Do patent applications affect the exit of venture-backed investments?

University of Amsterdam, Amsterdam Business School

Master in International Finance

Do Patent Applications Affect the Exit of Venture-Backed

Investments?

September 2017

Abstract

Backed by evidence that patents influence the investment decision of venture capitalists, I investigate if patent applications affect the exits of venture-backed investments. I employ three complementary experiments. I control the regressions for industry and time effects, and for investment effects such as total funding and total investment rounds received by the invested company. I find in the first hypothesis that applying for one or more patents after receiving the first round of venture-backed investment increases the probability successful exit route. I don’t find enough evidence for my second hypothesis. Patent applications do not show significant influence on the probability of an IPO or an M&A. In the last hypothesis, I observe that the number of patent applications does not show significant impact on the valuation of IPOs’ and M&As’ deals. I collect data from two different sources, and I match them by the name of the invested start-up. As a proxy for patent grants, I utilize the number of patent applications. This study provides new results on the effect patent applications have on venture-backed firms. I don’t find strong evidence of the impact of patent applications on the exits of venture-backed investments. I encourage further studies to use data of patent grants and compare the outcomes.

Supervisor: Prof. Tolga Caskurlu

Student: Julia Edelhoff de Almeida Cardoso Student Number: 11352930

Table of Contents

1 – Introduction ... 3

2 - Literature Review ... 7

3 - Data Analysis ... 12

3.1) Patent Data and Investment Data ... 12

3.2) Descriptive Data ... 14

3.2.1) Hypothesis 1 and Hypothesis 2 ... 14

3.2.2) Hypothesis 3 ... 14

3.3) Potential biases ... 15

4 - Hypotheses ... 17

4.1) Hypothesis 1: Patent applications influence the overall success of a venture capital investment ... 17

4.2) Hypothesis 2: Patent applications influence the exit route of a venture capital investment ... 17

4.3) Hypothesis 3: Patent applications influence the valuation of the investment exit ... 18

5- Regressions and Findings ... 19

5.1) Hypothesis 1: Patent applications influence the overall success of a venture capital investment ... 19

5.2) Hypothesis 2: Patent applications influence the exit route of a venture capital investment ... 20

5.3) Hypothesis 3: Patent applications influence the valuation of the investment exit ... 22

6- Conclusions ... 24

7 - Bibliography ... 26

8- Tables ... 29

1 – Introduction

“Why do I need you?”

“Why can’t I just go and make this myself?”

“Is there anything proprietary about your product?”

In practically every episode of Shark Tank1_{, spectators hear a Shark asking at least} one of the questions above, inquiring about the entrepreneur’s patent protection.2 _The scenario of the reality show represents the venture capital environment faithfully: Investors seek target firms that have innovative ideas with the potential to disrupt and conquer new markets. Granting a patent for such ideas is the best, and possibly only, way to guarantee they remain proprietary at least during the following years. Venture capitalists represent investment firms willing to take the risk of investing in young companies holding promising ideas, aiming at future high rates of return. By doing so, these investment firms stimulate innovation and fosters economic growth (Gompers & Lerner, 2001). In this regard, the correlation between patent practices and the venture capital investments is a topic that offers important inputs not only to venture capitalists and entrepreneurs but also to the whole economic society.

In this research, I investigate the effect of patent applications (used as a proxy for innovation) on the success of venture capital investments. I use the exit route as a proxy for investment success, and I consider IPOs and M&As as successful exits. Empirical evidence shows that patents are relevant to the investment decision of a venture capital (Mann and Sager, (2007), Graham et al. (2009), Kortum and Lerner (2000). In this study, I want to explore if patent practices also influence the exit of the investment. For venture-backed

1 _{Shark Thank is a business themed reality series, broadcasted in different countries. It features the}

sharks - tough, self-made, multi-millionaire and billionaire – who give budding entrepreneurs the chance to make their dreams come true and potentially secure business deals. - http://abc.go.com/shows/shark-tank/about-the-show

companies, the exit route is related to the exit decision of the venture capitalist (Schwienbacher, 2009). If the VC wants to exit, the start-up has better chances to survive if it goes public or is acquired by another company. Other exit routes, repurchases, secondary sales and especially write-offs, are not the best forecast for the company’s health.

This study combines three complementary analyses of the relationship between patent practices and venture capital investments. First, I investigate a broader hypothesis that patent applications influence the success of VC investments. Second, I focus on the successful cases of the first hypothesis and inspect if patent applications affect the probability of the successful exit being an IPO or an M&A. Lastly, I question if the number of patent applications impacts on the valuation of the exits.

Broadly, the primary hypotheses of this study are i) Patent applications influence the overall success of a venture capital investment, ii) Patent applications influence the exit route of a venture capital investment, and iii) Patent applications influence the valuation of the investment exit. I find that a higher number of patent applications increases the probability of a successful exit. I also observe that patent applications don’t affect the probability or the valuation of IPOs and M&As.

Literature shows that IPOs and M&As are the preferred exits of venture capitalists. Therefore, I choose these exit methods to represent successful exits in this study. Cumming and MacIntonsh’s (2003) have these exit methods on the top of their venture capital exit ranking, Smith and Wall (1998) name IPOs the “Holy Grail” of exits, and Schwienbacher (2002) states that IPOs bring significant reputational benefits.

I combine data from different sources to build the dataset for this research: bulk data of patent applications from USPTO (United States Patent and Trademark Office) and venture capital data from Thomson One. I match both datasets using the name of the target company – assignee company for USPTO and target company for Thomson One. To do so, I standardize the entries of these names in each dataset. The input for USPTO’s assignee name comes from the Patent Application Form (Figure 1) that is filled in manually by the inventor, allowing different writings for the same inventor name. Such variations must be standardized to guarantee that all patent applications of every company are taken into consideration. I create the variable CompanyName that is filled in with company names in

the standardized format. The variable TotalPatents is also calculated after matching both datasets. For each CompanyName in the Thomson One data set, I count how many identical entries are there in the USPTO data. The relevant patent application must have been filled between the first investment received by the invested company and the investment exit. If the filling is done before the first investment round or after the exit, the application is not taken into consideration.

I use a logit regression to investigate the first hypothesis. The dependent variable is called Success and assumes the value of 1 if the exit route of the investment is an IPO or an M&A, and 0 otherwise. For the second hypothesis, I use another logit regression. The dependent variable is called ExitRoute and assumes the value of 1 if the exit route is an IPO and 0 if it is an M&A. I separate the third hypothesis in two cases according to the exit route of the investment, for both I use an OLS regression. In the case of the exit route being an IPO, the dependent variable is the IPO valuation, which I calculate as (Offer Price * Number of Shares Offered). In case the exit route is an M&A, the dependent variable is the valuation of the deal, provided in the Thomson One data. I control these regressions for variate investment effects such as the year of the exit, the number of funds that invested in the target company, the target company’s industry, the investment size, and the time to exit.

For hypothesis 1, I conclude that patent applications are positively correlated with the success of venture-backed investments; I observe that patent applications increase the probability of the exit route being either an IPO or an M&A. Other possible cases are secondary sales and write-offs. For hypothesis 2 and 3, I don't find evidence that patent applications influence the probability or the valuation of IPOs and M&As. I attribute this finding to the fact that venture capitalists seek for high profitability in the first place and therefore are willing to take the deal that guarantees the highest return, no matter if an IPO or an M&A. After observing the results of Hypothesis 3, that patent applications show no significant influence on the valuation of IPOs and M&As, the findings of Hypothesis 2, that patents don’t influence the probability of IPOs and M&As, seem coherent.

This research offers relevant contributions to the literature as well as inputs for further studies. This study adds to the literature on patents and their influence on venture capital investment. First, previous literature documents that patent practices are highly

positively correlated with venture-backed investments (Graham et al. (2009), Kortum and Lerner (2000). Relating the success of the investment exits, manifested in the exit routes, with patents complements the literature by showing that the patent practices not only indicates more venture-backed investments but also better outcomes for these investments.

Secondly, this research adds to the literature that relates IPOs and M&As with patents. Following Deeds et al. (1997), I don’t find significant influence of patent applications on IPOs’ valuations. However, I also find no significant influence of patent applications on M&As’ valuations and this finding contrasts with the hypothesis of patent motivated acquisitions, as explored by Caskurlu (2015).

Thirdly, this research does not focus on one or few specific industries. While Mann and Sager (2007) focus on the software industry, Hall and Ziedonis (2001) observe the semiconductor industry and Lerner (2015) examines the biotechnological industry, this research presents results that encompass every industry.

Finally, I believe there are not many studies using patent applications as a proxy for patent grants. This approach can be deeply analyzed and generalized to other studies.

The remainder of this document is organized as follows: In chapter 2, I briefly summarize some concepts and previous findings on venture capital, patents and their relationship. Chapter 2 presents the data utilized for this research. In chapter 4 and 5, I present the hypotheses, regressions, and findings of this study. Finally, in chapter 6, the conclusions and implications of this research are exposed together with suggestions for possible further researches.

2 - Literature Review

Innovation is a phenomenon present in every industry. Utterback (1994) states that understanding and adapting to innovation is essential for companies to survive, once they are facing increasing competition and demand for high cost and globally appealing products.

Hall and Ziedonis (2001) show that patents are recognized as an important instrument to

promote innovation and technological progress and Kortum and Lerner (2000) find that

venture capital is associated with a substantial increase in patenting.

A patent grants the inventor of a property with the right to exclude others from

making, using, offering for sale, selling or importing the invention3 _{for at least 20 years (in}

the United States), therefore it acts as protection for innovative creations. Farre-Mensa, Hedge and Ljungqvist (2017) mention deterioration of copycats, defense against litigation suits, and attracting investors and customers, as some other benefits provided by patents.

Issuing a patent is an investment, it requires dedication of time and money. A patent application is subject to the payment of a basic fee and additional ones, that include a search fee, an examination fee, and an issue fee. Some applications are also subject to

excess claims fees.4 _{In addition to monetary costs, the process takes a long time to be}

concluded. As of August 2017, a traditional patent pendency is 24.4 months.5 _{This number}

has shown a decreasing trend over the last two years, in October 2015 it was 27.4 months. Due to this fact, young and small firms commonly face the dilemma of patenting or not. On one side, it might be too costly for a young company that already faces high initial costs and low cash flows; on the other side, not claiming the rights over an invention might increase competition in the long term and even take the company to extinction.

The term start-up commonly refers to a new business venture or a new commercial project6_{.6Normally, it is a company that has recently begun operating and is part of an}

3_{https://www.uspto.gov/patents-getting-started/general-information-concerning-patents}

4 _{https://www.uspto.gov/patents-getting-started/patent-process-overview#step4}

https://www.uspto.gov/learning-and-resources/fees-and-payment/uspto-fee-schedule

5 _{“Pendency is the average number of months from the patent application filing date to the date the}

application has reached final disposition (e.g., issued as a patent or abandoned) which is called

“disposal”. - https://www.uspto.gov/corda/dashboards/patents/main.dashxml?CTNAVID=1004 6_{http://www.dictionary.com/browse/startup}

innovative technological industry. One common attribute of start-ups is their ability to grow; they have been designed to scale very quickly.7 _{Often the idea of a product or service is the} most valuable part of the start-up business. The more innovative and attractive it is, the higher the chances of the start-up growing fast and attracting the attention of customers and investors. Accordingly, having all the rights of executing this idea should be a high priority for these young companies.

Venture capital firms invest in firms in their early stages, taking high risks and aiming to obtain a high rate of return. This return is capitalized during the disinvestment process, which can take the form of IPOs, strategical acquisitions (M&As), repurchases, refinancing or write-offs. Schwienbacher (2009) describes these exit as follows:

(i) Initial Public Offering (IPO): the company achieves a stock market listing so that the venture capitalist can also sell her shares to the public;

(ii) Trade sale (or acquisition): the sale of the portfolio company to an existing firm, possibly in the same industrial sector;

(iii) Management buy-out (or repurchase): the venture capitalist sells her shares back to the entrepreneur;

(iv) Refinancing (or secondary sale): the venture capitalist's stocks are purchased by another institutional investor (e.g., another VC); and

(v) Liquidation (or write-off): the company files for bankruptcy.

As in Hypothesis 1, literature shows a consensus when comparing performance and reputational gains; there are exits are better than others, IPOs and M&As are considered the most desired exit routes. Lopez-de-Silanes, Phalippou, and Gottschalg (2009) find the median IRR for an IPO is 46% versus 26% for other fully exited investments. Schwienbacher (2002) mentions reputational benefits for the venture firm when investments are exited via IPOs. Cumming and Johan (2009) mention that choosing to sell to a trade buyer normally allows a full exit of the investment, that is transformed in cash for the investor. Smith and

Wall (1998) observe that strategical buyers might be willing to pay a premium for the company since they expected to create substantial synergies.

There’s also evidence showing that secondary buyouts are less preferred among venture capitalists since they are normally used as a replacement IPOs (Sousa, 2010). Cumming and Johan (2009) note that buy-backs exit route is less preferred by the investor since the returns obtained are the lowest excluding bankruptcy.

Briefly, as explained by Chaplinski (2015), a typical venture capital fund is organized as a limited partnership: general partners manage the funds while limited partners act as equity contributors. The relationship between the partners, the partnership structure, and the compensation arrangements are stated in a limited partnership agreement. Normally, a PE fund has a life of 10 to 12 years. During the life of a VC fund, there are two main stages to be observed: the investment and the harvest period. During the first stage, managers invest in the companies chosen previously through a careful due diligence process. Managers aim to forecast the probability of the start-up reaching maturity and being sold to the public or another company by a higher price than the amount invested. Normally, a portion of the fund's capital is reserved for the case of additional investment rounds are performed or if any incurred losses need to be covered. The investment period lasts around 5 years, during which mainly cash outflows happen in the fund. For the next 5 years, managers look for ways to exit their investments. Once start-ups rarely pay interest or dividends due to a lack cash flow and profitability, most of the venture capital’s return tend to arise in the form of capital gains, which occur when the investment is exited (Cumming and MacIntosh, 2003).

The investment decision of a VC firm, aiming future high rates of returns, is based not only on financial documents and numbers but also on the product or service provided by the target firm and the industry the company belongs to (Macmillan, 1985). This analysis is necessary because the young target companies rarely have significant information in their financial statements; in such early stages, there haven’t been many opportunities to create a meaningful performance record yet.

All in all, patents, acting as a protection and an incentive to innovative ideas, should have relevant influence on the investment decision and the success of the venture capital investments. This influence has been widely studied in the last decades. Hedge and

Ljungqvist (2017) state that winning the first patent facilitates access to funding from venture capital and other ways of funding, such as bank loans. Lerman (2015) finds similar results in the biotech industry. She shows that patent practicing leads to significant increase in the number of rounds of funding a company receives and that having patents before the first investment round leads to an increase in the amount of funding in subsequent rounds.

In the American semiconductor industry, Hall and Ziedonis (2001) find that executives consider patent rights to be important in attracting venture capital funding. Similarly, Baum and Silverman (2004) show a positive correlation between patent applications and early stage private equity financing. They also conclude that the importance of patents in this regard varies across industries. Also, Hsu and Ziedonis (2013) argue that patents increase the perception of the value of the firms that hold them. Furthermore, they observe that start-ups benefit more from patent practices in earlier stages because patents act as a proof of their potential when there’s high uncertainty regarding their future. Lastly, they find that venture-backed start-ups with more patent applications are more likely to attract initial funding from prominent venture capital funds. Regarding the software industry, Mann and Sager (2007) find a significant and positive correlation between patenting and variables that measure a firm’s performance such as the number of investment rounds, the total funding received, as well as the amount received on later stages of funding and the company’s longevity. They also observe that the effect of patents depends more on the firm having at least one patent than on the firm having a vast patent portfolio. Additionally, they conclude that patents influence positively the number of investment rounds – for each additional patent a company holds, it gets 0.144 additional rounds of investment, and having at least one patent infers to 0.71 additional investment rounds. At the same time, they calculate that firms that have one or more patents receive, on average, $2.7 Million more in total investment.

Examining the way in which venture capital investments influence patenting practices, Graham et al. (2009) point out that venture-backed companies tend to patent more than non-venture-backed ones. In addition to that, Lerner and Tirole (2002) state that small and capital constrained firms have a higher propensity to patent than larger firms. Kortum and Lerner (2000) conclude that increases in venture capital activity in an industry

are associated with significantly higher patenting rates. They calculate that the impact of one US-Dollar of venture capital influences patent practices 3.1 more than one US- Dollar of investment in R&D. Hellman and Puri (2000) indicate that young companies backed by venture capital differ from those financed by other traditional alternatives in which they follow more innovative strategies and take a shorter time to introduce their products to the market, especially when the company is an innovator and not an imitator.8

For Hypothesis 2, literature presents different aspects for what influences exit methods. Cumming and MacIntosh (2003) argue that cyclicality significantly influences IPO valuations, especially due to psychological factors that play an important role next to investment fundamentals and Schwienbacher (2009) states that investment exits tend to vary between years. Schwienbacher (2008b) also shows that IPOs are limited to the most promising investments while M&As happen amidst both, more and less promising ones. He also states that the probability of a venture-backed company going public relates to factors such as the number of financing rounds, the investment duration and the reporting requirements of the target company to the venture capital manager.

For Hypothesis 3, when observing patents and exits, Deeds et al. (1997) attest that patents may only be noisy signals when observing their impact on funds raised by an IPO.

8 _{“Innovators are firms that are first to introduce new products or services for each no close}

substitute is yet offered in the market. Imitators are also engaged in relatively new products and technologies, but they are not the first movers in the markets, and therefore tend to compete on aspects other than innovation”.

3 - Data Analysis

3.1) Patent Data and Investment Data

I gather data from two different sources for this research: USPTO (United States Patent and Trademark Office) and Thomson One. The first provides information about patent applications in the United States and the second presents information about venture capital investments.

USPTO

From the USPTO website, I obtained detailed information about patent applications recorded between January 1970 and January 2017. The data available contains for each application: a reference id, the assignee’s name, the description of the assignment, the assignor’s name, the patent document, the date when the claim was recorded and the date of the last update. For this study, the most important variables are assignee’s name, which corresponds to the name of the company, and the record date.

Instead of gathering data only for patent grants, I used data containing records of all patent applications. USPTO does not provide a bulk data for patent grants. These must be individually hand-collected from their database (PatFT or AppFT9_{). Another option is to} download weekly files that documents the patents granted the prior week. Since the inception of digital patent record release in 1975, the USPTO has used no less than 8 different formats for representing granted patents alone. Hand-collecting data for more than 7 Million patents using any of this methods is barely possible. Therefore, I choose to use data on patent application. I believe the high costs associated with patent applications – not only with the application itself but also with the development of the idea-, which are very relevant for a restrained budget like the one of a start-up, will hinder the cases of these

9 _{http://patft.uspto.gov/netahtml/PTO/search-bool.html and}

companies filing frivolous applications that will end up not being granted.

The application form is filled in by the applicant (Figure 1 shows an example of a filing form). Consequently, there's no pattern to be seen in the fields. The same company may have its entries written in different ways. For the same assignee, the name might appear with different writings on the database. Therefore, I run a code to standardize the variable assignee name in USPTO data. I delete characters such as “.”, “,” and “!” and words like “LTD”, and “LIMITED”. They are common to company names and have many different abbreviations. One example of the standardization is the firm Zymogenetics. It appears as “Zymogenetics Inc”, “Zymogenetics Inc.”, “Zymogenetics Incorporated” and “Zymogenetics, Inc”. The code transforms all variations in “Zymogenetics”. Even after this procedure, the same company can appear with names written differently, since there are some peculiarities that might not have been included in the code.

To calculate the total number of patent applications of a company, I first count how many times the same CompanyName (already in the standard format) appears in the USPTO bulk data. After that, I compare the record date of each patent application with the date of the first round of investment received by the company and the date of the exit. I consider only the applications filed between these dates.

THOMSON ONE

From Thomson One, I extract two different datasets. First, one that contains information of all venture capital investments realized in the United States, during the same period as the patent applications. I selected a set of variables as follows: name of the company, SIC Code, the number of investment rounds, the company’s age at each investment round, the number of funds that invested in the company and the current status of the company. Second, I extract data only for venture-backed investment exits. It contains information about every deal that was recorded in the United States during the period analyzed. I select the variables date of the deal, how many years between the first investment round and the exit (time to exit), value of the deal for M&As, offer price and

number of shares outstanding for IPOs. I standardize the company names using the same procedure as before.

Table I.1 presents the description of the variables and table I.2 present a statistical summary of them.

3.2) Descriptive Data

3.2.1) Hypothesis 1 and Hypothesis 2

In the sample used for Hypothesis 1 and 2, there are 95,231 companies. The total number of patent applications of a company varies from 0 to 5,240, with an average of 0.165 per company. The total amount of funding received by the target firms ranges from USD 319,200 to USD 17.5 Million. The average amount of venture funding received is USD 20,200 per company. The time between the first investment received by the start-up and the exit (time to exit) varies from 0.1 to 30.5 years. On average, it takes 5.2 years for the company to either go public or be acquired. The average number of venture capital funds that invested in the same start-up is 3.2, and it ranges from 1 to 40. Meanwhile, a company gets, on average, 2.5 rounds of investment. The number of investment rounds received by the companies in the sample varies from 1 to 32. When they receive the first investment, companies are, on average 27.7 years old. (See Table I.2.1)

About 10% of the companies in the venture capital sample go public while 20% are acquired by or merged with another company. During the analyzed period, 10% are extinct, while 60% are still backed by venture capital investments (Figure 1 presents these numbers).

3.2.2) Hypothesis 3

For IPOs, the sample contains 5,241 companies. The average valuation is USD 111 Million, the valuation of the deals varies from USD 8,436 to USD 25 Billion. The total amount of funding received by the target firms that went public ranges from USD 10,000 to USD 8.9 Billion. The average amount of venture funding received until the exit is USD 74.65 Million per company. The time between the first investment received by the start-up and the IPO

(time to exit) varies from 0.1 to 37.5 years. On average, it takes 4.7 years for the company to either go public. The average number of venture capital funds that invested in the same start-up is 6.14, and it ranges from 1 to 40. Meanwhile, a company gets, on average, 0.97 rounds of investment. The number of investment rounds received by the companies in the sample varies from 1 to 20. When they receive the first investment, companies are, on average 24.3 years old. The youngest to receive the first investment is 0 years old, while the oldest is 99 years old. (See Table I.2.2)

For M&As, the sample contains 8,307 companies. The deal value ranges from USD 0.011 to USD 5,473. The average deal value is USD 234.4 Million. The total amount of funding received by the firms that were acquired ranges from USD 0.01 Million to USD 7 Billion. The average amount of venture funding received per company, until the M&A is USD 34.58. The time between the first investment received by the start-up and the M&A (time to exit) varies from 0.1 to 48.7 years. On average, it takes 5.4 years for the company to be acquired. The average number of venture capital funds that invested in the same start-up is 5.24, and it ranges from 1 to 31. Meanwhile, a company gets, on average, 0.36 rounds of investment. The number of investment rounds received by the companies in the sample varies from 1 to 16. When they receive the first investment, companies are, on average 9.45 years old. The youngest to receive the first investment is 0 years old, while the oldest is 99 years old. (See Table I.2.3)

3.3) Potential biases

3.3.1) Calculation of number of patent applications of a given company

Even after standardizing the company names in the datasets used for this research, there might be peculiarities that were not included in the code, resulting in one or more versions of a company name not being standardized. There's also the chance that companies changed names during the period analyzed. These facts may incur in an underestimation of patent applications.

3.3.2) Private Equity Data

Private equity firms do not have to publicly disclose information about their deals. The data that is available may suffer from potential self-reporting biases. One example is in Thomson One data used for this research, information on company’s financials such as (Equity Value / Net Income) and (Enterprise Value / EBITDA) is available for less than 10% of the M&A sample.

3.3.3) Patent application as proxy for number of patents

I utilize patent applications as a proxy for patent grants. These numbers are not equal for most the companies. Moreover, the number calculated in this study for patent applications might be distant from the real number of patents owned by a firm, implicating in incorrect results.

4 - Hypotheses

4.1) Hypothesis 1: Patent applications influence the overall success of a venture

capital investment

I base this hypothesis on three facts: i) venture capital can spur innovation (Kortum, Lerner (2000), ii) patents are incentive to innovation (Hall and Ziedonis (2001), and iii) IPOs and M&As are the preferred exit methods of venture capitalists (Cumming and MacIntonsh (2003).

Facing evidence that venture capital investment decisions and patent practices are highly correlated, I want to observe if such a relationship is also present between venture capital exits and patent practices. Patents facilitate the access to funding from venture capital (Hedge and Ljungqvist (2017). Do they also have any impact on the exit of these investments?

In this hypothesis, I investigate the influence of patent applications on the success of the investment, where I consider IPOs and M&As successful exits. I hope to find a positive correlation between the success of the investment and the number of patents applications of the invested company.

4.2) Hypothesis 2: Patent applications influence the exit route of a venture

capital investment

One way to observe this hypothesis is to say that if a company goes public, it becomes more independent, whereas if it is acquired, it proves to have become more attractive to other industry players. Cumming and MacIntosh (2003) classify a strategic acquirer as a player that belongs to the same business as that of the target firm. Therefore, the question here can be translated to “After applying for patents, do venture-backed companies become more independent or become a better target for other firms?”.

IPOs are the preferred exit route. One argument is that the average annual rate of return for a venture capital in the US is higher when the exit occurs through an IPO than for a trade sale (Gompers 1995). From the target company’s perspective, Roëll (1996) provides a list of advantages for companies to go public. The main ones are access to new finance, enhanced company image and publicity, and the possibility for cashing in. And Cockburn and Wagner (2007) find that firms that apply for more patents are less likely to be acquired, though obtaining unusually highly cited patents may make them more attractive acquisition target.

In this hypotheses, I investigate if patent applications influence the probability IPOs or M&As being the exit method.

4.3) Hypothesis 3: Patent applications influence the valuation of the investment

exit

Patents are considered a protection for innovative ideas and innovation is what keeps the economy strong (Caskurlu, 2015). In this hypothesis, I investigate if patent applications reflect their benefits in the valuation of IPOs and M&As. Hsu and Ziedonis (2013) show that a doubling in a start-up’s patent portfolio increases the company’s valuation by 24%, on average. Anecdotal evidence shows that many acquisitions are patent motivated. For example, in 2012, Google acquired Motorola for USD 12.4 billion. The acquirer paid USD 5.5 billion for patents, USD 2.9 billion for cash and USD 4 billion for the business and other assets. After a couple of months, Motorola Home Business was sold for USD 2.4 billion. Analyses of the transaction point out that Google was only interested in Motorola’s Android patents (Caskurlu, 2015). I expect to find a positive correlation for both.

5- Regressions and Findings

5.1) Hypothesis 1: Patent applications influence the overall success of a venture

capital investment

To test this hypothesis, I consider companies backed by venture capital investments, which received the first investment round between January 1970 and January 2017 and are no longer backed by venture capital. I limit my analysis to investments that have already been exited. The exit routes can be IPOs, M&As or write-offs. Secondary sales are not included in the sample. The dependent variable is the variable Success, a dummy that assumes the value of 0 if the outcome of the investment is negative and 1 otherwise. The outcome is positive if the exit route is an IPO or an M&A and negative in case of a write-off. For this analysis, I use a logit regression as in Equation (1).

𝑆𝑢𝑐𝑐𝑒𝑠𝑠_𝑖 = 𝛽1 ∗ log (𝑃𝑎𝑡𝑒𝑛𝑡𝑠)_𝑖+ 𝛽2 ∗ 𝑙𝑜𝑔𝑇𝑜𝑡𝑎𝑙𝐹𝑢𝑛𝑑𝑖𝑛𝑔_𝑖 + 𝛽3 ∗ 𝐴𝑔𝑒𝐴𝑡𝐹𝑖𝑟𝑠𝑡𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑛𝑔_𝑖 + 𝛽4 ∗ 𝑁𝑢𝑚𝑏𝑒𝑟𝑂𝑓𝑅𝑜𝑢𝑛𝑑𝑠_𝑖 + 𝛽5 ∗

𝑁𝑢𝑚𝑏𝑒𝑟𝑂𝑓𝐹𝑢𝑛𝑑𝑠𝑖+ 𝛽6 ∗ 𝐹𝑖𝑟𝑠𝑡𝐼𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡𝑌𝑒𝑎𝑟 + 𝛽7 ∗ 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦𝑖 + 𝜀𝑖

In this model, the term 𝑙𝑜𝑔𝑃𝑎𝑡𝑒𝑛𝑡𝑠𝑖 represents the logarithm of (1 + the number of patent applications the target company has filled during the investment period). The variable 𝑙𝑜𝑔𝑇𝑜𝑡𝑎𝑙𝐹𝑢𝑛𝑑𝑖𝑛𝑔𝑖 equals the logarithm of the sum of all venture capital investments the target firm has received until the exit. 𝑁𝑢𝑚𝑏𝑒𝑟𝑂𝑓𝐹𝑢𝑛𝑑𝑠𝑖 represents the number of funds that invested in the target company throughout all investment rounds it received. 𝐴𝑔𝑒𝐴𝑡𝐹𝑖𝑟𝑠𝑡𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑛𝑔𝑖 represents, in year, how old the company is when it receives the first investment round. The number of investment rounds is represented by 𝑁𝑢𝑚𝑏𝑒𝑟𝑂𝑓𝑅𝑜𝑢𝑛𝑑𝑠_𝑖. 𝐹𝑖𝑟𝑠𝑡𝐼𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡𝑌𝑒𝑎𝑟 is a dummy variable that equals 1 if the year is the same as the year of the first investment round, and 0 otherwise. 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦𝑖is also a

dummy variable that assumes the value of 1 if the industry is the same as the one the target company belongs to and 0 otherwise (Figure 4 presents information on industries).

Table II shows the regression results of Equation (1). The coefficient of interest 𝑙𝑜𝑔𝑃𝑎𝑡𝑒𝑛𝑡𝑠_𝑖 is statistically significant at 1% level. This result implies that for additional patent applications made during the investment period, there’s an increase in the probability of the investment going through a successful exit. A logit is a non-linear regression; therefore, the magnitude of the increase is not reflected in the coefficients.

This result is according to the expectation that patents do influence the success of investments positively.

5.2) Hypothesis 2: Patent applications influence the exit route of a venture

capital investment

To test this hypothesis, I consider exit routes that are either IPOs or an M&As and that the invested company filed at least one patent application during the investment period. I drop from the data set companies that have the variable CompanyStatus equal to "Defunct" or "Bankruptcy", these represent cases of write-offs. The dependent variable 𝐸𝑥𝑖𝑡𝑅𝑜𝑢𝑡𝑒𝑖 is a dummy that assumes the value of 1 if the exit route is an IPO, and 0 if it is an M&A. For this analysis, I employ a Logit regression as in Equation (2).

𝐸𝑥𝑖𝑡𝑅𝑜𝑢𝑡𝑒𝑖 = 𝛽1 ∗ 𝑙𝑜𝑔𝑃𝑎𝑡𝑒𝑛𝑡𝑠𝑖 + 𝛽2 ∗ 𝑙𝑜𝑔𝑇𝑜𝑡𝑎𝑙𝐹𝑢𝑛𝑑𝑖𝑛𝑔𝑖+ 𝛽3 ∗ 𝑇𝑖𝑚𝑒𝑇𝑜𝐸𝑥𝑖𝑡𝑖 + 𝛽4 ∗ 𝐴𝑔𝑒𝐴𝑡𝐹𝑖𝑟𝑠𝑡𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑛𝑔_𝑖 + 𝛽5 ∗ 𝑁𝑢𝑚𝑏𝑒𝑟𝑂𝑓𝐹𝑢𝑛𝑑𝑠_𝑖 + 𝛽6 ∗ 𝑁𝑢𝑚𝑏𝑒𝑟𝑂𝑓𝐷𝑒𝑎𝑙𝑠_𝑖+

𝛽7 ∗ 𝐸𝑥𝑖𝑡𝑌𝑒𝑎𝑟 + 𝛽8 ∗ 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦𝑖 + 𝜀𝑖

In this model, the term 𝑙𝑜𝑔𝑃𝑎𝑡𝑒𝑛𝑡𝑠𝑖 represents the logarithm of the number of patent applications the target company has performed during the investment period. The variable 𝑙𝑜𝑔𝑇𝑜𝑡𝑎𝑙𝐹𝑢𝑛𝑑𝑖𝑛𝑔𝑖 equals the logarithm of the sum of all venture capital investments the target firm has received until the exit. The term 𝑇𝑖𝑚𝑒𝑇𝑜𝐸𝑥𝑖𝑡𝑖 is the

difference between the exit date and the date when the target company received its first investment. 𝐴𝑔𝑒𝐴𝑡𝐹𝑖𝑟𝑠𝑡𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑛𝑔𝑖 represents, in year, how old the company is when it receives the first investment round. 𝑁𝑢𝑚𝑏𝑒𝑟𝑂𝑓𝐹𝑢𝑛𝑑𝑠𝑖 represents the number of funds that invested in the target company throughout all investment rounds it received, which is represented by 𝑁𝑢𝑚𝑏𝑒𝑟𝑂𝑓𝑅𝑜𝑢𝑛𝑑𝑠𝑖 . ExitYear is a dummy variable that equals 1 if the year is the same as when the exit occurred and 0 otherwise. 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦𝑖is also a dummy variable that assumes the value of 1 if the industry is the same as the one the target company belongs to and 0 otherwise (Figure 4 presents information on industries).

Table III shows the results of Equation (2). The coefficient of interest 𝑙𝑜𝑔𝑃𝑎𝑡𝑒𝑛𝑡𝑠𝑖 is not statistically significant is all test but one, when the variable 𝐴𝑔𝑒𝐴𝑡𝐹𝑖𝑟𝑠𝑡𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑛𝑔𝑖 is excluded from the regression. The sign of 𝑙𝑜𝑔𝑃𝑎𝑡𝑒𝑛𝑡𝑠_𝑖 is positive. This result implies that, once the investment is successful, patent applications point to a higher probability of IPOs, however, this influence is not significant. This result shows that although patents have significant influence in the overall success of venture capital investments (hypothesis 1), they don’t have significant impact on the probability of the exit route being an IPO or an M&A.

I attribute this result to the fact that the choice of the exit route is a decision primarily made by the investors, seeking the highest possible return, which depends on many other different factors. Cyclicality affects the liquidity of the exit. Hence it also affects the probability of IPOs and M&As.

The coefficients 𝑙𝑜𝑔𝑇𝑜𝑡𝑎𝑙𝐹𝑢𝑛𝑑𝑖𝑛𝑔𝑖 𝑎𝑛𝑑 𝑇𝑖𝑚𝑒𝑇𝑜𝐸𝑥𝑖𝑡𝑖 are statistically significant at the 1% level. The first takes a positive value. This result means that a higher amount of venture capital funding is associated with a higher probability of an IPO when compared to the probability of an M&A. 𝑇𝑖𝑚𝑒𝑇𝑜𝐸𝑥𝑖𝑡𝑖 takes a negative value, meaning that the longer it takes between the first investment round and the exit, the higher is the probability of the exit being an M&A instead of an IPO.

5.3) Hypothesis 3: Patent applications influence the valuation of the investment

exit

To test this hypothesis, I separate my analysis in two, exit routes that are IPOs and exit routes that are M&As. For IPOs, I calculate the dependent variable as log(Price * Number of Shares Offered). The valuation of M&A deals is provided by Thomson One. To test this hypothesis, I use the OLS regression as in Equation (3).

𝐸𝑥𝑖𝑡𝑉𝑎𝑙𝑢𝑎𝑡𝑖𝑜𝑛𝑖 = 𝛼 + 𝛽1 ∗ 𝑙𝑜𝑔𝑃𝑎𝑡𝑒𝑛𝑡𝑠𝑖 + 𝛽2 ∗ 𝑙𝑜𝑔𝑇𝑜𝑡𝑎𝑙𝐹𝑢𝑛𝑑𝑖𝑛𝑔𝑖+ 𝛽3 ∗ 𝑇𝑖𝑚𝑒𝑇𝑜𝐸𝑥𝑖𝑡𝑖 + 𝛽4 ∗ 𝐴𝑔𝑒𝐴𝑡𝐹𝑖𝑟𝑠𝑡 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑛𝑔𝑖+ 𝛽5 ∗ 𝑁𝑢𝑚𝑏𝑒𝑟𝑂𝑓𝐹𝑢𝑛𝑑𝑠𝑖 + 𝛽6 ∗ 𝑁𝑢𝑚𝑏𝑒𝑟𝑂𝑓𝑅𝑜𝑢𝑛𝑑𝑠_𝑖+ 𝛽7 ∗ 𝐸𝑥𝑖𝑡𝑌𝑒𝑎𝑟 + 𝛽8 ∗ 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦_𝑖 + 𝜀𝑖

In this model, the term 𝑙𝑜𝑔𝑃𝑎𝑡𝑒𝑛𝑡𝑠_𝑖 represents the logarithm of (1+ the number of patent applications the target company has performed during the investment period). The variable 𝑙𝑜𝑔𝑇𝑜𝑡𝑎𝑙𝐹𝑢𝑛𝑑𝑖𝑛𝑔_𝑖 equals the logarithm of the sum of all venture capital investments the target firm has received until the exit. The term 𝑇𝑖𝑚𝑒𝑇𝑜𝐸𝑥𝑖𝑡𝑖 is the difference between the exit date and the date when the target company received its first investment. 𝐴𝑔𝑒𝐴𝑡𝐹𝑖𝑟𝑠𝑡𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑛𝑔_𝑖 represents, in year, how old the company is when it receives the first investment round. 𝑁𝑢𝑚𝑏𝑒𝑟𝑂𝑓𝐹𝑢𝑛𝑑𝑠𝑖 represents the number of funds that invested in the target company throughout all investment rounds it received, which is represented by 𝑁𝑢𝑚𝑏𝑒𝑟𝑂𝑓𝑅𝑜𝑢𝑛𝑑𝑠𝑖 . ExitYear is a dummy variable that equals 1 if the year is the same as when the exit occurred and 0 otherwise. 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦𝑖is also a dummy variable that assumes the value of 1 if the industry is the same as the one the target company belongs to and 0 otherwise (Figure 4 presents information on industries).

Table IV shows the results of Equation (3) for IPOs and Table V for M&As. For IPOs, the coefficient of interest 𝑙𝑜𝑔𝑃𝑎𝑡𝑒𝑛𝑡𝑠𝑖 is not statistically significant. This result shows that patent applications do not present a significant impact on the valuation of such exit. The sign of the coefficient interest is positive for all regressions but one. This shows that, (3)

although not significant, the observed effect of patent applications on IPO valuations is positive. Deeds et al. (1997) find similar results in the biotech industry. They do not find any effect of patents on the funding raised in an IPO and infer that patents may only be noisy signals that do not affect the expectations investors.

For M&As, the coefficient of interest 𝑙𝑜𝑔𝑃𝑎𝑡𝑒𝑛𝑡𝑠𝑖 is also not statistically significant. This result shows that patent applications do not present a significant impact on the valuation of acquisitions of venture-backed companies. The sign of the coefficient interest is negative for all regressions. This result is not expected and shows that patent applications have a negative correlation with the valuation of M&A deals.

For both cases, the coefficients of the variables 𝒍𝒐𝒈𝑻𝒐𝒕𝒂𝒍𝑭𝒖𝒏𝒅𝒊𝒏𝒈𝒊 and 𝑵𝒖𝒎𝒃𝒆𝒓𝑶𝒇𝑭𝒖𝒏𝒅𝒔_𝒊 are statistically significant at 1% level. The term 𝒍𝒐𝒈𝑻𝒐𝒕𝒂𝒍𝑭𝒖𝒏𝒅𝒊𝒏𝒈_𝒊 assumes a positive sign. This result shows that the amount of funding raised by a start-up is positively correlated the valuations of their exits. On the other hand, the 𝑵𝒖𝒎𝒃𝒆𝒓𝑶𝒇𝑭𝒖𝒏𝒅𝒔_𝒊 assumes a negative sign. This result shows that, on average, the more funds investing in a start-up leads to a lower valuation of their exits.

6- Conclusions

In this study, I investigate the role of patents in the disinvestment process of venture capital investments using a combination of two different sources of data. First, using a logit regression and controlling for size and year of the investment, industry, age at first investment round, number of investment rounds and number of funds investing in the company, I find that patent applications are positively correlated with successful exits, once they show significant impact on the probability of IPOs or M&As being the exit route. Second, using another logit regression and controlling for size of the investment, year of the exit, industry, time to exit, age at first investment round, number of rounds and number of funds investing in the company, I observe that patent applications don’t influence which exit, an IPO or an M&A, has higher probability. Lastly, I also don’t find statistical significance when investigating the correlation between patent applications and the valuation of the exit.

This study contributes to the existing literature by providing a quantitative link between patent applications and venture capital investment success. First, it adds to the literature on venture capital and patents. Other studies mainly focus on how patents affect the company's attractiveness to venture backed funding. Kortum and Lerner (2002) find a positive correlation between patents and the amount of venture-backed investments. Hedge and Ljungqvist (2017) state that winning the first patent facilitates access to funding from venture capital. Mann and Sager (2007) find a significant and positive correlation between patenting and variables that measure a firm’s performance such as the number of investment rounds, the total capital received as investment, as well as the amount received on later stages of funding and the company’s longevity. In this study, I contribute to this literature showing that patents positively influence the success of venture capital investments exits.

Secondly, the sample covers all industries (represented by the 2 first digits of the SICCode, in this case), whereas most studies on patents are limited to one or a few industries. Mann and Sager (2007) study the software industry; Lerman (2015) finds results

for the biotech industry, and Hall and Ziedonis (2001) investigate the American semiconductor industry.

The last findings of this research are quite striking. First, I find patent applications do not significantly impact the probability and the valuation of IPOs and M&As. Second, I observe that, although not significant, the correlation between patent applications and the valuation of M&A deal is negative. Different studies point to cyclicality and liquidity as important factors that influence the choice of the exit route (MacIntosh (2003), Schwienbacher (2009). These factors might have influenced the outcome of my tests. I encourage further studies to consider different business cycles and compare the results. The term 𝒍𝒐𝒈𝑻𝒐𝒕𝒂𝒍𝑭𝒖𝒏𝒅𝒊𝒏𝒈𝒊 appears to impact significantly the results of both studies. I suggest further studies to address this relationship from another perspective, for example utilizing the IRR of IPOs and M&As instead of valuations of the deals.

Other research may seek to address other limitations of this study. Instead of utilizing patent applications, patents grants could be used. A time constraint would be helpful in the process of data collection. Industry constraint may also influence the outcomes.

7 - Bibliography

Baum, J., Silverman, B. (2004). ‘Picking winners or building them? Alliance, intellectual, and human capital as selection criteria in venture financing and performance of biotechnology startups’. Journal of Business Venturing. Vol. 19. 411-466.

Carley, M., Hedge, D., and Marco, A. (2013). ‘What is the Probability of Receiving a US Patent?’. Working Paper No. 2013-2.

Caskurlu, T. (2015). ‘Effects of Patent Rights on Industry Structure and R&D.’. Available at SSRN: https://ssrn.com/abstract=2427191 or http://dx.doi.org/10.2139/ssrn.2427191

Chaplinsky, S. (2015). ‘The Basics of Private Equity Funds’, Darden Business Publishing UV 6986

Cockburn, I., Wagner, S. (2007). ‘Patents and the Survival of Internet-Related IPOs’, Working paper 13146

Conti, A., Thursby, M.C., and Thursby, J. (2013). ‘Patents as Signals for Startup Financing’. The Journal of Industrial Economics, 61(3), 592–622.

Cumming, D., Johan, S. (2009). ‘Venture Capital and Private Equity Contracting: An International Perspective’.

Cumming, D., MacIntosh, J. (2003). ‘Venture-capital exits in Canada and the United States’. University of Toronto Journal 53, 101-200.

Cumming, D., Fleming, G. and Schwienbacher, A., (2005).’Liquidity Risk and Venture Capital Finance’. Financial Management, (77-105).

Deeds, D., DeCarolis, D. and Coombs, J. (1997). ‘The Impact of Firm Specific Capabilities on the Amount of Capital Raise in an Initial Public Offering: Evidence From the Biotechnology Industry’, Journal of Business Venturing 12, 165–187.

Farre-Mensa, J., Hegde, D. and Ljungqvist, A. (2016). ‘The Bright Side of Patents’. CEPR Discussion Paper No. DP11091.

Graham S.J.H., Merges, S.P., Samuelson, P. and Sichelman, T. (2009). ‘High technology entrepreneurs and the patent system: results of the 2008 Berkeley patent survey’. Berkeley

Technology Law Journal, 24(4), Available at:

Giot, P., Schwienbacher, A. (2007). ‘IPOs, trade sales and liquidations: Modelling Venture Capital Exits Using Survival Analysis’. Journal of Banking & Finance, 31(3), 679-702.

Gompers, P., (1995). ‘Optimal Investment, Monitoring, and the Staging of Venture Capital’. The Journal of Finance, 50(5), 1461-1489.

Gompers, P. (1996). ‘Grandstanding in the venture capital industry’. Journal of Financial Economics, 43, 133-156.

Gompers, P., Lerner, J., (2001). ‘The Venture Capital Revolution’., The Journal of Economic Perspectives, 15(2), 145-168.

Hall, B. H., Ziedonis, R.H. (2001). ‚The Patent Paradox Revisited: an empirical study of patenting in the U.S. semiconductor industry’. Journal of Economics, 32(1), 101-128.

Hellmann, T., Puri, M. (2000). ‘The Interaction between Product Market and Financing Strategy: The Role of Venture Capital’. The Review of Financial Studies, 13(4), 959-984.

Hottenrott, H., Bronwyn, H. and Czarnitzki D. (2015). ‘Patents as Quality Signals? The Implications for Financing Constraints on R&D’. DICE Discussion Papers 133. Düsseldorf Institute for Competition Economics (DICE), University of Düsseldorf.

Hsu, D., Ziedonis, R. (2013). ‘Resources as Dual Sources of Advantage: Implications for Valuing Entrepreneurial-Firm Patents’, Strategic Management Journal 34, 761–781.

Kaplan, N., Strömberg, P. (2000). ‘Financial Contracting Theory Meets The Real World: An Empirical Analysis Of Venture Capital Contracts’. CRSP Working Paper No. 513.

Kortum, S., Lerner, J. (2000). ‘Assessing the Contribution of Venture Capital to Innovation’. Journal of Economics, 31(4), 674-692.

Lerman, C. (2015). ‘Patent Strategies of Technology Startups: An Empirical Study’. Working Paper, Universidad Austral.

Lerner, J. (1994). ‘Venture capitalists and the decision to go public’. Journal of Financial Economics, 35, 293-316.

Lerner, J., Tirole J. (2002). ‘Efficient Patent Pools’. Working Paper 9175 http://www.nber.org/papers/w9175.

Long, C. (2002). ‘Patent Signals’. University of Chicago Law Review, 69, 625–679.

Lopez de Silanes, F., Phalippou, L., Gottschalg, O., 2009, Giants at the Gate: On the Cross-Section of Private Equity Investment Returns. AFA 2010 Atlanta Meetings Paper; EFA 2009 Bergen Meetings Paper. Available at SSRN: http://ssrn.com/abstract=1363883

Macmillan I., Siegel R., Subba Narasimha P.N. (1985). ‘Criteria used by venture capitalists to evaluate new venture proposals ‘. Journal of Business Venturing, 1(1), 119-128.

Mann, R. J., Sager, T. W. (2007). ‘Patents, venture capital and software startups’. ScienceDirect, 36, 193-208.

Mansfield, E. (1986). ‘Patents and Innovation: An Empirical Study’. Management Science, 32(2), 173-181.

Pakes, A. (1985). ‘On Patents, R&D, and the Stock Market Rate of Return’. Journal of Political Economy. 93(21), 390-408.

Smith, J., Wall, J. (1998). ‘Better Exits. European Venture Capital Association’., Technical Report.

Sousa, M. (2010). ‘Why Do Private Equity Firms Sell to Each Other?’. Working paper.

Stromberg, P. (2008). ‘The new demography of private equity’. SIFR, Working paper.

Schwienbacher, A. (2008), ‘Innovation and Venture Capital Exits’. The Economic Journal 118(533), 1888-191.6

Utterback, J. (1994). ‘Mastering the Dynamics of Innovation: How Companies Can Seize Opportunities in the Face of Technological Change’. University of Illinois at Urbana-Champaign's Academy for Entrepreneurial Leadership Historical Research Reference in Entrepreneurship. Available at SSRN: https://ssrn.com/abstract=1496719

8- Tables

I.1. Variables Description

This table presents the description of the variables relevant for this study. Their names, description, and units of measure are presented below.

VARIABLE DESCRIPTION UNIT

Success This variable represents the outcome of the investment. If the outcome is an IPO or an M&A, it assumes the value 1, otherwise 0.

Dummy

TotalPatents This variable represents the number of patent applications of a company. It is calculated based on the number of entries the company has in the USPTO data during the investment period.

Unit

TotalFunding This variable represents the total amount of venture capital funding the target company has received since the first round of investment.

USD Mio FirstInvestmentYear NumberofRounds AgeAtFirstFinancing ExitYear TimeToExit

This variable represents the year when the start-up received the first venture backed investment. This variable represents the number rounds of venture capital investment a company received. This variable represents the age of the start-up when she receives the first round of venture capital investment.

This variable represents the year when the exit took place.

This variable represents the amount of time between the First Investment Year and Exit Year.

Date

Units

Years

Date

M&A Deal Value This variable represents the value of the M&A deal. USD Mio IPO Valuation NumberofFunds CompanyStatus

This variable represents the Valuation of the IPO. It is calculated as (Offer Price * Number of Shares Offered)

This variable represents the number of venture capital funds that invested in the same start-up.

This variable represents the status of the target company on the last day of the research period. In this case, 31/12/2016. USD Mio Unit “Defunct” “Bankruptcy” “Went Public” “Acquired” “Merged”

I.2 Summary Statistics

I.2.1 - This table presents the statistical values of the sample used to test Hypothesis 1 and Hypothesis 2

Hypothesis 1 and Hypothesis 2

VARIABLES N Mean Min Max StDev

TotalPatents 9,559 0.165 0 5,240 20.014 TotalFunding 9,519 20.189 0.3192 17,457.93 138.321 TimeToExit 9,416 5.157 0.1 30.5 3.454 NumberofFunds 9,283 3.221 1 40 3.181 NumberofRounds 9,353 2.512 1 32 2.423 AgeAtFirstFinancing 9,353 27.651 0 99 24.355

I.2.2 - This table presents the statistical values of the sample used to test Hypothesis 3 for IPOs

Hypothesis 3 – IPOs

VARIABLES N Mean Min Max StDev

TotalFunding 4,808 74.64 .01 8,900 286.06

TimeToExit 5,100 4.679 .1 37.6 3.752

Offer Price 5,127 20.78 .011 5,473.65 175.82

Number of Shares Offered 5,128 67*10 1,000 1.72*10¹¹ 2.41*10

IPO Valuation 5,127 1.1*10 8,436 2.5*10 4.95*10

NumberofFunds 5,221 6.14 1 40 5.21

NumberofRounds 2,635 0.97 1 20 2.11

AgeAtFirstFinancing 2,635 24.31 0 99 25.12

I.2.3 - This table presents the statistical values of the sample used to test Hypothesis 3 for M&As

Hypothesis 3 – M&As

VARIABLES N Mean Min Max StDev

Total Funding 7,858 34.58 .01 7,000 118.83

TimeToExit 8,153 5.392 .1 48.7 3.82

M&A Deal Value 3,532 234.45 .05 29,404.37 1,010.11

NumberofFunds 8,292 5.24 1 31 4.09

NumberofRounds 4,794 0. 36 1 16 1.35

AgeAtFirstFinancing 4,687 9.45 0 99 18.55

6 ₉

Table II

Patent applications influence the overall success of a venture capital investment

This table presents the results illustrating the relationship between the success of a venture capital investment and the number of patent applications the target company has during the investment period. The exact calculation is defined by Equation (1). The dependent variable represents the success of an investment. The investment exit route as a proxy for success. The dependent variable is the dummy Success that assumes the value 1 if the exit route is an IPO or an M&A and 0 otherwise. 𝐥𝐨𝐠𝐏𝐚𝐭𝐞𝐧𝐭𝐬𝐢 is defined as the logarithm of (1+ the sum of all patent applications of company between the first investment round and the exit). Definition of the variables are provided in Table I. Robust standard errors are reported in parenthesis; *, ** and *** denote the significance of 10%, 5%, and 1%, respectively.

Dependent Variable: Success

VARIABLES (1) (2) (3) (4) (5) logPatents 0.621*** (0.174) 0.457*** (0.150) 0.492*** (0.156) 0.451*** (0.137) 0.387*** (0.141) logTotalFunding 0.713* 0.618** 0.741** 0.545** (0.375) (0.256) (0.374) (0.249) AgeAtFirstFinancing 0.0126 -0.000809 0.0110 0.0107 (0.0197) (0.0115) (0.0192) (0.0163) NumberOfRounds 0.226 0.0247 0.147 0.0854 (0.267) (0.159) (0.217) (0.186) NumberOfFunds -0.118 -0.00550 -0.106 -0.0431 (0.102) (0.0663) (0.0816) (0.0800) First Investment Year

Dummies Industry Dummies YES YES YES YES YES YES YES YES YES YES Observations 7,156 6,179 6,160 8,929 5,888 Pseudo R-Squared 0.135 0.191 0.061 0.069 0.191

Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table III

Patent applications influence the exit route of a venture capital investment

This table presents the results illustrating the relationship between the exit route of a venture capital investment and the number of patent applications of the target company during the investment period. The exact calculation is defined by Equation (2). The dependent variable refers to the exit route of the investment. It is a dummy called ExitRoute that assumes the value 1 if the exit route is an IPO and 0 if it is an M&A. 𝐥𝐨𝐠𝐏𝐚𝐭𝐞𝐧𝐭𝐬𝐢 is defined as the logarithm of (1 +the sum of all patent applications of company between the first investment round and the exit). Definition of the variables are provided in Table I. Robust standard errors are reported in parenthesis; *, ** and *** denote the significance of 10%, 5%, and 1%, respectively.

Dependent Variable: Exit Route

VARIABLES (1) (2) (3) (4) (5) (6) logPatents 0.317 0.486 0.171 0.748*** 0.272 0.288 (0.424) (0.356) (0.366) (0.289) (0.424) (0.424) logTotalFunding 1.053*** 1.005*** 0.836*** 1.233*** 1.232*** (0.116) (0.116) (0.0473) (0.104) (0.104) TimeToExit -0.524*** -0.334*** -0.338*** -0.430*** -0.442*** (0.133) (0.110) (0.0543) (0.132) (0.125) AgeAtFirstFinancing 0.00167 -0.00189 0.00217 0.00123 -0.431*** (0.00380) (0.00342) (0.00368) (0.00390) (0.131) NumberOfFunds 0.0703*** 0.229*** 0.0445* 0.00535 0.00117 (0.0248) (0.0233) (0.0232) (0.0102) (0.00390) NumberOfRounds 0.0159 0.0113 0.0178 -0.0496*** 0.0168 0.0704*** (0.0390) (0.0386) (0.0387) (0.0136) (0.0383) (0.0248)

Industry Dummy YES YES YES YES YES YES

Year Dummy YES YES YES YES YES YES

Observations 4,348 4,348 4,370 4,370 4,351 4,351 Pseudo R-Squared 0.367 0.345 0.356 0.322 0.367 0.287

Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table IV

Patent applications influence the valuation of the investment exit

This table presents the results illustrating the relationship between the number of patent applications during the investment period and the valuation of an IPO. The exact calculation is defined by Equation (3). The dependent variable is calculated as log (Offer Price * Number of Shares Offered) and refers to the valuation of the IPO. 𝐥𝐨𝐠𝐏𝐚𝐭𝐞𝐧𝐭𝐬𝐢 is defined as the logarithm of the sum of all patent applications of company between the first investment round and the exit. Definition of the variables are provided in Table I. Robust standard errors are reported in parenthesis; *, ** and *** denote the significance of 10%, 5%, and 1%, respectively.

Dependent Variable: IPO Valuation

VARIABLES (1) (2) (3) (4) (5) logPatents 0.0430 -0.0245 0.0408 0.0376 0.0672 (0.0755) (0.0953) (0.0753) (0.0738) (0.0721) logTotalFunding 0.349*** 0.347*** 0.299*** 0.339*** (0.0264) (0.0264) (0.0222) (0.0191) TimetoExit 0.00795 -0.00102 -0.00287 0.00450 (0.00962) (0.0114) (0.00942) (0.00684) AgeAtFirstFinancing -0.000382 -0.00278** -0.000519 -5.68e-05 -0.000265 (0.00122) (0.00132) (0.00122) (0.00123) (0.000865) NumberOfFunds -0.0318*** 0.0320*** -0.0300*** -0.0296*** (0.00628) (0.00614) (0.00584) (0.00444) NumberOfRounds 0.00123 0.00354 0.00147 0.00107 (0.0132) (0.0151) (0.0133) (0.0134)

Industry Dummy YES YES YES YES YES

Year Dummy YES YES YES YES YES

Observations 1,184 1,300 1,189 1,184 1,271

R-squared 0.459 0.273 0.460 0.447 0.432

Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table V

Patent applications influence the valuation of the investment exit

This table presents the results illustrating the relationship between the number of patent applications during the investment period and the valuation of an IPO. The exact calculation is defined by Equation (3). The dependent variable is referred to the valuation of the M&A deal. 𝐥𝐨𝐠𝐏𝐚𝐭𝐞𝐧𝐭𝐬𝐢 is defined as the logarithm of (1+ the sum of all patent applications of company between the first investment round and the exit). Definition of the variables are provided in Table I. Robust standard errors are reported in parenthesis; *, ** and *** denote the significance of 10%, 5%, and 1%, respectively.

Dependent Variable: M&A Deal Value

VARIABLES (1) (2) (3) (4) (5) (6) logPatents -0.606 -0.471 -0.574 -0.739 -0.638 -0.584 (0.691) (0.551) (0.668) (0.703) (0.686) (0.689) logTotalFunding 0.587*** 0.475*** 0.585*** 0.587*** 0.584*** (0.0605) (0.0522) (0.0604) (0.0606) (0.0604) NumberOfFunds -0.0787*** 0.0311 -0.0800*** -0.0799*** -0.0828*** (0.0215) (0.0192) (0.0213) (0.0212) (0.0213) NumberOfRounds -0.0468 -0.0390 -0.0525 -0.0454 -0.0458 (0.0337) (0.0322) (0.0326) (0.0335) (0.0340) AgeAtFirstFinancing 0.00205 0.00272 0.00286 0.00183 0.00201 (0.00287) (0.00307) (0.00284) (0.00287) (0.00287) TimetoExit -0.0130 0.00790 -0.0327* -0.0119 -0.0128 (0.0191) (0.0207) (0.0195) (0.0191) (0.0191)

Industry Dummy YES YES YES YES YES YES

Year Dummy YES YES YES YES YES YES

Observations 2,422 2,498 2,436 2,432 2,426 2,438 R-squared 0.169 0.248 0.308 0.324 0.326 0.326

Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Appendix

Figure 1

Utility Patent Application Transmittal

Figure 2

This figure presents the distribution of the variable CompanyStatus extracted from Thomson One. 19.65% 9.63% 9.56% 61.16% M&A Write-Off IPO Active