Does Going Public Affect Innovation?

• Date: 01 August 2015
• Author: SHAI BERNSTEIN
• DOI: http://doi.org/10.1111/jofi.12275
• Published date: 01 August 2015

THE JOURNAL OF FINANCE •VOL. LXX, NO. 4 •AUGUST 2015

Does Going Public Affect Innovation?

SHAI BERNSTEIN∗

ABSTRACT

This paper investigates the effects of going public on innovation by comparing the

innovation activity of firms that go public with firms that withdraw their initial

public offering (IPO) filing and remain private. NASDAQ fluctuations during the book-

building phase are used as an instrument for IPO completion. Using patent-based

metrics, I find that the quality of internal innovation declines following the IPO, and

firms experience both an exodus of skilled inventors and a decline in the productivity

of the remaining inventors. However, public firms attract new human capital and

acquire external innovation. The analysis reveals that going public changes firms’

strategies in pursuing innovation.

DOES THE TRANSITION TO PUBLIC equity markets affect innovation? This question

is important given the critical role of innovation in promoting economic growth

(Solow (1957)) and the prevalence of technological firms in the initial public

offering (IPO) market over recent years.1Although a large body of research

examines the performance of firms around their IPO, little is known about the

effects of going public on innovation. This paper’s main contribution is to show

that going public affects three important dimensions of innovation activity: the

creation of internally generated innovation, the productivity and mobility of

individual inventors, and the acquisition of external innovation.

Theoretically, in frictionless financial markets, selling equities publicly

should have no bearing on subsequent innovation activity.Under financial fric-

tions, however, the transition to public equity markets improves firms’ access

to capital, which can lead to an increase in innovation activity, because such

∗Shai Bernstein is with Stanford University. I am deeply grateful to Fritz Foley, Josh Lerner,

Andrei Shleifer, and Jeremy Stein for their invaluable comments. I also thank Philippe Aghion,

Effi Benmelech, Laura Field, Paul Gompers, Robin Greenwood, Sam Hanson, Oliver Hart, Victoria

Ivashina, Dirk Jenter, Bill Kerr,Jacob Leshno, Gustavo Manso, Ramana Nanda, Francisco Perez-

Gonzalez, David Scharfstein, Antoinette Schoar, Amit Seru, Adi Sunderam, Rick Townsend, and

Jeff Zwiebel for helpful comments. I am grateful to seminar participants at Chicago University,

Columbia University, Dartmouth College, Entrepreneurial Finance and Innovation Conference,

Harvard Business School, Hebrew University, London Business School, London School of Eco-

nomics, NBER Productivity Lunch, Northwestern University, Searle Conference on Innovation

and Entrepreneurship, Stanford University, Tel-Aviv University, University of British Columbia,

and University of Pennsylvania for helpful comments and suggestions. Andrew Speen provided

superb research assistance. I am grateful for funding from the Ewing Marion Kauffman Founda-

tion.

1Approximately 40% of all firms that have gone public in the last 40 years are technological

firms.

DOI: 10.1111/jofi.12275

1365

1366 The Journal of Finance R

activity is particularly likely to be sensitive to financing constraints (Arrow

(1962), Hall and Lerner (2010)). On the other hand, agency problems associated

with the transition to public equity markets may undermine firm incentives to

innovate (Berle and Means (1932), Jensen and Meckling (1976)).

Ultimately, empirical investigation is necessary to disentangle the positive

and adverse effects of public listing on innovation. Estimating these effects,

however, is challenging due to an inherent selection bias associated with the

decision to go public. A standard approach to studying the dynamics of firm

outcomes around the IPO is to use within-firm variation. But as Jain and Kini

(1994) note, firms choose to go public at a specific stage in their life cycle, and

as a result this approach may produce biased estimates of the IPO effect. For

instance, if firms choose to go public following an innovation breakthrough,

as argued by P´

astor, Taylor, and Veronesi (2009),2then post-IPO performance

may reflect reversion to the mean and thereby mix life cycle effects with the

IPO effect.

To overcome this selection bias and estimate the IPO effect on innovation, I

construct a sample of innovative firms that file an initial registration statement

with the Securities and Exchange Commission (SEC) in an attempt to go public,

and then either complete or withdraw their filing. Combined with standard

patent-based metrics, this sample allows me to compare the innovation activity

of firms that go public with that of private firms that intended to go public and

hence are at a similar stage in their life cycle. Of course, comparing complete

and withdrawn IPO filings introduces a new bias associated with the decision

of firms to withdraw the IPO filing and remain private.

To address this concern, I use NASDAQ fluctuations over the two months fol-

lowing the IPO filing date as an instrument for IPO completion, relying on fil-

ers’ sensitivity to aggregate stock market movements during the book-building

phase. Consistent with prior literature, I find that these short-run NASDAQ

fluctuations strongly predict IPO completion, with the effect concentrated at

market declines (e.g., Busaba, Benveniste, and Guo (2001), Benveniste et al.

(2003)). In the analysis, the IPO effect is identified from differences in long-run

innovation (i.e., innovation over a five-year period) between firms that file to go

public in the same year but experience different postfiling NASDAQ returns.

For the instrument to be valid, it needs to satisfy the exclusion restriction

condition, that is, two-month NASDAQ returns must be related to the long-run

innovation measures only through the IPO completion choice. I find empirical

support for the validity of the instrument (see Section II.C for a detailed dis-

cussion). While reduced-form analysis shows that the instrument significantly

explains long-run innovation, it may also affect innovation through other chan-

nels. To the extent that this is the case, two-month NASDAQ returns should

also explain long-run innovation outside the book-building phase. A placebo

test reveals that outside the book-building phase, when ownership choice is

fixed, NASDAQ returns have no effect on long-run innovation. This finding is

2Chemmanur, He, and Nandy (2010) find that firms go public following productivity improve-

ments.

Does Going Public Affect Innovation? 1367

consistent with the notion that short-term NASDAQ returns during the book-

building phase affect long-run innovation only through the IPO completion

choice.

Using the instrumental variables approach, I find a significant link between

public ownership and innovation: going public causes a substantial 40% decline

in innovation novelty as measured by patent citations. At the same time, I find

no change in the scale of innovation, as measured by the number of patents.

These results suggest that the transition to public equity markets leads firms

to reposition their R&D investments toward more conventional projects.

Having shown that going public affects the composition of innovative relative

to conventional projects, I study the effects of going public on individual inven-

tors’ productivity and mobility over time. I find that the quality of innovation

produced by inventors who remain at the firm declines following the IPO, and

key inventors are more likely to leave. The firms that go public are also more

likely to generate spin-off companies, suggesting that inventors who leave re-

main entrepreneurial. These effects are partially mitigated by the ability of

public firms to attract new inventors.

I also find a stark increase in the likelihood that newly public firms acquire

companies in the years following an IPO. To better understand whether these

acquisitions are used to purchase new technologies, I collect information on

targets’ patent portfolios. I find that public firms acquire a substantial number

of patents through M&A: acquired patents constitute almost a third of firms’

total patent portfolio in the five years following the IPO. The acquired patents

are of higher quality than the patents produced internally following the IPO.

These results suggest that the transition to public equity markets affects

the strategies that firms employ in pursuing innovation. While publicly traded

firms generate more incremental innovation internally, they also rely more

heavily on acquiring technologies externally. This shift takes place concurrent

with a substantial inventor turnover after the IPO.

What leads to these changes? A stock market listing may increase the scope

for agency problems, leading firms to pursue less innovation following the IPO

(Berle and Means (1932), Jensen and Meckling (1976)). Alternatively, agency

problems may be concentrated in private firms that pursue too much inno-

vation. Finally, it may be the case that the results are not driven by agency

problems, but rather the improved access to capital allows firms to focus on

commercialization after the IPO. Such a strategy is not viable for firms that

remain private and therefore focus on innovation only.I discuss these hypothe-

ses in detail in Section IV. I find supportive evidence for managerial career

concerns, consistent with recent papers on agency problems in public equity

markets (e.g., Aghion, Van Reenen, and Zingales (2013), Fisman et al. (2013),

Asker, Farre-Mensa, and Ljungqvist (2014), Gao, Hsu, and Li (2014)).

The paper is related to several strands of the literature. First, by propos-

ing an identification strategy to isolate the IPO effect, and focusing on firms’

innovation activities around the IPO, this paper contributes to the IPO liter-

ature that explores firm behavior following the IPO and documents a decline

in firm performance measures such as profitability and productivity (Degeorge