MASTER THESIS
OIL PRICE IN THE AIR
By
Valentina Zorzetto
A thesis submitted to the Faculty of ECONOMICS AND BUSINESS
Rijksuniversiteit Groningen
MSc International Financial Management
Supervisor: dr. L.J.R. Bert Scholtens Co-Assessor: dr. Victoria Purice v.zorzetto@student.rug.nl
ABSTRACT
Table of contents
1. Introduction 7
2. Literature and Hypotheses 9
3. Methodology and Discussion 13
4. Descriptive statistics and Results 20
5. Limitations and Further researches 26
6. Conclusions 26
References 28
Websites 30
1. Introduction
The most critical issue to be considered in the future economic environment of the transportation sector (roar, maritime, aviation) is the high volatility nature of oil prices (Elyasiani, 2011). In particular, this research focuses on the airline sector which is fourth-placed among industries with the most significant demand for oil all over the world according to Statista1 (2016). The International Air Transport Association (IATA) recognizes the oil as one of the most significant single cost items for the entire aviation industry sector2 (airline, manufacturing and services). Narayan and Narayan, (2007) find out that the cost of fuel is particularly crucial for the airline companies. The high correlation between crude oil and kerosene (0.91), encourage firms operating in the airline sector to become more sensitive concerning oil prices changes (Indexmundi, 2016)3.
The poor acknowledge and management of oil price fluctuations expose the operations of the whole sector to financial risks that influence their operations (Narayan et al., 2007). Treanor et al. (2014) state that 10%-15% of the total operating costs for the airline sector is attributed to jet fuel (kerosene). Morrell and Swan (2006) show that oil price volatility and the correlated jet fuel costs could be managed through hedging practices.
This research studies the impact of oil prices on airline companies under three different perspectives namely, managerial, financial and international. Under the managerial perspective, the thesis tries to investigate whether adoption of jet fuel hedging practices creates superior firm value for airline companies operating in countries spread all over the world. Tobin’s Q is the model used to proxy the firm value of a sample of 76 global listed airline companies collected from Orbis database. The timeframe taken into consideration goes from 2008 until 2016. Variables have annual frequency since fuel hedging ratios are issued yearly in each firm’s annual reports. In this first hypothesis, it is found that there is a positive relationship between jet fuel hedging and airline firm-value. However, because of F-statistic equals 0.000, the first hypothesis is rejected at any level of significance (reject the hypothesis with a 99% probability).
1 This document is available at
https://www.statista.com/statistics/307194/top-oil-consuming-sectors-worldwide/ Last accessed December 27th 2017.
2 This report is available at
http://www.iata.org/publications/economics/Reports/Industry-Econ-Performance/IATA-Economic-Performance-of-the-Industry-end-year-2017-report.pdf Last accessed December 27th 2017.
https://www.indexmundi.com/commodities/?commodity=jet-The financial perspective of this research examines the sensitiveness of airline companies’ stock market returns in case of an oil price increases or in case of an oil price decreases. To analyze this second assumption, the thesis uses the same population of airline 76 companies over a broader timeframe, from 1991 to 2016 because of a greater data availability on Datastream. After this analysis, the results report a negative correlation between airline companies stock returns and oil price fluctuations. And the asymmetry term used in E-GARCH model to detect the magnitude of oil price fluctuations confirms that airline stock returns are more sensitive to an oil price increases than to an oil price decreases. Durbin-Watson statistic proves absence of autocorrelation hence, the second hypothesis is not rejected.
With the international perspective, the research tries to orientate the topic into a more focused and narrow direction presenting two hypotheses. The third, which divides the sample into developing and developed countries according to the World Bank income level for 20174 (High Income/High Middle Income and Low Middle Income/Low Income). It is expected that countries classified as developed less perceive the impact of oil price increases. The findings confirm these forecasts and the interpretation of the outcomes show that oil price increases have a greater impact on airline companies stock returns operating in developing countries than on airline companies stock returns operating in developed countries.
The fourth hypothesis divides the sample on the basis of oil countries resource endowment, namely oil-rich and oil-poor countries classification from the Organization of the Petroleum Exporting Countries5 (OPEC, 2017). The population and the period analyzed are the same as in the second and third assumptions. Expectations want that oil-rich countries are less affected by oil price increases than oil-poor territories and, are statistically confirmed at a later stage.
This thesis will contribute to the existing literature in evaluating and measuring the impact of oil price changes on firm’s stock return in the airline sector. This path has already been taken by Werner and Concha (2016), but it will differ as it will take into consideration not only the US airline companies but a global sample of listed airline firms, interesting to stimulate comparisons. Moreover, it will take into consideration different effects of oil price on airline stock return which specific multi-country characteristics as the level of income and resource endowment.
4 The information is available at
https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country-and-lending-groups Last accessed December 27th 2017.
5 The document is accessible at http://www.opec.org/opec_web/en/data_graphs/330.htm
The study is divided in three main sections. The literature, which presents the problem statement, grounds the hypothesis and gives an overview of the topic and purpose of the study. A secondary section of this thesis is dedicated to the methodology and to its justification, namely why and how the chosen approaches are preferred over other procedures, followed by the empirical results. Lastly, are presented the concluding remarks, limitations, suggested further researches and conclusions.
2. Literature and Hypothesis
Uncertainty and risk dominate every moment of the operation of all companies and they represent a constant condition of innate entrepreneurial action to economic activity. Risk assumption is the obvious premise of waiting for an economic return (MacCrimmon et al., 1988). Erb et al. (1996) in their study wrote that the risk in the economic-financial field refers to the possibility that the result of an operation performed by an economic individual is different from that presumed ex-ante. The complexity that involves companies operating in the market, brings the need to assess the risk in order to mitigate the impact or exploit its benefits. In fact, operating in a strongly dynamic environment also implies the intensification of risks (MacCrimmon et al., 1988). Risk can arise because of several factors such as events stimulated by humans, the inflation, the wars, governmental changes, business cycles (Jorion 2007) and can be of different nature and grouped into two categories according to Dafir (2016). Operational risk (correlated to the product market in which the companies perform) and risks related to the financial markets (Dafir, 2016). This thesis has a focus on financial risks and more specificity on the commodity risk which is the most important motive for the adoption of advanced investor strategies for airline companies. That is because the airline industry is one of the most oil demanding sector (Statisata, 2106) and because fuel costs cover the bulk of their operating expenses (Dafir, 2016). The major risks of financial nature are:
• Equity risk arise when firms are unable to increase the capital through “the equity markets for funding expansion, acquisitions, restructuring, or other activities of the company”,
• Debt risk arise when firms “cannot borrow in the debt markets for financing ongoing activities or for long-term investments”,
• Foreign exchange (FX) risk is linked to a change in the exchange ratio between two currencies, which affects the value of a good expressed in a foreign currency and,
Airline companies faced significant oil price swings with the increase of oil prices during the last decade, moreover fuel expenses are said to be unpredictable because of the poor control on fluctuations of oil prices (Dafir, 2016). The exposure to fuel price risk approached airline companies to the field of risk management and in particular to the use of tools able to hedge different types of risks associated with assets’ price fluctuations. These methods, also known as derivatives, can be traced back to the 12th century in the commercial island of Venice (Chiu, 2012). A consistent number of studies give proper attention on whether undertaking hedge activities create value for the companies.
Rahnema (1990) states that it is obvious that decisions concerned with the reduction of risk exposure are important to create firm’s value. Several studies which investigate the relationship between derivative practices and firm value adopting Tobin’s Q as a proxy for the latter variable. An example is the one conducted by Carter et al. (2006) who investigate on whether jet fuel hedging techniques adopted by US airline industry during 1992-2003, are a source of value for these firms. They confirm their early assertion finding a positive relationship between jet fuel hedging and airline firm value. Jet fuel plays a key role for companies involved in the air-transports because of the associated high demand. Since the nature of jet fuel is highly volatile, companies might engage in hedging practices to manage it and limit related risk exposures. The major hedging techniques taken into consideration by companies are forward contracts, future contracts, options, collars, swaps, etc. (Morrel and Swan, 2006). This thesis will focus on future contracts whose function is to lock the price of crude oil ensuring a known and stable cash flows to the companies for the contract period. Future contracts are preferred to other hedging methods able to manage fuel price risk because they appear with most frequency in every firm’s airline annual report consequently, for this study there is more data availability.
This research inspired by the existing literature proposed by Carter et al. (2006) aims to explore whether hedging activities bring value to airline companies which undertake derivatives techniques to mitigate jet fuel risk. Then, it will differ as it will take into consideration not just a US but a global sample of airline companies over the recent time-period 2008-2016. Expectations whose foresee a positive relationship between hedged fuel costs and airline firm value lead to the following hypothesis:
To ground the second hypotheses, this research present the study by Narayan et al. (2014), which demonstrate the existence of the correlation between oil price shocks and US stock return volatility considering historical data spanning over 150 years. Key findings suggest that the oil price is an important indicator and predictor of firm return fluctuations hence, of relevant importance from an investors perspective. A very recent study with the objective of analyzing the role of oil prices in predicting stock returns is the one conducted by Christoffersen and Pan (2017). The authors state that in 2006, after the “financialization” of commodity futures market, oil volatility become a relevant indicator and predictor of returns. However, Huang et al. (2005) prove that there is no relationship between the oil futures return and return on stocks, suggesting that oil futures are good factors for diversifying stock portfolios. The second aim of this thesis is to examine whether airline stock returns are more sensitive to an oil price increase than to an oil price decrease. That is, to capture asymmetric stock returns behavior due to oil prices fluctuations. There exist a broad and growing literature on the impact of the oil price on stock returns. Many studies investigate how the sudden changes in oil prices impact the economies at global and country level. Past studies demonstrated that fluctuations in oil prices were not the only responsible factor for the U.S. post World War II decline period (Hamilton, 1983). Several studies take into consideration demand and supply shocks in the crude oil market. Kilian and Park (2009), investigate whether fluctuations in oil price reflect oil demand or oil supply shocks, highlighting that the impact on the stock market return is of different magnitude. Indeed, supply shocks have more impact than demand shocks. Narayan and Sharma (2011) and later Aggarwal et al. (2012) study the influence of oil price shocks on the transport sector (road, maritime and air). The findings in both studies demonstrate that as oil price increases, transportation stock returns decrease and, if the oil price decreases, exposure to fuel risks rise.
Early studies proposed by Mork (1989) and by Hamilton at a later stage in 2003 are more consistent with the second hypothesis, they suggest that an oil price increase has a significant impact on stock returns than an oil price decrease. Relying on Mork (1989) and Hamilton (2003) findings, the thesis wants to test for the asymmetric effect of oil price fluctuations on airline stock returns with a timeframe spanning from January 1991 to December 2016 over a population which includes 76 airline companies all over the world. It is expected that a rise in oil price has a higher impact on airline stock returns than a decrease in the oil price of the same magnitude. Based on this scenario is proposed the second hypotheses:
Similarity, there are no studies which attempt to study the impact of oil prices on airline stock returns in rich and poor countries. It can be expected that because of limited resource endowment, oil-poor countries will be more sensitives to oil price increases since they could incur in costs, for example, due to the geographic distance with the oil supplier country. At this stage are presented the following two hypotheses:
H3: An oil price increase has more impact on airline companies stock returns operating in developing countries than on airline companies stock returns operating in developed countries.
H4: An oil price increase has more impact on airline companies stock returns operating in oil-poor countries than on airline companies stock returns operating in oil-rich countries.
3. Methodology and Discussion
This section will contain the different methodical approaches used to test the hypotheses. The design can be described as highly quantitative since all the assumptions are founded on regressions and specific statistical criteria. For the first hypotheses is tested using a sample of 76 listed airline companies6 operating worldwide and which appears in the International Airline Traffic Association (IATA). This association includes the 83% of the total air traffic, making my results more robust and reliable. The timeframe contains eight years from 2008 to 2016 because it is the most extensive range with the highest number of information available for all variables taken into consideration for the first assumption. The collection of information on hedging percentages is not easy to gather and most of the time is not disclosed by airline firms. The period to be analyzed also comprises two important breaks for the history as the 9/11 terroristic attack and the financial crisis. The first event it is relevant since the 25% of the US oil imports comes from Middle East (EIA, 2016), the price of oil increased after the terroristic attack to reflect the fact that individuals were worried about a possible war and consequently of an inevitable Middle East oil supply disruption. The 2008/2009 financial crisis had a negative impact on the oil sector as it led to a decrease in oil prices (Investopedia, 20157). To
conclude, this period (not very extended) will help to detect a significant effect of hedging approach to firm value due to the presence of consistent oil prices fluctuations.
First of all, there have been collected data about risk management at a general level from literature reviews on the topic. From here, emerged the critical nature of hedging techniques for the airline industry sector, a good starting point where to begin the investigation on the relationship between firm value and fuel price hedging practices, by exploiting market imperfections faced in the real world. Carter et al., (2006) study face problems derived from causality and endogeneity because it regresses firm value on several proxies for derivatives and other control variables. Using this methodology one might encounter some causality issues since it is hard to detect if the hedging practices influence the firm value or vice versa. The data on airline companies have been collected merging two databases, namely Orbis (looking at 4512 SIC) to get ISIN numbers and establish the sample and Datastream to collect information about the variables included in the models to assess the hypotheses8.
Airline hedging ratios have been hand collected from the website of each of the 76 airline companies for eight years. Airline companies hold different accounting periods and ending date for the fiscal year (for example 31st March and 31st December), in this case, it was taken into consideration the hedge ratio of the second part of the year for companies publishing their annual report at the half year. The derivative used by airline companies are futures contracts. The 15th December 2004, the European Parliament and the Council issued the 2004/109/EC directive on the harmonization of transparency, where all listed companies adapt to the accounting standards proposed by the International Financial Reporting Standards (IFRS)9.
Hence, all listed companies, in our study the airline firms disclose their fair value of derivatives in their annual reports. Because of this, it was possible to collect the amount of fuel hedged (measured in percentage of the upcoming fiscal year jet fuel purchases), necessary to test the first hypothesis. Despite this facilitate extraction of hedge ratios, the final panel results to be unbalanced because of missing percentages in some years as airline decide to do not enter into hedging activities for some periods. While collecting hedging ratio, the extraction was carefully oriented at the currency (all converted in USD) since firms often issue their derivatives in the currency of the country in which they are operating.
The dependent variable, firm value is tested using Tobin’s q approximation used in the paper of El Ghoul et al. (2016) and calculated as:
8 All variables used in the models to test the hypotheses are reported in the descriptive statistics
table 2 in the appendix.
9 The document is available at
TQ= !"#$%& (")*% +, "--%&-./++$ (")*% +, "--%&-0/++$ (")*% +, %1*2&3/++$ (")*% +,
"--%&-Where the market value of equity is derived multiplying the share outstanding by the price per share. The difference between the book value of asset and book value of capital which is liabilities market value. Breaking down this approach, a TQ between 0 and 1 indicated that the stock is undervalued since the value of the stock is less than the cost of firm assets replacement. In turn, if TQ is greater than 1, the stock is overvalued.
The independent variable is represented by the percentage of fuel hedged (HEDGE), which measure the rate of jet fuel hedge contracts in which the company engaged with. Over the two primary variables, namely substantial value and hedging ratio and Tobin’s Q variable (dependent), other five control variables have been selected for the regression analysis following the study of Aggarwal et al., (2012) and Kvello et al., 2009.
Size of the Company (SIZE) which is determined by the natural logarithm of the firm’s market value. What it should be expected is that small businesses tend to adopt more derivatives than larger counterpart, since they are more subject to highly financial distress costs.
Return on Total Assets (ROA) control variable is used to evaluate the potential firm’s effectiveness in using their assets. The ratio is calculated as Earnings Before Interest and Taxes (EBIT) over the total book value of assets.
CAPEX to Assets (CAPEX) variable, is used to measure growth firm’s opportunities. It is derived using as the nominator the as capital expenditures and as the denominator the total book value of assets. A high ratio will indicate that the company is facing short-term growth opportunities. From the association between hedging and CAPEX, it is expected that firms presenting a high growth fraction would benefit engaging in more hedging activities.
Cash to Sales (CTS) item is included in the model as it represents the capacity of a company in generating cash. Data about this variable are collected in Datastream looking at the cash flow o sales information. The more a firm can produce cash, the less will be its willingness in undertaking hedging strategies. What is expected here, it is a negative relationship between derivatives activities for hedge jet fuel price and cash to sales ratio.
of assets. What it is in here, is that companies that exhibit low DTA ratio, run into a lower probability of financial distress10.
The first model, reflecting the early hypothesis calculated as panel data Ordinary Least Square multivariate regression.
The sample consists of 76 airline companies over eight-year period from 2008 to 2016. Model 1:
TQijt=b0 + b1 HEDGEijt + b2 SIZEijt + b3 ROAijt + b4 CAPEXijt + b5 CTSijt + b6 DTAijt + µi + eijt, (1)
where eijt ~ i.i.d N(0,s2)
where i indexes the industry, j the company and t indexes years since the study is using panel data. All variables are reported in Table 2. of the appendix, µi denotes firm-fixed effects (which remove
the bias of the random effects model) and e is the error term which is independent and identically distributed random variable. For this model, is assumed a linear relationship between the firm’s value (dependent variable) and the covariates. Test for residual independency running Durbin-Watson test so to detect possible autocorrelation among residuals. Moreover, it is significant to investigate on heteroscedasticity using White's analysis. Lastly, prove for the goodness of fit through Jarque-Bera test to see whether airline sample has skewness and kurtosis correspond to a normal distribution. To study the second hypotheses, hence impact of oil prices on the stock return of the air-transport companies, the thesis relies upon daily time series date since it will analyze and compare the variables based on different companies’ stocks return as well as fluctuation at oil prices only once within a more extensive timeframe from 1st January 1991 to 31st December 2016. Bestow to data availability, this is the largest rage of time.
In this context, it is of relevant importance to note that changes in the price of oil are partially deterministic (petroleum reserves, oil supply and demand, OPEC policy decisions) and in part somewhat random and unpredictable ("oil commodity" as financial assets exchanged in financial instruments). There are several types of oil that can be used11, the study will use crude oil spot prices
data from West Texas Intermediate (WTI) since it is recognized as one of the primary markers and it
10 See Table 2. in the appendix for descriptive statistics.
11 Brent Blend is the most widely used market of all since 2/3 of all crude oil contracts worldwide
is conducted in Phan et al., (2014) study. This research takes into consideration oil spot prices daily returns calculated with the following formula:
ROILt =ln(xt/xt-1)
where xt mean oil price at day t and xt-1 represent oil price at day t-1 (Kristjanpoller & Concha, 2016).
Moreover, it is added the market return risk-free rate (RMktRF) from Fama-French website12 as control variable used by El Hedi Arouri, (2011) where the RMktRF equals the slope of the security market line (SML), the graphical representation of the capital asset pricing model (CAPM). Normally it is useful for investors in evaluating a security for inclusion in an investment portfolio in terms of whether the security offers a favorable expected return against its level of risk. Expectations support the outcome in which RMktRF is positively related to airline firm stock returns, meaning that all
stocks are properly priced or overvalued. Vice versa, if market return risk free rate is negatively related to airline stock price return, it means that the stocks are undervalued.
Relying on what Aggarwal et al. say in their research, changes in oil prices are measured using the spot price and not the futures prices.
From the literature, we have seen that fluctuation in oil prices might impact the stock return of airline companies, because of the demand for fuel. The research will attempt to model, measure and evaluate the volatility of daily spot oil price return and its impact on airline stocks return, using an E-GARCH model. Oil price information are daily time series data over January 1991 to December 2016 as Narayan and Sharma (2011) did in their research to capture proper stock returns heteroscedasticity. Model 2:
RAIRit = a0 + a1 ROILit + a3 RMktRFit + eit (2)
where eit ~ i.i.d N(0,s&4 )
RAIR is the dependent variable for airline stock return, ROIL the independent variable for the daily spot oil price return and RMktRF the market return risk free. All the descriptive statistics are reported in Table 5. of the appendix. The indexes i and t indicates the industry and the time, respectively since
12 Data are available at Kenneth R. French – Data Library
data will be treated as pooled time series for this model. The error term e is independent and identically distributed.
The E-GARCH main equation (3) is the following:
ln (σ2t) = ω + βln (σ2t-1) + ϒ *567
85679 + α *567
85679
− 4
; + q1 ROILit+q2 RMktRFit+ eit (3)
where, σ2t = α0 + α1 𝑢4&0>+ ϒ Dt 𝑢&0>4 + β σ2t-1
with Dt = 1 if 𝑢&0> < 0 and Dt = 0, otherwise.
In this case, an E-GARCH model is preferred over the GARCH(1,1) and the model (2) is associated with the second hypotheses.
The specification of E-GARCH model, proposed by Nelson in 1991, arose following the derivation of estimation procedures that have imposed the positivity of the parameters. The main characteristics of this type of representation presented by Brooks (2014) are:
1. The impossibility of obtaining a negative variance, without the need to impose any restriction on the parameters;
2. The presence of asymmetry for the reaction of volatility to positive or negative shocks; 3. The possibility of measuring an asymmetric effect proportional to the extent of innovations. The model is specified regarding the logarithm of the conditional variance, and the exponential transformation ensures the non-negativity of the variance. Since the expression has an autoregressive term, the coefficient 1 captures the presence effect of volatility, the stationarity is provided by the condition 0 < 1 <1 and its size will determine how fast the absorption of past shocks will be. The second parameter of the expression is a causal variable with an average of 0 if the standardized innovations are typically distributed. This permits to consider the possibilities of an asymmetric reaction proportional to oil price changes, which is what we want to capture in the second hypothesis. The effect measured by the term expressed in absolute value is apparently independent of whether the shocks are positive or negative. The asymmetric impact is captured by the last element of the expression, this type of specification is very refined because the term at-1 can assume any sign and the exponential transformation still maintains the positive variance.
the E-GARCH response is more sensitive, as exponential unification dominates. This characteristic will suit the volatility of oil prices increases and decreases.
The model will consider a parsimonious representation of the conditional variance of returns. Indeed, it consists in just three terms: ARCH, asymmetry term, and GARCH (Brooks, 2014). Data should be treated as panel since the study expects an investigation of multiple firms over multiple years, but tools of calculus as Eviews, Stata or GRETL do not allow this setting. This will be valid also for the third and fourth hypotheses.
Analyzing the third hypotheses, the thesis will focus on differences of this impact in countries with different income, classified by the world bank for the current 2018 fiscal year. Countries will be divided into two groups namely Low-income economies/Lower-middle-income economies (LI) and upper-middle-income economies/high-income economies (HI). To test the third hypotheses, the following model is proposed:
Model 3:
RAIRit = a0 + a1 ROILit + a3 RMktRFit + a4 INCOMEit + eit (4)
where eit ~ i.i.d N(0,s&4 )
Equation (4) presents the same variables as for the second model but it adds the variable INCOME which represent the high income/high middle income country classification (HI) and low medium income/low income (LI) which are set as dummies 0 and 1 respectively. The error term is
represented by e which is independent and identically distributed. The data will be treated as pooled and not panel data because of limitation in tools of calculus i indexes the industry and t the time. A high relative fraction means a more significant profitability. From this relationship, I expect top profitable companies to incur less in hedging techniques than corporations showing low-profit ratio and comparable inferior earnings.
The E-GARCH main equation is:
ln (σ2t) = ω + βln (σ2t-1) + ϒ *567
85679 + α *567
85679 − 4
; + q1 ROILit +q2 MktRFit +q3 INCOMEit+ eit
(5) where, σ2t = α0 + α1 𝑢&0>4 + ϒ Dt 𝑢&0>4 + β σ2t-1
To add a further dimension, the study will also measure how oil price changes differently affect countries with different level of resource endowment. The sample of airline companies will be grouped in oil-rich and oil-poor countries. This classification is extracted from the Organization of the Petroleum Exporting Countries (OPEC)13 Annual statistical bulletin. The following model (4) is proposed:
Model 4:
RAIRit = a0 + a1 ROILit + a3 RMktRFit + a4 OILENDit + eit (6)
where eit ~ i.i.d N(0,s&4 )
In this model, OILEND represent the oil endowment country characteristics that is oil rich (OR) oil poor (OP) represented as dummies 0 ad 1 respectively. The error term is represented by e, independent and identically distributed. The data, as for the third model, will be treated as pooled and not panel data, that is because tools of calculus do not allow this setting. The expectations foresee a higher impact on airline companies stock return operating in oil poor countries than oil rich countries. The respective E-GARCH main equation is:
ln (σ2t) = ω + βln (σ2t-1) + ϒ *567
85679 + α *567
85679 − 4
; + q1 ROILit +q2 MktRFit +q3 OILENDit+ eit
(7) where, σ2t = α0 + α1 𝑢4&0>+ ϒ Dt 𝑢&0>4 + β σ2t-1
with Dt = 1 if 𝑢&0> < 0 and Dt = 0, otherwise.
4. Descriptive statistic and Results
Table 2. of the appendix contains the descriptive statistics of the data relative to all variables used to analyze the first hypotheses. Before use the data directly collected from the financial statements (hedge percentage) and Datastream, an accurate process of winsorization was made in order to deal with univariate outliers. To detect and manage outliers was used the SPSS platform for statistical analysis and all data were winsorized at 5% to exclude the present extreme values (outliers). From
13 Information available at http://www.opec.org/opec_web/en/data_graphs/330.htm Last accessed
the descriptive statistic, it is possible to notice that the independent variable HEDGE has a wide range that goes from 5% to 95%. This can be attributed to the highly volatility of oil prices. For example, during the financial crisis, indeed from the annual reports most of the companies were hedging a higher percentage of fuel consumption during the years 2008, 2009, 2010. It is known that in these years, oil prices decrease and it is assumed that airline companies welcomed these cheaper prices issuing hedging contracts with gasoline producers (higher fuel hedge percentages). While during recent years, companies hedge less or at all, due to a (relative) oil price stability.
Table 3. Reports results from panel data ordinary least square Tobin’s Q regression. By observing the table the hedging percentage (HEDGE) demonstrates a positive coefficient of 0.056 with a corresponding insignificant p-value 0.485. This is in line with the initial expectations because results state that airline companies who hedge have a higher value if measured by Tobin’s Q. Control variables as size present a coefficient of 0.009 and a p-value statistically significant at a 5% level, meaning that the size of the firm positively influence the firm value. This can be attributed to economies of scale. Large firms have lower costs of production compared to smaller firms. Moreover, hedging programs can be costly to start and manage, hence big firms holding more cash, in theory, should be more likely to use derivatives in risk management than small firms. CAPEX is positively related and statistically insignificant (p-value 0.375), this confirm what expected at the beginning since results attest that firms would benefit in growth terms engaging in more hedging activities. The variable that affect negatively Tobin’s Q dependent variables is the return on assets (ROA) -0.085 with statistically insignificant p-value (0.540). This result is opposed to what expected since profitable firms typically have higher values than start-up for example which could present low ROA and negative cash flows but they could also have higher profitability launching a new product in the market. Cash to sales (CTS) variable present a positive relationship with firm value proxy Tobin’s Q (0.129) and insignificant p-value (0.236), again opposite to previews expectations since the more a firm can produce cash, the less will be its willingness in undertaking hedging strategies.
As regard debt to assets (DTA) control variable, it exhibits a positive relationship with firm value 0.063 and statistically insignificant p-value (0.649) confirming the ex-ante forecasts since more debt incentive airline companies to hedge more because of a larger amount of distress costs.
a positive relationship between jet fuel hedging and airline firm value. The adjusted R squared 0.832 indicates a high goodness of fit, hence a high reliability of results.
Table 3. Regression summary: Jet fuel hedging percentage and firm value
Variable Coefficient
Constant - C 0.688*** (0.081)
Fuel Hedge Percentage - HEDGE 0.056
(0.037)
Size of the Company - SIZE 0.009**
(0.005)
Return on Total Assets - ROA -0.085
(0.241) Capital Expenditure to Assets - CAPEX (0.262) 0.066
Cash to Sales - CTS (0.152) 0.129
Debt to Assets - DTA 0.063
(0.118) Number of observations: 305 F-statistic: 11.413*** Durbin-Watson statistic: 0.427 R-squared: 0.912 Adjusted R-squared: 0.832
Notes: This table report all the variables used in model (1) to test the first assumption. Dependent variable: Tobin’s Q proxy for firm value (TQ). Independent variables: Fuel hedge percentage (HEDGE). Control variables: SIZE, ROA, CAPEX, CTS, DTA. For each variable is reported the coefficient value, standard error, t-statistic and p-value. F-statistic under which all coefficient related to the independent variables, excluding the constant, is equal to zero. P-values smaller than 0.01, 0.05, 0.10 are indicated by ***,**,* , respectively.
Table 4. reports the results from the second hypothesis tested with E-GARCH model to detect the different magnitude of oil price impact on airline companies stock returns. The objective is to allow them for the possibility that negative and positive shocks have a different effect on the stock returns. All descriptive statistics relative to the following hypotheses are in Table 5. in the appendix. Focusing on coefficient variables it can be noticed that values are in line with the expectations. The return on oil is negatively related to airline stock market return, hence results are consistent to what found by Narayan and Sharma (2011) and later Aggarwal et al. (2012), as oil price increases, airline firm stock returns decrease and, if the oil price decreases, exposure to fuel risks rise by oil price declines. Therefore, airline companies stock returns are negatively affected by oil price changes.
(***). The most important parameter is the asymmetry term (ASYM.) which allows determining the proportional firm stock return reaction proportional to oil price changes. The term is positive -0.057, meaning that the variance goes down and the negative oil price has a greater impact on airline stock return than a positive oil price of the same magnitude. This result is in concordance with the earlier expectation and what has been found in the existing literature by Phan, et al. (2014) which predict a heavier impact on stock return with a decrease in oil prices. Individually all p-values are statistically significant at any conventional level except for the market return risk free that with a p-value equal to 0.983 is said to be statistically insignificant. Looking at the Durbin-Watson statistics (2.001) it is possible to state that the second assumption is not rejected. Since DW coefficient is @ 214 there is no autocorrelation.
Table 4. The different impact of oil price increase/decrease on airline stock returns
LOG(GARCH)=C(4)+C(5)*ABS(RESID(-1)/@SQRT(GARCH(-1)))+C(6)*RESID( 1) /@ SQRT (GARCH(-1)) + C(7)*LOG(GARCH(-1)).
Variable Coefficient
Constant - C 45.796***
(0.220)
Airline Stock Return - RAIR -0.021*** (0.003)
Market Return Risk Free - RMktRF 0.386
(18.224)
Arch term - ARCH 1.299*** (0.176)
Asymmetry term - ASYM -0.057*** (0.006)
GARCH term - GARCH 0.834***
(0.022) Number of observations: 116,473
Durbin-Watson statistic: 2.001 Adjusted R-squared: 0.044 R-squared: 0.046
Note: Table 4. reports the results from pooled data airline companies stock return affected by oil price fluctuations from model (2). The equation and outcomes consider the constant, the airline stock return (RAIR), the market return risk free rate (RMktRF) and the three parameters of EGARCH model in the variation equation section namely, ARCH term, asymmetric term and GARCH term. The asymmetric term is the most important because let positive and negative shocks varies according to their magnitudes. P-values smaller than 0.01, 0.05, 0.10 are indicated by ***,**,* , respectively.
significant. The coefficient for the income level variable, demonstrate a positive output (0.288) which is in line with the expectations since oil price increase has more impact on airline companies stock return operating in developing countries than on airline companies’ stock returns operating in developed countries. Low-income level countries would suffer most from higher oil price because their economies present a more depended behavior on imported oil and its use is less efficient than developed countries. Looking at the RAIR variable (-0.021), we can state that results are consistent with what found by Narayan and Sharma (2011) and later Aggarwal et al. (2012). Oil price increase have a negative effect on airline stock returns. In this case, the market return risk-free is positive (0.446) and statistically insignificant (0.980). The Durbin Watson statistic also in this case equals 2.001, hence we do not reject the third hypotheses. Adjusted R square is close to 0.05, which means that the regression model fits the data only at the 5%.
Table 6. The different impact of oil price increase/decrease on airline stock returns operating in developing/developed countries.
Variable Coefficient
Constant - C 45.742*** (0.230)
Airline Stock Return - RAIR -0.021*** (0.003)
Market Return Risk Free - RMktRF (18.230) 0.446
Developing/Developed Income Level - INCOME 0.288
(0.348)
Arch term - ARCH 1.301***
(0.177)
Asymmetry term - ASYM -0.057*** (0.006)
GARCH term - GARCH 0.834*** (0.022)
Number of observations: 116,473 Durbin-Watson statistic: 2.001
R-squared: 0.046 Adjusted R-squared: 0.044
Note: Table 6. reports the results from pooled data airline companies stock return operating in developing and developed countries affected by oil price fluctuations from model (3). The equation and outcomes consider the constant, the airline stock return (RAIR), the market return risk free rate (RMktRF), INCOME dummy to denote developing and developed countries and the three parameters of EGARCH model in variance equation section namely, ARCH term, asymmetric term and GARCH term. The asymmetric term is the most important because let positive and negative shocks varies according to their magnitudes. P-values smaller than 0.01, 0.05, 0.10 are indicated by ***,**,* , respectively
significance level. The return market risk-free rate is positive (0.695) with a corresponding significant probability. Introducing the resource endowment variable, the output coefficient demonstrate a negative relationship with oil price increase (-2.298) which is not in line to the expectations, since oil price changes have more impact on airline companies stock returns operating in oil-rich countries than on airline companies stock returns operating in oil-poor countries. This can be justified saying that on average, oil resource-rich countries are often the cash poor ones (eg. Venezuela), hence an increase in oil price has greater impact on oil-rich countries since these have less strong economic resources (Rajan, 2011). Once again, the asymmetry term is negative -0.057, meaning that the variance goes down and the negative oil price has a greater impact on airline stock return than a positive oil price of the same magnitude, which is what was logically expected at the beginning. Durbin-Watson stat allows for the non-rejection of the hypotheses (2.001). ARCH and GARCH terms are both positive with 1.300 and 0.834 outcome coefficients respectively and, statistically significant p-values (0.000). The adjusted R squared coefficient 0.057 ensure a 5.7% ability to explain the dependent variable.
Table 7. The different impact of oil price increase/decrease on airline stock returns operating in oil rich/oil poor countries.
Variable Coefficient
Constant - C 48.008***
(0.689)
Airline Stock Return - RAIR -0.021***
(0.003) Market Return Risk Free - RMktRF 0.695*** (18.229)
Oil Endowment Country - OILEND (0.680) -2.298
Arch term - ARCH 1.300*** (0.176)
Asymmetry term - ASYM -0.057*** (0.006)
GARCH term - GARCH 0.834***
(0.022) Number of observations: 116,473
Durbin-Watson statistic: 2.001
R-squared: 0.057 Adjusted R-squared: 0.054
5. Limitations and Further researches
This thesis presents several limitations in each of the hypothesis. In fact, when controlling for superior value creation for airline companies that hedge against jet fuel prices, the analysis relies only on Tobin’s Q proxy for firm performance. A more reliable study should contain other firm value variables as return on equity and return on assets. As regard the second, third and fourth hypothesis the major limitation is that it was not possible to test the data from variables presents in the model as panel data. Further research could use more advanced tools to test the same assumption in more precise, detailed and reliable outcomes. Again, the hypotheses tested with EGARCH model lack of control variables. Therefore, use other models as cumulative abnormal returns (CAR) by Aggarwal et al., (2012) or FF-Carhart model which detect common macroeconomic risks and it includes in the model factors as excess return on small capital stocks (SMB), excess return of low price-to-book ratio stocks (HML) and the momentum factor (Mom).
6. Conclusions
This thesis throughout investigated on four hypotheses regarding the relationship between jet fuel hedging activities and firm value and the impact of oil price on the specific sector of the airline industry.
Initially, the study was introduced and grounded in theoretical expectations and prior empirical findings, whose function was to represent a “benchmark” for early expectations and final results. Different methods have been used to assess the hypotheses. Data were collected through a combination of two databases: Orbis, to create the population and Datastream to gather all information necessary for the variables used in all four models.
Tobin’s Q proxy for firm performance was used for the first hypothesis on whether fuel hedging creates value for airline companies, findings report a positive relationship between jet fuel hedging and airline firm value. Moreover, the hedging intensity differs over time form many airlines.
References
Adetunji Babatunde, M., Adenikinju, O., & Adenikinju, A. F. (2013). Oil price shocks and stock market behaviour in Nigeria. Journal of Economic Studies, 40(2), 180-202.
Aggarwal, R., Akhigbe, A., & Mohanty, S. K. (2012). Oil price shocks and transportation firm asset prices. Energy Economics, 34(5), 1370-1379.
Apergis, N., & Miller, S. M. (2009). Do structural oil-market shocks affect stock prices?. Energy
Economics, 31(4), 569-575.
Basher, S. A., & Sadorsky, P. (2006). Oil price risk and emerging stock markets. Global finance
journal, 17(2), 224-251.
Brooks, C. (2014). Introductory econometrics for finance. Cambridge university press.
Carter, D. A., Rogers, D. A., & Simkins, B. J. (2006). Does hedging affect firm value? Evidence from the US airline industry. Financial management, 35(1), 53-86.
Campbell, C. J. (1991). The golden century of oil 1950–2050: the depletion of a resource (Vol. 19). Springer Science & Business Media.
Chiu, C. S. (2012). Derivative and integral terminal sliding mode control for a class of MIMO nonlinear systems. Automatica, 48(2), 316-326.
Christoffersen, P., & Pan, X. N. (2017). The State Price Density Implied by Crude Oil Futures and Option Prices.
Dafir, S. M., & Gajjala, V. N. (2016). Fuel Hedging and Risk Management: Strategies for Airlines,
Shippers and Other Consumers. John Wiley & Sons.
Dick, H., Gupta, S., Vincent, D., & Voigt, H. (1984). The effect of oil price increases on four oil-poor developing countries: A comparative analysis. Energy Economics, 6(1), 59-70.
El Ghoul, S., Guedhami, O., & Kim, Y. (2017). Country-level institutions, firm value, and the role of corporate social responsibility initiatives. Journal of International Business Studies, 48(3), 360-385. El Hedi Arouri, M. (2011). Does crude oil move stock markets in Europe? A sector investigation. Economic Modelling , 1716-1725.
Elyasiani, E., Mansur, I., & Odusami, B. (2011). Oil price shocks and industry stock returns. Energy
Economics, 33(5), 966-974.
Erb, C. B., Harvey, C. R., & Viskanta, T. E. (1996). Political risk, economic risk, and financial risk. Financial Analysts Journal, 52(6), 29-46.
Fang, C. R., & You, S. Y. (2014). The impact of oil price shocks on the large emerging countries' stock prices: Evidence from China, India and Russia. International Review of Economics &
Ghosh, S. (2011). Examining crude oil price–Exchange rate nexus for India during the period of extreme oil price volatility. Applied Energy, 88(5), 1886-1889.
Hamilton, J. D. (2003). What is an oil shock?. Journal of econometrics, 113(2), 363-398.
Huang, B. N., Hwang, M. J., & Peng, H. P. (2005). The asymmetry of the impact of oil price shocks on economic activities: an application of the multivariate threshold model. Energy Economics, 27(3), 455-476.
Jobling, A., & Jamasb, T. (2017). Price volatility and demand for oil: A comparative analysis of developed and developing countries. Economic Analysis and Policy, 53, 96-113.
Jones, C. M., & Kaul, G. (1996). Oil and the stock markets. The Journal of Finance, 51(2), 463-491. Jorion, P. (2007). Financial risk manager handbook (Vol. 406). John Wiley & Sons.
Kilian, L. (2009). Not all oil price shocks are alike: Disentangling demand and supply shocks in the crude oil market. American Economic Review, 99(3), 1053-69.
Kilian, L., & Park, C. (2009). The impact of oil price shocks on the US stock market. International
Economic Review, 50(4), 1267-1287.
Kristjanpoller, W. D., & Concha, D. (2016). Impact of fuel price fluctuations on airline stock returns. Applied Energy, 178, 496-504.
Kvello, C., & Stenvik, H. N. (2009). Jet fuel hedging in the European airline industry: determinants
and value of hedging(Master's thesis).
MacCrimmon, K. R., Wehrung, D., & Stanbury, W. T. (1988). Taking risks. Simon and Schuster. Mork, K. A. (1989). Oil and the macroeconomy when prices go up and down: an extension of Hamilton's results. Journal of political Economy, 97(3), 740-744.
Morrell, P., & Swan, W. (2006). Airline jet fuel hedging: Theory and practice. Transport
Reviews, 26(6), 713-730.
Narayan, P. K., & Narayan, S. (2007). Modelling oil price volatility. Energy Policy, 35(12), 6549-6553.
Narayan P. K. & Sharma S. (2011). New evidence on oil price and firm returns. Journal of Banking & Finance 35 (12) 3253–62.
Narayan, P. K., & Sharma, S. (2014). Firm return volatility and economic gains: the role of oil prices. Economic Modelling, 38, 142-151.
Hamilton, J. D. (1983). Oil and the macroeconomy since World War II. Journal of political
economy, 91(2), 228-248.
Rahnema, A. (1990). An Overview of Exchange and Interest Rate Risk Management: How to Make
Risk Management Strategy a Competitive Weapon. Instituto de Estudios Superiores de la Empresa,
Rajan, S. C. (2011). Poor little rich countries: another look at the ‘resource curse’. Environmental
Politics, 20(5), 617-632.
Sadorsky, P. (1999). Oil price shocks and stock market activity. Energy economics, 21(5), 449-469. Treanor, S. D., Rogers, D. A., Carter, D. A., & Simkins, B. J. (2014). Exposure, hedging, and value: New evidence from the US airline industry. International Review of Financial Analysis, 34, 200-211.
Websites
U.S. Energy Information Administration, EIA (2016). How much oil consumed by the United States comes from foreign countries?
https://www.eia.gov/tools/faqs/faq.php?id=32&t=6 Kenneth R. French – Data Library
http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html#International Last accessed December 27th 2017.
IATA (2017). Economic performance of the airline industry, IATA 2017
http://www.iata.org/publications/economics/Reports/Industry-Econ-Performance/IATA-Economic-Performance-of-the-Industry-end-year-2017-report.pdf
Indexmundi (2016). Jet Fuel vs Crude Oil (petroleum) - Price Rate of Change Comparison - IndexMundi
https://www.indexmundi.com/commodities/?commodity=jet-fuel&months=120&commodity=crude-oil
Statista.com (2016). Top oil consuming sectors worldwide
https://www.statista.com/statistics/307194/top-oil-consuming-sectors-worldwide/ World Bank income level classification
Appendix
Table 1. Sample of airline companies
Airline company Home Country ISIN Developed/Developing
country
Oil rich/Oil poor countries
ABU DHABI AVIATION UAE AEA001001014 0 0
ACE AVIATION HOLDINGS Canada CA00440P4096 0 0
AIR ARABIA UAE AEA003001012 0 0
AIR ASIA CO., LTD. Taiwan TW0002630001 0 1
AIR BERLIN Germany GB00B128C026 0 1
AIR CANADA Canada CA0089118776 0 0
AIR CHINA LTD China CNE1000001S0 0 1
AIR FRANCE - KLM France FR0000031122 0 1
AIR MAURITIUS LTD Mauritius MU0010N00002 0 1
AIR NEW ZEALAND New Zeland NZAIRE0001S2 0 1
AIRLINE KYRGYZSTAN Russia KG0101172417 0 1
ALASKA AIR GROUP USA US0116591092 0 1
ALIA - THE ROYAL
JORDANIAN AIRLINE Jordan JO3121311018 1 1
AMERICAN AIRLINES USA US02376R1023 0 1
ANA HOLDINGS Japan JP3429800000 0 1
ASIA AVIATION Thailand TH3437010004 0 1
ASIANA AIRLINES South Korea KR7020560009 0 1
AZUL ADR 1:3 Brasil US05501U1060 0 1
BALTIA AIR LINES USA US0588231052 0 1
BANGKOK AIRWAY Thailand TH4403010002 0 1
BERMUDA AVIATION SERVICES LIMITED
British Overseas
Territories BMG103351000 0 1
BRITISH AIRWAYS UK GB0056794497 0 1
CATHAY PACIFIC AIRWAYS Hongkong HK0293001514 0 1
CEBU AIR Philippine PHY1234G1032 1 1
CHINA AIRLINES China TW0002610003 0 1
DELTA AIR LINES USA US2473617023 0 1
DEUTSCHE LUFTHANSA Germany DE0008232125 0 1
EASYJET UK GB00B7KR2P84 0 1
EL AL ISRAEL AIRLINES
LTD Israel IL0010878242 0 1
ENTER AIR S.A. Poland PLENTER00017 0 1
EVA AIRWAYS Taiwan TW0002618006 0 1
FASTJET Tanzania GB00BWGCH354 1 1
FINNAIR Finland FI0009003230 0 1
GRUPO AEROMEXICO Mexico MX01AE010005 0 1
HAINAN AIRLINES
HOLDING CO., LTD. China CNE000000RT0 0 1
HAWAIIAN HOLDINGS Hawaii US4198791018 0 1
ICELANDAIR GROUP Iceland IS0000013464 0 1
JAGSON AIRLINES India INE685B01018 1 1
JAPAN AIRLINES Japan JP3705200008 0 1
JAZEERA AIRWAYS Kuwait KW0EQ0602452 0 0
JAZZ AIR/CHORUS
AVIATION Canada CA17040T2011 0 0
JEJUAIR South Korea KR7089590004 0 1
JET AIRWAYS India INE802G01018 1 1
JETBLUE AIRWAYS USA US4771431016 0 1
JUNEYAO AIRL.'A' China CNE100001ZY0 0 1
KENYA AIRWAYS Kenya KE0000000307 1 1
KINGFISHER AIRLINES
LIMITED India INE438H01019 1 1
KOREAN AIR LINES Republic of Korea KR7003490000 0 1
LATAM AIRLINES GROUP Chile CL0000000423 0 1
MALAYSIAN AIRLINE SY. Malaysia MYL3786OO000 0 1
NOK AIRLINES Thailand TH4601010002 0 1
NORWEGIAN AIR SHUTTLE Norway NO0010196140 0 1
PAKISTAN INTL.AIRLINES Pakistan PK0003401012 0 1
PEGASUS HAVA
TASIMACILIGI A LTD. Turkey TREPEGS00016 0 1
QANTAS AIRWAYS Australia AU000000QAN2 0 1
RYANAIR HOLDINGS Ireland IE00BYTBXV33 0 1
SAMARA AIRLINES Russia RU0005295194 0 1
SAS DANMARK A/S Danimark DK0010223775 0 1
SINGAPORE AIRLINES Singapore SG1V61937297 0 1
SKYWEST USA US8308791024 0 1
SPICEJET India INE285B01017 1 1
SPIRIT AIRLINES USA US8485771021 0 1
SPRING AIRLINES LTD China CNE100001V45 0 1
STAR FLYER Japan JP3399320005 0 1
THAI AIRWAYS INTL. Taiwan TH0245010002 0 1
TIGER AIRWAYS HDG. Australia SG1Z26952619 0 1
UNITED AIRWAYS USA BD0001UTDAR9 0 1
VIETJET AVIATION Vietnam VN000000VJC7 1 1
VIRGIN AUSTRALIA HDG. Australia AU000000VAH4 0 1
WESTJET AIRLINES Canada CA9604105044 0 0
WIZZ AIR HOLDINGS Hungary JE00BN574F90 0 1
Table 1. includes all airline companies used for the study of the hypotheses, their home country, ISIN numbers used to
derive variables from Datastream database and their classification as developed/developed countries classified with 0 and one respectively and oil rich/oil poor countries classified with 0 and 1 respectively.
Table 2. Descriptive statistics Hypotheses 1.
Table 2. reports the descriptive statistics for all variables in hypotheses 1, namely Tobin’s Q ratio which is the proxy for
the firm performance (TQ), hedge percentage issued by airline annual reports (HEDGE), size of the airline companies which is measured with the natural logarithm of the single firm’s market value (SIZE), the return on assets of an individual airline companies is measured as EBIT over assets of the firm (ROA), CAPEX represents the opportunities of growth for the airline company and it is measured as the ratio of capital expenditures to assets of the firm, Cash to Sales ratio is used to measure the capacity of an airline company to generate cash and Debt to assets determines the operations from external funding. It is the ratio of long-term debt to assets of an airline company.
TQ HEDGE SIZE ROA CAPEX CTS DTA
Table 5. Descriptive statistics Hypotheses 2, 3 and 4
Table 5. reports the descriptive statistics for all variables in hypotheses 2, 3 and 4 namely RMktRF which is the market
return risk free rate, the return on oil price (ROIL), the airline stock returns (RAIR), INCOME which is the dummy relative country characteristics developed/developing countries (0 and 1 respectively) and, OILEND which is the dummy for oil rich/oil poor countries (0 and 1 respectively).
RMktRF RAIR ROIL INCOME OILEND
