Protection of trade secrets and capital structure decisions*

Sandy Klasaa, Hernán Ortiz-Molinab, Matthew Serflingc, and Shweta Srinivasand

May 2017

Abstract:
Firms strategically choose more conservative capital structures when they face greater
competitive threats stemming from the potential loss of their trade secrets to rivals. Following
the recognition of the Inevitable Disclosure Doctrine by U.S. state courts, which exogenously
increases the protection of a firm’s trade secrets by reducing the mobility of its workers who
know its secrets to rivals, the firm increases its leverage relative to unaffected rivals. The
effect is stronger for firms with a greater risk of losing key employees to rivals, for those facing
financially stronger rivals, and for those in industries where competition is more intense.

Keywords: capital structure; trade secrets; intellectual property; competitive threats.

* We thank an anonymous referee, Rajesh Aggarwal, Nilanjan Basu, Jan Bena, Cesare Fracassi, Lorenzo Garlappi,
Ron Giammarino, Jean Helwege, Jayant Kale, Kai Li, Lubo Litov, Neslihan Ozkan, Carolin Pflueger, Gordon Phillips,
Berk Sensoy, Elena Simintzi, Philip Valta, Wei Wang, Ryan Williams, Toni Whited (the Editor), Steven Xiao, and
seminar participants at the University of Arizona, University of British Columbia, University of New South Wales,
University of Sydney, Hong Kong University of Science & Technology, Laval University, Nanyang Technological
University, National University of Singapore, Northeastern University, Universidad de Los Andes (Chile), and also
participants at the One-Day Corporate Finance Conference (2013) at the University of Manchester, the Finance,
Organizations, and Markets (FOM) research group meeting at the University of Southern California (2013), the
Financial Intermediation Research Society (FIRS) Conference (2014), the University of Alberta Frontiers in Finance
Conference (2014), the Northern Finance Association Conference (2014), the European Winter Finance Summit
Conference (2015), the Western Finance Association Conference (2015), and the European Finance Association
Conference (2015) for helpful suggestions. We thank Andrew Bird and Robert Knopf for sharing with us their state-
level data on the strength of enforcement of non-compete agreements and Tobias Berg for making his data on firms’
total cost of bank debt available on his website. We also appreciate the excellent research assistance provided by
Maryam Fathollahi. Ortiz-Molina acknowledges the financial support from the Social Sciences and Humanities
Research Council of Canada.
a Eller College of Management, University of Arizona, Tucson, AZ 85721, sklasa@eller.arizona.edu.
b Sauder School of Business, University of British Columbia, Vancouver, BC V6T1Z2, ortizmolina@sauder.ubc.ca.
c Haslam College of Business, University of Tennessee, Knoxville, TN 37996, mserflin@utk.edu.
d School of Management, State University of New York at Binghamton, Binghamton, NY 13902,

srinivasan@binghamton.edu.
1. Introduction

Financial economists generally agree that risks stemming from a firm’s competitive

environment, such as the risk of predation by rivals, new entry, or competitors improving

their products, can lead the firm to choose its capital structure strategically, i.e., taking into

account how its financial decisions affect rival firms’ future moves and its ability to react to

these moves. Surprisingly, little is known about the relevance for capital structure choices of

competitive threats that originate from a firm’s inability to fully protect its intellectual

property from appropriation by opportunistic rivals. Yet, intellectual property, which

accounts for roughly one-third of the aggregate market equity value of U.S. publicly traded

firms, is among the critical revenue-generating assets that determine a firm’s competitive

advantages and performance in its product market (Shapiro and Hassett (2005)).

We study how a firm’s capital structure decisions are affected by the ex-ante risk that its

industry rivals could try to obtain its intellectual property in the form of trade secrets and

hurt its competitive position. These secrets consist of sensitive information that is not easily

ascertainable by outside parties, such as detailed information about a firm’s customers, price

lists, cost information, information about future business plans (e.g., future products and

services), and also formulas, practices, processes, or designs. Trade secrets are pervasive in

all industry sectors and are very valuable because they provide firms with competitive

advantages over their industry rivals.1 In line with the prior survey evidence in Cohen,

Nelson, and Walsh (2000) and Arundel (2001), a recent survey by Marsh and Liberty

Underwriters reports that trade secrets are firms’ most important form of revenue-

generating intellectual property, followed by trademarks and patents.2

Trade secrets are an important source of risk because the divulgence of such secrets can

erode a firm’s competitive advantages over its rivals and cause the firm significant economic

harm. Highlighting the importance of this risk, a survey sponsored by the U.S. Chamber of

1 For example, Hall, Helmers, Rogers, and Sena (2012) estimate an average value of $6.3 million per trade secret
based on the ruling by a federal court in Virginia in September 2011 that Kolon Industries Inc. was liable to
paying DuPont Co. $919.9 million for the theft of 149 trade secrets related to DuPont’s Kevlar business.
2 See “The 2011 Intellectual Property Survey Report”, Marsh and Liberty International Underwriters, 2011.

1
Commerce estimates that firms lose over $50 billion annually due to the divulgence of their

trade secrets. It also reports that the most common types of trade secrets lost to rivals are

secrets related to a firm’s customers, strategic plans, and financial data.3 Likewise,

underscoring the economic importance of trade secrets, prior work shows that trade secrets

are the most litigated form of intellectual property (Lerner (2006)) and that litigation of trade

secrets is on the rise in both state and federal courts (Almeling et al. (2010, 2011)).

Our main hypothesis follows from prior work which highlights the strategic value of

financial flexibility in the face of intense competition (e.g., Telser (1966), Poitevin (1989), and

Bolton and Scharfstein (1990)), the value of preserving unused debt capacity to meet

unexpected future funding needs (DeAngelo and DeAngelo (2007) and DeAngelo, DeAngelo,

and Whited (2011)), and that competitive threats lead to more conservative financial policies

(e.g., Haushalter, Klasa, and Maxwell (2007), Frésard (2010), Valta (2012), and Hoberg,

Phillips, and Prabhala (2014)). We argue that unused debt capacity is particularly valuable

for a firm facing a high risk that its rivals would try to obtain its trade secrets for two reasons.

First, it allows the firm to raise the funds needed to aggressively protect its competitive

position when its trade secrets are divulged to rivals (e.g., by increasing capital investment,

R&D, advertising, or hiring, or by engaging in costly litigation). Second, it credibly signals to

its rivals that the firm has the financial ability to respond aggressively to the appropriation

of its trade secrets, which in turn reduces rivals’ incentives to obtain these secrets. Hence,

we hypothesize that a greater risk that a firm’s rivals could seek to obtain its trade secrets

and harm its business increases the firm’s strategic benefit of maintaining unused debt

capacity and thus leads to a lower debt ratio.

The challenge in estimating the causal effect of a higher risk of losing trade secrets to

rivals on a firm’s capital structure decisions is to identify exogenous variation in this risk. To

this end, we focus on the mobility of employees with knowledge of trade secrets, which is the

main channel through which firms lose trade secrets to rivals. In the survey sponsored by the

U.S. Chamber of Commerce noted earlier, CEOs report that former employees are the

3 See “Trends in Proprietary Information Loss,” ASIS International, September 2002.

2
greatest source of risk associated with the loss of proprietary information. Also, Almeling et

al. (2011) report that in most legal cases involving trade secrets the misappropriator of a

firm’s trade secrets is a former employee. Importantly, the legal environment surrounding

the mobility of key employees and protection of firms’ trade secrets differs across states and

within a state over time. This allows us to use a novel empirical setting for identification.

Our difference-in-differences tests use the staggered adoption, and in a few cases the

subsequent rejection, of the Inevitable Disclosure Doctrine (IDD) by U.S. state courts over

the 1977-2011 period to gauge changes in the legal protection of firms’ trade secrets. The IDD

is a legal doctrine which states that a firm’s former employee can be prevented from working

for a rival firm if this would “inevitably” lead the employee to divulge the firm’s trade secrets

to the rival. It is applicable even if the employee did not sign a non-compete or non-disclosure

agreement with the firm, there is no evidence of bad faith or actual wrongdoing, or the rival

is located in another state. Hence, by increasing a firm’s ability to prevent employees who

know its trade secrets from working for rivals, the recognition of the IDD in the firm’s state

reduces the risk that rivals will obtain its trade secrets.

We show that the recognition of the IDD in a given state significantly reduces the

mobility to rival firms of workers in managerial and related occupations (which entail access

to a firm’s trade secrets) relative to the mobility to rival firms of workers in other occupations.

The variation in the protection of firms’ trade secrets associated with this legal event is

arguably exogenous and thus a potentially useful source of identification that can aid finance

research. Specifically, the legal background surrounding state courts’ recognition of the IDD

(discussed in Section 4.1) suggests that the occurrence of this event is unlikely to be

systematically associated with changes in business or political conditions in a state, lobbying,

or to be anticipated by firms. Empirically, we find that a state’s labor laws or adoption of the

Uniform Trade Secrets Act, worker characteristics, and economic or political conditions do

not affect state courts’ decisions to adopt the IDD.

Our key result is that the recognition of the IDD leads to increases of about 14-19% in

the net book and market leverage ratios of firms headquartered in recognizing states relative

3
to those of their industry rivals in non-affected states. Consistent with the validity of the

parallel trends assumption that is central to a causal interpretation of our results, we show

that the pre-treatment trends in the leverage of treated and control firms are statistically

indistinguishable, i.e., the increases in the financial leverage of treated firms relative to

control firms occur after the adoption of the IDD, but not before. These results suggest that

a decrease in their risk of losing trade secrets to rivals causes firms to increase their leverage.

As in any difference-in-differences setup, an additional concern is whether post-

treatment trends would have remained parallel had it not been for the legal change.

Specifically, the worry is that omitted variables correlated with both the adoption of the

IDD and capital structure decisions could drive the results by affecting post-treatment

trends in debt ratios. Our empirical tests mitigate this concern: our results hold after the

inclusion of standard controls used in capital structure tests, controls for the economic and

political conditions prevailing in a state and for the enforcement of covenants not to compete

by the courts in a state, firm fixed effects that control for time-invariant firm-level factors,

and industry-year fixed effects that control for time-varying industry heterogeneity.

Admittedly, we cannot entirely rule out that other omitted variables might drive the

results through channels other than the protection of firms’ trade secrets. For example, we

cannot control for unobservable state-level changes in investment opportunities that coincide

with the recognition of the IDD and require debt financing. However, inconsistent with

omitted variables driving our results, the recognition of the IDD has a larger impact on the

net leverage of firms that face a greater ex-ante risk of losing employees who know their

trade secrets to rivals. Further, supporting the economic mechanism behind our main

hypothesis, the recognition of the IDD has a larger effect on net leverage for financially weak

firms that face financially strong rivals and for firms that face heightened competitive threats

due to higher asset specificity or lower barriers to entry in their industry. It also has a smaller

effect for firms with more differentiated products and thus stronger competitive positions

relative to their industry rivals. Overall, these results highlight the strategic value of

maintaining unused debt capacity for firms that face competitive threats associated with the

4
risk of losing trade secrets to rivals, as implied by our hypothesis.

We provide evidence inconsistent with other (non-strategic) explanations of our main

results. First, the recognition of the IDD might lead firms to lever up by raising their future

profitability and the tax benefits of debt financing. But our results hold controlling for future

profitability and the recognition of the IDD does not affect firms’ future profitability or have

a stronger effect on the leverage of firms with higher expected tax benefits of debt financing.

Second, the recognition of the IDD could cause firms to lever up by reducing general cash

flow uncertainty. However, our results hold controlling for contemporaneous changes in

general cash flow uncertainty and the recognition of the IDD does not impact measures for

this uncertainty. Finally, the recognition of the IDD might lead firms to lever up to finance

increased spending in the development of trade secrets (e.g., R&D, capital, or advertising

expenditures). However, the recognition of the IDD does not affect such expenditures.

We also examine the association between firms’ net leverage ratios and two firm-level

indicators for whether a firm reports that it faces a significant risk of losing trade secrets to

its competitors that are based on textual analyses of firms’ 10-K filings. One uses the

approach in Eisfeldt and Papanikolaou (2013) and the other follows the approach in Hoberg

and Maksimovic (2015). Consistent with our hypothesis and with our results based on the

recognition of the IDD, we find a statistically and economically significant negative

association between both of these indicators and net financial leverage.

Further tests show that affected firms earn positive abnormal returns over the days

surrounding a state court’s final decision to recognize the IDD, and that following this

recognition their cost of bank debt decreases. These results further support the notion that

the recognition of the IDD reduces the competitive risk these firms face. Lastly, the

recognition of the IDD is unrelated to changes in the product fluidity measure of Hoberg et

al. (2014) and in other firm-level measures of competitive risk, suggesting that IDD rulings

capture a distinct aspect of competitive risk.

Our paper brings attention to important changes in the competitive landscape over the

last 50 years. As reported in Almeling (2012), firms are increasingly reliant on their

5
intellectual capital, in particular trade secrets, as a source of competitive advantage over

their rivals and are concerned with the protection of this capital. Also, employees who know

trade secrets are ever more mobile and less attached to their current employers, and rivals

have incentives to poach them to gain access to those trade secrets. Hence, competitive

threats stemming from labor markets are increasingly important and worth studying.

Closely related work examines how competitive threats related to firms’ output (products

and services) affect a firm’s financial policies (e.g., MacKay and Phillips (2005), Lyandres

(2006), Haushalter et al. (2007), Frésard (2010), Valta (2012), and Hoberg et al. (2014)). Our

contribution is to highlight the increasing importance of threats related to a scarce key input,

namely, the firm’s intellectual property. Specifically, losing trade secrets to rivals is a major

competitive threat that leads firms to choose their capital structures strategically.

Our paper is also related to recent work on how labor market frictions affect debt ratios

(e.g., Matsa (2010), Agrawal and Matsa (2013), and Simintzi, Vig, and Volpin (2015)). These

studies show that, when firms are concerned with their labor costs, debt ratios depend on

labor unions’ bargaining power, employee unemployment risk, and the rigidity of labor costs.

Because workers who know trade secrets typically account for a small fraction of a firm’s

total labor costs, they are unlikely to affect capital structure through these cost-related

mechanisms. Still, we find that the mobility of such workers can have a large impact on a

firm’s capital structure by affecting the protection of its valuable trade secrets.

Last, our paper is related to prior work on the use of non-compete agreements to protect

trade secrets and the effect of these agreements on labor mobility, investment in human

capital, and innovation (e.g., Marx, Strumsky, and Fleming (2009), Garmaise (2011), and

Samila and Sorenson (2011)). Although we focus on strategic capital structure issues and a

broader legal doctrine affecting trade secret protection, both these papers and ours highlight

the interplay between labor mobility, trade secret protection, and corporate outcomes.

The rest of the paper is organized as follows. Section 2 develops our hypothesis and

predictions. Section 3 discusses the IDD. Section 4 discusses identification. Section 5 reports

our key results. Section 6 reports further investigations. Section 7 concludes.

6
2. Conceptual framework and empirical predictions

Theoretical research relating capital structure choices and product market competition

emphasizes that a firm’s financial situation can affect its conduct as well as that of its rivals,

and ultimately its competitive position in its product market. The “deep pockets” argument,

put forward in Telser (1966) and formalized in Poitevin (1989) and Bolton and Scharfstein

(1990), highlights the importance of maintaining financial flexibility for a firm facing large

competitive threats. In Bolton and Scharfstein (1990), the optimal debt contract balances the

benefits of deterring predation by reducing a firm’s financial constraints against the cost of

exacerbating managerial incentive problems. In Poitevin (1989), a high quality entrant must

finance with debt to separate itself from a low quality entrant, while an incumbent who

finances with equity can exploit the entrant’s financial weakness to drive it into bankruptcy.

A key insight from Poitevin (1989) and Bolton and Scharfstein (1990) is that when a firm

holds more debt this increases opportunistic rivals’ incentives to launch competitive attacks

on the firm and that firms’ capital structure decisions take such strategic issues into account.4

The theory suggests that unused debt capacity is strategically valuable because it allows

a firm to aggressively respond to competitive attacks by its rivals, e.g., by increasing capital

investment, advertising, R&D spending, or employee recruiting and retention efforts. It can

also help a firm to deter such attacks by signaling to its rivals that the firm has the resources

to retaliate if attacked, which reduces the expected benefit to its rivals of engaging in such

tactics. Consistent with this notion, the empirical evidence shows that the financial flexibility

provided by lower financial leverage (relative to rivals’ leverage) allows firms to react to

opportunistic behavior by their rivals (e.g., Phillips (1995), Chevalier (1995), Kovenock and

Phillips (1997), Zingales (1998), Campello (2003, 2006), and Khanna and Tice (2000, 2005)).

Further, it suggests that firms will strategically take into account the competitive threats

they face when deciding on their financial policies (e.g., Mackay and Phillips (2005),

Haushalter et al. (2007), Frésard (2010), Valta (2012), and Hoberg et al. (2014)).

4 See also Maurer (1999) and Faure-Grimaud (2000) for additional theoretical support for the notion that when a
firm increases its outstanding debt this can weaken its competitive position in its product market.

7
 The novelty of our paper is to focus on how capital structure choices are affected by

competition among firms with regards to the access to trade secrets, a key form of revenue-

generating intellectual capital, that can give the proprietor of the trade secrets important

competitive advantages over its rivals. Our key insight is that the risk of losing trade secrets

to rivals is a key competitive threat that, by increasing the strategic value of maintaining

financial flexibility to react when these threats materialize or to deter them ex ante, can

affect a firm’s capital structure choices.

The conceptual framework behind our empirical analyses builds on prior work which

brings to light the strategic value of having unused debt capacity in the face of intense

competition (e.g., Telser (1966), Poitevin (1989), and Bolton and Scharfstein (1990)), the value

of preserving unused debt capacity to meet unexpected future funding needs (DeAngelo and

DeAngelo (2007) and DeAngelo et al. (2011)), and that larger competitive threats lead to more

conservative financial policies (e.g., Haushalter et al. (2007), Frésard (2010), Valta (2012),

and Hoberg et al. (2014)). Specifically, we draw attention to two roles of unused debt capacity.

First, unused debt capacity is strategically valuable for a firm that faces a larger risk of

losing trade secrets because, by maintaining a firm’s ability to borrow in the future, a lower

debt ratio allows the firm to raise funds and swiftly take actions to protect its competitive

position in the event its trade secrets are divulged to rivals. For instance, if the lost trade

secrets were related to a firm’s future products, the firm could make additional investments

in product development to further differentiate its products from those of its rivals. The firm

could also raise funds to ensure that, regardless of any adverse consequences from the loss of

its trade secrets, it can continue to fully invest in its growth opportunities and make its

payments to creditors and suppliers. Further, the firm could also respond to the loss of its

trade secrets by entering into costly litigation against rivals.

Second, unused debt capacity is also strategically valuable because it could deter a firm’s

rivals from actively seeking to obtain and commercially exploit its trade secrets. Maintaining

unused debt capacity increases the firm’s ex-ante ability to aggressively retaliate to protect

its competitive position against such attempts, for example, by increasing investment or by

8
entering into costly litigation, and it credibly conveys this ability to its rivals. This directly

reduces rivals’ expected benefit of opportunistically seeking to obtain the firm’s trade secrets.

These two arguments imply that, if a firm faces competitive threats associated with the

risk of its rivals obtaining its trade secrets, its optimal capital structure choice trades off the

strategic benefits of unused debt capacity discussed above against the costs of less borrowing

(e.g., forgone tax shields). Hence, ceteris paribus, the firm’s optimal debt ratio is higher when

it faces less such threats. This leads to our main hypothesis: a decrease in the risk that a

firm’s rivals could obtain its trade secrets and harm its competitive position reduces its

strategic benefit of keeping unused debt capacity and thus leads the firm to optimally increase

its debt ratio. The empirical assessment of this hypothesis benefits from a clear identification

of the mechanism through which a firm might lose trade secrets to rivals.

As noted in our introduction, the main source of the risk that a firm’s trade secrets will

be divulged to its rivals is the mobility of its employees with knowledge of its trade secrets.

Anecdotal evidence further suggests that the poaching of employees with access to a rival’s

trade secrets is a key element of firms’ competitive strategies. For example, in early 2016

XPO Logistics sued its competitor YRC Worldwide on the basis that YRC deliberately

“targeted and raided” XPO employees in key executive and sales roles as a means to access

the carrier’s “most valuable trade secrets.”5 Further, in the spirit of Telser’s (1966) deep

pockets argument, Kim (2014) shows theoretically that financially strong firms can engage

in “predatory poaching” against a financially weaker rival, i.e., offer higher wages to poach a

rival’s employees who know its trade secrets and deprive the rival of its competitive

advantage. Hence, our empirical tests aim to gauge how changes in a firm’s risk of losing

trade secrets to rivals through the employee mobility channel affect its capital structure.

As we explain in Section 3, in our tests we rely on exogenous changes in the legal

protection of trade secrets that reduce the mobility of employees with knowledge of trade

secrets to rival firms. Intuitively, an increase in the protection of a firm’s trade secrets

5See the article “XPO Logistics Sues Trucker YRC, Charging Rival ‘Poached’ Executives, Trade Secrets” by
Loretta Chao in the Wall Street Journal online published on February 5, 2016 for more detail.

9
increases the cost to rivals of obtaining the firm’s trade secrets and thus reduces the

competitive threats the firm faces. Hence, the optimal adjustment of the firm’s capital

structure following an increase in the protection of its trade secrets could lead the firm to

transit from one product market equilibrium with high competitive threats (high benefit of

unused debt capacity) and low leverage to another with lower competitive threats (lower

benefit of unused debt capacity) and higher leverage.

Importantly, our empirical tests need to isolate the strategic effect underlying our main

hypothesis from other potential non-strategic effects that could confound the inferences. In

particular, within the standard tradeoff theory, an increase in the legal protection of a firm’s

trade secrets could lead the firm to lever up simply by increasing its future profits and thus

the tax benefits of debt financing or by reducing uncertainty about the firm’s future

profitability. Alternatively, in dynamic models of leverage, higher expected future

profitability can make a firm conserve debt capacity to ensure it is able to finance good

investment projects that may arrive in the future (e.g., DeAngelo et al. (2011)). Thus, if

increased protection of trade secrets were to increase affected firms’ expected profits, this

could lead these firms to decrease their financial leverage. Further, another potential

confounding effect is that increased protection of their trade secrets could lead firms to lever

up to finance increased spending in the development of trade secrets.

The general empirical prediction generated by our hypothesis is that an increase in the

legal protection of a firm’s trade secrets from appropriation by its rivals leads the firm to

lever up relative to all other firms whose protection is unaffected. There are also testable

implications of our hypothesis which help to isolate the strategic effect from other effects.

First, the theoretical and empirical work previously discussed which relates capital

structure choices and product market competition suggests that a firm should lower its debt

ratio relative to that of its rivals when it faces a higher risk that these rivals will launch

competitive attacks on it. Consequently, if there is heterogeneity in the changes in the legal

protection of trade secrets across firms in an industry, the strategic view implies that firms

which experience increased legal protection of their secrets should lever up relative to their

10
rivals that do not experience this.

Second, the strategic benefit of unused debt capacity is greater for a firm with limited

access to external capital that faces financially strong rivals. The reason is that such rivals

could more aggressively seek and exploit trade secrets obtained from the firm and inflict

greater damage on its competitive position. Moreover, they are more likely to do so when they

perceive that the firm has a limited financial ability to respond to competitive attacks.

Noteworthy, in their 10-Ks firms often report that, because their rivals are financially

stronger than they are, they face a higher risk of losing trade secrets to rivals, their rivals

could significantly exploit their trade secrets if they obtained these secrets, and that they

may be unable to adequately respond to the acquisition of their trade secrets by rivals if this

requires significant financial resources.6 This suggests that changes in the legal protection of

a firm’s trade secrets should have a larger impact on its capital structure when it has limited

access to external capital and it faces financially stronger rivals.

Third, the strategic benefit of unused debt capacity is larger in industries with more

specific assets. Firms in such industries face heightened competitive threats, since they find

it more difficult to sell assets to raise funds to meet unexpected financing needs (Valta (2012)

and Ortiz-Molina and Phillips (2014)), for example if they lose trade secrets to rivals. Asset

specificity is also a barrier to exit for the less productive firms in the industry, leading to

excess capacity and more aggressive behavior by incumbents (Porter (2008)). Hence, changes

in the legal protection of a firm’s trade secrets are likely to have a larger impact on the firm’s

capital structure in industries with higher asset specificity.

Fourth, a key threat a firm faces is that the divulgence of its trade secrets could facilitate

the entry of new firms into its product market and the erosion of its competitive position.

However, barriers to entry make it difficult for new firms to enter the industry even if they

gain access to the incumbent firms’ trade secrets and thus reduce such threats.7 Hence, the

strategic benefit of keeping unused debt capacity is smaller in industries with higher barriers

6For instance, see Vanguard Minerals’ 2012 10-K and Oragenics’ 2014 10-K.
7Barriers to entry also soften competition leading to higher profits and survival rates. Thus, the financial distress
a firm would suffer from losing its trade secrets may be less severe in industries with higher barriers to entry.

11
to entry, and thus changes in the legal protection of a firm’s trade secrets should have a

smaller impact on firms’ leverage in industries with higher barriers to entry.

Finally, prior work argues and shows that firms use R&D and advertising to differentiate

their products from those of their competitors and that when a firm increases these

expenditures it experiences reductions in competition (e.g., Hoberg and Phillips (2016)). As a

result, the strategic benefit of maintaining unused debt capacity is smaller for firms for whom

these expenditures are high relative to those of their industry rivals, and consequently

changes in the legal protection of a firm’s trade secrets should have a smaller impact on the

leverage ratios of such firms.

3. The inevitable disclosure doctrine and construction of the IDD indicator

3.1. Origin and application of the IDD

The protection of trade secrets is largely governed by state law. The Uniform Trade

Secrets Act (UTSA) defines a trade secret as any information that (i) derives independent

economic value, actual or potential, from not being generally known to, and not being readily

ascertainable by proper means by other persons who can obtain economic value from its

disclosure or use, and (ii) is the subject of efforts that are reasonable under the circumstances

to maintain its secrecy. Misappropriation occurs when the trade secret is acquired by (i)

improper means (e.g., theft or breach of a duty to maintain secrecy) or (ii) disclosure without

express or implied consent by a person who acquired the trade secret under circumstances

giving rise to a duty to maintain its secrecy or limit its use.

The key legal concept in understanding the applicability of the IDD is “threatened

misappropriation”, which occurs when an employee who has knowledge of a firm’s trade

secrets goes to work for a direct competitor in a similar position. The IDD is a legal doctrine

based on a strong interpretation of the concept of threatened misappropriation which does

not immediately follow from the general principles in trade secrets law (e.g., as codified in

the UTSA). It maintains that, if the new employment would inevitably lead to the disclosure

of the firm’s trade secrets to a competitor and cause the firm irreparable harm, then upon

the firm’s request state courts can prevent the employee from working for the firm’s
12
competitor or can allow it but limit the responsibilities the worker can undertake.

The adoption of the IDD by a state court enhances the protection of trade secrets for firms

located in the state by reducing the risk that departing employees who know their firm’s

trade secrets will reveal them to rival companies (in any state) or use them to start a rival

company. Under the IDD, a firm’s suit can rest on the mere threat of irreparable harm. To

obtain an injunction, the firm must only establish that (i) the employee had access to its trade

secrets, (ii) the employee’s duties at the new employer would be so similar to those she had

at the firm that in performing them she will inevitably use or disclose the trade secrets, and

(iii) the disclosure of the trade secrets would produce irreparable economic harm to its

business. However, the firm need not establish actual wrongdoing by the employee or disclose

the actual details of the underlying trade secrets in the lawsuit.8

Noteworthy, lawsuits related to employment contracts are filed in the context of

employment law, and thus the relevant jurisdiction for a lawsuit seeking to protect a firm’s

trade secrets when employees switch employers is typically the state where the former

employee worked (Malsberger (2004) and Garmaise (2011)).9 As a result, the IDD protects a

firm’s trade secrets even if the new employer of a firm’s former worker is in another state

whose courts have not adopted the IDD.

3.2. Non-disclosure and non-compete clauses in employment contracts

Employment contracts sometimes contain a non-disclosure agreement (NDA) and/or a

covenant not to compete (CNC). Both are designed to protect the firm’s trade secrets in cases

in which employees wish to switch jobs or start competing firms and allow the firm to bolster

its suit by including a claim of breach of contract.

The protection offered by NDAs is somewhat limited because violations must be detected

and proved before the firm can initiate legal action against a former employee. Further, even

if the firm can prove that an employee violated an NDA, by that time the harm has already

8In the Internet Appendix, we discuss two legal cases that illustrate how state courts apply the IDD.
9The former or new employer’s state of incorporation (even if different from the state where the employer’s main
operations are located) or the employee’s state of residence do not play a role in the application of the IDD.

13
been done. As noted by Garmaise (2011), CNCs are most effective when workers seek to

switch jobs within a state. CNCs are much less effective when workers try to switch to a new

job in another state because the geographical area where a firm’s former employee may not

compete is typically limited to a state or a part of a state (Malsberger (2004)).

The IDD provides significant additional protection of a firm’s trade secrets even if the

firm’s employees sign NDAs and/or CNCs. First, it does not entail specific geographic

restrictions and thus it is more far reaching than CNCs. Second, it increases the

enforceability of NDAs and CNCs. For instance, it allows courts to prohibit an individual’s

employment at a rival firm if this would inevitably lead to a future violation of an NDA, i.e.,

before an actual violation is detected. The IDD is also a powerful means to establish a key

element in any legal action to enforce a CNC, i.e., a significant likelihood of irreparable harm

to the firm if the employee is allowed to work for the rival. Finally, the IDD allows courts to

grant an injunction even if the employee did not sign an NDA or CNC with the former

employer, i.e., solely on the basis that disclosure of the trade secrets is inevitable.

3.3. Adoption/rejection of the IDD by state courts and construction of the IDD indicator

To create the IDD indicator variable we use in our empirical tests, we first create a list

of the main legal cases addressing the IDD in each state based on historical accounts in prior

legal studies (Kahnke, Bundy, and Liebman (2008), Waldref (2012), Wiesner (2012), and

Malsberger (2011)). We then read all court rulings and identify the precedent-setting case

adopting the IDD as the earliest case in which the court’s decision (i) acknowledges that the

IDD can be used to prevent a firm’s former employee from working at a rival firm and (ii)

does not justify the use of the IDD by referring to an earlier case in the same state that used

the IDD.10 To identify the precedent-setting cases rejecting the IDD in a state that had

previously adopted it, we examine the legal cases that the studies above flag as reversals of

courts’ prior adoptions of the IDD and confirm that (i) the IDD was indeed rejected in these

10Some of the cases do not explicitly refer to the “Inevitable Disclosure Doctrine”, but as in prior legal work we
interpret them as adoptions of the IDD because the rulings are based on identical principles. We also note that in
some of the cases the court rulings explicitly recognized the general applicability of the IDD, but did not use it
due to special circumstances (e.g., the plaintiff failed to establish the existence of a trade secret).

14
cases and (ii) the case decision entails the first rejection of the IDD in the state.

Table 1 lists the 21 precedent-setting cases in which state courts adopt the IDD and the

three cases in which state courts later reject it. The earliest adoption was in New York in

1919, followed by three adoptions in the 1960s, one in the 1970s, four in the 1980s, nine in

the 1990s, and three in the 2000s. Three states (Florida in 2001, Michigan in 2002, and Texas

in 2003) reject the IDD after recognizing it in prior years.

A precedent-setting case becomes case law and courts in the state will subsequently

follow its ruling on the applicability of the IDD. Hence, we use the dates of these cases to

construct our indicator for whether state courts recognize the IDD in any given year. For the

21 states whose courts adopted the IDD, we set the IDD indicator equal to zero in all years

preceding the date of the precedent-setting case, and equal to one the year of the case and

afterwards. We allow the value of the IDD indicator to revert to zero in the three cases in

which a subsequent court decision reverses the state’s position regarding the IDD and

explicitly rejects the IDD. For the 29 states whose case law did not explicitly consider or

considered but rejected the IDD, we set the IDD indicator equal to zero in every year.

4. Identification strategy

4.1. Approach and conceptual discussion

We use a difference-in-differences approach to examine how the recognition of the IDD

affects the capital structure of firms in affected states. As already noted, the relevant

jurisdiction for the application of the IDD is the state where the employee works. Firms often

employ workers in different states, but data restrictions only allow us to identify a firm’s

state of headquarters. But, within our conceptual framework, only the employment location

of workers who know trade secrets matters for capital structure decisions. Hence, our tests

assume that workers who know trade secrets are higher-level employees who are employed

for the most part at firms’ headquarters (see the Internet Appendix for robustness tests).

Our pooled OLS regression models use net financial leverage as the dependent variable

and the indicator for if courts recognize the IDD in the firm’s state of headquarters (IDD) as

the key independent variable. In all models, we include firm fixed effects and industry-year
15
fixed effects. We cluster standard errors by state of headquarters because IDD is a state-level

variable. We report specifications with and without control variables, because some of these

variables could be affected by the recognition of the IDD (i.e., be endogenous) and bias our

estimates. The coefficient on IDD gauges the effect of changes in state courts’ position on the

IDD on a firm’s leverage relative to the leverage of rival firms in unaffected states.

The recognition of the IDD protects a firm’s trade secrets as long as its employees who

know these secrets have signed employment contracts under the labor laws in the firm’s

state. To illustrate what our difference-in-differences estimates capture, consider the

adoption of the IDD in state X, but not in state Y. A firm in state X becomes more protected

from the risk of losing trade secrets to rivals both in state X and in state Y. Hence, the firm

in state X faces a lower benefit of maintaining unused debt capacity to respond in case it loses

trade secrets to rivals.11 In contrast, for a firm in state Y the risk of losing trade secrets to

rivals, and thus the benefit of maintaining unused debt capacity to respond to such events,

is unaffected. This implies that the adoption of the IDD in state X, should lead firms in state

X to lever up relative to rivals in state Y, and this is what the estimated coefficient of IDD

captures in the specification with firm fixed effects and industry-year fixed effects.

The crucial assumption behind the identification strategy described above is that

changes in state courts’ positions regarding the IDD over time provide an exogenous source

of variation in the protection of firms’ trade secrets in the context of our capital structure

tests. Below we provide several reasons why this assumption is likely to be valid.

First, in changing their views on the applicability of the IDD, state courts do not directly

aim to affect firms’ capital structure choices. Instead, the judicial decisions in the precedent-

setting cases involving the IDD are mainly aimed at striking a balance between employers’

interests in protecting their trade secrets and public policy concerns related to employee

mobility and freedom of employment (see Godfrey (2004) and Harris (2000)).

Second, we note that our natural experiment is not based on state laws whose passage

11 The firm in state X is also less likely to obtain secrets from its rivals in state X. Hence, it might also face a lower
benefit of maintaining unused debt capacity to be able to exploit trade secrets obtained from rivals.

16
could be influenced by the lobbying of affected parties with political clout in the state, such

as organizations representing workers or companies. Instead, the experiment is based on

judicial decisions that are typically driven only by the merits of the specific case. The reason

for this is that the judges serving in state courts are deemed to be independent of both the

state and federal government, and thus largely immune to political pressure.

Third, changes in courts’ positions regarding the IDD are unlikely to be anticipated by

corporations. In the context of state courts’ decisions on legal cases related to the protection

of trade secrets, a court’s issuance of a new precedent is typically an idiosyncratic function of

the particular case and the disposition of the justices. Also, because the loss of trade secrets

can cause significant harm to a firm, courts’ decisions in these cases are often rendered

quickly, which further makes it difficult for firms to anticipate their outcomes.

Nevertheless, in Internet Appendix Table A1 we examine whether state-level factors

affect the likelihood that a state court will adopt the IDD. We consider state-level variables

related to labor and trade secrets laws, labor unionization, characteristics of the workforce,

local economic and political conditions, and local shocks to growth opportunities. Not

surprisingly, we find that the adoption of the IDD is more likely in states with stronger

enforcement of covenants not to compete, but no other variable affects the likelihood that

state courts will adopt the IDD. This evidence supports the notion that the adoption of the

IDD by a state court is a plausibly exogenous event.

4.2. The recognition of the IDD and the mobility of workers who know trade secrets

The key to our identification strategy is that, by restricting the mobility of a firm’s

workers who know its trade secrets to rival firms, the recognition of the IDD in a firm’s state

increases the protection of its trade secrets. In this section, we show that the recognition of

the IDD significantly reduces the mobility of a firm’s workers who know its secrets to rival

firms using data from the Census Bureau’s Survey of Income and Program Participation

(SIPP).The SIPP is a nationally representative sample of individuals interviewed over 8-16

consecutive periods that are in most cases four months apart. The data identifies a worker’s

employer, the employer’s 3-digit Census Industry Classification (CIC), and the Integrated
17
Public Use Microdata Series (IPUMS) code for the worker’s occupation. We exclude

individuals who are less than 18 years old, those observed in less than five survey periods,

and those employed in the financial or utility industries. We identify individuals who left

their firms to work for rival (non-rival) firms as those who switched between employers in

the same (different) 3-digit CIC industry from the prior to the current survey period. Our

monthly data spans the 1983-2011 period and contains 799,533 observations.

In Table 2, we use linear probability models and a difference-in-difference-in-differences

approach to study whether the recognition of the IDD affects the mobility of individuals in

managerial and related occupations (who have access to trade secrets) relative to that of

individuals in other occupations (with less likely access to trade secrets).12 The dependent

variables are binary indicators for whether an individual leaves her former employer to join

a rival employer in any state (models 1-2), a rival employer in another state (models 3-4), a

rival employer in the same state (models 5-6), or a non-rival employer in any state (models

7-8). The key independent variables are IDD, Mgr&Rel (an indicator equal to one for

individuals in managerial and related occupations, and zero otherwise), and IDD  Mgr&Rel.

All models include individual fixed effects, year-month fixed effects, and indicators for the

number of months between the current and past survey interviews. Also, in models 2, 4, 6,

and 8, we include an individual’s average monthly income and hours worked per week during

the prior interview period (in logs), an indicator for whether the individual attended school

during the interview period and the following state-level control variables, Strength of CNCs,

an index for the enforcement in a state of covenants not to compete, State GDP Growth, the

one-year growth rate of state GDP, and Political Balance, the fraction of a state’s congress

members in the U.S. House of Representatives that belong to the Democratic Party, which

captures the political leaning in the state. The standard errors are clustered by state.

The coefficient on IDD  Mgr&Rel gauges the difference-in-difference-in-differences

estimate of the impact of the recognition of the IDD in a state on the mobility of individuals

12We exclude workers in engineering and scientific occupations from the set of other occupations because these
employees may know some technical trade secrets.

18
in managerial and related occupations relative to its impact on the mobility to rival firms of

individuals in other occupations. Our key tests, reported in models 1 and 2, study how the

recognition of the IDD affects the mobility of workers to industry rivals in any state. The

coefficients on IDD  Mgr&Rel are negative and statistically significant. Hence, the

recognition of the IDD reduces the mobility to rival firms of individuals in managerial and

related occupations relative to that of individuals in other occupations. The coefficient on IDD

is insignificant, indicating that the recognition of the IDD does not affect the mobility to rival

firms of individuals in other occupations, who are less likely to know trade secrets. The

coefficients on IDD and IDD  Mgr&Rel imply that after the recognition of the IDD in a state

the probability that individuals employed in managerial and related occupations will become

employed at a rival firm decreases by 0.7 percentage points. Relative to the 2.8% of workers

in managerial and related occupations in our sample that are employed in states that have

not adopted the IDD and move to rival firms between survey interviews, this represents a

25% decrease in the mobility of workers who are likely to know trade secrets.

The recognition of the IDD in a state should reduce the mobility of workers in that state

who know trade secrets to rival firms in any state. Nevertheless, we separately study how

the recognition of the IDD affects the mobility of workers to rival firms in other states (models

3-4) and to rivals in the same state (models 5-6). The coefficient estimates for IDD  Mgr&Rel

indicate that the recognition of the IDD reduces the mobility of workers in managerial and

related occupations relative to that for workers in other occupations both to rivals in other

states and to rivals in the same state. Hence, the IDD has a far reaching effect on the mobility

of workers who know trade secrets.

In models 7-8, we conduct a falsification test, i.e., we examine if the recognition of the

IDD reduces the mobility to non-rival firms of individuals in managerial and related

occupations. By design, the recognition of the IDD cannot restrict the mobility of workers

who know trade secrets to non-rival firms. We find that the recognition of the IDD does not

reduce the mobility to non-rival firms of individuals in managerial and related occupations

relative to that of individuals in other occupations. Thus, the effect of the recognition of the

19
IDD on the mobility of workers who know trade secrets to rival firms that we document in

models 1-6 is not driven by general trends in labor mobility. Overall, the Table 2 results

provide further support for our identification strategy and also highlight the mechanism

through which the recognition of the IDD increases the protection of a firm’s trade secrets.

5. Recognition of the IDD and capital structure

5.1. Data and sample selection

Our sample consists of all firms in the merged CRSP-Compustat database (excluding

utilities and financials) headquartered in the U.S. for which we can construct the variables

used in our main capital structure tests. The sample period is 1977-2011, and it starts five

years before Pennsylvania adopts the IDD in 1982 and ends five years after Kansas adopts

the IDD in 2006. Over this period, courts in 16 states adopt the IDD and courts in three states

reject the IDD they had previously adopted. Our sample period excludes the events associated

with the adoption of the IDD by a few states in earlier years because the coverage of these

years in Compustat is sparser, and earlier recognition events thus do not affect many firms

and have little power for identification. The final sample has 125,895 firm-year observations.

Our data looks similar to that used in prior capital structure research. Internet Appendix

Table A2 reports summary statistics for the variables used in our main tests.

5.2. Main results

Table 3 reports the difference-in-differences estimates of the impact of the recognition of

the IDD by state courts on the capital structures of firms in affected states. The estimates

reflect the adoption of the IDD in 16 states and the subsequent rejection of the IDD in three

states, but for simplicity we interpret them as the impact of the “recognition” of the IDD on

capital structure. Models 1-2 (3-4) report the results for net book leverage (net market

leverage). Focusing on debt net of cash holdings accounts for the financial flexibility from

cash holdings. In models 1 and 3 we include the IDD indicator as well as firm and three-digit

20
SIC times year fixed effects.13 In models 2 and 4 we add standard control variables in capital

structure tests (natural logarithm of book assets, market-to-book assets, return on assets,

proportion of assets that are fixed, cash flow volatility, and an indicator for if a firm pays

common dividends), as well as Strength of CNCs, State GDP Growth, and Political Balance.

The recognition of the IDD has a positive and statistically significant impact on the net

leverage of firms in an affected state. The results in models 2 and 4 provide strong support

for the key implication of the strategic view of capital structure choices underlying our main

hypothesis, namely, that an increase in the legal protection of a firm’s trade secrets should

lead the firm to lever up relative to its industry rivals whose protection remains unchanged.

The estimated coefficients imply that after the recognition of the IDD affected firms increase

their ratio of total debt (net of cash holdings) per dollar of book (market) assets by 1.3 (1.2)

cents relative to that of their industry rivals in non-affected states. This represents a 19.1%

(14.6%) increase relative to the sample mean for net book (market) leverage of 0.068 (0.082). 14

These magnitudes are economically significant and comparable to the 23% decrease in net

book leverage following the adoption of the good faith exception to the employment at will

doctrine documented by Serfling (2016).

Models 5-8 use the standard measures of book and market leverage (cash holdings are

not subtracted from total debt).15 The estimates in models 6 and 8 imply that the recognition

of the IDD leads to a 5.2% (5.6%) increase in the book (market) leverage ratios of firms in

recognizing states relative to their sample means. The magnitudes of these effects are

comparable to those reported in prior work on how shocks affect firms’ debt ratios. For

example, Xu (2012) finds that a 2 percentage-point increase in an industry’s import

penetration ratio (a 10% increase over its mean) leads to a 8%-9% decrease in firms’ leverage;

13 All results in the paper are similar if we use year fixed effects instead of industry-year fixed effects. Internet
Appendix Table A3 reports results analogous to those in Table 3 but using year fixed effects.
14 The recognition of the IDD does not affect a firm’s cash holdings, so our net leverage results are driven by

changes in debt financing. An explanation for this is that, due to agency problems, uncertainty about future cash
flows, or costly external financing, cash is not negative debt (Jensen (1986), Acharya, Almeida, and Campello
(2007), and Gamba and Triantis (2008)). Frésard (2010) discusses this issue in the context of product market
competition and argues that it is likely that cash and debt play distinct roles in influencing competitive outcomes.
15 Table A4 further shows that our results are robust to measuring financial leverage using only the portion of

firms’ total debt that is long-term at issuance.

21
Li, Whited, and Wu (2016) show that the passage of antirecharacterization laws (which

increase the value of collateral) leads firms to increase their leverage by 13%; and Heider and

Ljungqvist (2015) document that after a rise in state corporate tax rates firms increase their

leverage by almost 6%. Last, Serfling (2016) (discussed above) reports decreases in leverage

of 3.6%-6.1% following the shock he considers.

We also separately studied the impact of adoptions and rejections of the IDD on leverage

and report the results in Internet Appendix Table A5. We find strong and statistically robust

evidence that firms raise their leverage when state courts adopt the IDD. As expected, the

estimated impact of rejections of the IDD (which reduce the protection of trade secrets) on

firms’ leverage is negative in most specifications, but it is statistically insignificant (these

tests arguably have less statistical power given there are only three rejection events). Hence,

our results are largely driven by the 16 adoptions of the IDD over our sample period.

5.3. Timing of changes in capital structure surrounding adoptions of the IDD

In Table 4, we study the timing of changes in capital structure relative to the timing of

adoptions of the IDD. If reverse causality drives our results, we should observe an increasing

trend in the leverage of firms in affected states prior to the adoption of the IDD. Such evidence

would cast doubt on the validity of our empirical approach, as it would imply a violation of

the parallel trends assumption that the trends in the leverage of treatment firms in adopting

states and control firms in non-adopting states are parallel prior to the adoption of the IDD.

The key variables of interest are IDD Adoption-3, IDD Adoption-2, IDD Adoption-1, IDD

Adoption0, IDD Adoption+1, IDD Adoption+2, and IDD Adoption3+, which are equal to one if

the firm is headquartered in a state that will adopt the IDD in three years, will adopt the

IDD in two years, will adopt the IDD in one year, adopts the IDD in the current year, adopted

the IDD one year ago, adopted the IDD two years ago, or adopted the IDD three or more years

ago, respectively, and zero otherwise. We also control for whether the state where a firm is

headquartered has rejected the previously adopted IDD by year t.

We find that the coefficients on IDD Adoption-3, IDD Adoption-2, IDD Adoption-1, and IDD

Adoption0 are close to zero and statistically insignificant, while the coefficients on IDD
22
Adoption+1, IDD Adoption+2 , and IDD Adoption3+ are positive and significant. In Internet

Appendix Table A6, we further show that a similar pattern emerges if we extend the window

to +/- 4 years around the event. Overall, these results show that firms in adopting states

increase their financial leverage relative to that of the control firms only after the adoption

of the IDD, but not before. Hence, reverse causality or a violation of the parallel trends

assumption do not explain our key result that an increase in the legal protection of a firm’s

trade secrets leads the firm to increase its financial leverage.

5.4. Cross-sectional variation in the impact of the recognition of the IDD on capital structure

We now augment our main specification and interact IDD with variables that affect the

strategic value of unused debt capacity. In addition to regressions with firm and industry-

year fixed effects, we use a specification with state-year fixed effects instead of industry-year

fixed effects. This latter specification ensures that changes in unobservable state-level factors

correlated with changes in state courts’ positions on the IDD and financing decisions do not

spuriously drive our results. These tests of the cross-sectional predictions developed in

Section 2 shed further light on the economic mechanism behind our results and underscore

the strategic role of financial leverage when firms face competitive threats due to the

potential loss of their trade secrets to rivals. The tests also provide further evidence that our

main results are causal, i.e., if a variable omitted from our main regression model were to

drive the results in Table 3, then such a variable would also have to explain the cross-

sectional results we report here.

In Panel A of Table 5, we show that the recognition of the IDD leads to a larger increase

in the debt ratios of firms with a greater ex-ante risk of losing employees who know their

trade secrets to rivals. This risk is greater when a larger fraction of the firm’s employees

know its trade secrets and when the firm faces geographically close rivals who could more

easily poach its workers. Our first proxy for this risk is the fraction of workers in the firm’s

state and industry that are employed in “managerial & related occupations”, which entail

23
access to the firm’s trade secrets (% Workers in Managerial Occupations).16 The second proxy

is the weighted-average distance between the firm’s headquarters and each of its industry

rivals’ headquarters (Average Distance to Rivals), with weights based on each rival’s sales.

We demean both measures before forming the interactions. In five of the eight models in this

panel, we find that the recognition of the IDD leads to a significantly greater increase in net

financial leverage for firms that face a higher ex-ante risk of losing employees who know

trade secrets to rival firms.

In Panel B of Table 5, we provide empirical support for our cross-sectional predictions

regarding a firm’s financial strength relative to its rivals and the specificity of the assets in

its industry. In models 1-4, we follow Valta (2012) and use credit ratings to identify if

financially weak firms face financially strong rivals. Specifically, we focus on the subsample

of unrated firms (with no access to public debt markets) and interact IDD with the average

credit rating of the firm’s rated rivals in the same three-digit SIC industry (Average Rating

of Rivals). In three of the four models, we find that the recognition of the IDD has a

significantly larger effect on financial leverage when a financially weaker firm faces

financially stronger rivals. In models 5-8, we interact IDD with an indicator equal to one if a

firm’s three-digit SIC industry has asset specificity above the sample median and zero

otherwise (High Asset Specificity Ind.). As in Acharya, Bharath, and Srinivasan (2007) and

Valta (2012), we measure asset specificity using the industry median ratio of machinery and

equipment to book assets. Supporting our prediction, in all four models we find that the

recognition of the IDD leads to a significantly larger increase in financial leverage in

industries with higher asset specificity.

In Panel C of Table 5, we document empirical support for our predictions relating to

industry barriers to entry and within-industry product differentiation. In models 1-4, we

measure industry barriers to entry using the three-digit SIC industry-median values of

16The data come from the IPUMS-USA database available for the census years 1980, 1990, and 2000. We use the
data from the 1980, 1990, and 2000 censuses for the periods 1977-1985, 1986-1995, and 1996-2011, respectively.

24
capitalized R&D plus advertising expenditures divided by assets as in Valta (2012).17 This is

motivated by Shaked and Sutton (1987) and Sutton (1991), who argue that firms use R&D

and advertising to differentiate their products and to make the entry of new competitors into

their product markets more difficult, and by Hoberg and Phillips (2016) who further show

that firms spending more on R&D and advertising face less competition. We interact IDD

with an indicator equal to one for firms in industries with capitalized R&D plus advertising

expenditures above the sample median and zero otherwise (High R&D and Advertising Ind.).

In three of the four models, we find that the recognition of the IDD has a significantly smaller

impact on the net debt ratios of firms in industries with high barriers to entry.

Last, in models 5-8 of Panel C we interact IDD with Relative R&D and Advertising,

defined as the ratio of a firm’s capitalized R&D and advertising expenditures to assets minus

the median value of this ratio across all firms in the firm’s three-digit SIC industry (excluding

the firm itself). This is motivated by the research discussed above, which also suggests that

firms which have invested more in differentiating their products from their rivals have a

stronger competitive position within their industry. For all four models, we find that the

recognition of the IDD has a significantly smaller impact on the net debt ratios of firms with

more differentiated products.

5.5. Are there alternative explanations for the study’s results?

5.5.1. The profitability and tax benefits of debt channel

The evidence so far supports our hypothesis that the recognition of the IDD leads a firm

to increase its leverage because it decreases the firm’s ex-ante risk of losing trade secrets to

rivals, thereby reducing the competitive risk faced by the firm and the strategic benefit of

maintaining unused debt capacity. However, an additional mechanism through which the

recognition of the IDD might lead a firm to lever up which follows from static-tradeoff models

of capital structure is by increasing the firm’s expected future profitability and thus the tax

17 We calculate the capitalized value of R&D as in Peters and Taylor (2017), and we compute the capitalized value
of advertising using an analogous approach assuming zero initial value and a depreciation rate of 20%.

25
benefits of debt financing. Also, as discussed earlier, in dynamic models of capital structure

(e.g., DeAngelo et al. (2011)) higher expected future profitability can make a firm conserve

debt capacity to ensure that it can finance good investment projects that may arise. Thus,

under this scenario, higher expected future profitability subsequent to the recognition of the

IDD could bias us against finding that firms increase their leverage subsequent to this

recognition. We conduct several tests to assess whether higher expected profitability

subsequent to the recognition of the IDD may confound our inferences.

In Panel A of Table 6, we augment our baseline specification regressing net leverage on

the IDD indicator to control for future profitability. Specifically, we include not only the

current value, but also one- and two-year leads of the profitability measure, and conduct

separate analyses using two alternative profitability variables: Profit Margin (the sum of pre-

tax income, interest expense, and depreciation and amortization divided by sales) and Return

on Assets (operating income before depreciation scaled by assets). All regressions include firm

and industry-year fixed effects. We find that the estimated impact of the IDD indicator on

both net book and net market leverage remains unaffected after controlling for future

profitability. These findings mitigate some of the concern that the estimated effect of the

recognition of the IDD on financial leverage might somehow be confounded by firms

anticipating increases in future profitability after their state courts recognize the IDD.

Second, we examine if the recognition of the IDD affects a firm’s future profitability.

Here, we measure the dependent variable using both Profit Margin and Return on Assets. In

the models with control variables, the controls include Log Book Assets, Fixed Assets, Cash

Flow Volatility, Strength of CNCs, State GDP Growth, and Political Balance. All regressions

include firm and industry-year fixed effects. The results reported in Internet Appendix Table

A7 show that the coefficient on the IDD indicator is not statistically significant in any of the

specifications. Hence, the recognition of the IDD does not seem to have a material impact on

a firm’s expected future profitability relative to that of its rivals in non-affected states.

Third, if the recognition of the IDD leads firms to lever up because it increases expected

future profitability and the tax benefits of debt financing, then this effect should be larger for

26
firms with higher marginal tax rates and less non-debt tax shields. In Internet Appendix

Table A8, we augment our baseline specification regressing net leverage on the IDD indicator

to include interactions between this indicator and the firm’s marginal tax rate (measured as

in Blouin, Core, and Guay (2010)) or proxies for the existence of non-debt tax shields

(depreciation/assets, tax loss carryforwards/assets, and investment tax credits/assets). We

estimate models without control variables, but the inclusion of control variables does not

affect our results. The coefficients on all interaction variables are statistically insignificant.

5.5.2. The reduction in uncertainty not related to competitive threats

It could be that the recognition of the IDD leads to a reduction in general uncertainty for

a firm that is unrelated to product market threats, and this explains our findings on the

impact of this recognition on firms’ capital structures. Below we investigate this possibility.

First, we note that our main results (reported in Table 3) show that firms lever up after

the recognition of the IDD controlling for historical cash flow volatility, which prior work uses

to account for general uncertainty that affects leverage choices (e.g., Lemmon, Roberts, and

Zender (2008)). In Panel B of Table 6, we augment this specification to control not only for

the current, but also the future values of cash flow volatility up to two years ahead, but this

does not affect our results. A concern with this test is that the cash flow volatility measure,

which is computed using the prior five years of data and thus exhibits limited year-to-year

variation, might not capture discrete changes in risk around the year of IDD rulings

accurately. Hence, in models 2 and 4 of Panel B of Table 6, we replace this variable by its

analogous counterpart based on unlevered stock return volatility, which is computed

annually using daily stock returns and thus is better suited to capture time-series variation

in cash flow risk. We find that controlling for current and future unlevered stock return

volatility also does not affect the estimated effect of the IDD indicator on leverage.

Second, we directly examine whether the recognition of the IDD affects firms’ unlevered

stock return volatility using an approach similar to that in our profitability tests. The results

reported in Internet Appendix Table A9 show that, regardless of whether we include control

27
variables, the recognition of the IDD does not affect a firm’s unlevered stock return volatility.

The two sets of results reported above do not support the notion that, following the

recognition of the IDD in a firm’s state, a reduction in general uncertainty that is unrelated

to product market threats can explain our main capital structure findings. Further, these

results show that standard measures of uncertainty do not capture changes in a firm’s ex-

ante competitive threats due to legal rulings that affect the protection of its trade secrets.

Prior work also highlights that measured cash flow risk and other firm risk proxies do not

capture competitive threats. For example, Hoberg et al. (2014) show that product fluidity

affects financial policies after controlling for cash flow risk. They explain that a firm facing

large competitive threats may take steps to maintain ex-post stable cash flows, and thus for

such a firm measured cash flow risk could be low even if the firm faces significant ex-ante

competitive threats. Similarly, Valta (2012) finds that the competitive threats a firm faces

are priced into its cost of debt after controlling for its ex-ante cash flow volatility and

accounting- and market-based measures of its default risk. He highlights that traditional

proxies for default risk do not capture banks’ assessment of firms’ competitive risk exposure.

5.5.3. Recognition of the IDD and investment in the development of trade secrets

The recognition of the IDD increases a firm’s ability to appropriate the economic rents

generated by its trade secrets and it could increase the marginal benefit of investment in the

creation of trade secrets. In consequence, the recognition of the IDD might increase the firm’s

demand for external financing that is needed to fund these expenditures.18 This raises the

question of whether the observed increase in financial leverage following the recognition of

the IDD could be driven by increased financing needs rather than a decrease in the strategic

benefit of maintaining unused debt capacity as implied by our main hypothesis. Below we

discuss the results of several tests that suggest this is unlikely to be the case.

18 We note, however, that survey evidence shows that most trade secrets lost to rivals are not technical in nature
(e.g., customer lists, strategic plans, or financial data) and it is unclear if the development of such secrets requires
significant expenditures. Further, Png (2017) argues and shows that better trade secret protection can increase
or decrease a firm’s R&D spending, because it can also reduce R&D spillovers from other firms, which often act
as a complement to a firm’s own R&D spending.

28
 In Internet Appendix Table A10, we explore whether the recognition of the IDD affects

spending in activities that are potentially associated with the creation of trade secrets. To

this end, we use our difference-in-differences specification with firm and industry-year fixed

effects, and consider models with and without the controls from our main capital structure

model. We first examine the impact of IDD rulings on a firm’s R&D spending (scaled by sales),

the expense most commonly associated with the development of new ideas, and find no effect.

Because the development of new ideas could also lead to higher capital expenditures, we

consider the impact of IDD rulings on the sum of R&D and capital expenditures, and find no

effect. The development of new ideas might also lead to more advertising, so we examine the

impact of IDD rulings on the sum of R&D, capital, and advertising expenditures, and again

find no effect. Finally, in Internet Appendix Table A11, we augment our baseline capital

structure specification to further include the above proxies for a firm’s spending in activities

associated with the creation of trade secrets. However, the estimated effect of the recognition

of the IDD on net leverage is unaffected by the inclusion of these additional control variables.

5.5.4. Discussion

Given the limitations of our tests reported in Sections 5.5.1., 5.5.2., and 5.5.3., the

evidence seems insufficient to entirely rule out the possibility that the recognition of the IDD

in a firm’s state might also have some impact on the firm’s capital structure by affecting its

future profitability, reducing the general uncertainty it faces, or increasing its incentives to

invest in the development of trade secrets. However, the evidence suggests that these non-

strategic mechanisms are unlikely to be the main driver of the increase in leverage we

observe after the recognition of the IDD. Further, unlike the strategic view behind our

hypothesis, these non-strategic mechanisms cannot easily explain our evidence in Section

5.4, which indicates that the effect of the recognition of the IDD on capital structure depends

on the relative financial strength of the firms in an industry, asset specificity in an industry,

barriers to entry, and product differentiation.

Overall, the results are consistent with the notion that firms optimally adjust their debt

29
ratios after the recognition of the IDD in their state in a manner that is consistent with a

transition between two product market equilibria, as implied by the study’s hypothesis.

Specifically, firms transition from an equilibrium with weak legal protection of their trade

secrets and low leverage (high value of financial flexibility) to another one with stronger legal

protection of their trade secrets and higher leverage (lower value of financial flexibility).19

5.6. Impact of firm-level measures of the risk of losing trade secrets on capital structure

Our prior capital structure tests rely on the recognition of the IDD to gauge changes in

the risk that a firm will lose trade secrets to its rivals. We also conduct tests that instead rely

on two firm-level indicators for whether a firm faces a significant risk of losing trade secrets

to its competitors, as reported by the firm itself in its 10-K filings. A shortcoming of these

tests is that they are based on a shorter sample period for which 10-K filings are available in

EDGAR (1993-2011) and they lack a clear identification strategy.

The first indicator considers if a firm reports a risk of losing trade secrets to rival firms

in its 10-K filing (Trade Secret Risk). It is constructed similar to the approach in Hoberg and

Maksimovic (2015). The second indicator considers if a firm reports a risk of losing “key

employees” in its 10-K filing (Key Employee Risk) and is constructed in a way similar to the

approach in Eisfeldt and Papanikolaou (2013). They argue that if a firm reports such a risk

then it faces a greater risk that its rivals could try to gain competitive advantages over it by

hiring some of its key employees who have inside knowledge of its corporate practices.

Table 7 reports the results of regressions of net leverage on Trade Secret Risk and Key

Employee Risk, with the same firm-level control variables as in our prior tests. We also

include state and industry-year fixed effects. Because there is limited within-firm time-series

variation in these indicators, we do not control for firm fixed effects. We find a negative

association between Trade Secret Risk or Key Employee Risk and net leverage. Firms that

report a risk of losing trade secrets (key employees) hold about 9.4 (4.4) cents less debt (net

of cash holdings) per dollar of book assets than do their rivals. This evidence further supports

19 We thank an anonymous referee for pointing out this interpretation of the results.

30
our hypothesis that a firm maintains a lower debt ratio when it faces a greater risk of losing

trade secrets to rivals.20

5.7. Protection of trade secrets or labor-related mechanisms: Discussion

The recognition of the IDD increases the protection of firms’ trade secrets by reducing

the mobility of workers with access to their trade secrets to rival firms. This raises the

question of whether reduced mobility of these workers to rival firms could have a large impact

on capital structure through pure labor mechanisms that are unrelated to better protection

of trade secrets and drive our results. We are unable to separate the effect on capital structure

due to increased protection of trade secrets from additional effects that might operate

independently through a reduced mobility of workers with access to trade secrets to rival

firms. Below we discuss why pure labor mobility effects that are unrelated to the protection

of trade secrets do not seem likely to be the main driver of our findings.

Most arguments linking IDD rulings to capital structure through pure labor mechanisms

(as in Agrawal and Matsa (2013), Simintzi et al. (2015), or Serfling (2016)) hinge on this event

having a large impact on firms’ total labor costs. However, the IDD only affects the mobility

of a small number of workers who know a firm’s trade secrets and whose wages typically

account for a minor share of the firm’s total wage bill. Another argument is that IDD rulings

could increase a firm’s debt capacity by reducing the risk that it will lose workers who possess

valuable skills. However, reduced labor mobility could instead lower a firm’s debt capacity,

by reducing workers’ effort and investment in human capital (Garmaise (2011)) and by

hampering the firm’s ability to recruit high quality workers who are averse to job lock.

6. Additional investigation

We now briefly discuss the results of additional investigations and robustness tests that

are contained and more extensively discussed in the Internet Appendix to preserve space.

20 In untabulated tests over the period 1997-2011, we do not find a change in the likelihood that a firm mentions
a significant risk of losing trade secrets to its competitors in its 10-K filings after the recognition of the IDD in
its state. Noteworthy, only 3 of the 16 adoption events can be used in these tests.

31
 First, we show that the recognition of the IDD in a firm’s state reduces its cost of bank

debt (see Table A12). This is further support for the proposition that the recognition of the

IDD in a firm’s state reduces the competitive risk that it faces.

Second, we find that affected firms experience significant positive abnormal returns over

the days surrounding the date when a state court adopts the IDD (see Table A13), which is

additional evidence consistent with the notion that market participants believe the adoption

of the IDD decreases the competitive risks these firms face. Supporting the view that the

changes in state courts’ positions regarding the IDD are unlikely to be anticipated events, we

also document that abnormal returns are not significantly different from zero during the

weeks prior to a state court’s decision to adopt the IDD.

Third, we provide evidence that suggests the IDD rulings capture a unique source of

competitive risk. We find no impact of IDD rulings on the Hoberg et al. (2014) measure of

product fluidity (see Table A14). This potentially implies that IDD rulings capture changes

in competitive threats related to the loss of trade secrets to rivals, while product fluidity

captures other competitive threats, e.g., related to improvements in competitors’ products.

We also find no effect of IDD rulings on other measures of competitive risk used in Haushalter

et al. (2007), Frésard (2010), and Valta (2012) (see Table A15).

Fourth, we find that the recognition of the IDD in a state increases the number of trade

secrets cases litigated in that state (see Table A16). This suggests that, in spite of reducing

the ex-ante mobility to rival firms of workers who know their employers’ trade secrets, the

recognition of the IDD increases firms’ propensities to litigate to protect their trade secrets,

possibly because they expect more favorable outcomes after this recognition.

Fifth, because in Bolton and Scharfstein (1990) the optimal debt contract balances the

benefits of deterring predation by reducing a firm’s financial constraints against the cost of

exacerbating managerial agency problems, we examine if the effect of the recognition of the

IDD on a firm’s capital structure is robust to controlling for its corporate governance. To do

so, we proxy for a firm’s governance with an indicator for if it is incorporated in a state that

has passed business combination laws or with the concentration of the ownership of its equity

32
shares by institutional investors. In Table A17, we show that the effect of the recognition of

the IDD on a firm’s capital structure is unaffected by the inclusion of these variables.

Sixth, in Table A18 we report findings which suggest measurement error in the

protection of trade secrets afforded by the IDD indicator variable arising from firm

headquarter relocations to another state or from geographical dispersion of a firm’s work

force does not drive our results.

Seventh, in Table A19 we show that our main results are robust to using propensity-

score matching to ensure that the treatment and control firms have similar observable

characteristics before the onset of the treatment.

Finally, in Table A20 we show that firms significantly increase their outstanding debt

during years when they litigate to protect their trade secrets. This evidence supports our

argument that having unused debt capacity is strategically useful for firms that face a

greater risk of losing their trade secrets to rivals, because it allows them to rapidly raise the

funds they need to protect their competitive position when such events materialize.

7. Conclusion

Our message is that the risk a firm’s rivals could gain access to its trade secrets induces

the firm to choose its capital structure strategically. In the spirit of Telser (1966), Poitevin

(1989), and Bolton and Scharfstein (1990), we hypothesize that firms facing a greater risk

that their rivals will try to gain access to their trade secrets and hurt their competitive

position benefit more from strategically maintaining unused debt capacity, either to be able

to respond to such competitive attacks or to deter them ex ante, and thus hold less debt. We

test our hypothesis using a difference-in-differences research design that exploits the

staggered recognition of the Inevitable Disclosure Doctrine (IDD) by state courts over the

1977-2011 period. The recognition of this doctrine causes an arguably exogenous decrease in

the risk that a firm’s rivals might gain access to its trade secrets because it increases the

firm’s ability to prevent its workers who know its trade secrets from working for rivals.

We first show that the recognition of the IDD significantly reduces the mobility of

workers who know trade secrets to rival firms and thus a firm’s risk of losing trade secrets to
33
rivals. Supporting our main hypothesis, we then show that firms significantly decrease their

net leverage after the recognition of the IDD by courts in their state of headquarters. This

effect is stronger for firms with a higher ex-ante risk of losing trade secrets to rivals through

the labor mobility channel. Importantly, highlighting the strategic dimension of a firm’s

capital structure choices when it faces competitive risk associated with its inability to fully

protect its trade secrets and competitive advantages, we find that the effect is stronger for

financially weak firms that face financially stronger rivals, for firms in industries with more

specific assets or lower barriers to entry, and for firms with less differentiated products.

We also find a negative association between firms’ debt ratios and firm-level measures

of the risk of losing trade secrets to rivals, as self-reported by firms in their 10-K filings. The

results of additional tests show that affected firms earn positive abnormal returns over the

days surrounding a state court’s final decision to recognize the IDD and that after this

recognition their cost of bank debt decreases. These results support the view that the

recognition of the IDD reduces the competitive threats these firms face. Importantly, we also

provide several pieces of evidence that are inconsistent with alternative (non-strategic)

interpretations of the paper’s results.

Our paper calls to attention the strategic value of financial flexibility for firms that face

competitive threats associated with their inability to fully protect their intellectual property,

and suggests that further research on the topic might be fruitful. For some firms, significant

threats of this kind can also arise through mechanisms other than the mobility of key

employees, such as industrial espionage and outright theft (even by foreign agents), an issue

that has become increasingly important in recent years (e.g., with this in mind the U.S.

Congress passed the Economic Espionage Act in 1996). Similar threats could also naturally

arise when firms find it difficult to enforce their property rights on other forms of intellectual

capital besides trade secrets, such as patents, copyrights, or trademarks. An important

challenge in pursuing this broader research agenda is to gauge both the relative importance

of different kinds of intellectual property for any given firm and the associated competitive

risk from losing this property to rival firms.

34
References

Acharya, V., Almeida, H., Campello, M., 2007. Is cash negative debt? A hedging perspective
on corporate financial policies. Journal of Financial Intermediation 16, 515–554.

Acharya, V., Bharath, S., Srinivasan, A., 2007. Does industry-wide distress affect defaulted
firms? Evidence from creditor recoveries. Journal of Financial Economics 85, 787–821.

Agrawal, A.K., Matsa, D.A., 2013. Labor unemployment risk and corporate financing
decisions. Journal of Financial Economics 108, 449–470.

Almeling, D.S., 2012. Seven reasons why trade secrets are increasingly important. Berkeley
Technology Law Journal 27, 1090-1118.

Almeling, D.S., Snyder, D.W., Sapoznikow, M., McCollum, W.E., Weader, J., 2010. A
statistical analysis of trade secret litigation in federal courts. Gonzaga Law Review 45,
291–334.

Almeling, D.S., Snyder, D.W., Sapoznikow, M., McCollum, W.E., Weader, J. 2011. A
statistical analysis of trade secret litigation in state courts. Gonzaga Law Review 46, 57–
101.

Arundel, A., 2001. The relative effectiveness of patents and secrecy for appropriation.
Research Policy 30, 611–624.

Bird, R.C., Knopf, J.D., 2015. The impact of local knowledge on banking. Journal of Financial
Services Research 48, 1-20.

Blouin, J., Core, J.E., Guay, W., 2010. Have the tax benefits of debt been overestimated?
Journal of Financial Economics 98, 195–213.

Bolton, P., Scharfstein, D.S., 1990. A theory of predation based on agency problems in
financial contracting. American Economic Review 80, 93–106.

Campello, M., 2003. Capital structure and product markets interactions: Evidence from
business cycles. Journal of Financial Economics 68, 353–378.

Campello, M., 2006. Debt financing: Does it boost or hurt firm performance in product
markets? Journal of Financial Economics 82, 135–172.

Chevalier, J., 1995. Do LBO supermarkets charge more? An empirical analysis of the effects
of LBOs on supermarket pricing. Journal of Finance 50, 1095–1112.

Cohen, W.M., Nelson, R.R., Walsh, J.P., 2000. Protecting their intellectual assets:
Appropriability conditions and why U.S. manufacturing firms patent (or not).
Unpublished working paper. NBER.

DeAngelo, H., DeAngelo, L., 2007. Capital structure, payout policy, and financial flexibility.
Unpublished working paper. University of Southern California.

DeAngelo, H., DeAngelo, L., Whited, T.M., 2011. Capital structure dynamics and transitory
debt. Journal of Financial Economics 99, 235–261.

35
Eisfeldt, A., Papanikolaou, D., 2013. Organizational capital and the cross-section of expected
returns. Journal of Finance 68, 1365–1406.

Faure-Grimaud, A., 2000. Product market competition and optimal debt contracts: The
limited liability effect revisited. European Economic Review 44, 1823–1840.

Frésard, L., 2010. Financial strength and product market behavior: The real effects of
corporate cash holdings. Journal of Finance 65, 1097–1122.

Gamba, A., Triantis, A., 2008. The value of financial flexibility. Journal of Finance 63, 2263–
2296.

Garmaise, M.J., 2011. Ties that truly bind: Noncompetition agreements, executive
compensation, and firm investment. Journal of Law, Economics, and Organization 27,
376–425.

Godfrey, E.R., 2004. Inevitable disclosure of trade secrets: Employee mobility v. employer's
rights. Journal of High Technology Law 3, 161–179.

Hall, B.H., Helmers, C., Rogers, M., Sena, V., 2012. The choice between formal and informal
intellectual property: A literature review. Unpublished working paper. NBER.

Harris, J.O., 2000. The doctrine of inevitable disclosure: A proposal to balance employer and
employee interests. Washington University Law Review 78, 325–345.

Haushalter, D., Klasa, S., Maxwell, W.F., 2007. The influence of product market dynamics
on a firm’s cash holdings and hedging behavior. Journal of Financial Economics 84, 797–
825.

Heider, F., Ljungqvist, A., 2015. As certain as debt and taxes: Estimating the tax sensitivity
of leverage from state tax changes. Journal of Financial Economics 118, 684–712.

Hoberg, G., Maksimovic, V., 2015. Redefining financial constraints: A text–based analysis.
Review of Financial Studies 28, 1312–1352.

Hoberg, G., Phillips, G., 2016. Text-based network industries and endogenous product
differentiation. Journal of Political Economy 124, 1423–1465.

Hoberg, G., Phillips, G., Prabhala, N., 2014. Product market threats, payouts, and financial
flexibility. Journal of Finance 69, 293–324.

Jensen, M.C., 1986. Agency costs of free cash flow, corporate finance, and takeovers.
American Economic Review 76, 323–329.

Kahnke, R.E., Bundy, K.L., Liebman, K.A., 2008. Doctrine of inevitable disclosure. Faegre
and Benson LLP.

Khanna, N., Tice, S., 2000. Strategic responses of incumbents to new entry: The effect of
ownership structure, capital structure, and focus. Review of Financial Studies 13, 749–
779.

Khanna, N., Tice, S., 2005. Pricing, exit, and location decisions of firms: Evidence on the role
of debt and operating efficiency. Journal of Financial Economics 75, 397–427.
36
Kim, J-H, 2014. Employee poaching: Why it can be predatory. Managerial and Decision
Economics 35, 309–317.

Kovenock, D., Phillips, G., 1997. Capital structure and product market behavior. Review of
Financial Studies 10, 767–803.

Lemmon, M.L., Roberts, M.R., Zender, J.F., 2008. Back to the beginning: Persistence and the
cross-section of corporate capital structure, Journal of Finance 63, 1575–1608.

Lerner, J., 2006. Using litigation to understand trade secrets: A preliminary exploration.
Unpublished working paper. Harvard Business School.

Li, S., Whited, T.M., Wu, Y., 2016. Collateral, leverage, and taxes. Review of Financial
Studies 29, 1453–1500.

Lyandres, E., 2006. Capital structure and interaction among firms in output markets: Theory
and evidence. Journal of Business 79, 2381–2421.

MacKay, P., Phillips, G., 2005. How does industry affect firm financial structure? Review of
Financial Studies 18, 1433–1466.

Malsberger, B.M., 2004. Covenants Not to Compete: A State-by-State Survey. BNA Books,
Arlington, Virginia.

Malsberger, B.M., 2011. Trade Secrets: A State-by-State Survey, 4th ed. BNA Books,
Arlington, Virginia.

Marx, M., Strumsky, D., Fleming, L. 2009. Mobility, skills, and the Michigan noncompete
experiment. Management Science 55, 875-889.

Matsa, D.A., 2010. Capital structure as a strategic variable: Evidence from collective
bargaining. Journal of Finance 65, 1197–1232.

Maurer, B., 1999. Innovation and investment under financial constraints and product market
competition. International Journal of Industrial Organization 17, 455–476.

Ortiz-Molina, H., Phillips, G., 2014. Real asset illiquidity and the cost of capital. Journal of
Financial and Quantitative Analysis 49, 1–32.

Peters, R.H., Taylor, L., 2017. Intangible capital and the investment-q relation. Journal of
Financial Economics 123, 251–272.

Phillips, G., 1995. Increased debt and industry product markets: An empirical analysis.
Journal of Financial Economics 37, 189–238.

Png, I.P.L., 2017. Law and innovation: Evidence from state trade secrets laws. Review of
Economics and Statistics 99, 167–179.

Poitevin, M., 1989. Financial signalling and the “deep–pocket” argument. RAND Journal of
Economics 20, 26–40.

Porter, M.E., 2008. The five competitive forces that shape strategy. Harvard Business Review
86, 78–93.

37
Samila, S., Sorenson, O., 2011. Noncompete covenants: Incentives to innovate or
impediments to growth. Management Science 57, 425-438.

Serfling, M., 2016. Firing costs and capital structure decisions. Journal of Finance 71, 2239–
2285.

Shaked, A., Sutton, J., 1987. Product differentiation and industrial structure. Journal of
Industrial Economics 26, 131–146.

Shapiro, R.J., Hassett, K.A., 2005. The Economic Value of Intellectual Property. USA for
Innovation, Washington D.C.

Simintzi, E., Vig, V., Volpin, P., 2015. Labor protection and leverage. Review of Financial
Studies 28, 561–591.

Sutton, J., 1991. Sunk Costs and Market Structure. MIT Press, Cambridge, Massachusetts.

Telser, L.G., 1966. Cutthroat competition and the long purse. Journal of Law and Economics
9, 259–277.

Valta, P., 2012. Competition and the cost of debt. Journal of Financial Economics 105, 661–
682.

Waldref, V., 2012. Inevitable disclosure doctrine & protection from the threat of trade secret
misappropriation: State law overview & analysis. Lee & Hayes, PLLC.

Wiesner, R.M., 2012. A state-by-state analysis of inevitable disclosure: A need for uniformity
and a workable standard. Marquette Intellectual Property Law Review 16, 211–231.

Xu, J., 2012. Profitability and capital structure: Evidence from import penetration. Journal
of Financial Economics 126, 427–446.

Zingales, L., 1998. Survival of the fittest or the fattest? Exit and financing in the trucking
industry. Journal of Finance 53, 905–938.

38
Table 1
Precedent-setting legal cases adopting or rejecting the Inevitable Disclosure Doctrine.
The table lists the precedent-setting legal cases in which state courts adopted the Inevitable Disclosure Doctrine (IDD) or rejected it after adopting it.
The states omitted from the table did not consider or considered but rejected the IDD. The text of all court decisions is available from Google Scholar.

State Precedent-Setting Case(s) Date Decision

AR Southwestern Energy Co. v. Eickenhorst, 955 F. Supp. 1078 (W.D. Ark. 1997) 3/18/1997 Adopt
CT Branson Ultrasonics Corp. v. Stratman, 921 F. Supp. 909 (D. Conn. 1996) 2/28/1996 Adopt
DE E.I. duPont de Nemours & Co. v. American Potash & Chem. Corp., 200 A.2d 428 (Del. Ch. 1964) 5/5/1964 Adopt
FL Fountain v. Hudson Cush-N-Foam Corp., 122 So. 2d 232 (Fla. Dist. Ct. App. 1960) 7/11/1960 Adopt
Del Monte Fresh Produce Co. v. Dole Food Co. Inc., 148 F. Supp. 2d 1326 (S.D. Fla. 2001) 5/21/2001 Reject
GA Essex Group Inc. v. Southwire Co., 501 S.E.2d 501 (Ga. 1998) 6/29/1998 Adopt
IL Teradyne Inc. v. Clear Communications Corp., 707 F. Supp. 353 (N.D. 111. 1989) 2/9/1989 Adopt
IN Ackerman v. Kimball Int’l Inc., 652 N.E.2d 507 (Ind. 1995) 7/12/1995 Adopt
IA Uncle B’s Bakery v. O’Rourke, 920 F. Supp. 1405 (N.D. Iowa 1996) 4/1/1996 Adopt
KS Bradbury Co. v. Teissier-duCros, 413 F. Supp. 2d 1203 (D. Kan. 2006) 2/2/2006 Adopt
MA Bard v. Intoccia, 1994 U.S. Dist. LEXIS 15368 (D. Mass. 1994) 10/13/1994 Adopt
MI Allis-Chalmers Manuf. Co. v. Continental Aviation & Eng. Corp., 255 F. Supp. 645 (E.D. Mich. 1966) 2/17/1966 Adopt
CMI Int’l, Inc. v. Intermet Int’l Corp., 649 N.W.2d 808 (Mich. Ct. App. 2002) 4/30/2002 Reject
MN Surgidev Corp. v. Eye Technology Inc., 648 F. Supp. 661 (D. Minn. 1986) 10/10/1986 Adopt
MO H&R Block Eastern Tax Servs. Inc. v. Enchura, 122 F. Supp. 2d 1067 (W.D. Mo. 2000) 11/2/2000 Adopt
NJ Nat’l Starch & Chem. Corp. v. Parker Chem. Corp., 530 A.2d 31 (N.J. Super. Ct. 1987) 4/27/1987 Adopt
NY Eastman Kodak Co. v. Powers Film Prod., 189 A.D. 556 (N.Y.A.D. 1919) 12/5/1919 Adopt
NC Travenol Laboratories Inc. v. Turner, 228 S.E.2d 478 (N.C. Ct. App. 1976) 6/17/1976 Adopt
OH Procter & Gamble Co. v. Stoneham, 747 N.E.2d 268 (Ohio Ct. App. 2000) 9/29/2000 Adopt
PA Air Products & Chemical Inc. v. Johnson, 442 A.2d 1114 (Pa. Super. Ct. 1982) 2/19/1982 Adopt
TX Rugen v. Interactive Business Systems Inc., 864 S.W.2d 548 (Tex. App. 1993) 5/28/1993 Adopt
Cardinal Health Staffing Network Inc. v. Bowen, 106 S.W.3d 230 (Tex. App. 2003) 4/3/2003 Reject
UT Novell Inc. v. Timpanogos Research Group Inc., 46 U.S.P.Q.2d 1197 (Utah D.C. 1998) 1/30/1998 Adopt
WA Solutec Corp. Inc. v. Agnew, 88 Wash. App. 1067 (Wash. Ct. App. 1997) 12/30/1997 Adopt

39
Table 2
Recognition of the IDD and labor mobility of workers with knowledge of trade secrets.
This table reports results from linear probability models that estimate the impact of the recognition of the IDD on the mobility of individuals in
managerial and related occupations (with access to trade secrets) relative to that of individuals in other occupations (who are less likely to have access to
trade secrets). The sample comes from the Census’ Survey of Income and Program Participation and spans the period 1983-2011. It includes individuals
who are 18+ years old who are observed at least five times during the sample period and are not employed in the financial or utility industries. Managerial
and related occupations correspond to the Integrated Public Use Microdata Series occupation codes 3-37 and the set of other occupations includes all other
occupation codes except for scientific and engineering occupations codes (workers in science and engineering occupations may know some technical trade
secrets). The binary indicators that identify job switches take a value of one if the individual has left her employer recorded in the preceding survey period
to join a new employer in the current survey period (survey periods are four months apart in the vast majority of cases, but the time between surveys
varies between one and twenty-four months), and zero otherwise. Employers are rivals if they operate in the same three-digit Census Industry
Classification industry. In models 1 and 2, the sample includes workers who move to rival employers located in any state (in the same state or in another
state) and those who remain with their current employers. In models 3 and 4, the sample includes workers who move to rival employers located in another
state and those who remain with their current employers. In models 5 and 6, the sample includes workers who move to rival employers located in the same
state and those who remain with their current employers. In models 7 and 8, the sample includes workers who move to non-rival employers and those who
remain with their current employers. The key independent variables are as follows: IDD is equal to one if the individual is employed in a state that
recognizes the IDD, and zero otherwise; and Mgr&Rel, is equal to one if the individual is employed in a “management and related occupation”, and zero
otherwise. The control variables are: Log(Income), the natural logarithm of the individual’s average monthly income recorded in the previous survey (in
$); Log(Hours), the natural logarithm of the average number of weekly hours the individual works recorded in the previous survey; InSchool, an indicator
equal to one if the individual attended school during the prior four months, and zero otherwise; Strength of CNCs, an index of the strength of the
enforcement of covenants not to compete (CNCs) by courts in a state from Bird and Knopf (2015) (higher values imply stronger enforcement); State GDP
Growth, the annual GDP growth rate in the state; Political Balance, the fraction of a state’s congress members representing their state in the U.S. House
of Representatives that belong to the Democratic Party. We also include twenty-four indicator variables for whether the current and the prior survey
records for an individual are 1,2,...,24 months apart. Dollar values are expressed in 2009 dollars. Each model includes individual fixed effects and year-
month fixed effects. Standard errors are corrected for heteroskedasticity and clustering at the state level (t-statistics are in parentheses). *, **, and ***
denote significance at the 10%, 5%, and 1% levels, respectively.

40
Table 2
Continued.
Mobility to Rival Firms Mobility to Non-Rival Firms
In Any State In Other States In Same State In Any State
(1) (2) (3) (4) (5) (6) (7) (8)
IDD -0.003 -0.003 -0.002* -0.002* -0.002 -0.001 -0.003 -0.003
(1.06) (1.02) (1.71) (1.76) (0.61) (0.56) (1.17) (1.10)
Mgr&Rel -0.004 -0.003 0.002** 0.001** -0.005* -0.005* -0.026*** -0.024***
(1.42) (1.29) (2.44) (2.40) (1.94) (1.80) (5.94) (5.57)
IDD × Mgr&Rel -0.007** -0.007** -0.003*** -0.003*** -0.005* -0.005* -0.002 -0.002
(2.30) (2.28) (2.78) (2.76) (1.71) (1.69) (0.39) (0.35)
Log(Income) -0.005*** 0.000 -0.005*** -0.017***
(6.46) (0.99) (6.76) (14.13)
Log(Hours) -0.004*** -0.000 -0.004*** -0.024***
(3.96) (0.53) (3.92) (14.16)
InSchool -0.013*** -0.001*** -0.012*** -0.015***
(5.88) (2.91) (5.52) (6.19)
Strength of CNCs -0.000 -0.000 0.000 0.001
(0.07) (0.34) (0.02) (0.34)
State GDP Growth 0.011 0.008** 0.006 0.019
(0.55) (2.13) (0.33) (0.99)
Political Balance -0.007* -0.001 -0.005 -0.007
(1.68) (0.86) (1.29) (1.38)

Year × Month Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes
Individual Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes
Observations 799,533 799,533 771,300 771,300 799,019 799,019 820,866 820,866
Adjusted R2 0.241 0.241 0.067 0.067 0.239 0.239 0.191 0.193

41
Table 3
Recognition of the IDD and capital structure.
This table reports results from OLS regressions of financial leverage on the indicator for the recognition of the IDD. The sample spans the 1977-2011
period. The dependent variables are Net Book Leverage (models 1 and 2), Net Market Leverage (models 3 and 4), Book Leverage (models 5 and 6), and
Market Leverage (models 7 and 8). Net Book Leverage is the book value of long-term debt (dltt) plus debt in current liabilities (dlc) minus cash holdings
(che) divided by book value of assets (at). Net Market Leverage is the book value of long-term debt (dltt) plus debt in current liabilities (dlc) minus cash
holdings (che) divided by market value of assets (prcc_f*csho + at - ceq). Book Leverage and Market Leverage are analogously defined, except that cash
holdings are not subtracted in the numerator. IDD is an indicator variable equal to one if the firm is headquartered in a state whose courts recognize the
IDD, and zero otherwise. Log Book Assets is the natural logarithm of total assets (at). Market-to-Book Assets is the market value of assets (prcc_f*csho +
at - ceq) divided by the book value of assets (at). Return on Assets is operating income before depreciation (oibdp) divided by the book value of assets (at).
Fixed Assets is the book value of property, plant, and equipment (ppent) divided by the book value of assets (at). Cash Flow Volatility is the standard
deviation of a firm’s Return on Assets over the previous five years (firms are required to have at least three years of data during the prior five years to
enter the sample). Dividend Payer is an indicator variable equal to one if a firm pays common dividends (dvc), and zero otherwise. Strength of CNCs, State
GDP Growth, and Political Balance are defined in Table 2. Dollar values are expressed in 2009 dollars. Industry fixed effects are defined at the three-digit
SIC level. Continuous variables, except state-level variables, are winsorized at their 1st and 99th percentiles. Standard errors are corrected for
heteroskedasticity and clustering at the state level (t-statistics are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1% levels,
respectively.

Net Book Leverage Net Market Leverage Book Leverage Market Leverage
(1) (2) (3) (4) (5) (6) (7) (8)
IDD 0.011** 0.013** 0.010*** 0.012*** 0.011*** 0.012*** 0.007** 0.010***
(2.38) (2.54) (2.73) (3.09) (3.07) (3.17) (2.34) (3.02)
Log Book Assets 0.039*** 0.047*** 0.031*** 0.036***
(8.87) (10.84) (9.03) (10.48)
Market-to-Book Assets -0.012*** 0.009*** -0.003*** -0.017***
(-13.95) (2.77) (-3.92) (-7.93)
Return on Assets -0.181*** -0.099*** -0.160*** -0.126***
(-16.13) (-6.25) (-14.23) (-8.04)
Fixed Assets 0.672*** 0.432*** 0.236*** 0.176***
(16.43) (26.78) (15.79) (15.02)
Cash Flow Volatility 0.001 0.003 0.057*** 0.013
(0.04) (0.26) (4.83) (1.59)
Dividend Payer -0.057*** -0.049*** -0.047*** -0.043***
(-13.04) (-12.68) (-12.39) (-13.94)
Strength of CNCs 0.002 0.002 0.001 0.001
(0.70) (0.74) (0.25) (0.37)
State GDP Growth -0.021 -0.089** 0.006 -0.086***
(-0.49) (-2.55) (0.24) (-3.33)
Political Balance -0.012 -0.017* -0.007 -0.012**
(-0.88) (-1.87) (-0.80) (-2.11)
Industry × Year Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes
Observations 125,895 125,895 125,895 125,895 125,895 125,895 125,895 125,895
Adjusted R2 0.701 0.739 0.674 0.707 0.625 0.650 0.665 0.708
42
Table 4
The timing of changes in firms’ capital structures around adoptions of the IDD.
This table reports results from OLS regressions of Net Book Leverage and Net Market Leverage on
indicators for the timing of state courts’ adoptions of the IDD. The sample spans the 1977-2011 period. IDD
Adoption-3, IDD Adoption-2, IDD Adoption-1, IDD Adoption0, IDD Adoption+1, IDD Adoption+2, and IDD
Adoption3+ are equal to one if the firm is headquartered in a state that will adopt the IDD in three years, adopts
the IDD in two years, adopts the IDD in one year, adopts the IDD in the current year, adopted the IDD one
year ago, adopted the IDD two years ago, or adopted the IDD three or more years ago, respectively, and zero
otherwise. IDD Rejection is an indicator variable set to one beginning the year when the state where the firm
is headquartered rejects the previously adopted IDD, and zero otherwise. Control variables are defined in
Tables 2 and 3 and include Log Book Assets, Market-to-Book Assets, Return on Assets, Fixed Assets, Cash Flow
Volatility, Dividend Payer, Strength of CNCs, State GDP Growth, and Political Balance. Industry fixed effects
are defined at the three-digit SIC level. Continuous variables, except state-level variables, are winsorized at
their 1st and 99th percentiles. Standard errors are corrected for heteroskedasticity and clustering at the state
level (t-statistics are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1% levels,
respectively.

Net Book Leverage Net Market Leverage

(1) (2) (3) (4)
IDD Adoption -3 -0.008 -0.005 -0.006 -0.004
(-1.37) (-0.85) (-1.34) (-0.97)
IDD Adoption -2 0.000 0.004 -0.002 0.000
(0.00) (0.76) (-0.42) (0.06)
IDD Adoption -1 0.004 0.008 -0.000 0.002
(0.52) (1.12) (-0.00) (0.38)
IDD Adoption 0 0.006 0.009 0.004 0.006
(0.88) (1.32) (0.68) (1.01)
IDD Adoption +1 0.018*** 0.019*** 0.012** 0.013**
(2.74) (2.76) (2.14) (2.33)
IDD Adoption +2 0.019** 0.020** 0.016*** 0.016***
(2.36) (2.48) (2.94) (2.99)
IDD Adoption 3+ 0.020* 0.020** 0.014* 0.015**
(1.95) (2.00) (1.92) (2.10)
IDD Rejection 0.005 -0.003 -0.002 -0.008
(0.58) (-0.31) (-0.22) (-1.06)

Control Variables No Yes No Yes

Industry × Year Fixed Effects Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes
Observations 125,895 125,895 125,895 125,895
Adjusted R2 0.702 0.739 0.674 0.707

43
Table 5
Cross-sectional variation in the effect of the IDD.
This table reports results from OLS regressions of Net Book Leverage and Net Market Leverage on the indicator for the recognition of the IDD. The
sample spans the 1977-2011 period in all models except models 1-4 of Panel B, which is from 1985-2011. IDD is an indicator variable equal to one if the
firm is headquartered in a state whose courts recognize the IDD, and zero otherwise. Panel A examines the cross-sectional effect of the ex-ante risk of
losing trade secrets to rivals. % Workers in Managerial Occupations is the fraction of workers employed in a “management and related occupation” (as
defined by the IPUMS occupation codes 3-37) in the firm’s three-digit NAICS industry and state. Average Distance to Rivals is the sales-based weighted
average of the number of miles between a firm’s headquarters and each of its rivals’ headquarters, where rivals are defined by three-digit SIC codes. To
compute the distance between the firm and each of its three-digit SIC industry rivals, we first use Compustat to extract each firm’s zip code from its
headquarters location to identify its longitude and latitude, and then use the “geodist” command in SAS. Panel B examines the effect of the financial
strength of rivals and asset specificity. In models 1-4, the sample includes only unrated firms (firms without a long-term S&P credit rating) and spans the
years 1985-2011 (credit rating coverage is incomplete prior to 1985). Credit ratings are assigned a numerical value from 1 to 10, with higher values
indicating a stronger credit rating (AAA = 10, AA- to AA+ = 9, A- to A+ = 8, BBB- to BBB+ = 7, BB- to BB+ = 6, B- to B+ = 5, CCC- to CCC+ = 4, CC- to
CC+ = 3, C- to C+ = 2, and D or SD = 1). Asset Specificity Ind. is the median ratio of machinery and equipment to total assets in a three-digit SIC industry
across all years in our sample. High Asset Specificity Ind. is an indicator variable equal to one if Asset Specificity Ind. is above the sample median, and
zero otherwise. Panel C examines the effect of industry barriers to entry and within-industry product differentiation. R&D and Advertising Ind. is the
median ratio of capitalized R&D and advertising expenses to total assets for all companies in a firm’s three-digit SIC industry. High R&D and Advertising
Ind. is an indicator variable equal to one if R&D and Advertising Ind. is above the sample median, and zero otherwise. Relative R&D and Advertising is a
firm’s ratio of capitalized R&D and advertising expenses to total assets minus the median of this ratio across all of its rival companies in its three-digit
SIC industry. % Workers in Managerial Occupations, Average Distance to Rivals, Average Rating of Rivals, and Relative R&D and Advertising are
demeaned to ease the interpretation of coefficient estimates on the interaction term. Industry fixed effects are defined at the three-digit SIC level. Standard
errors are corrected for heteroskedasticity and clustering at the state level (t-statistics are in parentheses). *, **, and *** denote significance at the 10%,
5%, and 1% levels, respectively.

44
Table 5
Continued.

Panel A: The effect of the ex-ante risk of losing trade secrets to rivals
Net Book Leverage Net Market Leverage Net Book Leverage Net Market Leverage
(1) (2) (3) (4) (5) (6) (7) (8)
IDD 0.012** 0.011*** 0.009** 0.009***
(2.53) (2.82) (1.96) (2.61)
IDD × % Workers in Managerial Occupations 0.221** 0.162* 0.173*** 0.100
(2.51) (1.68) (2.88) (1.33)
% Workers in Managerial Occupations 0.067 -0.169*** 0.045 -0.080
(0.56) (-2.61) (0.53) (-1.56)
IDD × Average Distance to Rivals -0.015 -0.027*** -0.008 -0.016**
(-1.63) (-3.41) (-1.18) (-2.43)
Average Distance to Rivals 0.006 0.009 0.005 0.003
(0.79) (1.24) (0.99) (0.46)

Industry × Year Fixed Effects Yes No Yes No Yes No Yes No

State × Year Fixed Effects No Yes No Yes No Yes No Yes
Firm Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes
Observations 122,944 122,944 122,944 122,944 120,949 120,949 120,949 120,949
Adjusted R2 0.701 0.693 0.673 0.658 0.702 0.695 0.671 0.657

45
Table 5
Continued.

Panel B: The effect of the financial strength of rivals and asset specificity
Net Book Leverage Net Market Leverage Net Book Leverage Net Market Leverage
(1) (2) (3) (4) (5) (6) (7) (8)
IDD 0.009 0.011** 0.001 0.005
(1.50) (2.13) (0.13) (0.81)
IDD × Average Rating of Rivals 0.011*** 0.009** 0.005* 0.004
(2.61) (2.02) (1.66) (1.06)
Average Rating of Rivals -0.001 0.000
(-0.22) (0.13)
IDD × High Asset Specificity Ind. 0.021** 0.021*** 0.012* 0.012***
(2.36) (2.65) (1.95) (2.67)
High Asset Specificity Ind. -0.008 0.002
(-0.88) (0.32)

Industry × Year Fixed Effects Yes No Yes No Yes No Yes No

State × Year Fixed Effects No Yes No Yes No Yes No Yes
Firm Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes
Observations 73,417 73,417 73,417 73,417 124,641 124,641 124,641 124,641
Adjusted R2 0.699 0.694 0.667 0.658 0.702 0.694 0.674 0.659

46
Table 5
Continued.

Panel C: The effect of industry barriers to entry and within-industry product differentiation
Net Book Leverage Net Market Leverage Net Book Leverage Net Market Leverage
(1) (2) (3) (4) (5) (6) (7) (8)
IDD 0.021*** 0.016*** 0.011** 0.010***
(2.99) (2.72) (2.23) (2.62)
IDD × High R&D and Advertising Ind. -0.019** -0.017* -0.011* -0.011
(-2.50) (-1.65) (-1.70) (-1.36)
High R&D and Advertising Ind. 0.004 -0.001
(0.69) (-0.13)
IDD × Relative R&D and Advertising -0.025** -0.028*** -0.019*** -0.019***
(-2.54) (-2.64) (-3.24) (-2.81)
Relative R&D and Advertising 0.085*** 0.079*** 0.025*** 0.023***
(12.03) (9.42) (5.88) (3.96)

Industry × Year Fixed Effects Yes No Yes No Yes No Yes No

State × Year Fixed Effects No Yes No Yes No Yes No Yes
Firm Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes
Observations 125,895 125,895 125,895 125,895 123,688 123,688 123,688 123,688
Adjusted R2 0.700 0.691 0.672 0.657 0.703 0.695 0.673 0.659

47
Table 6
Recognition of the IDD and leverage: Controlling for profitability and general uncertainty.
This table reports results from OLS regressions of Net Book Leverage and Net Market Leverage on the
indicator for the recognition of the IDD. The sample spans the 1977-2011 period. IDD is an indicator variable
equal to one if the firm is headquartered in a state whose courts recognize the IDD, and zero otherwise. In Panel
A, the set of control variables used in the benchmark regression in Table 3 is expanded to include in addition to
current profitability, the one-year ahead, and two-years ahead values of profitability measured alternatively as
Profit Margin (the sum of pre-tax income, interest expense, and depreciation and amortization (pi+xint+dp), all
divided by sales (sale)) or Return on Assets (defined in Table 3). In Panel B, the set of control variables used in
the benchmark regression in Table 3 is expanded to include in addition to the current value for general
uncertainty, the one-year, and two-years ahead values measured alternatively as Cash Flow Volatility (defined
in Table 3) or Unlevered Return Volatility (the annualized standard deviation of a firm’s daily stock returns over
a fiscal year (we require the firm to have at least 180 observations during the fiscal year) multiplied by the ratio
of the firm’s market value of equity to its market value of assets ((csho*prcc_f)/(at-ceq+csho*prcc_f))). All other
control variables (Log Book Assets, Market-to-Book Assets, Fixed Assets, Cash Flow Volatility, Dividend Payer,
Strength of CNCs, State GDP Growth, and Political Balance) are defined in Tables 2 and 3. Industry fixed effects
are defined at the three-digit SIC level. Continuous variables, except state-level variables, are winsorized at their
1st and 99th percentiles. Standard errors are corrected for heteroskedasticity and clustering at the state level (t-
statistics are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

Panel A: Controlling for future profitability

Net Book Leverage Net Market Leverage
(1) (2) (3) (4)
IDD 0.013** 0.013** 0.012*** 0.012***
(2.50) (2.37) (2.86) (2.82)
Profit Margint 0.003*** 0.001
(3.29) (1.21)
Profit Margint+1 0.006*** 0.003***
(4.35) (4.87)
Profit Margint+2 0.005*** 0.002***
(5.01) (4.51)
Return on Assetst -0.212*** -0.118***
(-14.43) (-6.55)
Return on Assetst+1 0.076*** 0.045***
(10.43) (12.99)
Return on Assetst+2 0.054*** 0.035***
(8.25) (6.06)

All other Control Variables Yes Yes Yes Yes

Industry × Year Fixed Effects Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes
Observations 108,752 109,217 108,752 109,217
Adjusted R2 0.741 0.744 0.707 0.709

48
Table 6
Continued.

Panel B: Controlling for general uncertainty

Net Book Leverage Net Market Leverage
(1) (2) (3) (4)
IDD 0.014** 0.010** 0.013*** 0.010**
(2.46) (1.98) (2.94) (2.43)
Cash Flow Volatilityt -0.013 0.006
(-0.89) (0.87)
Cash Flow Volatilityt+1 0.005 -0.002
(0.36) (-0.17)
Cash Flow Volatilityt+2 0.003 -0.003
(0.13) (-0.13)
Unlevered Return Volatilityt -0.268*** -0.244***
(-18.59) (-11.50)
Unlevered Return Volatilityt+1 -0.037*** -0.032***
(-5.41) (-5.58)
Unlevered Return Volatilityt+2 -0.047*** -0.023***
(-5.90) (-6.25)

All other Control Variables Yes Yes Yes Yes

Industry × Year Fixed Effects Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes
Observations 110,212 101,013 110,212 101,013
Adjusted R2 0.741 0.770 0.707 0.738

49
Table 7
Firm-level measures of self-reported risk of losing trade secrets and financial leverage.
This table reports results from OLS regressions of Net Book Leverage and Net Market Leverage on firm-
level measures of self-reported risk of losing trade secrets to rivals estimated over the 1993-2011 period. Trade
Secret Risk is an indicator variable set to one if a firm mentions the risk of losing trade secrets in its 10-K filing,
and zero otherwise. Specifically, we search each firm’s 10-K filing for the key phrases “trade secret”, “trade
secrets”, “confidential information”, or “proprietary information” and then set the indicator equal to one if the
firm also mentions “protect”, “protection”, or “safeguard” within a 20-word window before or after one of the
previous key phrases. Key Employee Risk is an indicator variable set to one if a firm mentions the risk of losing
key employees in its 10-K filing, and zero otherwise. Specifically, we search each firm’s 10-K filings for the key
phrases “key personnel”, “key employees”, “key executives”, “key members”, “key managers”, “key officers”, “key
staff”, or “key talent” and then set the indicator equal to one if the firm also mentions “loss of”, “departure of”,
or “retain” within a 20-word window before or after one of the previous key phrases. Control variables are
defined in Table 3. Industry fixed effects are defined at the three-digit SIC level. Continuous variables are
winsorized at their 1st and 99th percentiles. Standard errors are corrected for heteroskedasticity and clustering
at the state level (t-statistics are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1%
levels, respectively.

Net Book Leverage Net Market Leverage

(1) (2) (3) (4)
Trade Secret Risk -0.094*** -0.060***
(-14.44) (-13.55)
Key Employee Risk -0.044*** -0.024***
(-7.66) (-6.31)
Log Book Assets 0.032*** 0.031*** 0.022*** 0.022***
(17.50) (17.08) (17.28) (16.87)
Market-to-Book Assets -0.027*** -0.028*** 0.004*** 0.003***
(-15.79) (-16.52) (4.14) (3.02)
Return on Assets -0.035** -0.023 -0.012 -0.004
(-2.35) (-1.52) (-1.59) (-0.54)
Fixed Assets 0.532*** 0.548*** 0.332*** 0.343***
(26.55) (27.03) (22.62) (23.24)
Cash Flow Volatility -0.088*** -0.100*** -0.068*** -0.076***
(-3.78) (-4.25) (-5.10) (-5.63)
Dividend Payer -0.082*** -0.075*** -0.079*** -0.073***
(-12.05) (-10.88) (-15.62) (-14.54)

Industry × Year Fixed Effects Yes Yes Yes Yes

State Fixed Effects Yes Yes Yes Yes
Observations 62,066 62,066 62,066 62,066
Adjusted R2 0.417 0.409 0.368 0.361

50
 Internet Appendix for
Protection of trade secrets and capital structure decisions

Sandy Klasa, Hernán Ortiz-Molina, Matthew Serfling, and Shweta Srinivasan

A.1. Two legal cases applying the IDD to protect firms’ trade secrets

Below we discuss two legal cases in which state courts applied the IDD. In the first, the

IDD was used to enforce a CNC. In the second, the IDD was used to protect trade secrets

when a CNC did not exist. The court rulings are available from Google Scholar.

A.1.1. Procter & Gamble Co. v. Stoneham, 747 N.E.2d 268 (Ohio Ct. App. 2000)

Stoneham was in charge of international marketing at the Haircare Division of Procter

& Gamble (P&G) and knew confidential information about its global business goals and

strategies (e.g., market research, financial data, new products, and technological

developments). He had signed a CNC with P&G, but he accepted a job offer to work for

Alberto-Culver (AC), who competed with P&G in the market for haircare products, to run AC

International. P&G then sued Stoneham for breach of his CNC, alleging that his employment

at AC would pose an immediate threat that P&G’s trade secrets would be disclosed to AC.

Reversing a prior decision, the Court of Appeals of Ohio enforced the CNC and prohibited

Stoneham from working at AC’s haircare department for three years.

The Court stated that the CNC was reasonable and invoked the IDD to establish the

existence of a threat of irreparable harm warranting injunctive relief, noting that Stoneham

knew P&G’s trade secrets, AC was P&G’s competitor, and his job at AC would be similar to

his prior job at P&G. The ruling also highlighted how the harm was likely to take place. First,

the evidence indicated that after joining AC Stoneham would use his knowledge of P&G’s

trade secrets to increase AC’s competition with P&G on the same line of products he was

responsible for while employed at P&G. Second, the testimonies of P&G’s managers indicated

that AC could use Stoneham’s knowledge to obtain a financial advantage, exploit any

weakness in P&G’s products, easily replicate its pipeline of products without any research or

testing, or pre-empt P&G’s entry into the market for new products.

A.1.2. Air Products & Chemical Inc. v. Johnson, 442 A.2d 1114 (Pa. Super. Ct. 1982)

Air Products & Chemical (APC) and Liquid Air Corporation (LAC) were large

manufacturers and distributors of industrial gases. Johnson, who was employed in

1
Pennsylvania, was in charge of APC’s on-site gas delivery business and knew confidential

information, such as technical data on the methods of delivery, the status of negotiations

with customers, marketing strategies, and market opportunities. He had not signed a CNC

with APC and took a job in California at LAC that involved all of its industrial gas operations,

including on-site delivery. APC feared that Johnson might disclose its trade secrets to LAC

and filed a lawsuit seeking an injunction to prevent Johnson from working at LAC for two

years. The Superior Court of Pennsylvania affirmed a prior injunction issued by a trial court

that prohibited Johnson from working in LAC’s on-site operations and from disclosing APC’s

trade secrets.

In establishing a threat of irreparable harm and thus the need for injunctive relief, the

trial court concluded that Johnson did know APC’s trade secrets and that “It would be

impossible [for Johnson] to perform his managerial functions in on-site work without drawing

on the knowledge he possesses of Air Product’s confidential information.” The ruling also

discussed how the harm was likely to occur. First, it noted that knowledge of APC’s plans for

pipeline delivery of gases in the domestic market could allow a competitor to thwart APC’s

plans or to compete without the burden of testing and market analysis born by APC. Second,

it noted that Johnson knew APC’s costs and pricing methods and in some cases its capital

investment, which would be of great interest and benefit to a competitor.

A.2. The determinants of a state court’s decision to adopt the IDD

In Table A1, we use a Cox proportional hazard model to examine whether state-level

factors affect the likelihood that a state court adopts the IDD. All models include year fixed

effects and cluster standard errors by state. These analyses are helpful in assessing the

validity of the assumption that state courts’ decisions to adopt the IDD are exogenous events.

In model 1, we include variables related to labor and trade secrets laws and labor

unionization. These are an index for the enforcement in a state of covenants not to compete

(Strength of CNCs), an index for the number of wrongful discharge laws the state has adopted

(Wrongful Discharge), an indicator for if the state has passed right-to-work laws (Right-to-

2
Work), an indicator for if the state has adopted legal principles based on the UTSA (State

UTSA), and the fraction of workers in a state that are covered by a collective bargaining

agreement (Union Membership). In model 2, we add proxies for the education and age of the

workers in a state. Log P.C. Enrollment and Log P.C. Colleges measure enrollment in

institutions of higher education and the number of institutions of higher education in a state

per thousand residents, respectively. Bachelor’s Degree is the fraction of workers in the labor

force with at least a bachelor’s degree, and Age of Workers is the average age of the workers

in the labor force. In model 3, we add proxies for the economic and political conditions in the

state. State GDP Growth is the one-year growth rate of state GDP. Log P.C. GDP is the

natural logarithm of per capita GDP. Unemployment Rate is the fraction of workers in the

state labor force that are unemployed. Political Balance is the fraction of a state’s congress

members representing their state in the U.S. House of Representatives that belong to the

Democratic Party, which captures the political leaning in the state.1 In model 4, we add State

M/B and Change in State M/B, the median annual value of market-to-book assets for the

publicly-traded firms in a state and the median annual change to this ratio, to proxy for

shocks to the growth opportunities of the firms in a state. Finally, to proxy for the arrival of

a new large firm in a state, which could lead to a shock to the legal and business environment

in the state, in model 5 we include Large Establishment Entry Rate. This variable is

constructed by the U.S. Census and it is measured as the number of establishments with

1000+ employees that are created or arrive in a state within the last year divided by the

average number of establishments with 1000+ employees in the state over years t and t-1.

We find a positive association between the strength of the enforcement of covenants not

to compete in a state and the likelihood that courts in that state will adopt the IDD. This

suggests state courts that more strictly enforce covenants not to compete may also be more

1 Strength of CNCs is from Bird and Knopf (2014); Wrongful Discharge is from Autor, Donohue, and Schwab
(2006); Right-to-Work is from the Department of Labor; State UTSA is from Malsberger (2011); Union Membership
is from Unionstats (Hirsch and Macpherson (2003)); Enrollment and Colleges is from the Statistical Abstract of
the United States; Bachelor’s Degree and Age of Workers is from the Current Population Survey; State GDP
Growth and P.C. GDP are from the Bureau of Economic Analysis; Unemployment Rate is from the Bureau of Labor
Statistics; and Political Balance is from History, Art & Archives, U.S. House of Representatives.

3
likely to seek stronger protection of firms’ trade secrets by recognizing the IDD. But, none of

the other variables in the Table A1 models are associated with the likelihood courts in the

state will adopt the IDD. Hence, the overall evidence from these state-level analyses supports

the notion that the adoption of the IDD by a state court is an exogenous event.

A.3. Expanded discussion of the results reported in Section 6

A.3.1. Recognition of the IDD and changes in firms’ cost of debt

To further investigate whether the recognition of the IDD reduces the competitive risk a

firm faces, we examine if it affects the firm’s cost of debt. Following Valta (2012), who shows

that credit markets price a firm’s competitive risk into its cost of debt, we focus on the cost of

bank debt because bank debt is the key source of debt financing for most firms and data are

available for a large sample of firms. We explore this issue using data for the period 1987-

2011 provided by Tobias Berg on a firm’s total cost of bank debt, which includes the expected

credit spread and loan contract fees (see Berg, Saunders, and Steffen (2016)). In the cost of

bank debt models we include loan type and state fixed effects, as well as industry-year fixed

effects. We do not include firm fixed effects in these models due to a lack of enough annual

observations per firm.

The results reported in Table A12 show that the recognition of the IDD reduces a firm’s

total cost of bank debt, although this is only the case in the models that include control

variables. The estimated coefficients imply that subsequent to the recognition of the IDD in

a given state, the average total cost of bank debt for the firms in the state decreases by 3.1%

relative to that of these firms’ rivals in non-recognizing states.2 This is additional evidence

that the recognition of the IDD lowers the competitive risks a firm faces.

A.3.2. Announcement returns surrounding a state court’s decision to adopt the IDD

We also examine firms’ abnormal stock returns around the date when a state court adopts

the IDD. The sample includes firms headquartered in the 16 states that adopt the IDD over

2 The results in models 2 and 4 are similar whether book leverage is included as a control variable or not.

4
our sample period with complete data for our main capital structure tests. We estimate

cumulative abnormal returns (CARs) around the adoption date (day 0) using both the market

model and the four-factor model to estimate beta/factor parameters. The estimation window

is the [-280, -61] trading days before day 0. We then calculate CARs over the event window

([-1, +3] trading days) and the pre-event window ([-31, -2] trading days). Because an IDD

ruling in a state simultaneously affects all firms headquartered in the state and this can bias

the standard errors downward, we correct the standard errors for cross-sectional correlation

following the methodology used in Kolari and Pynnönen (2010).

The results reported in the first two columns of Table A13 show that affected firms

experience significant positive abnormal returns over the days surrounding a state court’s

final decision to adopt the IDD. Specifically, the average CARs over the event window [-1, +3]

are 0.70% based on the market model and 0.55% based on the 4-factor model, and both are

statistically significant. These results are consistent with the view that market participants

believe the adoption of the IDD increases the protection of the trade secrets of firms in

adopting states and thereby decreases the competitive risks they face. Supporting the notion

that the changes in state courts’ positions regarding the IDD are unlikely to be anticipated

events, we find that the CARs are not significantly different from zero over the pre-event

window [-31, -2]. The last two columns show that the results are unaffected if we exclude

from the sample those firms that had earnings or distribution announcements during the ±5

trading days around a state court’s final decision adopting the IDD.

A.3.3. Do existing product market competition measures capture the effects of IDD rulings?

Our main tests include industry-year fixed effects and thus our results cannot be driven

by a correlation of IDD rulings with changes in industry-level measures of product market

competition (e.g., industry concentration). However, the change in a firm’s competitive risk

resulting from IDD rulings could be captured by firm-level measures of this risk. We

investigate this issue and first consider the Hoberg et al. (2014) measure of product fluidity,

which measures the competitive threats a firm faces when its rivals are more actively

5
changing the wording they use in their 10-Ks to describe their products that compete with

the firm’s products. If the product fluidity measure also captures competitive threats arising

from an imperfect protection of a firm’s intellectual property, then we might expect a

reduction in product fluidity following the recognition of the IDD.

In Table A14, we run difference-in-differences tests to gauge the impact of the recognition

of the IDD on product fluidity, employing specifications with firm and industry-year fixed

effects. A limitation of these tests is that the product fluidity measure is only available

starting in 1997 and thus our analyses are confined to the period 1997-2011. This reduces

the power of these tests because although in the tests we are still able to use the three IDD

rejection events, we are left with only three of the sixteen IDD adoption events. In all

specifications, we find no impact of IDD rulings on product fluidity. This suggests that

product fluidity and the recognition of the IDD capture competitive threats from different

sources. Specifically, IDD rulings capture changes in the competitive threats a firm faces that

are directly associated with the loss of trade secrets to rivals through the labor mobility

channel, while product fluidity likely better captures other sources of competitive threats,

such as those related to improvements in competitors’ products. Of course, the results in

Table A14 might also be due to weak statistical power of the empirical tests.

We also examine if changes in competitive risk after IDD rulings are captured by two

other measures of this risk. Following MacKay and Phillips (2005), Haushalter et al. (2007),

Frésard (2010), and Valta (2012) we considered the similarity of operations between a firm

and its rivals, measured for each year by the absolute difference between its capital-labor

ratio and the median capital-labor ratio in its three-digit SIC industry. A smaller difference

(more similar operations) implies that the firm faces a greater risk that its rivals will try to

steal some of its market share. As in Haushalter et al. (2007) and Frésard (2010), we also

considered the covariance between a firm’s stock returns and the returns of its three-digit

SIC industry rivals. A higher covariance implies that a firm’s growth opportunities are more

interdependent with those of its rivals, and therefore that it faces a greater risk that its rivals

will try to steal some of its market share.

6
 The results of our difference-in-differences regressions over our full 1977-2011 sample

period reported in Table A15 show that IDD rulings have no impact on either of the firm-

level measures of product market competition discussed above. These results are further

evidence that IDD rulings capture changes in the competitive risk a firm faces that are not

captured by other product market competition measures.

A.3.4. Impact of the IDD on litigation involving trade secrets

We also examine whether IDD rulings affect the number of legal cases involving trade

secrets litigated in the recognizing state, but note there are two opposite forces at play. On

one hand, the recognition of the IDD discourages workers who know their employers’ trade

secrets from seeking employment at rival firms and thus reduces employers’ needs to sue

departing employees to protect their trade secrets. On the other hand, this recognition makes

it easier for firms to obtain injunctions to prevent their workers who know trade secrets from

working for rivals, thus increasing firms’ propensities to sue to protect their trade secrets.

To this end, we compile a dataset containing the number of trade secrets cases filed in

each state during every year over the period 1977-2011 using an approach similar to that in

in Almeling et al. (2011). Specifically, we search LexisNexis for trade secrets cases litigated

in state courts, district courts, and federal courts of appeals in the circuit to which each state

belongs. We identify as trade secrets cases those in which the strings “trade secret(s)” appears

at least three times in the whole document, which yields a total of 11,120 trade secrets cases

across all states. We next identify the state in which each case was litigated and the resulting

dataset with the number of cases by state contains a total of 1,750 state-year observations.

In Table A16, we regress the natural logarithm of one plus the total number of trade

secrets cases in a state on the IDD indicator and include state and year fixed effects. We

consider specifications with and without the following control variables: Strength of CNCs,

State GDP Growth, and Political Balance. We find that the recognition of the IDD in a state

increases the number of trade secrets cases litigated in that state. This implies that, in spite

of reducing the ex-ante mobility to rival firms of workers who know their employers’ trade

7
secrets, the recognition of the IDD increases firms’ propensities to litigate to protect their

trade secrets, possibly because they expect more favorable outcomes after this recognition.

A.3.5. The effect of IDD rulings on capital structure controlling for corporate governance

In Bolton and Scharfstein (1990), the optimal debt contract balances the benefits of

deterring predation by reducing a firm’s financial constraints against the cost of exacerbating

managerial agency problems. Thus, to investigate whether the study’s findings are robust to

keeping a firm’s corporate governance constant we examined if the impact of the recognition

of the IDD on a firm’s capital structure is affected by whether we control for its corporate

governance. To proxy for a firm’s corporate governance environment, we first consider

whether the firm is incorporated in a state that has passed business combination laws which

reduce the fear for managers of a hostile takeover, and consequently weaken the firm’s

corporate governance (e.g., Bertrand and Mullainathan (2003)). Second, we consider the

concentration of the ownership of the firm’s equity shares by institutional investors given

that when this ownership concentration is higher institutional investors are expected to

monitor a firm more intensely (e.g., Hartzell and Starks (2003)).

In Table A17, we report the results of the analysis described above. In models 1 and 4 we

report the results when we add the indicator for whether a firm is incorporated in a state

that has passed business combination laws as a control variable to our baseline capital

structure specification that includes control variables, and firm and industry-year fixed

effects. The results show that this control variable has no impact on the effect of the

recognition of the IDD in a firm’s state on its net leverage. Next, in models 2-3 and 5-6 we

control for corporate governance with institutional ownership concentration. In models 2 and

5, it is measured as the fraction of total institutional ownership accounted for by the five

largest institutional investors, while in models 3 and 6 it is measured as the Herfindahl-

Hirschman Index based on the percentages of institutional holdings by all institutional

investors. The findings for models 2-3 and 5-6 show that controlling for institutional

ownership concentration does not have an impact on the effect of the recognition of the IDD

8
on a firm’s net leverage.3 Put together, the Table A17 findings suggest that the effect of the

recognition of the IDD in a firm’s state on its capital structure decisions is robust to keeping

its corporate governance constant.

A.3.6. Measurement error in the IDD indicator

In Table A18, we examine whether measurement error in the protection of trade secrets

afforded by the IDD indicator variable affects our results. First, in our main capital structure

tests we use firms’ most recent state of headquarters from Compustat to determine the

applicability of the IDD, but firms may relocate their headquarters across states during our

sample period. Hence, we collect data on the historical state of location of firms’ headquarters

from their 10-K filings available on EDGAR. Our results are unchanged if we use historical

addresses obtained from 10-K filings to update a firm’s state of headquarters when available4

or if we discard from the sample those firms that could have relocated headquarters due to

major restructuring events (measured by sales or asset growth in excess of 100% in any year

from 1977-2011).5 Second, the recognition of the IDD in a firm’s state of headquarters could

be a noisy proxy for the protection of its trade secrets if the firm has employees in foreign

countries or it has a geographically dispersed workforce. Discarding from the sample firm-

years in which a firm reports foreign activities or firms in industries with geographically

dispersed workforces does not affect our results.

A.3.7. Propensity-score matching

In Table A19, we re-estimate the effect of the adoption of the IDD on financial leverage

using propensity-score matching to ensure that the treatment and control firms have similar

3 Although not tabulated, we also tried including total institutional ownership as an additional control variable
in models 2-3 and 5-6 and find this has little impact on the results for these models. Likewise, if we control for
both the indicator for whether a firm is incorporated in a state that has passed business combination laws and
the concentration of institutional ownership this does not affect the result that controlling for a firm’s corporate
governance does not impact the effect of the recognition of the IDD in a firm’s state on its capital structure.
4 We obtain the historical address for most firms from 1996-2011 and for some since 1992. Relocations of

headquarters across states are infrequent. Of the 8,637 firms for which we obtain the historical location of
headquarters over the 1992-2011 period, only 9% of these firms relocated headquarters from one state to another.
5 Pirinsky and Wang (2006) argue that major restructuring events are the main trigger of headquarter relocations

and Almeida, Campello, and Weisbach (2004) highlight that such events are usually associated with large
increases in sales or assets.

9
observable characteristics before the onset of the treatment. We match each treatment firm

(in an adopting state) to at least one and at most two control firms (in never adopting states)

in the year before the adoption (with replacement), first exactly on year and two-digit SIC

and then on the closest propensity score arising from a logistic regression using Log Book

Assets, Market-to-Book Assets, and Return on Assets as predictors of treatment. The

treatment and matched control groups do not differ in any of the three matching

characteristics or propensity scores. We continue to find that the recognition of the IDD has

a positive impact on financial leverage and that firms increase their leverage following the

adoption of the IDD in their state (but not before).

A.3.8. Litigation involving trade secrets and debt issuance

In our paper we argue that having unused debt capacity is strategically useful for a firm

that faces a higher risk of losing its trade secrets to rivals. A key reason is that unused debt

capacity allows the firm to increase its borrowing and use the funds to protect its competitive

position when it loses trade secrets to rivals. In Table A20 we examine whether firms

significantly increase the level of their outstanding debt during years when they litigate to

protect their trade secrets, relative to firms that do not engage in such litigation during those

years. Evidence that over such years firms indeed raise their outstanding debt would support

the argument that having unused debt capacity is incrementally beneficial when a firm faces

a higher likelihood that its competitors will try to obtain its trade secrets.

To build the sample for these analyses, we first consider all trade secrets cases litigated

in state, district, and federal courts over the 1992-2011 period (which is the last twenty years

of the sample period we use in our main tests). We follow Almeling et al. (2011) and identify

trade secrets cases from LexisNexis, defining a case as a trade secrets case if the strings

“trade secret(s)” appears at least three times in the whole legal case document.6 This yields

a total of 9,459 trade secrets cases over the 1992-2011 period. For each of these cases, we next

6 This is analogous to what we did in our analyses of the effect of the recognition of the IDD on the number of
trade secrets cases in a state (also discussed in this Internet Appendix).

10
manually identify who is the plaintiff in the case and also manually verify whether the

plaintiff is a publicly traded firm that is included in the Compustat database. Finally, we also

require that the plaintiff firm is included in the main sample used in the paper. Having

imposed these requirements we are left with 511 trade secrets cases in which the plaintiff is

a publicly traded firm included in our main sample. The overall sample we use in our Table

A20 tests contains those firms that acted as plaintiffs in trade secrets cases as well as all

other firms that did not, for a total of 71,191 firm-year observations.

In the first four models of Table A20, we regress the change in a firm’s net debt over a

given year scaled by lagged book or market assets on indicators for whether the firm was the

plaintiff in a trade secrets case during the prior year, the current year, or the next year. We

report the results of models without or with control variables, where the control variables are

the firm-level controls included in our Table 3 models, but lagged one year.7 In all models we

include firm and three-digit SIC industry-year fixed effects.

The results for the first four models of Table A20 show that across all four specifications,

a firm significantly increases its net debt during the year when it is the plaintiff in a trade

secret’s case. In models 5-8 of Table A20, we change the dependent variable to be the one-

year change in total debt over a given year scaled by lagged book or market assets. This

allows us to directly examine whether a firm increases the level of its outstanding debt during

the year when it is a plaintiff in a trade secrets case. Across all four specifications, we find

that firms significantly increase their outstanding debt during the year when they are a

plaintiff in a trade secrets case.

The estimated coefficients in models 6 and 8 of Table A20, which include control variables

and firm and three-digit SIC industry-year fixed effects, imply that during the year when a

firm is a plaintiff in a trade secrets case it increases its outstanding debt per dollar of book

(market) assets by 1.8 (1.1) cents relative to that of its industry rivals who are not the plaintiff

in a trade secrets case that year. This represents an 8.2% (7.1%) increase in book (market)

7The results of the Table A20 regression models that include control variables are very similar if we use the one-
year changes in the control variables instead of their levels, or if we include the control variables both in levels
and in changes.

11
leverage relative to the mean values for the sample used in the Table A20 tests for book

(market) leverage of 0.219 (0.156).8

In sum, the Table A20 findings show that firms significantly increase their outstanding

debt during years when they litigate to protect their trade secrets. These findings provide

support to our argument that having unused debt capacity is strategically useful for firms

that face a greater risk of losing their trade secrets to rivals, because it allows such firms to

rapidly raise the funds they need to protect their competitive position in the event their rivals

actually obtain some of these secrets.

8The results for models 5-8 in Table A20 are very similar if we consider only the increase in long-term borrowing
during the year when a firm litigates to protect its trade secrets.

12
Additional References

Almeida, H., Campello, M., Weisbach, M.S., 2004. The cash flow sensitivity of cash. Journal
of Finance 59, 1777-1804.

Autor, D.H., Donohue, J.J., Schwab, S.S., 2006. The costs of wrongful-discharge laws. Review
of Economics and Statistics 88, 211–231.

Berg, T., Saunders, A., Steffen, S., 2016. The total cost of corporate borrowing in the loan
market: Don’t ignore the fees. Journal of Finance 71, 1357-1392.

Bertrand, M., Mullainathan, S., 2003. Enjoying the quiet life? Corporate governance and
managerial preferences. Journal of Political Economy 111, 1043-1075.

Campello, M., Lin, C., Ma, Y., Zou, H., 2011. The real and financial implications of corporate
hedging. Journal of Finance 66, 1615-1647.

Hartzell, J.C., Starks, L.T., 2003. Institutional investors and executive compensation.
Journal of Finance 58, 2351-2374.

Hirsch, B.T., Macpherson, D.A., 2003. Union membership and coverage database from the
Current Populations Survey: Note. Industrial and Labor Relations Review 56, 349–354.

Kolari, J.W., Pynnönen, S., 2010. Event study testing with cross-sectional correlation of
abnormal returns. Review of Financial Studies 23, 3996-4025.

Pirinsky, C.A., Wang, Q., 2006. Does corporate headquarters location matter for stock
returns? Journal of Finance 61, 1991-2015.

13
Table A1
Determinants of the adoption of the IDD by state courts.
This table reports results from Cox proportional hazard models analyzing the hazard of a state court
adopting the IDD. The sample period is from 1977 to 2011. A “failure event” is the adoption of the IDD in a
given state. States are excluded from the sample once they adopt the IDD. Independent variables are measured
at the state level as of year t-1 and include: Strength of CNCs, an index of the strength of the enforcement of
covenants not to compete (CNCs) by courts in a state (higher values imply stronger enforcement); Wrongful
Discharge, the number of wrongful discharge laws (i.e., good faith, implied contract, and public policy
exceptions) a state has adopted; Right-to-Work, an indicator variable set to one if the state has passed right-to-
work laws, and zero otherwise; State UTSA, an indicator variable set to one if the state has adopted legal
principles based on the Uniform Trade Secrets Act, and zero otherwise; Union Membership, the fraction of
nonagricultural wage and salary employees covered by collective bargaining agreements; Log P.C. Enrollment,
the natural logarithm of enrollment in institutions of higher education in a state per thousand residents; Log
P.C. Colleges, the natural logarithm of the number of degree-granting institutions of higher education (colleges)
in a state per thousand residents; Bachelor’s Degree, the percentage of individuals in the labor force with at
least a bachelor’s degree; Age of Workers, the average age of individuals in the labor force; State GDP Growth,
the annual GDP growth rate in the state; Log P.C. GDP, the natural logarithm of per capita GDP in thousands;
Unemployment Rate, the fraction of workers in the labor force that are unemployed; Political Balance, the
fraction of a state’s congress members representing their state in the U.S. House of Representatives that belong
to the Democratic Party; State M/B, the median value of Market-to-Book Assets across all firms in a state
during a given year, where Market-to-Book Assets is the market value of assets (prcc_f*csho + at - ceq) divided
by the book value of assets (at); Change in State M/B, the median value of the one-year change in Market-to-
Book Assets across all firms in a state during a given year; and Large Establishment Entry Rate, the number of
establishments with 1000+ employees that are created or arrive in a state within the last year divided by the
average number of establishments with 1000+ employees in the state over years t and t-1. Data on
establishments are from the Census’ Business Dynamic Statistics Database. In models 4 and 5 the number of
observations drops because for a few state-years there are no firms included in Compustat. All independent
variables, except indicator variables, are standardized to have a mean of zero and a standard deviation of one.
Dollar values are expressed in 2009 dollars. Standard errors in parentheses are clustered by state. *, **, and
*** denote significance at the 10%, 5%, and 1% levels, respectively.

14
Table A1
Continued.
(1) (2) (3) (4) (5)
Strength of CNCs 1.064*** 1.111*** 1.352*** 1.380*** 1.392***
(4.10) (4.16) (3.38) (3.45) (3.48)
Wrongful Discharge -0.036 -0.052 -0.297 -0.316 -0.305
(-0.10) (-0.16) (-0.81) (-0.88) (-0.86)
Right-to-Work -0.431 -0.303 -0.714 -0.713 -0.732
(-0.65) (-0.34) (-0.75) (-0.75) (-0.78)
State UTSA 0.279 0.176 0.195 0.178 0.238
(0.31) (0.19) (0.20) (0.19) (0.25)
Union Membership 0.456 0.501 0.381 0.354 0.346
(1.06) (1.00) (0.58) (0.55) (0.54)
Log P.C. Enrollment -0.097 -0.103 -0.119 -0.123
(-0.26) (-0.21) (-0.25) (-0.26)
Log P.C. Colleges -0.180 -0.309 -0.274 -0.284
(-0.63) (-1.07) (-0.91) (-0.98)
Bachelor's Degree 0.754 0.413 0.363 0.324
(0.88) (0.43) (0.37) (0.33)
Age of Workers -0.737 -0.573 -0.632 -0.608
(-1.12) (-0.77) (-0.78) (-0.76)
State GDP Growth -0.197 -0.266 -0.273
(-0.95) (-1.13) (-1.15)
Log P.C. GDP 0.449 0.576 0.573
(1.15) (1.45) (1.46)
Unemployment Rate -0.357 -0.350 -0.357
(-0.58) (-0.56) (-0.57)
Political Balance -0.303 -0.295 -0.290
(-0.85) (-0.82) (-0.81)
State M/B 0.089 0.106
(0.36) (0.48)
Change in State M/B 0.283 0.271
(0.94) (0.91)
Large Establishment Entry Rate -0.177
(-0.63)

Year Fixed Effects Yes Yes Yes Yes Yes

Observations 1,297 1,297 1,297 1,293 1,293
Pseudo R2 0.215 0.228 0.237 0.240 0.240

15
Table A2
Summary statistics for variables in main capital structure tests (Table 3).
This table reports summary statistics for the variables used in our main capital structure tests in Table 3.
All variable definitions are in Tables 3 and A1.

Obs. Mean Std. Dev. P25 Median P75

Net Book Leverage 125,895 0.068 0.349 -0.140 0.110 0.303
Net Market Leverage 125,895 0.082 0.250 -0.066 0.070 0.242
Book Leverage 125,895 0.234 0.213 0.044 0.200 0.359
Market Leverage 125,895 0.182 0.180 0.022 0.134 0.291
IDD 125,895 0.421 0.494 0.000 0.000 1.000
Book Assets (millions) 125,895 1397 4208 42.04 163.0 704.7
Market-to-Book Assets 125,895 1.921 1.667 1.031 1.369 2.087
Return on Assets 125,895 0.062 0.226 0.032 0.113 0.175
Fixed Assets 125,895 0.287 0.221 0.110 0.232 0.409
Cash Flow Volatility 125,895 0.094 0.138 0.026 0.049 0.099
Dividend Payer 125,895 0.338 0.473 0.000 0.000 1.000
State GDP Growth 125,895 0.064 0.036 0.042 0.062 0.087
Political Balance 125,895 0.568 0.184 0.500 0.578 0.644
Strength of CNCs 125,895 3.805 2.167 3.000 4.000 5.000

16
Table A3
Recognition of the IDD and capital structure.
This table reports results from OLS regressions of financial leverage on the indicator for the recognition of the IDD. The sample spans the 1977-2011
period. The dependent variables are Net Book Leverage (models 1 and 2), Net Market Leverage (models 3 and 4), Book Leverage (models 5 and 6), and
Market Leverage (models 7 and 8). Net Book Leverage is the book value of long-term debt (dltt) plus debt in current liabilities (dlc) minus cash holdings
(che) divided by book value of assets (at). Net Market Leverage is the book value of long-term debt (dltt) plus debt in current liabilities (dlc) minus cash
holdings (che) divided by market value of assets (prcc_f*csho + at - ceq). Book Leverage and Market Leverage are analogously defined, except that cash
holdings are not subtracted in the numerator. IDD is an indicator variable equal to one if the firm is headquartered in a state whose courts recognize the
IDD, and zero otherwise. The control variables are defined in Tables 2 and 3. Dollar values are expressed in 2009 dollars. Continuous variables, except
state-level variables, are winsorized at their 1st and 99th percentiles. Standard errors are corrected for heteroskedasticity and clustering at the state level
(t-statistics are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

Net Book Leverage Net Market Leverage Book Leverage Market Leverage
(1) (2) (3) (4) (5) (6) (7) (8)
IDD 0.015** 0.018*** 0.012*** 0.016*** 0.013*** 0.015*** 0.009** 0.012***
(2.56) (3.30) (2.60) (3.63) (3.08) (3.91) (2.27) (3.59)
Log Book Assets 0.039*** 0.045*** 0.030*** 0.034***
(8.87) (11.25) (8.85) (11.00)
Market-to-Book Assets -0.014*** 0.007** -0.004*** -0.019***
(-16.39) (2.18) (-5.01) (-8.27)
Return on Assets -0.180*** -0.102*** -0.163*** -0.134***
(-16.88) (-6.27) (-15.66) (-8.62)
Fixed Assets 0.642*** 0.430*** 0.230*** 0.183***
(15.85) (27.67) (14.73) (17.51)
Cash Flow Volatility -0.025 -0.019* 0.042*** -0.001
(-1.50) (-1.85) (3.81) (-0.12)
Dividend Payer -0.062*** -0.052*** -0.051*** -0.045***
(-14.76) (-14.19) (-12.37) (-13.10)
Strength of CNCs 0.003 0.003 0.001 0.002
(0.67) (0.73) (0.33) (0.44)
State GDP Growth -0.062 -0.190*** -0.043 -0.188***
(-1.43) (-3.91) (-1.28) (-4.33)
Political Balance -0.005 -0.015 0.003 -0.008
(-0.33) (-1.43) (0.28) (-1.18)

Year Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes
Observations 125,895 125,895 125,895 125,895 125,895 125,895 125,895 125,895
Adjusted R2 0.691 0.729 0.655 0.690 0.609 0.637 0.636 0.685

17
Table A4
Recognition of the IDD and financial leverage measured using only long-term debt.
This table reports results from OLS regressions of financial leverage on the indicator for the recognition of the
IDD. The sample spans the 1977-2011 period. The dependent variables are Long-Term Book Leverage (models 1 and
2) and Long-Term Market Leverage (models 3 and 4). Long-Term Book Leverage is the book value of long-term debt
(dltt) plus the current portion of long-term debt (dd1) divided by book value of assets (at). Long-Term Market Leverage
is the book value of long-term debt (dltt) plus the current portion of long-term debt (dd1) divided by market value of
assets (prcc_f*csho + at - ceq). IDD is an indicator variable equal to one if the firm is headquartered in a state whose
courts recognize the IDD, and zero otherwise. The control variables are defined in Tables 2 and 3 and include Log
Book Assets, Market-to-Book Assets, Return on Assets, Fixed Assets, Cash Flow Volatility, Dividend Payer, Strength of
CNCs, State GDP Growth, and Political Balance. Industry fixed effects are defined at the three-digit SIC level.
Continuous variables, except state-level variables, are winsorized at their 1st and 99th percentiles. Standard errors
are corrected for heteroskedasticity and clustering at the state level (t-statistics are in parentheses). *, **, and ***
denote significance at the 10%, 5%, and 1% levels, respectively.

Long-Term Book Leverage Long-Term Market Leverage

(1) (2) (3) (4)
IDD 0.008*** 0.009*** 0.005** 0.008***
(2.74) (2.82) (2.03) (2.76)

Control Variables No Yes No Yes

Industry × Year Fixed Effects Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes
Observations 125,895 125,895 125,895 125,895
Adjusted R2 0.626 0.648 0.657 0.694

18
Table A5
Adoption of the IDD vs. rejection of the IDD after adoption.
This table reports results from OLS regressions of Net Book Leverage and Net Market Leverage on
indicators for the adoption or rejection of the (previously adopted) IDD in the state where a firm is
headquartered. The sample spans the 1977-2011 period. IDD Adoption is an indicator variable set to one
beginning the year when the state where the firm is headquartered adopts the IDD (it remains equal to one for
all subsequent years), and zero otherwise. IDD Rejection is an indicator variable set to one beginning the year
when the state where the firm is headquartered rejects the previously adopted IDD, and zero otherwise. The
control variables are defined in Tables 2 and 3 and include Log Book Assets, Market-to-Book Assets, Return on
Assets, Fixed Assets, Cash Flow Volatility, Dividend Payer, Strength of CNCs, State GDP Growth, and Political
Balance. Industry fixed effects are defined at the three-digit SIC level. Continuous variables, except state-level
variables, are winsorized at their 1st and 99th percentiles. Standard errors are corrected for heteroskedasticity
and clustering at the state level (t-statistics are in parentheses). *, **, and *** denote significance at the 10%,
5%, and 1% levels, respectively.

Net Book Leverage Net Market Leverage

(1) (2) (3) (4)
IDD Adoption 0.017** 0.016** 0.014*** 0.013***
(2.56) (2.54) (3.02) (2.97)
IDD Rejection 0.005 -0.003 -0.002 -0.008
(0.56) (-0.30) (-0.22) (-1.05)

Control Variables No Yes No Yes

Industry × Year Fixed Effects Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes
Observations 125,895 125,895 125,895 125,895
Adjusted R2 0.702 0.739 0.674 0.707

19
Table A6
The timing of changes in firms’ capital structures around adoptions of the IDD.
This table reports results from OLS regressions of Net Book Leverage and Net Market Leverage on
indicators for the timing of state courts’ adoptions of the IDD. The sample spans the 1977-2011 period. IDD
Adoption-4, IDD Adoption-3, IDD Adoption-2, IDD Adoption-1, IDD Adoption0, IDD Adoption+1, IDD Adoption+2,
Adoption+3 and IDD Adoption4+ are equal to one if the firm is headquartered in a state that will adopt the IDD
in four years, adopts the IDD in three years, adopts the IDD in two years, adopts the IDD in one year, adopts
the IDD in the current year, adopted the IDD one year ago, adopted the IDD two years ago, adopted the IDD
three years ago, or adopted the IDD four or more years ago, respectively, and zero otherwise. IDD Rejection is
an indicator variable set to one beginning the year when the state where the firm is headquartered rejects the
previously adopted IDD, and zero otherwise. The control variables are defined in Tables 2 and 3 and include
Log Book Assets, Market-to-Book Assets, Return on Assets, Fixed Assets, Cash Flow Volatility, Dividend Payer,
Strength of CNCs, State GDP Growth, and Political Balance. Industry fixed effects are defined at the three-
digit SIC level. Continuous variables, except state-level variables, are winsorized at their 1st and 99th
percentiles. Standard errors are corrected for heteroskedasticity and clustering at the state level (t-statistics
are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

Net Book Leverage Net Market Leverage

(1) (2) (3) (4)
IDD Adoption -4 -0.005 -0.003 -0.006 -0.004
(-0.64) (-0.40) (-1.07) (-0.77)
IDD Adoption -3 -0.009 -0.005 -0.008 -0.005
(-1.25) (-0.79) (-1.32) (-0.97)
IDD Adoption -2 -0.001 0.004 -0.003 -0.001
(-0.15) (0.59) (-0.60) (-0.11)
IDD Adoption -1 0.003 0.007 -0.001 0.002
(0.36) (0.98) (-0.18) (0.25)
IDD Adoption 0 0.005 0.008 0.003 0.005
(0.69) (1.16) (0.46) (0.83)
IDD Adoption +1 0.017** 0.018** 0.011* 0.012**
(2.32) (2.54) (1.82) (2.13)
IDD Adoption +2 0.018** 0.019** 0.015** 0.015***
(2.08) (2.27) (2.44) (2.67)
IDD Adoption +3 0.017* 0.019** 0.011 0.013*
(1.83) (2.11) (1.51) (1.90)
IDD Adoption 4+ 0.019 0.020* 0.014 0.015*
(1.64) (1.69) (1.56) (1.76)
IDD Rejection 0.005 -0.003 -0.002 -0.008
(0.58) (-0.31) (-0.22) (-1.07)

Control Variables No Yes No Yes

Industry × Year Fixed Effects Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes
Observations 125,895 125,895 125,895 125,895
Adjusted R2 0.702 0.739 0.674 0.707

20
Table A7
Recognition of the IDD and profitability.
This table reports results from OLS regressions of profitability on the indicator for the recognition of the
IDD. The sample spans the 1977-2011 period. The two alternative dependent variables are as follows: Profit
Margin is the sum of pre-tax income, interest expense, and depreciation and amortization (pi+xint+dp), all
divided by sales (sale); and Return on Assets is operating income before depreciation (oibdp) divided by book
value of assets (at). IDD is an indicator variable equal to one if the firm is headquartered in a state whose
courts recognize the IDD, and zero otherwise. The control variables are defined in Tables 2 and 3. Industry
fixed effects are defined at the three-digit SIC level. Continuous variables, except state-level variables, are
winsorized at their 1st and 99th percentiles. Standard errors are corrected for heteroskedasticity and clustering
at the state level (t-statistics are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1%
levels, respectively.

Profit Margin Return on Assets

(1) (2) (3) (4)
IDD -0.012 -0.002 -0.005 -0.001
(-0.46) (-0.07) (-1.02) (-0.16)
Log Book Assets 0.138*** 0.047***
(7.49) (6.73)
Fixed Assets 0.086 -0.120***
(0.54) (-4.15)
Cash Flow Volatility -2.092*** -0.433***
(-10.26) (-14.56)
Strength of CNCs -0.007 -0.000
(-0.97) (-0.37)
State GDP Growth -0.205 0.036
(-1.00) (1.63)
Political Balance 0.059 -0.003
(0.92) (-0.33)

Industry × Year Fixed Effects Yes Yes Yes Yes

Firm Fixed Effects Yes Yes Yes Yes
Observations 125,895 125,895 125,895 125,895
Adjusted R2 0.567 0.575 0.647 0.689

21
Table A8
Recognition of the IDD and capital structure: Effect of tax benefits of debt.
This table reports results from OLS regressions of net financial leverage (Net Book Leverage in Panel A
and Net Market Leverage in Panel B) on the indicator for the recognition of the IDD. The sample period spans
the 1980-2011 period for model 1 and the 1977-2011 period for models 2-4. IDD is an indicator variable equal
to one if the firm is headquartered in a state whose courts recognize the IDD, and zero otherwise. Marginal
Tax Rate is a firm’s marginal tax rate before interest deductions estimated by Blouin, Core, and Guay (2010).
Depreciation is depreciation and amortization divided by book value of assets (dp/at). Tax Loss Carry Forward
is the value of tax loss carry forwards divided by book value of assets (tlcf/at). Investment Tax Credit is the
accumulated tax deferrals of investment tax credits divided by book value of assets (itcb/at). Marginal Tax
Rate, Depreciation, Tax Loss Carry Forward, and Investment Tax Credit are demeaned to ease the
interpretation of coefficient estimates on the interaction term. Industry fixed effects are defined at the three-
digit SIC level. Standard errors are corrected for heteroskedasticity and clustering at the state level (t-statistics
are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

Panel A: Dependent variable is net book leverage

Net Book Leverage
(1) (2) (3) (4)
IDD 0.009* 0.011** 0.010** 0.012**
(1.95) (2.30) (2.37) (2.46)
IDD × Marginal Tax Rate 0.024
(0.79)
Marginal Tax Rate -0.343***
(-12.70)
IDD × Depreciation 0.082
(0.56)
Depreciation 1.864***
(10.00)
IDD × Tax Loss Carry Forward 0.002
(0.43)
Tax Loss Carry Forward 0.037***
(14.18)
IDD × Investment Tax Credit 3.019
(0.40)
Investment Tax Credit -8.005*
(-1.70)

Industry × Year Fixed Effects Yes Yes Yes Yes

Firm Fixed Effects Yes Yes Yes Yes
Observations 116,312 125,881 93,822 124,698
Adjusted R2 0.705 0.717 0.716 0.693

22
Table A8
Continued.

Panel B: Dependent variable is net market leverage

Net Market Leverage
(1) (2) (3) (4)
IDD 0.009** 0.011*** 0.008** 0.011***
(2.53) (2.67) (2.34) (2.82)
IDD × Marginal Tax Rate 0.026
(0.86)
Marginal Tax Rate -0.162***
(-7.17)
IDD × Depreciation 0.023
(0.37)
Depreciation 0.859***
(16.79)
IDD × Tax Loss Carry Forward -0.001
(-0.22)
Tax Loss Carry Forward 0.012***
(6.24)
IDD × Investment Tax Credit 1.856
(0.28)
Investment Tax Credit -6.038*
(-1.66)

Industry × Year Fixed Effects Yes Yes Yes Yes

Firm Fixed Effects Yes Yes Yes Yes
Observations 116,312 125,881 93,822 124,698
Adjusted R2 0.674 0.662 0.688 0.673

23
Table A9
Recognition of the IDD and unlevered stock return volatility.
This table reports results from OLS regressions of a firm’s unlevered stock return volatility on the indicator
for the recognition of the IDD. The sample spans the 1977-2011 period. The dependent variable Unlevered
Return Volatility is the annualized standard deviation of a firm’s stock returns over a fiscal year (we require
the firm to have at least 180 observations during the fiscal year) multiplied by the ratio of the firm’s market
value of equity to its market value of assets ((csho*prcc_f)/(at-ceq+csho*prcc_f)). IDD is an indicator variable
equal to one if the firm is headquartered in a state whose courts recognize the IDD, and zero otherwise. The
control variables are defined in Tables 2 and 3. Industry fixed effects are defined at the three-digit SIC level.
Continuous variables, except state-level variables, are winsorized at their 1st and 99th percentiles. Standard
errors are corrected for heteroskedasticity and clustering at the state level (t-statistics are in parentheses). *,
**, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

Unlevered Return Volatility × 100

(1) (2)
IDD 0.167 -0.148
(0.26) (-0.26)
Log Book Assets -6.018***
(-23.01)
Fixed Assets -9.533***
(-5.16)
Strength of CNCs -0.015
(-0.08)
State GDP Growth 10.321***
(3.14)
Political Balance 0.692
(0.80)

Industry × Year Fixed Effects Yes Yes

Firm Fixed Effects Yes Yes
Observations 122,170 122,170
Adjusted R2 0.708 0.724

24
Table A10
Recognition of the IDD and spending in the development of trade secrets.
This table reports the results from OLS regressions of expenditures related to the development of trade
secrets on the indicator for the recognition of the IDD. The sample spans the 1977-2011 period. The dependent
variables are as follows: R&D/Sales is research and development expenditures scaled by sales (xrd/sale);
(R&D + Capex)/Sales is research and development expenditures plus capital expenditures scaled by sales
((xrd+capx)/sale); (R&D + Capex + Adv)/Sales is the sum of research and development expenditures, capital
expenditures, and advertising expenditures scaled by sales ((xrd+capx+xad)/sale). xrd and xad are set to zero
if the value is missing in Compustat. The control variables are defined in Tables 2 and 3. Industry fixed effects
are defined at the three-digit SIC level. Continuous variables, except state-level variables, are winsorized at
their 1st and 99th percentiles. Standard errors are corrected for heteroskedasticity and clustering at the state
level (t-statistics are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1% levels,
respectively.

(R&D + Capex + Adv)

R&D / Sales (R&D + Capex) / Sales
/ Sales
(1) (2) (3) (4) (5) (6)
IDD 0.009 0.006 0.017 0.014 0.018 0.014
(0.82) (0.54) (1.29) (1.12) (1.37) (1.17)
Log Book Assets 0.024*** 0.059*** 0.062***
(2.81) (5.22) (5.44)
Market-to-Book Assets 0.018*** 0.030*** 0.031***
(4.40) (3.69) (3.72)
Return on Assets -0.864*** -1.185*** -1.233***
(-8.14) (-10.26) (-10.55)
Fixed Assets -0.154** 0.128 0.114
(-2.38) (1.25) (1.08)
Cash Flow Volatility 0.084 0.084 0.093
(1.39) (0.90) (1.00)
Dividend Payer 0.009** 0.005 0.005
(2.03) (0.79) (0.79)
Strength of CNCs 0.003 0.004 0.004
(0.77) (0.82) (0.82)
State GDP Growth 0.047 0.181 0.208
(0.65) (1.32) (1.50)
Political Balance 0.003 -0.008 -0.008
(0.14) (-0.26) (-0.27)

Industry × Year Fixed Effects Yes Yes Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes Yes Yes
Observations 125,895 125,895 124,674 124,674 124,674 124,674
Adjusted R2 0.683 0.706 0.618 0.638 0.614 0.635

25
Table A11
Recognition of the IDD and capital structure: Controlling for spending in the
development of trade secrets.
This table reports results from OLS regressions of net financial leverage (Net Book Leverage in models 1-3
and Net Market Leverage in models 4-6) on the indicator for the recognition of the IDD. The sample spans the
1977-2011 period. IDD is an indicator variable equal to one if the firm is headquartered in a state whose courts
recognize the IDD, and zero otherwise. R&D/Sales is research and development expenditures scaled by sales
(xrd/sale). (R&D + Capex)/Sales is research and development expenditures plus capital expenditures scaled
by sales ((xrd+capx)/sale). (R&D + Capex + Adv)/Sales is the sum of research and development expenditures,
capital expenditures, and advertising expenditures scaled by sales ((xrd+capx+xad)/sale). xrd and xad are set
to zero if the value is missing in Compustat. The other control variables are defined in Tables 2 and 3. Industry
fixed effects are defined at the three-digit SIC level. Continuous variables, except state-level variables, are
winsorized at their 1st and 99th percentiles. Standard errors are corrected for heteroskedasticity and clustering
at the state level (t-statistics are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1%
levels, respectively.

Net Book Leverage Net Market Leverage

(1) (2) (3) (4) (5) (6)
IDD 0.013*** 0.012** 0.012** 0.012*** 0.012*** 0.012***
(2.58) (2.42) (2.42) (3.14) (3.09) (3.10)
Log Book Assets 0.040*** 0.040*** 0.040*** 0.047*** 0.047*** 0.047***
(9.55) (9.46) (9.50) (11.17) (11.14) (11.16)
Market-to-Book Assets -0.011*** -0.011*** -0.011*** 0.009*** 0.010*** 0.010***
(-13.34) (-12.45) (-12.46) (2.89) (2.95) (2.96)
Return on Assets -0.215*** -0.212*** -0.213*** -0.113*** -0.114*** -0.115***
(-21.90) (-19.93) (-20.19) (-7.63) (-7.25) (-7.31)
Fixed Assets 0.665*** 0.677*** 0.677*** 0.429*** 0.435*** 0.435***
(17.18) (17.66) (17.68) (27.32) (27.37) (27.35)
Cash Flow Volatility 0.004 -0.001 -0.000 0.004 0.002 0.003
(0.24) (-0.03) (-0.02) (0.40) (0.24) (0.26)
Dividend Payer -0.056*** -0.057*** -0.057*** -0.049*** -0.049*** -0.049***
(-12.96) (-13.36) (-13.37) (-12.63) (-12.84) (-12.85)
Strength of CNCs 0.002 0.002 0.002 0.002 0.002 0.002
(0.75) (0.68) (0.69) (0.77) (0.74) (0.74)
State GDP Growth -0.019 -0.023 -0.023 -0.088** -0.092*** -0.092***
(-0.45) (-0.57) (-0.55) (-2.54) (-2.71) (-2.70)
Political Balance -0.012 -0.011 -0.011 -0.017* -0.017* -0.017*
(-0.87) (-0.81) (-0.81) (-1.86) (-1.87) (-1.87)
R&D / Sales -0.040*** -0.017***
(-13.32) (-9.91)
(R&D + Capex) / Sales -0.026*** -0.013***
(-10.18) (-12.37)
(R&D + Capex + Adv) / Sales -0.026*** -0.014***
(-10.25) (-12.65)

Industry × Year Fixed Effects Yes Yes Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes Yes Yes
Observations 125,895 124,674 124,674 125,895 124,674 124,674
Adjusted R2 0.742 0.742 0.742 0.707 0.708 0.708

26
Table A12
Recognition of the IDD and the cost of bank debt.
The sample used in the cost of bank debt tests includes all firms with non-missing data for the period 1987-
2011. Models 1 and 2 report the results from OLS regressions of Log Total Cost of Borrowing on the indicator
for the recognition of the IDD. Log Total Cost of Borrowing is from Berg, Saunders, and Steffen (2016) and is
the natural logarithm of a measure of total borrowing costs (in basis points) that accounts for the expected
credit spread and loan contract fees. IDD is an indicator variable equal to one if the firm is headquartered in a
state whose courts recognize the IDD, and zero otherwise. Log Loan Maturity is the natural logarithm of the
number of months until the loan matures. Log Loan Size is the natural logarithm of the loan amount (in
millions). Other control variables are defined in Tables 2 and 3. All models include loan-type fixed effects for
each loan type (defined as in Campello, Lin, and Zou (2011)); the categories are term loan, revolver greater
than one year, revolver shorter than one year, and 364-day facility. Industry fixed effects are defined at the
three-digit SIC level. Continuous variables, except state-level variables, are winsorized at their 1st and 99th
percentiles. Standard errors are corrected for heteroskedasticity and clustering at the state level (t-statistics
are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively

Log Total Cost of Borrowing

(1) (2)
IDD 0.014 -0.031**
(0.61) (-2.27)
Log Book Assets -0.142***
(-15.99)
Market-to-Book Assets -0.098***
(-11.16)
Return on Assets -1.051***
(-8.93)
Fixed Assets -0.126**
(-2.44)
Cash Flow Volatility 0.778***
(5.02)
Dividend Payer -0.287***
(-22.00)
Book Leverage 1.085***
(39.33)
Log Loan Maturity -0.140***
(-14.83)
Log Loan Size -0.049***
(-7.29)
Strength of CNCs -0.024**
(-2.14)
State GDP Growth 0.005
(0.02)
Political Balance 0.061
(0.96)

Industry × Year Fixed Effects Yes Yes

State Fixed Effects Yes Yes
Loan-Type Fixed Effects Yes Yes
Observations 18,262 18,262
Adjusted R2 0.575 0.771

27
Table A13
CARs to announcement of the adoption of the IDD.
This table reports the cumulative abnormal returns (CARs) surrounding the announcement that a state
court adopts the IDD for firms headquartered in adopting states. The sample spans the 1977-2011 period. The
CARs are calculated over the event window [-1,3] and pre-event window [-31,-2], where t=0 is the date the court
adopts the IDD. The sample used in columns 1 and 2 includes all available observations, and the sample used
in columns 3 and 4 excludes all firms with an earnings or distribution announcement during the ±5 trading
days around the announcement of the adoption of the IDD in their state. In columns 1 and 3, CARs are
calculated from the market model using CRSP value-weighted market returns. In columns 2 and 4, CARs are
calculated using the 4-factor model, in which firm returns are regressed on CRSP value-weighted market
returns as well as the returns to zero-investment long-short portfolios formed from small cap stocks minus
large cap stocks, high book-to-market stocks minus low book-to-market stocks, and high momentum stocks
minus low momentum stocks. The parameters for the market and 4-factor models are estimated over the
window [-280,-61] relative to the announcement date. CARs are winsorized at their 1st and 99th percentiles.
T-statistics reported in parentheses are corrected for cross-sectional correlation (i.e., event-day clustering)
following Kolari and Pynnönen (2010). *, **, and *** denote significance at the 10%, 5%, and 1% levels,
respectively.

Sample Includes All Firms Sample Excludes Confounding Events

(Obs. = 1,877) (Obs. = 1,549)
CAR Window Market Model CARs 4-Factor CARs Market Model CARs 4-Factor CARs
(1) (2) (3) (4)

[-1,3] 0.702%** 0.551%** 0.704%** 0.524%**

(2.54) (2.48) (2.41) (2.20)
[-31,-2] -0.435% 0.156% -0.733% 0.067%
(-0.80) (0.38) (-1.24) (0.20)

28
Table A14
Recognition of the IDD and product fluidity.
This table reports results from OLS regressions of product fluidity on the indicator for the recognition of
the IDD. The dependent variable is the product fluidity measure of Hoberg, Phillips, and Prabhala (2014). The
sample spans the 1997-2011 period. IDD is an indicator variable equal to one if the firm is headquartered in a
state whose courts recognize the IDD, and zero otherwise. The control variables are defined in Tables 2 and 3.
Industry fixed effects are defined at the three-digit SIC level. Continuous variables, except state-level variables,
are winsorized at their 1st and 99th percentiles. Standard errors are corrected for heteroskedasticity and
clustering at the state level (t-statistics are in parentheses). *, **, and *** denote significance at the 10%, 5%,
and 1% levels, respectively.

Product Fluidity
(1) (2)
IDD 0.087 0.095
(1.35) (1.46)
Log Book Assets 0.410***
(11.35)
Market-to-Book Assets 0.034***
(2.76)
Return on Assets -0.657***
(-8.80)
Fixed Assets 0.328
(1.52)
Cash Flow Volatility 0.603***
(3.59)
Dividend Payer -0.206***
(-3.59)
Strength of CNCs -0.034
(-1.22)
State GDP Growth 0.287
(0.55)
Political Balance -0.096
(-0.62)

Industry × Year Fixed Effects Yes Yes

Firm Fixed Effects Yes Yes
Observations 55,579 55,579
Adjusted R2 0.813 0.817

29
Table A15
Recognition of the IDD and other firm-level measures of competitive risk.
This table reports the results from OLS regressions of firm-level measures of competitive risk on the
indicator for the recognition of the IDD. The sample spans the 1977-2011 period. The dependent variable in
models 1 and 2 is abs(K-L Ratio Deviation), which is the absolute value of the difference between a firm’s
capital-labor ratio and the industry-year median capital-labor ratio of all other firms in its three-digit SIC
industry. This measure is scaled by the industry-year range of the capital-labor ratio. The dependent variable
in models 3 and 4 is Industry Beta, which is a measure that captures the correlation of a firm’s stock returns
with its industry stock returns. To calculate this measure, we regress a firm’s daily stock return on the daily
CRSP equally weighted market return and the daily equally weighted return for the index of all other firms in
the same three-digit SIC industry as the firm. We use daily returns over a firm’s fiscal year and require the
firm to have at least 180 observations during the fiscal year. Industry Beta is the regression coefficient on the
industry return index. The control variables are defined in Tables 2 and 3. Industry fixed effects are defined at
the three-digit SIC level. Continuous variables, except state-level variables, are winsorized at their 1st and
99th percentiles. Standard errors are corrected for heteroskedasticity and clustering at the state level (t-
statistics are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

abs(K-L Ratio Deviation) Industry Beta

(1) (2) (3) (4)
IDD 0.001 0.001 0.002 0.002
(0.14) (0.24) (0.18) (0.23)
Log Book Assets 0.009*** 0.068***
(4.29) (9.48)
Market-to-Book Assets -0.000 0.018***
(-0.47) (10.20)
Return on Assets -0.007 -0.011
(-0.92) (-0.60)
Fixed Assets 0.320*** 0.013
(12.33) (0.60)
Cash Flow Volatility 0.006 0.105***
(0.56) (3.18)
Dividend Payer -0.000 -0.029***
(-0.11) (-6.77)
Strength of CNCs 0.000 -0.004
(0.06) (-1.17)
State GDP Growth -0.046 0.060
(-1.35) (0.76)
Political Balance -0.007 -0.001
(-0.68) (-0.06)

Industry × Year Fixed Effects Yes Yes Yes Yes

Firm Fixed Effects Yes Yes Yes Yes
Observations 121,488 121,488 120,738 120,738
Adjusted R2 0.738 0.749 0.336 0.340

30
Table A16
Recognition of the IDD and the number of trade secrets legal cases.
This table reports results from OLS regressions of the number of legal cases involving trade secrets
litigated in a state on the indicator for the recognition of the IDD. The sample spans the 1977-2011 period and
contains a total of 1,750 state-year observations. As in Almeling et al. (2011), for each state, the number of
cases includes all cases litigated in state courts, district courts, and in the federal court of appeals in the circuit
to which the state belongs for which the text of the legal ruling available in LexisNexis contains the phrase
“trade secret(s)” at least three times. IDD is an indicator variable equal to one if the firm is headquartered in
a state whose courts recognize the IDD, and zero otherwise. The control variables are defined in Table 2.
Standard errors are corrected for heteroskedasticity and clustering at the state level (t-statistics are in
parentheses). *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

Log(# of Trade Secrets Cases)

(1) (2)
IDD 0.200** 0.215**
(2.20) (2.41)
Strength of CNCs 0.031
(0.70)
State GDP Growth -0.322
(-1.00)
Political Balance -0.099
(-1.08)

Year Fixed Effects Yes Yes

State Fixed Effects Yes Yes
Observations 1,750 1,750
Adjusted R2 0.797 0.797

31
Table A17
Recognition of the IDD and capital structure: Controlling for governance.
This table reports results from OLS regressions of net financial leverage (Net Book Leverage in models 1-3
and Net Market Leverage in models 4-6) on the indicator for the recognition of the IDD. The sample spans the
1977-2011 period in models 1 and 4 and the 1980-2011 period in models 2-3 and 5-6. IDD is an indicator variable
equal to one if the firm is headquartered in a state whose courts recognize the IDD, and zero otherwise. Business
Combination Laws is an indicator variable equal to one if the firm is incorporated in a state that has passed
business combination laws, and zero otherwise. Inst. Ownership Top 5 is the number of a firm’s shares held by
the five largest institutional investors divided by the firm’s total institutional ownership. Inst. Ownership HHI
is the Herfindahl-Hirschman Index based on the percentages of institutional holdings by all institutional
investors. The other control variables are defined in Tables 2 and 3. Industry fixed effects are defined at the
three-digit SIC level. Continuous variables, except state-level variables, are winsorized at their 1st and 99th
percentiles. Standard errors are corrected for heteroskedasticity and clustering at the state level (t-statistics
are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

Net Book Leverage Net Market Leverage

(1) (2) (3) (4) (5) (6)
IDD 0.013*** 0.011** 0.011** 0.013*** 0.011*** 0.011***
(2.78) (2.29) (2.34) (3.36) (3.06) (3.08)
Log Book Assets 0.039*** 0.046*** 0.045*** 0.046*** 0.052*** 0.051***
(8.92) (9.54) (9.74) (10.90) (10.85) (10.89)
Market-to-Book Assets -0.012*** -0.010*** -0.011*** 0.009*** 0.011*** 0.010***
(-14.14) (-11.22) (-12.14) (2.80) (3.45) (3.21)
Return on Assets -0.176*** -0.171*** -0.173*** -0.095*** -0.088*** -0.090***
(-15.60) (-15.49) (-15.77) (-6.12) (-5.84) (-5.99)
Fixed Assets 0.667*** 0.679*** 0.680*** 0.425*** 0.429*** 0.430***
(16.03) (15.67) (15.66) (26.54) (24.87) (24.85)
Cash Flow Volatility 0.001 -0.002 -0.005 0.002 0.004 0.002
(0.06) (-0.10) (-0.27) (0.17) (0.38) (0.16)
Dividend Payer -0.056*** -0.053*** -0.053*** -0.049*** -0.045*** -0.045***
(-11.61) (-11.82) (-11.64) (-10.92) (-11.55) (-11.34)
Strength of CNCs 0.003 0.002 0.002 0.002 0.001 0.001
(0.83) (0.58) (0.56) (0.81) (0.40) (0.35)
State GDP Growth -0.016 -0.034 -0.039 -0.080** -0.097*** -0.102***
(-0.38) (-0.81) (-0.92) (-2.26) (-2.86) (-3.12)
Political Balance -0.010 -0.009 -0.009 -0.017* -0.016* -0.017*
(-0.76) (-0.64) (-0.70) (-1.77) (-1.78) (-1.87)
Business Combination Laws 0.018** 0.006
(2.45) (0.96)
Inst. Ownership Top 5 0.075*** 0.066***
(7.19) (7.44)
Inst. Ownership HHI 0.078*** 0.060***
(10.52) (10.31)

Industry × Year Fixed Effects Yes Yes Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes Yes Yes
Observations 119,270 118,262 118,262 119,270 118,262 118,262
Adjusted R2 0.736 0.743 0.792 0.701 0.710 0.710

32
Table A18
The effect of measurement error in the IDD indicator.
This table reports results from OLS regressions of Net Book Leverage and Net Market Leverage on the indicator for the recognition of the IDD in the
state where a firm is headquartered defined in Table 3. In model 1, we correct the location of headquarters (HQ) over the 1992-2011 period to reduce
measurement error in the IDD indicator. To do so, we use the state of headquarters information from 10-K filings over the 1992-2011 period when it is
available, and when not available, we assume there were no relocations prior to the earliest date when headquarters information is available. For models
2-4, the sample spans the 1977-2011 period. In model 2, we exclude firms whose annual sales or book asset growth exceeded 100% in any year during the
sample period. In model 3, we exclude all firm years with positive foreign income (pifo) or foreign taxes (txfo). In model 4, we exclude firms in geographically
dispersed industries. Following Agrawal and Matsa (2013), we define geographically dispersed industries as the retail, wholesale, and transportation
industries. Industry fixed effects are defined at the three-digit SIC level. Standard errors are corrected for heteroskedasticity and clustering at the state
level (t-statistics are in parentheses). *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

Panel A: Dependent variable is net book leverage

Exclude if firm growth ever Exclude if firm reports Exclude if firm is in
Corrected location of HQ
exceeds 100% foreign income or taxes dispersed industry
(1) (2) (3) (4)
IDD 0.013*** 0.015** 0.018** 0.011*
(4.26) (2.26) (2.29) (1.92)

Industry × Year Fixed Effects Yes Yes Yes Yes

Firm Fixed Effects Yes Yes Yes Yes
Observations 125,895 60,497 77,944 101,746
Adjusted R2 0.702 0.712 0.703 0.698

Panel B: Dependent variable is net market leverage

Exclude if firm growth ever Exclude if firm reports Exclude if firm is in

Corrected location of HQ
exceeds 100% foreign income or taxes dispersed industry
(1) (2) (3) (4)
IDD 0.012*** 0.013** 0.016*** 0.010**
(3.80) (2.32) (3.19) (2.40)

Industry × Year Fixed Effects Yes Yes Yes Yes

Firm Fixed Effects Yes Yes Yes Yes
Observations 125,895 60,497 77,944 101,746
Adjusted R2 0.702 0.712 0.703 0.698

33
Table A19
IDD and net financial leverage using a propensity score matched sample.
This table examines the impact of adoptions of the IDD on firms’ net financial leverage using a propensity
score matched sample and using data over the window +/- 5 years around the adoption of the IDD. The sample
spans the 1977-2011 period. The treatment group is firms headquartered in states that adopt the IDD, and the
control group is firms headquartered in states that never adopt the IDD. We require firms in the treatment
and control groups to have at least one observation in the pre- and post-period (5 years before and after the
adoption of the IDD). Using a probit regression and data in years t-1 before the adoption of the IDD, we estimate
the probability (i.e., propensity score) of being in the treatment group using Log Book Assets, Market-to-Book
Assets, and Return on Assets (defined in Table 3). We then match each treatment firm in year t-1 to at least one
control firm and at most two control firms (with replacement), first matching exactly on year and two-digit SIC
industry, and then on the closest propensity scores (with maximum difference between propensity scores of
0.01). Panel A tabulates the means of the matching variables and propensity scores for the treatment and
control groups (the differences across the two groups are not statistically significant at the 10% significance
level). Panel B presents results of tests examining the impact of the adoption of the IDD on firms’ net financial
leverage. In models 1 and 2, the dependent variable is Net Book Leverage. In models 3 and 4, the dependent
variable is Net Market Leverage. IDD Adoption is defined in Table A5. IDD Adoption-3, IDD Adoption-2, IDD
Adoption-1, IDD Adoption0, IDD Adoption+1, IDD Adoption+2, IDD Adoption3+ are defined in Table 4. Industry
fixed effects are defined at the three-digit SIC level. Standard errors in Panel B are corrected for
heteroskedasticity and clustering at the state level (t-statistics are in parentheses). *, **, and *** denote
significance at the 10%, 5%, and 1% levels, respectively.

Panel A: Comparison of means across matched samples in year t-1

Matched Sample for Adoptions

Treatment Group Control Group
(Obs. = 1,552) (Obs. = 3,030)
Propensity Score 0.114 0.114
Log Book Assets 5.254 5.164
Market-to-Book Assets 1.774 1.728
Return on Assets 0.101 0.103

34
Table A19
Continued.

Panel B: Adoption of the IDD and financial leverage

Net Book Leverage Net Market Leverage
(1) (2) (3) (4)
IDD Adoption 0.028*** 0.020***
(3.83) (3.73)
IDD Adoption -3 -0.003 -0.002
(-0.57) (-0.52)
IDD Adoption -2 0.006 0.004
(0.69) (0.73)
IDD Adoption -1 0.011 0.010
(1.23) (1.42)
IDD Adoption 0 0.015 0.012
(1.36) (1.47)
IDD Adoption +1 0.032*** 0.024***
(3.11) (2.97)
IDD Adoption +2 0.034*** 0.025***
(2.70) (3.29)
IDD Adoption 3+ 0.042*** 0.031***
(2.94) (3.23)

Industry × Year Fixed Effects Yes Yes Yes Yes

Firm Fixed Effects Yes Yes Yes Yes
Observations 42,178 42,178 42,178 42,178
Adjusted R2 0.751 0.751 0.724 0.724

35
Table A20
Litigation involving trade secrets and debt issuance.
This table reports results from OLS regressions of changes in debt on dummy variables indicating whether a firm is a plaintiff in a trade secrets case
during a given year. Trade secrets cases are identified from LexisNexis. The sample spans the 1992-2011 period and is made up of firms included in the
study’s main sample. The dependent variables are (Δ Net Debt)t / Book Assetst-1 (models 1 and 2), (Δ Net Debt)t / Market Value Assetst-1 (models 3 and 4),
(Δ Total Debt)t / Book Assetst-1 (models 5 and 6), and (Δ Total Debt)t / Market Value Assetst-1 (models 7 and 8). (Δ Net Debt)t / Book Assetst-1 is the change
in total debt minus cash holdings from year t-1 to t [(dlttt+dlct-chet) - (dlttt-1+dlct-1-chet-1)] divided by book value of assets at the beginning of the year (att-
1). (Δ Net Debt)t / Market Value Assetst-1 is the change in total debt minus cash holdings from year t-1 to t [(dlttt+dlct-chet) - (dlttt-1+dlct-1-chet-1)] divided by
market value of assets at the beginning of the year (prcc_ft-1*cshot-1+att-1-ceqt-1). (Δ Total Debt)t / Book Assetst-1 is the change in total debt from year t-1 to
t [(dlttt+dlct) - (dlttt-1+dlct-1)] divided by book value of assets at the beginning of the year (att-1). (Δ Total Debt)t / Market Value Assetst-1 is the change in
total debt from year t-1 to t [(dlttt+dlct) - (dlttt-1+dlct-1)] divided by market value of assets at the beginning of the year (prcc_ft-1*cshot-1+att-1-ceqt-1). Trade
Secret Caset-1 is an indicator variable that is set to one if the firm will be a plaintiff in a trade secrets case in the following year and zero otherwise. Trade
Secret Caset is an indicator variable that is set to one if the firm is the plaintiff in a trade secrets case in the current year and zero otherwise. Trade Secret
Caset+1 is an indicator variable that is set to one if the firm was the plaintiff in a trade secrets case in the prior year and zero otherwise. Control variables
are defined in Table 3. Industry fixed effects are defined at the three-digit SIC level. Continuous variables are winsorized at their 1st and 99th percentiles.
Standard errors are corrected for heteroskedasticity and clustering at the firm level (t-statistics are in parentheses). *, **, and *** denote significance at
the 10%, 5%, and 1% levels, respectively.

(Δ Net Debt)t / (Δ Net Debt)t / Market (Δ Total Debt)t / (Δ Total Debt)t / Market
Book Assetst-1 Value Assetst-1 Book Assetst-1 Value Assetst-1
(1) (2) (3) (4) (5) (6) (7) (8)
Trade Secret Case t-1 -0.010 -0.010 -0.001 0.000 -0.002 -0.001 0.001 0.002
(-1.13) (-1.03) (-0.13) (0.01) (-0.32) (-0.17) (0.15) (0.50)
Trade Secret Case t 0.023*** 0.019** 0.008* 0.008* 0.013* 0.018*** 0.008** 0.011***
(2.69) (2.17) (1.77) (1.66) (1.91) (2.66) (2.07) (2.77)
Trade Secret Case t+1 0.008 0.002 0.002 0.002 -0.005 0.001 -0.004 -0.001
(0.99) (0.30) (0.51) (0.35) (-0.74) (0.23) (-1.23) (-0.37)
Log Book Assets t-1 0.019*** -0.005*** -0.050*** -0.029***
(6.14) (-3.27) (-24.08) (-25.49)
Market-to-Book Assets t-1 -0.022*** -0.007*** 0.009*** -0.000
(-14.25) (-12.96) (11.27) (-0.16)
Return on Assets t-1 0.052*** 0.013** 0.049*** 0.042***
(3.68) (2.18) (6.19) (11.24)
Fixed Assets t-1 -0.297*** -0.149*** -0.030** -0.021***
(-14.97) (-14.63) (-2.14) (-2.69)
Cash Flow Volatility t-1 -0.066*** -0.020** -0.019 -0.008
(-3.26) (-2.33) (-1.61) (-1.42)
Dividend Payer t-1 0.031*** 0.021*** 0.027*** 0.017***
(7.69) (8.78) (8.53) (8.72)
Industry × Year Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes
Observations 71,191 71,191 71,191 71,191 71,191 71,191 71,191 71,191
Adjusted R2 0.027 0.054 0.052 0.063 0.064 0.094 0.080 0.102

36