What Is a Patent Worth? Evidence from the U.S. Patent “Lottery”

• Published date: 01 April 2020
• Author: ALEXANDER LJUNGQVIST,DEEPAK HEGDE,JOAN FARRE‐MENSA
• DOI: http://doi.org/10.1111/jofi.12867
• Date: 01 April 2020

THE JOURNAL OF FINANCE •VOL. LXXV, NO. 2 •APRIL 2020

What Is a Patent Worth? Evidence from the U.S.

Patent “Lottery”

JOAN FARRE-MENSA, DEEPAK HEGDE, and ALEXANDER LJUNGQVIST∗

ABSTRACT

We provide evidence on the value of patents to startups by leveraging the quasi-

random assignment of applications to examiners with different propensities to grant

patents. Using unique data on all first-time applications filed at the U.S. Patent

Office since 2001, we find that startups that win the patent “lottery” by drawing

lenient examiners have, on average, 55% higher employment growth and 80% higher

sales growth five years later. Patent winners also pursue more, and higher quality,

follow-on innovation. Winning a first patent boosts a startup’s subsequent growth

and innovation by facilitating access to funding from venture capitalists, banks, and

public investors.

PATENTS AWARD TEMPORARY MONOPOLY RIGHTS over inventions to their inven-

tors. In theory, patents can benefit their holders by deterring copycats and by

serving as defensive shields in litigation suits, as bargaining chips in licens-

ing negotiations, and as signaling devices to attract investors and customers

(Budish, Roin, and Williams (2016)). Policymakers and scholars argue that

these benefits offset the social cost of patent rights (Nordhaus (1969), Scherer

(1972)). startups may derive particularly large private benefits from patents,

∗Joan Farre-Mensa is with the University of Illinois at Chicago. Deepak Hegde is with New

YorkUniversity. Alexander Ljungqvist is with the Stockholm School of Economics and the Swedish

House of Finance. A previous version of this paper was circulated under the title “The Bright Side of

Patents.” We are grateful to Emily Blanchard, Lauren Cohen, WesleyCohen, Michael Ewens, Matt

Fiedler, Lee Fleming, Teresa Fort, Ken French, Alberto Galasso, Bronwyn Hall, Zorina Khan, Josh

Lerner, Ross Levine, Alan Marco, Ramana Nanda, Bhaven Sampat, Robert Seamans, Amit Seru

(the Editor), Ted Sichelman, Scott Stern, Rick Townsend, Heidi Williams, Rosemarie Ziedonis, an

anonymous Associate Editor, two anonymous reviewers, and audiences at the NBER Productivity

Lunch Seminar, the NBER Summer Institute, the SFS Cavalcade, the WFAmeetings, and various

universities for helpful comments and to Scott Bornstein, George Elliott, Jamie Kucab, Adam

Mossoff, Barry Schindler, and Ted Sichelman for patiently explaining the nuances of the patent

examination process. We also thank the Institute for Exceptional Growth Companies for granting

access to the NETS database. Hegde gratefully acknowledges the support of the United States

Patent and Trademark Office’s Thomas Alva Edison VisitingScholars program and the Kauffman

Junior Faculty Fellowship. The views and comments expressed herein are solely the opinion of

the authors, do not reflect the performance of duties in the authors’ official capacities, and are

not endorsed by, nor should be construed as, any viewpoint official or unofficial of the United

States Patent and Trademark Office. The authors confirm to the best of their knowledge that no

information contained herein is privileged, confidential, or classified. The authors have read The

Journal of Finance’s disclosure policy and have no conflict of interest to disclose.

DOI: 10.1111/jofi.12867

C2019 the American Finance Association

639

640 The Journal of Finance R

given that these firms typically have few assets other than their ideas early

in their lives and an absence of alternative mechanisms to protect their in-

ventions (Lamoreaux and Sokoloff (1999), Federal Trade Commission (2011)).

At the same time, mounting evidence suggests that patents are abused and

costly to enforce, and that they impose a tax on follow-on inventors (Heller and

Eisenberg (1998), Jaffe and Lerner (2004), Lemley and Shapiro (2005), Becker

and Posner (2013), Galasso and Schankerman (2015)). How valuable, then, are

patents to startups?

The value of a patent is the incremental economic benefit accruing to its

holder from the legal right to exclude others from exploiting the invention,

beyond what would be earned if the invention were not granted a patent.

Separating this incremental benefit from the economic value of the underlying

invention is challenging. Previous studies document that firm value increases

in the number of patents a firm holds or at the time a firm is awarded a patent,

but these correlations conflate the value of the underlying inventions and the

incremental benefits of the exclusion rights conferred by the patents.1In this

study, we aim to provide an estimate of these incremental economic benefits by

focusing on how patents affect the growth and success of startups. To do so, we

employ novel data and an identification strategy that disentangles the causal

effects of patent rights from the economic value of the underlying inventions.

To isolate the incremental benefit of patents over and above the value of

the underlying inventions, we assemble novel data on patent rejections. Until

recently, researchers have lacked access to data on rejections and thus have

instead compared firms with and without patents. Of course, firms could lack

patents for many reasons: they may have no inventions to patent in the first

place, they may have had their patent applications denied, or they may have

chosen to protect their inventions in other ways (such as trade secrets). As a

result, comparing firms with and without patents cannot identify the value

of a patent right, as distinct from the value of the underlying invention. We

overcome this challenge thanks to special access to the internal databases of

the U.S. Patent and Trademark Office (USPTO), which cover all applications,

whether granted or rejected.

We combine the applications data with data on the performance of star-

tups assembled from a variety of sources. We focus on five types of outcome

variables: (i) growth in sales and employment (from Dun & Bradstreet’s Na-

tional Establishment Time Series or NETS database), (ii) follow-on patenting

and patent citations (from the USPTO’s patent database), (iii) the pledging of

patent applications as collateral to raise debt (from the USPTO’s patent assign-

ment database), (iv) venture funding (from VentureXpert), and (v) fundraising

by startups through initial public offerings (IPOs) (from VentureXpert and

Thomson-Reuters’s SDC database).

To disentangle the value of a patent from that of the underlying invention,

we exploit exogenous variation in the USPTO’s review process through an

1For example, Hall, Jaffe, and Trajtenberg(2005), Balasubramanianand Sivadasan (2011), and

Kogan et al. (2017) report a positive correlation between patents and measures of firm value.

What Is a Patent Worth? 641

instrumental variables approach pioneered by Sampat and Williams (2019).

The validity of the approach rests on two features of the process. First, the

USPTO assigns applications in each field (or “art unit”) to examiners ran-

domly with respect to the characteristics of the underlying invention (Lemley

and Sampat (2012), Sampat and Williams (2019)). Second, examiners vary in

their propensity to approve applications—some are more lenient, while others

are stricter (Cockburn, Kortum, and Stern (2002)). The assignment of appli-

cations to examiners who quasi-randomly differ in their leniency imparts a

lottery-like element to the process: among applicants with patentable inven-

tions, some win patent rights, while others do not simply because the former

were lucky enough to draw more lenient examiners.2This setting allows us

to isolate the economic benefits of patent rights from those of the underlying

invention.3

Our sample covers all 34,215 first-time patent applications filed by U.S.

startups at the USPTO since 2001 that received a final decision by December

31, 2013. Our estimates suggest that winning the patent lottery, by randomly

drawing an examiner more likely to approve applications, increases the average

startup’s growth in employment and sales over the next five years by 54.5 and

79.5 percentage points, respectively, compared to startups with only randomly

different underlying inventions that fail to obtain a patent. For the average

sample firm, these estimates imply that receiving a patent leads to a cumu-

lative 16 additional employees and $10.6 million in additional sales over the

five years after winning the patent lottery. Receiving a patent also increases

both the number of subsequent patents the firm is granted (by 56.5%) and

their quality (the average number of citations per subsequent patent increases

by 33%).

In contrast to the large benefits associated with receiving a first patent,

subsequent patents are much less obviously beneficial. We find little evidence

that approval of a second patent application boosts employment or sales growth

significantly (even for startups whose first application was rejected), although

it does increase the number of subsequent patents (by 49.8%). Approval of

a third patent shows similar patterns. These results suggest that the value

of patents to startups changes over time. Securing a first patent, when they

typically have few other assets, allows startups to innovate and embark on a

high-growth trajectory. Subsequent patents encourage further innovation but

play less of a direct role in boosting growth, which instead presumably depends

on other resources the startups have acquired along the way.

What explains the large economic benefits that startups derive from ob-

taining a first patent, over and above the economic benefits arising from the

2Gaule (2018) uses a similar instrument to study the effect of patents on the likelihood of

venture capital (VC)-backed startups going public or being acquired. In a similar spirit, Galasso

and Schankerman (2015) exploit variation in the leniency of judges to study the effect of patent

invalidation on follow-on innovation.

3Applications of marginal quality are likely to be the most affected by examiner leniency.In the

language of the treatment effects literature, these marginal applications comprise our instrument’s

compliant subpopulation.