Revenue Management in the Airline Industry

(1)

Revenue Management in the Airline Industry

Measuring the contribution of dynamic pricing to airline performance

Alice van Riel

(2)

2

Name: A.J.M. van Riel

Student number: S2280310

Supervisor: Dr. J.W.J. de Kort

University of Groningen

Faculty of Economics and Business Administration

Abstract

This thesis studies the gap in literature with regard to the effect of dynamic pricing on firm performance in the airline industry. The question raised in this paper is whether airlines do perform better due to the use of dynamic pricing techniques or that the negative effects of revenue management diminish the positive effects. To answer this question a conceptual model has been derived from academic literature and a hierarchical multiple regression analysis has been performed. The conceptual model includes airline performance drivers for which the effects on firm performance are eliminated in the first block of the hierarchical regression. Subsequently we analysed whether dynamic pricing tactics are more effective for low-cost carriers then they are for full-service carriers. The main finding of this thesis is that there is no evidence that dynamic pricing is able to predict a significant amount of the variance in EBITDA Margin above and beyond the control variables. We also found no evidence that the effect of dynamic pricing on firm performance for low-cost carriers is different from the effect for full-service carriers. The results indicate that the non-significant relation between firm performance and dynamic pricing is negative. Our framework learns that this could be a result of the spiral-down effect. In addition a significant negative relation is found between dynamic pricing and customer satisfaction.

Keywords

(3)

3

Acknowledgements

This work would not have been possible without the advice and support of many people. First of all I would like to thank Dr. J.W.J de Kort for his guidance and good advice. My thanks must also go to Klaas Jongsma, who, in spite of having practically no spare time, still managed to find time to provide help and advice on SPSS. Not to forget Henk Otten, for critically evaluating the work. Finally, I would like to thank my family and friends for all their valuable support.

(4)

4

1 Introduction

1.1 Statement of the problem

While travelling on an airplane, have you ever asked the people sitting next to you what they paid for their flight? Most probably they paid a different price then you did. Welcome to the world of dynamic pricing in which totally different prices are paid for perfectly identical products depending on demand at the moment in time.

Dynamic pricing is one of several revenue management techniques and is considered to help airlines to optimize revenue. Dynamic pricing comprises management of airline seat inventory prices to get the most out of performance on the long run. Summarized: Dynamic pricing involves selling the correct product to the correct buyer at the correct moment at the correct price by means of demand data gathering, demand modelling, demand prediction and pricing optimization. The emphasis of this thesis will be on demand-based pricing. When demand is low, lower rates are quoted. When demand surges, higher rates are offered. Although the principle is not innovative, its widespread use today has been facilitated by broadband internet networks.

There are many points of view in favour and against dynamic pricing. Although revenue management is expected to result in higher revenue, competitive forces could lead to a so-called spiral-down effect in which prices get lower and lower as competitors try to sell below competition. Additionally, the notion of consistency by customers is an important issue. Customer satisfaction may be reduced if airlines’ service quality, setting, brand, and facilities price is all of a sudden much less or much more just because demand changed within a few hours.

(8)

8 1.2 Purpose

The focus in this thesis is on the gap in literature with regard to the impact of dynamic pricing strategies on firm performance. Do airlines perform better due to the use of dynamic pricing techniques or do the negative effects of revenue management diminish the positive effects on revenue? To answer this question two regression analyses will be performed. The first analysis will include a mixture of low-cost carriers and full-service carriers. The second analysis will make a distinction between the effects of dynamic pricing on firm performance for low-cost carriers and full-service carriers in order to analyse the differences.

The thesis will contribute to the field of knowledge in the area of revenue management and dynamic pricing and will explore, challenge, and increase the understanding of the impact of pricing on the bottom line of airlines.

1.3 Significance of the study

This work focusses on the passenger segment of airlines. In this segment 2.000 carriers operate more than 24.000 airplanes providing service to over 3800 airports worldwide. The companies carry over 2 billion passengers each year. Every year the numbers grow with approximately 5%. Simultaneously many airlines struggle to survive. A large number of the major airlines have been making losses for many years. The industry is capital intensive and highly competitive. This thesis should help decision makers to understand the contribution of dynamic pricing strategies to firm performance and should help them to make better trade-offs on possible performance improvements.

1.4 Research question and hypothesis

1.4.1 Main questions:

(9)

9

2) Is there a difference between low-cost carriers and full-service carriers with regard to the effect of dynamic pricing on firm performance?

1.4.2 Sub questions

1. What is dynamic pricing?

2. What are the performance drivers of airline performance according to academic literature? 3. What are the negative and positive effects of dynamic pricing according to academic literature? 4. What are the fundamental differences between low-cost carriers and full-service carriers?

1.5 Structure of the thesis

(10)

10

2 Literature Review

Dynamic pricing is a subject which has received significant academic attention. In this section an extensive literature review will be provided on the subject. The first section will contain a short background on dynamic pricing systems. The second section will explain the difference between low-cost carriers and full-service carriers and their respective pricing strategies. The third section covers the positive effects of dynamic pricing on firm performance followed by a second part that will discuss some of the negative aspects. In the fourth section major determinants of airline performance are discussed. A conceptual model will be derived from this literature which will form the basis for a quantitative analysis in the subsequent chapters.

2.1 Dynamic pricing

Dynamic pricing, sometimes referred to as real-time pricing, is a method to set the price for a highly elastic product or service (Elmaghraby & Keskinocak, 2003). Its objective is to permit a business to alter is prices quickly in direct response to changes in market demand.

These alterations are usually organized by pricing robots which are software agents that collect records and use algorithms to correct prices in accordance with predefined business rules. These rules are generally based on the time of day, the day of the week, competitor’s pricing, current demand and specific customer characteristics like the customers’ geographical location. With the introduction of increased amounts of data and improved data analytics business rules for price modifications become more effective. Through data gathering and analysis sellers can forecast prices that customers are willing to pay and adjust prices accordingly. In the airline industry, this process results in different fares charged over time for seats on the same single flight (Elmaghraby & Keskinocak, 2003).

2.2 Low-cost carriers versus full-service carriers

(11)

11

base fare. An additional difference is the lower cost base of low-cost carriers. Low-cost carriers tend to save on labour cost and fly to cheaper airports (O’Connell & Williams, 2005; Williams, 2001). Traditionally low-cost carrier customers are tourists while most business travellers used to choose for full-service airlines. Low-cost carriers generally have only one class while full-service carriers tend to divide their seats into different classes. The most relevant difference relates to pricing structures. Low-cost airlines tend to have simple pricing structures that are very transparent. They adopt the so-called single fare mechanism. This includes a fare at any point in time based on the number of seats available. The price then controls demand. The prices are usually low when the departure date is far ahead and increases when time passes. Full-service carriers on the other hand sell bundles of products making it difficult for customers to value. In addition full-service carriers differentiate with regard to cancellation policies and seating classes. Traditionally full-service carriers controlled demand by seat allocation to the different classes. They established multi-fare mechanisms with pre-set fares (Kuancheng, Yi-Ya & I-Wen, 2013).

Although these differences seem very clear they are subject to change over time. Many low-cost carriers are beginning to offer more services on their flights and there are full-service carriers that adopt more aspects of the low-cost business model. This change is also visible in the pricing strategies (CAPA, 2013). Full-service carriers have started implementing low-cost carrier pricing strategies and started to dynamically adjust their prices based on demand.

Schrader & Constantinides (2013) found that full-service carriers modified their fares more often than low-cost carriers. Schrader & Constantinides performed an experiment in which they followed the prices of both low-cost carriers and full-service carriers for one month and four months prior to departure. They found that the full-service airlines, during this research period, executed 86% of all price adjustments that were observed. Mantin & Koo (2009) argue that when the full-service carriers are confronted with low-cost carrier competition they will likely embrace a more hostile high-low pricing strategy.

(12)

12

save costs too. They argue that price is currently probably the most important factor in choosing carriers. We expect that dynamic pricing has a greater effect on the performance of low-cost carriers because of the higher level of price awareness of their customers.

2.3 The positive effect of dynamic pricing

Several scholars did an attempt to determine the impact of revenue management and dynamic pricing. Şen (2013) studied the issue of selling a fixed inventory of objects over a predetermined sales period and looked for supplementary revenue increases due to dynamic pricing. Şen’s study put forward two dynamic heuristics that constantly bring up prices based on residual inventory and time in the sales period. The conclusion of the research is that dynamic pricing is a better option in practice. The study acknowledges that dynamic pricing has an effect on revenue improvements. However, although higher revenue is one of the ingredients of better firm performance, the study does not proof a relation between dynamic pricing and firm performance.

Kuancheng, Yi-Ya & I-Wen (2013) study the differences in approach between low-cost carriers that lean towards single-fare structures and dynamically alter fares on one hand and old-style airlines that in addition use multi-fare systems on the other hand. Based on their statistical model, Kuancheng, Yi-Ya & I-Wen (2013) find that dynamic pricing results in higher revenues.

Williams (2013) made a model of active airline pricing with price discrimination and dynamic adjustments to random demand. He finds that dynamic tuning to stochastic demand is above all useful as a way to safe seats for customers of high value that decide on flying at a late stage. It results in higher revenue improvements in comparison to pricing systems which depend on moment of buying but not on residual volume. This study confirms the belief that dynamic pricing contributes positively to revenue growth.

2.4 The negative effect of dynamic pricing

(13)

13

Burger & Fuchs (2005) study the financial consequences of a dynamic pricing approach for the airline industry taking the effect of competitors into account. Their findings indicate that the use of a dynamic pricing system has a neutral or increasing effect on revenues. Whether the effect is neutral or positive was depending on the pricing system of the competition and on the time horizon. When competition used the traditional revenue management system, the airline with the dynamic pricing system would outperform the other. The disadvantage of dynamic pricing exists in the situation where two opposing airlines are using dynamic pricing strategies, where they will push each other into lowering prices (spiral-down effect. Burger & Fuchs (2005) also show that dynamic pricing in situations of high market concentration does not always lead to more revenue.

Anderson & Blair (2004) sketch a performance monitoring tool to monitor the influence of applying dynamic pricing through a detailed analysis of the lost revenue opportunities of significant decisions in the past. The usefulness of dynamic pricing systems is still dubious Anderson and Blair (2004) state. Behavioural academics have researched customers' perceived price unfairness when experiencing dynamic pricing and its damaging influence on customer trust and future buying plans. In prior studies dissimilar price stages were used on a comparison base to scrutinize customers' reactions. Research results of many price fairness studies are rather comparable in the sense that shoppers experience price unfairness, suspicion, and negative (re)purchase plans when they pay a larger amount than others for identical merchandise.

Xia, Monroe & Cox (2004) argue that a pricing practice that is at first alleged to be unfair might gradually spread and develop into a novel standard that is acknowledged by most individuals and is not as much expected to be alleged as unfair. They mention the example of the airlines’ practice of dynamic pricing with yield management technology. As this practice turns into an acknowledged practice by most consumers it is more likely to be alleged as fair (Xia, Monroe & Cox, 2004). Hence, perceived unfairness of a price or method may well weaken over the years. Airfares unfairness may therefore be not as much of a problem anymore.

(14)

14

model to examine the practice through which airlines predict demand and enhance booking controls over a sequence of ﬂights. The study shows that dynamic pricing does not always positively contribute to performance.

Based on the first part of the literature review on the impact of dynamic pricing we find that a study on the impact of dynamic pricing on firm performance is worthwhile. Research on the positive effects of dynamic pricing has only explored possible revenue impacts. There is no evidence of dynamic pricing effects on bottom line performance. A negative effect of dynamic pricing is that it may cause a spiral-down effect on prices. In addition there is evidence that dynamic pricing may negatively affect customers’ perceived fairness.

2.5 Determinants of airline performance

To investigate the effect of dynamic pricing on firm performance we need to identify other factors that affect performance too in order to eliminate these effects in our analysis. If we control for the possible effects of other performance drivers, is dynamic pricing still able to predict a significant amount of the variance in performance? In the following section determinants of airline performance according to academic literature are identified. These are: fuel cost, labour cost, executive performance, market concentration, gross domestic product and customer satisfaction.

Banker & Johnston (1993) offer a model for cost driver examination in the U.S. airline industry. They detailed and projected a structure of cost functions with various cost drivers for the industry. They discovered that volumes as well as operation based cost drivers are statistically significant. Banker & Johnston (1993) also verified the latent managerial significance of the operations-based drivers by clarifying discrepancies in marginal costs across airlines in terms of operating tactics mirrored in the cost driver values. Banker & Johnston identify in their study fuel and labour as significant operations-based cost drivers of the airline industry. M.F. (2008) examines the effect of fuel price on airline operational performance. Fuel cost turned out to be a main factor in total operating costs.

(15)

15

Tarry (2014) found a link between developments in the airline industry and progress in the worldwide economy in 2014. The Conference Board forecasts international gross domestic product (GDP) to rise around 3.5% in 2014, while data also shows that the total amount of airline seats will grow with 4.8% during the year. The article implies that GDP has a significant effect on airline performance. We therefore choose to eliminate the effects of GDP in our model in order to measure the effect of dynamic pricing on firm performance.

Dresner, Dong & Steven (2012) examine relations between customer service, customer satisfaction, and firm performance for US airlines. They study these relations with a possible moderating effect of market concentration. Their main outcome is that market concentration weakens the linkage between customer satisfaction and airline profitability. The article confirms that customer satisfaction affects firm performance in a positive way. Therefore we will eliminate the effect of customer satisfaction in our regression model in order to be able to measure the effect of dynamic pricing. For the purpose of this thesis we will ignore the effect of market concentration as it is very difficult to quantify.

Schefczyk (1993) gives an overview of factors of high profitability and performance in the airline industry. These factors contain operational factors such as operational cost and revenues that include labour and fuel. Additional factors are marketing and sales activities.

Kumar, Johnson and Lai (2009) concentrate on problems that U.S. airlines need to tackle in order to control costs and improve operations. They made a diagram to demonstrate the relevant factors. Their diagram shows a cause and effect ﬁshbone figure of cost drivers including salaries, fuel and other operational costs.

Flouris & Walker (2005) studies the stock and accounting performance of three carriers in the United States after the terrorist attacks on September 11, 2001. They discovered that low-cost carrier Southwest's performance was better than that of full-service airlines Continental and Northwest and argued that Southwest's business model offers the firm considerably more financial and operational flexibility than full-service carriers. Southwest's lower operating costs and customer trust seemed to be two of the aspects that contribute to Southwest's achievement. This study confirms the importance of consumer trust for firm performance.

(16)

16

to evaluate airline performance. Behn & Riley (1999) empirically observed these metrics and discovered that on-time performance, mishandled baggage and in-flight service were related to their substitute variable for customer satisfaction. They found that customer satisfaction is related to operating income and revenues and that customer satisfaction is related to expenses.

2.6 Conclusion and Conceptual Model

The literature review in the previous chapters facilitates the creation of a conceptual model. Academic literature shows evidence for both positive as well as negative effects of dynamic pricing on firm performance. Literature showed that although low-cost carriers and full-service carriers traditionally used different pricing strategies they now adapt similar strategies. Literature identifies several drivers of firm performance that should be eliminated from the model in order to measure the direct effects of dynamic pricing on performance. From the literature four main airline performance drivers have been identified being: Customer Satisfaction (CS), Fuel Cost (FC), Gross Domestic Products (GDP) and Labour Cost (LC). All of these variables have - according to academic literature - significant impact on airline firm performance (EM) and need to be eliminated in order to measure the actual effect of dynamic pricing (DP) on firm performance.

Figure 1: Conceptual model

(17)

17

3 Methodology

This chapter will start with a discussion of the theory testing process including the relevant details of this study. It is followed by a definition of conceptual model variables.

3.1 Theory testing

The research approach in this thesis will be theory testing which focusses on the second part of the empirical cycle (Saunders, Lewis and Thornhill, 2007). This process follows the deduction, testing and evaluation steps of the empirical cycle. The outcome of our research process is insight in the effects of dynamic pricing on firm performance.

Figure 2: theory-testing process

After the literature review, the second step in our theory-testing research is to derive a conceptual model and hypotheses. In addition relevant variables must be identified to operationalize the conceptual model. A definition of the different variables will follow in the next section.

Effect of dynamic pricing on firm performance: No conclusive evidence in literature. Identification of important variables, generation of conceptual model and

hypotheses.

Data collection Statistical Analysis

Results Theoretical implications & future

(18)

18

3.1.1 Hypothesis

Based on the conceptual model we have defined two main hypotheses given below:

H_01: Dynamic pricing does not affect firm performance above and beyond the effect of customer

satisfaction, fuel cost, gross domestic products and labour cost.

H_11: Dynamic pricing does affect firm performance above and beyond the effect of customer

satisfaction, fuel cost, gross domestic products and labour cost.

H_02: Dynamic pricing does not correlate stronger with firm performance for low-cost carriers above

and beyond the effect of customer satisfaction, fuel cost, gross domestic products and labour cost. H_12: Dynamic pricing does correlate stronger with firm performance for low-cost carriers above and

beyond the effect of customer satisfaction, fuel cost, gross domestic products and labour cost.

3.2 Model Variables

In this section the variables that are used in the model will be defined.

3.2.1 Firm performance (EBITDA MARGIN, EM)

EBITDA Margin (EM) is selected to represent firm performance. EBITDA margin is defined as EBITDA/OPERATING REVENUE. EBITDA margin is a proper variable as it gives insight in the operational performance of the airline and ignores effects of taxes, depreciation, amortization and financing costs (Kaen & Baumann, 2003). EBITDA margin can offer shareholders a better understanding of a company's

core productivity. EBITDA margins for our sample airlines are extracted from Reuters DataStream. In our analysis we work with the mean EBITDA margin over the last three years in order to eliminate possible extraordinary performance results in the last reporting periods.

3.2.2 Dynamic pricing (DP)

(19)

19

separate airlines into ones that use dynamic pricing and airlines that do not use it. We assume that all airlines practice dynamic pricing to a certain extent. The objective of the thesis is finding evidence for the positive effects of price deviation on airline performance.

In literature dynamic price dispersion can be described as the ‘variations of daily ticket prices across the histories of fares’ (Mantin & Koo, 2009). There have been a number of recommended methods to measure price dispersion. Examples are the coefficient of variation, the price gap between the two lowest prices and the range of prices between the maximum and minimum tariffs (Mantin & Koo, 2009). In this study we will use the coefficient of variation (CV), which can be defined as the proportion of the standard deviation of a ticket (σ) to the mean of a ticket (µ). We will use the absolute value of the CV, which is also known as relative standard deviation and which is stated as a percentage.

The formula is as follows:

= Where:

CV = Coefficient of variation for specific ticket [%]

σ = Standard deviation for a specific ticket over a two week period [-] µ= Mean for a specific ticket over a two week period [-]

The formula for calculating the standard deviation is as follows:

= ₋1

Where:

= Ticket price over time

σ = Standard deviation for a specific ticket over a two week period [-] µ = Mean for a specific ticket over a two week period [-]

(20)

20

3.2.3 Customer satisfaction (CS)

Customer Satisfaction (CS) is measured based on data from airlinequality.com. This website contains a comprehensive airline rating system established in 1999. The rating is built on a detailed quality examination of more than 800 different items across an airline's front-line product and staff service standards, applied to the Airport and Cabin Service environments. A maximum of 5 stars are assigned to airlines depending on their quality. The star rating of airlinequality.com will be converted into a percentage (e.g. 2 stars = 2/5 = 40% = 0.4)

3.2.4 Fuel Cost (FC)

Fuel costs (FC) per airline are derived from the profit and loss section of annual reports. The fuel costs are calculated as a percentage of operating revenues in order to measure the proportional effect of fuel expense in the operations of the airline.

=_{3 $%#
&}3 !"#

3.2.5 Gross domestic product per capita (GDP)

GDP per capita is calculated as the gross domestic product divided by midyear population. GDP is the sum of gross value added by each resident producer in the economy plus any product taxes and minus any subsidies not included in the value of the products. It is calculated without making deductions for depreciation of fabricated assets or for reduction and deprivation of natural resources. It is presented per country. Data are in current U.S. dollars and derived from data.worldbank.org

3.2.6 Labour Cost (LC)

Labour costs (LC) per airline are derived from the profit and loss section of annual reports. The labour costs are calculated as a percentage of operating revenues in order to measure the proportional effect of labour expense in the operations of the airline.

(21)

21

4 Data collection and analysis

This chapter starts with explaining the hierarchical multiple regression technique. Thereafter validity and reliability issues are discussed followed by a description of the sample selection and sample size. The section finishes with a description of some of the most important assumptions that underlie this technique.

4.1 Hierarchical regression

Hierarchical regression is used to assess the relationship between a set of independent variables and the dependent variable while controlling for the effect of a set of control variables (Gelman & Hill, 2007). We choose for the hierarchical multiple regression technique because we expect a possible relation between customer satisfaction and dynamic pricing. Hierarchical multiple regression permits us to use a fixed order of entry for independent variables in order to control for the effects of anticipated covariates. In our study the dependent variable is EBITDA Margin (EM), the independent variable is dynamic pricing (DP) and the control variables are Gross domestic product (GDP), labour cost (LC), fuel cost (FC) and customer satisfaction (CS).

In our hierarchical multiple regression independent variables are entered in two blocks into the analysis. One block contains the control variables and the other block contains these control variables including the variable dynamic pricing (DP). In general our focus is on the change in R² resulting for entering the next block. A significant increase of R² after adding dynamic pricing into the equation implies that the variable dynamic pricing contributes to the explanatory power of the variance in EBITDA margin above and beyond the effect of the control variables. The null hypothesis is that the change in R² is zero, implying no contribution to the explanation power of the variance of dynamic pricing. If the null hypothesis is rejected, then our interpretation indicates that dynamic pricing may have a relationship to EBITDA Margin, above and beyond the effects of the control variables.

4.2 Reliability and Validity

(22)

22

In order to increase internal validity we defined all variables as described in the previous section. The objective is to measure averages of multiple years in order to average out incidental extremes. Cost and margins are calculated according to different accounting standards. This could result in undesired differences. We rely on the quality of Reuters DataStream standardized reporting format to minimize this effect.

In order to increase external validity we have used an index of multiple flights at multiple times for a large variety of airlines from different countries. This should safeguard external validity.

4.2.1 Selection of sample

Data is derived from the International Air Transport Association (IATA). For the purpose of this thesis a selection is made of the 73 largest airlines in the world measured in revenue. This selection method results in a bias as only large commercial airlines are included in the sample. Conclusions may therefore not be relevant for smaller airlines.

For the comparison between low-cost carriers and full-service carriers, the 73 airlines have been categorized in one of these two respective groups. This has been done by checking their individual websites. When a carrier uses a low-cost business model, this is explicitly stated on their website. Appendix 3 gives an overview of the total airline sample with the low-cost or full-service tags included.

4.2.2 Sample size

(23)

23 4.3 Assumptions of multiple regression

Osborne & Waters (2002) mention four assumptions of multiple regressions that researchers should test for. When these assumptions are not met the results derived from the analysis may not be trustworthy. These assumptions are related to multicollinearity, homoscedasticity, normality and outliers and are discussed below. Output from SPSS with regard to these checks can be found in appendices 1 and 2 of this paper.

4.3.1 Multicollinearity

Multicollinearity occurs when there is a perfect linear relationship or strong correlation between two or more independent variables. It is important to avoid this as it could cause R2 values to be large but the individual beta weights to be statistically insignificant. It could produce bizarre beta weights for example with the wrong sign. Multicollinearity can be measured through the tolerance and variance inflation factor. Tolerance is the percentage of variance in the independent variable that is not accounted for by the other independent variables (1 - R2). Values of less than 0.10 are quoted as problematic. The variance inflation factor is the reciprocal of tolerance 1/ (1 - R2). It indicates whether the standard errors are inflated due to the levels of collinearity. Values over 10 are cited as problematic. In our study we have not found tolerance levels of less than 0.10 or variance inflation factors larger than 10. We have not violated the assumption of multicollinearity.

4.3.2 Homoscedasticity

With homoscedasticity is meant homogeneity of our errors. This means that the variance of our errors is constant. With homoscedasticity, all the sample data lie with about the same distance from a line which is drawn through a scatterplot (figure 3, appendix 1). With heteroscedasticity, the errors are much larger. The scatterplot indicates no problems. Therefore the assumption of homoscedasticity has not been violated.

4.3.3 Normality

The data should be approximately normally distributed. In hierarchical multiple regression it is

(24)

24

check for deviations from normality (figure 1, appendix 1). The straight line in the plot represents a normal distribution and the points represent observed residuals. In a perfectly normally distributed data set, all points will lie on the line. In our plot all the residuals cluster reasonably well around the line suggesting that the assumption of normality has been met. A histogram has also been drawn to check for non-normality. The EBITDA margin is roughly normally distributed in this study. Therefore we do not consider non-normality as a concern.

4.3.4 Outliers

(25)

25

5 Results

This chapter sets out the results of the hierarchical multiple regression analysis for both the total sample as well as for the comparison between low-cost carriers and full-service carriers. The results are presented in two steps. First, a hierarchical multiple regression analysis has been performed to investigate whether dynamic pricing affects firm performance above and beyond the effect of Customer Satisfaction (CS), Fuel Cost (FC), Gross Domestic Products (GDP) and Labour Cost (LC).

Second, a hierarchical multiple regression analysis has been performed in order to assess whether there is a stronger correlation between dynamic pricing and performance for low-cost carriers then there is for full-service carriers, above and beyond the effect of the before mentioned variables.

5.1 Analysis of the total sample

5.1.1 The sample

The table below shows the mean and standard deviation of each variable in the dataset. N presents the total sample size.

Variable Abbr. Mean Std.

Deviation N EBITDA MARGIN EM 0,047 0,103 73 LABOUR COST LC 0,165 0,052 73 FUEL COST FC 0,312 0,086 73 CUSTOMER SATISF. CS 0,668 0,13 73

GROSS DOM. PROD. GDP 34847 20800 73

DYNAMIC PRICING DP 0,071 0,067 73

Table 1: Descriptive Statistics total sample

5.1.2 Correlations

(26)

26 EM LC FC CS GDP DP Pearson Correlation EM 1,000 LC -,107 1,000 FC -,137 -,084 1,000 CS ,253 ,165 -,033 1,000 GDP -,034 ,274 -,347 ,053 1,000 DP -,242 ,042 -,069 -,356 ,102 1,000 Sig. (1-tailed) EM LC ,185 FC ,123 ,240 CS ,015 ,082 ,391 GDP ,388 ,010 ,001 ,327 DP ,020 ,363 ,281 ,001 ,194

Table 2: Correlations total sample

(27)

27

5.1.3 Model Summary

This section describes the hierarchical regression model itself. Model 1 in the table below refers to the first stage in the hierarchical regression in which only the block with control variables is entered into the regression equation. Fuel costs (FC), labour costs (LC), customer satisfaction (CS) and GDP are used as independent variables.

Model 2 refers to the situation in which dynamic pricing is added to the previous four independent control variables. R² is a measure of how much of the variability of EBITDA is explained by independent variables. In the first model R² is .110, which means that fuel costs, labour costs, customer satisfaction and GDP account for 11% in the variation in EBITDA margin. When dynamic pricing is included in model 2, the value of R² increases to 13.4% of the variance in EBITDA margin. Dynamic pricing therefore accounts for an additional 2.4% to the explanatory power of the model.

Model R R Square Std. Error of the Estimate Change Statistics R Square Change F Change Sig. F Change 1 ,332a 0,110 0,100387 0,11 2,105 0,09 2 ,366b 0,134 0,099769 0,024 1,845 0,179

a. Independent variables: (Constant), GDP, CS, LC, FC b. Independent variables: (Constant), GDP, CS, LC, FC, DP Dependent Variable: EM

Table 3: Model Summary total sample

(28)

28

5.1.4 Regression coefficients

The table below shows the model parameters of the complete model with all variables.

Model 2 Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta (Constant) ,069 ,093 ,735 ,465 LC -,270 ,239 -,135 -1,127 ,264 FC -,205 ,146 -,171 -1,406 ,164 CS ,169 ,099 ,213 1,716 ,091 GDP -000 ,000 -,050 -,398 ,692 DP -,257 ,189 -,167 -1,358 ,179

Table 4: Regression coefficients total sample

There are two types of regression coefficients: B and Beta. B is given in terms of the units of the specific variable. Beta uses a standard unit that is similar for all variables in the equation. Beta weights are useful as they allow comparison between two variables that were originally measured in different units. B-values tell us about the relationship between EBITDA margin and each independent variable. They indicate whether there is a positive or negative relationship between the independent variable and the dependent variable. More importantly, they describe the change in the dependent variable corresponding to a 1-unit increase in the independent variable.

Labour cost (t = -1,127, p > 0.05), fuel cost (t = -1,406, p > 0.05), customer satisfaction (t = 1,716, p > 0.05), GDP (t = -0,398, p > 0.05) and dynamic pricing (t = -1,358, p > 0.05) are all insignificant predictors of EBITDA margin at an α of 0.05.

5.2 Comparison analysis low-cost carriers versus full-service carriers

(29)

29

Sample Low-cost Carriers Sample Full-service Carriers

Mean Std. Deviation N Mean Std. Deviation N

EM ,074 ,079 15 EM ,040 ,108 58 LC ,135 ,028 15 LC ,173 ,054 58 FC ,358 ,059 15 FC ,300 ,088 58 CS ,587 ,092 15 CS ,690 ,131 58 GDP 36711 26913 15 GDP 34366 19169 58 DP ,109 ,078 15 DP ,061 ,061 58

Table 5: Descriptive statistics low-cost carriers and full-service carriers

The average EBITDA margin for low-cost carriers is substantially higher than it is for full-service carriers in the measuring time period. Labour costs as a percentage of revenue are lower for the low-cost carriers which can be explained by the lower cost based of low-cost airlines. Fuel costs as a percentage of revenue are higher for low-cost airlines which can be explained by the fact that fuel makes up a larger part of the product offered. Customer satisfaction is slightly higher for full-service carriers. As expected, dynamic pricing variance is higher for low-cost carriers, implying a more active strategy with regard to dynamic pricing.

5.2.1 Correlations

(30)

30

Correlations low-cost carriers

EM LC FC CS GDP DP Pearson Correlation EM 1 LC -0,026 1 FC -0,037 -0,12 1 CS 0,152 0,094 0,085 1 GDP 0,018 0,075 -0,661 -0,087 1 DP -0,533 0,006 -0,124 -0,053 0,025 1 Sig. (1-tailed) EM LC 0,463 FC 0,448 0,335 CS 0,294 0,37 0,382 GDP 0,474 0,395 0,004 0,379 DP 0,02 0,492 0,329 0,426 0,465 N EM 15 15 15 15 15 15

Table 6: Correlations low-cost carriers

Correlations full-service carriers

EM LC FC CS GDP DP Pearson Correlation EM 1 LC -0,074 1 FC -0,202 0,009 1 CS 0,338 0,074 0,059 1 GDP -0,057 0,366 -0,329 0,114 1 DP -0,252 0,171 -0,176 -0,351 0,123 1 Sig. (1-tailed) EM LC 0,291 FC 0,064 0,473 CS 0,005 0,29 0,33 GDP 0,335 0,002 0,006 0,198 DP 0,028 0,099 0,093 0,003 0,178 N 58 58 58 58 58 58

Table 7: Correlations full-service carriers

(31)

31

Other notable difference can be found between GDP and labour cost percentage. For full-service carriers there is a significant moderate positive relationship. When GDP rises, labour costs as a percentage of revenue also rise. As earlier mentioned, this might be because labour costs as a percentage of revenue are generally higher in developed countries. Low-cost carriers however have lower labour costs and might not experience this effect very much. Finally, dynamic pricing has a significant negative moderate relationship with customer satisfaction for full-service carriers. This might be because passengers of full-service airlines appreciate it less when these airlines vary more with prices. The relation between dynamic pricing and customer satisfaction is less for low-cost airlines.

5.2.2 Model Summary Model R R Square Std. Error of the Estimate Change Statistics R Square Change F Change Sig. F Change 1 ,167a 0,028 0,093 0,028 0,072 0,989 2 ,563b 0,317 0,082 0,289 3,815 0,083

a. Independent variables: (Constant), GDP, LC, CS, FC b. Independent variables: (Constant), GDP, LC, CS, FC, DP Dependent Variable: EM

Table 8: Model Summary low-cost carriers

Model R R Square Std. Error of the Estimate Change Statistics R Square Change F Change Sig. F Change 1 ,445a 0,198 0,100 0,198 3,269 0,018 2 ,472b 0,222 0,099 0,024 1,635 0,207

a. Independent variables: (Constant), GDP, CS, FC, LC b. Independent variables: (Constant), GDP, CS, FC, LC, DP Dependent Variable: EM

Table 9: Model Summary full-service carriers

(32)

32

The tables above learn that dynamic pricing explains substantially more variation in EBITDA margin for low-cost carriers then it does for full-service airlines. Unfortunately, none of the changes in R² is significant. Therefore, there is no evidence to reject the null hypothesis and accept the alternative hypothesis which states that dynamic pricing correlates stronger with firm performance for the low-cost carrier.

5.2.3 Regression coefficients

Low-cost carriers Full-service carriers

Beta P-value Beta P-Value

(Constant) 0,560 0,803 LC -0,051 0,859 -0,001 0,991 FC -0,161 0,675 -0,308 0,025 CS 0,137 0,636 0,315 0,022 GDP -0,059 0,876 -0,172 0,230 DP -0,544 0,083 -0,174 0,207

Table 10: Regression coefficients

For the full-service carrier sample labour cost (p > 0.05), GDP (p > 0.05) and dynamic pricing (p > 0.05) are all insignificant independent variables of EBITDA margin. Significant independent variables are fuel cost (p < 0.05) and customer satisfaction (p < 0.05).

For the low-cost carrier model, labour cost (p > 0.05), FC (p > 0.05), CS (p > 0.05), GDP ( p > 0.05) and DP ( p > 0.05) are all insignificant independent variables of EBITDA margin. No variable is significant.

(33)

33

6 Discussion and recommendations

In this chapter we will discuss the results, limitations, theoretical and managerial implications and give recommendations for further research. We will start with answering our two main questions.

6.1 Answer to main question 1

The first main research question of this study is:

‘If we control for the possible effects of other performance drivers (fuel cost, labour cost, customer satisfaction and GDP) is dynamic pricing still able to predict a significant amount of the variance in EBITDA Margin?

Academic literature gave clear insight into the concept of dynamic pricing and showed reasons to believe in both positive as well as negative effects of dynamic pricing on firm performance. Several studies showed positive revenue impacts but no direct relation with firm performance was found in literature. According to some scholars, competitive forces can lead to spiral-down effects when airlines push each other into lowering prices. Spiral-down effects could also occur when prices are set to wrong historical levels.

The main finding of our empirical analysis is that there is no statistically significant evidence that makes it possible to reject our H0 hypothesis. In the first model, fuel costs, labour costs, customer satisfaction and GDP explain 11% of the variation in EBITDA margin. When dynamic pricing is added to the model, the value of R² increases to 13,4%. Dynamic pricing therefore accounts for an additional 2,4% to the explanatory power of the model. Unfortunately this change in R2 is not significant.

Therefore there is no evidence that dynamic pricing contributes to firm performance on the long run or that the positive effects will diminish the negative effects.

6.1.1 Correlations

(34)

34

between GDP and labour cost indicates that GDP growth results in an increase in labour cost. This can be explained by the idea that labour costs are generally higher in developed countries. In addition there is a significant moderate negative relationship between GDP and fuel cost which could be explained by the idea that relatively rich countries have newer and more fuel efficient airplanes.

A significant moderate negative relationship exists between dynamic pricing and customer satisfaction. This could imply that perceived unfairness still plays a role in the airline industry and may not have weakened yet, as literature argued.

Negative non-significant correlations exist between labour cost percentage and performance and between fuel cost percentage and performance. Customer satisfaction is positively correlated with performance. GDP and dynamic pricing are negatively correlated with performance, implying that dynamic pricing has a negative effect on performance.

6.1.2 Summary

Based on the results of the hierarchical regression we can conclude that there is no significant evidence to assume that there is a relation between dynamic pricing activities and firm performance. Remarkable is that both correlations and beta’s are negative implying a negative effect of dynamic pricing on firm performance. This could be explained by the several spiral-down theories that describe the concept of decreasing sales prices due to competitive forces and wrong historical data usage as described in the literature review.

6.2 Answer to main question 2

The second main question is formulated as follows:

‘Is there a difference between low-cost carriers and full-service carriers with regard to the effect of dynamic pricing on firm performance?’

(35)

low-35

cost carrier customers being more price-sensitive. When controlling for the effect of other performance drivers low-cost carriers show a larger change in R2 implying that dynamic pricing is a stronger predictor of variance in EBITDA margin.

Although adding dynamic pricing to the model improved the explanatory power of the model a lot more for low-cost carriers then it did for full-service airlines, the increase is not significant. The findings of the comparison analysis therefore do not show statistically significant evidence to reject our H0 hypothesis. No evidence can be found that dynamic pricing is stronger related with firm performance for low-cost carriers.

6.2.1 Descriptives

Average EBITDA margin for low-cost carriers is substantially higher than for full-service carriers in the measuring time period. There is not a clear explanation for this. It could be that low-cost carriers are more successful in times of economic downturn. Labour costs as a percentage of revenue are lower for the low-cost carriers which can be explained by the lower cost base of low-cost airlines. Fuel costs as a percentage of revenue are higher for low-cost airlines, which can be explained by the fact that fuel makes up for a larger part of the product offered. Customer satisfaction is slightly higher for full-service carriers. As expected, dynamic pricing variance is higher for low-cost carriers implying a more active strategy with regard to dynamic pricing.

The correlation between customer satisfaction and EBITDA margin is moderate and significant for full-service carriers. An increase in satisfaction probably results in higher returns. For low-cost carriers this relationship is negligible.

There is a significant strong negative relationship between dynamic pricing and EBITDA margin for low-cost carriers (r = -0,533, p < 0.05). The correlation is negative for both low-low-cost carriers and full-service airlines. Possible explanations for this according to literature are the spiral-down effects.

(36)

36 6.3 Limitations

A reason that could have resulted in not getting any statistical significant evidence for several relations is that the sample size is too small. Coefficients and sample size are the only factors that are needed to compute the standard error, and hence assess statistical significance. To be able to test again for results in the future the sample size should be enlarged.

Although several scholars used standard deviation-based factors to define dynamic pricing activity we argue that this approach is questionable. It could be that effective dynamic pricing strategies do not involve large deviations but are based on small sensitive corrections instead.

Another drawback of our study lies in the selection of the sample cases. The selection of airlines was based on a minimum amount of operating revenue to separate the large passenger transport airlines from smaller helicopter companies. Unfortunately this excludes the possible effect of dynamic pricing executed by smaller airlines.

An important reason for not finding a positive correlation between dynamic pricing and performance could be the time lag between the measurement of the performance (3 years average) and the measurement of dynamic pricing activity of the airline (3 weeks average). To reduce this gap a larger dataset is needed with a longer timeframe.

Another reason for the results being insignificant is that data has been gathered from multiple databases and annual reports. Because the databases often had limited data availability, it was difficult to calculate the mean for multiple years of several variables. This could have led to results of airlines being more extreme. Additionally, different accounting standards could disturb comparability.

6.4 Theoretical and managerial implications

(37)

37

performance and that a positive impact has not been proven. The study is a contribution to the field of knowledge on revenue management and a step forward in making trade-offs in performance management policies. Airline management and investors should realize that the link between firm performance and revenue management is not as clear as assumed. When making investments in revenue management tools, airlines should realize that there is no guarantee for an increase in performance.

6.5 Further research

Our study shows that a possible reason for the negative correlation between performance and dynamic pricing may be caused by the so-called spiral-down effect. Therefore it is important to take the effect of competition and market concentration into account in further analyses. A lot of scholars emphasize the importance of the effect of market concentration.

Future scholars should also review the determinants of airline performance because in this study we found no statistically significant correlations between any of the individual performance drivers and the EBITDA.

Scholars can focus on the effect of dynamic pricing on sales and profit growth of different pricing approaches in order to find a better explanation for the possible negative effects of dynamic pricing as suggested in this study. We propose that future research also performs a cross industry analysis to find out how other industries organize the pricing function.

(38)

38

7 References

Anderson, C. K., & Blair, M. (2004). Performance monitor: the opportunity costs of revenue management. Journal of Revenue and Pricing Management, 2(4), 353-367.

Banker, R. D., & Johnston, H. H. (1993). An Empirical Study of Cost Drivers in the U.S. Airline Industry. Accounting Review, 68(3), 576-601.

Behn, B. K., & Riley, R. A. (1999). Using nonfinancial information to predict financial performance: The case of the US airline industry. Journal of Accounting, Auditing & Finance, 14(1), 29-56.

Burger, B., & Fuchs, M. (2005). Dynamic pricing--A future airline business model. Journal Of Revenue & Pricing Management, 4(1), 39-53.

Cooper, D. R., & Schindler, P. S. (2003). Business research methods.

Cooper, W. L., Homem-de-Mello, T., & Kleywegt, A. J. (2006). Models of the spiral-down effect in revenue management. Operations Research, 54(5), 968-987.

Dunn, G., & InSIGhT, F. (2012). How to get a Head in airlines. Airline Business, 28(1).

Elmaghraby, W., & Keskinocak, P. (2003). Dynamic pricing in the presence of inventory considerations: Research overview, current practices, and future directions. Management Science, 49(10), 1287-1309. Field, A. (2009). Discovering statistics using SPSS. Sage publications.

Garbarino, E., & Lee, O. F. (2003). dynamic pricing in Internet Retail: Effects on Consumer Trust. Psychology & Marketing, 20(6), 495-513. doi:10.1002/mar.10084

Gelman, A., & Hill, J. (2007). Data analysis using regression and multilevel/hierarchical models. Cambridge University Press.

Green, S. B. (1991). How many subjects does it take to do a regression analysis. Multivariate behavioral research, 26(3), 499-510.

Hawkins, D. M. (1980). Identification of outliers (Vol. 11). London: Chapman and Hall.

Kaen, F., & Baumann, H. (2003). Firm size, employees and profitability in US manufacturing industries. Employees and Profitability in US Manufacturing Industries (January 13, 2003).

Kuancheng, H., Yi-Ya, P., & I-Wen, W. (2013). A Framework and Models for Evaluating the Multifare dynamic pricing Mechanism of Low-cost Carriers. Transportation Journal (Pennsylvania State University Press), 52(3), 308-322.

Kumar, S., Johnson, K. L., & Lai, S. T. (2009). Performance improvement possibilities within the US airline industry. International Journal of Productivity and Performance Management, 58(7), 694-717. M., F. F. (2008). Fuel taking ever-greater bite out of airlines' income. Travel Weekly, 67(29), 39.

Mantin, B., & Koo, B. (2009). Dynamic price dispersion in airline markets. Transportation Research Part E: Logistics and Transportation Review, 45(6), 1020-1029.

O’Connell, J. F., & Williams, G. (2005). Passengers’ perceptions of low-cost carriers and full service carriers: A case study involving Ryanair, Aer Lingus, Air Asia and Malaysia Airlines. Journal of Air Transport Management, 11(4), 259-272.

Sameer, K., Kevin L., J., & Steven T., L. (2009). Performance improvement possibilities within the US airline industry. International Journal Of Productivity & Performance Management, 58(7), 694-717. Saunders, M. N., Saunders, M., Lewis, P., & Thornhill, A. (2011). Research methods for business students, 5/e. Pearson Education India.

Schefczyk, M. (1993). Operational performance of airlines: an extension of traditional measurement paradigms. Strategic Management Journal, 14(4), 301-317.

(39)

39 Profiling.

Şen, A. (2013). A comparison of fixed and dynamic pricing policies in revenue management. Omega, 41(3), 586-597.

Steven, A. B., Dong, Y., & Dresner, M. (2012). Linkages between customer service, customer satisfaction and performance in the airline industry: Investigation of non-linearities and moderating

effects. Transportation Research: Part E, 48(4), 743-754.

Tarry, C. (2014). ANALYSIS: GDP growth will determine airline performance as market improves. Airline Business, 30(3), 13.

Weisstein, F. L., Monroe, K. B., & Kukar-Kinney, M. (2013). Effects of price framing on consumers’ perceptions of online dynamic pricing practices. Journal of the Academy of Marketing Science, 41(5), 501-514.

Williams, K. R. (2013). Dynamic Airline Pricing and Seat Availability.

(40)

40

APPENDIX 1: Sample Descriptives

Assumption 1: Multicollinearity Coefficients Model Unstandardized Coefficients Standardized Coefficients t Sig. Collinearity Statistics B Std. Error

Beta Tolerance VIF

1 (Constant) 0,021 0,087 0,241 0,81 LC -0,295 0,24 -0,148 -1,229 0,223 0,902 1,108 FC -0,195 0,146 -0,163 -1,336 0,186 0,879 1,137 CS 0,219 0,092 0,276 2,377 0,02 0,972 1,028 GDP -3,21E-07 0 -0,065 -0,511 0,611 0,819 1,22 2 (Constant) 0,069 0,093 0,735 0,465 LC -0,27 0,239 -0,135 -1,127 0,264 0,897 1,115 FC -0,205 0,146 -0,171 -1,406 0,164 0,877 1,14 CS 0,169 0,099 0,213 1,716 0,091 0,838 1,193 GDP -2,50E-07 0 -0,05 -0,398 0,692 0,814 1,229 DP -0,257 0,189 -0,167 -1,358 0,179 0,852 1,174 a. Dependent Variable: EM Table 11: Multicollinearity

The table above shows that none of the independent variables are correlated with a tolerance level lower than 0.1. The VIF values are also below 10. Therefore the assumption that the independent variables are not highly correlated is satisfied.

Assumption 2: Normality

(41)

41

Non-normality could also be presented in a histogram. The histogram is supposed to look like a normal distribution, or a bell-shaped curve. Any deviation from this curve is a sign of non-normality and the greater the deviation, the more non-normally distributed the residuals are. The EBITDA margin is roughly normally distributed in this study.

Figure 1: P-P plot total sample of airlines Figure 2: Histogram residuals

(42)

42 Assumption 3: Homoscedasticity

The graph looks like a random array of dots evenly dispersed around zero. This pattern is indicative for a situation in which the assumption and homoscedasticity have been met.

Assumption 4: Outliers

(43)

43

APPENDIX 2: Low-cost carriers and Full-service carriers

Assumption 1: Multicollinearity

Coefficients

Model Collinearity Statistics

Low-cost Carrier

Collinearity Statistics Full-service Carrier Tolerance VIF Tolerance VIF

1 (Constant) LC 0,975 1,026 0,847 1,181 FC 0,557 1,795 0,864 1,157 CS 0,98 1,02 0,976 1,024 GDP 0,562 1,779 0,744 1,344 2 (Constant) LC 0,975 1,026 0,819 1,221 FC 0,546 1,832 0,847 1,181 CS 0,978 1,022 0,843 1,186 GDP 0,559 1,79 0,742 1,347 DP 0,977 1,024 0,811 1,233 a. Dependent Variable: EM Table 12: Multicollinearity

In this analysis, none of the independent variables are correlated with a tolerance level lower than 0,1, and the VIF values are below 10. Therefore the assumption that the independent variables are not highly correlated has been satisfied.

Assumption 2: Normality

(44)

44

Figure 4: Normal P-P Plot Low-cost Carriers Figure 5: Normal P-P Plot Full-service Carriers

Figure 6: Histogram Low-cost Carriers Figure 7: Histogram Full-service Carriers

In the plots that are showed above all residuals cluster reasonably well around the line, suggesting that the assumption of normality has been met.

The histogram below should look like a normal distribution, or a bell-shaped curve. Any deviation from this curve is a sign of non-normality and the greater the deviation, the more non-normally distributed the residuals are.

(45)

45 Assumption 3: Homoscedasticity

In the first scatterplot that is shown below we can see that the points are not as evenly dispersed as we have seen in Appendix 1 and in the second scatterplot. However, we also don’t see a funnel shape or curvilinear relationship. Therefore there is no heteroscedasticity. This second pattern is indicative of a situation in which the assumption of homoscedasticity has been met.

Figure 8: Scatterplot Low-cost Carriers Figure 9: Scatterplot Full-service Carriers

Assumption 4: Outliers

(46)

46

APPENDIX 3: Airline Sample

COMPANY COUNTRY LOW-COST / FULL-SERVICE

Southwest Airlines US Low-cost

Easyjet Great Britain Low-cost

Norwegian Air Shuttle Norway Low-cost

Air Asia Malaysia Low-cost

Spicejet India Low-cost

Cebu Air Philippines Low-cost

Tiger Airways Singapore Low-cost

Gol Linhas Brazil Low-cost

Frontier Airlines US Low-cost

Spirit Air Lines US Low-cost

JetBlue Airways US Low-cost

Allegiant Air US Low-cost

Air Arabia UAE Low-cost

Vueling Spain Low-cost

Nok Air Thailand Low-cost

Lufthansa Germany Full-service

Air France France Full-service

British Airways Great Britain Full-service

China Southern Airlines China Full-service

China Airlines Taiwan Full-service

Qantas Airways Australia Full-service

China Eastern Airlines China Full-service

US Airways US Full-service

Cathay Pacific Airways Hong Kong Full-service

Singapore Airlines Singapore Full-service

Air Canada Canada Full-service

Aeroflot Russia Full-service

SAS Scandinavian Airlines Sweden Full-service

Thai Airways Thailand Full-service

Air Berlin Germany Full-service

Alaska Air US Full-service

Malaysian Airlines Malaysia Full-service

Avianca Colombia Full-service

Garuda Indonesia Indonesia Full-service

Virgin Australia Australia Full-service

Air New Zealand New Zealand Full-service

Jet Airways India Full-service

(47)

47

TAP Portugal Portugal Full-service

Aer Lingus Ireland Full-service

Kenya Airways Kenya Full-service

Icelandair Iceland Full-service

Aegean Airlines Greece Full-service

Croatia Airlines Croatia Full-service

Iberia Spain Full-service

Alitalia Italy Full-service

Eva Airways Taiwan Full-service

Virgin Atlantic Airways Great Britain Full-service

Air Corsica FR Full-service

Air Serbia RS Full-service

Meridiana Fly IT Full-service

Enter Air PL Full-service

Bulgaria Air BG Full-service

United Air Lines US Full-service

American Airlines US Full-service

Delta Air Lines US Full-service

Hawaiian Airlines US Full-service

ExpressJet Airlines US Full-service

World Airways US Full-service

Virgin America US Full-service

Sun Country Airlines US Full-service

North American Airlines US Full-service

Miami Air International US Full-service

SkyWest Airlines US Full-service

Air Wisconsin Airlines US Full-service

Mesa Airlines US Full-service

Royal jordanian Jordan Full-service

ANA Japan Full-service

El Al Israel Full-service

Korean Air South Korea Full-service

Transaero Airlines Russia Full-service

Japan Airlines Japan Full-service

Revenue Management in the Airline Industry