0.0194**

0.0201***

0.0183

0.0191*

(0.0091)

(0.0075)

(0.0125)

(0.0105)

Ln(GSP)

Ln(Income per Capita)

Unemployment Rate

Ln(Patents)

0.510**

0.791***

(0.192)

(0.203)

0.0997

0.208

(0.256)

(0.302)

-0.0074*

-0.0099*

(0.0041)

(0.0056)

0.0093

0.0324

(0.0364)

(0.0448)

State FE

yes

yes

yes

yes

Year-quarter FE

yes

yes

yes

yes

Observations

788

788

788

788

0.032

0.161

0.015

0.189

Within R-squared

– 33 –

Table 5: Bankruptcies This table reports the effect of state anti-patent-troll laws on the number of bankruptcies in a state. The dependent variables are the natural logarithm of the number of business bankruptcies (Columns 1 and 2) and non-business bankruptcies (Columns 3 and 4). Patent Law is an indicator for whether a state has passed anti-patent-troll legislation on or before that quarter. The list of states and corresponding signing dates is provided in Table A1. The control variables are defined in the text. Each specification includes state and year-quarter fixed effects. Robust standard errors are clustered by state. ***p<1%. ** p<5%, * p<10%.

Business bankruptcies

Patent Law

Non-business bankruptcies

(1)

(2)

(3)

(4)

-0.0383

-0.0486*

0.0254

0.0237

(0.0294)

(0.0267)

(0.0239)

(0.0236)

Ln(GSP)

Ln(Income per Capita)

Unemployment Rate

Ln(Patents)

-0.416

-0.324

(0.496)

(0.695)

-1.976**

-0.355

(0.769)

(0.447)

0.0350**

0.0221

(0.0165)

(0.0135)

-0.0945

-0.104

(0.165)

(0.108)

State FE

yes

yes

yes

yes

Year-Quarter FE

yes

yes

yes

yes

Observations

900

900

900

900

0.003

0.047

0.006

0.054

Within R-squared

– 34 –

Table 6: VC Financing This table reports the effect of state anti-patent-troll laws on venture capital (VC) investments in a state. The dependent variables are the natural logarithm of the number of VC-funded firms in a state and of the dollar amount of VC financing raised by these firms. We focus on early-stage VC rounds raised by firms in high-tech industries. Specifically, we measure VC-fundraising by high-tech firms at the “startup/seed,” “early stage,” or “expansion” phase that were founded in 2005 or later (our results are very similar if we include firms founded before 2005). In Columns 1-4, the sample includes all 50 states, while in Columns 5-8 we focus on states with a level of VC activity in 2012 at or above the median of all states. Patent Law is an indicator for whether a state has passed antipatent-troll legislation on or before that quarter. The list of states and corresponding signing dates is provided in Table A1. The control variables are defined in the text. Each specification includes state and year-quarter fixed effects. Robust standard errors are clustered by state. ***p<1%. ** p<5%, * p<10%.

All States Ln(# Firms)

Patent Law

High VC States

Ln($ Amount)

Ln(# Firms)

(1)

(2)

(3)

(4)

0.0678

0.0562

0.0757

0.0775

0.187** 0.144** 0.289* 0.229*

(0.0590) (0.0579) (0.116) (0.112)

(0.0795) (0.0698) (0.148) (0.124)

Ln(GSP)

Ln(Income per Capita)

Unemployment Rate

Ln(Patents)

(5)

(6)

Ln($ Amount) (7)

(8)

0.808

0.401

-0.107

0.595

(0.849)

(1.763)

(2.174)

(4.262)

-0.291

1.525

-3.219

-3.995

(1.483)

(2.498)

(3.044)

(4.520)

-0.00247

-0.0700

-0.00641

-0.0397

(0.0316)

(0.0870)

(0.0523)

(0.136)

0.503*

0.684

1.716**

3.101**

(0.263)

(0.673)

(0.690)

(1.447)

State FE

yes

yes

yes

yes

yes

yes

yes

yes

Year-Quarter FE

yes

yes

yes

yes

yes

yes

yes

yes

Observations

900

900

900

900

486

486

486

486

0.002

0.010

0.001

0.008

0.019

0.042

0.011

0.030

Within R-squared

– 35 –

Table 7: Small Business Employment by VC Activity – High-Tech This table reports the effect of state anti-patent-troll laws on employment at small firms (<20 employees) in high-tech industries for states with high and low VC activity. In Columns 1-2, we focus on states with a level of VC activity in 2012 at or above the median of all 50 states, while in Columns 3-4 we focus on states with below-median VC activity. Patent Law is an indicator for whether a state has passed anti-patent-troll legislation on or before that quarter. The list of states and corresponding signing dates is provided in Table A1. The control variables are defined in the text. Each specification includes state and year-quarter fixed effects. Robust standard errors are clustered by state. ***p<1%. ** p<5%, * p<10%.

High VC States

Patent Law

Low VC States

(1)

(2)

(3)

(4)

0.0218**

0.0240***

0.0114

0.0104

(0.0103)

(0.0089)

(0.0164)

(0.0137)

Ln(GSP)

0.597**

0.643**

(0.218)

(0.252)

0.333

-0.156

(0.272)

(0.311)

0.0038

-0.0086

(0.0036)

(0.0064)

0.0347

-0.0145

(0.0415)

(0.0436)

Ln(Income per Capita)

Unemployment Rate

Ln(Patents)

State FE

yes

yes

yes

yes

Year-Quarter FE

yes

yes

yes

yes

Observations

428

428

360

360

0.058

0.221

0.007

0.129

Within R-squared

– 36 –

Table 8: Number of Establishments by VC Activity – High-Tech This table reports the effect of state anti-patent-troll laws on the number of high-tech business establishments for states with high and low VC activity. In Columns 1-2, we focus on states with a level of VC activity in 2012 at or above the median of all 50 states, while in Columns 3-4 we focus on states with below-median VC activity. Patent Law is an indicator for whether a state has passed antipatent-troll legislation on or before that quarter. The list of states and corresponding signing dates is provided in Table A1. The control variables are defined in the text. Each specification includes state and year-quarter fixed effects. Robust standard errors are clustered by state. ***p<1%. ** p<5%, * p<10%.

High VC States

Patent Law

Low VC States

(1)

(2)

(3)

(4)

0.0235**

0.0287***

-0.0099

-0.0119

(0.0097)

(0.0096)

(0.0121)

(0.0085)

Ln(GSP)

Ln(Income per Capita)

Unemployment Rate

Ln(Patents)

0.403*

0.688***

(0.216)

(0.143)

0.750**

-0.482***

(0.358)

(0.125)

0.0040

-0.0112**

(0.0062)

(0.0052)

-0.0904*

-0.0037

(0.0478)

(0.0257)

State FE

yes

yes

yes

yes

Year-Quarter FE

yes

yes

yes

yes

Observations

486

486

414

414

0.048

0.155

0.011

0.318

Within R-squared

– 37 –

Patent Trolls and Small Business Employment Ian Appel, Joan Farre-Mensa, and Elena Simintzi

INTERNET APPENDIX

–1–

Figure A1: Sample Demand Letter MPHJ Technology Investments, LLC, through its over 100 subsidiaries, sent patent demand letters to approximately 16,465 small businesses located in all fifty states and the District of Columbia between September 2012 and May 2013. Over the course of the campaign, MPHJ used different versions of the letter that share a core text. One such letter, redacted to remove the name and address of the recipient, is shown below. (Source: https://www.ftc.gov/system/files/documents/cases/150317mphjtechexhibitsa-c.pdf.)

–2–

–3–

–4–

–5–

–6–

Table A1: Patent Troll Laws

State

Date Signed

Alabama

3/18/2014

Arizona

3/24/2016

Colorado

6/5/2015

Florida

6/2/2015

Georgia

4/15/2014

Idaho

3/26/2014

Illinois

8/26/2014

Indiana

5/5/2015

Kansas

5/20/2015

Louisiana

5/28/2014

Maine

4/14/2014

Maryland

5/5/2014

Mississippi

3/28/2015

Missouri

7/8/2014

Montana

4/2/2015

New Hampshire

7/11/2014

North Carolina

8/6/2014

North Dakota

3/26/2015

Oklahoma

5/16/2014

Oregon

3/3/2014

Rhode Island

6/4/2016

South Carolina

6/9/2016

South Dakota

3/26/2014

Tennessee

5/1/2014

Texas

6/17/2015

Utah

4/1/2014

Vermont

5/22/2013

Virginia

5/23/2014

Washington

4/25/2015

Wisconsin

4/24/2014

Wyoming

3/11/2016

–7–

Table A2: Predictive Regressions This table reports the effect of state economic conditions on the timing of adoption of state antipatent-troll laws. Observations for states that adopt a patent law are excluded from the sample after the law is signed. The explanatory variables are lagged by one quarter in Columns 1-4 and by four quarters in Column 5. Each specification includes state and year-quarter fixed effects. Robust standard errors are clustered by state. ***p<1%. ** p<5%, * p<10%.

Patent Law Signed

Ln(GSP )t-1

(1)

(2)

(3)

(4)

0.563

0.897

0.844

0.738

(0.554)

(0.579)

(0.632)

(0.636)

-1.532

-1.601

-1.795

(1.236)

(1.204)

(1.205)

-0.00726

-0.00521

(0.0182)

(0.0177)

Ln(Income per Capita)t-1 Unemployment Ratet-1 Ln(Patents)t-1

(5)

0.272 (0.194)

Ln(GSP)t-4

0.275 (0.764)

Ln(Income per Capita)t-4

0.805 (1.351)

Unemployment Ratet-4

0.0150 (0.0258)

Ln(Patents)t-4

0.0654 (0.244)

State FE

yes

yes

yes

yes

yes

Year-Quarter FE

yes

yes

yes

yes

yes

Observations

662

662

662

662

512

0.200

0.204

0.204

0.208

0.235

R-squared

–8–

Table A3: Falsification Test: Placebo Event This table repeats the analysis in Table 2, but it falsely assumes that each state that passed an anti-patent-troll law did so three years before the actual law was passed in the state. The dependent variable in Columns 1-3 is employment in high-tech industries, and the dependent variable in Columns 4-6 is employment in non-high-tech industries. Each specification includes state and year-quarter fixed effects. Robust standard errors are clustered by state. ***p<1%. ** p<5%, * p<10%.

Tech Firms Firm Size=

Non-Tech Firms

0-19

20-50

51-500+

0-19

20-50

51-500+

(1)

(2)

(3)

(4)

(5)

(6)

0.00718

-0.00796

0.00604

0.000296

0.000752

0.00325

(0.0111)

(0.0168)

(0.00929)

(0.00529)

(0.00614)

(0.00479)

State FE

yes

yes

yes

yes

yes

yes

Year-Quarter FE

yes

yes

yes

yes

yes

yes

Observations

796

796

796

796

796

796

0.003

0.001

0.002

0.000

0.000

0.002

Patent Law (-12 quarters)

Within R-squared

–9–

Table A4: Robustness Test: Excluding California This table excludes California from our baseline employment and VC analyses. Specifically, Panel A repeats the same analysis as Table 2 excluding California from the sample, and Panel B repeats the same analysis as Table 6 excluding also California. Patent Law is an indicator for whether a state has passed anti-patent-troll legislation on or before that quarter. The list of states and corresponding signing dates is provided in Table A1. The control variables are defined in the text. Each specification includes state and year-quarter fixed effects. Robust standard errors are clustered by state. ***p<1%. ** p<5%, * p<10%.

Panel A Tech Firms Firm Size=

Patent Law

Non-Tech Firms

0-19

0-19

20-50

51-500+

0-19

20-50

51-500+

(1)

(2)

(3)

(4)

(5)

(6)

(7)

-0.0028

0.0010

-0.0010

-0.0016

0.0023

0.0205** 0.0199** (0.0099)

Ln(GSP)

(0.0082) (0.0206) (0.0106)

(0.0052) (0.0059) (0.0045)

0.650*** (0.203)

Ln(Income per Capita)

0.134 (0.248)

Unemployment Rate

-0.0071 (0.0044)

Ln(Patents)

0.0164 (0.0382)

State FE

yes

yes

yes

yes

yes

yes

yes

Year-Quarter FE

yes

yes

yes

yes

yes

yes

yes

Observations

772

772

772

772

772

772

772

0.028

0.175

0.000

0.000

0.000

0.000

0.001

Within R-squared

– 10 –

Panel B All states Ln(# Firms) (1)

Patent Law

0.0688

(2)

High VC States

Ln($ Amount) (3)

(4)

Ln(# Firms) (5)

(6)

Ln($ Amount) (7)

(8)

0.0539 0.0910 0.0856

0.191** 0.139* 0.317** 0.237*

(0.0594) (0.0582) (0.117) (0.114)

(0.0803) (0.0711) (0.150) (0.126)

Ln(GSP)

Ln(Income per Capita)

Unemployment Rate

Ln(Patents)

0.852

0.257

0.0145

0.397

(0.873)

(1.799)

(2.348)

(4.563)

-0.256

1.487

-3.369

-4.185

(1.491)

(2.502)

(3.089)

(4.549)

-0.00362

-0.0657

-0.0122

-0.0395

(0.0327)

(0.0892)

(0.0566)

(0.142)

0.513*

0.651

1.789**

3.089**

(0.269)

(0.672)

(0.717)

(1.458)

State FE

yes

yes

yes

yes

yes

yes

yes

yes

Year-Quarter FE

yes

yes

yes

yes

yes

yes

yes

yes

Observations

900

900

900

900

486

486

486

486

0.002

0.010

0.001

0.007

0.019

0.043

0.013

0.030

Within R-squared

– 11 –

Table A5: Falsification Test: Small Business Employment – Non-High-Tech This table repeats the same analysis as Table 4, but focuses on small firms (<20 employees) that are not in high-tech industries and so for which we do not expect the anti-patent-troll laws to have any effect. The dependent variables are the natural logarithm of employment of college graduates (Columns 1 and 2) and employment of non-college graduates (Columns 3 and 4). Patent Law is an indicator for whether a state has passed anti-patent-troll legislation on or before that quarter. The list of states and corresponding signing dates is provided in Table A1. The control variables are defined in the text. Each specification includes state and year-quarter fixed effects. Robust standard errors are clustered by state. ***p<1%. ** p<5%, * p<10%.

Ln(College Employment)

Patent Law

Ln(Non-College Employment)

(1)

(2)

(3)

(4)

-0.0009

-0.0007

-0.0056

-0.0050

(0.0045)

(0.0045)

(0.0058)

(0.0058)

Ln(GSP)

Ln(Income per Capita)

Unemployment Rate

Ln(Patents)

0.172***

0.126

(0.0640)

(0.0754)

0.312**

0.449***

(0.127)

(0.162)

0.0037

0.0039

(0.00399)

(0.0052)

0.0165

0.0151

(0.0198)

(0.0244)

State FE

yes

yes

yes

yes

Year-quarter FE

yes

yes

yes

yes

Observations

788

788

788

788

0.000

0.053

0.003

0.047

Within R-squared

– 12 –

Table A6: Falsification Test: Employment by VC Activity – Non-High-Tech This table repeats the same analysis as Table 7, but focuses on small firms (<20 employees) that are not in high-tech industries and so for which we do not expect the anti-patent-troll laws to have any effect. In Columns 1-2, we focus on states with a level of VC activity in 2012 at or above the median of all 50 states, while in Columns 3-4 we focus on states with below-median VC activity. Patent Law is an indicator for whether a state has passed anti-patent-troll legislation on or before that quarter. The list of states and corresponding signing dates is provided in Table A1. The control variables are defined in the text. Each specification includes state and year-quarter fixed effects. Robust standard errors are clustered by state. ***p<1%. ** p<5%, * p<10%.

High VC States

Patent Law

Low VC States

(1)

(2)

(3)

(4)

-0.0020

-0.0012

-0.0040

-0.0043

(0.0055)

(0.0040)

(0.0088)

(0.0095)

Ln(GSP)

Ln(Income per Capita)

Unemployment Rate

Ln(Patents)

0.295

0.0878*

(0.187)

(0.0488)

0.538***

0.189

(0.174)

(0.209)

0.0107*

0.0060

(0.0046)

(0.0081)

0.0487

-0.0128

(0.0412)

(0.0281)

State FE

yes

yes

yes

yes

Year-Quarter FE

yes

yes

yes

yes

Observations

428

428

360

360

0.001

0.142

0.001

0.017

Within R-squared

– 13 –