Does Corporate Hedging Increase Firm Value?

An Empirical Analysis 1
John R. Grahama, Daniel A. Rogers∗,b
a

Fuqua School of Business, Duke University, Durham, NC 27708-0120, USA
b
Northeastern University, Boston, MA 02115, USA

January 20, 2000

Abstract
We study the derivative holdings of firms facing interest rate and/or currency risk. We net long and short
positions to measure the extent of hedging with net notional values. Our results are consistent with firms
hedging in response to expected financial distress costs, firm size, and investment opportunities, and also
to increase debt capacity and therefore firm value. Corporate hedging is not related to an explicit
estimate of the convexity in each firm's tax function. The tax gain associated with increased debt
capacity is much larger than the potential increase in value related to tax convexity.

JEL classification: G39; G32
Keywords: Corporate hedging; Derivatives; Capital structure; Taxes

1

We thank Tim Adam, George Allayannis, Hank Bessembinder, Gordon Bodnar, Nick Bollen, David Haushalter, Eric
Hughson, Ron Lease, Mike Lemmon, Uri Loewenstein, Steve Manaster, Mitchell Petersen, Jim Schallheim, Cathy
Schrand, Betty Simkins, Cliff Smith, Liz Tashjian, Wayne Thomas, Sheridan Titman, Bob Whaley, Jaime Zender, and
seminar participants at the 2000 AFA meeting, California State University-Northridge, the 1999 EFA meeting, the 1999
FMA meeting, Fordham University, Marquette University, Northeastern University, Oklahoma State University,
Southern Methodist University, University of Utah, Washington State University, and Wilfrid Laurier University for
helpful comments and suggestions. We also appreciate the research assistance of Yong Cai, Suzanne Perlee, and Ge
Zhang. This paper builds upon Rogers’ dissertation at the University of Utah.

Corresponding author. Tel.: 617/373-4707; fax: 617/373-8798; e-mail: drogers@cba.neu.edu.

If capital markets are perfect, hedging with derivative financial instruments (“derivatives”) does
not add to firm value, so firms should not hedge. Yet, an increasing number of firms use derivatives each
year. The International Swaps and Derivatives Association (ISDA) reports (on its Internet site,
http://www.isda.org )

that the notional value of outstanding OTC derivative contracts more than

quadrupled from 1994 to 1999, increasing from $11.3 to over $50 trillion. A large body of theoretical
research identifies costly market imperfections that, if they exist, explain why firms might hedge (see
Section I for a literature review). A number of empirical studies examine derivative holdings to identify
which of the theoretical incentives and imperfections lead to corporate hedging. Our paper adds to the
empirical literature by determining the factors that are related to corporate hedging for a unique sample
of firms facing interest rate and currency risk.
We use net notional derivative holdings to measure hedging for a broad cross-section of U.S.
firms with ex ante risk exposure. Our data are different from, and in many cases better than, those used
in previous studies of corporate hedging. Effective December 15, 1994, SFAS 119 requires firms to
report detailed information on the direction and purpose of notional derivative holdings, which makes it
possible to net long and short positions. (Prior to SFAS 119, notional values in financial statements
were aggregated across derivative type, and therefore it was not possible to reliably net derivative
positions.) Among research that uses financial statements to identify the determinants of hedging, ours is
the only study to determine net positions based on SFAS 119.2 Our “net” derivatives variable measures
the extent of hedging activity more precisely than variables used in past research.
Our dependent variable allows us to study the extent of derivative holdings, which differentiates
our paper from many corporate hedging papers. As summarized in Table I, most of the early empirical
hedging papers either use survey data (e.g., Nance, Smith, and Smithson, 1993; and Dolde, 1995) or
study the binary decision of whether or not to hedge (e.g., Mian, 1996; and Géczy, Minton, and
Schrand, 1997). Some recent papers use aggregated notional values of derivative holdings to study the
extent of hedging (e.g., Allayannis and Ofek, 1998; Berkman and Bradbury, 1996; Gay and Nam,
1999; and Howton and Perfect, 1999). However, these aggregated measures are not modified for
offsetting long and short payoffs, and so they are less precise than our dependent variable.

2

Wong (1997) uses SFAS 119 data to examine whether net notional value reliably measures currency exposure.

1

We restrict our sample to firms that face ex ante currency or interest rate risk. This is important
because we can interpret a lack of derivative holdings as a decision by a firm to hedge none of its risks,
which is distinct from not holding derivatives because of a lack of exposure to hedgeable risks. Other
studies that restrict their samples to firms with ex ante exposure include Géczy, Minton, and Schrand
(1997), Tufano (1996) and Haushalter (2000). These papers examine either the binary hedging decision
or hedging within a single industry, while we study the extent of hedging for a broad cross-section of
firms. Studying a broad cross-section also sets our paper apart from most other empirical hedging
papers, many of which examine large, Fortune 500-type firms (see Table I).
In sum, we use a relatively precise dependent variable to determine which market imperfections
lead to corporate hedging among firms with ex ante interest rate or currency risk exposure, and for a
broader cross-section than studied in most hedging research. We believe that using this sample and
hedging data contribute to the body of knowledge, even if we were to only examine the same
hypotheses studied in other papers.
In addition to the advantages offered by the data, our paper makes three contributions to the
literature investigating the causes of corporate hedging. First, we study whether firms hedge in response
to tax function convexity. Smith and Stulz (1985) show that if the function that maps income into tax
liability is convex, then by Jensen's Inequality, firms can reduce expected tax liabilities by hedging to
reduce income volatility. As summarized in Table I, virtually every empirical study that investigates the
use of derivatives includes a variable to proxy for tax function convexity. As we argue below, these
variables are, at best, imprecise measures of tax function convexity. We use a variant of the Graham and
Smith (1999) approach to explicitly calculate tax function convexity, and find no relation between
derivative holdings and tax function convexity. Although not finding evidence that firms hedge in
response to tax function convexity is a "non-result", we think it is important because it implies that other
research either needs to be reinterpreted or else use a more precise measure of convexity.
The second contribution is our investigation of whether firms hedge in response to an alternative
tax-related incentive to hedge, namely, hedging to increase debt capacity. Ross (1997) and Leland
(1998) model the primary benefit of debt financing as the tax deductibility of interest and show that by
hedging firms can increase debt capacity and therefore firm value. Stulz (1996) argues that hedging is
used to reduce the probability of "left-tail" outcomes, thereby increasing debt capacity and interest

2

write-offs. We investigate the hedging and debt policy choices jointly using simultaneous equations
regressions. We find that leverage exerts a positive influence on the use of both interest rate (IR) and
currency (FX) derivatives. Importantly, we also find that the debt-hedging relation runs the other way:
hedging leads to greater debt capacity. For the average firm, hedging with interest rate (currency)
derivatives increases the debt ratio by 2.85% (4.52%), with the capitalized value of the incremental tax
shields resulting from this increased debt equaling 1.4% (1.4%) of firm value. (Our estimates indicate
that tax incentives to increase debt capacity are approximately 10 times larger than convexity
incentives.) To the best of our knowledge, this is the first evidence that hedging increases debt capacity
and firm value, and moreover, the only explicit empirical estimate of value added due to derivatives
hedging associated with any of the hypothesized explanations of why firms hedge.
The third contribution of the paper involves contrasting, across the IR and FX samples, other
explanations of how costly volatility can lead to corporate hedging. We find evidence that firms use
derivatives in a manner consistent with value maximization. In addition to pursuing tax savings through
greater debt, we find that large firms subject to underinvestment problems and high expected distress
costs use IR derivatives to hedge. We also find strong relations between the extent of FX hedging and
both size and leverage. The FX results are weakly consistent with theories that suggest hedging reduces
underinvestment costs. In both samples, hedging increases with institutional ownership. Overall, in both
samples, we find that firms hedge in response to tax incentives to increase debt capacity,
underinvestment costs, financial distress costs, and firm size. Finding these results in both the IR and FX
samples indicates that these basic factors that drive corporate hedging can emanate from different
sources of risk, and also that firms hedge with more than one type of derivative security to reduce the
effects of costly volatility.
The remainder of the paper is organized as follows. Section I reviews risk management theories
and the related empirical evidence about why firms hedge, and defines the explanatory variables.
Section II describes how we construct the IR and FX samples and the hedging variables. Section III
performs univariate analysis of the differences between derivative users and non-users. Section IV
provides multivariate tests of the likelihood and extent of derivatives hedging, and also examines the
relation between derivatives hedging and tax incentives in greater detail. We conclude in Section V.

3

I. Value-maximizing rationales for hedging
An extension of the Modigliani and Miller (1958, 1961) propositions on capital structure and
dividend policy suggests that, in the absence of market imperfections, hedging does not add to firm
value. If capital markets are perfect, shareholders possess the requisite information about a firm’s
idiosyncratic risk exposures, and the necessary tools, to create their desired risk profiles; therefore, in
this environment, there is no reason for a firm to hedge. For corporate hedging to increase firm value,
market imperfections must exist. If a firm is exposed to economic risks in an imperfect environment,
these exposures, if unhedged, can impose costs on the corporation. Hedging helps reduce these costs.
For example, market imperfections can create an environment in which exposure to volatile interest
rates is costly.
Schrand (1998) points out that hedging research can be grouped into at least two broad
categories: 1) papers that investigate which market imperfections make volatility costly and therefore
lead to corporate hedging, and 2) papers that investigate why one method of reducing volatility is
cheaper than another. With respect to 1), financial economists have identified at least four market
imperfections that make volatility costly: the corporate income tax, costs of financial distress, managerial
risk aversion, and information asymmetry. (In the remainder of this section, we describe how these
imperfections provide an incentive to hedge.) Notably, the theories that model how these imperfections
impose costs on the corporation generally do not specify the source of the volatility, nor which type of
derivative instrument should be used to hedge. As examples of 2), Fenn, Post, and Sharpe (1996) and
Visvanathan (1998) study which factors and imperfections make it cheaper to use interest rate swaps in
conjunction with short-term debt instead of issuing long-term debt.
Our primary purpose is to determine which factors make cash flow or income volatility costly
and so our paper belongs in category 1. (See Table I for a review of empirical papers in this category.)
Given that the theories do not specify which type of derivative should be used to reduce volatility, we
test whether the factors outlined below affect the incentive to hedge using either interest rate or currency
derivatives. The remainder of this section describes imperfections and incentives that can affect
corporate hedging.

4

A. Tax function convexity and the incentive to hedge
Smith and Stulz (1985) demonstrate that volatility is costly for firms with convex effective tax
functions. Fig. 1 provides a simple illustration of their argument. A firm that realizes taxable income
levels of A or B, each with a probability of 1/2, has an expected tax liability of E[T]. If the firm
eliminates taxable income uncertainty (perhaps through hedging), it earns income level C with certainty,
and lowers its expected tax liability to E[T|H]. Reducing the volatility of taxable income reduces the
firm’s expected tax liability and increases market value. The argument also works in reverse: If a firm
has a concave tax function, hedging to reduce the volatility of taxable income increases expected tax
liabilities and decreases firm value.
If the tax function were linear over the entire income range, there would not be a convexitybased tax incentive to hedge. Ignoring progressivity, this would be the case if loss firms received an
immediate payment from the government equal to the product of the corporate tax rate and the dollar
amount of the loss. Instead of receiving such a payment, firms can carry losses backward and forward
to offset taxable income in other years. If firms could immediately sell tax preference items (such as net
operating loss carryforwards) at full price, or if the firms were certain they would eventually use their tax
preference items and the discount rate were zero, there would be no tax incentive to hedge. Therefore,
it is the limitations on the ability to sell or immediately use tax preference items that lead to convex tax
functions and (convexity-based) tax incentives to hedge.
Graham and Smith (1999) simulate important features of the corporate tax code to explicitly
measure tax function convexity for a large sample of U.S. firms. They find that roughly one-half of
corporations face a convex tax schedule. For these firms, the average tax savings achievable by
reducing income volatility by 5% are approximately $120,000. However, the savings are material for
only about 10% of Compustat firms. Graham and Smith also find that the tax function is effectively
concave for 25% of their sample, although tax incentives to increase volatility are typically small.
Most empirical derivatives papers measure tax function convexity with a variable based on
existing net operating loss (NOL) carryforwards (see Table I). Such variables imply that firms with
existing NOLs have convex tax functions, although the Smith and Stulz (1985) argument is about losses
that firms expect to experience in the future. In fact, Graham and Smith (1999) document that existing
NOLs provide a tax disincentive to hedge for firms with small expected losses (if a firm expects to lose

5

money, hedging reduces "right tail" outcomes and the chance that the firm will use its existing NOLs) but
provide an incentive to hedge for firms that expect to be profitable. Thus, variables based on existing
NOLs are too simple to capture incentives that result from the shape of the tax function, and may even
work backwards for expected loss firms. In addition, existing NOLs may measure financial distress or
some other firm characteristic, rather than a tax incentive to hedge.
Rather than using a NOL variable, we use the Graham and Smith approach to explicitly
measure tax function convexity. This technique quantifies the convexity-based benefits of hedging by
determining the tax savings that result from reducing volatility. We first calculate the expected tax liability
for a "full volatility" case, and then recalculate the expected tax liability with volatility reduced by 5%; the
difference between these two numbers represents the convexity-based tax benefit of hedging. We
choose a 5% volatility reduction to be consistent with the risk reduction observed by Guay (1999) when
firms introduce hedging programs. Our qualitative results do not change if we use a 1% or 3% volatility
reduction.
For the full-volatility case, we calculate the expected tax liability from t-3 (to account for
carrybacks) to t+18 (to account for carryforwards). To determine tax liabilities over this range, we
forecast earnings through period t+18. The forecasts assume that taxable earnings follow a random
walk with drift and that innovations are normally distributed, with mean and variance calculated from
firm-specific historical data through period t.3 The tax liability calculation incorporates carrybacks,
carryforwards, the Alternative Minimum Tax, and the progressive corporate tax schedule. We repeat
the experiment 49 additional times for each firm-year in the sample, starting each time with a new
forecast. We average across the 50 different scenarios to determine the expected tax liability. By
considering 50 different paths, we capture the interaction between earnings uncertainty and the
aforementioned features of the tax code.
Next, we recalculate the expected tax liability after modifying each of the 50 paths by reducing
volatility by 5% in year t+1. We measure tax function convexity as the expected tax savings resulting

3

Our reported analysis bases volatility on sales revenue. Although Graham and Smith (1999) show that convexity
does not differ substantially for different measures of earnings, previous readers of our paper suggested that taxable
earnings might be affected by a firm's past hedging program, while sales revenue should not be directly affected.
Therefore, we base our analysis on sales revenue volatility. Our qualitative results are unchanged if we base volatility
on taxable earnings.

6

from the 5% volatility reduction, scaled by sales revenue during year t. Positive values for tax convexity
indicate that hedging to decrease volatility reduces expected tax liabilities, implying a convex tax
function. Negative values imply tax function concavity. Our variable is a more precise measure of tax
function convexity than has been available to previous empirical derivatives studies.
There are countervailing influences that work against detecting corporate hedging in response to
tax function convexity. Graham and Smith (1999) show that the firms that are most likely to have
convex functions are small, have expected income near zero, and alternate between profits and losses.
These firms may find the fixed costs of setting up a hedging program prohibitive, and consequently not
hedge.4 In addition, the option value of equity may provide the shareholders of these firms with
incentives to increase volatility, opposite the incentive provided by the tax code.

B. The relation between hedging and debt ratios
B.1. Debt ratios measure financial distress
If financial distress is costly (whether from direct bankruptcy costs, such as legal fees, or indirect
costs, such as business disruption), and volatility increases the chance of encountering distress, then
volatility can reduce firm value. Smith and Stulz (1985) argue that hedging can increase firm value by
reducing volatility and the probability of distress. (As we discuss in more detail below, Froot,
Scharfstein, and Stein (1993) extend the Smith and Stulz argument by endogenizing distress costs and
show that hedging can reduce the underinvestment problem resulting from the deadweight costs
associated with external financing.) Stulz (1996) emphasizes the role of hedging in preventing left-tail
outcomes that would cause distress and force firms to bypass investment opportunities or otherwise fail
to exploit their competitive advantages.
Dolde (1995), Berkman and Bradbury (1996), Haushalter (2000), Gay and Nam (1999), and
Howton and Perfect (1999) use the debt ratio to measure expected costs of distress and find that higher
debt ratios lead to greater hedging (see Table I). Most studies interpret this relation as evidence that
greater expected financial distress costs cause greater hedging. This interpretation assumes that firms
with higher debt ratios face higher probabilities of encountering financial distress. We include the debt
4

Bodnar et al. (1996) find the “costs of establishing and maintaining a derivatives program exceed the expected
benefits” as the second most common explanation of why some firms do not use derivatives.

7

ratio in our analysis, defined as total debt divided by the book value of assets.
We also directly measure expected distress costs with a variable that accounts for both the
probability of distress and the associated cost if distress occurs. To calculate this variable, we multiply
the debt ratio by the equity market-to-book ratio. The probability of financial distress increases with the
debt-asset ratio, and the cost of distress (if encountered) increases with the market-book ratio. We
expect to observe a positive relation between this variable and derivatives use if firms hedge in response
to costs of financial distress. Géczy et al. (1997) use the product of the debt ratio and the market-book
ratio in their analysis, although they interpret the variable as a proxy for underinvestment costs. Géczy et
al. find a positive relation between the decision to use foreign currency derivatives and the product of
the debt and market-book ratios.
We examine two other variables related to costs of distress. Firm profitability might be inversely
related to hedging if less profitable firms have a higher probability of encountering distress. Conversely,
the option value of equity might encourage unprofitable firms to hedge less than their non-distressed
counterparts. We measure profitability as the pre-tax return on assets (ROA). We also denote firms that
might be experiencing current (or recent) financial distress with NOL carryforwards scaled by the book
value of assets.5

B.2. Hedging leads to increased debt capacity
When we examine explanatory variables based on the debt ratio, we follow other studies and
interpret the debt ratio as a measure of the expected costs of distress, which is a causal determinant of
hedging policy. However, Stulz (1996), Ross (1997), and Leland (1998) suggest an alternative reason
that debt ratios and hedging practices may be positively correlated: hedging, by reducing the volatility of
income and/or reducing the probability of financial distress, increases debt capacity. If firms add
leverage in response to greater debt capacity, the associated increase in interest deductions reduces tax
liabilities and increases firm value. Thus, the ability to increase debt capacity provides a tax incentive to
hedge.
5

We also examine the credit rating of senior debt as a measure of the probability of distress. Including credit rating in
our empirical specification reduces the sample size, and yet the variable is not statistically significant, so we do not
include credit rating in our reported specifications. We also do not report a specification that includes a dummy
variable indicating if a firm has a credit rating because the variable is not statistically significant.

8

Leland (1998) implies that hedging increases value through two different channels related to the
debt ratio. The principal gains from hedging come from "the fact that lower average volatility allows
higher leverage with consequently greater tax benefits." A secondary hedging gain comes from "lower
expected default rates" and distress costs, resulting from unused debt capacity. That is, the majority of
the gain comes from increased leverage/tax deductions but a portion of the increased debt capacity goes
unused, resulting in lower distress costs and higher value. Ross (1997) also argues that the reduction in
expected distress costs is less important than the tax shield available from increased leverage.
In sum, theory suggests that the hedging/leverage causality can go both ways: hedging can lead
to increased debt capacity, but higher leverage (to the extent that it increases the probability of distress)
can increase the incentive to hedge. Therefore, in part of our analysis, we model the hedging/capital
structure decision as a simultaneous system, which is appropriate if these two corporate policies are
jointly determined. (Haushalter (2000) and others argue that this is the case.) In the regressions with
hedging as dependent variable and debt on the right hand side, we interpret a positive coefficient as
evidence that leverage leads to increased hedging due to higher expected costs of distress. In the
regressions with debt as dependent variable and derivatives use on the right hand side, we interpret a
positive coefficient as evidence that hedging leads to increased debt capacity and tax deductions.

C. Underinvestment costs and the incentive to hedge
Myers (1977) and Myers and Majluf (1984) describe circumstances in which a firm might
reject positive net present value (NPV) projects (i.e., underinvest). Bessembinder (1991) and Froot et
al. (1993) propose hedging as a solution to the underinvestment problem. 6
Bessembinder (1991) assumes that a firm simultaneously determines a hedging policy and debt
level before selecting its optimal level of investment. He shows that in the absence of hedging, the firm
may underinvest because too much of the incremental value from investment accrues to debtholders.
However, if the firm can credibly commit to a hedging policy at the time of the debt decision, the
underinvestment problem is attenuated because the value of debt becomes less sensitive to incremental
investment. Shareholders are beneficiaries of the improved contracting terms with debtholders (or other

6

Mayers and Smith (1987) model the purchase of insurance as a means of reducing underinvestment.

9

fixed claimants).
In Froot et al. (1993) firms can use internal and/or external funds to finance investment projects.
The authors assume that financing costs increase with the level of external financing, and that the
marginal benefit of investment declines with the level of investment. Volatility is costly in this environment
because if internal funds are relatively scarce in some states of nature, positive NPV projects may be
rejected (if the marginal cost of external funds exceeds the marginal benefit to shareholders in these
states). Hedging effectively allows a firm to shift internal funds into states where they would otherwise be
scarce. If internal funds are cheaper than external funds, hedging therefore permits the company to
finance more valuable investment projects and increase firm value.7
Theory implies that underinvestment is most severe for firms with valuable investment
opportunities, which we quantify with the market-book ratio. Several other empirical studies examine
the relation between market-book and hedging and find no relation (see Table I). As mentioned in
Section I.B.1, we include the product of the market-book ratio and the debt ratio in our specification,
which Géczy et al. (1997) define as a measure of underinvestment costs. We also use R&D spending
scaled by the book value of assets to measure growth options. Nance et al. (1993), Dolde (1995),
Géczy et al. (1997), Allayannis and Ofek (1998), and Gay and Nam (1999) all find that hedging
increases with R&D spending. Finally, we include cash capital expenditures on property, plant, and
equipment, scaled by the book value of assets to control for the relation between hedging and current
investment spending.

D. Managerial risk aversion and the incentive to hedge
Risk aversion among managers and/or shareholders can provide an incentive to hedge (Stulz,
1984; and Smith and Stulz, 1985). If managers have concave utility functions, and the variability of their
compensation is related to the volatility of corporate income or cash flows, then corporate volatility can
be costly. If managers can not effectively hedge corporate volatility in their personal accounts, or if it is

7

Tufano (1998) argues that managers might hedge to avoid scrutiny of negative NPV “pet projects” by external
capital markets. Providers of external capital would not fund these pet projects. However if managers use hedging to
ensure the availability of internal capital, then the projects might be funded. Tufano thus suggests that hedging can
lead to overinvestment. However, Tufano notes that if "pet project" agency costs are relatively low, then
underinvestment concerns will dominate.

10

cheaper for the firm to hedge than it is for managers, then corporate hedging can improve managerial
welfare. Therefore, it can be value maximizing for a firm to hedge for the benefit of managers if doing so
reduces the risk premium managers demand, and therefore reduces their required compensation. A
similar argument can justify corporate hedging of the risk faced by ill-diversified corporate insiders with
large shareholdings.
We use three variables to test whether manager/shareholder risk aversion affects corporate
hedging: 1) the number of executive stock options held by the CEO, 2) the natural logarithm of CEO
stock value, and 3) the presence of multiple classes of common stock. Stock options introduce a
convexity between managerial wealth and stock value, which offsets the concavity in the manager's
utility function. This in turn makes the manager behave in a less risk averse manner, which reduces the
need for corporate hedging. Therefore, hedging should be negatively related to CEO option holdings.
In contrast, the value of common stock increases linearly with firm value. Thus, a manager’s
holdings of common stock do not counteract the concavity of his/her utility function. A manager with
large stockholdings has incentive to engage in risk averse behavior. Therefore, we predict that hedging
activity is positively related to the value of CEO stockholdings. Finally, firms with multiple classes of
shares often have a controlling group with superior voting rights. If management or ill-diversified
shareholders are represented in the controlling group, managerial or shareholder risk aversion is more
likely to affect corporate actions. We include a dummy variable that equals one if a firm has multiple
classes of shares and hypothesize a positive relation between this variable and corporate hedging.
Tufano (1996) and Schrand and Unal (1998) find evidence that hedging increases with
managerial shareholdings and decreases with managerial option ownership, consistent with the
hypotheses outlined above. Other studies (e.g., Géczy et al., 1997; and Haushalter, 2000) find no
evidence that managerial risk aversion or shareholdings affect corporate hedging. We use the variables
defined above to maintain consistency with these prior studies. However, such simple variables may not
measure precisely the incentive effects of stock and option holdings. Therefore, we also calculate the
more elaborate “delta” and “vega” variables defined in Core and Guay (1999). These variables measure
the change in a manager’s wealth as stock price and volatility change, respectively, given his or her
stock and option portfolios. We find (in unreported analysis) that the Core and Guay variables are not
significantly related to corporate hedging in our samples, so, to permit comparison of our analysis to

11

previous work, the results we report below are based on the three simpler variables.

E. Other incentives to hedge
DeMarzo and Duffie (1991) and Breeden and Viswanathan (1998) argue that there are
informational asymmetries between managers and shareholders. DeMarzo and Duffie conclude that
corporations should sometimes hedge based on private information that can not be costlessly conveyed
to shareholders. Breeden and Viswanathan argue that high-quality managers have an incentive to hedge
away uncertainty about managerial performance so that the market can draw more precise inference
about their ability. To measure the degree of informational asymmetry, we use the percentage of each
firm’s shares that are owned by institutions as of September 1994, based on 13-f filings reported by
CDA Spectrum. If firms owned primarily by institutions face less informational asymmetry of the type
assumed by DeMarzo and Duffie (1991) and Breeden and Viswanathan (1998), their theories imply
that high-institution-ownership firms should hedge less. Géczy et al. (1997) find the opposite, namely
that firms with high institutional ownership are more likely to hedge with currency derivatives.
Most prior studies find that the likelihood of using derivatives increases with firm size. A positive
size effect is consistent with firms not hedging with derivatives unless the benefits are larger than the fixed
costs of establishing a hedging program. In contrast, Haushalter (2000) finds a negative relation between
size and hedging, given that the firm hedges. Conditional on hedging, a negative relation is consistent
with the extent of hedging increasing with informational asymmetry and financial distress costs. We
measure firm size as the natural logarithm of total assets.
Nance et al. (1993) note that "hedging substitutes" can reduce the need for hedging with
derivatives. For example, dividend restrictions might allow a firm to retain sufficient liquidity to make
hedging unnecessary. Likewise, holding liquid assets may reduce the need to hedge. Dividend yields
(Nance et al., 1993) and liquidity ratios (Tufano, 1996; and Géczy et al., 1997) have both been shown
to be negatively related to derivatives usage. We also examine the effect of dividend yield and liquidity
on derivatives hedging.
Géczy et al. (1997) find that the likelihood of using currency derivatives is positively affected by
the use of other types of derivatives, suggesting that there are scale economies to running a derivatives
program. Therefore, we include a dummy variable indicating interest rate (currency) derivatives usage in

12

the tests of currency (interest rate) hedging. In unreported multivariate analysis, we find that the
existence of interest rate derivatives positively affects currency hedging but not vice versa. Therefore,
our reported regressions include an interest rate dummy variable in tests of currency hedging but not a
currency derivatives dummy in tests of interest rate hedging.
As mentioned at the beginning of Section I, some papers study the factors that make using
interest rate swaps in conjunction with short term debt preferable to using long term debt. While our
focus is on what makes volatility costly, we control for this other explanation for the use of swaps. We
include the sum of short-term and floating-rate debt (as a percentage of total debt) in our regressions to
control for the effect of such debt on interest rate derivatives use. (Analogously, for the foreign currency
sample, we control for the importance of foreign sales relative to total sales.) To control for another
explanation of why firms use swaps and short term debt, we examine the level of credit ratings to see if
low-quality firms use IR derivatives to avoid the "credit quality spread" associated with issuing long-term
debt (Bicksler and Chen, 1993). Finally, we also examine changes in future credit ratings, to see if firms
with favorable private information use short-term debt and IR derivatives until their credit rating
improves (Titman, 1992).8 Neither of these last two hypotheses is supported by our interest rate
derivatives tests, and yet including the credit rating reduces the sample size, so we do not include credit
ratings in our reported results.

II. Derivatives data and hedging variables
A. Sample formation
We analyze corporate derivative holdings as of calendar-year-end 1994 or fiscal 1995. This
time frame is chosen because, prior to December 1994, financial statement disclosures were generally

8

Simkins (1997) tests Titman’s (1992) hypothesis that firms use interest rate swaps because of asymmetric
information about their credit quality. Titman's theory implies that firms with credit ratings just below investment
grade, that know their rating will be upgraded soon, borrow short-term with floating rates. They then enter into
interest rate swaps that effectively fix the interest rates on their debt obligations until the good news about their
credit quality is revealed. Our interest rate sample is not designed to test Titman's hypothesis; therefore, it is not
surprising that we do not find a relation between credit rating changes and IR derivatives hedging. Simkins' (1997)
sample is designed to test this hypothesis, and she finds evidence consistent with Titman's (1992) predictions.

13

inadequate to analyze the extent of derivatives hedging.9 In 1994, the Financial Accounting Standards
Board (FASB) issued SFAS 119 in an attempt to improve derivative disclosures by U.S. corporations.
SFAS 119 mandates that firms disclose information on notional values of derivative contracts,
including the direction of the position, across several categories.10 SFAS 119 became effective for fiscal
years ending after December 15, 1994 for firms with assets greater than $150 million. 11 Firms are
required to specify if derivatives are held for trading purposes or, if for non-trading reasons, the purpose
of the holdings. We examine only derivatives held for non-trading purposes.
We obtain the information about each firm’s fiscal year-end derivatives ownership from 10-K
forms filed electronically in the SEC’s Electronic Data Gathering and Retrieval (EDGAR) database
during March-December 1995. (An example of one firm’s derivatives disclosure is shown in Appendix
A.) This population consists of 3,232 firm-filings. We search the footnotes and Management’s
Discussion and Analysis (MD&A) of each 10-K filing for word strings such as “hedg”, “swap”, and
“derivative”. If a reference is made to any of the search terms, we read the surrounding text to confirm
that it refers to derivative holdings. Given the time-consuming nature of data collection from these filings,
we examine a random sample of 855 firms selected from the initial list of filers. Out of these 855 firms,
we retain observations that meet the following criteria:
(1) fiscal year ends December 15, 1994 through October 31, 1995,
(2) complete set of annual financial statements including footnotes is available,
(3) the firm is not a subsidiary of another firm in the sample,
(4) the firm's income is taxable at the corporate level,
(5) the firm is listed on the Compustat annual database, and
(6) the firm discloses the notional value of its derivative holdings, if any.
Out of the 531 observations that meet these six criteria, we divide the data into two non-mutually

9

Géczy et al. (1997) note this limitation in their analysis of corporate foreign currency derivatives during 1991. They
consider notional values of derivatives from financial statements to be unreliable due to aggregation across different
types of derivatives (as was commonly done prior to December, 1994).
10

Smithson et al. (1995, p. 147) define notional value (or notional principal) as “principal … that is not paid or received
at contract maturity but is instead used only to calculate the cash flows paid and received.” The notional value of a
derivatives contract is typically similar, if not equal, to that of the underlying asset or liability being hedged.
11

SFAS 105 and 107 require firms to declare if they use derivatives, even if their assets are less than $150 million.

14

exclusive groups (as described in the next section): firms that face ex ante interest rate risk (404 firms),
and firms that face ex ante currency risk (242 firms). Seventy-four of the firms in the FX sample are not
included in the IR sample, and 168 firms are in both samples.
Thirty-five percent of the firms in the IR sample use interest rate derivatives. Over 43% of FX
sample firms report that they use currency derivatives (see Table II). Both samples contain observations
from all major industrial classifications. Manufacturing firms (SIC codes 22-39) are common in both
samples, 40% and 62% of the IR and FX firms, respectively. Not surprisingly, banking and investment
firms are the most frequent users of IR derivatives. Transportation, construction, and mining firms also
use IR derivatives more often than average. Other than the large number of manufacturing firms, the FX
sample is represented by a relatively small numbers of firms in other industries.

B. Definition of ex ante risk exposure
By focusing on firms that face ex ante risk, we can interpret the absence of derivatives as a
choice not to use derivatives, rather than possibly indicating a lack of exposure to hedgeable risks. So
that we can study the use of both interest rate and currency derivatives, we define two separate
samples: 1) firms with ex ante foreign exchange risk exposure; and 2) firms with ex ante interest rate risk
exposure. Similar to previous research, we define firms to have ex ante currency exposure if they
disclose foreign assets, sales, or income in the Compustat Geographic segment file, or disclose positive
values of foreign currency adjustment, exchange rate effect, foreign income taxes, or deferred foreign
taxes in the annual Compustat files.
We define ex ante interest rate exposure in two separate ways. We start by noting that interest
rate derivatives are often used to hedge the interest cash flows from debt liabilities. Therefore, financial
leverage is a possible source of interest rate risk. In the first definition of interest rate risk, we classify
firms as facing interest rate risk if their debt-assets ratio is greater than 10%.12 While this cutoff point is
arbitrary, only ten of the 127 firms with debt-assets ratios less than 10% use IR derivatives. Our results
are qualitatively similar if we define IR exposure using a cut-off of 0%, 5%, or 15%.
The alternative definition of ex ante interest rate risk is based in part on the sensitivity of
12

Fenn et al. (1996) find that debt holdings significantly affect the use of IR derivatives, which we interpret as
evidence that debt holdings are a reasonable measure of IR exposure.

15

operating income to interest rates. Specifically, we regress changes in operating income on changes in
the six-month LIBOR rate. (Note that operating income does not usually contain income resulting from
interest rate derivatives, and therefore should not be directly affected by the endogenous decision to
hedge with interest rate derivatives.) Based on the sign of the regression coefficient, we classify firms as
having positive, negative, or zero operating exposure to interest rates. Zero exposure occurs when the
regression coefficient is not significant at the 10% level. For this definition, a firm faces ex ante interest
rate risk if it meets any of the following criteria:
1) zero operating exposure to interest rate changes, and positive amounts of floating debt (i.e.,
short-term and/or floating-rate debt),
2) negative operating exposure to interest rate changes, or
3) positive operating exposure to interest rate changes, and less than 50% of debt is floating.
(If there is sufficient floating debt, it can offset the positive operating exposure to interest
rates, so a firm would not face ex ante interest rate risk.)
Our qualitative results are similar regardless of how we define ex ante interest rate risk exposure. The
results we present below are based on the first definition of interest rate risk.

C. Measuring derivatives hedging
We gather notional values for each of three major derivatives categories: interest rate, currency,
and commodity.13 We classify derivative positions as “long” or “short”. A long interest rate position is
one that benefits from rising interest rates, such as an interest rate swap that pays a fixed rate and
receives a floating rate. Conversely, a short interest rate position benefits from declining interest rates. A
long currency derivative position benefits from price increases of a currency other than the U.S. dollar,
while a short position benefits from decreasing foreign currency prices. If a position is not clearly “long”
or “short”, we classify it as “unsure”.
We measure derivative holdings several ways. One measure is the total notional value of
derivative contracts held by each firm for non-trading purposes. Total notional values have recently been
13

We include the commodity holdings for informational purposes but they should be interpreted cautiously. Many
commodity instruments are not considered derivative financial instruments under SFAS 119 because of the
possibility of physical delivery (such as with commodity futures contracts). Thus, the use of commodity derivatives
is likely understated in our data. We do not investigate commodities in detail.

16

used in Berkman and Bradbury (1996), Allayannis and Ofek (1998), and Gay and Nam (1999). We
calculate total notional values for IR holdings and, separately, total notional values of FX holdings. We
present summary information about total notional values (in Tables III and V) but do not use this
variable in our regression analysis (except in one robustness check).
While total notional value effectively gauges derivatives ownership, it may not accurately
estimate derivatives hedging if a firm holds offsetting contracts. The distinction between ownership and
hedging is important when testing theories of risk management. Therefore, for our second derivatives
variable, we calculate the absolute value of the “net” derivatives position in each category. The net
position is the difference between each firm’s long and short positions in interest rate and, separately,
individual currency derivatives (and thirdly, commodities).14 For interest rate hedging, we add basis
swaps (interest rate swaps that are essentially an exchange of floating rate indices) to the absolute
difference between long and short positions. The net derivatives variable is the dependent variable in our
primary regression analyses.15
Finally, to contrast our analysis with studies that examine the yes/no hedging decision, we
construct a binary variable. Firms that hold derivatives for non-trading purposes are assigned a value of
one for the binary variable, and all other firms are assigned a zero.
Three important issues affect how precisely our variables measure corporate hedging. First,
derivative holdings may measure speculative activity, not hedging. Taken literally, SFAS 119 requires
firms to explicitly state if they speculate with derivatives. Many firms provide statements such as
“derivatives are used for risk management purposes only”; none of our sample firms state that they
speculate. Further, our variables are defined using only those derivatives disclosed as being held for
non-trading purposes. Consequently, we classify firms that use derivatives for non-trading purposes as

14

The total notional value used in previous derivatives studies might classify $100 million long and $50 million short
as $150 million. In contrast, we net the two figures to determine a net position of $50 million. Wong (1997) finds that
net notional values are related to foreign currency exposure in 1995.
15

For example, consider a firm with notional values of $100 million long interest rate, $50 million short interest rate, $20
million long British pounds, and $75 million short Japanese yen. This firm’s net position in IR (FX) derivatives is $50
($95) million. We use absolute values because hedging to maximize firm value may require a firm to go long in one
derivatives category but short in another.

17

"hedgers" and those that do not use derivatives as "non-users."16 Second, firms can hedge with
operational strategies, such as building a manufacturing facility in a locale that is the source of foreign
currency risk, or issuing convertible debt. Our study focuses on hedging with derivatives, rather than on
operational hedging strategies.
Finally, as noted by Smith (1995), different firms can hold the same notional value of derivatives
and still have very different hedging practices. For example, two firms may hold $10 million in
derivatives, but one has one-year swaps and the other seven-year swaps. The firm with the one-year
swap might appear to hedge less than its counterpart with the seven-year swap. However, if the first
firm is hedging a liability that matures in one year while the other firm is hedging a liability that matures in
seven years, then both firms are hedging appropriately. Current financial reporting guidelines do not
require firms to disclose the underlying asset and/or liability that is being hedged with a derivative
contract, so we are unable to determine whether firms employ a maturity-matching hedging policy. To
the extent that notional values do not fully reflect corporate hedging practices, our data are noisy, which
works against our ability to document hedging incentives.
Table III presents aggregate measures of total and net derivative positions for the IR and FX
samples (Panels A and B, respectively). The mean and median notional values of all derivative positions
are $2.8 billion and $210 million in the IR sample, and $2.75 billion and $171 million in the FX sample.
The net derivative positions are significantly smaller: both samples have mean and median values of
approximately $700 million and $125 million, respectively.
Table III shows that 180 out of 404 firms in the IR sample disclose some type of derivative
holdings, and 142 hold IR derivatives. We can calculate net IR derivative positions for 131 firms. For
the 142 IR derivative users in this sample, the mean (median) notional values of IR derivatives is $3.0
($0.213) billion, or 16.9% (8.2%) of total assets. The mean (median) net IR position is $728 ($110)
million, or 9.5% (5.5%) of total assets.
One hundred thirty-eight out of the 242 firms in the FX sample use some type of derivative. Of
16

Several studies lead us to believe that most corporate derivatives are held as hedging instruments. Guay (1999)
finds that initiation of corporate derivatives use is associated with declines in various measurements of firm risk.
Tufano (1996) uses extremely accurate data to describe the derivative holdings of U.S. gold mining firms and finds no
evidence of speculation. Hentschel and Kothari (1998) find no evidence that derivatives use increases firm risk.
However, Bodnar et al. (1996) find that corporate management teams sometimes allow their "market view" to influence
hedging decisions, implying that there may be an element of speculation in samples of corporate derivative holdings.

18

these 138 firms, 105 use currency derivatives, with mean (median) holdings of $684 ($80) million. From
this set of currency derivative holders, we can calculate net currency derivatives for 64 firms.17 The
mean (median) net currency position among these firms is $195 ($67) million.

III. Univariate analysis of hedgers vs. non-users
Panel A of Table IV summarizes the characteristics of the firms in the IR and FX samples.
Comparisons between derivative hedgers and non-users are presented in Panel B.
Hedgers are much larger than non-users, which is consistent with the fixed costs of derivatives
hedging acting as a barrier to small firms (see Panel B). Interest rate hedgers are more likely to use
currency derivatives, and vice versa, which is consistent with there being scale economies to corporate
hedging. Institutional investors are more important investors for hedgers than non-users, contrary to
information asymmetry rationales. CEOs of derivative-using firms hold more options and wealth in their
firms’ stocks than do CEOs of non-users. Hedgers have significantly lower book-market ratios (based
on the Wilcoxon rank-sum test). This suggests that hedgers have greater investment opportunities, which
is consistent with the Froot et al. (1993) underinvestment argument and the analysis of Allayannis and
Weston (1998). Dividend yields are higher for derivative hedgers than for non-users, consistent with the
Nance et al. (1993) argument that dividend curtailment is a substitute for hedging. Currency hedging is
negatively related to liquidity, which is consistent with liquidity serving as a hedging substitute. In the FX
sample, the likelihood of hedging increases with the percent of sales from foreign sources.
We find weak evidence of financial distress motives to hedge, as suggested by results for the
interaction of debt ratio with the market-book ratio. In both samples, derivative hedgers have higher
values for this variable, with the differences significant using means in the IR sample and a Wilcoxon
rank-sum test in the FX sample. The results for the tax convexity variable and debt ratio are not
consistent with tax-related hedging rationales. The differences in the R&D ratio indicate that firms with
greater growth options are more likely to hedge with FX derivatives.
Table V summarizes the correlations between the independent variables and the three measures

17

A relatively large number of currency positions are "unsure" because the disclosures for currency derivatives
often lack the detail of the IR disclosures. We exclude from regression analysis all firms that have positive total
derivatives but are missing a net derivatives position.

19

of scaled IR and FX derivative holdings. Firm size and institutional ownership are positively related to all
three derivatives variables in both the IR and FX categories. CEO stock value and option holdings are
also positively related to the hedging variables. The financial distress variable (interaction of debt and
market-book) is positively related to the IR variables and net currency derivatives. The debt ratio is
positively correlated with the two continuous measures of IR derivative holdings, as well as with net
currency derivatives. FX derivative use increases with foreign sales. The correlation coefficients are not
consistent with the hypothesis that firms hedge in response to tax function convexity. In unreported
analysis, we find that total and net derivatives positions have a correlation of 0.49 (0.69) in the IR (FX)
sample.

IV. Multivariate analysis of corporate hedging
In this section, we use multivariate analyses to test the joint influence of the explanatory variables
on corporate derivatives hedging. We start by examining the extent of hedging using a Tobit analysis.
These results are then contrasted with those from a combination of probit analysis and truncated
regression as suggested by Cragg (1971). The section concludes with an in-depth analysis of tax
motives to hedge.

A. The extent and likelihood of derivatives hedging
A.1. IR derivatives hedging
Tobit regression: We use a Tobit specification to analyze the factors that affect the extent of
derivatives ownership. This econometric model is appropriate when the dependent variable is leftcensored at zero. The Tobit model takes the form
Yi* = β'Xi + ui where ui ~ NID(0, σ2)
Yi = Yi* if Yi* > 0
Yi = 0 if Yi* ≤ 0

(1)

The dependent variable, Yi*, is unobservable if its true value is negative.18 The vector of independent

18

Although the dependent variable never takes a value below zero, "negative hedging" can occur if firms use
derivatives to speculate. If this occurs, the parameter estimates may be biased towards zero, in which case the Tobit
specification lacks power to reject the null hypothesis of no relation between the dependent and explanatory

20

variables, Xi, includes the explanatory variables discussed previously and industry indicator variables for
the classifications described in Table II. We conduct unreported Tobit regressions on the full model
specification. However, for brevity, the reported results are based on specifications that exclude the
insignificant industry dummy variables, as well some other insignificant explanatory variables. Our major
conclusions are not affected by excluding these variables.
Model IR 1 of Table VI, Panel A presents slope estimates from a Tobit regression using net IR
derivatives (scaled by the book value of assets) as dependent variable.19 The debt ratio and the
interaction of debt with market-book are both positively related to IR hedging. These findings are
consistent with firms hedging in response to large expected costs of distress. Firms with low ROA
hedge more, which is also consistent with the financial distress motive to hedge. In contrast, the negative
coefficient on the NOL carryforward variable indicates that firms reduce hedging if they have recently
accumulated losses. Although inconsistent with hedging in response to past distress, this relation may
indicate that the option value of equity discourages severely distressed firms from hedging. Our results
are more supportive of a financial distress motive to hedge than those found in most empirical studies,
perhaps because we examine a broad sample of firms. Haushalter (2000), Gay and Nam (1999), and
Howton and Perfect (1999) find that leverage ratios, but not other measures of distress, are positively
related to hedging activity. Tufano (1996), Géczy et al. (1997), and Allayannis and Ofek (1998) find
weak or no evidence that distress costs affect hedging.
We find a positive relation between R&D expenses and hedging. This is consistent with firms
hedging to minimize underinvestment problems when they have growth options (Bessembinder, 1991;
Froot et al., 1993). However, our analysis indicates that there is no relation between the market/book
ratio and hedging. Our results are consistent with other studies that find evidence consistent with the
underinvestment hypothesis based on R&D expenses, but no evidence based on market/book (see

variables. Given the requirements of SFAS 119, we believe that misidentification of derivatives positions in this
manner is at most a minor problem (see discussion in Section II.C).
19

The parameter estimates from a Tobit regression represent the marginal effect of each regressor on the unobserved
dependent variable, Yi*. We are more interested in the marginal effect on the observed dependent variable, Yi. The
marginal effect of a change in the kth regressor is calculated by F(z)βk, where F(z) is the normal CDF evaluated at
z=β'Xi/σ and Xi is the vector of independent variables. The marginal effects are calculated at the means of the
independent variables (as described in Maddala, 1983).

21

Table I). Like Géczy et al. (1997), we find a positive relation between hedging and the product of the
debt and market-book ratios.
The parameter estimate for the tax function convexity variable is negative and insignificant. This
result does not support the tax convexity motive for hedging. In Section IV.B, we investigate tax
incentives to hedge in more detail.
CEO stockholdings are positively related to IR hedging in our sample. This result is consistent
with the linear payoff from equity contributing to risk-averse managerial behavior (Smith and Stulz,
1985). The reported IR results thus confirm in the broad cross-section what Tufano (1996) finds for
gold-mining firms and Schrand and Unal (1998) find among savings and loan companies. However, we
find no relation between stock holdings and hedging when we scale the holdings, or when we use the
Core and Guay (1999) compensation variables (not tabulated). Nor do we find a relation between IR
hedging and the number of options held by the CEO. Overall, the relation between hedging and
managerial risk aversion is at best weak.
Hedging increases with firm size. This result is consistent with fixed costs limiting hedging by
small firms, but not consistent with informational asymmetry leading to increased hedging. The positive
relation between hedging and institutional ownership is also not consistent with the informational
asymmetry hypothesis. This latter finding confirms in the broad cross-section what Géczy et al. (1997)
find for Fortune 500 firms. We also find that IR derivatives hedging increases with the importance of
floating rate debt. Finally, firms with dual classes of stock hedge more, although this result is not
statistically significant.
Robustness of Tobit analysis: Many extant hedging studies exclude financial firms (SIC codes
60-69) and utilities (SIC code 49) because financial firms use derivatives for both risk management and
trading purposes and utilities are regulated. In our paper, we only analyze derivatives that are held for
non-trading purposes to avoid problems associated with financial firm’s derivative dealer operations
(most banks explicitly disclose derivatives held for asset-liability risk management separately from other
derivative positions). Nonetheless, to benchmark our analysis against the extant literature, the Tobit
analysis is repeated excluding financial and utility firms. Excluding these firms does not materially alter
the reported results.
We repeat the Tobit regressions using total notional value as dependent variable, rather than net

22

notional value, to see if netting long and short positions improves our specification. The log-likelihood is
significantly lower for this unreported regression. Also, the ROA and debt times market-book variables
are not significant when total notional value is used as dependent variable. These results indicate that
corporate hedging activity is more precisely measured by net positions than by total positions.
There could be a positive relation between the use of IR derivatives and the debt variables
simply because high-debt firms have more liabilities to hedge and not because of financial distress. To
address this concern, we repeat the Tobit analysis, scaling the dependent variable by debt, rather than
by assets (see Model IR 1B). This specification thus measures hedging per dollar of debt. (Note that
this might "over correct" the problem and induce a negative coefficient on the independent variables that
involve the debt ratio because debt is in the denominator of the dependent variable). The results still
show a positive, significant relation between leverage and IR hedging, corroborating the Model IR 1
results.20
Many cross-sectional studies (Nance et al., Mian, and Géczy et al.) examine the yes/no hedging
decision. Given the large variation in extent of derivatives ownership (see Table II), a strength of our
analysis is that we use a continuous dependent variable, enabling us to incorporate more information into
the dependent variable. However, Tobit analysis implicitly assumes that firms make one hedging
decision, the choice of how much hedging to perform. Strictly speaking, in the Tobit analysis, a zero for
the dependent variable reflects the outcome of the extent of hedging decision, rather than a decision not
to hedge.
Alternatively, the hedging decision might involve two steps. First, the firm decides whether or
not to hedge, and then, how much to hedge (if it hedges). The econometric approach suggested by
Cragg (1971) is used to model this two-step process (see Haushalter, 2000; and Allayannis and Ofek,
1998). The first step examines whether or not a firm hedges (using a probit model) and the second step
examines the extent of hedging, given that a firm hedges (using a truncated regression). In the Cragg
procedure, zero derivative holdings indicate a decision not to hedge, rather than the extent of hedging,
and so the truncated regression only uses observations with non-zero derivative holdings. Our goal here

20

There could be a spurious negative relation between debt and derivatives use if foreign debt substitutes for
currency hedging, as argued by Allayannis and Ofek (1998) and Géczy et al. (1997). To the extent that this
substitution occurs, it reduces the power of our tests to detect a positive relation between hedging and debt.

23

is not to determine how many steps are involved in the corporate decision to hedge. Rather, we present
analysis based on the Cragg two-step approach and contrast these results with the Tobit analysis.
Probit regression: Models IR 2 and 3 of Table VI present results of a binomial probit and
truncated regression, respectively. The signs and significance for the probit coefficients are generally
similar to those for the Tobit analysis. The probit results indicate that large firms with greater investment
opportunities (as measured by R&D) are more likely to hedge with IR derivatives. The probability of
using IR derivatives is also positively related to the debt ratio. This result is consistent with expected
financial distress costs affecting the decision of whether to hedge with IR derivatives. Like in the Tobit
analysis, NOL carryforwards decrease the likelihood that a firm hedges with IR derivatives. Finally, the
value of CEO stockholdings is positively related to whether a firm uses IR derivatives, consistent with
the managerial risk aversion hypothesis. The decision of whether to hedge is not affected by tax function
convexity.
Truncated regression: The truncated regression in the second stage of the Cragg analysis
measures the extent of hedging activity. As in the Tobit analysis, the truncated regression indicates that
the extent of IR derivatives hedging is significantly affected by leverage and the debt ratio interacted with
market-book. Profitability is also negatively related with IR hedging in the truncated regression. These
findings suggest that the extent of corporate hedging is affected by the expected costs of financial
distress.
There are several differences between the Tobit and truncated regression results. The tax
convexity coefficient is significantly negative in the truncated regression, opposite what is predicted by
theory. The truncated regression coefficients for the R&D spending, NOL carryforwards, institutional
ownership, and CEO stock value are all insignificant, unlike in the Tobit analysis. As in Haushalter
(2000), size is negatively related to the extent of hedging with IR derivatives in the truncated regression.
The negative relation between size and hedging is consistent with the amount of hedging increasing with
the degree of informational asymmetry or distress costs, given that a firm hedges. Or, perhaps larger
firms are more naturally diversified through their operations and thus require smaller hedging positions.
The size result should be interpreted cautiously, however, because firm size (total assets) is the
denominator of the dependent variable.
One interpretation of the differences between the Tobit and truncated regressions is that the

24

Tobit analysis blurs the distinction between the choice of whether to hedge with the decision about how
much to hedge. An alternative interpretation is that the truncated analysis loses valuable information
about the chosen extent of (zero) hedging for some firms. To investigate this latter possibility, we look
carefully at the SFAS 119 derivative disclosures. Thirty-four firms report recent hedging activity but
currently (as of the financial statement date) are not holding derivatives. This is consistent with at least
some firms currently choosing “zero” for the extent of current hedging, a possibility that is ignored in the
truncated regression.

A.2. FX derivatives hedging
Tobit regression:21 Model FX 1 in Panel B of Table VI summarizes the Tobit regression of net
FX derivatives (scaled by assets) on the explanatory variables. The FX sample is only half the size of
the IR sample, and we lose a fair number of FX derivative users because we cannot calculate net
currency positions. Nonetheless, like in the IR sample, we find hedging with currency derivatives
increases with size, debt ratios, R&D spending, and institutional ownership. FX hedging is also higher
when the firm uses interest rate derivatives. We find no relation between tax function convexity and FX
hedging. Unlike in the IR sample, we find a positive relation between hedging and pretax ROA, but no
significant relation with the product of debt and market-book. These FX results generally support the
distress and underinvestment rationales for hedging but not the informational asymmetry, tax convexity,
or managerial risk aversion hypotheses.
Probit regression: The probit regression results in Model FX 2 show that firm size, R&D
spending, and institutional ownership affect the decision to hedge with FX derivatives. Similar to Géczy
et al. (1997), we find that the use of interest rate derivatives and the importance of foreign sales are
positively related to the probability that a firm uses foreign currency derivatives (they examine foreign
income, not sales). Also like Géczy et al. (1997), the CEO stock and options variables are not
significantly related to currency hedging.
Truncated regression: Model FX 3 shows that, for firms that hedge with FX derivatives, the
21

In addition to running separate IR and FX regressions, we perform an unreported regression pooling the IR and FX
samples, with the dependent variable equal to the sum of IR and FX hedging. This is appropriate if one believes that
IR and FX derivatives work equally well in reducing the effects of costly volatility. These results are qualitatively
similar to those shown.

25

debt ratio is positively related to the extent of hedging. This is consistent with the financial distress
hypothesis, however the coefficients on other variables do not support the distress argument.
Our results can be contrasted with Allayannis and Ofek (1998), who also perform a truncated
regression on the use of currency derivatives. Perhaps because we sample a broader cross-section of
firms and calculate net positions, we find that the debt ratio, size, and ROA are all significantly related to
currency hedging, while Allayannis and Ofek do not. Like those authors, we find that currency hedging
is unrelated to CEO stock and options holdings.

B. Do tax incentives encourage firms to hedge?
We find a strong positive relation between leverage and the extent of hedging using both IR and
FX derivatives. Thus far, we have interpreted this as evidence that high leverage leads to increased
hedging because firms hedge more when the expected costs of distress are high. Stulz (1996), Ross
(1997), and Leland (1998) theorize that causality can also run the other way: hedging with derivatives
increases debt capacity, allowing firms to use more debt and increase firm value through increased tax
deductions. In this section, we jointly model the hedging/financing decision using simultaneous equations
in an effort to determine the direction of causality.

B.1 Simultaneous equations investigation of the link between hedging and debt capacity
In our simultaneous equations model, we use a two-stage estimation procedure similar to that
used by Géczy et al. (1997). In the first stage, separate regressions are performed for derivatives
hedging and the debt ratio. In the second stage, structural equations are estimated using the predicted
values from the first-stage regressions as explanatory variables.
For the first stage, the hedging specification is estimated with a Tobit model, and the debt ratio
equation is estimated with ordinary least squares (OLS). The debt specification uses independent
variables suggested by Titman and Wessels (1988) and Graham, Lemmon, and Schallheim (1998).22
The independent variables in the debt specification are described in Appendix B. Among other things,
22

Titman and Wessels (1988) posit that a collection of variables may affect corporate capital structure: non-debt tax
shields, asset structure, growth opportunities, asset uniqueness, firm size, income volatility, and profitability. Graham
et al. (1998) use variables measuring marginal tax rates, expected financial distress costs, probability of financial
distress, growth opportunities, asset structure, and industry.

26

the debt equation explores whether the extent of derivatives hedging is an important determinant of debt
policy.
The estimated coefficients in the second-stage hedging equation (see Table VII, Panels A and
B) are similar to those presented in Table VI. Of particular interest, the coefficient on the predicted
value of the debt ratio in Column 1 is positive and significant, indicating that high-debt ratios contribute
to the incentive to hedge with both IR and FX derivatives.
In the second-stage debt regression, the predicted extent of both IR and FX derivatives hedging
is positively related to the debt ratio (Column 2 of Table VII, panels A and B, respectively).23 To our
knowledge, this is the first evidence that hedging increases the debt ratio and consequently increases
value through the tax benefit of interest deductions.
While Leland (1998) and Ross (1997) suggest that firms hedge to increase debt capacity
because of tax incentives, this relation could also be driven by non-tax factors. For example, a
nontaxable firm might hedge to increase debt capacity and use the external funding to invest in profitable
projects, even though it does not benefit from the interest tax deductions. To investigate the relative
importance of tax and non-tax incentives in our "hedging causes debt" result, we include a variable that
interacts the hedging variable with the marginal tax rate, and repeat the second stage regressions. (We
use the simulated marginal tax rates computed by Graham et al. (1998).) For example, in the Table VII,
Panel A specification, we include predicted interest rate hedging (NET_INTp*), the marginal tax rate
(MTR_BEF), and a new variable that interacts NET_INTp* with MTR_BEF.
In an untabulated regression for the IR sample, the results strongly suggest that tax incentives are
behind the hedging causes debt result: the coefficient on NET_INTp* is –0.096 (t-score of 2.8) and the
coefficient on NET_INTp* x MTR_BEF is 2.14 (t-score of 8.05). This indicates that the incentive to
hedge to increase debt capacity is positively related to the corporate marginal tax rate. Taking the partial
derivative with respect to NET_INTp*, the net effect of hedging to increase debt capacity is positive for
firms with tax rates greater than 4.4%, with the positive effect driven by tax incentives. In the FX
sample, the results are more difficult to interpret because the tax rate and the hedging variable
(NET_CURp*) are highly correlated (the correlation coefficient is 0.98 because most of the FX
23

We find the same positive effect of "hedging causing debt" in an unreported regression that measures leverage
with debt-to-value, instead of debt-to-assets.

27

hedgers have high tax rates). When both variables are included in a specification analogous to that
reported in Table VII, Panel B, the coefficient on NET_CURp* is positive and significant and the
interactive variable is insignificant. If we drop NET_CURp* and include only the interactive term, the
estimated coefficient is 2.21 (t-score of 2.51). These results are generally consistent with tax incentives
driving the “hedging causes debt” result in the FX sample but we can not conclude statistically that this is
the case.
Finally, we contrast our simultaneous results with previous research. Géczy et al. (1997) do not
find that currency hedging significantly increases the debt ratio in their second-stage regression. Note
that our simultaneous system of equations tests whether the extent of hedging affects the debt ratio,
while Géczy et al. (1997) test whether the probability of hedging affects the debt ratio. In unreported
analysis, when we change our simultaneous system and use a binary hedging variable in the first stage
(like in Géczy et al. (1997)), we no longer find that currency hedging significantly increases the debt
ratio in the second-stage regression. This analysis indicates that it is not the yes/no decision of whether
to hedge, but rather how much a firm hedges, that increases debt capacity as suggested by Stulz (1996),
Ross (1997), and Leland (1998).

B.2. Comparing tax benefits from increased debt capacity vs. those from tax function convexity
Our evidence implies that firms hedge to increase debt capacity but not in response to tax
function convexity. Although there may be other benefits, the central motive for increasing debt capacity
in several models of corporate hedging is to increase tax deductions (e.g., Stulz, 1996; Ross, 1997; and
Leland, 1998). In this section, we quantify the size of the tax benefits provided by increased debt
capacity and contrast them with the potential (but apparently unexploited) benefits associated with tax
function convexity.
Hedging increases the average (median) firm's debt ratio by 2.9% (2.2%) in the IR sample, and
by 4.5% (3.0%) in the FX sample (see the first column of Table VIII). We calculate these figures by
multiplying the estimated influence of hedging on the debt ratio (i.e., the estimated coefficients for net
derivatives in the second-stage debt regressions) by each firm’s actual scaled net holdings of IR or FX
derivatives. The mean and median increases in the debt ratios are statistically different from zero
according to a t-test and a Wilcoxon rank-sum test, respectively.

28

In column 2 of Table VIII we determine the tax benefit of the extra debt firms use when they
hedge by multiplying the incremental debt usage of each firm by its marginal tax rate. This is just an
application of the traditional “corporate tax rate times debt” formula to estimate the capitalized tax
benefits provided by debt (see APV approach in Brealey and Myers, 1996). The mean incremental tax
benefit provided by hedging is approximately $81 ($63) million in the IR (FX) samples, with median
values of $11.7 ($17.7) million.
In the third column of Table VIII, we scale each firm’s incremental debt tax benefit (DTB) from
hedging by its market value from the prior year. The mean (median) tax benefits of increased debt
capacity amount to 1.4% (0.7%) of firm value for IR sample firms and 1.4% (1.0%) for FX firms. If FX
and IR hedging are independent, this indicates that typical tax benefits of hedging from increased debt
capacity contributes between 1.7% and 2.8% to firm value, a value gain consistent with the numerical
examples in Leland (1998). The tax benefits are greater than 5% of value for some firms. To the extent
that there are also nontax benefits, these numbers are conservative estimates of the benefits of hedging
to increase debt capacity.
As a point of comparison, we calculate the tax benefits that would be provided if firms hedged
(i.e., reduce earnings volatility) in response to tax function convexity (see column 4 of Table VIII).24
For the average firm, hedging to reduce the volatility of earnings by 5% would increase the value of the
firm by approximately 0.15% (assuming that the dollar savings from convexity are a perpetuity
discounted at 10%). This indicates that the tax incentive to hedge to increase debt capacity is more than
ten times larger than the incentive provided by tax function convexity. Therefore, it is not surprising that
the regression analysis documents a tax incentive to hedge to increase debt capacity but not an incentive
to hedge in response to tax function convexity.

V. Conclusion
We investigate whether firms use derivatives to implement risk-management strategies in a
manner consistent with increasing firm value. Although derivatives offer only one means for managing

24

If we assume that firms with concave tax functions increase earnings volatility to reduce their expected tax bill, the
negative numbers in the last column of Table VIII would be positive. This would increase the mean slightly but not
by enough to change our conclusions.

29

risk, the relatively low transactions costs of engaging in a derivatives program make this an ideal setting
to study corporate hedging practices. For example, Brown (1999) studies an anonymous firm active in
currency hedging with derivatives. He estimates the annual costs associated with the currency hedging
program to be $3.8 million annually.25 This cost estimate is quite low compared to the average tax
benefits of hedging illustrated in Table VIII. Another reason to study derivatives when analyzing
corporate risk management is because they are disclosed in financial statements, while other hedging
strategies are more difficult to observe.
We find evidence that corporate hedging practices are consistent with optimal risk management.
Our results indicate that firms hedge in response to high costs of underinvestment and financial distress.
We also find that hedging increases firm value by increasing debt capacity and interest deductions. In
particular, we estimate that the tax benefits resulting from hedging add between 1.7% and 2.8% to firm
value.
In general, we find that small firms hedge less than do large firms. This result is inconsistent with
the notion that small firms face substantial informational asymmetry costs and therefore should hedge
more than large firms. Instead, this result may indicate that there are large fixed costs to implementing a
derivatives hedging program, and small firms are less likely to achieve sufficient benefits to offset this
cost.
We do not find any evidence that hedging increases with the convexity of the corporate income
tax function. This result is notable because we use the most precise measure of convexity presently
available. Our interpretation is that firms do not hedge in response to convexity because the incentive is
small relative to other hedging incentives. Another possibility is that the firms with the sharpest convexity
in their tax functions are small and have near-zero expected income. The option value of equity may
discourage these firms from hedging, offsetting the incentive to reduce expected tax liabilities. It is also
possible that firms reduce income volatility by means other than using derivatives. For example,
accounting policies can be used to smooth taxable income through time. Petersen and Thiagarajan
(1997) note that Homestake Mining, which faces substantial gold price risk but does not hold
25

The firm studied by Brown (1999) generated currency derivative transactions in 1997 with total notional value of
$15 billion and has eleven full-time employees dedicated to foreign currency risk management. The $3.8 million total
annual cost consists of $1.5 million in “fixed” costs (employee compensation and overhead expenses) and $2.3 million
in transactions costs.

30

derivatives, uses accounting policies that effectively increase taxable income during periods of low gold
prices and decrease income during periods of high prices.
Using a continuous measure of derivatives hedging allows us to identify certain aspects of risk
management that we do not detect when we use a binary variable. For example, we always find that
firms hedge in response to high debt levels when we measure the extent of derivatives usage. This is not
the case when we examine the binary decision of whether to hedge with FX derivatives. These results
are consistent with debt levels affecting the extent of corporate hedging but not the decision of whether
or not to hedge. Likewise, we find that the extent of currency hedging (but not the yes/no decision to
hedge with currency derivatives) increases debt capacity and firm value.
To summarize, with respect to the IR sample, our paper is the first to examine the "what makes
volatility costly" question on a stand-alone interest rate sample; therefore, the IR results are new. With
respect to the FX sample, some of our evidence confirms results found in other studies. Our analysis
shows that these results also hold on a broad sample of firms and when hedging is measured with a
relatively precise, continuous dependent variable. For example, like Géczy et al. (1997), we find that
size and underinvestment costs lead to increased hedging but informational asymmetry and managerial
risk aversion do not. However, we also find results that add to Géczy et al. (1997). We find evidence
that hedging is affected by expected financial distress costs. We use a direct measure of tax function
convexity and find no evidence that firms hedge in response to the shape of the tax function. Most
importantly, we find that the extent of hedging leads to increased debt capacity and increased firm value
due to tax benefits. Aggregating our results across the two samples, we find that tax incentives to
increase debt capacity, underinvestment costs, financial distress costs, and firm size are the factors that
drive corporate hedging to reduce costly volatility. Finding that these factors drive corporate hedging in
both samples is important because it suggests similar forms of costly volatility despite differing types of
risk exposure.
We close by noting that FASB is scheduled to implement SFAS 133 on June 15, 2000, to
supersede SFAS 119. SFAS 133 requires firms to report the fair market values of derivative contracts
but does not require the disclosure of notional values. Fair values of derivatives measure the amount that
the contract holder would receive, or pay, to liquidate a contract. Therefore, fair values provide
information on the extent of price movements in derivative contracts, rather than the amount of

31

derivatives held. (For example, many derivatives have a market value of zero at origination, in which
case fair market value reveals little, if anything, about derivative usage.) The ability of investors,
regulators, and researchers to determine the extent of corporate derivative holdings could be
undermined if FASB no longer requires firms to report notional values.

32

Appendix A. Derivatives disclosure of Georgia-Pacific Corporation
From Management Discussion and Analysis (MD&A):
At December 31, 1994, the Corporation's weighted average interest rate on its total debt, including the $700
million accounts receivable sale program (which is currently scheduled to expire in May 1995 although all or a portion
might be refinanced on a long-term basis in 1995) and floating rate debt, was 8.3%. At December 31, 1994, the
Corporation had outstanding interest rate exchange agreements which effectively converted $946 million of floating
rate obligations with a weighted average interest rate of approximately 5.5% to fixed rate obligations with an average
effective interest rate of approximately 9.2%. As of December 31, 1994, the Corporation's total floating rate debt,
including the accounts receivable sale program, exceeded related interest rate exchange agreements by approximately
$1.3 billion. Interest rate exchange agreements of $800 million, outstanding at December 31, 1993, expired during 1994.
Interest rate exchange agreements of $450 million, outstanding at December 31, 1994, will expire during 1995.
The Corporation has disclosed the fair value of its short-term and long-term debt and its interest rate
exchange agreements in accordance with Financial Accounting Standard Number 107 (SFAS 107), "Disclosures about
Fair Value of Financial Instruments." In addition, the Corporation adopted Financial Accounting Standard Number
119, "Disclosure about Derivative Financial Instruments and Fair Value of Financial Instruments," effective December
31, 1994, which requires entities to provide in their financial statements information about derivative financial
instruments and to provide a summary of this information together with FAS 107 disclosures. Refer to Note 6 of the
Notes to Financial Statements for this presentation.
From Note 6 (FINANCIAL INSTRUMENTS) in Footnotes to Financial Statements:
The carrying amount and fair value of the Corporation's financial instruments are as follows:
December 31, 1994
December 31, 1993
----------------------------------------Carrying
Fair
Carrying
Fair
(Millions)
Amount
Value
Amount
Value
- -----------------------------------------------------------------------------Commercial paper and
other short-term notes
(Note 5)
$ 868
$ 868
$ 650
$ 650
Notes and debentures
(Note 5)
3,528
3,520
3,778
4,172
Revenue bonds
(Note 5)
389
385
375
375
Other loans
(Note 5)
53
53
96
96
Interest rate exchange
agreements
*
12
*
99
Accounts receivable
sale program
(Note 4)
700
700
700
700
- -----------------------------------------------------------------------------* The Corporation accrued interest of $10 million and $31 million at December 31, 1994 and 1993, respectively, related
to these agreements.
INTEREST RATE AND FOREIGN CURRENCY EXCHANGE AGREEMENTS. The Corporation has used interest rate
and foreign currency exchange agreements in the normal course of business to manage and reduce the risk inherent
in interest rate and foreign currency fluctuations.
Under the interest rate exchange agreements, the Corporation makes payments to counterparties at fixed
interest rates and in turn receives payments at variable rates. The Corporation entered into interest rate exchange
agreements in prior years to protect against the increased cost associated with a rise in interest rates. During 1994,

33

$800 million in interest rate exchange agreements expired. At December 31, 1994, the Corporation had outstanding
interest rate exchange agreements which effectively converted $946 million of floating rate obligations with a
weighted average interest rate of 5.5 % to fixed rate obligations with an average effective interest rate of
approximately 9.2%. These agreements have a weighted average maturity of approximately 2.3 years. During 1995,
$450 million of these agreements will expire. As of December 31, 1994, the Corporation's total floating rate debt,
including the accounts receivable sale program, exceeded related interest rate exchange agreements by approximately
$1.3 billion.
The estimated fair value of the Corporation's liability under interest rate exchange agreements at December
31, 1994 and 1993 was $12 million and $99 million, respectively, and represents the estimated amount the Corporation
could have paid to terminate the agreements. The fair value at December 31,1994 and 1993 was estimated by
calculating the present value of anticipated cash flows. The discount rate used was an estimated borrowing rate for
similar debt instruments with like maturities. The Corporation accrued interest of $10 million and $31 million at
December 31, 1994 and 1993, respectively, related to these agreements. The Corporation enters into foreign exchange
contracts, futures and options, the amounts of which were not material to the consolidated financial position of the
Corporation at December 31, 1994.
The Corporation may be exposed to losses in the event of nonperformance of counterparties, but does not
anticipate such nonperformance.

Appendix B. Independent variables used in debt specification
ITCp: investment tax credits scaled by book value of assets.
INTANGp: book value of intangible assets as a percentage of book value of total assets.
RD_EXPp: research and development expense as a percentage of net sales revenues.
SGAp: selling, general, and administrative expense as a percentage of net sales revenues.
LN_SALES: natural logarithm of net sales revenues.
VOL: absolute coefficient of variation of taxable income.
MTR_BEF: before-financing simulated marginal tax rate used in Graham et al. (1998).
OENEG: dummy variable indicating that the book value of shareholders’ equity is negative.
PPE_Np: book value of property, plant, and equipment, net of depreciation as a percentage of book
value of total assets.

34

References
Allayannis, Y., Ofek, E., 1998, Exchange rate exposure, hedging, and the use of foreign currency
derivatives, unpublished working paper, Darden School of Business, University of Virginia.
Allayannis, Y., Weston, J., 1998, The use of foreign currency derivatives and firm market value,
unpublished working paper, Darden School of Business, University of Virginia.
Berkman, H., Bradbury, M.E., 1996, Empirical evidence on the corporate use of derivatives, Financial
Management 25(2), 5-13.
Bessembinder, H., 1991, Forward contracts and firm value: Investment incentive and contracting effects,
Journal of Financial and Quantitative Analysis 26, 519-532.
Bicksler, J., Chen, A.H., 1986, An economic analysis of interest rate swaps, Journal of Finance 41, 645655.
Bodnar, G.M., Hayt, G.S., Marston, R.C., 1996, 1995 Wharton survey of derivatives usage by US nonfinancial firms, Financial Management 25(4), 113-133.
Brealey, R.A., Myers, S.C., 1996, Principles of Corporate Finance, fifth edition. McGraw-Hill, New
York.
Breeden, D., Viswanathan, S., 1998, Why do firms hedge? An asymmetric information model,
unpublished working paper, Fuqua School of Business, Duke University.
Brown, G., 1999, Managing foreign exchange risk with derivatives, unpublished working paper, KenanFlagler School of Business, University of North Carolina.
Core, J., Guay, W., 1999, A new, low-cost proxy for the incentive effects of stock option portfolios,
unpublished working paper, Wharton School, University of Pennsylvania.
Cragg, J.G., 1971, Some statistical models for limited dependent variables with application to the demand
for durable goods, Econometrica 39, 829-844.
DeMarzo, P., Duffie, D., 1991, Corporate financial hedging with proprietary information, Journal of
Economic Theory 53, 261-286.
Dolde, W., 1995, Hedging, leverage, and primitive risk, Journal of Financial Engineering 4, 187-216.
Fenn, G., Post, M., Sharpe, S., 1996, Debt maturity and the use of interest rate derivatives, unpublished
working paper, Federal Reserve Board.
Financial Accounting Standards Board, 1994, Disclosure about derivative financial instruments and fair
value of financial instruments, Statement No. 119.

35

Financial Accounting Standards Board, 1998, Accounting for derivative instruments and hedging activities,
Statement No. 133.
Froot, K.A., Scharfstein, D.S., Stein, J.C., 1993, Risk management: coordinating corporate investment and
financing policies, Journal of Finance 48, 1629-1658.
Gay, G.D., Nam, J., 1999, The underinvestment problem and corporate derivatives use, Financial
Management 27(4), 53-69.
Géczy, C., Minton, B.A., Schrand, C., 1997, Why firms use currency derivatives, Journal of Finance 52,
1323-1354.
Graham, J.R., Lemmon, M.L., Schallheim, J.S., 1998, Debt, leases, taxes, and the endogeneity of
corporate tax status, Journal of Finance 53, 131-162.
Graham, J.R., Smith, C.W., Jr., 1999, Tax incentives to hedge, Journal of Finance 54, 2241-2262.
Guay, W.R., 1999, The impact of derivatives on firm risk: An empirical examination of new derivatives
users, Journal of Accounting and Economics 26, 319-351.
Haushalter, G.D., 2000, Financing policy, basis risk, and corporate hedging: Evidence from oil and gas
producers, Journal of Finance forthcoming.
Hentschel, L., Kothari, S.P., 1998, Are corporations reducing or taking risks with derivatives?, unpublished
working paper, William E. Simon Graduate School of Business Administration, University of
Rochester.
Howton, S.D., Perfect, S.B., 1999, Currency and interest-rate derivatives use in US firms, Financial
Management 27(4), 111-121.
Leland, H.E., 1998, Agency costs, risk management, and capital structure, Journal of Finance 53, 12131243.
Maddala, G.S., 1983. Limited Dependent and Qualitative Variables in Econometrics. Cambridge
University Press, Cambridge, UK.
Mayers, D., Smith, C.W., Jr., 1987, Corporate insurance and the underinvestment problem, Journal of
Risk and Insurance 54, 45-54.
Mian, S., 1996, Evidence on corporate hedging policy, Journal of Financial and Qualitative Analysis 31,
419-439.
Miller, M.H., Modigliani, F., 1961, Dividend policy, growth, and the valuation of shares, Journal of Business
34, 411-433.
Modigliani, F., Miller, M.H., 1958, The cost of capital, corporation finance and the theory of investment,
American Economic Review 48, 261-297.

36

Myers, S.C., 1977, Determinants of corporate borrowing, Journal of Financial Economics 5, 147-175.
Myers, S.C., Majluf, N.S., 1984, Corporate financing and investment decisions when firms have
information that investors do not have, Journal of Financial Economics 13, 187-221.
Nance, D.R., Smith, C.W., Jr., Smithson, C.W., 1993, On the determinants of corporate hedging, Journal
of Finance 48, 267-284.
Petersen, M.A., Thiagarajan, S.R., 1997, Risk measurement and hedging, unpublished working paper, J.
L. Kellogg Graduate School of Management, Northwestern University.
Ross, M.P., 1997, Corporate hedging: What, why and how?, unpublished working paper, Walter A. Haas
School of Business, University of California, Berkeley.
Schrand, C., 1998, Discussion: Who uses interest rate swaps? A cross-sectional analysis, Journal of
Accounting, Auditing, and Finance 13(3), 201-205.
Schrand, C., Unal, H., 1998, Hedging and coordinated risk management: Evidence from thrift conversions,
Journal of Finance 53, 979-1013.
Simkins, B.J., 1997, Asymmetric information, credit quality, and the use of interest rate derivatives,
unpublished working paper, Oklahoma State University.
Smith, C.W., Jr., Stulz, R.M., 1985, The determinants of firms’ hedging policies, Journal of Financial and
Quantitative Analysis 20, 391-405.
Smith, C.W., Jr., 1995, Corporate risk management: Theory and practice, Journal of Derivatives 2(4), 2130.
Smithson, C.W., Smith, C.W., Jr., Wilford, D.S., 1995, Managing Financial Risk: A Guide to Derivative
Products, Financial Engineering, and Value Maximization. Irwin Professional Publishing, Chicago.
Stulz, R.M., 1984, Optimal hedging policies, Journal of Financial and Quantitative Analysis 19, 127-140.
Stulz, R.M., 1996, Rethinking risk management, Journal of Applied Corporate Finance 9(3), 8-24.
Titman, S., 1992, Interest rate swap and corporate financing choices, Journal of Finance 47, 1503-1516.
Titman, S., Wessels, R., 1988, The determinants of capital structure choice, Journal of Finance 43, 1-19.
Tufano, P., 1996, Who manages risk? An empirical examination of risk management practices in the gold
mining industry, Journal of Finance 51, 1097-1137.
Tufano, P., 1998, Agency costs of corporate risk management, Financial Management 27(1), 67-77.
Visvanathan, G., 1998, Who uses interest rate swaps? A cross-sectional analysis, Journal of Accounting,
Auditing, and Finance 13(3), 173-200.

37

Wong, M.H. Franco, 1997, The usefulness of SFAS 119 derivatives disclosures for assessing the foreign
exchange risk exposure of manufacturing firms, Journal of Accounting Research forthcoming.

38

Fig. 1. The effect of convexity in the tax function on expected tax liability. The figure provides a simple
illustration of the Smith and Stulz (1985) theory. Consider a firm that realizes taxable income levels of A
or B, each with a probability of 1/2. Its income is taxed according to a convex tax function with
expected tax liability shown at E[T]. If the firm eliminates uncertainty about the realization of taxable
income, it earns income level C with certainty, and lowers its expected tax liability to E[T|H]. The
reduction in volatility of taxable income causes a reduction in expected tax liability and a corresponding
increase in market value.

39

Table I
Summary of empirical papers that study why firms hedge with derivatives
Table I summarizes eleven recent papers (including our paper) that investigate the determinants of corporate hedging. The method column indicates the type of
multivariate analysis performed. Derivative type describes the dependent variable used to quantify hedging activity. Net indicates whether the firms in the sample
were required to report the direction of the derivative position, which permits calculation of net hedging positions. Ex ante describes whether the sample is
restricted to firms with ex ante risk exposure of a type that can be managed via hedging. Sample identifies the population from which the firms in the sample are
drawn and the data source. The last nine columns summarize the variables, predicted signs, and consistency of the evidence for hypotheses linking corporate
hedging to costs of distress, underinvestment, and tax convexity.
Study

Method

Derivative type

Net

Ex ante

Sample

Distress

Nance, et al.
(1993)

Logit

All types
binary

no

no

Fortune500/S&P400
survey

Interest coverage
Leverage
Market value

+
-

no
no
no

Underinvestment
R&D
Book/market

+
-

weak yes
no

Tax convexity
NOL cf./value
ITC/value
Prog. Region

+
+
+

no
no
no

Dolde (1995)

Logit

All types
binary

no

yes

Fortune 500
survey

Leverage

+

yes

R&D

+

yes

NOL cf.

+

no

Berkman &
Bradbury (1996)

Tobit

All types
notional and fair values

no

no

NZ public firms
fin. statements

Interest coverage
Leverage

+

yes
yes

E/P ratio
Asset growth

+

no
no

NOL cf. dummy

+

yes

Mian (1996)

Logit

All, currency, interest rate
binary

no

no

3,022 firms
fin. statements

Firm value

-

no

Market/book

+

no

Prog. Region
NOL cf. dummy
FTC dummy

+
+
+

yes
no
no

Tufano (1996)

Tobit

Gold
% production hedged

yes

yes

Gold mining firms
survey

Cash costs of gold
Leverage

+
+

no
weak yes

various

no

NOL cf./value

+

no

Geczy, et al.
(1997)

Logit

Currency
binary

no

yes

Fortune 500
fin. statements

Leverage
credit quality

+
-

no
no

R&D
Debt ratio
Debt*(M/B)

+
+
+

yes
no
yes

NOL cf./TA

+

no

Allayannis &
Ofek (1998)

Tobit
Cragg

Currency
notional values

no

no

S&P 500
fin. statements

Leverage
EBIT/Assets

+
-

no
no

R&D
Market/book

+
+

yes
no

NOL/ITC dummy

+

no

Haushalter
(1998)

Tobit
Cragg

Oil & gas
% production hedged

yes

yes

Oil & gas firms
survey

Leverage

+

yes

Inv. expend.
+
Fin. policy vars

no
yes

MTR
Prog. Region

+

no
no

Gay & Nam
(1999)

Tobit

All types
notional values

no

no

Bus. Week 1000 and
Swaps Monitor database

Leverage
Interest coverage

+
-

yes
no

Various

+

yes

NOL cf./TA

+

no

Howton &
Perfect (1999)

Tobit

All, currency, interest rate.
notional values

no

no

Fortune/S&P500
and 461 random firms
fin. statements

Interest coverage
Leverage
Tangible assets

+
-

no
yes
no

R&D
Cash flow/TA

+
-

no
yes

NOL cf. dummy
Prog. Region

+
+

no
no

Our paper

Tobit
Cragg

currency, interest rate.
net notional values

yes

yes

all sizes, random.
fin. statements

Leverage
Debt*(M/B)

+
+

yes
weak yes

B/M
R&D

+

no
yes

Convexity of
tax function

+

no

Table II
Derivatives hedging by industry
Table II provides descriptive statistics of the derivatives hedging of sample firms by industry. Panel A describes the
number of interest rate derivative hedgers and non-users in each industry classification in the interest rate (IR) sensitive
sample of 404 firms. Panel B describes the number of foreign currency derivative hedgers and non-users in each industry
classification in the foreign currency (FX) sensitive sample of 242 firms. A derivatives hedger is defined as a firm that
discloses derivative holdings at the end of its fiscal year (1994 or 1995) for non-trading purposes, while a non-user does
not disclose any such holdings.

Panel A: IR Sample
Industry

SIC
Code

Food
Mi n i n g
Construction
Manufacturing
Transportation
Communi c a t i o n
Ut i l i t y
Whol e s a l e
Retail
Bank
Investment
Insurance
Financial Service
Service

0100 & 2000-2099
1000-1499
1500-1999
2200-3999
4000-4799
4800-4899
4900-4999
5000-5199
5200-5999
6000-6099
6100-6299
6300-6499
6500-6599
7000-9999

Total

N

IR deriv.
hedgers

Non-users

% I R deriv.
hedgers

Percentage
non-users

15
10
7
162
15
17
41
20
26
34
16
5
4
32

3
4
3
55
7
5
8
4
8
23
9
3
0
10

12
6
4
107
8
12
33
16
18
11
7
2
4
22

20.00%
40.00%
42.86%
33.95%
46.67%
29.41%
19.51%
20.00%
30.77%
67.65%
56.25%
60.00%
0.00%
31.25%

80.00%
60.00%
57.14%
66.05%
53.33%
70.59%
80.49%
80.00%
69.23%
32.35%
43.75%
40.00%
100.00%
68.75%

404

142

262

35.15%

64.85%

Panel B: FX Sample
Industry

SIC
Code

Food
Mi n i n g
Construction
Manufacturing
Transportation
Communi c a t i o n
Ut i l i t y
Whol e s a l e
Retail
Bank
Investment
Insurance
Financial Service
Service

0100 & 2000-2099
1000-1499
1500-1999
2200-3999
4000-4799
4800-4899
4900-4999
5000-5199
5200-5999
6000-6099
6100-6299
6300-6499
6500-6599
7000-9999

Total

N

FX deriv.
hedgers

Non-users

% FX deriv.
hedgers

Percentage
non-users

5
12
2
151
6
2
3
12
4
10
4
6
0
25

3
2
0
74
4
1
1
5
1
4
2
2
0
6

2
10
2
77
2
1
2
7
3
6
2
4
0
19

60.00%
16.67%
0.00%
49.01%
66.67%
50.00%
33.33%
41.67%
25.00%
40.00%
50.00%
33.33%
0.00%
24.00%

40.00%
83.33%
100.00%
50.99%
33.33%
50.00%
66.67%
58.33%
75.00%
60.00%
50.00%
66.67%
0.00%
76.00%

242

105

137

43.39%

56.61%

Table III
Notional values of contracts held by derivative hedgers
Table III, Panel A (Panel B) provides descriptive statistics of the derivative positions held by interest rate, IR, (foreign
currency, FX) sensitive sample firms at the end of fiscal 1994 or 1995. Total derivative positions are defined as the sum of
notional values of all derivative contracts. Net positions are defined as the sum across the three derivative categories
(interest rate, currency, and commodity) of the absolute values of the difference between “long” and “short” positions in
each of the categories. Panel A (Panel B) provides detail on IR (FX) derivative holdings by position (long, short, unsure,
etc.), total notional values, and net positions. All values are in millions of dollars.
Panel A: IR Sample
Derivative Type

Position
Type

Total positions
Scaled by assets

N

Mean

Std dev

Min

Max

Median

180

2,803
18.59%

8,908
37.95%

1
0.05%

68,009
463.65%

210
10.67%

Net positions
Scaled by assets

Total net

159

699
11.21%

1,655
15.61%

0
0.00%

12,312
134.35%

125
7.34%

Interest rate

Long
Scaled by assets
Short
Scaled by assets
Basis swaps
Scaled by assets
Unsure
Scaled by assets
Total
Scaled by assets
Net
Scaled by assets

115

739
10.76%
1,943
11.74%
558
3.82%
8,825
13.66%
3,008
16.90%
728
9.50%

2,259
26.55%
4,181
26.40%
754
4.27%
14,694
19.02%
9,027
41.37%
1,783
14.53%

3
0.10%
7
0.08%
50
0.19%
3
0.05%
3
0.05%
0
0.00%

13,950
250.02%
18,400
211.30%
3,000
13.65%
53,800
69.13%
66,183
463.65%
12,194
134.35%

113
5.04%
400
6.32%
236
1.88%
330
4.65%
213
8.20%
110
5.52%

Interest rate
only users

Total
Scaled by assets

69

1,978
21.95%

5,399
57.56%

3
0.05%

26,025
463.65%

125
8.82%

Currency only
users

Total
Scaled by assets

24

130
6.54%

192
7.50%

1
0.16%

760
28.15%

67
3.06%

Commodity only
users

Total
Scaled by assets

11

358
17.69%

663
25.42%

14
0.88%

1,972
85.50%

63
7.34%

Interest rate
and
Currency users

IR Total
Scaled by assets
Currency Total
Scaled by assets

61

4,696
13.30%
1,067
7.17%

12,364
14.10%
2,180
9.39%

5
0.15%
1
0.00%

66,183
76.45%
12,600
51.35%

377
8.52%
192
3.16%

Interest rate
and
Commodity users

IR Total
Scaled by assets
Commodity Total
Scaled by assets

7

90
5.68%
162
8.76%

85
3.96%
210
10.57%

33
0.60%
4
0.12%

275
11.62%
473
25.42%

55
5.52%
21
1.39%

Users of all
3 types

IR Total
Scaled by assets
Currency Total
Scaled by assets
Commodity Total
Scaled by assets

5

702
6.76%
486
6.71%
107
0.92%

729
4.16%
826
12.59%
101
0.59%

150
2.20%
25
0.14%
5
0.08%

1,980
11.50%
1,942
29.18%
223
1.60%

465
6.99%
62
1.85%
54
1.07%

67
16
23
142
131

Table III (continued)
Panel B: FX Sample
Derivative Type

Position
Type

Total positions
Scaled by assets
Net positions
Scaled by assets

Total net

Currency

Long
Scaled by assets
Short
Scaled by assets
Unsure
Scaled by assets
Total
Scaled by assets
Net
Scaled by assets

N

Mean

Std dev

Min

Max

Median

138

2,751
14.93%

9,346
14.50%

1
0.16%

68,009
86.32%

171
10.37%

116

634
10.13%

1,693
11.69%

0
0.00%

12,312
85.50%

121
7.41%

48

215
4.73%
246
6.61%
988
5.33%
684
7.38%
195
5.96%

401
5.47%
374
7.59%
2,320
5.51%
1,733
8.22%
318
6.45%

1
0.00%
1
0.11%
1
0.05%
1
0.14%
1
0.13%

1,909
22.88%
1,495
36.14%
12,600
23.84%
12,600
51.35%
1,605
36.14%

69
2.73%
62
4.89%
110
3.03%
80
5.17%
67
4.11%

41
52
105
64

Interest rate
only users

Total
Scaled by assets

27

1,996
11.27%

5,464
8.39%

3
0.29%

26,025
35.23%

193
7.80%

Currency only
users

Total
Scaled by assets

41

134
6.84%

257
5.87%

1
0.16%

1,453
23.56%

55
6.08%

Commodity only
users

Total
Scaled by assets

5

347
26.65%

570
33.52%

14
3.97%

1,348
85.50%

63
17.13%

Interest rate
and
Currency users

IR Total
Scaled by assets
Currency Total
Scaled by assets

57

4,373
13.06%
1,124
7.88%

12,131
11.73%
2,245
9.44%

5
0.69%
1
0.21%

66,183
43.79%
12,600
51.35%

377
9.07%
192
4.61%

Currency
and
Commodity users

Currency Total
Scaled by assets
Commodity Total
Scaled by assets

4

48
3.67%
272
11.84%

30
2.48%
505
17.81%

15
1.88%
7
0.76%

74
7.34%
1,029
38.13%

51
2.72%
26
4.24%

Users of all
3 types

IR Total
Scaled by assets
Currency Total
Scaled by assets
Commodity Total
Scaled by assets

3

460
7.14%
676
10.39%
94
0.97%

72
4.28%
1,096
16.30%
114
0.80%

386
2.95%
25
0.14%
5
0.08%

530
11.50%
1,942
29.18%
223
1.60%

465
6.99%
62
1.85%
54
1.24%

Table IV
Summary statistics and tests of differences between derivative hedgers and non-users
Table IV provides summary information for the independent variables. Panel A shows descriptive statistics for the full IR
and FX samples. Panel B presents the results of statistical tests differentiating the independent variables between
derivative hedgers vs. non-users. The first two columns of Panel B show the results of a t-test of the differences in means
between derivative hedgers and non-users. The two rightmost columns contain the results of Wilcoxon rank-sum tests on
the differences between derivative hedgers and non-users.

Panel A: Descriptive statistics of full samples
IR Sample
Variable
Tax convexity (dollars)
Tax convexity (scaled by sales)
Debt-to-assets
Debt ratio x market/book
R&D-to-assets
Capital expenditures-to-assets
Book/market of equity
Dividend yield
Quick ratio
Floating rate debt-to-total debt
Log (BV of assets)
Pretax ROA
NOL carryforwards-to-assets
Institutional ownership pct.
Dual class dummy
Log (CEO stock value)
# of CEO options
Currency hedging dummy

N
375
375
404
379
391
391
379
379
336
404
404
404
404
360
360
338
380
404

Mean
0.4391
0.0004
0.3421
0.8096
0.0179
0.0613
0.4686
0.0206
0.3700
0.3741
6.5752
0.0434
0.0426
0.4054
0.0333
0.3443
0.3106
0.2302

Median
0.0278
0.0001
0.3005
0.4285
0
0.0486
0.6285
0.0140
0.0983
0.2400
6.4647
0.0530
0
0.4095
0
1.0588
0.1302
0

Std Dev
2.7752
0.0009
0.2154
3.1790
0.1015
0.0650
2.5478
0.0237
0.8887
0.3479
2.1096
0.1302
0.2459
0.2266
0.1798
3.6074
0.5266
0.4215

Min
-6.9485
-0.0017
0.1003
-9.8650
0
0
-46.7768
0
0
0
-0.1936
-1.7866
0
0.0005
0
-13.8155
0
0

Max
40.9785
0.0113
2.3835
36.9232
1.8884
0.7210
2.8588
0.1444
8.2052
1.7901
12.2008
0.3768
3.0364
0.9038
1
6.8048
5.1815
1

224
224
242
234
241
236
234
235
218
235
242
242
242
231
231
224
235
242

Mean
0.6070
0.0003
0.2492
0.8712
0.0303
0.0594
0.5276
0.0206
0.4799
0.1914
6.7781
0.0715
0.0227
0.4959
0.0433
0.8014
0.3819
0.3636

Median
0.0431
0.0001
0.2142
0.3473
0.0075
0.0507
0.5018
0.0130
0.1827
0.1264
6.5102
0.0743
0
0.5296
0
1.2646
0.2000
0

Std Dev
3.5339
0.0010
0.1974
3.6857
0.0527
0.0459
0.7371
0.0626
0.9395
0.2125
1.9542
0.0894
0.1082
0.2226
0.2040
3.1952
0.5193
0.4820

Min
-6.9485
-0.0017
0
-7.6747
0
0
-7.6902
0
0
0
2.0517
-0.4000
0
0.0026
0
-13.8155
0
0

Max
40.9785
0.0113
1.1448
36.9232
0.4025
0.3528
6.1415
0.9174
7.7263
1.0000
12.2008
0.2905
1.0069
0.9038
1
6.8048
3.1778
1

FX Sample
Variable
Tax convexity (dollars)
Tax convexity (scaled by sales)
Debt-to-assets
Debt ratio x market/book
R&D-to-assets
Capital expenditures-to-assets
Book/market of equity
Dividend yield
Quick ratio
Foreign sales-to-total sales
Log (BV of assets)
Pretax ROA
NOL carryforwards-to-assets
Institutional ownership pct.
Dual class dummy
Log (CEO stock value)
# of CEO options
Interest rate hedging dummy

N

Table IV (continued)
Panel B: Analysis of Differences in Derivative Hedgers vs. Non-users
IR Sample

Variable
Tax convexity (dollars)
Tax convexity (scaled by sales)
Debt-to-assets
Debt ratio x market/book
R&D-to-assets
Capital expenditures-to-assets
Book/market of equity
Dividend yield
Quick ratio
Floating rate debt-to-total debt
Log (BV of assets)
Pretax ROA
NOL carryforwards-to-assets
Institutional ownership pct.
Dual class dummy
Log (CEO stock value)
# of CEO options
Currency hedging dummy

Difference
in Means
0.5869
-0.0002
0.0152
0.8200
-0.0066
-0.0067
0.0052
0.0086
-0.1515
0.0242
2.2123
0.0101
-0.0540
0.1699
0.0055
2.1669
0.2488
0.3617

t-test of difference
Higher
Wilcoxon
p-value
Wilcoxon score p-value
0.1426
Non-users
0.5622
0.0078
Non-users
0.0040
0.4994
Deriv. users
0.7017
0.0547
Deriv. users
0.1446
0.4201
Deriv. users
0.1163
0.3763
Non-users
0.0335
0.9810
Non-users
0.0877
0.0006
Deriv. users
0.0001
0.0736
Non-users
0.3993
0.4240
Deriv. users
0.1188
0.0001
Deriv. users
0.0001
0.3483
Non-users
0.1472
0.0047
Non-users
0.1865
0.0001
Deriv. users
0.0001
0.7782
Deriv. users
0.7792
0.0001
Deriv. users
0.0001
0.0002
Deriv. users
0.0001
0.0001
Deriv. users
0.0001

Difference
in Means
1.0577
-0.0002
0.0106
0.5731
0.0119
-0.0017
-0.1904
0.0056
-0.1515
0.0739
1.7432
0.0162
-0.0209
0.1615
-0.0448
1.1589
0.2375
0.3321

t-test of difference
Higher
Wilcoxon
p-value
Wilcoxon score p-value
0.0506
Non-users
0.9465
0.1004
Non-users
0.1363
0.6912
Deriv. users
0.9533
0.2872
Deriv. users
0.0735
0.0833
Deriv. users
0.0007
0.7841
Deriv. users
0.2963
0.0406
Non-users
0.0148
0.0623
Deriv. users
0.0024
0.0836
Non-users
0.0116
0.0107
Deriv. users
0.0019
0.0001
Deriv. users
0.0001
0.1613
Deriv. users
0.4093
0.1471
Deriv. users
0.0824
0.0001
Deriv. users
0.0001
0.1068
Non-users
0.1089
0.0054
Deriv. users
0.0003
0.0013
Deriv. users
0.0125
0.0001
Deriv. users
0.0001

FX Sample

Variable
Tax convexity (dollars)
Tax convexity (scaled by sales)
Debt-to-assets
Debt ratio x market/book
R&D-to-assets
Capital expenditures-to-assets
Book/market of equity
Dividend yield
Quick ratio
Foreign sales-to-total sales
Log (BV of assets)
Pretax ROA
NOL carryforwards-to-assets
Institutional ownership pct.
Dual class dummy
Log (CEO stock value)
# of CEO options
Interest rate hedging dummy

Table V
Pearson correlation coefficients
Table V presents the Pearson correlations of each independent variable with the derivative variables. Tax convexity is the dollar tax benefit from a 5% volatility
reduction, scaled by revenues. Debt ratio is the ratio of debt to book value of assets. Debt x mkt/book is the debt ratio multiplied by the market-book ratio of
equity. R&D is research and development expense scaled by book value of assets. Cap expend is cash paid for capital expenditures scaled by book value of
assets. Book/Market is the book-market ratio of equity. Float rate debt is the sum of debt in current liabilities plus long-term floating rate debt scaled by total debt.
Foreign sales is the ratio of foreign sales to total sales. Total assets are the natural logarithm of book value of assets. Pretax ROA is earnings before taxes scaled
by book value of assets. NOL cf’s are net operating loss carryforwards scaled by book value of assets. Institutional pct is the percentage of common shares
owned by institutional investors. Dual class dummy is an indicator variable equal to 1 if multiple classes of common equity exist and zero otherwise. Log(CEO
stock) is the natural logarithm of the value of CEO stock holdings. # CEO options are the number of stock options held by the CEO.

Total IR deriv. (scaled)
Net IR deriv. (scaled)
Binary IR deriv. var.
Total FX deriv. (scaled)
Net FX deriv. (scaled)
Binary FX deriv. var.

Total IR deriv. (scaled)
Net IR deriv. (scaled)
Binary IR deriv. var.
Total FX deriv. (scaled)
Net FX deriv. (scaled)
Binary FX deriv. var.

Tax convexity

Debt ratio

Debt x mkt/book

R&D

Cap. expend.

Book/Market

Dividend yield

Float rate debt

-0.0336
-0.0469
-0.01111**
-0.0742
-0.0660
-0.0996

0.1760***
0.1966***
0.0337
0.1021
0.2052***
0.0275

0.1306**
0.3544***
0.1245**
0.0673
0.1311*
0.0828

-0.0240
-0.0362
-0.0310
0.0924
0.0250
0.0940

-0.0315
0.0386
-0.0493
0.0087
0.0401
-0.0111

-0.0088
-0.0142
0.0010
-0.0877
-0.0897
-0.1420**

0.0116
-0.0354
0.1760***
-0.0406
-0.0442
-0.0124

0.0922*
0.0481
0.0399

Foreign sales

Total assets

Pretax ROA

NOL cf's

Institutional pct.

Dual class dummy Log (CEO stock) # CEO options

0.1857***
0.1282*
0.1576**

0.1683***
0.1210**
0.5013***
0.2411***
0.1667**
0.4297***

-0.0177
-0.0186
0.0372
0.1263**
0.1493**
0.0939

-0.0362
-0.0544
-0.1050**
-0.0227
-0.0414
-0.0815

0.1641***
0.1362**
0.3641***
0.2928***
0.2575***
0.3357***

0.0276
0.1049**
0.0149
-0.0345
0.0057
-0.1052

* denotes significance at the 10% level.
** denotes significance at the 5% level.
*** denotes significance at the 1% level.

0.1298**
0.1777***
0.2931***
0.1164*
0.1277*
0.1801***

0.1071**
0.0861*
0.2275***
0.3041***
0.1095
0.2272***

Table VI
Multivariate analysis of derivatives hedging
Table VI summarizes multivariate regression analysis of derivatives hedging. Panel A shows results for the interest
rate (IR) sample. Model IR 1 shows slope estimates and p-values from a Tobit regression of net IR derivatives scaled
by book value of assets on the listed independent variables. Model IR 1B shows results from a Tobit regression of
net IR derivatives scaled by book value of debt. Model IR 2 summarizes a binomial probit regression of a binary
variable that equals one if total IR derivatives are positive, and zero otherwise, on the listed variables. Model IR 3
shows coefficient estimates and p-values from a truncated regression of positive values of net IR derivatives scaled
by book value of assets on the listed variables. Panel B shows analogous results for the foreign currency (FX)
sample. Slope estimates are shown (rather than the maximum likelihood parameter estimates) for the Tobit and probit
regressions. The slope estimates represent the marginal effects of changes in the independent variables on the value
of the observed dependent variable, calculated at the mean levels of the independent variables.
Panel A: IR Sample
Model Type:
Dependent Variable:
Variable
Intercept
Tax Convexity
Debt-to-assets
Debt ratio x market/book
R&D-to-assets
Capital expenditures-to-assets
Book/market of equity
Floating rate debt-to-total debt
Log(BV of assets)
ROA
NOL carryforwards-to-assets
Institutional ownership pct.
Dual class dummy
Log (CEO stock value)
# of CEO options
SIC 63 dummy
Log likelihood
Number of observations
IR derivative hedgers
IR non-users

Model IR 1
Model IR 1B
Tobit Regression
Tobit Regression
Net IR scaled by assets Net IR scaled by debt
Slope
Slope
estimate
p-value estimate p-value
-0.1128
-8.0098
0.0811
0.0030
0.1934
0.6726
-0.0016
0.0257
0.0062
-0.1302
-0.1292
0.0370
0.0137
0.0033
0.0008
0.0320
-27.78
307
104
203

0.0001
0.2252
0.0001
0.0004
0.0626
0.2463
0.6428
0.0044
0.0009
0.0121
0.0212
0.0150
0.3778
0.0111
0.8690
0.0482

-0.2800
-20.3841
0.1023
0.0041
0.4740
0.0898
-0.0043
0.0880
0.0201
-0.2941
-0.3269
0.0855
0.0549
0.0093
0.0024
0.0479
-124.06
307
104
203

0.0001
0.2416
0.0439
0.0702
0.0803
0.3747
0.6284
0.0002
0.0001
0.0299
0.0247
0.0323
0.1618
0.0058
0.8444
0.2658

Model IR 2
Probit Regression
1 if Total IR > 0, or zero
Slope
estimate
p-value
-1.4621
1.0224
0.5032
0.0168
2.2292
0.1149
-0.0271
0.3115
0.1359
-0.7041
-1.1169
0.2673
0.1122
0.0331
0.0652
0.1314
-131.87
307
107
200

0.0001
0.9871
0.0239
0.4164
0.0330
0.7838
0.5288
0.0016
0.0001
0.2152
0.0351
0.0904
0.5224
0.0070
0.3598
0.4989

Model IR 3
Truncated Regression
Net IR scaled by assets
Coefficient
estimate
p-value
0.2083
-235.7400
0.7659
0.0188
1.2692
0.4532
0.0344
0.2304
-0.1182
-3.2020
0.0899
0.2743
0.0573
0.0178
0.0165
0.4750
167.03
104
104
0

0.4215
0.0177
0.0061
0.0124
0.5955
0.2447
0.2613
0.0767
0.0001
0.0016
0.9544
0.2856
0.7575
0.5262
0.7931
0.0049

Table VI (continued)
Panel B: FX Sample
Model Type:
Dependent variable:
Variable
Intercept
Tax Convexity
Debt-to-assets
Debt ratio x market/book
R&D-to-assets
Capital expenditures-to-assets
Dividend yield
Book/market of equity
Foreign sales-to-total sales
Log(BV of assets)
ROA
NOL carryforwards-to-assets
Institutional ownership pct.
Log (CEO stock value)
# of CEO options
Interest rate hedging dummy
Log likelihood
Number of observations
FX derivative hedgers
FX non-users

Model FX 1
Tobit Regression
Net FX scaled by assets
Slope
estimate
p-value
-0.0576
-0.2541
0.0270
-0.0002
0.0664
-0.0355
-0.3252
0.0011
0.0117
0.0040
0.0581
-0.0239
0.0236
-0.0004
-0.0044
0.0106
21.08
161
52
109

0.0001
0.9320
0.0102
0.6360
0.0501
0.3650
0.0190
0.6038
0.1977
0.0079
0.0349
0.4282
0.0116
0.4841
0.1864
0.0117

Model FX 2
Probit Regression
1 if Total FX > 0, or zero
Slope
estimate
p-value
-1.0452
17.1910
0.0819
0.0060
1.5574
-0.0951
-3.9843
-0.0623
0.4513
0.0787
0.2930
-0.6591
0.4248
-0.0023
-0.0411
0.2449
-74.50
161
52
109

0.0001
0.7782
0.7468
0.6001
0.0475
0.9047
0.1791
0.4547
0.0238
0.0168
0.6512
0.3328
0.0357
0.8619
0.6189
0.0091

Model FX 3
Truncated Regression
Net FX scaled by assets
Coefficient
estimate
p-value
-0.2194
-58.7080
0.2725
-0.0065
0.2512
-0.6244
-3.7177
0.0275
-0.1422
0.0272
0.8316
1.4476
0.0140
-0.0064
-0.0263
0.0071
114.54
52
52
0

0.0691
0.2854
0.0004
0.3646
0.3854
0.2119
0.0189
0.2175
0.0831
0.0875
0.0010
0.0647
0.8750
0.2551
0.2551
0.8432

Table VII
Simultaneous equations analysis of debt and hedging decisions
Table VII shows the results of structural models linking the extent of derivatives hedging with the debt ratio. The
first column of Panel A (Panel B) shows the results of second-stage Tobit estimation of net IR (FX) derivatives scaled
by book value of assets on predicted values of debt-to-assets ratio, and the predicted debt ratio multiplied by the
market-book ratio. The second column of Panel A (Panel B) shows the results of second-stage OLS estimation of
debt-to-asset ratios on predicted values of net IR (FX) derivatives scaled by book value of assets and other variables
explaining debt levels as suggested by Titman and Wessels (1988) and Graham et al. (1998). The debt policy variables
in the second column are defined in Appendix B. Column 2 p-values are computed with heteroskedastically
consistent standard errors. * means predicted value from an untabulated first stage regression.
Panel A - IR Sample
Column 1
Dependent Variable:
Variable
Intercept
Tax convexity
Debt-to-assets*
Debt ratio* x market/book
R&D-to-assets
Capital expenditures-to-assets
Book/market of equity
Floating rate debt-to-assets
Log(BV of assets)
ROA
NOL carryforwards-to-assets
Institutional ownership pct.
Dual class dummy
Log (CEO stock value)
# of CEO options
SIC 63 dummy
NET_INTp*
ITCp
INTANGp
RD_EXPp
SGAp
LN_SALES
VOL
MTR_BEF
OENEG
PPE_Np
Log likelihood
Number of observations
Non-limit observations
Censored observations
Adjusted R-squared

Net IR scaled by assets
Slope
estimate
p-value
-0.1291
0.0001
-3.8616
0.4288
0.1576
0.0001
0.0021
0.0716
0.1627
0.0585
0.0280
0.3293
0.0063
0.0256
0.0207
0.0007
0.0065
0.0001
-0.1012
0.0141
-0.1944
0.0486
0.0286
0.0068
0.0344
0.0043
0.0024
0.0028
-0.0001
0.9725
0.0121
0.2510

Column 2
Debt-to-assets
OLS
estimate
p-value
0.4455
0.0001

0.3185
-5.6792
0.1821
-0.6714
-0.0973
-0.0149
0.0045
-0.1696
0.4703
0.1494
17.62
238
80
158

N/A
238
N/A
N/A
0.4691

0.0065
0.0786
0.0190
0.0248
0.1452
0.0087
0.0272
0.2572
0.0001
0.0004

Table VII (continued)
Panel B - FX Sample
Column 1
Dependent Variable:
Variable
Intercept
Tax convexity
Debt-to-assets*
Debt ratio* x market/book
R&D-to-assets
Capital expenditures-to-assets
Book/market of equity
Dividend yield
Foreign sales-to-total sales
Log(BV of assets)
ROA
NOL carryforwards-to-assets
Institutional ownership pct.
Log (CEO stock value)
# of CEO options
Interest rate hedging dummy
NET_CURp*
ITCp
INTANGp
RD_EXPp
SGAp
LN_SALES
VOL
MTR_BEF
OENEG
PPE_Np
Interest rate hedging dummy
Log likelihood
Number of observations
Non-limit observations
Censored observations
Adjusted R-squared

Net FX scaled by assets
Slope
estimate
p-value
-0.0722
0.0001
-2.8638
0.5301
0.0986
0.0001
0.0012
0.0877
0.1964
0.0001
-0.0028
0.9524
0.0034
0.0520
-0.0950
0.5107
0.0091
0.2766
0.0024
0.1232
0.0240
0.3743
-0.0313
0.3040
0.0282
0.0045
-0.0002
0.6921
-0.0066
0.0814
0.0046
0.2599

Column 2
Debt-to-assets
OLS
estimate
p-value
0.1252
0.2060

0.7977
4.9576
0.3477
-0.9506
0.0231
-0.0027
0.0096
0.0920
0.5386
0.2927
0.0386
26.87
128
38
90

N/A
128
N/A
N/A
0.5248

0.0062
0.4921
0.0013
0.0064
0.7916
0.8048
0.0089
0.5493
0.0001
0.0005
0.2393

Table VIII
Quantifying the tax advantage of hedging with derivatives
Table VIII summarizes tax savings associated with derivatives hedging. The first column of Panel A (Panel B) shows
the distribution of the product of firm-specific values of net IR (FX) derivatives scaled by book value of assets and
the OLS coefficient on the derivatives variable from the second-stage debt level regression (the coefficient taken from
column 2, Table 7). This product represents the increase in the long-term debt ratio that is attributable to derivatives
hedging. The numbers in the second column are the products of values from the first column and marginal tax rates,
where DTB stands for debt tax benefit. The third column scales the values in the second column by the market value
of the firm. The rightmost column shows the distribution of DSAVE5 (the expected dollar reduction in expected tax
liabilities resulting from a 5% reduction in the volatility of sales revenue) discounted as a perpetuity (using a 10%
discount rate) scaled by market value. Comparing the two rightmost columns, the tax benefit of hedging to increase
debt capacity is approximately 10 times as large as the tax benefit associated with tax function convexity.

Panel A - IR Sample

N
Mean
Median
Std Dev
99th percentile
95th percentile
90th percentile
75th percentile
25th percentile
10th percentile
5th percentile
1st percentile

Portion of debt ratio
due to IR hedging
83
2.85%
2.22%
3.16%
19.58%
8.09%
6.17%
3.60%
0.78%
0.30%
0.04%
0.00%

Dollars of DTB due
to IR hedging (mil)
83
$81.31
$11.70
$190.95
$996.46
$491.64
$213.34
$51.71
$4.31
$2.05
$0.98
$0

Contribution of DTB
to market value
82
1.38%
0.73%
1.99%
11.86%
4.74%
3.40%
1.42%
0.25%
0.08%
0.04%
0.00%

Contribution of convexity
to market value
82
0.153%
-0.010%
0.372%
1.656%
1.102%
0.533%
0.183%
-0.050%
-0.080%
-0.090%
-0.150%

Portion of debt ratio
due to FX hedging
38
4.52%
3.04%
5.30%
28.82%
17.92%
8.79%
5.80%
1.50%
0.43%
0.34%
0.33%

Dollars of DTB due
to FX hedging (mil)
38
$63.30
$17.69
$105.29
$448.11
$416.37
$158.72
$83.81
$4.22
$0.63
$0.42
$0

Contribution of DTB
to market value
38
1.36%
0.97%
1.37%
6.07%
4.79%
3.65%
1.90%
0.33%
0.12%
0.03%
0.00%

Contribution of convexity
to market value
38
0.162%
0.011%
0.418%
1.656%
1.603%
0.533%
0.192%
-0.050%
-0.070%
-0.110%
-0.470%

Panel B - FX Sample

N
Mean
Median
Std Dev
99th percentile
95th percentile
90th percentile
75th percentile
25th percentile
10th percentile
5th percentile
1st percentile