Fuel Hedging and Debt Capacity in the
U.S. Airline Industry
Abstract
Does hedging affect the debt ratio? If so, is the direction consistent with the theory regarding hedging and derivatives? This thesis investigates the relationship between jet fuel hedging and the debt ratios of U.S. airlines. Using a sample of 21 U.S. passenger airlines between 2001 and 2007, this paper aims to provide an answer. The research of corporate risk
management mainly focuses on what motivates the use of derivatives and if hedging increases firm value. Graham and Rogers (2002) find that leverage has a positive influence on the use of derivatives and that this relationship also runs backwards. The airline industry offers a good perspective, since airlines have a common exposure to volatile fuel prices and fuel costs represent a substantial part of total costs. Analyzing theories of risk management using fuel hedging could provide interesting insights. The coefficients on the fuel hedging variable are negative in the first two models using ordinary least squares and significant at the 10% and 5% level. A third model shows that fuel hedging is positively related to the debt ratio and could be interpreted as hedging increases debt capacity and tax deductions (Graham and Rogers, 2002). The fixed effects model does not find evidence for the hypothesis that fuel hedging is positively related to the debt ratio.
Bachelor thesis
Emma den Held
10079599
Finance
Supervised by dr. J.E. Ligterink
Faculty of Economics and Business
University of Amsterdam
Table of contents
1. Introduction p. 3
2. Literature review p. 5
2.1. Corporate risk management p. 5
2.2. U.S. airline industry p. 7
2.3. Fuel hedging p. 8 3. Data p. 11 3.1. Sources p. 11 3.2. Sample description p. 11 4. Methodology p. 13 4.1. Variables p. 13 4.2. Models p. 15
5. Does fuel hedging affect the debt ratio? Empirical evidence p. 16
6. Conclusion and suggestions for further research p. 20
References p. 21
1. Introduction
The famous Modigliani-Miller theorem states that, without market imperfections, firm value is independent of a firm’s financing decisions and risk management. Violation of the strong assumptions made by Modigliani and Miller (1958) leads to another view. Theoretical research shows that corporate risk management can be value increasing when there are capital market imperfections such as bankruptcy costs, a convex tax schedule (Smith and Stulz, 1985) or underinvestment problems due to costly external finance (Froot et al., 1993). Corporate risk management is important to most firms and they are likely to engage in hedging activities. In the airline industry for instance, Southwest Airlines is known for hedging activities to manage fuel price risk. The company hedged up to a maximum of 95% of next year’s fuel
requirements between 2001 and 2007. Even though hedging has been thoroughly investigated for many years, questions remain on what exactly motivates the use of derivatives. Research mainly examines the relation between hedging and firm value. This thesis highlights hedging in relation to debt capacity. In their paper, Guay and Kothari (2003) discussed several studies that examined this subject. Some of them find evidence for a positive relation but others fail to do so. Hence there is no clear-cut answer. This paper contributes to the research of corporate risk management by providing an answer to the question whether fuel hedging, as the independent variable, affects the debt capacity of U.S. airlines. The outcome could help to understand why airlines become involved in hedging activities and give direction for further research like investigating the magnitude of the relation.
To reduce their exposure to different sources of risk, airlines make use of hedging strategies. Rising jet fuel prices are an important source of risk facing airlines. Fuel costs often represent a substantial part of total operating expenses. Airlines can hedge their future fuel requirements thereby locking in a part of their future cash flows. There are two main reasons for choosing the airline industry. First, the airline industry is competitive and airlines face similar risk as a result of fluctuating oil prices. Second, Carter et al. (2006) note that in comparison to other underlying assets, for example currencies, jet fuel prices are more volatile. The sample used in their study consists of 28 U.S. passenger airlines. The use of interest rate derivatives among the airlines suggests that risk associated with interest rates is less than risk associated with jet fuel prices.
So how does debt capacity fit into all of this? If a firm uses a good hedging strategy the volatility of future cash flows can be reduced. One way to create value is to increase the debt level to benefit from the increased tax shield. Important questions emerge; Does lower
volatility mean allowing higher leverage? And when a part of the increased debt capacity remains unused, does this decrease the financial distress costs? This thesis aims to provide answers, taking in account the relation between hedging and debt capacity.
Using panel data of 21 U.S. airlines from December 2001 to December 2007, this paper examines the effect of fuel hedging on debt capacity. The mean price of jet fuel is $1.39 per gallon during the period, and the standard deviation nearly 60 cents. Airlines state in their annual reports that their exposure to jet fuel price risk is significant. The results of the
regression analysis are hard to interpret. The first two models estimate a negative coefficient on the fuel hedging variable. This result is significant at the 10% level at least and in
accordance with Carter et al. (2006) who find that firm leverage is negatively related to the amount of fuel hedged. In their article leverage is defined as long-term debt divided by assets. On the other hand this result is opposed to the results of Graham and Rogers (2002) and Bartram (2009) that there is a positive correlation between hedging and leverage.
The paper proceeds as follows. In Section 2 literature regarding hedging and
derivatives is reviewed. Background information on the U.S. airline industry is given and fuel hedging explained. Section 3 describes the sample that is used for this thesis and provides the reader with some relevant information regarding the airlines of the sample. Next, Section 4 explains the models as well as the variables used to investigate if fuel hedging affects the debt capacity. The results of the regression analysis are discussed in Section 5. The paper ends with a summary and conclusion in Section 6. An important note to this thesis is that all derivatives that companies hold are solely for hedging purposes and not for trading or speculating purposes. This is stated in their annual reports.
2. Literature review
2.1. Corporate risk management
Corporate risk management refers to the use of financial instruments, cash instruments and/or derivative instruments, to manage exposure to risk and creating economic value by a firm. Stocks and bonds are examples of cash instruments. Their value is determined directly by the market. Derivative instruments are rather complicated. Their value depends on an underlying entity such as an asset, interest rate or index. Hedging activities are a part of firms’ corporate risk management, who manage risk to increase value. However, it is well known that in a perfect financial market setting decisions regarding risk management do not affect firm value as investors can copy or undo this themselves. Therefore market imperfections drive potential value creation. The remainder of this paragraph discusses what motivates the use of derivatives and earlier evidence with respect to hedging.
Incentives for derivatives use
A question subject to research is if hedging influences firm value. “The determinants of firms’ hedging policies” by Smith and Stulz, published in 1985, was pioneering in this field of research. The article identifies three reasons a value-maximizing firm could hedge: taxes, financial distress costs and managerial incentives. They only consider non-financial firms. If firms with a convex tax function reduce the volatility of taxable income, this can lower expected taxes. In agreement with Smith and Stulz, Hayne Leland (1998) concludes that the benefits of hedging are larger with greater tax convexity. In this case hedging will reduce the expected taxes more. Also, the greater tax benefits resulting from increased leverage make hedging valuable. Graham and Smith (1999) take a closer look at tax convexity using a sample of over 80,000 firm-year observations. Approximately 50% of the firms from their sample face a convex tax function and one-quarter of the firms with a convex tax function have potential tax savings from hedging that appear material. However, the benefits are not large for most firms. In addition, Graham and Rogers (2002) find no evidence that firms hedge with regards to tax convexity.
One of the other reasons a firm might want to hedge is because of reduced financial distress costs. Cash flow volatility can be the reason for a firm to get into a situation where the firm has trouble meeting its payment obligations. If being in financial distress is costly
then reducing the probability of encountering such state will reduce the financial distress costs and thereby increase firm value (Smith and Stulz, 1985). The effect is larger for firms with higher costs of financial distress. Costs associated with the failure to invest in profitable projects are included in the financial distress costs. Besides, if having debt in the capital structure is advantageous, hedging could be used to increase debt capacity (Froot et al., 1993). Research by Haushalter (2000) examined the risk management activities of 100 oil and gas producers from 1992 to 1994. He finds a positive correlation between the extent of hedging and financial leverage. This result supports findings that corporate risk management is used to lessen financial contracting costs (Haushalter, 2000).
Managerial incentives can also be a determinant of a firm’s hedging decisions. There is an asymmetric information problem between managers and shareholders. Managers might act in their own self-interest. Smith and Stulz (1985) point out that managerial compensation should be designed in a way that increasing the value of the firm also increases managers’ expected utility. If managerial compensation is a concave function of firm value, the manager is incentivized to reduce the variability of the firm’s cash flow. Furthermore, if the manager owns a significant fraction of the firm, the firm is expected to hedge more.
When external financing is costly in comparison to internal financing, hedging could make sense for firms. Firms may underinvest if raising external funds is costly. The use of derivatives can increase shareholder value by organizing the need for and availability of internal funds (Froot et al., 1993). Overall, there are several incentives for firms to engage in hedging activities and different strategies and instruments can be used for risk management. Yet the goal remains the same: reducing the risk of adverse price movements and creating value for the firm.
Evidence on the use of derivatives
The hypothesis that hedging increases firm value is investigated in empirical research. The study of Allayannis and Weston (2001) looks into foreign currency hedging and firm value, as measured by Tobin’s Q. Their sample consists of 720 large non-financial firms for the period 1990-1995. Allayannis and Weston investigate the so-called “hedging premium” and find that firms with exposure to exchange rates have a statistically and economically significant hedging premium of 4.87% on average. On the contrary, Jin and Jorion (2006) examined 119 U.S. oil and gas producers from 1998 to 2001 and could not find evidence for
the existence of a hedging premium. In the article is written that there is generally no difference in firm values between firms that hedge and firms that do not hedge.
Guay and Kothari (2003) investigate the hypothesis that financial derivatives are an economically important component of corporate risk management. Changing exchange rates, commodity prices and interest rates by three standard deviations generates at most $15 million in cash and $30 million in value for the median firm’s derivatives portfolio. These changes are relatively small compared to benchmarks such as operating and investing cash flows or firm size (Guay and Kothari, 2003). The first to find evidence with respect to hedging and debt capacity are Graham and Rogers (2002). Their study of a broad cross section of firms indicates that leverage has a positive influence on the use of derivatives but that this relation also runs backwards, meaning that hedging leads to greater debt capacity. More precisely, hedging increases the debt ratio by 3 percent on average. Bartram et al. (2009) find results that are consistent with the findings of Graham and Rogers. In their study of 7,319 companies in 50 countries, covering about 80% of global market capitalization of
non-financial firms, derivatives use has a significant positive effect on leverage.
To sum up, there is quite some literature about hedging. Even though there is not one view on the impact of derivatives on firm value or debt capacity, the existing literature provides a good understanding of the investigated subject of this thesis.
2.2. U.S. Airline industry
Airlines are exposed to risk in different ways. Movements in interest rates, exchange rates and commodity prices all form part of that risk. Hedging strategies are used by airlines to reduce their exposure. Airlines can manage all three of these risks. Particularly the risk of rising jet fuel prices is something they have to deal with since it applies to every firm
operating in the industry. Risk resulting from changes in foreign currency prices, for example, applies to a limited set of airlines that operate in foreign markets (Carter et al., 2006).
In the news the poor economic condition of airlines is frequently discussed. The terrorist attacks of September 11, 2001 had a negative effect on the airline industry and the rising jet fuel prices make it increasingly difficult for airlines to operate. To give an idea, in December 2001 the jet fuel spot price of U.S. Gulf Coast Kerosene was $0.52 per gallon and in December 2007 it was $2.60 per gallon (source: U.S. Energy Information Administration). Airlines can operate more competitively in the market when they are able to control their fuel costs. Further, airlines operate in a capital-intensive industry. The primary assets of airlines,
aircraft, hold substantial collateral value for lenders (Carter et al., 2006). High levels of debt in the capital structure are often observed for firms operating in the industry. Besides buying aircraft, airlines engage in leasing. The firm can then obtain the use of an aircraft for which it makes a series of contractual payments. When a lease qualifies as an “operating lease” for accounting purposes, it affects the balance sheet. In this case, the asset and the underlying capital are not reported (Carter et al., 2006). If airlines make a lot of use of operating leases, reported debt and assets are understated.
2.3. Fuel hedging
Fuel costs usually represent a great amount of operating expenses. By hedging a part of next year’s fuel requirements, airlines have more control over their fuel expenses and are less dependent on fluctuating oil prices.
Evidence on fuel hedging
Carter et al. (2006) study the U.S. Airline industry to find out if hedging affects firm value. Their sample consists of 28 firms and 18 firms report fuel hedging for at least one year during the period 1992-2003. Fuel costs represent on average 13.6% of operating expenses. The average hedged percentage of next year’s fuel requirements for hedging firms is roughly 15% but a wide variation is observed. The choice of dependent variable is the simple
approximation of Tobin’s Q, developed by Chung and Pruitt (1994). A total of 251 firm-year observations of Tobin’s Q is used. In the models that estimate the effect of jet fuel hedging on airline firm value either a binary fuel hedging variable or a continuous fuel hedging variable is used. They are estimated using pooled ordinary least squares (OLS). To demonstrate that the continuous hedging variable is robust to differing econometric specifications, fixed effects and time-series feasible generalized least squares (FGLS) are also used in the regression analysis. The results show a positive relation between jet fuel hedging and airline firm value. The coefficient of the binary fuel-hedging variable is not statistically significant. The results suggest that the average hedging premium for airlines is somewhere between 5% and 10%. This premium is larger than the 4.87% hedging premium that Allayannis and Weston (2001) find in their study of currencies and firm value. Carter et al. identify the determinants of jet fuel hedging before estimating the relation between firm values and hedging behavior. Their results raise questions about causality. Firms with higher values of Tobin’s Q are shown to
hedge more and firms that hedge more are associated with higher values of Tobin’s Q. Therefore in the conclusion they do not claim that airlines can increase firm value by increasing the amount of fuel hedged. The hedging premium reflects that firms, able to take advantage of the benefits associated with hedging, have higher firm value if they choose an optimal hedging percentage. Also interesting is that leverage is negatively related to jet fuel hedging while a positive relation is expected if firms facing higher probabilities of distress hedge more (Carter et al., 2006).
Another relevant article testing if hedging affects firm value is the working paper of Chang and Lin (2009). They use data containing 69 airline companies across 32 countries and the volatility of jet fuel prices in the sample period is larger than in previous studies. They also find that jet fuel hedging is valuable for airline companies and estimate a hedging premium of 12%. For U.S. airlines, engaging in hedging activities even increases firm value by approximately 13.61%.
Hedge contracts
Fuel hedging basically means locking in the cost of future fuel purchases. This way the airline protects itself against losses resulting from rising fuel prices. But it also prevents gains from decreasing fuel prices. Thus, airlines hedge to stabilize their fuel costs (Morrell and Swan, 2006). They can engage in different types of hedge contracts to manage jet fuel price risk. Common types of derivatives are discussed below.
Forward contracts are over-the-counter contracts between two parties. One of the parties agrees to buy a fixed amount of the other party, at a fixed price and at a specified future time. The party that commits to buying the underlying is holding the long position and the party that commits to selling the underlying is holding the short position. For third parties or speculators their tailor-made nature is not very convenient (Morrell and Swan, 2006). Futures contracts are standardized contracts and are traded on exchanges. There are no exchange-traded futures available in aviation fuel. However crude oil is closely related and traded on a liquid market. Next, an option is a contract that gives the owner the right to buy or sell the underlying at a fixed “strike price” on or before a future date. Options give more flexibility since they give the owner the possibility to protect against adverse price movements while still profiting from favorable movements (Morrell and Swan, 2006). Options can be taken out against other parties in aviation fuel. Airlines can make use of collars, which is combining a put and a call option, thereby locking in the jet fuel price
between two values. This is a popular way of hedging by airlines. Finally, swaps are tailor-made and can be used by airlines for hedging purposes. The airline buys a swap for a period (say one year) at a certain strike price and receives a specified amount of jet fuel each sub period (say one month). The average price of each sub period is compared to the strike price, and the airline either pays or receives the difference (Morrell and Swan, 2006). To conclude, jet fuel can only be hedged with over-the-counter contracts. Hedging of exchange-traded crude oil is possible: it eliminates counter-party risk and is more liquid.
There are some alternatives to fuel hedging by airlines like fuel pass-through agreements or charter agreements. These types of agreements can facilitate airlines to pass through the risk of volatile fuel prices to partner airlines or customers (Chang and Lin, 2009). If airlines use these alternatives to protect against jet fuel price risk, they could be less
inclined to hedge.
Hypotheses
Based on the existing literature and the theoretical framework developed above, I derive the final hypothesis about fuel hedging and the debt capacity of airlines. It is difficult to define debt capacity because only numbers of realized debt can be observed. The
hypothesis is based on the debt ratio and the empirical study thus looks at the actual debt position and conclusions are based on that. Using a good hedging strategy, thereby lowering the volatility of future cash flows, could imply a higher level of debt in the capital structure. If an airline hedges a high, not speculative, percentage of next year’s fuel requirements, it is expected to allow a higher debt ratio since exposure to a particular risk is reduced. This is put in words in the final hypothesis:
Hypothesis I: The amount of next year’s fuel requirement hedged is positively related to the debt ratio of U.S. passenger airlines.
In the next section I introduce the sample of airlines that is used for this thesis and continue with the empirical research.
3. Data
3.1. Sources
The sources used to gather data are the database Compustat and the annual reports of airlines from the sample. I consult the 10-K forms, which are the required annual reports by the U.S. Securities and Exchange Commission. In the Compustat database the SIC code 4512 for scheduled air transportation is used to find airlines. Out of the 138 results, only U.S. passenger airlines that operate during the time period used in this thesis are selected. In total 21 U.S. airlines are selected. Information regarding total assets, fixed assets, debt, Tobin’s Q, capital expenditures, sales, cash and S&P credit rating is collected from Compustat. The U.S. dollar is the currency used in the sample. For information concerning fuel hedging the 10-K forms are consulted. These contain detailed information on fuel hedging: whether the firm hedged a part of next year’s fuel requirement, what the percentage fuel hedged is and fuel costs as a percentage of operating expenses. To be more precise, the item 7A of the 10-K form called “Quantitative and Qualitative Disclosures about Market Risk” discloses this specific information. If the firm holds interest rate derivatives and/or foreign exchange derivatives is also stated in the annual reports. A total of 119 observations is used.
3.2. Sample description
To estimate the effect of fuel hedging on debt capacity a time period from December 2001 to December 2007 is used. The period begins after the terrorist attacks of September 11, 2001 and ends before the financial crisis became really severe in late 2008. Alaska Air Group mentions the impact of the terrorist attacks in their 2003 annual report: “The September 11, 2001 terrorist attacks negatively impacted our industry and our business and further
threatened or actual terrorist attacks, or other hostilities involving the U.S., may significantly harm our industry and our business in the future” (Alaska Air Group, Form 10-K, 2003).
From the 21 airlines in the sample, 13 airlines are still active meaning eight airlines ceased operations by now. Furthermore, none of the airlines have a Standard & Poor’s credit rating that is considered Investment Grade. The sample consisting of 21 airlines is believed to be a representative sample of the U.S. airline industry. It contains 9 out of 10 current largest U.S. passenger airlines and 5 out of 10 world’s largest passenger airlines based on total
passengers carried. Eighteen firms end their fiscal year in December. Only two firms end their fiscal year in March and one firm in September. Taking December 2001 as starting point, the
fiscal year 2001 is included for all firms except Mesa Air Group. The last observations are from the fiscal year 2007 except for MAIR Holdings and Frontier Airlines Holdings. General information about the airlines can be found in Table 1 of the Appendix. As mentioned in Section 2, jet fuel prices had risen substantially during the period. The spot price of U.S. Gulf Coast Kerosene in December 2007, $2.6 per gallon is five times the spot price in December 2001, $0.52 per gallon. Gulf Coast, New York Harbor and Los Angeles are three major trading hubs in the U.S. (Carter et al., 2006). Over this time period the mean price of jet fuel is $1.39 per gallon. The standard deviation of average monthly fuel prices is 59.5 cents per gallon. It should be an attractive time period for airlines to engage in fuel hedging.
I find that airlines manage jet fuel price risk, interest rate risk and foreign currency risk. However, some annual reports contain limited data and therefore do not cover all firm-year observations. From the 119 firm-year observations, there are 71 disclosures that a part of next year’s fuel requirements are hedged at fiscal-year end, 37 disclosures of interest rate (IR) derivatives and 15 disclosures of foreign exchange (FX) derivatives.
4. Methodology
4.1. Variables
This thesis explores the empirical relationship between jet fuel hedging and airlines’ debt ratios. The relevant literature is closely followed to construct the models for the
regression analysis. The article of Carter et al. is used as guidance since it examines fuel hedging in the U.S. airline industry and is related literature. There is however one crucial difference: the dependent variable is firm value. Rajan and Zingales (1995) identify factors that correlate with leverage and is consulted while searching for control variables.
The dependent variable that is used is the debt ratio, defined as total debt divided by total assets. The debt ratio strongly relates to debt capacity, but does not take into account the portion of debt that stays unused. As mentioned in Section 2, Graham and Rogers find that hedging increases the debt ratio by approximately 3 percent and point out that the theory indicates that the hedging/leverage causality can go both ways. They model the
hedging/capital structure decision as a simultaneous system, which is appropriate if they are jointly determined. Debt as the dependent variable and hedging on the right-hand side is interpreted as hedging increases debt capacity and tax deductions (Graham and Rogers, 2002).
The regression models should contain an independent variable for jet fuel hedging. I use the percentage of next year’s fuel requirements hedged. In the Appendix an example of how airlines disclose information regarding their percentage hedged is shown. Carter et al. also use this continuous variable for jet fuel hedging. Other explanatory variables are firm size, Tobin’s Q, Standard and Poor’s (S&P) credit rating, capital expenditures, property, plant and equipment (PP&E) and cash and equivalents. Table I summarizes the theory predictions of the variables on the debt ratio.
Firm size is included to control for a possible size effect. Large firms are likely to be better diversified and fail less often. Therefore size should have a positive impact on debt supply (Rajan and Zingales, 1995) and debt ratios increase. There are costs attached to hedging. This could imply that size is positively related to hedging because large firms can better bear these costs. Tobin’s Q is a proxy for growth opportunities. Similarly, capital expenditures provide information about investment opportunities but from a realized
perspective. According to the pecking order theory, firms prioritize their sources of financing. They prefer internal financing over debt and raising equity thereafter. Lang et al. (1996) show that there is a negative relation between growth and leverage for firms with low Tobin’s Q.
It suggests that the negative effect of leverage on growth only applies to firms with good investment opportunities that the market does not recognize and firms without good
investment opportunities but want to grow anyway. Carter et al. found evidence in this respect for a positive relation between Tobin’s Q and fuel hedging.
The S&P credit rating tells something about the probability of bankruptcy. Firms with good credit ratings are less likely to default. Since none of the airlines have a credit rating of BBB- or higher, an Investment Grade (IG) dummy cannot be used. Instead, two dummy variables are created: one indicates a credit rating of B (highest rating) and one indicates a credit rating of D (lowest rating). Firms that have a credit rating of D are very risky and expected to have higher debt ratios than firms with a credit rating of B. Next, cash and equivalents provide a buffer. If a firm has a lot of cash, it is expected to have a lower debt ratio. Cash is also expected to negatively relate to hedging because there is no urgent need to reduce volatility of cash flows by using derivatives when a firm holds a substantial amount of cash. PP&E should serve as collateral, meaning lenders are more willing to supply loans. This is due to the diminishing risk of the lender suffering the agency costs of debt and the assets retaining more value in case of liquidation (Rajan and Zingales, 1995). As a result leverage is expected to increase with PP&E.
Two other dummy variables are included. They indicate whether a firm holds interest rate (IR) derivatives and/or foreign exchange (FX) derivatives. A value of 1 means that the firm holds derivatives of any kind mentioned in the dummy.
Table I: Theory prediction of the variables
Variable Theory prediction Definition
Size + Natural logarithm of total assets
Tobin’s Q +/- (Market value of equity + book value of debt) / (Book value of equity + book value of debt) Capital expenditures +/- Capital expenditures / net sales
S&P credit rating +/- On a scale from B to D
Cash - (Cash + equivalents) / total assets
PP&E + (Property, plant and equipment) / total assets
4.2. Models
The data has the form of panel data. Still, three models are estimated using Ordinary Least Squares (OLS) with robust standard errors. The continuous variable for fuel hedging is used. The first model used for regression contains the following variables:
Model I: Debt ratio = β0+ β1 * ln(Assets) + β2 * Tobin’s Q + β3 * Capital expenditures + β4 * Cash + β5* PP&E + β6 *B-rating indicator + β7 * D-rating indicator + β8 * Percentage of next year’s fuel requirements hedged + ε
In the second model two dummy variables for interest rate and foreign exchange derivatives are added. The third model is similar to the first model but includes an interaction term: fuel hedging * Tobin’s Q. Subsequently, the fourth model is estimated using a different econometric specification, namely firm fixed effects. This further explores the relationship between fuel hedging and the debt ratio taking into account the different entities. It includes all variables except for the interaction term. The B-rating indicator as well as the D-rating indicator are omitted in the fixed-effects model because of collinearity.
5. Does fuel hedging affect the debt ratio? Empirical evidence
Research in the field of corporate risk management concentrates on factors that explain the use of derivatives. The goal of this thesis is to find out whether jet fuel hedging affects the debt ratio of U.S. passenger airlines. The results of the regression analysis could help airlines better understand the impact of their fuel hedging behavior on the debt ratio. Table II presents the summary statistics of variables used in the regressions.
Table II: Summary statistics of the variables
Variable Mean Median Std. Dev. Min. Max.
Debt ratio 0.800 0.781 0.315 0.109 2.299 ln(Assets) 7.751 7.625 1.723 4.411 10.399 Tobin’s Q 1.283 1.173 0.375 0.735 2.613 Capex/sales 0.120 0.056 0.168 -0.010 1.030 Cash/total assets 0.222 0.202 0.128 0.003 0.748 PP&E/total assets 0.563 0.592 0.185 0.012 0.906 B-rating indicator 0.227 0.000 0.421 0 1 D-rating indicator 0.319 0.000 0.468 0 1
% of next year’s fuel requirements hedged
17.858 12.000 21.997 0 95
IR derivatives indicator 0.311 0.000 0.465 0 1
FX derivatives indicator 0.126 0.000 0.333 0 1
% of next year’s fuel
requirements hedged * Tobin’s Q
23.964 13.150 33.113 0 145.796
In total 119 observations are used. The mean of next year’s fuel requirements hedged is 17.86% and the standard deviation is 22%. Tobin’s Q has a mean of 1.283, which is higher than 1, and a standard deviation of 0.375. The debt ratio has a mean of 0.800, implying that on average airlines have less total debt than assets. The standard deviation is 0.315.
Table III shows the correlations coefficients of several selected variables. An extensive correlation matrix can be found in Table 2 in the Appendix. Surprisingly, the correlation between the debt ratio and fuel hedging variable is negative. The other correlation coefficients are as expected. For example, fuel hedging and cash are negatively correlated with the possible explanation laid out in Section 4.
Table III: Correlation coefficients
Correlation coefficients
Debt ratio, fuel hedging -0.1776
Debt ratio, ln(Assets) 0.2834
Debt ratio, Tobin’s Q 0.1109
Fuel hedging, ln(Assets) 0.3975
Fuel hedging, Tobin’s Q 0.1288
Fuel hedging, Cash/total assets -0.1243
The results of the regression analysis are presented in Table IV. The models find some interesting results. In the first and second model the estimated coefficient on the hedging variable is not consistent with the expectations. It shows a negative relation between the percentage of next year’s fuel requirements hedged and the debt ratio. The coefficients are -0.0023 and -0.0028 in the first and second model and significant at the 10% and 5% level, respectively. This result implies that higher fuel hedging percentages are associated with lower debt ratios. The coefficient on the interest rate derivatives indicator, added in the second model, is statistically significant. This result suggests that the use of interest rate derivatives decreases the debt ratio by approximately 0.1719. The coefficient on the dummy variable used to indicate the use of foreign exchange derivatives is not significant. The R2 is 0.6145 in the first model and 0.6541 in the second model.
The third model contains an interaction term consisting of the percentage of next year’s fuel requirements hedged and Tobin’s Q. The coefficient is -0.0044 and significant at the 10% level. The correlation coefficient between fuel hedging and Tobin’s Q is positive (see Table II) and together the variables have a negative effect on the debt ratio. The higher the percentage of fuel hedging and the Tobin’s Q, the bigger is the negative effect on the debt ratio. This is also the only model that estimates the coefficient on the fuel hedging percentage in accordance with my expectations. It finds a coefficient of 0.0038, however not significant. The R2 of the third model is 0.6255.
Finally, the fixed-effects model is estimated. This model takes into account that the sample consists of 21 different entities. The results could imply that within the sample firms with higher fuel hedging percentages have higher debt ratios. Unfortunately, only the estimated coefficients on Tobin’s Q, capital expenditures and property, plant and equipment are statistically significant in this model. The coefficient on the fuel hedging percentage is -0.0005 and is not significant. The model makes it hard to draw conclusions.
All four models show that debt ratio increases with the natural logarithm of total assets. As expected, larger firms have higher debt ratios. Higher values of Tobin’s Q are associated with higher debt ratios. The coefficients on Tobin’s Q are significant at the 5% level at least in all models and vary between 0.2359 and 0.3357. Capital expenditures have a statistically significant negative effect on the debt ratio. A possible explanation is that firms with good investment opportunities and high capital expenditures choose internal financing. Cash also negatively affects the debt ratio. A high amount of cash is likely to coincide with lower debt ratio. The coefficients on cash are significant at the 1% level in the first three models. Property, plant and equipment are only significant in the fixed-effects model. In that model, the coefficient shows a strong positive relation between PP&E and the debt ratio as expected.
Table IV: Results of regressions of the debt ratio
Notes: Standard errors are reported in the parentheses. Statistical significance at 10%, 5% and 1% is indicated by *,** and ***, respectively. In Model 4 the B-rating and D-rating indicators are omitted because of collinearity.
n = 119 Model 1 Pooled OLS Model 2 Pooled OLS Model 3 Pooled OLS Model 4 Fixed Effects Constant 0.6695*** (0.1820) 0.4932** (0.1929) 0.5774*** (0.1999) -0.4490 (0.4323) ln(Assets) 0.0289** (0.0126) 0.0526*** (0.0179) 0.0278** (0.0125) 0.0312 (0.0454) Tobin’s Q 0.2503** (0.1041) 0.2359** (0.1127) 0.3179** (0.1230) 0.3357*** (0.0546) Capex/sales -0.4442*** (0.1051) -0.5246*** (0.1135) -0.4282*** (0.1125) -0.4283*** (0.1528) Cash/total assets -1.1223*** (0.3057) -1.1108*** (0.3168) -1.1206*** (0.3099) 0.2993 (0.3213) PP&E/total assets -0.0894 (0.1771) 0.0426 (0.1958) -0.0786 (0.1793) 1.0039*** (0.2356) B-rating indicator -0.2187*** (0.0541) -0.1947*** (0.0516) -0.2076*** (0.0547) D-rating indicator 0.0913* (0.0466) 0.0708 (0.0459) 0.0889* (0.0476) % next year’s fuel
requirements hedged -0.0023* (0.0012) -0.0028** (0.0012) 0.0038 (0.0034) -0.0005 (0.0013) IR derivatives indicator -0.1719*** (0.0546) 0.0390 (0.0592) FX derivatives indicator 0.0513 (0.0645) -0.0556 (0.0902) % next year’s fuel
requirements hedged * Tobin’s Q
-0.0044* (0.0025)
The dummy variable used for S&P credit rating implies that a “high” credit rating (in this case B) decreases the debt ratio by 0.2187 on average in the first model and has a significant negative effect in the first three models. It is omitted in the fixed-effects model because of collinearity. Altogether, the results of the regression analysis are diverse. The first and second model show a negative relation between the percentage of next year’s fuel requirements hedged and debt ratio while in the third model a positive relation is estimated. Only the latter result is consistent with the expectations. The fixed-effects model takes into account that this study is dealing with panel data and there are more observations for each airline. When this econometric specification is used, the coefficient on the hedging variable is almost zero and it seems that there is no evidence for any effect.
Likely the regression analysis suffers from endogeneity of the regressors. This occurs when regressors are correlated with the error term. As a result, estimators are biased and the Table IV results should be questioned. This could also explain why the coefficients shift from negative to positive numbers. Possible causes of endogeneity are omitted variables, sample selection and simultaneous causality and one has to be critical when interpreting these results. Earlier I mentioned that Graham and Rogers (2002) find that the hedging/leverage causality can go both ways. This could cause the problem of endogeneity. Granger causality tests are conducted. The null hypothesis that the debt ratio does not Granger-cause fuel hedging, could not be rejected (F-statistic is 1.908). And the null hypothesis that fuel hedging does not Granger-cause the debt ratio could also not be rejected (F-statistic is 0.5157). Hence the possibility remains that this analysis suffers from endogeneity due to simultaneous causality. The results of the Ganger causality tests are reported in Table V.
Table V: Granger causality Wald tests
Equation Excluded F-statistic Prob. > F Debt ratio Fuel hedging
Debt ratio ALL
1.908 1.908
0.1704 0.1704 Fuel hedging Debt ratio
Fuel hedging ALL
0.5157 0.5157
0.4744 0.4644
6. Conclusion and suggestions for further research
Fuel hedging is a common activity of airlines to manage risk resulting from volatile fuel prices. The airline industry offers a good perspective to investigate the relationship between fuel hedging and the debt ratio due to the high volatility of jet fuel prices and the fact that the industry is very competitive and airlines face similar risk as a result of fluctuating oil prices. Previous research by Carter et al. (2006) shows that firm value is positively associated with the amount of hedging. A value-maximizing firm would hedge for three reasons according to Smith and Stulz (1985): taxes, financial distress costs and managerial incentives. The amount of fuel hedging is expected to be positively related to the debt ratio. Reducing volatility of future cash flows should have a positive effect on the debt supply and therefore lead to higher observed debt ratios.
This thesis could not find evidence that fuel hedging is positively related to the debt ratio. The regression analysis even suggests there might be a small negative effect, however, this should be further investigated. Contradictory to findings of Haushalter (2000), Graham and Rogers (2002) and Bartram (2009) the coefficients on the hedging variable are negative and significant at the 10% and 5% level in model 1 and 2, respectively. Nonetheless, Carter et al. (2006) also find that firm leverage is negatively related to the amount of jet fuel hedged even though this is not the focal point of their study. Only the third model shows a positive relation between the amount of fuel hedging and the debt ratio but is not significant. The last model demonstrates that the negative effect in accordance with model 1 and 2 is not
significant when using fixed effects. The clear-cut answer this empirical research was hoping for could not be found. But with resources being limited for this thesis, I believe that the relation between fuel hedging and the debt ratio should be further examined. Understanding this relation could help airlines revise or refine their hedging strategies.
The problem of endogeneity should be closely considered when developing new models to test the hypothesis. For example, lagged percentages of next year’s fuel
requirements could be used in regressions. The U.S. airline industry is quite limited and I would suggest composing a global sample of airlines like Chang and Lin (2009). Fuel pass-through and other agreements could be another factor to threaten the models. Airlines that have agreements are different from airlines that do not hedge. This should be taken into account. A final suggestion is to use lease-adjusted values of several variables such as total assets and total debt. So regardless of hedging and derivatives being exhaustively investigated in the field of risk management these days, there still is a lot of work to do.
References
Allayannis, G. and Weston, J. (2001): “The use of foreign currency derivatives and firm market value”. Review of Financial Studies, 14, 243-276.
Bartram, S., Brown, G. and Fehle, F. (2009): “International evidence on financial derivatives usage”. Financial Management, 38, 185-206.
Carter, D., Rogers, D. and Simkins, B. (2006): “Does hedging affect firm value? Evidence from the U.S. airline industry”. Financial Management, 35, 53-86.
Chang, Y. and Lin, R. (2009): “Does hedging add value? Evidence from the global airline industry”. Working paper, National Chengchi University.
Chung, K. and Pruitt, S. (1994): “A simple approximation of Tobin’s Q”. Financial Management, 23, 70-74.
Froot, K., Scharfstein, D. and Stein, J. (1993): “Risk management: coordinating corporate investment and financing policies”. Journal of Finance, 48, 1629-1658.
Graham, J. and Rogers, D. (2002): “Do firms hedge in response to tax incentives?”. Journal of Finance, 57, 815–839.
Graham, J. and Smith, C. (1999): “Tax incentives to hedge”. Journal of Finance, 54, 2241-2262.
Guay, W. (1999): “The impact of derivatives on firm risk: An empirical examination of new derivative users”. Journal of Accounting and Economics, 26, 319-351.
Guay, W. and Kothari, S. (2003): “How much do firms hedge with derivatives?”. Journal of Financial Economics, 70, 423-461.
Haushalter, D. (2000): “Financing policy, basis risk, and corporate hedging: evidence from oil and gas producers”. Journal of Finance, 55, 107-152.
Jin, Y. and Jorion, P. (2006): “Firm value and hedging: evidence from U.S. oil and gas producers”. Journal of Finance, 61, 893-919.
Lang, L., Ofck, E. and Stulz, R. (1996): “Leverage, investment and firm growth”. Journal of Financial Economics, 40, 3-29.
Leland, H. (1998): “Agency costs, risk management, and capital structure”. Journal of Finance, 53, 1213-1243.
Mellow, A. and Parsons, J. (2000): “Hedging and liquidity”. Review of Financial Studies, 13,
127-153.
Modigliani, F. and Miller, M. (1958): “The cost of capital, corporation finance, and the theory of investment”. American Economic Review, 48, 261–297.
Morrell, P. and Swan, W. (2006): “Airline jet fuel hedging: theory and practice”. Transport Reviews, 6, 713-730.
Rajan, R. and Zingales, L. (1995): “What do we know about capital structure: some evidence from international data”. Journal of Finance, 50, 1421-1460.
Smith, C. and Stulz, R. (1985): “The determinants of firms’ hedging policies”. Journal of Financial and Quantitative Analysis, 20, 391-405.
Stock, H. and Watson, W (2007): “Introduction to econometrics”. Pearson Education, Pearson/Addison Wesley, Boston (third edition 2012).
Stulz, R. (1996): “Rethinking risk management”. Journal of Applied Corporate Finance, 9, 8-24.
Stulz, R. (2004): “Should we fear derivatives?”. Journal of Economic Perspectives, 18, 173-192.
Tufano, P. (1998): Agency costs of corporate risk management. Financial Management, 27, 67-77.
Appendix
Table 1: U.S. Airline Sample
Notes: Frontier Airlines Holdings was formed from a reorganization of Frontier Airlines in 2006; Frontier Airlines Holdings and Midwest Air Group are currently owned by Republic Airways Holdings; Global Aviation Holdings started as AMTRAN in 2002 and changed its name twice before it became Global Aviation Holdings in 2009; SkyWest was incorporated in 2005; UAL Corp., the parent company of United Airlines, merged with Continental Airlines in 2010 and was renamed United Continental Holdings; MAIR Holdings was dissolved in 2008; The same year Northwest Airlines merged with Delta Air Lines.
Company Founded Headquarter Fiscal year-end Current status
Airtran Holdings 1997 Orlando, Florida 12 Inactive
Alaska Air Group 1985 SeaTac, Washington 12 Active
Allegiant Travel 1998 Enterprise, Nevada 12 Active
AMR Corp. 1982 Fort Worth, Texas 12 Active
Continental Airlines 1934 Houston, Texas 12 Inactive
Delta Air Lines 1924 Atlanta, Georgia 12 Active
ExpressJet Holdings 1986 Atlanta, Georgia 12 Inactive
Frontier Airlines Holdings 2006* Denver, Colorado 3 Inactive Global Aviation Holdings 2009* Peachtree City, Georgia 12 Active
Great Lakes Aviation 1977 Cheyenne, Wyoming 12 Active
Hawaiian Holdings 1929 Honolulu, Hawaii 12 Active
Jetblue Airways 1999 Long Island City, New York 12 Active
MAIR Holdings 1993 Minneapolis, Minnesota 3 Inactive
Mesa Air Group 1982 Phoenix, Arizona 9 Inactive
Midwest Air Group 1989 Oak Creek, Wisconsin 12 Inactive
Northwest Airlines 1926 Eagan, Minnesota 12 Inactive
Republic Airways Holdings 1973 Indianapolis, Indiana 12 Active
SkyWest, Inc. 2005* St. George, Utah 12 Active
Southwest Airlines 1967 Dallas, Texas 12 Active
United Continental Holdings 2010* Chicago, Illinois 12 Active
Example of fuel hedging disclosure:
“In addition to the fuel purchase contracts discussed above, we have entered into fuel related derivative financial instruments with financial institutions to reduce the impact of fluctuations in jet-fuel prices on future fuel expense. Our primary objective of entering into derivative instruments is to reduce the impact on our operating results of the volatility of jet fuel prices. We do not hold or issue derivative financial instruments for trading purposes. As of December 31, 2007, the fair market value of our fuel related derivative assets was $13.0 million and was comprised of option and swap arrangements. As of December 31, 2007, these contracts pertain to 70.6 million gallons of our 2008 fuel purchases and 2.0 million gallons of our 2009 fuel purchases and represent 17.9 percent of our anticipated 2008 jet-fuel
requirement and 0.5 percent of our 2009 jet fuel requirements.”
(Airtran Holdings, Form 10-K, 2007).
Example of fuel pass-through agreement disclosure:
“In the past, the Company has not experienced difficulties with fuel availability and expects to be able to obtain fuel at prevailing prices in quantities sufficient to meet its future needs. Effective January 1, 2002, the Company's contracts with its major partners obligate its major partners to bear the majority of the economic risk of fuel price fluctuations. As such, during the terms of those contracts the Company anticipates that its results from operations will not be directly affected by fuel price volatility.”
(SkyWest, Inc., Form 10-K, 2003).
Table 2: Correlation matrix of variables used in the regressions
Variable Debt ratio Fuel
hedging
ln(Assets) Tobin’s Q Capex/
sales Cash/total assets PPE/total assets Debt ratio 1.0000 Fuel hedging -0.1776 1.0000 ln(Assets) 0.2834 0.3975 1.0000 Tobin’s Q 0.1109 0.1288 -0.1168 1.0000 Capex/sales -0.2603 0.2568 0.0510 0.1692 1.0000 Cash/total assets -0.5126 -0.1243 -0.3954 0.2105 -0.0631 1.0000 PPE/total assets 0.3081 0.1558 0.2202 0.1152 0.2600 -0.7422 1.0000
