Upload
vokhue
View
225
Download
0
Embed Size (px)
Citation preview
0
Protection of Trade Secrets and Capital Structure Decisions*
Sandy Klasa† , Hernán Ortiz-Molina‡ , Matthew A. Serfling , and Shweta Srinivasan$
ABSTRACT
We study whether a firm’s capital structure decisions are affected by the risk that its rivals in product markets could gain access to its non-patented “trade secrets.” Our tests exploit the staggered recognition of the Inevitable Disclosure Doctrine (IDD) by U.S. state courts as an exogenous event that increases the protection of a firm’s trade secrets by reducing the mobility of workers with knowledge of the trade secrets to similar jobs at rival firms. We find that firms increase their financial leverage following the recognition of the IDD, especially those in more competitive industries, those whose workers are more likely to know trade secrets, and those that face greater ex-ante risk of losing key employees to rivals. Also, the credit spread on firms’ bank loans decreases after the recognition of the IDD, implying that firms’ default risk is reduced. Finally, the recognition of the IDD is accompanied by gains in market share for firms in recognizing states. Overall, our findings imply that the risk of losing intellectual property to rivals affects a firm’s default risk, and through this channel, it has a significant impact on the firm’s financing decisions.
Keywords: capital structure; trade secrets; intellectual property; competitive risk.
* We thank Jan Bena, Lorenzo Garlappi, Ron Giammarino, Kai Li, Neslihan Ozkan, Carolin Pflueger, Elena Simintzi, Ryan Williams,tand seminar participants at the University of Arizona, the University of British Columbia, and the One-Day Corporate Finance Conference (2013) at the University of Manchester for helpful suggestions. We are grateful to Andrew Bird and Robert Knopf for sharing with us their state-level data on the strength of enforcement of non-compete agreements. † Eller College of Management, University of Arizona, Tucson, AZ 85721, [email protected]. ‡ Sauder School of Business, University of British Columbia, Vancouver, BC V6T1Z2, [email protected]. Eller College of Management, University of Arizona, Tucson, AZ 85721, [email protected]. $ Eller College of Management, University of Arizona, Tucson, AZ 85721, [email protected].
1
Financial economists generally agree that risks stemming from a firm’s competitive
environment, such as the risk of predation by rivals, can affect its capital structure decisions.
Surprisingly, little is known about the relevance for capital structure choices of competitive risks
associated with the protection of a firm’s intellectual property. Yet, intellectual property, which
accounts for roughly one-third of the market equity value of all American publicly traded firms1,
is among the critical assets that determine a firm’s competitive position and performance within
its product market and thus its ability to generate the cash flow needed to service debt payments.
In this paper, we study how a firm’s capital structure decisions are affected by the risk that
its industry rivals could gain access to its intellectual property in the form of non-patented “trade
secrets”. We hypothesize that a firm’s optimal financial leverage is higher when it faces less risk
that its rivals in product markets could obtain its trade secrets and, in doing so, harm its
competitive position and reduce its ability to service the payments associated with its debt
obligations. Our empirical tests based on exogenous variation in this risk support our hypothesis
and suggest that this risk has an economically important impact on capital structure decisions.
Trade secrets allow a firm to maintain an economic advantage over its competitors and are
pervasive in industrial activity. They consist of formulas, practices, processes, designs,
instruments, patterns, or any compilations of information which are not generally known or
easily ascertainable by outside parties such as customer lists, price lists, cost information, and
future business plans. Trade secrets are not protected by patents, because they are not patentable
(e.g., customer lists or business plans) or because revealing them to the public as required by the
patenting process is too costly. Hence, keeping their trade secrets confidential is a major concern
for many firms because the divulgation of the trade secrets can quickly erode a firm’s
competitive advantage in its product market and allow its rivals to appropriate some of the future
cash flows generated by these secrets, causing significant economic harm to the firm.
1 Shapiro, R., and K. Hassett, “The Economic Value of Intellectual Property,” USA for Innovation, October 2005.
2
The inherent mobility of key employees with knowledge of the firm’s trade secrets is a key
determinant of the risk of trade secret divulgation, since employees naturally carry such
knowledge with themselves when they leave the firm to work for a rival. Supporting this view,
Almeling et al. (2010) report that in most legal cases involving trade secrets the misappropriator
of a firm’s trade secrets is a former employee or someone related to the firm. A rival firm might
seek to gain access to the firm’s trade secrets by hiring away its employees with knowledge of
the trade secrets. Alternatively, the rival might hire away such employees to perform tasks
similar to those performed with the prior employer but with no explicit intent of obtaining the
former employer’s trade secrets. However, in performing their duties, the employees might
inevitably use their knowledge of their prior employer’s trade secrets to benefit their new
employer. Last, employees with knowledge of a firm’s trade secrets might leave the firm and use
them to start a new firm whose business directly competes with that of their prior employer.
Our empirical tests use a difference-in-differences design based on the staggered recognition
of the Inevitable Disclosure Doctrine (IDD) by U.S. state courts in 16 states over the 1977-2011
period. The recognition of the IDD represents an exogenous increase in a firm’s ability to
prevent its employees with knowledge of trade secrets from working at rival firms, and thus a
decrease in the risk that such employees will disclose its trade secrets to its rivals. The IDD
maintains that, when an employee has knowledge of a firm’s trade secrets, the employee can be
prevented from beginning employment at a rival firm if it would “inevitably” lead the employee
to disclose the prior employer’s trade secrets to the new employer. The IDD is applicable even if
the employee did not sign a non-compete agreement, there is no evidence of bad faith or actual
wrongdoing, and the rival firm is in a different state. Hence, a firm’s lawsuit seeking to prevent
an employee from beginning employment at a rival firm in order to protect its trade secrets is
much more likely to succeed when courts in the firm’s state recognize the IDD. Consistent with
this assertion, several studies and anecdotal evidence show that the recognition of the IDD makes
it difficult for a firm’s employees to leave their firm to go work for a rival firm. For example,
3
Png (2012) finds that in states that recognize this doctrine, there is a decrease in the mobility of
workers who are likely to have knowledge of their employer’s trade secrets.
We implement a difference-in-differences approach in a regression framework, which
compares changes in the financial leverage of firms headquartered in states that recognize the
IDD to changes in the financial leverage of firms headquartered in states that do not. We rely on
state-by-state analyses of case law involving trade secrets in prior legal research and in our own
reading of specific legal cases to identify the timing of changes in state courts’ position regarding
the IDD. We include firm and year fixed effects to control for time-invariant firm-level
unobserved factors and for secular trends in financial leverage. In addition to standard control
variables used in capital structure tests, we include a state’s GDP growth rate and a measure of
the state’s political leaning to control for underlying economic or political conditions in the state
that might affect both courts’ recognition of the IDD and firms’ capital structure decisions.
Our key result is that, following the recognition of the IDD by a state court, those firms
headquartered in the state increase their book value and market value leverage ratios by about
6.1% relative to their respective sample means. These findings are robust to using alternative
measures of financial leverage, including measures based only on long-term debt and measures
that subtract cash holdings from debt obligations. An examination of the timing of capital
structure changes supports a causal interpretation of our results, as we find that the increase in
financial leverage occurs after the recognition of the IDD but not before. Validating our
difference-in-difference approach, we also find that the capital structures of firms in recognizing
states and in non-recognizing states follow parallel trends before the recognition of the IDD.
Overall, these initial results support the view that a lower risk of losing valuable trade secrets to
product market rivals leads to an increase in a firm’s debt capacity and financial leverage.
To better understand the economic mechanism behind our results, we next study the cross-
sectional variation in the impact the recognition of the IDD has on capital structure. First, we
consider the extent of competition the firm faces in product markets. The financial distress a firm
would suffer upon the divulgation of its trade secrets is arguably more severe and more likely to
4
occur for a firm that operates in a more competitive industry. The reason is that firms in such
industries typically have lower operating margins and survival rates, and as a result, they are
typically more affected by adverse shocks. Thus, the recognition of the IDD should have a
stronger impact on the capital structures of firms operating in more competitive environments.
We gauge competitive pressure using the Census Bureau’s four-firm industry concentration ratio
and industry-average R&D plus advertising expenses, which Hoberg and Phillips (2011) show
creates important barriers to entry. We find that the recognition of the IDD causes an increase in
leverage ratios only for firms in industries with more intense competition, as measured by
concentration ratios or barriers to entry below the sample median.
Second, workers in managerial and science occupations and more educated workers are
more likely to know their firm’s trade secrets. Hence, firms that employ more of these workers
are more exposed to the risk that their product market rivals might gain access to their trade
secrets by poaching some of their workers. This implies that the recognition of the IDD should
have a larger impact on the capital structure of firms that employ more of these workers.
Consistent with this prediction, we find that the impact of the IDD on financial leverage exists
only for firms operating in industries for which the fraction of workers in managerial
occupations, in science occupations, or with at least a bachelor’s degree are above the sample
medians for these fractions.
Third, the recognition of the IDD should have a larger impact on firms’ capital structures
when firms face a higher ex-ante risk that their workers will accept jobs at rival firms. To test
this prediction, we gauge this ex-ante risk in two ways. First, the risk is smaller for firms with
defined benefit pension plans since retirement benefits from these plans tend to be less portable
and thus create incentives for workers to remain with the firm. Second, the risk is greater if the
firm’s rivals are located close to the firm and employ more workers compared to the firm,
because the firm’s workers are more likely to find employment at a rival firm that is close to
their current job and hence have a smaller cost of switching employers. Consistent with our
prediction, we find that the positive impact of the recognition of the IDD on firms’ debt ratios
5
only exists for firms that do not offer their employees defined benefit pension plans and when the
share of a firm’s own workers of the total number of workers in its state that work in its industry
is below the sample median.
We argue that an increase in the legal protection of a firm’s trade secrets increases the firm’s
desired leverage ratio because it reduces the probability of financial distress. Hence, we also
study whether the recognition of the IDD reduces a firm’s default risk measured by credit
spreads on bank loans, which are the main source of debt financing for most firms and can be
measured for a large sample. We find that the recognition of the IDD in a firm’s state is
associated with a 5.4% decrease in credit spreads. Supporting a causal interpretation of this
result, credit spreads decrease only after and not before state courts recognize the IDD. This
evidence implies that the increase in a firm’s leverage ratio following the recognition of the IDD
by state courts occurs because better protection of trade secrets reduces the firm’s risk of default.
We also examine whether the recognition of the IDD affects firms’ investment behavior. We
consider this issue because, by increasing the protection of trade secrets, the recognition of the
IDD could increase the marginal benefit of investment and thus increase firms’ demand of
external financing to fund investment. In that case, if firms were to fund the additional
investment with more debt financing, then the increases in financial leverage that we observe
could be caused by increased financing needs rather than by increased debt capacity as implied
by our main hypothesis. However, we do not find that the recognition of the IDD is associated
with subsequent increases in capital expenditures, acquisition activity, R&D expenses, or
advertising costs.
The trade secrets of firms located in states whose courts recognize the IDD are better
protected than those of their industry rivals in non-recognizing states. This might give a
competitive edge to firms headquartered in states with better protection. Hence, we explore
whether better protection of trade secrets boosts firms’ performance in product markets.
Suggesting that the IDD has competitive effects in product markets, we find that the recognition
6
of the IDD leads to subsequent moderate gains in market share for those firms headquartered in
recognizing states.
Last, we examine whether a firm’s market share and capital structure are affected when the
recognition of the IDD in states other than the firm’s state increases the protection of the trade
secrets of its out-of-state industry rivals. We find no impact on the firm’s market share or book
leverage, but some evidence of a negative effect on market leverage. This evidence suggests that
the indirect effect on a firm’s market share and capital structure caused by an increase in the
protection of the trade secrets of its out-of-state industry rivals is less important than the direct
effect associated with an increase in the protection of trade secrets in the firm’s own state.
Our main contribution is to show that the competitive risk associated with firms’ inability to
protect valuable trade secrets from appropriation by rivals significantly affects financial leverage.
As such, we increase the understanding of capital structure decisions (see Leary and Roberts
(2005) and Lemmon, Roberts, and Zender (2008) for recent papers and Harris and Raviv (1991)
or Frank and Goyal (2007) for surveys). In particular, prior work relates firms’ financial policies
to competition in product markets, mainly based on the idea that financially strong firms could
prey on financially weak rivals (Telser (1966) and Bolton and Scharfstein (1990)). Several
empirical studies highlight capital structure and product market interactions and show that
predation risk shapes financial policies (e.g., Phillips (1995), Chevalier (1995), Kovenock and
Phillips (2007), Campello (2003, 2006), MacKay and Phillips (2005), Lyandres (2006),
Haushalter, Klasa, and Maxwell (2007), Frésard (2010), Valta (2012), and Hoberg, Phillips, and
Prabhala (2013), among others). Noteworthy, the competitive risk behind our results is that rivals
might gain access to a firm’s trade secrets and not the presence of rivals with “deep pockets”.
Our paper is also related to recent work that shows how labor markets affect capital structure
decisions (e.g., Matsa (2010), Simintzi, Vig, and Volpin (2012), Kim (2012), and Agrawal and
Matsa (2013), among others). These studies show that financial leverage can depend on strategic
issues that arise in bargaining with labor unions, the rigidity of labor costs, the size of labor
markets, and employee unemployment risk. Although our focus - the protection of trade secrets -
7
is different, our work is related to these studies because the recognition of the IDD that we use in
our empirical tests increases the protection of a firm’s trade secrets by reducing the mobility of
the firm’s key employees to rival firms. Since workers with knowledge of trade secrets typically
do not account for a large fraction of a firm’s total labor costs, their mobility is unlikely to have a
large effect on capital structure solely through labor-related mechanisms like those outlined
above.2 However, our evidence suggests that the mobility of such workers can still have a
significant impact on a firm’s capital structure by affecting the protection of its intellectual
property.
The rest of the paper is organized as follows. Section I develops our main hypothesis and
related predictions. Section II discusses the IDD and how we identify its recognition by state
courts. Section III describes our data, variables, and empirical methodology. Section IV reports
our main empirical results. Section V reports robustness tests. Section VI concludes.
I. State-Level Protection of Trade Secrets and Capital Structure
Trade secrets are important strategic assets in generating cash flow and often give firms a
valuable economic advantage over their industry rivals. However, they are only guarded by firms
primarily through their secrecy. Hence, a firm is exposed to the risk that industry rivals might
gain access to its trade secrets, harm its competitive position, and adversely affect its cash flow.
Since the protection of trade secrets is governed by state law (see Section II for more detail), the
likelihood of losing trade secrets to rivals depends to an important extent on the degree of
protection of a firm’s trade secrets afforded by the legal system in the firm’s state.
We argue that an increase in the protection of trade secrets in a firm’s state is likely to
increase the firm’s optimal financial leverage by reducing the probability of financial distress.3 In
essence, the key mechanism is that a lower risk that industry rivals could gain access to the
firm’s trade secrets and hurt its business can both increase the average value and reduce the
2 In Section IV.F we discuss why pure labor-related stories are unlikely to explain our results. 3 In Section V.B we discuss and empirically examine the possibility that a firm’s capital structure might also be affected by changes in the legal protection of trade secrets in states that harbor the firm’s industry rivals.
8
variability of the firm’s future cash flow, ultimately reducing the probability that the firm will
not have enough cash flow to meet its debt payments. Below, we discuss in more detail these two
intertwined effects on the probability of financial distress, which go in the same direction.
To understand these effects, note that in principle any given firm’s cash flow can both be
harmed by trade secret losses to rivals or boosted by access to the trade secrets of rivals and that
the recognition of the IDD provides increased protection of trade secrets to all the firms in an
affected state but not to firms in other states. Thus, increased legal protection of trade secrets in a
firm’s state reduces the probability that the firm will lose trade secrets to rivals in any state and
the probability that it will gain access to the trade secrets of rivals in the same state, but not the
probability that it will gain access to the trade secrets of rivals in other states. Intuitively, this
implies that while better protection of trade secrets reduces both the left-tail and right-tail risk of
the distribution of future cash flows, the reduction in the left-tail risk is likely to be more
significant. Hence, an increase in the protection of trade secrets in a firm’s state is likely to
increase the firm’s average future cash flows and reduce the variability of these cash flows,
lowering the probability of financial distress.4 This leads to our main hypothesis relating the
protection of trade secrets and capital structure:
Hypothesis: An increase in the legal protection of trade secrets in a firm’s state increases
the firm’s optimal financial leverage by reducing the probability of financial distress.
This hypothesis gives rise to additional predictions that pertain to the cross-sectional
variation in the strength of the impact that better protection of trade secrets has on capital
structure along several important dimensions, including competition in product markets, the
characteristics of the labor force, costs of switching jobs, and competition in local labor markets.
First, better protection of trade secrets should have a larger impact on the capital structures
of firms in more competitive industries. Competitive industries are likely to exhibit higher
4 If better protection of trade secrets increases the firm’s profitability, then it could also provide the firm with additional incentives to increase its financial leverage by increasing the tax benefits of debt financing.
9
default risk since incumbents’ market positions are not protected by barriers to entry and firms
have low profit margins (e.g., Porter (1980)). Hence, the negative consequences for firms whose
trade secrets are divulged to competitors are expected to be more pronounced for firms in more
competitive industries, and therefore those firms are expected to benefit more from better
protection of their trade secrets.
Second, better protection of trade secrets should have a more important effect on the capital
structures of firms that employ more workers who, due to the nature of their jobs (e.g., managers,
engineers, and scientists), have access to their employers’ trade secrets. This is because firms
that employ more of these workers are more exposed to the risk that their industry rivals could
gain access to their trade secrets by poaching some of their workers. Hence, these firms are
expected to benefit more from better protection of their trade secrets.
Last, better protection of trade secrets should have a larger impact on the capital structure of
firms that face a higher ex-ante risk that their workers would accept a position with a rival firm.
Smaller job switching costs (e.g., in the form of geographical distance or portability of pensions)
increase the likelihood that workers will accept jobs at rival firms and thus increase the risk that
a firm’s key workers with knowledge of its trade secrets would go to work for a rival firm. This
risk is also higher when competition in local labor markets is more intense, for example, when
the firm is located in a state with a strong presence of rival firms that could poach its workers. In
this case a firm’s workers also have lower switching costs in taking a position with a rival firm.
To summarize, an increase in the legal protection of trade secrets in a firm’s state should
have a stronger positive effect on the firm’s financial leverage when the firm:
(a) operates in more competitive product markets.
(b) employs workers who are more likely to possess knowledge of its trade secrets.
(c) faces a greater risk that some of its key workers would take a position with a rival firm.
10
II. The Inevitable Disclosure Doctrine
A. Legal Background
Unlike most areas of intellectual property law, which are governed by federal statute, trade
secrets are governed by state law and there is wide variation across state courts in their treatment
of cases involving trade secrets. The IDD can be traced back to 1919 (see Eastman Kodak Co. v.
Powers Film Products, Inc.), but it has developed more recently in the broader context of trade
secret law. Trade secret law developed as common law, and thus, it did not follow universally
applicable principles until 1979. In that year, the National Conference of Commissioners on
Uniform State Laws issued the first Uniform Trade Secrets Act (UTSA), which codified the
existing common law and sought to promote uniformity of the legal treatment of these cases
across states.5 To date, 48 states and the District of Columbia have adopted laws based on the
principles outlined by the UTSA. The other two states, North Carolina and New York, have not
yet enacted laws based on the UTSA and continue to rely on case law. More generally, the state
laws do not encompass all aspects of civil trade secrets law. Consequently, in practice the
application of the law combines state law with case law and continues to evolve over time.
Trade secrets are any kind of confidential information that provides the firm with a
competitive advantage over the rest of the industry.6 Section 1(4) of the UTSA defines a trade
secret as information, including a formula, pattern, compilation, program, device, method,
technique, or process, 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. Section 1(2) of the UTSA
highlights that misappropriation occurs when the trade secret is acquired by (1) improper means
(e.g., theft or breach of a duty to maintain secrecy) or (2) disclosure without express or implied
consent by a person who acquired the trade secret under circumstances giving rise to a duty to
5 The Act was amended in 1985 to include some clarifications to the original act of 1979. 6 As noted in the introduction, trade secrets include not only things like formulas, recipes, or specific manufacturing processes, but also financial information, marketing information, customer lists, and future business plans.
11
maintain its secrecy or limit its use. Of particular importance is Section 2(a), which allows courts
to provide injunctive relief for “actual or threatened misappropriation” of trade secrets.
The term “threatened misappropriation” used in trade secret law is directly related to the
IDD.7 The problem of threatened misappropriation occurs when an employee who has acquired
knowledge of a firm’s trade secrets goes to work for a direct competitor in a similar position. The
IDD maintains that, if the new employment would inevitably lead to the disclosure of the firm’s
trade secrets to a competitor (i.e., the duties at the new employer cannot be performed without
making use of trade secrets acquired at the previous employer) and cause the firm irreparable
harm, then upon the firm’s request courts can prevent the employee from working for the firm’s
competitor or they can allow it but limit the responsibilities she can undertake.
The recognition of the IDD by state courts significantly enhances the protection of the trade
secrets of firms located in the recognizing state. This occurs because of a resulting reduction in
the risk faced by a firm that its departing employees could reveal its trade secrets to industry
rivals. Under the IDD, the plaintiff’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 are so similar to those she had at the firm
that working for the new employer 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 is not required to establish actual wrongdoing by the employee (disclosure,
misappropriation, or bad faith) or to disclose the actual details of the underlying trade secrets.
Employment contracts often contain a non-disclosure agreement (NDA) and/or a covenant
not to compete (CNC). By signing an NDA, the employee agrees not to use or disseminate the
firm’s confidential information, and under a CNC, the employee agrees not to enter into or start a
similar trade in competition against the firm. Both clauses, especially CNCs, are designed to
protect the firm’s trade secrets in cases in which employees wish to switch jobs or start
7 Some states that have not yet adopted the UTSA recognize the IDD, and states that have adopted the UTSA do not uniformly recognize the IDD (e.g., California adopted the UTSA but rejected the IDD).
12
competing firms. With these agreements in place, it is easier for the firm to seek injunctive relief
because it could bolster its suit by including a claim of breach of contract. However, the
enforcement of CNCs varies by legal jurisdiction and courts generally deem over-broad CNCs
unenforceable as a matter of public policy because they restrict employee mobility. Courts
enforce CNCs only when there are reasonable limitations as to the geographical area and time
period in which an employee of a company may not compete, and this shapes such contracts
(Malsberger (2004)).8
The IDD provides significant protection of trade secrets over and above any protection
firms’ might enjoy when their employees sign NDAs and CNCs. The reason is that it is much
broader in scope (e.g., unlike CNCs, the IDD does not explicitly entail specific geographic or
time restrictions) and allows the firm to bolster its suit seeking to prevent a former employee
from working for a rival firm, whether NDAs or CNCs were signed or not.
First, the IDD increases the enforceability of existing NDAs and CNCs. It allows the firm to
preempt the divulgation of its trade secrets in violation of the NDA before it occurs, which is
important because detecting and proving actual violation of NDAs takes time and is difficult. It
is also a powerful means of establishing an essential element in any legal action to enforce a
CNC, namely, that there is significant likelihood of irreparable harm to the firm if the employee
is allowed to work for the rival. State court’s application of the IDD in enforcing NDAs and
CNCs is illustrated in legal cases such as Procter & Gamble Co. v. Stoneham (Ohio, 2000),
Branson Ultrasonics Corp. v. Stratman (Connecticut, 1996), and Schlumberger Well Services v.
Blaker (Indiana 1985).
Second, the IDD serves as a powerful substitute for NDAs or CNCs when they do not exist.
It maintains that courts can grant an injunction if disclosure of the trade secrets is deemed
inevitable, even if the employee did not sign a NDA or CNC. State court’s application of the
IDD to protect a firm’s trade secrets in a situation in which the firm’s former employee was not
8 The scope of enforceable CNCs is often a state or a part of a state, for example, a county, a city or a 10 or 50 mile radius around the place of business (Malsberger, 2004).
13
bound by a CNC include PepsiCo, Inc. v. Redmond (Illinois, 1995), Air Products & Chemical,
Inc. v. Johnson (Pennsylvania, 1982), and Essex Group, Inc. v. Southwire Co. (Georgia, 1998).
Third, state courts can use the IDD to protect a firm’s trade secrets when some employees
with knowledge of its trade secrets wish to sever their employment with the firm and use the
firm’s trade secrets to set up a new company in direct competition with its operations. State
court’s application of the IDD to protect a firm’s trade secrets in this situation include Novell Inc.
v. Timpanogos Research Group Inc. (Utah, 1998), Surgidev Corp. v. Eye Technology, Inc.
(Minnesota. 1986), and Rugen v. Interactive Business Systems Inc. (Texas, 1993).
B. An Example of the Application of the IDD: PepsiCo, Inc. v. Redmond (Illinois, 1995)
Redmond was employed at the Pepsi Cola North America division of PepsiCo and was the
General Manager of the business unit covering all of California. As such, he had knowledge of
PepsiCo’s inside information and trade secrets, and had signed a NDA but not a CNC. After
being courted by a former PepsiCo executive who was then working for Quaker, Redmond
accepted a position at Quaker to serve as Vice President-Field Operations for “Gatorade”, a
product that was in direct competition with PepsiCo’s “All Sport” drink. PepsiCo immediately
filed suit seeking a temporary restraining order to enjoin Redmond from assuming his duties at
Quaker and to prevent him from disclosing trade secrets or confidential information to his new
employer. The trade secrets PepsiCo wanted to protect were mainly strategic sales, marketing,
logistics and financial information relating to the marketing of its drink “All Sport”.
The U.S. Court of Appeals for the 7th Circuit affirmed a preliminary injunction issued by an
Illinois court temporarily preventing Redmond from assuming any duties with Quaker relating to
beverage pricing, marketing, and distribution, and preventing him forever from divulging
PepsiCo trade secrets and confidential information to Quaker. The court noted that Redmond’s
departure from PepsiCo for a similar job at Quaker “lay against a backdrop of fierce beverage-
industry competition between Quaker and PepsiCo, especially in ‘sports drinks’ and ‘new age
drinks’.” The court maintained that “a plaintiff may prove a claim of trade secret
14
misappropriation by demonstrating that the defendant’s new employment will inevitably lead
him to rely on the plaintiff’s trade secrets.” In sum, the court ruling in favor of PepsiCo shows
that the IDD acted as a substitute for a non-existent CNC and also allowed PepsiCo to preempt a
likely future violation of the existing NDA should Redmond be allowed to work for Quaker.
C. Recognition of the Inevitable Disclosure Doctrine by State Courts and the IDD Index
To identify the timing of recognition of the applicability of the IDD by U.S. state courts, we
rely on state-by-state analyses of case law associated with trade secrets in legal research (e.g.,
Kahnke, Bundy, and Liebman (2008), Malsberger (2011), Png (2012a), Stevens (2001), Waldref
(2012), and Wiesner (2012)) and in our own reading of specific cases. Specifically, for each state
we identify the “precedent-setting case” in which a state court explicitly recognizes the
applicability of the doctrine in protecting trade secrets. We note that some court rulings explicitly
recognized the general applicability of the IDD but did not use it due to special circumstances.9
Other state courts have recognized the applicability of the IDD when the employee had signed an
NDA and/or CNC, but their position is unclear in the case such agreements do not exist. As in
prior legal research, we also code these cases as precedent-setting cases recognizing the IDD.
These precedent-setting cases are listed in Table I and described in the paper’s Appendix.
To build our IDD index, which captures the recognition of the IDD by U.S. state courts, we
assume that courts in a state are likely to use the IDD to protect firms’ trade secrets after the date
when the state’s precedent-setting case was decided, but not before. For the 21 states whose
courts recognized the IDD, we set the IDD index equal to zero in all years preceding the date of
the precedent-setting case, and equal to one afterwards. We allow the value of the IDD index to
revert to zero in the few cases in which a subsequent court decision reverses the interpretation of
the trade secrets law and rejects the IDD. For the 29 states whose case law did not explicitly
consider or considered but rejected IDD, we set the IDD index equal to zero in every year.
9 Examples of such special circumstances include when the plaintiff fails to establish the existence of a trade secret and when the defendant seeks employment at a firm that is not a direct competitor of the plaintiff.
15
III. Data and Methods
A. Sample Selection and Variable Definitions
Our data consists of all industrial firms in the merged CRSP-Compustat database (excluding
utilities and financials) incorporated in the U.S. and headquartered in a U.S. state for which there
is enough data to calculate the variables required in our main tests. The sample period is 1977-
2011, and it starts 5 years before the recognition of the IDD in Pennsylvania in 1982 and ends 5
years after Kansas recognized the doctrine in 2006. Our sample period excludes the events
associated with the recognition of the IDD by a few states in earlier years for two main reasons.
First, as discussed in Section II.A, the trade secret law surrounding the application of the IDD
did not follow the same principles in all states until the issuance of the UTSA in 1979. Second,
the coverage of earlier years in Compustat is sparser, especially in the 1960’s when Delaware,
Florida, and Michigan recognized the IDD (the data does not go back to 1919, when New York
recognized the IDD). Hence, earlier recognition events do not affect a significant number of firms
and have little power for identification.10 The final sample contains 134,428 firm-year
observations.
The key variables in our regression models are Book Leverage (book value of long-term debt
plus debt in current liabilities divided by book value of assets), Market Leverage (book value of
long-term debt plus debt in current liabilities divided by market value of assets), Inevitable
Disclosure (the IDD index), Log Book Assets (log of total assets), Market-to-Book (market value
of assets divided by book value of assets), Return on Assets (operating income before
depreciation divided by book value of assets), Fixed Assets (book value of property, plant, and
equipment divided by book value of assets), Industry Cash Flow Volatility (median of the
standard deviations of Return on Assets over the previous ten years for firms in the same 2-digit
SIC industry)11, Dividend Payer (indicator variable set to one if the firm pays a common
10 Our results are similar if we extend the sample back to 1971 and include the recognition in North Carolina in 1976 which affects only 38 firms in that state and occurred before the issuance of the UTSA. 11 Firms are required to have at least three years of data to enter the calculation.
16
dividend during the fiscal year and zero otherwise), State GDP Growth (state-level GDP growth
rate over the year, based on data from the Bureau of Economic Analysis), and Political Balance
(fraction of House of Representatives that belong to the Democrat Party for a given state and
year) 12.
Table II provides the definitions of these variables using Compustat designations and reports
summary statistics. All continuous variables are winsorized at their 1st and 99th percentiles.
Dollar values are expressed in 2009 dollars. Our data looks similar to that used in prior research
on capital structure. The mean (median) book leverage ratio is 0.23 (0.20) and the mean (median)
market leverage ratio is 0.18 (0.13). Firm’s total assets have a mean (median) value of $1,322
million ($146 million).
B. Difference-in-Differences Methodology
We use a difference-in-differences research design to examine how the recognition of the
IDD by state courts affects the financial leverage of firms headquartered in those states. The IDD
is applied in the context of employment law, not corporate law, so the relevant jurisdiction is the
state where the employee works and not the firm’s state of incorporation. Firms often operate
and thus employ workers in several different states, but data restrictions only allow us to identify
a firm’s state of headquarters. Nevertheless, within our conceptual framework, only the
employment location of workers with access to trade secrets matters for capital structure
decisions. Hence, our tests assume that workers with access to the trade secrets of publicly traded
firms are higher-level employees who are employed mostly at firms’ headquarters.
Our main specification is as follows:
Leverageist = α Inevitable Disclosurest + Xistβ + ωi + μt + εist (1)
where i denotes firm, s denotes state of a firm’s headquarters, and t denotes year. Leverage is a
measure of financial leverage, Inevitable Disclosure is a binary indicator for whether courts have
12 We obtain congress profile data on house representatives from the History, Art & Archives, U.S. House of Representatives available at http://history.house.gov/Congressional-Overview/Profiles/1st/.
17
recognized the IDD in the firm’s headquarters state by year t, X is a vector of control variables,
ωi is a firm fixed effect, and μt is a year fixed effect. The coefficient α is the difference-in-
difference estimate which gauges the effect of the IDD on firms’ capital structures.13 We note
that in our approach a firm in a given state can belong to both the “treatment” and “control”
groups at different points in time, and that our specification is not affected by the fact that some
states did not recognize the IDD during our sample period or recognized the IDD before 1977.
The vector X contains standard control variables in capital structure tests (e.g., Lemmon,
Roberts, and Zender (2008)), including log assets (a measure of firm size), the market-to-book
assets ratio (a proxy for growth opportunities), return on assets (a proxy for profitability and the
availability of internal funds), the proportion of assets that are fixed (a proxy for potential
collateral), industry cash flow volatility (a proxy for the likelihood of financial distress), and an
indicator variable for whether the firm pays common dividends (a proxy for financial
constraints). We also include two state-level control variables. The first is the one-year growth
rate of the GDP in the firm’s state, which captures business conditions in the state. The second 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. Including these variables addresses the concern that business conditions in the state or the
state’s political leaning might affect both the recognition of the IDD and financing decisions, and
thus cause a spurious association between financial leverage and the recognition of the IDD.
The firm fixed effects control for time-invariant unobservable firm characteristics and
ensure that estimates of α reflect actual changes in the adoption of IDD and the financial
leverage measure over time rather than simple cross-sectional correlations. The year fixed effects
13 Intuitively, suppose we want to estimate the effect of the recognition of the IDD by Massachusetts’ courts in 1994 on the leverage of firms in that state. For firms in Massachusetts (treatment state), we calculate a first difference by subtracting the average leverage of those firms before the IDD was recognized from their average leverage after it was recognized. We then calculate an analogous second difference but for firms in a control state (say California) that does not adopt the IDD in 1994 to capture changes in leverage unrelated to the recognition of the IDD, such as economy-wide shocks. Last, the estimated α results from subtracting the second difference from the first one.
18
account for transitory nation-wide factors, such as macroeconomic conditions, that could
possibly affect both financial leverage and state’s recognition of IDD.
The estimated standard errors in all our regressions are clustered at the state of headquarters
level, which assumes that observations are independent across states but not necessarily
independent within states. This is appropriate because Inevitable Disclosure is a state-level
variable and thus the regression errors may be correlated within state groupings. In addition to
accounting for heteroskedasticity, clustering at the state level addresses the concerns that the
residuals may be (i) serially correlated within a firm and (ii) correlated across firms within the
same state (in the same or different periods of time). Hence, the clustering method accounts for
the fact that firms headquartered in the same state are all simultaneously affected by the same
shock (the recognition of the IDD by the state’s court) and for any serial correlation induced by
the small time-series variation in the IDD indicator. See Bertrand, Duflo, and Mullainathan
(2004) for a discussion of these issues in the context of difference-in-difference estimation.
IV. Results
A. Recognition of the Inevitable Disclosure Doctrine and Capital Structure
There is some debate on whether capital structure tests should be based on book or market
leverage ratios, and prior work often uses one or the other. Market leverage is arguably more
appealing from a theoretical point of view, but many managers report that they base financing
decisions on book leverage (Graham and Harvey (2002)). Further, a substantial portion of the
variation in market leverage stems from variation in the market value of the firm rather than
changes in debt policies (Welch (2004)). Given this, throughout the paper, we measure a firm’s
capital structure using both book leverage and market leverage, but our results are similar.
Table III reports the difference-in-difference estimates of the impact the recognition of the
IDD by state courts has on the capital structures of firms in the recognizing state. Models 1-3 use
book leverage and models 4-6 use market leverage. For each dependent variable, we start with a
specification including Inevitable Disclosure, firm fixed effects, and year fixed effects. We then
19
add the typical firm-level control variables in capital structure tests (Log Book Assets, Market-to-
Book, Return on Assets, Fixed Assets, Industry Cash Flow Volatility, and Dividend Payer), and
finally add the two state-level control variables (State GDP Growth and Political Balance).
We find that the recognition of the IDD has a positive and statistically significant effect on
the financial leverage of firms in the recognizing state, which holds for both book and market
leverage measures and across all specifications we consider. In particular, the result is robust to
controlling not only for standard determinants of capital structure but also for the economic
conditions and the political leaning prevailing in the state, which could affect both the
recognition of the IDD and capital structure decisions. Further, the impact is economically
significant: the estimated coefficients in model 3 (model 6) imply that following the recognition
of the IDD firms increase their debt ratios by 1.4 (1.1) cents of additional debt per dollar of book
(market) assets, which represents a 6.0% (6.2%) increase relative to the sample mean.
In Table IV, we examine whether our findings are robust to using alternative measures of
financial leverage. First, in models 1 and 2, we measure book and market leverage net of cash
holdings, i.e., for both measures we calculate the numerators as the book value of long-term debt
plus debt in current liabilities less the book value of cash and short-term investments. We find
that the recognition of the IDD is also associated with an increase in net leverage. Net book
leverage increases by 1.7 cents for every dollar of book assets, which is equivalent to a 22.8%
increase relative to its sample mean of 0.057. Similarly, net market leverage increases by 1.5
cents for every dollar of market assets, which is equivalent to a 19.5% increase relative to its
sample mean of 0.077.14
Second, in models 3 and 4, we consider whether our results are robust to measuring financial
leverage using only the long-term debt portion of firms’ total debt, which includes both the
current portion of long-term debt and the portion of long-term debt maturing in more than one 14 The greater economic effect of the recognition of the IDD on net leverage compared to that for total leverage mostly reflects the fact that although the increases in these ratios are similar (e.g., an increase of 1.4 cents of additional debt per dollar of book assets for total book leverage and an increase of 1.7 cents of additional net leverage per dollar of book assets for net book leverage), the sample means for net book leverage and net market leverage are 0.057 and 0.077, while those for book leverage and market leverage are 0.23 and 0.18.
20
year. The results show that firms increase their book and market long-term debt ratios following
the recognition of the IDD and that the impact is economically important. Specifically, the
coefficient estimates on the IDD variable in models 3 and 4 imply that following the recognition
of the IDD firms increase their long-term book leverage ratios by 6.1% relative to the sample
mean of 0.197, and their long-term market leverage ratios by 5.9% (the sample mean is 0.152).
B. Further Evidence of Causality and Validity of Difference-in-Differences Approach
We now use the approach from Bertrand and Mullainathan (2003) to study the timing of
changes in capital structure associated with the recognition of the IDD. This addresses related
concerns about the interpretation of our results and about the validity of our empirical method.
First, although the recognition of the IDD appears to represent an exogenous event with
respect to capital structure decisions, it is conceivable that reverse causality might still affect our
results. For example, unobserved poor economic conditions might induce firms in some states to
increase their leverage ratios and financial distress risk, leading their valuable workers to seek
jobs at healthier competitors. Such firms may respond by exerting pressure on state courts to
protect their human capital through the recognition of the IDD. If this reverse causality exists,
one should see increases in the financial leverage of firms in recognizing states relative to that of
firms in non-recognizing states even before the IDD is actually recognized.
Second, a key assumption behind our difference-in-differences approach is that the trends in
the financial leverage of treatment firms (in recognizing states) and control firms (in non-
recognizing states) are parallel prior to the recognition of the IDD. If the “parallel trends”
assumption is violated, the estimated effect of the recognition of the IDD on financial leverage is
biased in an unknown direction. The reason is that the change in the capital structure of the
control firms does not correctly gauge the counterfactual for the change in the capital structure of
the treated firms, i.e., the change that treated firms would have experienced in the absence of
treatment. In this case, one should see changes in the financial leverage of firms in recognizing
states relative to that of firms in non-recognizing states before the IDD is actually recognized.
21
Table V presents the results of our timing tests, which replace Inevitable Disclosure (which
indicates whether a state has recognized the IDD by year t) with four indicator variables:
Inevitable Disclosure-1, Inevitable Disclosure0, Inevitable Disclosure1, and Inevitable
Disclosure2+. These indicators are equal to one if the firm is headquartered in a state that will
recognize the IDD next year, recognizes the IDD the current year, recognized the IDD the prior
year, and recognized the IDD two or more years ago, respectively.15 The coefficient on
Inevitable Disclosure-1 sheds light on both the possibility of reverse causality and the validity of
the parallel trends assumption. A statistically significant positive coefficient would indicate that
firms in recognizing states increased their leverage relative to those in non-recognizing states
even before courts in their states recognized the IDD, suggesting that the recognition of the IDD
is not the cause of the observed increase in leverage. More generally, a statistically significant
coefficient of any sign would indicate that the parallel trends assumption is violated, and thus
that the difference-in-difference estimates we report in Table III are biased.
Our results are similar regardless of whether we use book leverage or market leverage and
hold across all three specifications we consider (no controls, firm-level controls, firm-level and
state-level controls). Specifically, the coefficients on Inevitable Disclosure-1 and Inevitable
Disclosure0 are small and statistically insignificant in all cases, while the coefficients on
Inevitable Disclosure1 and Inevitable Disclosure2+ are positive and significant.16 Overall, the
results show that financial leverage increases only after the recognition of the IDD but not
before. Hence, reverse causality or a violation of the parallel trends assumption do not explain
our key result that the recognition of the IDD is associated with increases in financial leverage.
15 There are three instances in which a state rejected the IDD following a prior recognition: Florida in 2001, Michigan in 2002, and Texas in 2003. To ensure that these reversals do not confound our inferences, we drop all observations for these states from the date of the reversal onwards. This reduces our sample to 129,451 observations. 16 For robustness, we also tried including in the regression an indicator variable for whether a firm is headquartered in a state that will recognize IDD in two years. The coefficient on this variable is statistically insignificant and the coefficients on the remaining variables are unaffected by the inclusion of this additional control variable.
22
C. Cross-Sectional Variation in the Effect of the Recognition of the IDD on Capital Structure
We now study the cross-sectional variation in the impact of increased protection of trade
secrets on capital structure and test the predictions that follow from our main hypothesis as
discussed in Section I. Our approach is to split the sample in two groups based on whether the
value of a characteristic is above or below the sample median, estimate our main specification in
each group, and compare the estimated coefficients of Inevitable Disclosure across groups.17
These tests shed light on the economic mechanism behind our main results and provide
further evidence of identification, namely, that our results have a causal interpretation.
Specifically, we examine if the effect of the recognition of the IDD on financing decisions varies
predictably with the degree of competition in the industry, the type of workers employed by the
firm, and the risk that workers will become employed at rival firms. If one believes that an
omitted variable uncorrelated with all the control variables might drive our results in Table III,
then one would have to explain why such an omitted variable is correlated with all the
characteristics for which we find a stronger relation between IDD and financial leverage.
In Table VI, we test the prediction that better protection of trade secrets should have a larger
impact on the capital structures of firms in more competitive environments. Panel A uses book
leverage and Panel B uses market leverage, but both panels have the same structure. We first
gauge the extent of competition in the industry using the four-firm concentration ratio compiled
by the U.S. Economic Census for 5-digit NAICS industries, which captures the fraction of the
industry’s sales accounted for by the top four firms in the industry.18 In columns 1 and 2, we split
the sample according to whether the four-firm industry concentration ratio is above the sample
median (less competitive industries) or below the sample median (more competitive industries).
The sample for these tests excludes four industries for which the Census does not compile
17 Compared to an approach in which the full sample is used and in which we would interact the Inevitable Disclosure indicator variable with an indicator for whether a firm is in the high subsample for a given characteristic, a benefit of the approach we use is that it allows the coefficients of the control variables to differ across subsamples. 18 Concentration ratios are only available for years 1997, 2002, and 2007, but are stable over time. Hence, following a common approach (e.g., Campello (2006) and Haushalter, Klasa, and Maxwell (2007)), we assume that the ratios in 1997, 2002, and 2007 are valid for the periods 1977-1999, 2000-2004, and 2005-2011, respectively.
23
concentration ratios (agriculture, forestry, fishing, and hunting, mining, construction, and
management of company enterprises). Supporting our prediction, the results based on both book
and market leverage indicate that the positive effect of the adoption of the IDD on firms’ debt
ratios only exists for firms that operate in the less competitive industries.
We then gauge competition in an industry using a proxy for the magnitude of barriers to
entry, since fewer barriers to entry increase competition by facilitating the entry of new firms
into these industries. Following Valta (2012), we use the 3-digit SIC industry-average values of
R&D plus advertising expenditures divided by sales to measure barriers to entry. This is
motivated by Shaked and Sutton (1987) and Sutton (1991) who argue that firms use R&D and
advertising to differentiate their products from those of their competitors and to make it more
difficult for new competitors to enter the market. In addition, Hoberg and Phillips (2011) show
that firms spending more in R&D and advertising experience reductions in competition. In
columns 3 and 4 of both panels, we split the sample into industries with barriers to entry above
the sample median (less competitive industries) and below the sample median (more competitive
industries). In both panels, we find that the positive effect of the adoption of the IDD on firms’
debt ratios only exists for firms in industries with less significant barriers to entry. This provides
further evidence that the protection of trade secrets matters more in more competitive industries.
In Table VII we test the prediction that better protection of trade secrets should have a larger
impact on the capital structures of firms that employ more workers who likely know their trade
secrets. To this end, we assume that more of a firm’s workers are likely to know the firm’s trade
secrets in industries that employ a greater fraction of workers in managerial or science
occupations (occupations which typically entail access to firms’ trade secrets) and in industries
that employ a greater fraction of educated workers (such workers often know the firm’s trade
secrets). The skills and thus the occupations of the workers in the industry may vary by state.
Hence, we consider the occupational structure and education level of the workers in the industry
that are employed in the firm’s state. The data is from the Integrated Public Use Microdata Series
24
(IPUMS-USA) database, which reports the characteristics of workers by 3-digit NAICS industry
and state.19
We then repeat our main regression in subsamples which result from splitting our full
sample in three alternative ways, namely, according to whether the fraction of the workers in the
industry that are in managerial occupations (codes 4, 13, 22, or 33 in IPUMS), that are in science
occupations (codes 64, 68, 69, 73, 74, 75, 76, 77, 78, 79, or 83 in IPUMS), or that have at least a
bachelor’s degree is below or above the sample median for these fractions, respectively. Our
results in Panel A (book leverage) and Panel B (market leverage) are qualitatively similar, and
suggest that the recognition of the IDD indeed has a larger impact on financial leverage when
firms’ workers are more likely to know trade secrets. We find that the positive effect of the
recognition of the IDD on corporate debt ratios is prevalent only in samples of firms operating in
industries with above the median fraction of workers (i) in managerial occupations, (ii) in
science occupations, and (iii) with at least a bachelor’s degree.
Last, in Table VIII we test the prediction that better protection of trade secrets should have a
larger impact on the capital structure of firms that face a greater ex-ante risk that their employees
with knowledge of trade secrets would accept a position with a rival. To do so, using both book
leverage (Panel A) and market leverage (Panel B), we conduct two related tests based on sample
splits analogous to those in our prior analyses: one based on pure switching costs and another
based on the extent of competition among rival firms in local labor markets.
First, the cost of switching employers is higher for workers in firms with defined benefit
pension plans, since retirement benefits from these plans are less portable (Ippolito (1985)).20
Hence, firms with defined benefit pension plans face less risk of losing key employees to rival
19 The IPUMS-USA database is compiled from the American population federal censuses conducted every 10 years and is available for the years 1980, 1990, and 2000. We assume that the information from the 1980, 1990, and 2000 censuses is valid for the periods 1977-1985, 1986-1995, and 1996-2011, respectively. 20 Defined benefit pension plans create incentives for workers to remain with the firm because retirement income is based on years of service and final wage. For example, an employee who worked for a firm for many years will accrue higher retirement income benefit than someone who worked for the firm only a few years. If he or she leaves to work for another firm, the benefits are reset and the job switcher realizes a lower level of retirement benefits.
25
firms, and thus the recognition of the IDD should have a smaller impact for those firms. To test
this prediction, we identify firms with defined benefit pension plans as those that report positive
net pension benefit assets or accumulated pension benefit obligations. Because pension data are
available in Compustat only since 1980, in these tests we restrict our sample to the years 1980-
2011. Supporting our prediction, models 1 and 2 in both panels of the table show that the
recognition of the IDD only affects the debt ratios of firms without defined benefit pension plans.
Second, firms face higher ex-ante risk that rivals might gain access to their trade secrets
through labor mobility when there is a significant presence of geographically close rivals and
thus more competition in local labor markets. Such presence is associated with more outside job
opportunities for the firm’s workers that entail small switching costs due to the proximity
between their current and prospective jobs. Hence, the recognition of the IDD should have a
larger impact on the debt ratios of firms facing stronger competition in local labor markets. We
gauge the extent of competition in local labor markets a firm faces using the firm’s share in the
2-digit SIC industry’s employment located in the firm’s state, calculated using Compustat data.
A lower value of this variable indicates that the firm’s rivals account for a larger fraction of the
industry’s employment in the state, and thus the firm is likely to face more intense competition in
local labor markets. Supporting our prediction, the results in models 3 and 4 of both panels
indicate that the recognition of the IDD only impacts the debt ratios of firms whose share in the
industry’s employment in the state is below the sample median.
Overall, the results in Tables VI-VIII are consistent with predictions which follow naturally
from our main hypothesis that a firm’s optimal leverage is higher when there is less risk that
rivals might gain access to the firm’s trade secrets and damage its competitive position. The
results provide further evidence that the positive impact of the recognition of the IDD on
corporate debt ratios that we identify in our main tests is not spuriously driven by unobserved
heterogeneity.
26
D. Recognition of the IDD and Changes in Firms’ Default Risk
Our main hypothesis is that an increase in the legal protection of a firm’s trade secrets
increases the firm’s desired leverage ratio because it reduces the probability of financial distress.
Our prior evidence showing that firms increase their financial leverage following the recognition
of the IDD in their states supports our hypothesis, but does not directly show why this occurs. To
this end, we now study whether the recognition of the IDD reduces a firm’s default risk and thus
causes the observed increase in financial leverage through the mechanism behind our hypothesis.
In Table IX we investigate the effect of the recognition of the IDD on a firm’s default risk
measured by the interest rate a firm pays on its bank debt in excess of the risk-free rate. This
credit spread reflects lenders’ forward-looking assessment of the risk that the firm will default on
its debt obligations. We focus on the credit spreads of bank debt as in Valta (2012) because bank
debt is the principal source of debt financing used by most firms (Faulkender and Petersen
(2006)) and because the data is available for a large sample of firms. The data comes from
Dealscan and contains loans made to the firms in our main sample that are originated in the U.S.
and denominated in U.S. dollars. The sample period used for this analysis is 1987-2011.
The dependent variable in all regression models is the logarithm of firms’ credit spreads,
defined as the spread between a firm’s interest rate on a bank loan and the LIBOR rate.21 Since
Dealscan contains data only for new loans in the year they are granted and thus most firms enter
the data sporadically, we are unable to include firm fixed effects in these tests due to a lack of
enough annual observations per firm for proper estimation. Instead, in addition to year fixed
effects, we include both 3-digit SIC industry fixed effects and state of headquarters fixed effects.
This implies that the estimated coefficient on Inevitable Disclosure indicates the average impact
the recognition of the IDD has on the credit spread of those firms in an industry that are
headquartered in the recognizing state (relative to its impact on the borrowing costs of those
firms in the industry that are headquartered in non-recognizing states).
21 The credit spread is measured as the all-in-spread drawn in Dealscan, defined as the amount the borrower pays in basis points over LIBOR for each dollar drawn down (including annual fees paid to the bank group).
27
In Panel A, we regress the logarithm of credit spreads on Inevitable Disclosure and control
variables. In model 1, we include the control variables in our main specification (reported in
Table III) and leverage.22 In model 2, we also include the logarithms of loan size (in $ millions)
and of loan maturity (in months), which are usually used in cost of debt models, and loan-type
fixed effects. In model 3, we also control for the state GDP growth and political balance
variables. All specifications indicate that the recognition of the IDD in a state is associated with a
decrease in the average credit spreads of firms in that state. In terms of economic significance,
the results for model 3 imply that following the recognition of the IDD the credit spread firms in
the recognizing state pay over LIBOR decreases by approximately 5.4%.
In Panel B we report the results of timing tests (analogous to those reported in Table V), in
which we replace Inevitable Disclosure by the four dummy variables previously defined in
Section II.B: Inevitable Disclosure-1, Inevitable Disclosure0, Inevitable Disclosure1, and
Inevitable Disclosure2+. Supporting a causal interpretation of the effect of the recognition of the
IDD on firms’ credit spreads, the results for the three specifications we consider show that firms’
credit spreads decrease only after and not before the recognition of the IDD by state courts. In
addition, the coefficient on Inevitable Disclosure-1 is not statistically different from zero,
suggesting that the trends in the capital structure of firms in recognizing states and those in non-
recognizing states before the recognition of the IDD are parallel.
Overall, together with the evidence in prior tables, the results in Table IX strongly suggest
that the increase in a firm’s leverage ratio following the recognition of the IDD by state courts
occurs because better protection of trade secrets reduces the firm’s risk of default.
E. Recognition of the IDD and Changes in Firms’ Market Shares
Our main hypothesis and the evidence we presented so far imply that, by increasing the
protection of a firm’s trade secrets, the recognition of the IDD allows firms to more successfully
compete in product markets. Specifically, when the recognition of the IDD increases the
22 The results are very similar if we do not include leverage as a control variable in the regression.
28
protection of a firm’s trade secrets, it also improves the firm’s ability to maintain a competitive
advantage over its industry rivals derived from those trade secrets. In this section we thus explore
whether a better protection of trade secrets boosts firms’ performance in product markets.
In Table X, we examine the impact of the IDD on firms’ performance using an approach
similar to that in Opler and Titman (1004), Campello (2006), and Frésard (2010). Specifically,
the dependent variable is the firm’s sales growth rate, and the control variables include logarithm
of assets, return on assets, the market-to-book ratio, capital expenditures scaled by assets, R&D
expense scaled by sales, advertising expenses scaled by sales, and book leverage. Because we are
interested in the impact of firms’ performance within their product markets, in all regressions we
include industry times year fixed effects, which is equivalent to subtracting the corresponding
industry means from each variable in each year. With this approach we can interpret the
coefficients as the impact of the right-hand side variables on a firm’s sales growth relative to that
of its industry rivals, which roughly corresponds to changes in the firm’s market share.
We report specifications including and excluding financial leverage among the control
variables, to account for the finding in prior studies that capital structure might affect
performance in product markets. We also use alternative product market definitions based on 3-
digit and 4-digit SIC codes, but this does not have a material effect on our results. All
specifications we consider provide statistically significant evidence that firms perform better
within their product markets following the recognition of the IDD in their state. However, the
economic significance and the associated gains in market share appear modest: the results in
column 4 suggest that after the recognition of the IDD in a firm’s state, its sales grow 1.8
percentage points faster relative to the sales growth of its rivals in the same 4-digit SIC industry.
F. Discussion: Protection of Trade Secrets or Labor-Related Mechanisms?
We argue that the recognition of the IDD affects a firm’s capital structure because it
increases the protection of the firm’ trade secrets. However, the reduced mobility of workers
with access to trade secrets inherently associated with the recognition of the IDD could also
29
affect capital structure through pure labor-related mechanisms unrelated to better protection of
trade secrets. As we discuss below, pure labor-related mechanisms seem unlikely to explain our
findings.
Pure labor-related mechanisms linking the recognition of the IDD to capital structure hinge
on the assumption that this event would have a large effect on a firm’s labor costs. However, the
recognition of the IDD typically only affects the mobility of a small number of workers – those
with knowledge of the firm’s trade secrets – and not the mobility of the bulk of workers.
Although workers with access to trade secrets are usually paid a higher salary, the total wage bill
associated with the compensation of such workers is likely to be small relative to a firm’s total
wage bill. This suggests that labor-related stories for the study’s results that are based on
compensation costs are unlikely to explain the large effect the recognition of the IDD has on
firms’ capital structures. Notwithstanding this very important caveat, we now outline and discuss
possible mechanisms.
Recent empirical work by Agarwal and Matsa (2013) and Kim (2012) highlights that
unemployment is very costly for workers (e.g., due to lost wages or psychological damage) and
thus workers demand additional compensation when firms expose them to unemployment risk.
Intuitively, the additional compensation is the probability of layoffs times the cost of layoffs.
Since financial leverage increases the probability of distress and thus of layoffs, a firm can
increase its leverage without a large impact on wages when there is a reduction in the cost of
layoffs to workers. In our context, the recognition of the IDD might constrain the
reemployability of workers with knowledge of a firm’s trade secrets if they are laid off in
financial distress. Since this increases the cost that those workers will bear if laid off, those
workers would demand higher ex-ante compensation for unemployment risk. This suggests that
following the recognition of the IDD firms could decrease financial leverage to reduce the
30
probability of distress and thus to minimize the additional compensation they must pay.
Noteworthy, the predicted effect on capital structure is opposite to what we find.23
The recognition of the IDD reduces the outside employment opportunities of a firm’s
workers with access to trade secrets at industry rivals and thus could also reduce their bargaining
power in negotiating compensation for unemployment risk. How this affects capital structure is
not obvious, as the recognition of the IDD affects current and prospective employees differently.
On one hand, after the recognition of the IDD the firm might be able to increase its debt ratio and
retain its current workers with access to trade secrets without commensurate compensation for
the additional unemployment risk. On another hand, after the recognition of the IDD the firm
might need to hire new workers who will soon gain access to its trade secrets and become less
mobile. Such workers will demand extra compensation for unemployment risk to accept working
for the firm, leading the firm to reduce its leverage to moderate compensation demands.
Last, it could be argued that, because the recognition of the IDD “handcuffs” workers with
knowledge of trade secrets to their jobs with the firm, it potentially helps the firm retain its most
productive workers even if it does not pay such workers competitive wages. This, in turn, could
increase debt capacity by protecting the firm’s valuable human capital, i.e., by reducing the risk
of a loss of key talent that could hurt its performance, aside from any issues associated with trade
secrets. However, this mechanism is far from obvious. The reason is that the constraints on the
mobility of key workers imposed by the recognition of the IDD could also reduce the firm’s debt
capacity if these constraints discourage current workers from exerting effort, distort these
workers’ incentives to invest in their human capital, or hamper the firm’s ability to recruit new
workers who are averse to job lock.24 23 Our hypothesis further suggests that, by protecting the firm’s trade secrets, the recognition of the IDD reduces the probability of distress and thus of layoffs, which reduces the likelihood that workers with knowledge of trade secrets will incur unemployment costs. This could lower the compensation for unemployment risk those workers demand and thus could lead firms to increase the desired debt ratio. If this is taken into account, the net effect on financial leverage due to the recognition of the IDD resulting from pure labor-related mechanisms is a priori ambiguous. 24 The recognition of the IDD might also allow the firm to reduce the wages of workers with access to trade secrets in bad times without losing them to industry rivals, which reduces its operating leverage, and thus increases its debt capacity. However, as explained above, any effect on the firm’s total wage bill is likely to be too small to affect debt ratios.
31
V. Additional Investigation and Robustness Tests
A. Recognition of the IDD and Investment Policy
We also explore whether the recognition of the IDD affects firms’ investment and other
related expenses, which serves to shed light on two related issues. First, by increasing the
protection of trade secrets, the recognition of the IDD might increase the marginal benefit of
investment, especially of investment associated with the creation of trade secrets (such as R&D),
and thus increase firms’ demand of external financing which is needed to fund investment.25 In
that case, the observed increase in financial leverage could be caused by increased financing
needs rather than by increased debt capacity as implied by our main hypothesis.
Second, the recognition of the IDD might be associated with possibly time-varying
unobserved heterogeneity in firms’ investment opportunities which is not captured by the firm
fixed effects and our control variables. If the recognition of the IDD coincides with increases in
the investment opportunities of firms in the state, then one would expect to see increases in
investment and related expenses following the recognition of the IDD. In turn, such investment
likely requires external debt financing, and might cause the increase in leverage we observe.
In untabulated tests we study the impact of the recognition of the IDD on four investment
policy variables: R&D expenses scaled by sales, advertising expenses scaled by sales, capital
expenditures scaled by book assets, and acquisition expenses scaled by book assets. In addition
to firm fixed effects and year fixed effects, we include the same control variables as in our main
leverage specification: logarithm of assets, market-to-book ratio, return on assets, fixed assets
scaled by total assets, industry cash flow volatility, dividend payer dummy, state GDP growth,
and the state’s political balance. We find that the recognition of the IDD has no effect on any of
the four investment variables we consider, which suggests that increases in investment needs are
not the main force behind the observed increase in financial leverage.
25 We entertain this possibility but, as noted by Png (2012b), the impact of trade secret protection on innovation is a priori ambiguous because better protection increases the firm’s ability to appropriate the benefits of its investment, but it diminishes the firm’s ability to benefit from spillovers from other firms. His results based on R&D spending do not provide a general conclusion, as better protection increases R&D in some cases and decreases it in other ones.
32
B. Impact of Increases in the Protection of Rivals’ Trade Secrets
Our prior analyses study how a firm’s capital structure and competitive position are affected
by the recognition of the IDD in its own state, namely, by an increase in the legal protection of
its own trade secrets. We now consider the possibility that the firm might also be affected by the
recognition of the IDD in states other than the firm’s state. Such events do not affect the
protection of the firm’s trade secrets and thus do not affect the firm directly. However, they
increase the protection of the trade secrets of the firm’s industry rivals located in other states and
thus they can indirectly affect the firm’s competitive position and capital structure. Our tests
discussed below address this issue and aim to quantify its empirical relevance.
In Table XI we study the impact on a firm’s market share and capital structure of an increase
in the legal protection of the trade secrets of the firm’s industry rivals that are headquartered
outside of the firm’s state. We gauge the protection of the trade secrets of out-of-state industry
rivals using Out-of-State Rivals’ Inevitable Disclosure, which is defined for each 3-digit SIC
industry and state as the weighted-average IDD of out-of-state industry rivals (with weights
equal to each state’s share in the total sales of out-of-state industry rivals) multiplied by the share
of out-of-state industry rivals in the total sales of the industry.26 Hence, this measure takes into
account that better protection of out-of-state rivals’ trade secrets is likely to have a larger impact
on the firm when out-of-state rivals account for a larger share of the total sales of the industry.
In Panel A, we report analyses analogous to those reported in Table X, in which we regress a
firm’s sales growth rate on Out-of-State Rivals’ Inevitable Disclosure and control variables. We
include industry times year fixed effects and thus, as explained in Section IV.E, the estimated
coefficients can be interpreted as changes in the firm’s market share. Columns (1)-(2) use 3-digit
SIC industry definitions for the industry times year fixed effects and columns (3)-(4) use 4-digit
26 Specifically, let s = 1,...,n denote the set of states, j a particular state, i industry, and t year. For any given industry i with operations in state j in year t, the inevitable disclosure index of out-of-state industry rivals is calculated as:
‐ ‐ , , ∑ ,, ,
∑, ,
∑
, ,
∑ , ,
33
SIC industry definitions for these fixed effects. If an increase in the legal protection of the trade
secrets of a firm’s out-of-state industry rivals significantly weakens a firm’s competitive position
and in consequence leads to a decrease in its market share, we would expect a negative
coefficient on Out-of-State Rivals’ Inevitable Disclosure.
In columns (1) and (3), we include Out-of-State Rivals’ Inevitable Disclosure (but not
Inevitable Disclosure). For both industry definitions, the estimated coefficient on Out-of-State
Rivals’ Inevitable Disclosure is negative as predicted, but it is statistically insignificant when
industries are defined using 3-digit SIC codes and only significant at the 10% level when
industries are defined using 4-digit SIC codes. In columns (2) and (4), we include Inevitable
Disclosure, and the coefficient on Out-of-State Rivals’ Inevitable Disclosure becomes
statistically insignificant in both cases, while Inevitable Disclosure has a positive and statistically
significant coefficient as in Table X. Overall, these results suggest that increased protection of
the trade secrets of a firm’s out-of-state industry rivals does not cause a significant decrease in
the firm’s market share.
In Panel B, we report analyses analogous to those in Table III, in which we regress a firm’s
financial leverage on Out-of-State Rivals’ Inevitable Disclosure and control variables, including
firm fixed effects and year fixed effects. Columns (1)-(2) use book leverage and columns (3)-(4)
use market leverage. If an increase in the legal protection of the trade secrets of a firm’s out-of-
state industry rivals adversely affects the distribution of the firm’s future cash flow and increases
the probability of financial distress, it should decrease the firm’s desired leverage ratio. In this
case, the effect of Out-of-State Rivals’ Inevitable Disclosure on financial leverage should be
negative.
We find that an increase in the protection of the trade secrets of out-of-state industry rivals
does not affect a firm’s book leverage (the estimated coefficient on Out-of-State Rivals’
Inevitable Disclosure is negative but statistically insignificant, regardless of whether Inevitable
Disclosure is included in the regression). In contrast, there is some indication that it reduces a
firm’s market leverage: the coefficient on Out-of-State Rivals' Inevitable Disclosure is negative,
34
but it is only statistically significant at the 10% level when Inevitable Disclosure is included in
the regression. However, the effect is not economically important: a one standard deviation
increase in Out-of-State Rivals’ Inevitable Disclosure decreases the firm’s market leverage by
only 1% (which is small relative to the 6.1 % increase associated with the recognition of the IDD
in the firm’s own state). In all specifications, Inevitable Disclosure continues to have a positive
effect on financial leverage which is similar in both statistical and economic significance to that
reported in Table III.
In sum, the evidence in Table XI suggests that the indirect effect on a firm’s competitive
position and capital structure caused by an increase in the protection of the trade secrets of its
out-of-state industry rivals is markedly less important than the direct effect associated with an
increase in the protection of trade secrets in the firm’s own state.
C. Does Measurement Error in the Inevitable Disclosure Index Affect the Results?
C1. Relocation of Firms’ Headquarters from One State to Another
We study how the recognition of the IDD in a firm’s state of headquarters affects its capital
structure, and to this end, we identify a firm’s state of headquarters using the most recent address
of a firm’s headquarters because this is the only information provided in the Compustat database.
This assumes that firms are headquartered in their most recent state of headquarters during the
entire sample period, namely, that firms never relocated their headquarters from one state to
another. However, if many firms do relocate their headquarters to other states during our sample
period, then measurement error in the state of headquarters – and thus in Inevitable Disclosure –
could confound our inferences and possibly attenuate our results reported in Table III.
We use the programming language PHP to search the 10-K filings available on the SEC’s
website and collect the historical state of location of each firm’s headquarters. Given data
availability, we are able to obtain the information for most firms between 1996 and 2011 and for
some as early as 1992 (but not for our entire sample which spans 1977-2011). Of the 8,852 firms
for which we obtain historical location of headquarters during the period 1992-2011, only 826
35
(or 9.3%) changed relocated headquarters from one state to another. These findings imply that
relocations of corporate headquarters across states are relatively infrequent and that they only
affect a small fraction of the firms we study. These findings are also consistent with those
reported in Pirinsky and Wang (2006), who also report that relocations of corporate headquarters
are rare events. Given the low incidence of headquarter relocations, it seems unlikely that these
events could have a large impact on our results.
Nevertheless, in columns (1)-(3) of Table XII, we examine whether unobserved relocations
of firms’ headquarters could affect our main results for both book leverage (Panel A) and market
leverage (Panel B). First, in column (1) we use the location of headquarters that we collect from
the 10-Ks to reduce the measurement error in Inevitable Disclosure, while retaining our full
sample period. Specifically, we use the information from the 10-Ks when it is available, and
when it is not available we assume there were no relocations prior to the earliest date it is
available.27 Second, in column (2) we use only the subsample of firm-years for which the
information on the location of headquarters that we collect from the 10-Ks is available. The
sample is much smaller and spans only the period 1992-2011, but in this subsample Inevitable
Disclosure is measured without any error caused by headquarter relocations. Third, in column
(3), we exclude from the sample those firms that are likely to have experienced major
restructuring events (those which Pirinsky and Wang (2006) argue are the main trigger of
headquarter relocations) during the sample period. We identify such firms as those with sales or
assets growth in excess of 100% in any year during 1977-2011, because Almeida, Campello, and
Weisbach (2004) highlight that major corporate events are usually associated with large
increases in sales or assets.
In all three tests discussed above, Inevitable Disclosure continues to have a positive and
statistically significant impact on both book leverage and market leverage which is similar in
27 This approach removes the measurement error for firm-years with historical 10-K available and reduces it for earlier years in the sample by using the closest in time information available instead of the most recent one.
36
magnitude to that reported in Table III. Hence, we conclude that our inferences based on
Inevitable Disclosure are unlikely to be biased due to changes in firms’ state of headquarters.
C2.Other Potential Sources of Measurement Error
We further show that our main results in Table III are robust to the imposition of further
constraints to our sample which arguably reduce the potential measurement error in Inevitable
Disclosure caused by foreign operations and geographical dispersion of employment.
First, the recognition of the IDD affects firms to the extent that their workers with
knowledge of their trade secrets are employed in the U.S. However, the IDD offers no extra
protection if the trade secrets are revealed by employees in foreign countries. If firms have
substantial operations in foreign countries, the recognition of the IDD by U.S. courts is likely to
be less effective in increasing the protection of a firm’s trade secrets. Column (4) in both panels
shows that excluding firms which report foreign income or taxes from the sample leads to a
larger impact of Inevitable Disclosure on book leverage and a similar effect on market leverage.
Second, our tests rely on the recognition of the IDD in the firm’s state of headquarters,
where arguably most of the firm’s employees with access to trade secrets are employed.
However, we are likely to measure changes in trade secret protection with error for firms that
have a geographically dispersed workforce. Using an approach similar to that in Agarwal and
Matsa (2013), we address this concern by excluding from the sample industries whose workforce
is likely to be geographically dispersed, namely, retail, wholesale, and transportation. The results
in column (5) show that the impact of the recognition of the IDD is slightly larger when we use
book leverage as the dependent variable and unchanged when we use market leverage.
D. Additional Control Variables in Main Specification
In Table XIII we provide evidence on whether our results in Table III might be spuriously
driven by additional factors we have not controlled for in our main specification. First, it is
possible that changes in the extent of product market competition in the firm’s state affect both
courts’ decisions to recognize the IDD and firms’ capital structure decisions. We address this
37
concern by including the Herfindhal-Hirschmann Index of sales concentration within the firm’s
industry and state (State-Industry HHI) based on Compustat data as an additional control
variable. However, the estimated coefficient on Inevitable Disclosure is unaffected and that on
State-Industry HHI is statistically insignificant.
Second, in our main tests we use a significant discrete event – the recognition of the IDD by
state courts – to identify an increase in the protection of firms’ trade secrets. However, firms’
trade secrets are also protected by a slowly evolving legislation and enforcement of employment
contracts in their states. To assess whether the recognition of the IDD has a distinct impact on
capital structure, we further include two state-level measures of the extent of trade secret
protection in the firm’s state. These are Png’s (2012a) trade secret protection index (Strength of
Trade Secret Protection), which measures aspects of trade secret protection in the state other
than those captured by Inevitable Disclosure, and Bird and Knopf’s (2010) extension of the
Garmaise (2010) index (Strength of Non-Competes), which measures the extent to which
covenants not to compete are enforced in the state.28 The inclusion of these variables does not
affect the estimated coefficient on Inevitable Disclosure and the coefficients for both variables
are statistically insignificant.
Last, by increasing the protection of confidential information maintained in the form of trade
secrets, the recognition of the IDD might affect a firm’s incentives to patent its innovation. This
could, in turn, change the firm’s business risk and have an impact on its financing decisions. To
address this possibility, we include two firm-level measures of innovation output obtained from
the NBER patent database: the logarithm of the number of patents obtained by the firm and the
logarithm of the number of citations of those patents. Due to truncation issues with these data in
later years, the sample for these tests ends in 2000. We find that increases in a firm’s number of
patents are associated with decreases in its debt ratio. Nevertheless, controlling for changes in
patented innovation does not affect the estimated coefficient on Inevitable Disclosure.
28 The Bird and Knopf (2010) index covers the period 1976-2004. Since changes in the index are infrequent, we use the 2004 values to fill in the period 2005-2011. Our results are similar if we only use the period 1976-2004.
38
VI. Conclusion
Our main message is that the risk that a firm’s rivals might gain access to its intellectual
property in the form of trade secrets influences capital structure decisions. In particular, we
hypothesize that firms facing less risk that their product market rivals will gain access to their
trade secrets and hurt their competitive positions can hold more debt because they are less likely
to default on debt payments. We then test our hypothesis using a difference-in-differences
research design that exploits the staggered recognition of the Inevitable Disclosure Doctrine by
state courts over the period 1977-2011. The recognition of this doctrine causes an exogenous
decrease in the risk that industry rivals might gain access to a firm’s trade secrets, because it
increases a firm’s ability to prevent its workers who know trade secrets from working at rival
firms and conveying the trade secrets to their new employers.
Supporting our hypothesis, we find that firms significantly increase their leverage following
the recognition of the doctrine by courts in their states of headquarters. We further show that this
relation is causal using a timing test which indicates that leverage increases after the recognition
of the doctrine but not before. The cross-sectional variation in the impact of the recognition of
the doctrine on capital structure is also consistent with the economic mechanism we describe.
The impact is particularly strong for firms in competitive industries, for firms whose workers are
more likely to know trade secrets, and for firms that face a greater ex-ante risk of losing key
employees to competitors. We also show that the credit spreads a firm pays on its bank loans
decrease following the recognition of the doctrine, which suggests that the increase in leverage
we document is driven by a reduction in the firm’s risk of default as suggested by our main
hypothesis. Last, consistent with competition issues driving our results, we find that the
recognition of the doctrine causes moderate gains in market share for those firms headquartered
in recognizing states.
Our paper emphasizes the interplay of firms in product and labor markets, and that
competition in those markets creates business risks which affect firms’ financial decisions. In
particular, it calls to attention risks associated with the firm’s intellectual property, which is a
39
key element of firms’ competitive strategies, and highlights that the mobility of key employees
among rival firms exacerbates this risk. We believe this is a fruitful area for future research.
References
Agrawal, A. K., Matsa, D. A., 2013. Labor unemployment risk and corporate financing decisions. Journal of Financial Economics 108, 449-470.
Almeida, H., Campello, M., Weisbach, M.S., 2004. The cash flow sensitivity of cash. Journal of Finance 59, 1777-1804.
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.
Bertrand, M., Mullainathan, S., 2003. Enjoying the quiet life? Corporate governance and managerial preferences. Journal of Political Economy 111, 1043-1075.
Bertrand, M., Duflo, E., Mullainathan, S., 2004. How much should we trust differences-in-differences estimates? Quarterly Journal of Economics 119, 249-275.
Bird, R. C., Knopf J. D., 2010. The impact of labor mobility on bank performance. Working paper, University of Connecticut.
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 supermarkets’ pricing. Journal of Finance 50, 1095–1112.
Faulkender, M., Petersen, M.A., 2006. Does the source of capital affect capital structure? Review of Financial Studies 19, 45-79.
Frank, M. Z., Goyal, V. K., 2007. Trade-off and pecking order theories of debt. In B. Espen Eckbo Ed., Handbook of Empirical Corporate Finance, Vol. 2, Chapter 12, 135-202, North-Holland.
Frésard, L., 2010. Financial strength and product market behavior: The real effects of corporate cash holdings. Journal of Finance 65, 1097-1122.
40
Garmaise, M. J., 2011. Ties that truly bind: Noncompetition agreements, executive compensation, and firm investment. Journal of Law, Economics, and Organization 27, 376-425.
Graham, J., Harvey, C., 2002. How do CFOs make capital budgeting and capital structure decisions? Journal of Applied Corporate Finance 15, 8-23.
Harris, M., Raviv, A., 1991. The theory of capital structure. Journal of Finance 46, 297-355.
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.
Hoberg, G., Phillips, G., 2011. Text-based network industries and endogenous product differentiation. Working paper, University of Maryland.
Hoberg, G., Phillips, G., Prabhala, N., 2013. Product market threats, payouts, and financial flexibility. Journal of Finance, forthcoming.
Ippolito, R. A., 1985. The economic function of underfunded pension plans. Journal of Law and Economics 28, 611–51.
Kahnke, R. E., Bundy, K. L., Liebman, K. A., 2008. Doctrine of inevitable disclosure. Faegre and Benson LLP.
Kim, H., 2012. Labor market size and employer capital structure: Evidence from large plant openings. Working paper, Cornell University.
Kovenock, D., Phillips, G., 1997. Capital structure and product market behavior: an examination of plant exit and investment decisions. Review of Financial Studies 10, 767–803.
Leary, M. T., Roberts, M. R., 2005. Do firms rebalance their capital structures? Journal of Finance 60, 2575-2619.
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.
Lyandres, E., 2006. Capital structure and interaction among firms in output market: 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, VA.
Malsberger, B. M., 2011. Trade Secrets: A State-by-State Survey, 4th ed. BNA Books, Arlington, VA.
41
Matsa, D. A., 2010. Capital structure as a strategic variable: Evidence from collective bargaining. Journal of Finance 65, 1197-1232.
Opler, T., Titman, S., 1994. Financial distress and corporate performance. Journal of Finance 49, 1015–1040.
Phillips, G., 1995. Increased debt and industry product markets: an empirical analysis. Journal of Financial Economics 37, 189–238.
Pirinsky, C. A., Wang, Q., 2006. Does corporate headquarters location matter for stock returns? Journal of Finance 61, 1991 – 2015.
Porter, M., 1980. Competitive strategy. New York: Free Press.
Png, I., 2012a. Trade secrets, non-competes, and mobility of engineers and scientists: Empirical evidence. Working paper, National University of Singapore.
Png, I., 2012b. Law and innovation: Evidence from state trade secrets laws. Working paper, National University of Singapore.
Shaked, A., Sutton, J., 1987. Product differentiation and industrial structure. Journal of Industrial Economics 26, 131-146.
Simintzi, E., Vig, V., Volpin, P., 2012. Labor bargaining power and access to finance. Working paper, London Business School.
Stevens, L. K., 2001. Trade secrets and inevitable disclosure. Tort & Insurance Law Journal 36, 917-948.
Sutton, J., 1991. Sunk costs and market structure. MIT Press, Cambridge, MA.
Telser, L. G., 1966. Cutthroat competition and the long purse. Journal of Law and Economics 9, 259-77.
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. Working paper, Lee & Hayes, PLLC.
Welch, I., 2004. Capital structure and stock returns. Journal of Political Economy 112, 106-132.
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.
42
Table I Precedent-Setting Legal Cases Recognizing the Inevitable Disclosure Doctrine
The table lists the precedent-setting legal cases in which state courts recognized the applicability of the Inevitable Disclosure Doctrine (IDD) or rejected it after recognizing it. The states omitted from the table did not consider or considered but rejected the IDD. The Appendix contains a brief discussion of each case.
State Precedent-Setting Case(s) Date Decision
AR Southwestern Energy Co. v. Eickenhorst, 955 F. Supp. 1078 (W.D. Ark. 1997) 3/18/1997 Recognize
CT Branson Ultrasonics Corp. v. Stratman, 921 F. Supp. 909 (D. Conn. 1996) 2/28/1996 Recognize
DE E.I. duPont deNemours & Co. v. American Potash & Chem. Corp., 200 A.2d 428 (Del. Ch. 1964) 5/5/1964 Recognize
FL Fountain v. Hudson Cush-N-Foam Corp., 122 So. 2d 232 (Fla. Dist. Ct. App. 1960) Del Monte Fresh Produce Co. v. Dole Food Co., Inc., 148 F. Supp. 2d 1326 (S.D. Fla. 2001)
7/11/1960 5/21/2001
Recognize Reject
GA Essex Group Inc. v. Southwire Co., 501 S.E.2d 501 (Ga. 1998) 6/29/1998 Recognize
IL Teradyne Inc. v. Clear Communications Corp., 707 F. Supp. 353 (N.D. 111. 1989) 2/9/1989 Recognize
IN Schlumberger Well Services v. Blaker, 623 F. Supp. 1310 (S.D. Ind. 1985) 1/23/1985 Recognize
IA Uncle B’s Bakery v. O’Rourke, 920 F. Supp. 1405 (N.D. Iowa 1996) 4/1/1996 Recognize
KS Bradbury Co. v. Teissier-duCros, 413 F. Supp. 2d 1203 (D. Kan. 2006) 2/2/2006 Recognize
MA Bard v. Intoccia, 1994 U.S. Dist. LEXIS 15368 (D. Mass. 1994) 10/13/1994 Recognize
MI Allis-Chalmers Manuf. Co. v. Continental Aviation & Eng. Corp., 255 F. Supp. 645 (E.D. Mich. 1966) CMI Int’l, Inc. v. Intermet Int’l Corp., 649 N.W.2d 808 (Mich. Ct. App. 2002)
2/17/1966 4/30/2002
Recognize Reject
MN Surgidev Corp. v. Eye Technology Inc., 648 F. Supp. 661 (D. Minn. 1986) 10/10/1986 Recognize
MO H & R Block Eastern Tax Servs. Inc. v. Enchura, 122 F. Supp. 2d 1067 (W.D. Mo. 2000) 11/2/2000 Recognize
NJ Nat’l Starch & Chem. Corp. v. Parker Chem. Corp., 530 A.2d 31 (N.J. Super. Ct. 1987) 4/27/1987 Recognize
NY Eastman Kodak Co. v. Powers Film Prod., 189 A.D. 556 (N.Y.A.D. 1919) 12/5/1919 Recognize
NC Travenol Laboratories Inc. v. Turner, 228 S.E.2d 478 (N.C. Ct. App. 1976) 6/17/1976 Recognize
OH Procter & Gamble Co. v. Stoneham, 747 N.E.2d 268 (Ohio Ct. App. 2000) 9/29/2000 Recognize
PA Air Products & Chemical Inc. v. Johnson, 442 A.2d 1114 (Pa. Super. Ct. 1982) 2/19/1982 Recognize
TX Rugen v. Interactive Business Systems Inc., 864 S.W.2d 548 (Tex. App. 1993) Cardinal Health Staffing Network Inc. v. Bowen, 106 S.W.3d 230 (Tex. App. 2003)
5/28/1993 4/3/2003
Recognize Reject
UT Novell Inc. v. Timpanogos Research Group Inc., 46 U.S.P.Q.2d 1197 (Utah D.C. 1998) 1/30/1998 Recognize
WA Solutec Corp, Inc. v. Agnew, 88 Wash. App. 1067 (Wash. Ct. App. 1997) 12/30/1997 Recognize
43
Table II Summary Statistics
This table reports summary statistics for the main variables in our regression models. The sample consists of industrial firms (utilities and financials are excluded) during the period 1977-2011 and includes 134,428 firm-year observations. All continuous variables are winsorized at their 1st and 99th percentiles and dollar values are expressed in 2009 dollars. Variable definitions refer to Compustat designations when appropriate. Book Leverage is the book value of long-term debt (dltt) plus debt in current liabilities (dlc) divided by book value of assets (at). Market Leverage is the book value of long-term debt (dltt) plus debt in current liabilities (dlc) divided by market value of assets (book assets (at) plus market value of equity (prcc_f*csho) minus book value of equity (ceq)). Inevitable Disclosure is equal to one if the state where the firm is headquartered has recognized the Inevitable Disclosure Doctrine (IDD) by year t and zero otherwise. Book Assets is total assets (at, in $ millions). Market-to-Book is market value of assets (book assets (at) plus market value of equity (prcc_f*csho) minus book value of equity (ceq)) divided by book value of assets (at). Return on Assets is operating income before depreciation (oibdp) divided by book value of assets (at). Fixed Assets is the ratio of the book value of property, plant, and equipment (ppent) to book value of assets (at). Industry Cash Flow Volatility is the median of the standard deviations of the Return on Assets over the previous ten years for firms in the same 2-digit SIC industry (firms are required to have at least three years of data to enter the calculation). Dividend Payer is equal to one if the firm pays a common dividend (dvc) during the fiscal year and zero otherwise. State GDP Growth is the state-level GDP growth rate over the year. Political Balance is the fraction of House of Representatives that belong to the Democrat Party for a given state and year.
Mean Std. Dev. P25 Median P75
Dependent Variables
Book Leverage 0.23 0.21 0.04 0.20 0.36
Market Leverage 0.18 0.18 0.02 0.13 0.28
Main Explanatory Variable
Inevitable Disclosure 0.42 0.49 0.00 0.00 1.00
Control Variables
Book Assets 1322 4097 37.93 146.0 639.9
Market-to-Book 1.99 1.75 1.04 1.40 2.18
Return on Assets 0.06 0.23 0.03 0.11 0.17
Fixed Assets 0.28 0.22 0.11 0.23 0.41
Industry Cash Flow Volatility 0.07 0.03 0.05 0.07 0.09
Dividend Payer 0.32 0.47 0.00 0.00 1.00
State GDP Growth 0.06 0.04 0.04 0.06 0.09
Political Balance 0.57 0.18 0.50 0.58 0.64
44
Table III Inevitable Disclosure Doctrine and Financial Leverage
This table reports the results from OLS regressions in which the dependent variable is financial leverage (Book Leverage in models 1-3 and Market Leverage in models 4-6), the key variable of interest is Inevitable Disclosure, and the control variables are Log Book Assets, Market-to-Book, Return on Assets, Fixed Assets, Industry Cash Flow Volatility, Dividend Payer, State GDP Growth, and Political Balance. All variables are defined in Table II. All specifications include firm fixed effects and year fixed effects. The coefficient on Inevitable Disclosure captures the difference-in-difference estimate of the impact the recognition of the IDD has on financial leverage. 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% level, respectively.
Book Leverage Market Leverage
(1) (2) (3) (4) (5) (6)
Inevitable Disclosure 0.013*** 0.014*** 0.014*** 0.008** 0.010*** 0.011*** (3.61) (3.75) (3.73) (2.38) (-2.96) (3.28)
Log Book Assets 0.030*** 0.030*** 0.035*** 0.035*** (10.03) (10.00) (12.15) (12.10)
Market-to-Book -0.004*** -0.004*** -0.019*** -0.019*** (-6.10) (-6.08) (-9.15) (-9.21)
Return on Assets -0.164*** -0.164*** -0.130*** -0.129*** (-15.42) (-15.54) (-9.22) (-9.37)
Fixed Assets 0.244*** 0.244*** 0.187*** 0.187*** (18.39) (18.36) (21.53) (21.53)
Industry Cash Flow Volatility -0.098 -0.099 -0.153*** -0.156*** (-1.61) (-1.60) (-3.07) (-3.05)
Dividend Payer -0.050*** -0.050*** -0.045*** -0.045*** (-13.47) (-13.45) (-14.85) (-14.79)
State GDP Growth -0.049 -0.198*** (-1.58) (-5.38)
Political Balance -0.001 -0.010* (-0.09) (-1.89)
Year Fixed Effects Yes Yes Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes Yes Yes
Observations 134,428 134,428 134,428 134,428 134,428 134,428
Adjusted R2 0.597 0.628 0.628 0.623 0.677 0.678
45
Table IV Alternative Measures of Financial Leverage
This table reports the results from OLS regressions in which the dependent variable is financial leverage (Net Book Leverage, Net Market Leverage, Long-Term Book Leverage, or Long-Term Market Leverage), the key variable of interest is Inevitable Disclosure, and the control variables are Log Book Assets, Market-to-Book, Return on Assets, Fixed Assets, Industry Cash Flow Volatility, Dividend Payer, State GDP Growth, and Political Balance. Net Book Leverage is the book value of long-term debt (dltt) plus debt in current liabilities (dlc) less book value of cash and short-term investments (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) less book value of cash and short-term investments (che) divided by market value of assets (at + prcc_f*csho - ceq). 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 (at + prcc_f*csho - ceq). All other variables are defined in Table II. All specifications include firm fixed effects and year 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% level, respectively.
Net Book Leverage
Net Market Leverage
Long-Term Book Leverage
Long-Term Market Leverage
(1) (2) (3) (4)
Inevitable Disclosure 0.017*** 0.015*** 0.012*** 0.009*** (3.49) (3.57) (3.60) (3.18)
Log Book Assets 0.042*** 0.047*** 0.030*** 0.032*** (10.87) (12.67) (10.10) (11.51)
Market-to-Book -0.016*** 0.007** -0.004*** -0.015*** (-19.75) (2.41) (-6.17) (-8.79)
Return on Assets -0.179*** -0.096*** -0.105*** -0.091*** (-15.55) (-6.22) (-11.60) (-8.57)
Fixed Assets 0.707*** 0.444*** 0.237*** 0.184*** (18.75) (32.87) (21.69) (21.94)
Industry Cash Flow Volatility -0.139 -0.320*** -0.090 -0.141*** (-1.57) (-4.51) (-1.33) (-2.71)
Dividend Payer -0.060*** -0.053*** -0.047*** -0.042*** (-16.16) (-15.90) (-13.22) (-14.55)
State GDP Growth -0.072 -0.203*** -0.032 -0.167*** (-1.67) (-4.93) (-1.28) (-5.35)
Political Balance -0.006 -0.018** -0.003 -0.011** (-0.53) (-2.08) (-0.42) (-2.05)
Year Fixed Effects Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes
Observations 134,428 134,428 134,428 134,428
Adjusted R2 0.723 0.681 0.627 0.664
46
Table V Timing of Changes in Financial Leverage
This table reports the results from OLS regressions in which the dependent variable is financial leverage (Book Leverage in models 1-3 and Market Leverage in models 4-6), the key variables of interest are Inevitable Disclosure1, Inevitable Disclosure0, Inevitable Disclosure1, and Inevitable Disclosure2+, and the control variables are Log Book Assets, Market-to-Book, Return on Assets, Fixed Assets, Industry Cash Flow Volatility, Dividend Payer, State GDP Growth, and Political Balance. Inevitable Disclosure-1 is equal to one if the firm is headquartered in a state that will recognize the IDD in one year. Inevitable Disclosure0 is equal to one if the firm is headquartered in a state that recognizes the IDD in the current year. Inevitable Disclosure1 is equal to one if the firm is headquartered in a state that recognized the IDD one year ago. Inevitable Disclosure2+ is equal to one if the firm is headquartered in a state that recognized the IDD two or more years ago. All other variables are defined in Table II. All specifications include firm fixed effects and year fixed effects. The sample excludes all observations for firms in Florida, Michigan, and Texas starting the year in which they rejected the previously recognized IDD. 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% level, respectively.
Book Leverage Market Leverage
(1) (2) (3) (4) (5) (6)
Inevitable Disclosure-1 0.005 0.006 0.006 0.003 0.003 0.002 (1.06) (1.35) (1.36) (0.49) (0.52) (0.54)
Inevitable Disclosure0 0.003 0.004 0.004 0.001 0.003 0.004 (0.58) (0.84) (0.89) (0.24) (0.51) (0.74)
Inevitable Disclosure1 0.015*** 0.015*** 0.015*** 0.010** 0.011** 0.012*** (4.10) (3.52) (3.56) (2.40) (2.44) (3.08)
Inevitable Disclosure2+ 0.017*** 0.017*** 0.017*** 0.010** 0.012*** 0.013*** (3.18) (3.53) (3.50) (2.25) (2.80) (3.16)
Log Book Assets 0.031*** 0.031*** 0.035*** 0.035*** (9.24) (9.22) (10.96) (10.90)
Market-to-Book -0.004*** -0.004*** -0.019*** -0.019*** (-6.13) (-6.12) (-9.41) (-9.48)
Return on Assets -0.164*** -0.164*** -0.129*** -0.129*** (-15.05) (-15.17) (-9.02) (-9.16)
Fixed Assets 0.244*** 0.244*** 0.188*** 0.188*** (17.29) (17.26) (19.93) (20.00)
Industry Cash Flow Volatility -0.068 -0.068 -0.142*** -0.145*** (-1.23) (-1.23) (-2.82) (-2.86)
Dividend Payer -0.050*** -0.050*** -0.045*** -0.045*** (-12.84) (-12.81) (-14.30) (-14.25)
State GDP Growth -0.050 -0.199*** (-1.65) (-5.19)
Political Balance -0.005 -0.014*** (-0.73) (-2.74)
Year Fixed Effects Yes Yes Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes Yes Yes
Observations 129,451 129,451 129,451 129,451 129,451 129,451
Adjusted R2 0.601 0.632 0.632 0.628 0.682 0.682
47
Table VI Effect of Competition in Product Markets
This table reports the results from OLS regressions in which the dependent variable is financial leverage (Book Leverage in Panel A and Market Leverage in Panel B), the key variable of interest is Inevitable Disclosure, and the control variables are Log Book Assets, Market-to-Book, Return on Assets, Fixed Assets, Industry Cash Flow Volatility, Dividend Payer, State GDP Growth, and Political Balance. All variables are defined in Table II. All specifications include firm fixed effects and year fixed effects. In both panels, we split the sample according to whether the values of selected industry characteristics are below or above the sample median. The first characteristic is the Four Firm Concentration Ratio, defined as the fraction of total industry sales captured by the four largest firms in a 5-digit NAICS industry from the U.S. Economic Census. The second is Barriers to Entry, defined as the average value of firm’s R&D expenses (xrd) plus advertising expenses (xad) divided by sales (sale) across all firms in a 3-digit SIC industry. In columns 1-2, the sample excludes four industries for which the Census concentration data is not available (Agriculture, forestry, fishing, and hunting, Mining, Construction, and Management of company enterprises). 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% level, respectively.
Panel A: Dependent Variable is Book Leverage
Four Firm Concentration Ratio Barriers to Entry
Below Median Above Median Below Median Above Median
(1) (2) (3) (4)
Inevitable Disclosure 0.016** 0.003 0.018*** 0.004 (2.25) (0.59) (3.28) (0.60)
Log Book Assets 0.037*** 0.021*** 0.040*** 0.022*** (6.88) (4.41) (14.61) (4.69)
Market-to-Book -0.005*** -0.003** -0.005*** -0.003*** (-3.67) (-2.21) (-3.46) (-4.61)
Return on Assets -0.208*** -0.165*** -0.195*** -0.144*** (-21.79) (-10.83) (-20.13) (-12.61)
Fixed Assets 0.235*** 0.209*** 0.232*** 0.263*** (10.28) (9.07) (12.49) (16.79)
Industry Cash Flow Volatility -0.430*** 0.101 -0.146 0.045 (-4.35) (0.60) (-1.59) (0.53)
Dividend Payer -0.052*** -0.048*** -0.057*** -0.040*** (-9.19) (-10.99) (-11.50) (-8.88)
State GDP Growth -0.037 -0.005 -0.107*** 0.031 (-1.23) (-0.08) (-3.06) (0.80)
Political Balance 0.000 0.015 0.012 -0.016 (0.03) (1.49) (1.00) (-1.50)
Year Fixed Effects Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes
Observations 41,826 39,571 67,215 67,213
Adjusted R2 0.686 0.630 0.660 0.597
48
Table VI - (Continued)
Panel B: Dependent Variable is Market Leverage
Four Firm Concentration Ratio Barriers to Entry
Below Median Above Median Below Median Above Median
(1) (2) (3) (4)
Inevitable Disclosure 0.015** 0.006 0.015*** 0.003 (2.37) (1.47) (2.93) (0.71)
Log Book Assets 0.045*** 0.029*** 0.044*** 0.025*** (10.72) (8.41) (20.95) (6.42)
Market-to-Book -0.025*** -0.017*** -0.030*** -0.014*** (-11.40) (-9.87) (-14.34) (-10.61)
Return on Assets -0.188*** -0.135*** -0.183*** -0.097*** (-12.92) (-7.62) (-14.97) (-8.64)
Fixed Assets 0.203*** 0.150*** 0.193*** 0.173*** (11.84) (10.63) (14.70) (15.31)
Industry Cash Flow Volatility -0.311*** -0.117 -0.098 -0.084 (-2.94) (-0.90) (-1.22) (-1.34)
Dividend Payer -0.047*** -0.044*** -0.051*** -0.034*** (-11.19) (-13.60) (-13.90) (-10.45)
State GDP Growth -0.150*** -0.127** -0.246*** -0.085*** (-5.54) (-2.23) (-5.80) (-2.89)
Political Balance -0.008 0.001 -0.001 -0.013* (-0.90) (0.11) (-0.10) (-1.79)
Year Fixed Effects Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes
Observations 41,826 39,571 67,215 67,213
Adjusted R2 0.720 0.698 0.685 0.668
49
Table VII Effect of Employee Characteristics
This table reports the results from OLS regressions in which the dependent variable is financial leverage (Book Leverage in Panel A and Market Leverage in Panel B), the key variable of interest is Inevitable Disclosure, and the control variables are Log Book Assets, Market-to-Book, Return on Assets, Fixed Assets, Industry Cash Flow Volatility, Dividend Payer, State GDP Growth, and Political Balance. All variables are defined in Table II. All specifications include firm fixed effects and year fixed effects. In both panels, we split the sample according to whether the values of selected industry characteristics are below or above the sample median. The first is the Fraction of Workers in Managerial Occupations, defined as is the fraction of workers employed in managerial occupations in the firm’s 3-digit NAICS industry and state. The second is the Fraction of Workers in Science Occupations, defined as the fraction of workers that are employed in science-related occupations in the firm’s 3-digit NAICS industry and state. The third is the Fraction of Workers with a Bachelor’s Degree, defined as the fraction of workers with at least a bachelor’s degree that are employed in the firm’s 3-digit NAICS industry and state. 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% level, respectively.
Panel A: Dependent Variable is Book Leverage
Fraction of Workers in
Managerial Occupations Fraction of Workers in Science Occupations
Fraction of Workers with a Bachelor’s Degree
Above Median
Below Median
Above Median
Below Median
Above Median
Below Median
(1) (2) (3) (4) (5) (6)
Inevitable Disclosure 0.021*** 0.005 0.017*** 0.008 0.015*** 0.006 (6.46) (0.78) (3.47) (1.08) (2.73) (0.77)
Log Book Assets 0.026*** 0.035*** 0.025*** 0.037*** 0.025*** 0.037*** (6.04) (13.18) (5.80) (12.98) (6.15) (12.18)
Market-to-Book -0.003*** -0.006*** -0.003*** -0.006*** -0.003*** -0.005*** (-6.34) (-5.32) (-5.38) (-4.17) (-4.02) (-4.50)
Return on Assets -0.145*** -0.192*** -0.140*** -0.209*** -0.141*** -0.209*** (-10.10) (-19.72) (-11.60) (-22.55) (-15.27) (-18.64)
Fixed Assets 0.267*** 0.218*** 0.276*** 0.207*** 0.289*** 0.198*** (22.26) (12.87) (22.31) (11.48) (33.37) (10.49)
Industry Cash Flow Volatility -0.003 -0.206** 0.045 -0.233** 0.128 -0.289** (-0.02) (-2.26) (0.47) (-2.42) (1.14) (-2.64)
Dividend Payer -0.040*** -0.052*** -0.038*** -0.054*** -0.037*** -0.054*** (-6.63) (-9.26) (-6.49) (-9.99) (-6.84) (-11.24)
State GDP Growth -0.072 -0.019 -0.049 -0.041 -0.034 -0.069* (-1.38) (-0.69) (-0.95) (-1.52) (-0.73) (-2.00)
Political Balance 0.000 -0.005 0.004 -0.001 0.005 -0.003 (0.03) (-0.45) (0.35) (-0.06) (0.40) (-0.29)
Year Fixed Effects Yes Yes Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes Yes Yes
Observations 65,539 65,754 65,631 65,662 65,583 65,710
Adjusted R2 0.611 0.656 0.605 0.660 0.618 0.649
50
Table VII - (Continued)
Panel B: Dependent Variable is Market Leverage
Fraction of Worker in
Managerial Occupations Fraction of Workers in Science Occupations
Fraction of Workers with a Bachelor’s Degree
Above Median
Below Median
Above Median
Below Median
Above Median
Below Median
(1) (2) (3) (4) (5) (6)
Inevitable Disclosure 0.014*** 0.008 0.011*** 0.008 0.010*** 0.006 (5.31) (1.35) (3.00) (1.43) (2.70) (1.06)
Log Book Assets 0.029*** 0.041*** 0.028*** 0.044*** 0.028*** 0.044*** (7.19) (17.39) (7.41) (19.01) (8.52) (18.44)
Market-to-Book -0.015*** -0.027*** -0.015*** -0.027*** -0.015*** -0.026*** (-9.48) (-18.18) (-9.52) (-13.71) (-9.77) (-10.51)
Return on Assets -0.096*** -0.185*** -0.092*** -0.198*** -0.090*** -0.204*** (-7.73) (-17.00) (-8.54) (-13.91) (-10.70) (-11.45)
Fixed Assets 0.187*** 0.179*** 0.193*** 0.170*** 0.203*** 0.165*** (25.38) (13.55) (20.43) (13.62) (22.02) (12.19)
Industry Cash Flow Volatility -0.067 -0.257*** -0.067 -0.164* -0.008 -0.255** (-0.85) (-2.94) (-0.86) (-1.84) (-0.11) (-2.62)
Dividend Payer -0.034*** -0.048*** -0.033*** -0.049*** -0.032*** -0.049*** (-7.92) (-12.00) (-7.31) (-11.91) (-7.88) (-14.07)
State GDP Growth -0.209*** -0.142*** -0.207*** -0.142*** -0.193*** -0.172*** (-3.00) (-6.52) (-3.39) (-5.08) (-2.89) (-5.23)
Political Balance -0.005 -0.011 -0.007 -0.009 -0.009 -0.009 (-0.65) (-1.39) (-0.86) (-1.24) (-1.10) (-1.18)
Year Fixed Effects Yes Yes Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes Yes Yes
Observations 65,539 65,754 65,631 65,662 65,583 65,710
Adjusted R2 0.658 0.693 0.646 0.697 0.663 0.689
51
Table VIII Effect of Ex-Ante Employee Mobility
This table reports the results from OLS regressions in which the dependent variable is financial leverage (Book Leverage in Panel A and Market Leverage in Panel B), the key variable of interest is Inevitable Disclosure, and the control variables are Log Book Assets, Market-to-Book, Return on Assets, Fixed Assets, Industry Cash Flow Volatility, Dividend Payer, State GDP Growth, and Political Balance. All variables are defined in Table II. All specifications include firm fixed effects and year fixed effects. In columns 1-2 of both panels, we split the sample according to whether a firm has a Defined Benefit Pension Plan or not. We define a firm as having a defined benefit pension plan if it reports positive pension plan assets (pbnna>0) or accumulated obligations (pbaco>0). In columns 3-4, we split the sample according to whether the Employee Market Share, defined as the firm’s share in the 2-digit SIC industry’s employment located in the firm’s state, is below or above the sample median. 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% level, respectively.
Panel A: Dependent Variable is Book Leverage
Defined Benefit Pension Plan Employee Market Share
No Yes Below Median Above Median
(1) (2) (3) (4)
Inevitable Disclosure 0.014*** 0.005 0.016*** 0.007 (3.95) (0.54) (4.23) (1.27)
Log Book Assets 0.031*** 0.036*** 0.028*** 0.031*** (8.82) (6.60) (5.01) (14.43)
Market-to-Book -0.005*** 0.003 -0.003*** -0.007*** (-5.04) (1.25) (-4.33) (-5.62)
Return on Assets -0.151*** -0.352*** -0.135*** -0.237*** (-15.03) (-16.40) (-11.85) (-14.30)
Fixed Assets 0.275*** 0.099*** 0.307*** 0.171*** (19.96) (3.54) (18.18) (8.24)
Industry Cash Flow Volatility -0.138* 0.082 -0.058 -0.081 (-1.93) (0.71) (-0.62) (-1.01)
Dividend Payer -0.040*** -0.057*** -0.041*** -0.051*** (-11.98) (-8.79) (-7.95) (-10.29)
State GDP Growth -0.110*** 0.015 -0.058 -0.016 (-3.57) (0.35) (-0.96) (-0.61)
Political Balance -0.002 -0.002 -0.006 0.001 (-0.20) (-0.13) (-0.53) (0.16)
Year Fixed Effects Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes
Observations 98,124 28,157 65,850 65,851
Adjusted R2 0.639 0.660 0.612 0.673
52
Table VIII - (Continued)
Panel B: Dependent Variable is Market Leverage
Defined Benefit Pension Plan Employee Market Share
No Yes Below Median Above Median
(1) (2) (3) (4)
Inevitable Disclosure 0.012*** 0.004 0.012*** 0.006 (4.16) (0.66) (5.15) (1.26)
Log Book Assets 0.032*** 0.042*** 0.031*** 0.037*** (10.42) (9.22) (5.58) (23.66)
Market-to-Book -0.017*** -0.034*** -0.014*** -0.029*** (-9.03) (-11.05) (-8.75) (-18.77)
Return on Assets -0.104*** -0.388*** -0.088*** -0.248*** (-9.50) (-18.14) (-7.44) (-16.09)
Fixed Assets 0.196*** 0.115*** 0.218*** 0.151*** (24.59) (4.99) (22.52) (9.04)
Industry Cash Flow Volatility -0.213*** 0.016 -0.149* -0.137* (-4.09) (0.19) (-1.89) (-1.85)
Dividend Payer -0.040*** -0.045*** -0.038*** -0.046*** (-13.86) (-10.41) (-8.46) (-11.20)
State GDP Growth -0.238*** -0.118** -0.271*** -0.102*** (-6.40) (-2.51) (-4.21) (-4.47)
Political Balance -0.015** -0.000 -0.013 -0.009 (-2.47) (-0.02) (-1.57) (-1.43)
Year Fixed Effects Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes
Observations 98,124 28,157 65,850 65,851
Adjusted R2 0.682 0.715 0.668 0.710
53
Table IX Inevitable Disclosure Doctrine and Credit Spreads
This table reports the results from OLS regressions in which the dependent variable is Log Loan Spread and the key variables of interest are Inevitable Disclosure (in Panel A) and Inevitable Disclosure-1, Inevitable Disclosure0, Inevitable Disclosure1, and Inevitable Disclosure2+ (in Panel B). The control variables are Log Book Assets, Market-to-Book, Return on Assets, Fixed Assets, Industry Cash Flow Volatility, Dividend Payer, Log Loan Maturity, Log Loan Size, State GDP Growth, and Political Balance. Log Loan Spread is the logarithm of the amount the borrower pays over LIBOR for each dollar drawn down in basis points. Log Loan Maturity is the logarithm of the number of months until the loan matures. Log Loan Size is the logarithm of the loan amount (in millions). Inevitable Disclosure-1 is equal to one if the firm is headquartered in a state that will recognize the IDD in one year. Inevitable Disclosure0 is equal to one if the firm is headquartered in a state that recognizes the IDD in the current year. Inevitable Disclosure1 is equal to one if the firm is headquartered in a state that recognized the IDD one year ago. Inevitable Disclosure2+ is equal to one if the firm is headquartered in a state that recognized the IDD two or more years ago. All other variables are defined in Table II. The sample period is 1987-2011. All specifications include state fixed effects, 3-digit SIC industry fixed effects, and year fixed effects. Models 2-3 further include loan type 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% level, respectively. Panel A: Inevitable Disclosure and Credit Spreads (1) (2) (3)
Inevitable Disclosure -0.068*** -0.061*** -0.057** (-2.85) (-2.72) (-2.36)
Log Book Assets -0.220*** -0.142*** -0.142*** (-33.66) (-20.68) (-20.60)
Market-to-Book -0.079*** -0.061*** -0.061*** (-7.61) (-7.80) (-7.85)
Return on Assets -1.215*** -1.131*** -1.132*** (-10.37) (-10.70) (-10.73)
Fixed Assets -0.236*** -0.178*** -0.178*** (-4.28) (-3.53) (-3.54)
Industry Cash Flow Volatility 0.924** 1.276*** 1.257*** (2.23) (3.59) (3.56)
Dividend Payer -0.376*** -0.317*** -0.317*** (-18.86) (-18.85) (-18.67)
Book Leverage 1.105*** 0.884*** 0.883*** (33.12) (38.64) (38.67)
Log Loan Maturity 0.014 0.014 (1.37) (1.36)
Log Loan Size -0.082*** -0.082*** (-15.33) (-15.29)
State GDP Growth 0.334 (1.08)
Political Balance -0.034 (-0.76)
Year Fixed Effects Yes Yes Yes Industry Fixed Effects Yes Yes Yes State Fixed Effects Yes Yes Yes Loan Type Fixed Effects No Yes Yes Observations 25,017 25,017 25,017 Adjusted R2 0.555 0.619 0.619
54
Table IX - (Continued)
Panel B: Inevitable Disclosure and the Timing of Changes in Credit Spreads
(1) (2) (3)
Inevitable Disclosure-1 -0.012 -0.002 0.001 (-0.33) (-0.07) (0.04)
Inevitable Disclosure0 0.021 0.016 0.017 (0.76) (0.74) (0.71)
Inevitable Disclosure1 -0.040 -0.036 -0.036 (-0.63) (-0.65) (-0.63)
Inevitable Disclosure2+ -0.076*** -0.060** -0.056** (-3.07) (-2.40) (-2.21)
Log Book Assets -0.223*** -0.147*** -0.147*** (-33.30) (-18.60) (-18.52)
Market-to-Book -0.076*** -0.059*** -0.060*** (-7.22) (-7.30) (-7.42)
Return on Assets -1.237*** -1.141*** -1.140*** (-10.27) (-10.44) (-10.44)
Fixed Assets -0.255*** -0.197*** -0.197*** (-4.67) (-4.16) (-4.17)
Industry Cash Flow Volatility 0.992** 1.323*** 1.315*** (2.08) (3.54) (3.54)
Dividend Payer -0.379*** -0.319*** -0.319*** (-18.05) (-18.28) (-18.20)
Book Leverage 1.103*** 0.883*** 0.883*** (29.67) (34.15) (34.26)
Log Loan Maturity 0.016 0.016 (1.50) (1.47)
Log Loan Size -0.078*** -0.078*** (-13.66) (-13.64)
State GDP Growth 0.539* (1.81)
Political Balance -0.011 (-0.21)
Year Fixed Effects Yes Yes Yes
Industry Fixed Effects Yes Yes Yes
State Fixed Effects Yes Yes Yes
Loan Type Fixed Effects No Yes Yes
Observations 23,173 23,173 23,173
Adjusted R2 0.558 0.621 0.621
55
Table X Inevitable Disclosure Doctrine and Sales Growth
This table reports the results from OLS regressions in which the dependent variable is Sales Growth (Salest / Salest-1 – 1), the key variable of interest is Inevitable Disclosure, and the control variables are Log Book Assets, Return on Assets, Market-to-Book, Capital Expenditures, R&D Expenditures, Advertising Expenditures, and Book Leverage. Capital Expenditures is capital expenditures (capex) divided by book assets (at). R&D Expenditures is R&D expenses (xrd) divided by sales (sale). Advertising Expenditures is advertising expenses (xad) divided by sales (sale). All other variables are defined in Table II. The sample spans the period 1977-2011. All specifications include firm fixed effects and interacted year times industry fixed effects. In models 1 and 2 industries are defined at the 3-digit SIC level and in models 3 and 4 they are defined at the 4-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% level, respectively.
(1) (2) (3) (4)
Inevitable Disclosure 0.017*** 0.017*** 0.018*** 0.018*** (2.81) (2.84) (2.70) (2.71)
Log Assets 0.057*** 0.058*** 0.064*** 0.064*** (10.50) (11.32) (10.45) (11.21)
Return on Assets 0.463*** 0.457*** 0.453*** 0.447*** (28.38) (30.56) (27.59) (30.39)
Market-to-Book 0.048*** 0.048*** 0.048*** 0.048*** (14.75) (14.85) (13.08) (13.18)
Capital Expenditures 0.604*** 0.603*** 0.595*** 0.594*** (13.08) (13.02) (13.10) (13.06)
R&D Expenditures -0.091*** -0.092*** -0.094*** -0.094*** (-15.32) (-15.19) (-15.68) (-15.46)
Advertising Expenditures 0.444*** 0.438*** 0.487*** 0.482*** (3.08) (3.08) (3.00) (3.00)
Book Leverage -0.033* -0.028 (-1.93) (-1.41)
Firm Fixed Effects Yes Yes Yes Yes
Year Industry Fixed Effects Yes Yes Yes Yes
Industry Definition 3-Digit SIC 3-Digit SIC 4-Digit SIC 4-Digit SIC
Observations 131,979 131,979 131,979 131,979
Adjusted R2 0.261 0.261 0.226 0.226
56
Table XI Recognition of the Inevitable Disclosure Doctrine in States Harboring Industry Rivals
This table explores the impact on a firm’s market share and capital structure due to an increase in the protection of the trade secrets of the firm’s out-of-state industry rivals, which we gauge using Out-of-State Rivals’ Inevitable Disclosure. For each 3-digit SIC industry and state, Out-of-State Rivals’ Inevitable Disclosure is defined as the weighted-average IDD of out-of-state industry rivals (with weights equal to each state’s share in the total sales of out-of-state industry rivals) multiplied by the share of out-of-state industry rivals in the total sales of the industry (see text for formal definition). Panel A reports the results from OLS regressions in which the dependent variable is Sales Growth (Salest / Salest-1 – 1), the key variable of interest is Out-of-State Rivals’ Inevitable Disclosure, and the control variables are Inevitable Disclosure, Log Book Assets, Return on Assets, Market-to-Book, Capital Expenditures, R&D Expenditures, Advertising Expenditures, and Book Leverage. All specifications include firm fixed effects and interacted year times industry fixed effects. To calculate these fixed effects, in models 1 and 2 industries are defined at the 3-digit SIC level and in models 3 and 4 they are defined at the 4-digit SIC level. Panel B reports the results from OLS regressions in which the dependent variable is financial leverage (Book Leverage or Market Leverage), the key variable of interest is Out-of-State Rivals’ Inevitable Disclosure, and the control variables are Inevitable Disclosure, Log Book Assets, Market-to-Book, Return on Assets, Fixed Assets, Industry Cash Flow Volatility, Dividend Payer, State GDP Growth, and Political Balance. All specifications include firm fixed effects and year fixed effects. All other variables are defined in Tables II or X. In both panels, 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% level, respectively. Panel A: Effect on Sales Growth
(1) (2) (3) (4)
Out-of-State Rivals’ Inevitable Disclosure -0.018 0.001 -0.038* -0.019 (-1.03) (0.04) (-1.80) (-0.84)
Inevitable Disclosure 0.017*** 0.016** (2.75) (2.17)
Log Assets 0.058*** 0.058*** 0.064*** 0.064*** (11.33) (11.41) (11.13) (11.22)
Return on Assets 0.457*** 0.457*** 0.447*** 0.447*** (30.37) (30.53) (30.21) (30.35)
Market-to-Book 0.048*** 0.048*** 0.048*** 0.048*** (14.88) (14.85) (13.21) (13.18)
Capital Expenditures 0.603*** 0.603*** 0.594*** 0.594*** (13.06) (13.03) (13.07) (13.05)
R&D Expenditures -0.092*** -0.092*** -0.094*** -0.094*** (-15.18) (-15.19) (-15.43) (-15.45)
Advertising Expenditures 0.436*** 0.438*** 0.480*** 0.481*** (3.07) (3.09) (2.98) (3.00)
Book Leverage -0.033* -0.033* -0.028 -0.028 (-1.90) (-1.93) (-1.40) (-1.41)
Firm Fixed Effects Yes Yes Yes Yes
Year Industry Fixed Effects Yes Yes Yes Yes
Industry Definition 3-Digit SIC 3-Digit SIC 4-Digit SIC 4-Digit SIC
Observations 131,979 131,979 131,979 131,979
Adjusted R2 0.263 0.263 0.229 0.230
57
Table XI – (Continued)
Panel B: Effect on Financial Leverage
Book Leverage Market Leverage
(1) (2) (3) (4)
Out-of-State Rivals’ Inevitable Disclosure -0.005 -0.004 -0.007** -0.006* (-1.17) (-0.89) (-2.12) (-1.85)
Inevitable Disclosure 0.014*** 0.011** (3.68) (3.17)
Log Book Assets 0.030*** 0.030*** 0.035*** 0.035*** (9.96) (10.00) (12.13) (12.13)
Market-to-Book -0.004*** -0.004*** -0.019*** -0.019*** (-6.14) (-6.07) (-9.21) (-9.21)
Return on Assets -0.164*** -0.164*** -0.129*** -0.129*** (-15.45) (-15.55) (-9.35) (-9.38)
Fixed Assets 0.244*** 0.244*** 0.186*** 0.187*** (18.33) (18.38) (21.47) (21.53)
Industry Cash Flow Volatility -0.098 -0.097 -0.155*** -0.154*** (-1.54) (-1.58) (-2.99) (-3.01)
Dividend Payer -0.050*** -0.050*** -0.045*** -0.045*** (-13.50) (-13.47) (-14.90) (-14.87)
State GDP Growth -0.048 -0.049 -0.197*** -0.197*** (-1.54) (-1.58) (-5.33) (-5.40)
Political Balance 0.003 -0.001 -0.007 -0.010* (0.39) (-0.10) (-1.12) (-1.89)
Year Fixed Effects Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes
Observations 134,428 134,428 134,428 134,428
Adjusted R2 0.628 0.628 0.678 0.678
58
Table XII The Effect of Measurement Error in the Inevitable Disclosure Doctrine Index
This table reports the results from OLS regressions in which the dependent variable is financial leverage (Book Leverage in Panel A and Market Leverage in Panel B), the key variable of interest is Inevitable Disclosure, and the control variables are Log Book Assets, Market-to-Book, Return on Assets, Fixed Assets, Industry Cash Flow Volatility, Dividend Payer, State GDP Growth, and Political Balance. All variables are defined in Table II. All specifications include firm fixed effects and year fixed effects. In model 1, we correct the location of headquarters (HQ) over the period 1992-2011 to reduce measurement error in Inevitable Disclosure. To do so, we use the state of headquarters information from the 10-Ks over the 1992-2011 when it is available, and when it is not available we assume there were no relocations prior to the earliest date it is available. In model 2, we use only the subsample of firm-years for which we can confirm the location of headquarters from 10-K documents. This subsample spans only the period 1992-2011, but in this sample Inevitable Disclosure is not measured with any error due relocations of firms’ headquarters. In model 3, we exclude all firms whose annual sales or book asset growth exceeded 100% in any year during the sample period. In model 4, we exclude all observations when a firm reports positive foreign income or foreign taxes. In model 5, we exclude all observations when a firm is in a geographically dispersed industry (retail, wholesale, and transportation). 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% level, respectively. Panel A: Dependent Variable is Book Leverage
Corrected
location of HQ (1977-2011)
Corrected
location of HQ (1992-2011)
Exclude if firm
growth ever exceeds 100%
Exclude if firm reports foreign income or taxes
Exclude if firm is in dispersed
industry (1) (2) (3) (4) (5)
Inevitable Disclosure 0.016*** 0.015*** 0.010** 0.015*** 0.014*** (5.22) (3.01) (2.43) (3.96) (3.44)
Log Book Assets 0.030*** 0.026*** 0.023*** 0.032*** 0.032*** (11.48) (5.47) (5.50) (12.28) (9.56)
Market-to-Book -0.004*** -0.004*** -0.003*** -0.004*** -0.004*** (-4.97) (-4.38) (-2.47) (-3.86) (-5.58)
Return on Assets -0.164*** -0.138*** -0.249*** -0.156*** -0.165*** (-15.63) (-9.93) (-12.24) (-15.52) (-14.80)
Fixed Assets 0.244*** 0.212*** 0.198*** 0.276*** 0.252*** (19.30) (14.66) (12.79) (17.45) (18.04)
Industry Cash Flow Volatility -0.098* -0.100* -0.107 -0.019 -0.188** (-1.79) (-1.75) (-1.15) (-0.24) (-2.67)
Dividend Payer -0.049*** -0.031*** -0.050*** -0.052*** -0.047*** (-15.22) (-7.62) (-13.23) (-10.80) (-11.65)
State GDP Growth -0.046 -0.047 -0.018 -0.047 -0.052 (-1.66) (-1.64) (-0.68) (-1.23) (-1.67)
Political Balance -0.007 0.005 -0.007 -0.004 -0.006 (-0.74) (0.54) (-1.02) (-0.54) (-0.86)
Year Fixed Effects Yes Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes Yes
Observations 134,298 65,070 63,655 89,394 108,433
Adjusted R2 0.628 0.694 0.689 0.634 0.612
59
Table XII – (Continued)
Panel B: Dependent Variable is Market Leverage
Corrected
location of HQ (1977-2011)
Corrected
location of HQ (1992-2011)
Exclude if firm
growth ever exceeds 100%
Exclude if firm reports foreign income or taxes
Exclude if firm is in dispersed
industry (1) (2) (3) (4) (5)
Inevitable Disclosure 0.011*** 0.014*** 0.009** 0.011*** 0.011*** (3.85) (3.14) (2.35) (3.77) (2.97)
Log Book Assets 0.035*** 0.032*** 0.034*** 0.037*** 0.035*** (12.94) (7.06) (10.77) (17.47) (10.85)
Market-to-Book -0.019*** -0.015*** -0.033*** -0.019*** -0.017*** (-8.89) (-7.41) (-17.52) (-10.29) (-9.54)
Return on Assets -0.129*** -0.104*** -0.247*** -0.116*** -0.125*** (-9.31) (-6.55) (-11.24) (-10.03) (-9.26)
Fixed Assets 0.187*** 0.154*** 0.192*** 0.204*** 0.186*** (19.45) (14.91) (17.38) (23.29) (20.04)
Industry Cash Flow Volatility -0.157*** -0.203*** -0.050 -0.107** -0.211*** (-3.26) (-4.38) (-0.60) (-2.10) (-3.75)
Dividend Payer -0.045*** -0.032*** -0.045*** -0.049*** -0.043*** (-18.87) (-9.67) (-14.45) (-13.42) (-13.24)
State GDP Growth -0.193*** -0.137*** -0.133*** -0.220*** -0.214*** (-5.93) (-3.92) (-4.06) (-5.43) (-5.38)
Political Balance -0.014** -0.004 -0.009 -0.009 -0.013** (-2.01) (-0.55) (-1.33) (-1.44) (-2.49)
Year Fixed Effects Yes Yes Yes Yes Yes
Firm Fixed Effects Yes Yes Yes Yes Yes
Observations 134,298 65,070 63,655 89,394 108,433
Adjusted R2 0.678 0.725 0.719 0.684 0.673
60
Table XIII Additional Controls in Main Specification
This table reports the results from OLS regressions in which the dependent variable is financial leverage (Book Leverage in models 1-3 and Market Leverage in models 4-6) and the key variable of interest is Inevitable Disclosure. In addition to the control variables in our main specification (Log Book Assets, Market-to-Book, Return on Assets, Fixed Assets, Industry Cash Flow Volatility, Dividend Payer, State GDP Growth, and Political Balance), we sequentially include several other control variables defined next. State-Industry HHI is the sales-based Herfindhal-Hirschmann Index of concentration within the firm’s 2-digit SIC industry and state of headquarters. Strength of Trade Secret Protection is an index between zero and one that indicates the strength of trade secret protection in each state (higher values indicate greater protection). Strength of Non-Competes is an index between zero and twelve indicating the strength of the enforcement of covenants not to compete by courts in the state (higher values indicate stronger enforcement). Log Number of Patents is the logarithm of the number of patents owned by a firm. Log Number of Citations is the logarithm of the total number of citations of a firm’s patents. All other variables are defined in Table II. All specifications include firm fixed effects and year 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% level, respectively.
Book Leverage Market Leverage (1) (2) (3) (4) (5) (6) Inevitable Disclosure 0.014*** 0.015*** 0.012*** 0.011*** 0.011*** 0.008**
(3.79) (4.48) (2.88) (3.33) (4.23) (2.37) State-Industry HHI -0.036 -0.038 -0.026 -0.032 -0.032 -0.009
(-0.90) (-0.92) (-0.66) (-1.02) (-1.04) (-0.28) Strength of Trade Secret Protection -0.007 -0.005 0.000 0.001
(-0.95) (-0.79) (0.03) (0.16) Strength of Non-Competes 0.001 0.002 0.002 0.002
(0.43) (0.54) (0.56) (0.70) Log Number of Patents -0.011*** -0.012***
(-3.80) (-6.63) Log Number of Citations -0.000 0.001**
(-0.19) (2.17) Log Book Assets 0.030*** 0.030*** 0.037*** 0.035*** 0.035*** 0.041***
(10.00) (10.02) (11.46) (12.09) (12.09) (12.90) Market-to-Book -0.004*** -0.004*** -0.004*** -0.019*** -0.019*** -0.019***
(-6.08) (-6.05) (-5.10) (-9.21) (-9.21) (-9.60) Return on Assets -0.164*** -0.164*** -0.182*** -0.129*** -0.129*** -0.143***
(-15.54) (-15.55) (-18.63) (-9.37) (-9.38) (-10.16) Fixed Assets 0.244*** 0.244*** 0.262*** 0.187*** 0.187*** 0.202***
(18.37) (18.35) (19.79) (21.54) (21.54) (22.38) Industry Cash Flow Volatility -0.101 -0.101 -0.114 -0.158*** -0.158*** -0.184***
(-1.65) (-1.67) (-1.54) (-3.10) (-3.12) (-2.85) Dividend Payer -0.050*** -0.050*** -0.054*** -0.045*** -0.045*** -0.047***
(-13.45) (-13.42) (-14.45) (-14.80) (-14.72) (-14.71) State GDP Growth -0.049 -0.048 -0.039 -0.197*** -0.198*** -0.218***
(-1.60) (-1.57) (-1.16) (-5.47) (-5.46) (-6.13) Political Balance -0.000 -0.001 -0.007 -0.010* -0.010* -0.014***
(-0.03) (-0.11) (-1.06) (-1.80) (-1.84) (-2.97) Year Fixed Effects Yes Yes Yes Yes Yes Yes Firm Fixed Effects Yes Yes Yes Yes Yes Yes Observations 134,428 134,428 102,253 134,428 134,428 102,253 Adjusted R2 0.628 0.628 0.646 0.678 0.678 0.693