Firm Value and Corporate Hedging: A “Catch Me if You Can” Game
Natalia Golub*1
Master thesis in Finance
Faculty of Economics and Business, University of Groningen
Supervisor: Prof. Dr. R. E. Wessels
June 26, 2014
13,845 words
Abstract
This study examines derivatives use in a sample of 74 nonfinancial randomly selected S&P 500 firms. I examine the use of derivatives to differentiate among existing theories of hedging behavior and estimate a simultaneous equations model to investigate the determinants of how firm-level hedging influences firm value. The evidence is consistent with firms hedging to avoid expected financial distress costs and increase debt capacity. I also find that derivatives hedging and corporate diversification are rather complements than substitutes. The overall results suggest that there is a negative value effect induced by derivatives use.
JEL classification: G3, G32
Keywords: Derivatives, Hedging, Firm value, Financial distress
2
1. Introduction
The use of derivative securities for risk management purposes has been a worldwide phenomenon for several decades. The purpose of the firm’s risk management practices, in addition to compliance with legal minimum requirements, is to protect the firm’s human and physical resources, including its revenues, against potential loss and minimize the adverse effects should a loss occur.
In a competitive financial environment, firms widely use derivative instruments to alleviate a variety of risks associated with fluctuations in interest rates, currency and commodity prices. This gives rise to a highly debated topic in corporate finance of whether active risk management policies, such as corporate hedging, affect firm value.
Given that the use of derivative contracts increases so rapidly, practitioners and academics need to know how and why firms use these relatively new financial instruments.
It is not that there is a lack of theories to describe why firms might wish to hedge. Two strands of literature attempt to explain the motives of risk management. Firm value maximization theories (see, e.g., Mayers and Smith, 1982; Smith and Stulz, 1985; Nance et al., 1993; Geczy et al., 1997; Leland, 1998) indicate that firms hedge to reduce certain costs or capital market imperfections related to volatile cash-flows. The straightforward reason is that hedging activity can be motivated by tax incentives, reducing the deadweight costs of financial distress and increasing the firm’s debt capacity, generating greater tax advantages. Also, it increases value, when firms face convex costs, as progressive taxation and bankruptcy costs.
Also, when a firm has a lot of growth opportunities and external financing is more expensive compared to the funds generated within the company, hedging might also mitigate the potential underinvestment problem (Smith et al., 1990; Froot et al., 1993), which arises when the investment opportunities are negatively correlated with the cash flows.
From a different vantage point, manager’s utility maximization theory suggests that managers decide to hedge to maximize their personal utility functions. Corporate volatility can be costly if variability of managers’ compensation is related to the volatility of corporate income. Risk-averse managers engage in hedging if they cannot effectively hedge corporate volatility in their personal accounts or when the cost of hedging at the firm level is lower than doing it on their own account (Stulz, 1984; Smith and Stulz, 1985). Hence, it can be value maximizing for a firm to hedge if this reduces the risk premium and compensation that managers demand.
3 1997; Howton and Perfect, 1998; Hagelin, 2003; Allayannis and Ofek, 2000; Graham and Rogers, 2002; Haushalter, 2000).
Until recently, the overall effect of corporate risk management on firm value has not been researched thoroughly. Thus far, the reported evidence is broadly consistent with the notion that hedging activity increases value (see, e.g., Allayannis and Weston, 2001; Graham and Rogers, 2002; Hagelin, 2003; Nelson et al., 2005; Adam and Fernando, 2006; Carter et al., 2006; Davies et al., 2006; Mackay and Moeller, 2007; Bartram et al., 2009). But then, Nguyen and Faff (2007; 2010) reveal in direct contrast to their prediction, that a “discount” is imposed on firm value of derivative users.
Withal, some studies report inconsistent evidence. For instance, Guay and Kothari (2003), Clark et al. (2006), Bartam et al. (2007), and Khediri and Folus (2010) conclude that derivatives use is too small to have a discernible impact on firm value.
This study contributes to the previous research by not only examining firm characteristics associated with the use of derivatives, but also identifying whether hedging policy has an impact on firm value. More specifically, it analyzes whether the extent of hedging activity contributes to value creation. I examine these issues by investigating a sample of 74 large, nonfinancial, randomly selected US firms which are included in S&P 500 Index and hope to provide an adequate empirical answer to this question.
The results reveal somewhat unexpected outcome that the analyzed hedging theories have little predictive power in explaining corporate risk management decisions. The evidence based on multivariate empirical relations between the extent of hedging and firm value maximization theories, fails to provide any support for any of the tested relations, except for costs of financial distress and debt capacity, which are measured by leverage. Nevertheless, the main finding, contrary to the formulated prediction that derivatives use is associated with a “hedging premium” suggests that corporate hedging decreases firm value.
The outline of the paper is as follows: Section 2 reviews the theories of corporate risk management. Section 3 describes sample selection and hypotheses development. In Section 4, I report descriptive statistics on the economic characteristics of sample firms and their derivatives positions, as well as the main results. Section 5 concludes.
2. Review of the literature
4
2.1 Managerial wealth and risk aversion
Stulz (1984) claims that one reason for corporate hedging is the managers’ risk aversion. The ability of outside stockholders to diversify, determines them to be indifferent to the amount of hedging undertaken, however, this does not hold for managers that may hold a large portion of their wealth in the firm’s stock. Hence, by reducing the variance of the total firm, managers are considerably better off, as hedging can potentially reduce the required risk premium. However, managers that invest greater proportions of their wealth in firm’s shares engage more in hedging activity, while those with greater option holdings prefer less to hedge because options provide convex payoffs, whereas stocks provide linear payoffs.
A weakness of Stulz’s theory consists of the assumption that managers face significant costs, when they trade on their own account. Hence, they have to involve the company in any hedging activity, to adjust the risks they face. The risk aversion of managers induces them to engage in hedging due to the fact that their compensation is a function of firm value. Also, Smith and Stulz (1985) argue that shareholders can affect managers’ risk aversion through the design of compensation contracts, in the attempt to counteract the effects of risk aversion because management’s utility function is concave in expected wealth.
Compensation can be structured as a linear, concave, or convex function of firm value. If the contract is linear or concave in firm value, the manager has utility maximizing incentives to reduce risk. Contrariwise, a convex compensation contract indicates that utility is less concave in firm value. Hence, the extent of convexity determines the manager to be risk-averse, risk-neutral, or even risk-seeking.
In contrast to the above view, Breeden and Viswanathan (1990), and DeMarzo and Duffie (1995) put forward a very different managerial theory of hedging. Their models are based on asymmetric information and claim that labor market overhauls its view regarding the managers’ abilities, based on their firm’s performance (Breeden and Viswanathan, 1990). This can determine some managers to undertake hedging to influence the perceptions of the labor market.
The work of DeMarzo and Duffie (1995) cited earlier, shows that firms should hedge based on private information that cannot be easily communicated to shareholders without any costs.
5 Brown (2001) argues that managers of large firms with geographically diversified operations decide to hedge due to the agency and contracting cost considerations for the sake of removing the noise from their divisional performance.
Also, Oosterfhof (2001) links shareholder value maximization with managerial utility maximization. Still, if a manager owns a lot of employee stock options (ESOs) on his own firm, he will be inclined to increase the volatility of firm’s stocks. According to Black-Scholes model, this ceteris paribus, increases the value of his employee stock options. But the fixed wage part is in principle not affected by risk shifts of the firm’s activities. This can lead managers to take too much risk which decreases shareholder value but increases the value of the managers’ total wage package which partly consists of ESOs.
2.2 Tax function convexity
Several theoretical studies argue that it is optimal for firms to hedge if their taxes form a convex function of earnings, because hedging will decrease the expected taxes (Smith and Stulz, 1985). The reason is that more volatile earnings stream leads to higher expected taxes, due to convexity. Such being the case, the convexity of a tax function is entirely plausible for firms that have a high probability of negative earnings and are not able to carry completely over their tax losses to the next periods.
Graham and Rogers (2002) argue that tax code characteristics enable corporations to potentially increase their value by the decision to hedge.
Expected Taxes
Expected taxes
without hedging
Expected taxes Convex with hedging tax schedule
PTI1 PTI3 PTI2 Taxable Income Figure 1. Taxes
6
2.3 Costs of Financial Distress and Debt Capacity
For a given level of debt, hedging can reduce the probability that a firm will encounter financial distress. Hence, firms can decide to hedge, to increase their debt capacity, if financial distress is costly besides the advantage to have debt in capital structure, due to agency problems or taxes associated with “free cash flow”.
Jensen and Meckling (1976) highlight the relative importance of indirect costs of financial distress in the selection of the financial structure. A more sophisticated version of this argument, introduced by Smith and Stulz (1985), relies on the assumption that bankruptcy involves some exogenous transactions costs. Hedging can increase the value of the firm by narrowing the distribution of firm-value outcomes, and thereby the expected costs of financial distress. Consequently, hedging probability is expected to increase with increases in the probability of financial distress or costs associated with it.
Nance et al. (1993) suggest that as leverage increases, the probability of a firm to find itself in a situation where it is unable to repay its debt, also increases. This probability can be reduced by hedging, as well as by maintaining more liquid assets or dividend yields. Froot et al. (1993) argue that hedging can be value enhancing when external financing is more costly than internal financing, while matching fund outflows with inflows, therefore lowering the probability of a firm accessing the capital markets.
Frequency distribution
Probability of bankruptcy and financial distress
0 FPO E1(V) E2(V) Firm value
Figure 2. Bankruptcy and financial distress
7 Leland (1998) indicate that decreasing the volatility of cash-flows through hedging can increase the debt capacity and generate higher tax benefits.
In addition, Fig. 2 expresses a firm’s cash flows distribution that it is not able to pay off the fixed claimholders and employees because the cash flows are not sufficient to honor its fixed payment obligations (FPO). Therefore, the firm incurs bankruptcy costs due to the fact that the probability of falling below this point is positive. However, if the hedging activity ensures that cash flows are below FPO in fewer states of nature, then firm value increases (E1(V) to E2(V)).
2.4 Capital Market Imperfections and Inefficient Investment
A more innovative version of the discussion invokes Myers’s (1977) “debt overhang” underinvestment effect to endogenize the costs of financial distress. Moreover, he points out that firms’ equityholders tend to underinvest when there is issuance of claims with higher priority than equity and proves that financial distress can provide shareholders with incentives to forgo positive net present value projects if the gains accrue primarily to fixed claimholders. Therefore, hedging firm value lowers the probability of distress and that shareholders will find it useful to pass up valuable projects. However, Bessembinder (1991) suggests that employing derivatives for hedging reduces the impulse for underinvestment, as hedging shifts the individual states from default to nondefault outcomes. All in all, higher leveraged firms with more valuable growth opportunities are more likely to be affected by underinvestment, as well as to employ hedging.
USD in State ST
Value of payoff of investment project
Fixed payment obligations + Initial Investment
Initial Investment
ST0 ST1 States of the World Figure 3. The underinvestment problem
8 hedging are close to those expressed in Lessard (1990), as well as Froot et al. (1993). In the interest of brevity, Froot et al. (1993) argue that hedging is valuable if it can allow a firm to prevent needless fluctuations in any investment spending of resources raised from outside investors.
Although the actual mechanism is somewhat different, all the concepts depend on the basic observation that, without hedging, firms can be compelled to underinvest in some parts of the world because it is expensive or difficult to raise external funds.
Moreover, Fig. 3 shows an example with infinite states of nature. When there are no agency conflicts, it is optimal for a firm to invest, if the gains are higher than the initial outlay, this happening when ST > ST0. In the presence of agency conflicts, managers act in stockholders’ interests, whenever the gains exceed initial investment plus fixed payment obligations (ST > ST1). This suggests that underinvestment occurs between ST0 and ST1. However, the firm can stabilize its cash flows by engaging in hedging activity, thus ensuring that gains are less often below initial outlay plus fixed obligations. Therefore, positive NPV are accepted and firm value is increased.
Supply of annual USD in State ST cash flows
Hedging
Hedging Demand for investment funds
Declining business Stale business Improving business conditions conditions conditions
(a)
Supply of annual USD in State ST cash flows
Hedging Demand for investment funds Hedging
Declining business Stale business Improving business conditions conditions conditions
(b)
9 Furthermore, Fig. 4 (a) shows the gap between the demand for investment funds and the supply of internal cash flows. In this scenario, the decision to hedge can align internal generated funds with investment needs in the following way: by reducing the surplus when cash flows exceed investment expenditures (arrow pointing downwards) and providing cash when these are less than the required investment (arrow pointing upwards).
Nevertheless, firms can decide not to hedge, if the internal generated cash flows match the need for investment funds. Hence, Fig. 4 (b) illustrates that the available cash flows partially match the necessary funds for investment. In this case, the line is not flat anymore.
2.5 Other factors and substitutes for hedging
One of the main arguments put forward by Nance et al. (1993) is that conflict between stockholders and fixed claimholders can be reduced not only by hedging with derivative instruments. More specifically, the agency conflict can be reduced by issuing convertible bonds or preference capital, as opposed to straight debt. Unlike derivatives that reduce the variance of the firms’ net cash flows, convertible debt includes an embedded option on firms’ assets, making this liability more sensitive to changes in firm value, thus reducing the equity’s sensitivity to firm value changes.
Notwithstanding this, the preference capital reduces the probability of encountering financial distress costs. Even though they are similar to debt, preferred shares pay dividends periodically instead of paying interest. Therefore, while preference capital does not produce tax shields, firms can postpone the dividend payment without any threat of insolvency. On the other hand, if a firm does not meet its interest payment obligations, a bankruptcy filing cannot be avoided.
Furthermore, a firm’s need to hedge is also determined by its dividend policy. A lower dividend payout makes it more likely that there will be funds available to pay the bondholders, and therefore reduce the agency conflict.
Another hedging substitute is liquidity. Some studies claim that firms with higher levels of liquidity are less likely to engage in hedging activity, since they have larger financial buffers and funds will be available to pay the debt claims. By the same token, firms with more liquid assets will have less need to access costly external financing to fund their investment projects.
2.6 Overview of the existing empirical evidence
10 For instance, Warner (1977) finds that direct costs of financial distress are less than proportional to firm size, indicating that small firms are more likely to hedge than larger firms. Nevertheless, the empirical evidence indicates a positive relation between hedging and firm size, which conversely suggests that hedging activities are related to high fixed costs.
Withal, Nance et al. (1993) conclude in their study of 169 firms, as well as Géczy et al. (1997), that hedging companies face greater growth opportunities, which is substantial with the claim that hedging helps to relieve the underinvestment problem, employing fewer hedging substitutes.
In testing the bankruptcy and financial distress hypothesis, empirical studies mainly focus on leverage and profitability. Dolde (1995) and Haushalter (2000) distinguish a positive and significant relation between hedging and leverage, which is consistent with the theory that hedging reduces the probability of encountering financial distress costs.
Tufano (1996) does not find any support for value maximization theory, in a sample of gold mining firms, but obtains strong evidence for managerial risk-aversion theory. This claims that managers holding more stocks, undertake more hedging activities, but this argument has not been confirmed in broader-based samples. He relies on the assumption that management’s stock and option holdings are exogenous in testing the determinants of risk management. The result shows that option holdings are negatively related to hedging. Also, Fok et al. (1997) find empirically that nonhedged firms have larger managerial ownership than hedged firms. Other empirical studies of hedging fail to find any significant evidence of managerial motives in determining the firms to hedge, for instance, Geczy et al. (1997), Gay and Nam (1999), Haushalter (2000), Allayannis and Ofek (2001), and Judge (2003).
Empirical studies find little support for the tax argument, which is often tested by observing the availability of tax-loss carry-forwards or by calculating the potential tax savings from hedging. For instance, Gay and Nam (1999) find that incentives to hedge for tax purposes are small compared to other determinants. Graham and Rogers (2002) provide evidence that firms hedge to increase their debt capacity but not because of the tax convexity, along with the fact that on average, the debt capacity induced by derivatives, increases firm value by 1.1%.
According to Bodnar et al. (2003), institutional differences matter with respect to corporate risk management practices, claiming that Dutch firms hedge on average more financial risk than U.S. corporations, due to the fact that Dutch firms experience far more foreign exchange exposure than U.S. firms.
Furthermore, Judge (2003) finds low support for the importance of taxes or the presence of bankruptcy costs to determine the use of derivatives.
11 supplements U.S.-based literature on corporate hedging, highlighting that internal governance mechanisms can also play a role in derivatives usage, this role being influenced by regulatory environment.
In contrast to the previous theoretical models, Hahnenstein and Roder (2007) design a model which relies on firms’ hedging decision with endogeneous leverage and prove that optimally leveraged firms choose higher hedge ratios when facing convex indirect bankruptcy cost functions.
Despite the fact that diversification is a form of hedging, Lim and Wang (2007) prove that financial hedging and corporate diversification are more often complements than substitutes, concluding that firms’ hedging behavior can increase their incentives to manage risk through diversification.
Dolde and Mishra (2007), using a data set comprised of U.S. firms with sales exceeding 1 billion USD, show that complexity, managerial options ownership, financial distress, and primitive risk relate to hedging behavior. However, their estimates do not support the hypothesis that underinvestment and scale economies determine the firms to hedge.
Moreover, after performing an in-depth analysis of the foreign exchange rate exposure of large nonfinancial firms, Bartram (2008) finds that managers of these firms, whose operations are exposed to foreign currency risk, take savvy actions to reduce exposure to a level that is too low to be empirically detected.
Meanwhile, Purnanandam (2008), using a comprehensive data set, obtains a nonmonotonic relation between leverage and hedging, where the effect of leverage on hedging activity is higher for firms in more concentrated industries.
Unlike the previous empirical models, Lin et al. (2008) investigate the interaction between hedging, financing, and investment decisions. Their evidence indicates that as firms become more efficient by investing in more risky assets, they borrow less, invest more in those assets, as well as decide to engage more in hedging activities. The results also suggest that there is a positive relation between hedging and leverage, which is consistent with debt capacity arguments.
Due to the mixed results from previous research regarding the effect of financial distress costs, Clark and Judge (2008) use UK data to present strong empirical evidence. They argue that using leverage as a proxy for financial distress without distinguishing between foreign debt users and nonusers causes misleading results. Specifically, after dividing their sample of hedgers into firms that use and do not use foreign debt, they prove that leverage significantly relates to hedging activity but for those firms that use only foreign debt or together with derivative instruments, and not for firms that just hedge, highlighting the fact that foreign debt users influence these results.
12 From a different perspective, Acharya and Bisin (2009) examine managers’ ability to hedge their exposure to firm cash flows and show that risk-substitution moral hazard increases aggregate risk in stock markets, reducing investors’ ability to share risks.
More recently, researchers have been examining the direct relation between hedging and firm value. Nance et al. (1993), Géczy et al.(1997), Allayannis and Weston (2001), Graham and Rogers (2002), Hagelin (2003), Nelson et al. (2005), Adam and Fernando (2006), Carter et al. (2006), Davies et al. (2006), Mackay and Moeller (2007), and Bartram et al. (2009) report results in support of the value enhancing hypothesis.
Allayannis and Weston (2001), in their analysis of US nonfinancial firms with exposure to exchange rates, obtain the evidence that there is a positive relation between the use of foreign currency derivatives and Tobin’s Q, concluding that currency hedging improves firm value.
Guay and Kothari (2003) argue that many sample firms that were used in prior studies of hedging, may be unable to capture the benefits from holding currency or interest rate derivatives. They analyze the economic effects of derivatives positions for 234 large nonfinancial corporations and claim that the potential gains on derivatives are relatively small compared to cash flows or fluctuations in equity values, so that the registered increase in market values can be determined by operational hedge that increases value and is positively correlated with derivatives positions. Also, they question the validity of the results obtained by Graham and Rogers (2002).
The main finding of Pramborg (2004) when using a sample of Swedish firms from 1997 until 2001 is a positive value effect from hedging exposure, but still translation exposure hedging does not add value. Similarly, Bartram et al. (2009) study 7,263 nonfinancial companies from 48 countries and finds a positive valuation effect for those firms that use interest rate derivatives, as well as Hagelin et al. (2004) in their research of Swedish firms, obtain results confirming that hedging activity leads to an increase in firm value.
In the same vein, Adam and Fernando (2006) investigate the hedging activity of gold mining firms. Their results indicate that these firms realize significant cash flow gains from derivatives transaction, which in turn increases shareholder value.
13 effect discovered by Allayannis and Weston (2001), in their cross-sectional analysis, is hard to interpret due to the hurdle in adequately controlling for endogeneity of value and hedging.
Meanwhile, Nguyen and Faff (2007) analyze the use of foreign currency and interest rate derivatives by Australian firms and report a negative relation between hedging and firm value. Their results are similar to those obtained by Bartam et al. (2007) that examine a large sample of 6,888 firms and hardly find any relation between firm value and the extent of hedging, as well as Khediri and Folus (2010) in their analysis of 320 French firms do not find any support that hedging activity adds value to the firm.
Mackay and Moeller (2007) being inspired by Smith and Stulz (1985) that use Jensen’s Inequality to show that cash flow volatility has to be managed if the firm faces convex financial costs, conclude for a sample of 34 oil refiners, that hedging can add value if revenues are concave in product prices or costs are convex in factor prices.
Moreover, there is some evidence that the value impact of corporate risk management can depend on corporate governance structures. For instance, Allayannis et al. (2012) examine the impact of currency derivatives on firm value, by using a sample of firms from 39 countries. They argue that well-governed firms are more likely to use derivatives for hedging rather than speculate or pursue managers' own interests. Their results suggest that derivative users with strong internal firm-level or external country-level governance are associated with a significant value premium. The same comprehensive evidence is obtained by Lel (2012) that examines a sample of firms from 30 countries over the period 1990 to 1999. The results show that strongly governed firms appear to employ derivatives to hedge their foreign currency risk and overcome costly external financing. However, weakly governed firms tend to hedge mostly for managerial reasons.
Nevertheless, almost all empirical work in this area faces some significant challenges that need to be considered when interpreting the results. For instance, the empirical modeling of structural equations, identification of appropriate proxy variables to measure the extent of hedging, and endogeneity, are major issues that many studies fail to address, therefore limiting the conclusions to be drawn from their results.
3. Sample selection and hypotheses development
14 Statement of Financial Accounting Standards (SFAS) 105 requires firms to report information on financial instruments with off-balance sheet risk. Mainly, firms must report the face, contract, or notional amount of the financial instrument, and information on the credit or market risk of those contracts. In addition, SFAS 133 requires that all derivatives instruments must be shown on balance sheet at fair market value with accounting for changes in fair value depending on the purpose of derivative.
I gather detailed data regarding the fair values and notional amounts of derivative holdings from financial statement footnotes, from 10-K fillings in the SEC’s EDGAR database, along with a dichotomous measure of derivatives use that equals to one if the firm engages in derivatives hedging and zero, otherwise. The notional value is the face or principal value upon which the performance of a derivative contract is based. In contrast, the fair value gives the net gain or loss on the derivative contracts outstanding at balance date.
Afterwards, I exclude firms that employ derivatives for hedging but do not report their fair value or notional amounts. Since the financial statements are audited, the most likely reason for nondisclosure is that the amounts are immaterial. Also, I utilize only derivatives disclosed as being used for nontrading purposes, as this distinction is required by SFAS 119, and those that are designed as hedges for accounting purposes. The key purpose of hedge accounting is to reduce earnings volatility by recording in earnings in the same time period a gain or loss on a hedged item and the loss or gain on the related hedge. With hedge accounting, whether a gain or loss on the hedged item is recognized currently or deferred in accumulated other comprehensive income, an offsetting loss or gain on the hedge is recognized in a similar way. Furthermore, none of the sample firms state that they speculate.
For each disclosure, I attempt to classify derivative holdings as long or short. The net fair value is computed as the absolute value of the net gain or loss on derivative positions that are marked to market (Mardsen and Prevost, 2005). The total notional value is calculated by summing the long and short positions. Using the dollar notional amount of derivatives has several advantages over using a single measure like a binary variable to indicate whether or not firms hedge. For instance, by employing this continuous measure, I can test hypotheses regarding the determinants of the amount of hedging and investigate its impact on firm value.
15 I compute the three-year average for each of the financial variables to be used further in the analysis, due to the fact that accounting data does not vary significantly over time and to prevent the assumption of independent errors in the regressions from being violated. The resulting sample consists of 74 large nonfinancial firms.
As a dependent variable for firm value I use the natural logarithm of Tobin’s Q ratio, defined as the market value of assets over the book value of assets. The firm market value is calculated as the book value of total assets plus the market value of the equity minus the book value of equity
Q =BV total assets −BV common equity +MV common equity
BV total asets (1)
In testing whether the hedging activity increases the value of the firm or not, it is necessary to exclude the effect of all other variables that could have an impact on firm value. The effects from these factors must be controlled for before inference whether hedging affects value can be drawn. Below, I describe the control variables from multivariate regressions and the reasons that determine to use them.
Firm size. The previous empirical evidence is controversial, that is why it is used as an exogenous
variable in this analysis, especially because larger firms are more likely to hedge than small firms, because of the existence of large fixed start-up costs of hedging (Bodnar et al., 1997; Hagelin, 2003; and Pramborg, 2002). In addition, Allayannis and Weston (2001) find differences in Tobin’s Q for large firms as compared to small firms, where large firms are associated with lower Tobin’s Q.
Profitability. A profitable firm is likely to trade at a premium relative to a less profitable one. Thus
if hedgers are more profitable, they will have higher Tobin’s Qs.
Investment opportunities. Myers (1977) and Smith and Watts (1992) argue that firm value also
depends on future investment opportunities. Because hedgers are more likely to have larger investment opportunities, such control is important.
Access to financial markets. Allayannis and Weston (2001) claim that if hedgers forgo projects
because they are not able to obtain the necessary financing, their Tobin’s Q can remain high because they undertake only positive net present value (NPV) projects. If a firm pays a dividend, it is less likely to be capital constrained and may thus have a lower Q. Fama and French (1998) conclude that dividends convey information about future profitability (expected net cash flows).
Leverage. A firm's capital structure may also be related to its value. Fama and French (1998) and
Allayannis and Weston (2001) find evidence of a negative relation between leverage and Tobin’s Q.
Geographic diversification. Several theories suggest that geographic diversification increases
16 multinationality is positively related to firm value. However, Lang and Stulz (1994) emphasize that diversified firms receive a discount in valuation.
Liquidity/Financial distress. Liquidity affects the discrepancy between the book and market value
of an asset. Lower liquidity means lower value. The ability to turn assets into cash to cover the current liabilities is crucial for the survival of a firm in economic hardship. All things being equal, a firm with more liquid assets is worth more than a firm with less liquid assets, because of the risk associated with being able to raise capital in times of distress. In contrast, firms that are cash constrained can have a higher Tobin’s Q because they are more likely to invest in positive NPV projects. This results from the “free cash flow” argument of Jensen (1986) that firms with excess free cash flow are more likely to invest in projects with negative NPV.
Financial stability. Firms that face financial distress, bear bankruptcy costs, which in turn,
decrease the value of the firm. I employ a discriminant analysis to predict the likelihood of a firm’s bankruptcy, based on Altman’s Z-score (Altman, 1983). A score below 1.8 means that the company is probably headed for bankruptcy, while companies with scores above 3.0 are not likely to go bankrupt. This implies that the higher the score, the lower the probability to encounter financial distress.
Table 1
Hypothesized relation between the extent of hedging and firm value, considering the impact of other control variables Independent variables Sign Data description (Abbreviation) Hedging
variable
Aggregated (binary variable, net fair value and gross notional amount)
+ Principal component analysis (HV)
Other control variables
Firm size +/- Natural logarithm of total assets (FSZ) Profitability + Return on assets, ROA % (P)
Investment opportunities + Capital expenditure expressed as percentage of total assets (IO)
Access to financial markets
+/- A dummy variable that equals to one if the firm pays dividends on common equity in the current year and zero, otherwise.
Leverage - Book value of long-term debt over the market value of common equity, % (L)
Geographic diversification +/- Dummy variable that is set to one when firms expand their activities in other countries and zero, otherwise (GD)
Financial distress - Dividend yield at close, annual and calculated as dividend per share, divided by the market price - year end, % (FD)
Financial stability + Computed by using Altman’s Z-score formula (2) (FST)
Altman’s Z-score is calculated as
Z-Score = 1.2×WC/TA + 1.4×RE/TA + 3.3×EBIT/TA + 0.6×MVE/TL + 1.0×Sales/TA (2) where:
17 TA – total assets;
RE – retained earnings;
EBIT – earnings before interest and taxes; MVE – market value of equity;
TL – total liabilities.
Therefore, Table 1 presents a summary of the predicted direction for potential relation between hedging and firm value, considering also, other control variables.
Moreover, Table 2 summarizes the hypotheses regarding what determines the firms to hedge and shows the expected sign and proxy variables for each of the firm characteristics used in the regression where the dependent variable is the extent of hedging.
Table 2
Hypothesized relations between firm characteristics and the extent of risk management activities
Hypothesis Independent Variables Sign Data description (Abbreviation)
Financial distress
Leverage + Book value of long-term debt over the market value of common equity, % (L)
Altman Z-score - Computed by using Altman’s formula (2) (FST)
Dividend yield + Dividend yield at close, annual and calculated as dividend per share, divided by the market price at year-end, % (FD)
Underinvestment costs
Market-to-book + Calculated by dividing the annual closing price of the stock by the book value per share (PB)
R&D expenses + Research and development expenses expressed as percentage of sales (RD) Growth + Expressed as percentage of total assets growth (G)
Tobin’s Q + Natural logarithm of market value of assets over the book value of assets. The firm market value is calculated as the book value of total assets plus the market value of the equity minus the book value of equity (Q)
Substitutes to hedging with derivatives
Convertible debt - Book value of convertible debt scaled by market value of equity, % (CVD) Preferred stocks - Book value of preferred stocks scaled by market value of equity, % (PS) Quick ratio - Cash and cash equivalents plus net receivables, divided by total current
liabilities (QR) Exposure and
variation
Foreign assets + Foreign assets expressed as percentage of total assets (FA)
Firm size Market value of equity +/- Natural logarithm of market value of equity (MVE) Other Geographic
diversification
+/- Dummy variable that is set to one when companies expand their activities in other countries and zero, otherwise (GD)
4. Empirical Evidence
4.1 Descriptive statistics and univariate analysis
Table 3 presents summary statistics for proxy variables described in Table 1 from Section 3. Both, gross notional amounts and net fair values are scaled by the market value of equity. Only the scaled variables are used further in the research.
18 Table 3
Summary statistics for derivatives and control variables
Mean Std. Dev. Minimum Median Maximum
Panel A: Derivative variables
Binary derivatives variable 0.61 0.49 0.00 1.00 1.00
Net fair value (mln USD) 48 101 0.00 7.8 573
Scaled by MVE 2.32 7.22 0.00 0.54 51.66
Gross notional amount (mln USD) 1,379 2,299 0.00 295 11,333
Scaled by MVE 56.12 98.05 0.00 17.19 640.03
Panel B: Control variables
Firm size 16.55 0.88 14.33 16.60 18.47 Profitability, % 9.08 4.79 -6.95 8.61 21.19 Capital expenditure, % 4.75 4.77 0.50 3.46 30.85 Dividend dummy 0.85 0.36 0.00 1.00 1.00 Leverage, % 24.21 12.79 0.11 21.67 65.43 Geographic diversification 0.84 0.37 0.00 1.00 1.00 Dividend yield, % 2.00 1.63 0.00 1.96 10.97 Altman Z-score 3.94 2.08 -0.51 3.46 9.24
Panel A presents the descriptive statistics for 74 firms that report derivatives use for hedging purposes. The binary derivatives variable is a dummy variable that equals to one if the firm hedges and zero, otherwise. The net fair value is computed as the absolute value of the net gain or loss on derivative positions that are marked to market. The gross notional value is calculated by summing the long and short positions. Panel B reports descriptive statistics for control variables used in the structural model where firm value is specified as a dependent variable. Firm size is measured by the natural logarithm of total assets. Profitability is the return on assets. Capital expenditure is expressed as a percentage of total assets. Dividend dummy is a variable that equals to one if the firm pays dividends and zero, otherwise. Leverage is computed as the book value of long-term debt over the market value of common equity. Geographic diversification is expressed as a binary variable that is set to one when companies expand their activities in other countries and zero, otherwise. Dividend yield is at close, annual and calculated as dividend per share divided by the market price at year-end. Altman’s Z-score is computed by using Formula (2).
Table 4 presents summary statistics for proxy variables described in Table 2 from Section 3, and tests for differences between the means and medians of these variables for users and nonusers of derivatives. Approximately 61% of the sample firms are classified as derivative hedgers. A Wilcoxon test for difference in medians identifies a considerable number of differences between the firms classified as hedging and those that do not hedge. Also, Table 4 reports the mean differences between hedged and nonhedged firms, as well as the p-values associated with them.
19 Hedged firms have higher leverage than nonhedged firms, the difference being significant. Thus, there is a strong support for the hypothesis that firms hedge to reduce the transaction costs of financial distress. Also, firms that employ derivatives are more diversified than those that choose not to hedge, suggesting that hedging with derivatives and corporate diversification can be complements rather than substitutes (Lim and Wang, 2007).
Table 4
Summary of financial characteristics of 74 randomly selected S&P 500 Firms, conditioned on the financial risk management employed Hedgers (N = 45) Nonhedgers (N = 29) (H) – (NH) t-statistic (p-value) (H) – (NH) Z-statistic (p-value) Mean Median Std.
20 The percentage of foreign assets owned by hedgers is approximately three times higher than the percentage for nonhedgers, which is consistent with the argument that firms which conduct businesses in other countries, and hold huge amounts of foreign capital, are more likely to hedge.
Last but not least, examination of firm size reveals that hedgers are significantly larger as compared to nonhedgers. This evidence suggests that economies of scale in hedging dominate other considerations with respect to firm size.
Table 5 summarizes correlations between the independent variables and aggregated hedging variable. The smallest correlation of 0.19 is between the market value of equity and foreign assets, and the largest of 0.78 is between market value of equity and firm size. However, these two proxies are used in different structural models. Only four correlations have an absolute value larger than 0.5. This suggests that multicollinearity is present in the data, but that it is not severe.
Firm size, as well as geographic diversification is positively related to the extent of hedging. Leverage is positively correlated with the hedging variable which is consistent with the financial distress costs hypothesis. Specifically, firms with important gearing in their capital structure are more likely to face financial distress and if it is costly, firms are better off with hedging activity because it reduces its probability. Also, foreign assets are positively related to hedging, confirming hypothesis that firms conducting businesses in other countries and being exposed to foreign exchange risk, engage in derivatives hedging.
21 Table 5
Pearson Correlation Coefficients
Hedging
variable Firm size Profitability
Capital expenditure Dividend dummy Leverage Geograph. diversific. Dividend yield Altman Z-score Growth Market-to-book ratio Foreign assets Convertible debt Preferred stock Quick ratio MV of equity R&D expense Hedging variable 1.00 Firm size 0.20* 1.00 Profitability -0.24 -0.35*** 1.00 Capital expenditure 0.14 0.00 -0.04 1.00 Dividend dummy -0.25** 0.26** 0.10 -0.03 1.00 Leverage 0.33*** 0.09 -0.11 -0.04 0.03 1.00 Geographic diversification 0.25** 0.00 0.26** 0.08 -0.08 -0.08 1.00 Dividend yield -0.17 0.19 -0.02 -0.18 0.51*** 0.26** -0.25** 1.00 Altman Z-score -0.26** -0.49*** 0.77*** 0.02 0.06 -0.40*** 0.12 -0.19* 1.00 Growth -0.07 0.03 0.04 0.10 -0.01 0.08 0.04 0.10 -0.03 1.00 Market-to-book ratio -0.06 -0.29** 0.53*** -0.09 0.18 0.24** 0.15 0.11 0.37*** -0.04 1.00 Foreign assets 0.27** 0.14 0.12 0.31*** -0.01 0.13 0.35*** -0.13 -0.04 0.08 0.26** 1.00 Convertible debt 0.12 0.03 -0.10 0.04 -0.42*** 0.08 0.11 -0.25** -0.21* 0.07 -0.12 -0.02 1.00 Preferred stock -0.03 0.23* -0.15 -0.05 0.05 0.11 0.07 0.06 -0.15 -0.06 -0.10 -0.02 -0.04 1.00 Quick ratio 0.01 -0.14 0.27** -0.04 -0.12 -0.12 0.31*** -0.17 0.28** 0.11 -0.14 0.02 0.09 -0.03 1.00 Market value of equity 0.01 0.78*** 0.18 -0.04 0.39*** -0.01 0.15 0.14 0.02 0.04 0.11 0.19* -0.11 0.08 -0.01 1.00 R&D expense 0.04 0.06 0.11 -0.24** -0.31*** -0.19* 0.25** -0.20* 0.17 0.04 -0.11 -0.09 0.18 -0.04 0.63*** 0.21* 1.00
22
4.2 Principal component analysis
Table 6, Panel B shows the correlations between all three measures of hedging activity and indicates that there is fairly substantial variation in hedging. As such, the highest correlation coefficient of 0.76 is between the net fair value and gross notional amount. Subject to this caveat, the principal component analysis is employed to avoid multicollinearity and allow for the most important influences from all of these variables at the same time. Given the factor loadings of the principal components, from Panel A, I compute a single measure of the extent of hedging for each derivatives user, to be used further in the multivariate analysis.
Table 6
Factor analysis
Panel A: Factor loadings of the principal components for hedging variables
Hedging variables aj1 aj2 aj3
Binary derivatives variable 0.45 0.86 0.22
Net fair value (scaled) 0.60 -0.48 0.63
Gross notional amount (scaled) 0.66 -0.15 -0.74
Eigenvalue λi 2.02 0.77 0.21
Proportion of variability
explained by eigenvalue i, φi (%)
67.2 25.8 7.00
Panel B: Pearson correlation coefficients
Binary derivatives
variable
Net fair value Gross notional amount
Binary derivatives variable 1.00
Net fair value (scaled) 0.26 1.00
Gross notional amount (scaled) 0.46 0.76 1.00
Panel A shows the factor loadings of all three principal components, as well as the ordered eigenvalues for hedging variables and the proportion of variability it explains. The loadings on each factor making up the first principal component are all positive. Panel B shows the ordinary correlation coefficients between the three measures of hedging. The binary derivatives variable is a dummy variable that equals to one if the firm hedges and zero, otherwise. The net fair value is computed as the absolute value of the net gain or loss on derivative positions that are marked to market. The gross notional value is calculated by summing the long and short positions.
4.3 Multivariate analysis
23 The hedging variable and firm value, measured by Tobin’s Q create the most concern because the decision to hedge with derivatives and their use, in fact can affect the firm value, and it in turn can determine the choice for hedging.
To further control for simultaneity of the firm value and derivatives use, I estimate their determinants simultaneously with a two-stage estimation technique. There is no theoretical model explaining a firm’s joint choice of its derivatives use and how it impacts the value of the firm. The structural equations are
The firm value:
Q = 𝛼1+ 𝛽11𝐹𝑆𝑍 + 𝛽21𝑃 + 𝛽31𝐼𝑂 + 𝛽41𝐴𝐹𝑀 + 𝛽51𝐿 + 𝛽61𝐺𝐷 + 𝛽71𝐹𝐷 + 𝛽81𝐹𝑆𝑇 + 𝛽91𝐻𝑉 + 𝜀1 (3)
Derivatives use decision:
HV = 𝛼2+ 𝛽12𝑄 + 𝛽
22𝐺 + 𝛽32𝑃𝐵 + 𝛽42𝐿 + 𝛽52𝐺𝐷+𝛽62𝐹𝑆𝑇 + 𝛽72𝐹𝐷 + 𝛽82𝐹𝐴 + 𝛽92𝐶𝑉𝐷 + 𝛽102 𝑃𝑆 + 𝛽112 𝑄𝑅 + 𝛽
122 𝑅𝐷 + 𝛽132 𝑀𝑉𝐸 + 𝜀2 (4)
The reduced form equations are:
HV = (α1𝛽12+𝛼2) (1−𝛽91𝛽12) + 𝛽11 (1−𝛽91𝛽12)𝐹𝑆𝑍 + 𝛽21𝛽12 (1−𝛽91𝛽12)𝑃 + 𝛽31𝛽12 (1−𝛽91𝛽12)𝐼𝑂 + 𝛽41𝛽12 (1−𝛽91𝛽12)𝐴𝐹𝑀 + (𝛽51𝛽12+𝛽42) (1−𝛽91𝛽12) 𝐿 + (𝛽61𝛽12+𝛽52) (1−𝛽91𝛽12) 𝐺𝐷 + (𝛽71𝛽12+𝛽72) (1−𝛽91𝛽12) 𝐹𝐷 + (𝛽81𝛽12+𝛽62) (1−𝛽91𝛽12) 𝐹𝑆𝑇 + 𝛽22 (1−𝛽91𝛽12)𝐺 + 𝛽32 (1−𝛽91𝛽12)𝑃𝐵 + 𝛽82 (1−𝛽91𝛽12)𝐹𝐴 + 𝛽92 (1−𝛽91𝛽12) 𝐶𝑉𝐷 + 𝛽102 (1−𝛽91𝛽12) 𝑃𝑆 + + 𝛽112 (1−𝛽91𝛽12) 𝑄𝑅 + 𝛽122 (1−𝛽91𝛽12) 𝑅𝐷 + 𝛽132 (1−𝛽91𝛽12) 𝑀𝑉𝐸 +(𝜀1𝛽12+𝜀2) (1−𝛽91𝛽12) (5) Q =(𝛼1+𝛼2𝛽91) 1−𝛽91𝛽12 + 𝛽11 1−𝛽91𝛽12 𝐹𝑆𝑍 + 𝛽21 1−𝛽91𝛽12 𝑃 + 𝛽31 1−𝛽91𝛽12 𝐼𝑂 + 𝛽41 1−𝛽91𝛽12 𝐴𝐹𝑀 + (𝛽51+𝛽91𝛽42) 1−𝛽91𝛽12 𝐿 + (𝛽61+𝛽91𝛽52) 1−𝛽91𝛽12 𝐺𝐷 + (𝛽71+𝛽91𝛽72) 1−𝛽91𝛽12 𝐹𝐷 + (𝛽81+𝛽91𝛽62) 1−𝛽91𝛽12 𝐹𝑆𝑇 + 𝛽91𝛽22 1−𝛽91𝛽12 𝐺 + 𝛽91𝛽32 1−𝛽91𝛽12 𝑃𝐵 + + 𝛽91𝛽82 1−𝛽91𝛽12 𝐹𝐴 + 𝛽91𝛽92 1−𝛽91𝛽12 𝐶𝑉𝐷 + 𝛽91𝛽102 1−𝛽91𝛽12 𝑃𝑆 + 𝛽91𝛽112 1−𝛽91𝛽12 𝑄𝑅 + 𝛽91𝛽122 1−𝛽91𝛽12 𝑅𝐷 + 𝛽91𝛽132 1−𝛽91𝛽12 𝑀𝑉𝐸 + (𝜀1+𝛽91𝜀2) 1−𝛽91𝛽12 (6)
24
4.2.1 Hausman test for exogeneity
Of particular note, given the discussion on the issue of endogeneity, Geczy et al. (1997) control for simultaneity of the capital structure and derivatives use decisions and specify the model of the capital structure decision following Titman and Wessels (1988), and Opler and Titman (1993). In contrast to the previous work, Rogers (2002) recognizes that risk-taking incentive proxies utilized in the risk management model are also choice variables. He models this choice by first solving for a specification of CEO risk-taking incentives and then, incorporates the predicted value from this model as an explanatory variable into the risk management model.
As there are no empirical models explaining a firm’s joint choice of its derivatives use and how it affects the value of the firm, I apply the Hausman test for exogeneity to check whether the variables, namely firm value and the extent of hedging, need actually to be treated as endogeneous.
First, I rewrite the Eqs. (5) and (6) in the following way:
𝐻𝑉 = 𝜋10+ 𝜋11𝐹𝑆𝑍 + 𝜋12𝑃 + 𝜋13𝐼𝑂 + 𝜋14𝐴𝐹𝑀 + 𝜋15𝐿 + 𝜋16𝐺𝐷 + 𝜋17𝐹𝐷 + 𝜋18𝐹𝑆𝑇 + 𝜋19𝐺 + 𝜋20𝑃𝐵 + 𝜋21𝐹𝐴 + 𝜋22𝐶𝑉𝐷 + 𝜋23𝑃𝑆 + 𝜋24𝑄𝑅 + 𝜋25𝑅𝐷 + 𝜋26𝑀𝑉𝐸 + 𝑢1 (7)
𝑄 = 𝜋30+ 𝜋31𝐹𝑆𝑍 + 𝜋32𝑃 + 𝜋33𝐼𝑂 + 𝜋34𝐴𝐹𝑀 + 𝜋35𝐿 + 𝜋36𝐺𝐷 + 𝜋37𝐹𝐷 + 𝜋38𝐹𝑆𝑇 + 𝜋39𝐺 + 𝜋40𝑃𝐵 + 𝜋41𝐹𝐴 + 𝜋42𝐶𝑉𝐷 + 𝜋43𝑃𝑆 + 𝜋44𝑄𝑅 + 𝜋45𝑅𝐷 + 𝜋46𝑀𝑉𝐸 + 𝑢2 (8)
Where the π coefficients in the reduced forms are simple combinations of the original coefficients, so that: 𝜋10 = (α1𝛽12+𝛼2) (1−𝛽91𝛽12); 𝜋11 = 𝛽11 (1−𝛽91𝛽12); 𝜋12 = 𝛽21𝛽12 (1−𝛽91𝛽12); 𝜋13 = 𝛽31𝛽12 (1−𝛽91𝛽12); 𝜋14 = 𝛽41𝛽12 (1−𝛽91𝛽12); 𝜋15 = (𝛽51𝛽12+𝛽42) (1−𝛽91𝛽12); 𝜋16 = (𝛽61𝛽12+𝛽52) (1−𝛽91𝛽12); 𝜋17 = (𝛽71𝛽12+𝛽72) (1−𝛽91𝛽12); 𝜋18 = (𝛽81𝛽12+𝛽62) (1−𝛽91𝛽12); 𝜋19= 𝛽22 (1−𝛽91𝛽12); 𝜋20 = 𝛽32 (1−𝛽91𝛽12); 𝜋21 = 𝛽82 (1−𝛽91𝛽12); 𝜋22 = 𝛽92 (1−𝛽91𝛽12); 𝜋23 = 𝛽102 (1−𝛽91𝛽12); 𝜋24 = 𝛽112 (1−𝛽91𝛽12); 𝜋25 = 𝛽122 (1−𝛽91𝛽12); 𝜋26 = 𝛽132 (1−𝛽91𝛽12); 𝑢1 = (𝜀1𝛽12+𝜀2) (1−𝛽91𝛽12) 𝜋30 = (𝛼1+𝛼2𝛽91) 1−𝛽91𝛽12 ; 𝜋31 = 𝛽11 1−𝛽91𝛽12 ; 𝜋32 = 𝛽21 1−𝛽91𝛽12 ; 𝜋33 = 𝛽31 1−𝛽91𝛽12 ; 𝜋34 = 𝛽41 1−𝛽91𝛽12 ; 𝜋35 = (𝛽51+𝛽91𝛽42) 1−𝛽91𝛽12 ; 𝜋36 = (𝛽61+𝛽91𝛽52) 1−𝛽91𝛽12 ; 𝜋37 = (𝛽71+𝛽91𝛽72) 1−𝛽91𝛽12 ; 𝜋38 = (𝛽81+𝛽91𝛽62) 1−𝛽91𝛽12 ; 𝜋39= 𝛽91𝛽22 1−𝛽91𝛽12 ; 𝜋40 = 𝛽91𝛽32 1−𝛽91𝛽12 ; 𝜋41 = 𝛽91𝛽82 1−𝛽91𝛽12 ; 𝜋42 = 𝛽91𝛽92 1−𝛽91𝛽12 ; 𝜋43 = 𝛽91𝛽102 1−𝛽91𝛽 12 ; 𝜋44 𝛽91𝛽112 1−𝛽91𝛽 12 ; 𝜋45 = 𝛽91𝛽132 1−𝛽91𝛽 12 ; 𝜋45 = 𝛽91𝛽132 1−𝛽91𝛽 12 ; 𝜋46 = 𝛽91𝛽132 1−𝛽91𝛽 12 𝑢2= (𝜀1+𝛽91𝜀2) 1−𝛽91𝛽 12
Afterwards, I estimate the reduced form equations and save the fitted values 𝑌 1(= 𝑄 ) and 𝑌 2(= 𝐻𝑉 ) to include them in the structural models, (1) and (2), and estimate them.
25 HV = 𝛼2+ 𝛽12𝑄 + 𝛽22𝐺 + 𝛽32𝑃𝐵 + 𝛽42𝐿 + 𝛽52𝐺𝐷+𝛽62𝐹𝑆𝑇 + 𝛽72𝐹𝐷 + 𝛽82𝐹𝐴 + 𝛽92𝐶𝑉𝐷 + 𝛽102 𝑃𝑆 + 𝛽112 𝑄𝑅 + 𝛽122 𝑅𝐷 + 𝛽123 𝑀𝑉𝐸 +𝜆2𝑌1+ 𝜀2 (10) Table 7 Hausman test Independent variables Column (1) Tobin’s Q Column (2) Hedging variable
Coefficient p-value Coefficient p-value
Intercept -0.574 0.409 -190.590 0.230 Firm size 0.013 0.735 Profitability, % 0.012 0.263 Capital expenditure, % -0.004 0.457 Dividend dummy 0.045 0.647 Leverage, % 0.014 0.000 3.117 0.000 Geographic diversification 0.246 0.012 57.320 0.008 Dividend yield, % -0.016 0.408 -6.193 0.190 Altman Z-score 0.096 0.001 14.331 0.032 Growth, % -0.405 0.323 Market-to-book ratio 5.891 0.195 Foreign assets. % 0.491 0.255 Convertible debt, % -2.160 0.319 Preferred stock, % -10.201 0.058 Quick ratio -13.105 0.184
Market value of equity 8.953 0.339
R&D expenses, % 2.673 0.083
Hedging variable 0.001 0.320
Fitted hedging variable -0.004 0.005
Tobin’s Q 118.388 0.0714
Fitted Tobin’s Q -305.094 0.001
This table shows the final results of conducting a Hausman test for exogeneity, including the fitted values of Tobin’s Q and hedging variable. Tobin’s Q is computed by taking the natural logarithm of market value of assets over the book value of assets. The firm market value is calculated as the book value of total assets plus the market value of the equity minus the book value of equity. Hedging variable is calculated by conducting a principal component analysis, for three different measures of hedging, namely a binary variable, net fair value and gross notional amount. Firm size is measured by the natural logarithm of total assets. Profitability is the return on assets. Capital expenditure is expressed as a percentage of total assets. Dividend dummy is a variable that equals to one if the firm pays dividends and zero, otherwise. Leverage is computed as the book value of long-term debt over the market value of common equity. Geographic diversification is expressed as a binary variable that is set to one when companies expand their activities in other countries and zero, otherwise. Dividend yield is at close, annual and calculated as dividend per share divided by the market price at year-end. Altman’s Z-score is computed by using Formula (2). Growth is the percentage of total assets growth. Market-to-book ratio is calculated by dividing the annual closing price of the stock by the book value per share. Foreign assets are expressed as percentage of total assets. Convertible debt is the book value of convertible debt scaled by market value of equity. Preferred stock is the book value of preferred stock scaled by market value of equity. Quick ratio is calculated as cash and cash equivalents plus net receivables, divided by total current liabilities. Research and development expenses are expressed as percentage of sales.
26 HV from the reduced form equations. Thus, it will not be valid to simply estimate these equations on their own using OLS.
4.2.2 Simultaneity of derivatives usage and firm value
The model employed in Table 8 assumes that derivative holdings and firm value are partial functions of one another. In this subsection, I test the two hypotheses, first, if the use of derivatives affects the value of the firm, so that it should offer additional power in explaining risk management behavior and, second, what determines the firms to engage in derivatives hedging.
A key feature of the earlier defined system is that risk management choice is based upon variables that proxy for financial distress costs, underinvestment costs, exposure and variation, firm size, diversification and substitutes for hedging. These are proxies for “value-maximizing” hedging incentives. If the risk management decision imparts a positive impact on firm value, then the firm value can play a role in the extent of hedging.
Column (1) of Table 8 shows the results of the simultaneous system (Eqs. (1) and (2)), for the whole sample, whilst Column (2) indicates the estimates for hedging firms only.
Hereby findings hint that the predicted hedging incentives are negatively related to firm value, for both samples. This result suggests that firm value is partially driven because of relatively low risk management incentives. Additionally, the predicted firm value is a factor in setting the extent of hedging. Therefore, corporate derivatives usage is partially a function of firm value, and this value is driven partially by firms’ derivative holdings. The obtained statistically significant negative relations are consistent with the hypothesis that hedging affects the value of the firms, however, in an unexpected way for this research, namely “value-destroying”.
Furthermore, Table 8, Panel A shows that the coefficient for financial stability measured by Altman’s Z-score is in the hypothesized direction, suggesting that the higher the score, the lower the likelihood of bankruptcy, being significant at 0.01 level. Moreover, the coefficient of leverage is not consistent with the predicted sign but significant at 0.01 level, which is also valid for hedgers’ sample.
The positive sign for geographic diversification suggests that multinationality is positively related to firm value (see, e.g., Morck and Yeung, 1991; Bodnar et al., 1997), the coefficient being significant at 0.10 level, except for hedging firms’ sample. Still, firm size, profitability, investment opportunities and access to financial markets are not significant for any of the samples.
27 Table 8
Two-stage least squares
Independent variables
Column (1) All sample ( N = 74)
Column (2) Hedgers (N = 45)
Coefficient p-value Coefficient p-value Panel A. First-stage regression of firm value
Intercept -0.574 0.550 -0.526 0.610 Firm size 0.013 0.807 0.008 0.892 Profitability 0.012 0.418 0.013 0.353 Capital expenditure -0.004 0.591 -0.005 0.499 Dividend dummy 0.0450 0.741 -0.030 0.848 Leverage 0.014 0.005 0.012 0.003 Geographic diversification 0.246 0.067 0.269 0.256 Dividend yield -0.016 0.550 -0.004 0.929 Altman Z-score 0.096 0.010 0.121 0.002 Hedging variable -0.003 0.050 -0.002 0.049 Nr. of observations 74 45 Adjusted R-squared 38% 76%
Panel B. Second-stage regression of the extent of derivatives hedging
Intercept -190.590 0.302 17.500 0.960 Tobin’s Q -186.706 0.004 -268.148 0.009 Growth -0.405 0.396 0.480 0.592 Market-to-book ratio 5.891 0.265 4.892 0.478 Leverage 3.117 0.001 4.032 0.004 Geographic diversification 57.320 0.022 55.649 0.476 Altman Z-score 14.331 0.064 28.846 0.053 Dividend yield -6.193 0.260 -17.767 0.150 Foreign assets 0.491 0.328 0.563 0.399 Convertible debt -2.160 0.392 -6.413 0.048 Preferred stock -10.201 0.102 -8.658 0.231 Quick ratio -13.105 0.253 -24.816 0.127
Market value of equity 8.953 0.411 -2.232 0.913
R&D expenses 2.673 0.136 4.049 0.125
Nr. of observations 74 45
Adjusted R-squared 11% 42%
28 More importantly, two of the three proxies for financial distress in determining firms to hedge are significant. The coefficient for Altman’s Z-score is significant at 0.10 level but not in the predicted direction. Contrariwise, the coefficient for leverage is positive and significant at 0.01 level, supporting the argument that more highly leveraged firms have higher payment obligations, and experiencing difficulties in honoring these commitments, make greater use of derivatives, since they face higher expected costs of financial distress. This result is similar to those obtained by Dolde (1995), Berkaman and Bradbury (1997), Haushalter (2000), Gay and Nam (1999), and Howton and Perfect (1998).
On the other hand, Stulz (1996), Ross (1997), and Leland (1998) suggest an alternative explanation that debt ratio and hedging behavior can be positively related. The use of derivatives for hedging purposes reduces the volatility of income and in the same time reduces the probability of encountering financial distress costs, thus, increasing the debt capacity. Therefore, the ability to increase debt capacity provides an incentive to hedge.
The impact of geographic diversification is not predicted, but the results suggest a positive impact on the extent of hedging, significant at 0.05 level, for the whole sample only. Put another way, firms that conduct business in other countries, are exposed, inter alia, to foreign exchange risk which in turn induces them to hedge.
Another interesting result is that book value of preferred stock negatively relates to the extent of hedging, but is significant at 0.1 level only for the whole sample, whereas the book value of convertible debt also negatively relates to the hedging decision and is significant at 0.05 level, but only for derivatives users. These are indeed substitutes for hedging with derivatives.
I do not find any relation between the level of hedging activity and the proportion of overseas assets. I also find no relation between hedging and firm size. Therefore, there is no evidence supporting exposure and variation, and firm size hypotheses.
Finally, I would like to point out that the adjusted R-squared of 11% for the whole sample of firms, in explaining the extent of hedging, indicates a quite low goodness of fit statistic, suggesting that there are other factors that influence the corporate hedging. Notwithstanding this, the adjusted R-squared increases considerably to 42%, for hedging firms’ sample.
5. Conclusions
29 hedging based on the net fair value, gross notional amount and a binary variable, that equals to one if the firm hedges and zero, otherwise.
I simultaneously test the impact on firm value of the extent of hedging and what determines a firm to hedge.In a nutshell, the evidence supports the hypothesis that hedging activity indeed influences firm value, but destroys it. Nevertheless, this finding is important because it is consistent with the previous results provided by Nguyen and Faff (2010), that a “discount” is imposed on firms that decide to hedge. However, an important caveat to the supportive evidence from previous studies regarding that hedging increases the value of a firm, can be driven by the fact that derivatives’ usage proxies for other firm attributes which are known to affect its value. The traditional risk-management theory is unlikely to fully explain the motivation for this derivatives program. This suggests a need for additional analysis to be undertaken to understand the precise mechanisms by which the use of derivatives affects firm value. Perhaps corporate governance or managerial ability plays a role as suggested by Allayannis et al. (2012) and Lel (2012).
The results concerning the determinants of derivatives use are consistent with theoretical models of corporate risk management, with few exceptions. Especially, there is evidence that Tobin’s Q negatively influences the extent of hedging, suggesting that firms with higher growth opportunities are less likely to hedge. Given that the growth of total assets, market-to-book ratio, research and development expenses, and market value of equity do not act as the use of derivatives, there is no evidence supporting that underinvestment costs, exposure and variation, and firm size induce the firms to hedge.
More importantly, the multivariate results related to the proxies for financial distress are mixed. Despite the fact that the obtained relation between hedging activity and Altman’s Z-score is not consistent with the prediction, the extent of hedging relates to leverage through two different channels. First, lower average volatility allows higher leverage with consequently greater tax benefits and second, the benefit of hedge comes from minimizing the number of states in which a hedging firm can experience financial difficulty and distress costs, resulting from unused debt capacity.
Notably, corporate derivative use increases with geographic diversification, illustrating that a firm with higher levels of multinational operations has greater risk exposure, and therefore, receives superior benefits from hedging. This finding is consistent with Lim and Wang (2007) evidence that financial hedging and corporate diversification are more often complementary than substitutive.
However, the extent of hedging decreases with the two alternatives for hedging, such as convertible debt and preferred stocks, which indeed proves that these are substitutes for hedging with derivatives.
