Choosing financial Key Performance Indicators: the Airline Industry case Ganna Demydyuk Address: Herbertstr. 26, 14193,Berlin, Germany Tel. +49 1725874801 e-mail: [email protected]

Abstract Selecting relevant Key Performance Indicators (KPIs) involves an assessment of both costand revenue-driven measures. Cost driven allocation usually predominates, due primarily to a traditional accounting mindset coupled with the need for cost savings in the current economic environment. Using data from the airline industry in all of the major markets in the world, this paper demonstrates that revenue- or profit-driven KPIs, consistently applied, will likely lead to better financial performance than „flying‟ the business based on cost-driven metrics or those, representing a mixture of revenue target and cost-driven metric. Specifically it examines the effectiveness of models that characterize performance based on two performance indicators in particular – seats and passenger-kilometers. We document strong evidence indicating that Operating Profit per Passenger or per Passenger-Kilometer is the most significant variable when it comes to explaining the variation in airline profitability, arriving at the conclusion that despite the traditional belief that measuring per seat is only appropriate for point-to-point destination services typically provided by Low Cost Carriers, the same model also fits Full Service Network Carriers and, thus, can be used by them as a meaningful tool for financial targeting and strategic decision-making.

Acknowledgements This paper would be incomplete if it did not contain the fullest expression of my gratitude to Professor Dr. Hany Shawky, who has so generously given of his time and expertise; and not only with this paper but throughout my PhD project. He has prompted and mentored my attempts to achieve a level of excellence with an open and gentle manner that created an environment for ideas to bear fruit and for me to believe in my own potential. His contribution has been invaluable and I thank him most sincerely!

1. Introduction Selecting relevant Key Performance Indicators (KPIs) involves an assessment of both costand revenue-driven measures. Cost driven allocation usually predominates, due primarily to a traditional accounting mindset coupled with the need for cost savings in the current economic environment. Using data from the airline industry in all of the major markets in the world, this paper demonstrates that revenue- or profit-driven KPIs, consistently applied, will likely lead to better financial performance than „flying‟ the business based on cost-driven metrics or those, representing a mixture of revenue target and cost-driven metric. Specifically it examines the effectiveness of models that characterize performance based on two performance indicators in particular – Seats and Passenger-kilometers. Despite airline industry growth over decades, the majority of airline businesses remain consistently unprofitable over an entire business cycle. The topic of airline profitability leaves few disinterested, and almost all consumers feel able to participate. Therefore, every judgment that helps or hinders an airline‟s success, invariably results in an emotional discussion about the complexity and diversity of airline specifics depending on the size, region or business model. Indeed, various airlines operate in dramatically different external and internal environments, offering different products and complying with different legal requirements. The 2008 results (IATA Economics, 2010) showed that there is no evidence of a straightforward correlation between size and absolute profit, confirming results from earlier years. Recent years were particularly challenging through the worst-ever demand crisis and record fuel prices. In 2008, operating profitability was extremely weak across all regions. Out of 233 major airlines representing 89% of total passenger traffic, 90 airlines made operating losses with 32 losing more than US$ 100 million each. The larger size was associated with both bigger losses (e.g. American or United Airlines) and larger profits (e.g. Lufthansa or FedEx). Those able to make operating profit represented a range of business models. Large network airlines, smaller network airlines, regional feeder airlines, cargo airlines and Low Cost Carriers - all emerge among the top 30 airlines. Some Low Cost Carriers, previously generating +10% margins, saw weaker performance, demonstrating that all business models were challenged by the economic environment. Certainly, achieving a successful airline consists of a great number of aspects. Thousands of activities towards revenue increase or cost reduction are simultaneously taking place in airline business, but whatever is done it results in the black or red bottom line in the Profit and Loss Statement. We provide an empirical study that attempts to distinguish between cost driven and revenue driving financial performance indicators that may better help us predicate airline‟s financial performance. Our main assumption underlines the effect of using two different KPI‟s models. We examine the effectiveness of models characterizing performance based on two activity drivers – seats or passengers (revenue driving) and passenger-kilometers (cost driven). It has been traditionally considered that measuring by seats is only appropriate for point-to-point destination services, typically provided by Low Cost Carriers and should not be used by legacy airlines. Our key findings indicate that model based on kilometers fits the industry slightly better than the one based on passengers (seats). Furthermore, we find strong evidence indicating that Operating Profit per Passenger or per Passenger-Kilometer is the most significant variable explaining airline profitability. In spite of classical beliefs, we found it is more meaningful than revenue, cost and load factor traditionally used by the industry. We also found that relationship between profit margin and seats-based model is strong enough for both classes – LCCs and Full Service network carriers. Therefore we arrive at the conclusion that Operating Profit per Seat can be successfully used for targeting the financial performance of Full Service Network Carriers. 2

2. Industry application 2.1.

Selection of the airline industry

When analyzing the effectiveness of various KPIs having financial or operational origin the data should be as detailed as possible. To compare variables relating to cost and revenue generation one would require data overlooking long-term financial perspective down to product and customer level details. Airlines data provided a solution when we were able to get operational statistics (number of passengers, number of kilometers offered/flown, employees, aircrafts, etc.) from open sources, as for each specific company so for the industry. The data used for the analysis has been taken from published annual reports of commercial airlines over recent 5 years. These reports contain the main financial statements rich in operational statistics on a product and customer level. The data is presented in such a way that it can be processed and brought to a consistent basis. The reports also contain detailed cost and revenue explanations, which serve as an excellent demonstration, independently of the scope, focus region and business model used. We can also look at the stock market information such as share price development and Beta. Additionally, there are a number of very useful and detailed industry reports, covering European, American and global markets from specialized organizations. Therefore ongoing analysis and testing performs on airline industry data. 2.2.

Industry data

The classification of airlines in this paper will follow a model used by the DLR‟s Air Transport and Airport Research Unit, as we guided our sample selection based on the DLR‟ classification and overviews. For further analysis regarding airline types, airline companies are distinguished by those of (abbreviation in brackets): Full Service Network Carriers (“FSNCs”) Low Cost Carriers (“LCCs”) Regional Carriers (“Regionals”) Holiday / Charter Carriers (“Charters”) Full Service Network Carriers are scheduled airlines with a business model that focuses on providing a diverse and extensive service. These are typically international operating companies with a network-oriented system (normally with one or more hubs), covering a wide geographical area and providing transportation in several different classes. The Low Cost Carriers category comprises those airlines that offer low prices for the majority of flights and which mainly operate on short and medium-distance routes with low overheads and a relatively high load factor; these airlines use a no-frills business model. In our analysis we will only work with FSNC‟s and LCC‟s and not with Regionals or Charters, because the market is generally dominated by these two business models. For example in the year 2008, FSNCs supplied 58% of weekly seats available at European airports, followed by LCCs offering 34.1% of total capacity. Charter carriers and Regionals had respective shares of only 4.7% and 3.2%. On average, the top 40 airlines cover almost the whole market: in 2008 - 40 top FSNCs 91,1% and 40 top LCCs - 99.8% respectively. (DLR, Annual Report 2007, 2008)

3

2.3.

Literature review

According to Doganis (1985), the profitability of an airline depends on the interplay of three variables, the unit costs, the unit revenues or yields and the load factors achieved. Airline managers must adjust costs, fares and load factors to produce profitable combinations. This dynamic and complex process is very difficult due to pricing instability inherent in the airline industry. He (Doganis) characterizes the industry by short-run marginal costs which are close to zero. Marginal cost of carrying an extra passenger on a flight, which is due to leave with empty seats, is no more than a cost of additional meal, an airport passenger charge, the cost of ground handling and a few pounds of fuel burnt as a result of extra weight. The run of th ese costs is really short, because if the seats remain unsold, these seats flown or seat kilometers produced will be lost forever. Therefore it makes business sense to maximize revenues and load factors. For passenger airlines, the average revenue per output sold is called Yield and measures average revenue per passenger, per passenger kilometer, per passenger tone kilometer performed. Thus, he reasons the existence of low cost carriers, stating that by combining passenger yields with low cost and relatively high load factors one can achieve profitability. He also demonstrates that low cost itself does not provide big margins interacting with low revenues, nor does high cost necessarily mean low profits if the revenues are high enough. Doganis concludes that airlines deciding on their pricing strategy, and working out var ious tariffs, must balance and assess all these factors, which transform the various fares into average yield. He states that it is the yield in conjunction with the achieved load factor and the unit costs, which will determine whether an airline‟s revenue and financial targets can be met. To assure such process airlines apply revenue management process, underlying revenue management systems. McGill and Van Ryzin (1999) in their „Revenue Management: Research Overview and Prospects‟ reviewed forty-year history of research on transportation revenue management. They define Revenue management as practice of controlling the availability and/or pricing of travel seats in different booking classes with the goal of maximizing expected revenues or profits called. This term has largely replaced the original term yield management. The revenue management problem can be formulated as “given a current booking request, should that request be fulfilled at the current price or should it be held in anticipation of realizing a higher price in the future?”. In the case of single leg (Origin – Destination) and single product a solution can be found using the expected marginal seat revenue approach (Belobaba, 1987). Later, numerous extensions of this model have been investigated, since the development of revenue management systems has progressed from simple single leg control, through segment control, and finally to origin– destination control since the 80-s. (Dunleavy and Phillips, 2009). The objective in revenue management is to maximize profits; however, in most situations, it is considered sufficient to seek booking policies that maximize revenues. (McGill and Van Ryzin, 1999) Thus, revenue management focuses on revenue achievement without direct link between profit and revenue in a single system. Planned or targeted revenue is usually calculated to cover costs and achieve profit; the latter can be applied as a further layer of percentage. Calculation of projected revenue is usually done using traditional accounting concepts, which focus mainly on cost allocation and therefore based on cost driving metrics (kilometers). 4

When talking about accounting concepts, an interesting approach to management decisionmaking for improved bottom-line performance „Contribution Based Activity‟ or CBA was suggested by Cleland (1998). It does not directly relate to transport industry, but to business in general. According to CBA approach, the planned number of units of output is divided into the planned gross profit, defined as sales less cost of materials/goods or total expenses+ profit. The resulting benchmark is known as Target Average Gross Profit per Unit or TARI. The approach suggests a performance management system (including pricing and productivity), not denying cost or revenue management, but complementing them. It links financials (profit = revenue - costs) and operational metrics (units of output) and simplifies the process of comparing planned with actual gross profit per unit. As this method involves revenue, costs and activity, it allows management to overview the whole picture in a timely and relevant manner. Although CBA and TARI approach possesses substantial benefits compared with traditional accounting and costing systems, it has not yet been universally recognized as a viable alternative. CBA method critically demands proper definition of an output measure, called activity driver, considered to be fundamental to all other activities, in other words it suggests working with profit driving indicators. Latest trends and research suggest targeting Profit instead of Revenue; however we did not find studies, which extended into profit per unit of output, and answering what should be considered to be the right output measure? Even though marketing management literature in transport industry and the like, addresses problems of pricing, revenue per passenger and profit per passenger, such studies mainly deal with promotion and market communication, which is not the objective of this paper. Thus, an idea of CBA appeared for us being the most integral, comparing with the other accounting approaches, and therefore we will try to extend Cleland‟s approach to one of the possible applications in the airline industry. Summarizing the above we arrive with two potential models of performance measurement. First one consists of commonly used metrics described by Doganis. It is focused on revenue achievement and based on cost driven measures such as kilometers (more kilometers flown generate greater costs). Another model we are going to develop would consist of revenue driving indicators such as seats (more seats filled with passengers increase revenue), which according to Doganis are mainly attributed to LCCs. Following the Cleland‟s idea it is going to be united with focus on Profit instead of Revenue in order to incorporate costs in the suggested model.

3. Variables 3.1. Operational variables From our study of airlines annual reports as well as from literature on airline economics, we conclude that main operational measurements are as follows (definitions according to R. Doganis “Economics of international airlines”): ASKs – available seat kilometers are obtained by multiplying the number of seats available for sale on each flight by the stage distance flown (sometimes miles, then referred as ASM) RPKs – revenue passenger kilometers or passenger kilometers are obtained by multiplying the number of fare-paying passengers on each flight stage by flight stage distance. They are a measure of airline passenger traffic. Load factor (per cent) or Passenger load factor is passenger kilometers RPK expressed as a percentage of available seat kilometers ASK. Load factor is considered 5

to be one of the most important indicators of airline operations and for certain airlines it remains the main management focus. On a single sector, load factor is simplified to the number of passengers carried as a percentage of seats available for sale. Such approach could be used by LCCs because of the nature of their operations; however even by LCCs it is reported as RPK to ASK. Number of Passengers or Passengers carried (PAX) equals the number of passengers, which boarded each aircraft and summed over a certain period of time. As we mentioned earlier, FSNCs and LCCs in their reports demonstrate different attitudes to the role this variable plays, varying from key to just informative. Not knowing the internal KPIs system of each specific airline, one can judge from the reports that use of this variable as one of the core ones is more attributable to LCCs where it found to be mainly successful. Airlines with more traditional culture, mainly FSNCs, look upon this figure as having more statistical rather than operational value. 3.2.

Financial Ratios

Traditionally, airlines used the following ratios to indicate their profitability and efficiency: Yield is the average revenue collected per passenger kilometer or RPK. Passenger Yield is calculated by dividing the total passenger revenue on a flight by the passenger kilometers generated by that flight. It is a measure of the weighted average fare paid. It is considered that airlines should be focused at Yield increase. (It is commonly understood that yields decrease under the pressure of competition and consumer buying ability). Cost per ASK is a measure obtained by dividing total operating costs by total ASKs. Operating costs exclude interest payments, taxes and extraordinary items. Costs could also be measured by RPK, but measuring costs by ASK is more relevant and therefore very common. Cost reduction is another important focus of airlines. Recent fuel price fluctuations drove Cost per ASK nearly out of control. Logically, FSNCs have higher costs per ASKs comparing with LCCs, because of the different business model they use. Flying with very thin margins, LCCs are more cost sensitive than their FSNCs competitors. As we find further on, FSNCs in turn demonstrate higher sensitivity to revenues. As for any firm we also looked at the following financial variables or our analysis: Total revenue connected to scheduled passenger traffic Operating costs, excluding interest expenses, taxes, extraordinary items and other non-operating expenses, detailed and total Operating profit, absolute and operating margin percentage. While FSNCs rely traditionally on cost driven metrics, calculating everything per kilometer, LCCs often use revenue or profit driving indicators such as per seat or per passenger metrics. For example, the following „per seat measures‟ were used for reporting by European LCC easyJet (easyJet full year results 2009): Per seat measures (underlying) Profit before tax per seat (£) Revenue per seat (£) Cost per seat (£) Cost per seat excluding fuel (£) 6

For consumers of this report such approach looks and sounds clearer and more focused than the kilometers. This model is typical for LCC rather than FSNCs as they mostly sell one-way single leg restricted fares and consider such model being applicable. FSNCs in contrast sell far more complex product, and consider the kilometer version to be more appropriate. In the next chapters we are going to demonstrate that Seat or Passenger model is also applicable for FSNCs in planning and achieving firm‟s financial targets.

4. Methodology. For the airline industry, revenue management systems are in place to help decision-making. The main focus of such systems is to maximize revenue by finding a balance between selling cheap promotion and expensive regular fares. Usually, conventional airlines have more complex fare structure, because of booking classes and restrictions, while LCCs usually offer single class and fewer fares. Revenue management is certainly about pricing, but profit normally takes costs into account. Cost reduction and control plays one of the key roles in airlines operations. Keeping costs under control is vitally important to provide continuing operations. FSNCs represent extensive cost structures with numerous fund-consuming assets, branches and legacy requirements. LCCs often operate younger and more unified fleets with maximal possible aircraft utilization. Both classes are conscious about fuel price expenses, and constantly try to decrease on-board weight and hedge fuel prices. Depending on different financial strategies (buying or leasing aircraft) financial and operating costs will also be different from company to company. Airlines representing different business models have their own ways for revenue maximization. FSNCs mainly offer premium class services for higher paying passengers, and thus face a question concerning the right distribution of aircraft space between business/first and economy seats. FSNCs often sell more expensive tickets the closer the departure date. Intercontinental or long-haul flights, which presumably are more profit gaining, are mainly offered by FSNCs. In turn, LCCs differ by their creativity towards revenue maximization, implementing various extra fees and charges (check-in bags, priority boarding, food and beverages, credit cards fee, etc.) Numerous actions toward revenue increase or cost reduction are simultaneously taking place in airline business, but whatever is done – shifting the curtain between business and economy compartments or rushing passengers running into the plane without seat assignments – it results in a black or red bottom line in the Profit and Loss statement. In spite of latest trends focusing on profit instead of revenue, it is still very common for companies to focus on revenue. Planned or targeted revenue is usually calculated to cover costs and achieve profit; the latter applied as a further layer of percentage. Such focus on revenue is especially attributable to industries with low marginal costs, for example hotels or passenger transport carriers. In order to involve both participants in the profit process – revenue and costs – we suggest planning and targeting profit instead of revenue and costs, combined with planned load factor. Should we apply CBA approach and use the Gross Profit (sales minus cost of materials), GP would tend to Revenue, as direct material costs here are only marginal. Therefore in this case Operating Profit is the most informative and consistent variable to express financial contribution produced by key activity. We distinguish Operating Profit from Net Profit because the latter already contains extraordinary items, government grants, write offs and the like. Nevertheless, net profit is still an important indicator and it is incorporated in further used ratio Return on Assets ROA %. 7

Mainly, key metrics represent ratios, consisting of numerator and denominator. Numer ator indicates a targeted value and Denominator indicates a measure, in this case a measure of output. Instead of revenue we suggest expressing targets as financial contribution, which is made by activity driver or output (seats sold or revenue passenger kilometers): Target: Operating Profit per output rather than on Revenue per output Now we can distinguish between cost and revenue driving metrics and aim to compare effectiveness of two existing models to airline performance measurement per seat and per passenger kilometer. As we are measuring output, we are going to work with most commonly used metrics such as Seats sold and Revenue Passenger Kilometers (RPK). For the purpose of this research we identify seats sold with passengers carried, primarily because any existing difference between the two is insignificant and in any case, it is not possible to access the data from most company reports. Therefore we used the number of passengers carried in both – data collection and empirical testing. Thus, the second suggestion relates to output: Output: Target per Passenger carried and rather than Target per RPK These both suggestions specifically result in the following ratios, which we are going to use the in our empirical testing: Operating Profit per passenger carried Revenue per passenger carried Operating cost per passenger carried Future targets are based on past performance and therefore we are going to analyze these variables in available data sample. Further we develop the models involving the above ratios as well as the others used by airlines, including models consisting of the traditional KPI‟s. The goal of this analysis is to establish whether there is a measurable significance in profitable performance between focusing on Operating Profit per passenger or per RPK (passenger-kilometer) instead of Revenue. This suggested model will be compared with traditional models, consisting of revenue, load factor and RPKs.

5. Sample characteristics 5.1.

Data availability

For our analysis we focused on 20 top airlines in each class (FSNC and LCC), accessing financial data for 5 years (2004-2008). Companies were ranked as top in terms of revenue passenger kilometers in the first 9 months of 2008. The list of these companies and the ranking has been taken from DLR “Annual analyses of the European air transport market 2008”. Our sample includes 15 top FSNCs and 12 LCCs, reflecting worldwide geography. We did not manage to get data for all 40 companies because some of them are not publically traded and so don‟t publish reports while others are unavailable because of recent mergers and organizational changes. The data in the form of annual financial statements, annual filings and business reports were downloaded from airlines web-sites, transferred into US dollars and processed into a consistent basis. The numerical part of this data was clustered according to 3 criteria: Business model (FSNC or LCC) Region of origin and operations (Europe, Americas, Asia-Pacific and Middle East) Financial performance (High or Low) 8

Table 1. Airline data used in the analysis, rank Jan.-Sep.2008 Rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Passengers 71,341 56,260 49,000 54,280 37,710 43,590 78,040 24,510 42,020 16,670 14,390 18,780 35,280 43,140 18,310 17,540 36,660 20,970 29,200 16,810 18,860 16,800 9,020 13,030 8,190 8,450 4,050

RPK 163 192 157 398 137 705 130 231 103 438 96 392 90 705 83 242 75 584 75 083 70 330 68 600 62 657 61 591 61 223 58 777 41 575 34 634 33 582 32 465 23 723 19 678 16 745 14 424 12 422 9 685 7 865

Airline American Airlines Air France - KLM Group United Airlines Delta Air Lines Continental Airlines Lufthansa Southwest Airlines British Airways US Airways Emirates Airline Singapore Airlines Cathay Pacific Japan Airlines International China Southern Airlines Qantas Air Canada Ryanair Airberlin easyJet JetBlue Airways AirTran Airways GOL Transportes Aereos WestJet Virgin Blue Frontier Airlines AirAsia Transavia Airlines

Region North America Europe North America North America North America Europe North America Europe North America Middle East Asia-Pacific Asia-Pacific Asia-Pacific Asia-Pacific Asia-Pacific North America Europe Europe Europe North America North America Latin America North America Asia-Pacific Asia-Pacific Asia-Pacific Europe

Business model FSNC FSNC FSNC FSNC FSNC FSNC LCC FSNC FSNC FSNC FSNC FSNC FSNC FSNC FSNC FSNC LCC LCC LCC LCC LCC LCC LCC LCC LCC LCC LCC

One important observation is that it is not always possible to get some operational data from airlines reports. For example, such basic and fundamental figure as the number of passengers is sometimes not ascertainable. So, it was necessary to surf industry and country statistical reports in order to find this data, especially for the companies with North American origin. Taking into account that number of passengers is our main f ocus, the fact that a 120page report does not contain this basic figure reflects adversely on report quality, and thus information used by managers in their decision-making. Another point to mention is a slight inconsistency in data regarding the load factor, r eported by airlines. It is not clear from the reports whether they use passenger load factor or seat load factor, which counts not only fare-paying passengers, or includes also crew travelling to the point of their future destination. Therefore we calculated load factor ourselves assuring the consistency of this ratio used in our analysis. Finally, some airlines report their average net airfare, which is very interesting to analyze looking at the revenue and profit structure. However, the number of such companies is insufficient to present a valid sample. That is why we used total revenue for passenger airlines, which includes not only total passenger revenue, but also freight and mail, cargo revenue, MRO, catering as accompanying service for passenger airlines. 5.2.

Descriptive Statistic and Correlations.

As mentioned before, the sample includes 5 years data (2004-2008) for 27 companies, i.e. 135 cuts on annual basis. The statistical analysis overlooks all variables and ratios used in both Kilometers and Passenger modes and present such data as Minimum, Maximum, Standard Deviation, Skewness and Kurtosis. Sample data is very diverse, because of such a difference in scope, region and business model and therefore total sample required to be clustered accordingly. Analysis of separate clusters uncovered interesting facts, attributable to each cluster, and findings are summary follows. 9

We ran the correlation analysis for 2 models (Kilometers and Passengers) and 3 data clusters (Region, Model, Performance). Throughout the 16 correlations with various data clusters Operating Profit per Passenger and Number of Passengers were found to be the key variables. When describing the findings, we will focus on the passengers rather than the kilometers model. Below are some conclusions about the airline business in different classes and continents. 5.2.1. Region of origin and operations (Asia & Pacific and rest of the world, Europe, Americas) The analysis has been done on regional basis, namely 3 groups, which contain 40 annual cuts for Asia & Pacific and rest of the world, 35 annual cuts for Europe and 60 annual cuts for Americas. The first noticeable fact out of this analysis points to the high airfares in Asia compared with airfares in Europe and Americas. An important reason for this is the longer overall average length of passenger haul, attributable to this region. However, prices are higher, proven by higher Yields most probably through a lack of competition in this part of the world. This region demonstrates the greatest Operating profit per passenger despite the similarity in Operating margin % for Asia and Europe (6,9% and 6,1%). It means that carrying fewer passengers requires higher profit per passenger in order to cover costs and earn profit. Reviewing the total sample for profitability, we can now see the highest margins achieved by Asian companies and the lowest belong to American, with Europe in between. What is extremely interesting here, is very similar passenger revenues of 179,5$ in Europe and 172,9$ in Americas with noticeable difference in Operating profit per passenger (7,7$ and -0,2$ respectively) and an average Operating margin of 6% and 2%. For the European companies we found the following. There is a negative correlation between Load Factor and Revenue per Seat with the coefficient -0,411. Also, there is a negative correlation of -0,390 between Number of Passengers carried and Operating Profit per seat. Unlike Asia and America, in Europe Revenue and Profit drop while the Load Factor and Number of Passengers increase. This speaks for the competitiveness of European Market and represents an ideally functioning market, when passengers pay enough to cover costs and earn profit, that it is possible to decrease prices in order to handle competition. As we already mentioned, such a trend is not the case for Asia and America. Later we will go through some European companies, noting they are more successful than disastrous. In 2008, 6 of the top 10 carriers in FSNC class originated from North America, illustrating the importance of this mode of transport in the United States. The bulk of our sample belongs to American companies, which means the American national carriers dominate the sample. The maximal load factor 85,2% was noticed in Europe and minimal 61,6% in America. 5.2.2. Business model (FSNC versus LCC) The aim of this analysis confronts common opinion that LCCs fly with very cheap fares and are significantly more successful when compared with FSNCs, operating with high costs, less flexibility and less aggression. As it is immediately apparent, there is no significant difference in average load factors, nor in mean values or in the extremes. Thus, with the nearly identical mean of 77,3% for FSNCs and 77,9% for LCCs, the minimal value of 64,9% for FSNCs, which is even higher than 61,6% belonged to LCCs; however with the maximal load factor of 85,6% LCCs overcame FSNC‟s 82,7%. 10

Another interesting point that in spite of the range in passenger revenues (308,4$ versus 105,3$), which we rather predicted, operating profit per passenger – the value we are focusing on – is nearly identical and differs by only 9% (FSNCs 8,5$, LCCs 7,7$). For FSNCs there is a strong negative correlation between the number of passengers carried and Operating Profit per passenger as well as Operating Margin (Correlation coefficients 0,481 and -0,452 accordingly). In contrast, for LCCs correlations between the number of passengers and Operating profit per passenger as well as Operating profit percentage are insufficient, in other words for LCCs poor or high financial performance doesn’t depend on company’s scope of operations. This anomaly about profit decrease with increase of the number of passengers can be explained that FSNCs sample is dominated by US companies, and although carrying the most number of passengers, they are the poorest performers. Secondly, negative correlation between Load Factor and Operating Profit % (-0,259) is given for American companies only. In other words, in spite of high loads up to 85%, Amer ican companies did not manage to achieve operating profitability (decrease costs or increase revenues) unlike their Asian and European colleagues. Taking a closer look at operating profit per passenger in America we can see the mean -0,2$ compared with 7,7$ in Europe and 8,1$ industry average. We can also draw the conclusion that Load Factor is an important variable for LCCs profitability measured by Operating margin % and, ROA%. In contrast there is no strong corr elation of Load factor to profit for FSNC‟s. In turn, Revenue per seat is important for FSNC‟s profitability, but not that important for LCC‟s to achieve their financial targets. In other words, Low Cost carriers can afford to decrease prices for purposes of competition. Putting it all together we can say that LCCs with their thin margins and focus on earnings per passenger must watch their Load Factors, attracting more customers for the same number of flights. In other words, in order to manage the profit they must manage the key activity, and attract a sufficient number of passengers. Number of passengers is a key leverage for LCCs in conjunction with operating profit per passenger. FSNCs in turn appear have greater stability in number of passengers, and any efforts to increase their number or Load factor will not pay back if the revenue per passenger results in an inadequate operating profit. Moreover, if every passenger brings a negative profit (because of insufficient revenue), then multiplied by tenths of millions of passengers, their business results in financial disaster. We can reasonably conclude that for FSNCs the key leverage is Revenue per passenger or Operating profit per passenger in conjunction with passenger numbers. Finally, Beta negatively correlated with size (RPK) of FSNCs and indicates higher risk for bigger companies. For LCCs, however, lower costs and higher profits decrease the risk expressed in Beta. There were negative and positive profits in both classes FSNCs and LCCs; however on average they are both profitable, supporting the opinion that profitability does not depend on low or high cost operations, and suggesting the existence of one or more metrics, providing the right business focus that matters. 5.2.3. Financial performance (High performers versus Low performers) In this part of our analysis we sorted the data according to Operational margin %. At the mean value of 0,045 or 4,5% we separated the data under +0,045 and over +0,045. The result was 67 low performing annual cuts and 68 high performing annual cuts. 11

It is interesting to note the difference in mean Operating margin % and operating profit per passenger from -1,7% to 10,8% and -4,3$ to +20,4$, alongside with 206,2$ and 230$ revenue per passenger. Moreover, the smallest revenue per passenger of 42,8$ has been found among high performers, when at the same time, the lowest average revenue for low performers achieved 63,9$. We are going to analyze this connection between profit and revenue further, using correlation and regression analysis. Yields did not differ dramatically, but the highest Yield was found among Low performers. We can say the same about load factors which average between 76,6% and 78,5%. The, highest load factor of 85,6% was found among the low performers again. Operating costs per passenger for low performers were even 0,5$ lower than for high performers i.e. 210,8$ and 211,3$ respectively. Let’s summarize. Judging according to traditional KPI’s model, Low Performers appear to have out-performed High Performers, whereas in reality, they underperformed financially. In contrast, according to the passenger model, High performers, supported by higher operation margins, did better than Low performers. Drawing an advance conclusion from this part of analysis, we would note that financial performance for 27 companies studied out of 40 top global carriers, did not depend on the business model applied, or on the region of operation, nor on their size. And the traditional business drivers, such as Yield, Load factor, Air fare and costs did not appear to drive successful financial performance. Regarding geographical distribution and belonging to FSNC or LCC class for High and Low Performers, there is certain overbalance in that more failures can be found among FSNCs originating from America, however as previously stated, the big American FSNCs dominate our sample. There were winners and losers in all geographical areas, among traditional FSNCs and modern LCCs, flying with average load factor about 77,6%, charging on average 218$ and earning in average 8$ per passenger. In the following chapters we are going to use correlations and regression analysis in order to define what we claim to be a better model (Kilometers or Passenger) for airline financial performance.

6. Modeling KPI’s using regression analysis 6.1.

Models

For the regression analysis we suggest the existence of two KPIs models. In both models, the beta coefficients show that either “Operating profit per passenger” or “Operating profit per RPK” is shown to be the best predictor of firm performance. Operating Profit per Output appears to represent a powerful driver for firm success. First model traditionally suggests focusing on revenue increase (Yield) per item of c ostdriver (kilometer) in order to achieve financial targets. In this model management supposed to focus on increase of revenue and load factor as well as cost reduction. We will mention it further as Kilometers model. Second model suggests focusing on operational profit achievement, which expressed as a multiple of number of passengers carried and operating profit per passenger. In this case Operating Profit already incorporates Revenue and Cost, while Number of Passengers carried arrives from the relation between number of available seats and Load Factor. This model we will name as Passenger model. 12

For analytical purposes we also developed our own ratio, which connects operational and financial variables and helps us avoid absolute figures, differing from region to region. As two models involve different variables, then our ratios will be slightly differing from each other. For kilometers model it will be Revenue per ASK over Operating Cost per ASK and for the passenger model it is Operating cost per passenger over Revenue per passenger. We checked all the participating variables for multicolinearity and excluded variables stron gly correlating with the dependent variable. Below the present tests results for these two models in connection with profitability metrics, such as Operating margin %, Net margin %, Return on assets ROA % and Beta. The aim of analysis is to answer a question “Is Operating Profit per Passenger in conjunction with number of Passengers a better model than Yield (Revenue per RPK) in conjunction with number of RPKs for predicting airline profitability?” According to our data clustering, we have basically 12 models for each performance measure. The performance measures we used as dependent variable are Operating margin % and Return on Assets (ROA) %. Thus, out of total of 24 models we selected those, with the highest R squared and summarized in Table 2 (Operating Profit) and Table 3 (ROA). Table 2. Model Summary and Coefficients. Dependent variable Operating margin % Standardized Coefficients

Model

t

Beta Kilometers Model TOTAL sample Kilometers Model Europe Passenger Model Low Performers

Kilometers Model Low Performers Passenger model FSNC

Kilometers Model FSNC Passenger model LCC Kilometers Model LCC

(Constant) RPK million Operating Profit per RPK (Constant) YIELD, cents € Operating Profit per RPK (Constant) Number of passengers Revenue per seat sold Operating profit per passenger (Constant) Operating Profit per RPK (Constant) Number of passengers Revenue per seat sold Operating profit per passenger (Constant) Load Factor Operating Profit per RPK (Constant) Revenue per seat sold Operating profit per passenger (Constant) Operating Profit per RPK Rev. ASK/cost ASK

-,087 ,881 -,243 ,858 -,122 ,138 ,902 ,939

-,093 -,170 ,958 -,031 ,958 -,392 ,896 ,491 ,443

3,468 -2,329 23,498 3,022 -4,055 14,323 -,822 -2,223 2,508 16,890 -,435 22,001 2,300 -1,630 -3,001 18,438 ,941 -,926 28,726 6,323 -5,508 12,571 -3,923 4,734 4,276

Adjusted R Square

,833

,884

,821

,880 ,858

,918

,729

,783

Judging by the higher R-squared and in comparison with ROA the Operating Margin model is found to be a better model explaining variation in the performance of airlines. Out of our analysis Operating Profits per output sold (RPK or Passenger) is the dominant variable in explaining firm performance. However in the cases of ROA models, we can see more clearly potential impact of size as reflected by the “Number of Passengers” variable. 13

In the ROA models, number of passengers and Yield appear in a number of instances as important variables. As well as in above described correlations, in most cases the number of passengers variable has a negative relation to performance, suggesting that smaller airlines are more likely to be profitable than larger airlines. Indeed, it is generally easier to stay focused in a small company rather than in large one. However, it is not the case in our example – as we already explained in our sample - the biggest carriers originate from America and they are mostly unprofitable, while more successful carriers from other regions are several times smaller. Table 3. Model Summary and Coefficients. Dependent variable ROA % Standardized Coefficients

Model Passenger model, Total sample

Kilometers model, Total sample Passenger Model High Performers Kilometers Model High Performers Passenger Model Low Performers Kilometers Model Low Performers Passenger Model FSNC Kilometers Model FSNC Passenger Model LCC

Kilometers Model LCC Passengers model Europe

Kilometers model Europe

6.2.

t

Beta (Constant) Operating profit per passenger Operating Cost/Revenue per passenger (Constant) RPK million Operating Profit per RPK (Constant) Number of passengers Operating profit per passenger (Constant) Rev. ASK/cost ASK (Constant) Number of passengers Operating profit per passenger (Constant) RPK million Operating Profit per RPK (Constant) Operating profit per passenger (Constant) Operating Profit per RPK (Constant) Revenue per seat sold Operating profit per passenger (Constant) YIELD, cents € Operating Profit per RPK (Constant) Number of passengers Operating profit per passenger Operating Cost/Revenue per passenger (Constant) Operating Profit per RPK YIELD, cents €

,618 ,182 ,062 ,771

-,293 ,341 ,508 ,158 ,708

,177 ,738 ,841

,845 -,393 ,665 -,360 ,811

,133 ,399 ,381 ,711 ,093

-2,645 8,567 2,520 -2,123 1,013 12,687 5,262 -2,632 3,059 -3,415 4,793 -3,009 1,740 7,792 -3,286 1,961 8,171 -2,516 13,268 -1,317 13,508 2,943 -3,758 6,356 2,610 -3,344 7,527 -2,528 ,917 2,691 2,655 -,736 5,447 ,712

Adjusted R Square ,539

,560

,238

,247

,471

,495

,703

,710 ,415

,482

,348

,451

Full sample

The regression analysis shown that kilometers model involving Operating profit per RPK and number of RPK fits the industry better than the passenger model, involving Operating Profit per passenger. This conclusion is based on regression coefficients in both – Operating Margin and ROA models. Table 3 does not contain data on Passenger model for Full sample, because key coefficients were insufficient comparing with other models. 14

One of possible explanations why RPK might predict firm‟s performance better is that unlike number of passengers, RPK not only incorporates load factor, but also contains such important numbers as average haul length, aircrafts number/size, and other important numbers and thus, characterize the industry better. Another explanation is still there – American companies are the biggest, FSNCs, Low Performers and they dominate the sample. This can be valid if see the better fit of Kilometers model for Low Performers and FSNCs as well. 6.3.

High performers – Low performers

A very interesting and somewhat surprising result is found when we compare high versus low performing carriers. For both dependent variables, Operating profit and ROA, we find that we are able much better explain the performance of low performing airlines than the high performing ones. For both models, we observe a much lower R-squared for the high performing airlines than the regressions with low performing ones. While surprising at first glance, a possible explanation for this result might be that for the large, traditional and likely unprofitable airlines, use traditional financial performance metrics focusing on Yield, Kilometers and Load Factor are probably all what they are guided by. Again, large and unprofitable airlines originate mainly from America and successful European carriers (both FSNC and LCC) are of smaller size. In our case most large, traditional airlines appeared to be low performers and therefore traditional airlines performance model should explain their behavior better than the relatively fewer high performers which probably are more creative and expand traditional airline performance models with better KPIs. For example, US-operating LCC Frontier experienced loses over years. Looking into financial and operational statements it is noticeable that company was heavily focused on Load Factor increase. For the year ending Dec 31st, 2006, management committed to improve financial performance. Indeed, in 2007 the load factor increased dramatically from 75,2% to 79,6%. Unfortunately, company failed to control revenues and profits and thus went into even greater loses. On April 10, 2008, Frontier Airlines filed for Chapter 11 bankruptcy to solve financial issues and to ensure long-term viability. The airline continued its operations uninterrupted. In June 2008 Frontier announced a 17% reduction of its scheduled flights. 6.4.

FCNC versus LCC

Our models using Operating Margins as the dependent variable representing performance is shown to describe the performance of Full service carriers much better than low cost carriers. (Compare the R-squared of 0,918 and 0,858 with 0,783 and 0,729). Closer look at Operating Profit table shows that Kilometers model better fits FSNC sample while Passenger model better explains performance of LCCs, which was rather expected. According to theorists, „per passenger‟ models are used for single leg „Origin-Destination‟ routes, which is usually the case for LCCs and not the case for FSNCs. However, what we can also see from analysis for FSNCs in Operating Profit Table, that Operating profit per Passenger and Operating profit per RPK got same Beta of 0,958 and different but high enough t (18.438 and 28.726 accordingly). In other words, Kilometers model with Operating Profit per RPK better fits FSNCs, whereas Passenger model with Operating profit per Passenger fits both airlines classes – LCCs as well as FSNC. Judging from the Adjusted R Square, passenger model fits FSNCs sample (0,858) even better that it does LCCs (0,729), despite the traditional view that passenger model can be used only by LCCs providing single point to point destination services. 15

7. Conclusions and observations This paper is an empirical study that attempts to identify a reasonable model that may better help us predicate airline financial performance. In doing so we looked at the most commonly used metrics in the airline industry, specifically we examined the effectiveness of models that characterizes performance based on two activity drivers – passengers and kilometers, passengers based indicators represent revenue driving ones and kilometers based – cost driven. Our main results indicate that the kilometers model fits the industry slightly better than the passenger model, however we discovered that passenger model can be applied to FSNCs despite the traditional view that it is appropriate mainly for LCCs. Furthermore, we find strong evidence indicating that Operating Profit per Passenger or per Passenger-Kilometer is the most significant variable predicting airline profitability. It was found to be more important than revenue, unit cost and load factor traditionally used by the industry. In our analysis we did not find any significant correlation between size, business model or region, which would explain low or high profitability of an airline. 7.1.

“Seats” and “Passenger-kilometers” are both good as output target.

In our analysis we did not find that Seats is a better denominator as Passenger-kilometers. Certain correlation indexes between Profitability measures and Operating Profit per Passenger were higher than between Profit and Operating Profit per RPK. In turn, the regression analysis shown that Operating Profit per passenger-kilometer fits the industry better. Judging from the Adjusted R Square, passenger model fits FSNCs sample (0,858) even better that it does LCCs (0,729), despite the traditional view that passenger model can be used only by LCCs providing single point to point destination services. What we also found out of the regression analysis is that Operating Profit per Passenger is almost as good as Operating Profit per RPK. In light of the above this should be the most important finding. 7.2.

Operating profit per output is the most important variable.

Out of all the variables used for the analysis, we defined Operating Profit per RPK or per Passenger as the most important variable. Even though there was a significant correlation between Revenue per seat and profitability of FSNCs, it was the fact for this business model only. At the same time for the same class Operating Profit per seat and per RPK was found to be not less important than Revenue. Indeed, the importance of sufficient Revenue per passenger shows that company should price the airfares and target ancillaries in a way that permits covering cost and achieving planned profitability. Analysis of all the clusters, including FCNC, shown the importance of Operating profit per output (Passengers or Kilometers) and have not shown the importance of Revenue for operating and financial profitability. This finding is supported also by fact that in spite of the range in passenger revenues (FSNCs 308,4$ versus LCCs 105,3$) operating profit per passenger – the value we were focusing on – was nearly identical and differs by only 9% (FSNCs 8,5$ versus LCCs 7,7$). In even greater detail Table 4 contains selected key data for two very different airline companies, which in terms of profitability are very successful and are the 1st and 2nd large European carriers in their segments. The first part of Table 4 highlights the difference in scope of these two businesses. However, when it comes to operating profit per passenger, the figures are surprisingly equal. Schematically we depicted this difference in Diagram 1. 16

Table 4. Comparison of key financial data for two European air carriers, Ryanair and Lufthansa Ryanair

Lufthansa*

Europe LCC 2 714 2 177 537 50,931

Region of operations Business model Passenger Revenue, m. EUR Total Costs, m. EUR Operating Profit, m. EUR Passengers Carried, m. EUR Number of passenger aircrafts Average revenue per passenger

163 EUR 53,3

Europe FSNC 18 393 17 671 261 70,543 494 EUR 260,7

Average passenger fare

EUR 43,7

EUR 238,9

Operating Profit / passenger

EUR 10,6

EUR 10,2

*Lufthansa Group, passenger segment

This comparison has been made not to judge the success of one or another airline comp any, but to demonstrate how we can define the key driving activity and drill into the heart of the business, deriving KPI‟s which pinpoint and focus on the business goal achievement. In other words, if we liken the airline business to a machine driven by passengers, we find the operating profit from one turn of a small machine equals the operating profit from one turn of a big machine. Traditionally in their management accounts FSNC‟s consider that the machine is driven by kilometers and LCC‟s tend to view the machine as driven by passengers. We believe this difference is an important part of the success secret applicable to business in general. 7.3.

Unlike Low Performers, financially successful companies were consistent in focusing on Operating profit per passenger and number of Passengers.

The outlook of the financial statements of all 27 companies demonstrated rather stable relationship between Profit and KPI‟s. While the majority of companies achieved consistent gradual increase in Yields, RPKs and Load Factors, only few of them improved profitability. 17

During the data collection we did ensured that financial data is more focused in Europe and is of better quality (in terms of KPIs). Separating the data into High and Low Performing we searched for certain consistency in focusing on Operating Profit increase. The majority of high performing companies demonstrated this according to interplay of operating profit per Passenger and Number of passengers. In other words, by increasing its Total Operating Profit, a company can increase Operating profit per Passenger or Number of Passengers (Flights) or increase both. Some real examples are summarized in Table 5. Irish discounter Ryanair demonstrated its dedication to aggressive growth (Number of passengers increases about 20% annually) and strong focus on Operating Profit per Passenger (~ 11 US dollars). Its competitor FSNC British Airways in contrast, even slightly decreased number of passengers, probably optimizing the routes, but focused on increase and maximization of Operating profit per passenger. As an illustration, BA rec ently offered regular service between London and New York with business class seats on board exclusively. Finally, another British LCC easyJet demonstrated slight but stable increase in both – Operating Profit (~ 2,8 dollars) and Number of Passengers (~ 15% per annum), see Table 5.

Table 5. Financial KPIs for selected high performing airlines Variable

2004

2005

2006

2007

2008

205,25 203,63 15,12 35 717 107 892 74,83% 7,2

233,57 231,84 21,40 32 432 109 713 76,09% 7,5

238,72 239,99 18,20 33 068 112 851 76,09% 7,5

245,38 237,57 26,39 33 161 113 016 75,57% 7,7

251,20 278,17 6,64 33 117 114 346 77,00% 7,9

2004

2005

2006

2007

2008

44,80 42,73 2,07 24 351 21 562 84,73% 5,1

45,38 43,14 2,24 29 562 27 448 85,40% 4,9

49,13 45,86 3,27 32 969 31 621 85,26% 5,1

48,29 44,04 4,26 37 216 36 976 85,00% 4,9

54,07 51,99 2,08 43 700 47 690 85,64% 5,0

2004

2005

2006

2007

2008

48,44 36,49 11,94 27 594 24 003 83,75% 5,6

48,68 37,89 10,79 34 769 32 731 83,78% 5,2

52,62 41,52 11,10 42 509 43 352 84,09% 5,2

53,28 42,74 10,55 50 931 55 434 83,34% 4,9

50,23 47,77 2,46 58 566 63 076 83,23% 4,7

BritishAirways Revenue per seat sold Operating cost per passenger Operating profit per passenger Passengers, thousands Revenue passenger-km RPK, mill Passenger load factor % YIELD, cents € Variable

easyJet Revenue per seat sold Operating cost per passenger Operating profit per passenger Passengers, thousands Revenue passenger-km RPK, mill Passenger load factor % YIELD, cents € Variable

Ryanair Revenue per seat sold Operating cost per passenger Operating profit per passenger Passengers, thousands Revenue passenger-km RPK, mill Passenger load factor % YIELD, cents €

In contrast, there was no single airline out of the poor performing ones, which could display such consistency per passenger performance over 5 years. However, as we mentioned in previous chapters, highest Yields and Load Factors were attributable to Low performing airlines. Also, it was a consistent increase or stability over 5 years for these metrics, unlike per passenger ones. (Refer chapter 5, p.5.1 about the absence of passengers‟ number in long reports). 18

On the other hand, some companies, improving Total Operating Profit, tried to increase Operating profit per Passenger or Number of Passengers (Flights) or increase both (Table 5). Such consistent patterns were attributable only to High Performers and none of the poor performers demonstrated this in financial statements. Moreover, companies with above mentioned regularity had better share price performance than even profitable airlines without such regularity.

8. Recommendations. Out of the above summary we can accept that Operating Profit per unit of output is a stronger Performance indicator than Revenue per unit of output. The suggested denominator (Passengers) is equally good as passenger-kilometers and can be used independently of an airline‟s business model. Focus on number of Passengers and Target per passenger helped successful airlines to improve operating profitability and create company value. Looking at Total Operating Profit (OP) as a multiple we can formulate it in two ways.

OP = OP/RPK x RPK

or

OP=OP/PAX x PAX

were OP/RPK is average Operating Profit per 1 Revenue Passenger-kilometer OP/PAX is average Operating Profit per Passenger carried, PAX is Number of Passengers carried

Indeed, both multiplications result in Total Operating Profit, however targeting for example 10$ per passenger is more meaningful and clear rather than targeting 1 cent of Operating Profit per RPK. Being experts in their field and focusing on such figures, staff receive immediate awareness how they can improve 10$ per passenger and thus help the company achieving its financial profitability. For example, mean operating profit in America in 2004-2008 achieved 0,2$ per passenger compared with 7,7$ in Europe and 8,1$ industry average. A revenue increase about 8,0$ per passenger would hardly make a difference for the customer. Taking into account geographical specific of the US market where companies often have monopoly on routes, it raises the question: how difficult would it be to get extra 8$ per passenger in revenue? Increase in revenues does not necessary mean an increase in airfares. Many LCCs fly with zero or negative profits and earn from ancillaries. Following example data illustrates how every extra dollar in revenue multiplies profit by 51 million passengers. The example is made using the data taken from Irish discounter Ryanair (biggest LCC in Europe) 2007 reports: Table 6. Structure of passenger revenue, example of RYANAIR, 2007 Revenue per passenger (sales)€ of which:

53,3 of which:

Airfare € Ancillaries €

Operating costs € 43,7 9,6 Operating profit per passenger €

42,75 10,55

Number of passengers in 2007 – 50.931 thousands Table 7. Structure of operating profit example of Ryanair, 2007 19

Operating profit per passenger

10,55 €

of wich:

9,60 €

Ancillaries, of wich: Non-flight Scheduled

6,58 €

Profit from air fare

0,95 €

Car Rental

0,50 €

Ancillaries

9,60 €

In-flight Sales

1,44 €

Internet-related

1,08 €

On the other hand, flying with such thin margins is dangerous – should the number of passengers dropped, profit or loss will be leveraged accordingly. Nevertheless, recently American FSNCs successfully picked up some ideas from their LCCs competitors such as implementing ticketing and check-in luggage fees, which helps them now raising billions in ancillary revenue. Indeed, carrying most of the passengers worldwide, even 1 dollar in revenue per passenger multiplies by the greatest leverage. To apply such approach successfully one should really question himself: “What is my leverage and what really drives my business?” In other words, what is the effort and what the payback for me to attract more passengers or getting extra dollar in sales from each passenger? For example, FSNCs have greater stability in number of passengers, and any efforts to increase their number or Load factor will not pay back if the revenue per passenger r esults in an inadequate operating profit. Moreover, if every passenger brings a negative profit (because of insufficient revenue), then multiplied by tenths of millions of passengers, their business results in financial disaster. If we accept Passengers drive airline business, then we are looking for an overall target average profit per passenger by passenger numbers. This is then broken down to fit the various routes, flights, classes and load factors. Not denying the importance of per RPK measures, we suggest for big traditional companies to use Operating Profit per passenger carried or per Seat sold as a tool in achieving Operating profit RPK. Such a simple and meaningful benchmark would be very helpful in developing and taking strategic decisions towards revenue and profit increase. Operating Profit per Passenger in conjunction with Number of Passengers carried is a fundamental KPI, which we recommend to use for analysis, planning, benchmarking and certainly for internal reporting, in order to achieve operating profitability. Regular reports play valuable role in keeping the business on track. Airlines do track revenue, number of passengers and load factors quite frequently. If Average Operating Profit per Passenger will become a part of revenue management system this would greatly assist poor performing companies. Apart from revenue management, average Operating Profit per passenger can be tracked on daily and weekly basis against targets.

20

References AEA: Annual Reports (www.aea.be) Agatz, N., Campbell, A., Fleischmann, M., Nunen, J., Savelsbergh, M. (2008) "What E-Tailers Can Learn From Airline Pricing" Wall Street Journal, New York, N.Y Air Transport and Airport Research (2008) "Annual analyses of the European air transport market, Annual Report 2007" German Aerospace CenterDeutsches Zentrum für Luft- und Raumfahrt e.V. Air Transport and Airport Research (2009) "Annual analyses of the European air transport market, Annual Report 2008" German Aerospace CenterDeutsches Zentrum für Luft- und Raumfahrt e.V. Alexis George de Borges, P. and A. E. S. Respicio, Jr. (2004). "Developing a fleet standardization index for airline pricing." Journal of Air Transportation Anderson, C. K. (2002). "Revenue management: Issues in pricing and allocation." Anonymous (2009) "The outlook for travel", Airline Business, Feb 2009 Borestein, S., N. L. R. (1994). "Competition and price dispesion in the U.S. Airline Industry." Jornal of Political Economy Brueckner J. K., Lee D., Singer E. (2010). “Network vs. LCC competition and airfares: New US empirical evidence”. IATA eAnalyst Carpenter, G. S. and D. M. Hanssens (1994). "Market expansion, cannibalization, and international airline pricing strategy." International Journal of Forecasting Christian, H., D. Martin, et al. (2005). "Financial Distress and US Airline Fares." Journal of Transport Economics and Policy Cleland, K, (1998) "Contribution Based Activity – the flip side of Activity Based Costing"; CIMA Cleland, K, (2004) "As Easy as CBA – Limitations of Time-Driven ABC from a CBA perspective"; CIMA Curry, R. E. (2001). "A market-level pricing model for airlines." International Journal of Services Technology and Management Doganis, R. (2002). "Flying off Course – The Economics of International Airlines". Second Edition. New York: Routledge. Doganis, R.: The Airline Business 2nd ed. London 2006 Douglas, G. W. and J. C. Miller, II (1974). "The CAB's domestic passenger fare investigation." The Bell journal of Economics and Management Science Drury, C., (1994), "Management Accounting for Business Decisions", Thomson, London Dunleavy, H.; Phillips, Glen (2009) "The future of airline revenue management". Journal of Revenue and Pricing Management Ezard K. (2009) "Copy CATS", Airline Business First Research Industry Profiles (2009) "Airlines - Quarterly Update 11/16/2009." Graff, J. (2008). "Revenue management for the whole aircraft: Coordinating acceptance decisions for passenger and cargo transportation." Journal of Revenue and Pricing Management 21

Huefner, R. J, James A Largay III, (2008). "The role of accounting information in revenue management Business Horizons." Greenwich IATA Economics, (2009) "IATA 2009-10 forecasts and some 2008 estimates." IATA economic briefing, February 2010, “Airline Profitability - size and business model – 2008” IATA: Annual Reports (www.iata.org) Ian, D. (2005). "Do business class air fares capture added value for airline alliance membership?" Journal of Revenue and Pricing Management Journal of Air Transport Management Kaplan, R. S. and Anderson, Steven R., (2003), "Time-Driven Activity-Based Costing." Kaplan, R.S. and Bruns W, (1987) "Accounting and Management: A Field Study Perspective", Harvard Business School Press L.Phillips, R., Ed. (2005). "Pricing and Revenue optimisation" McCrary, M. (2009), "Enhanced customer targeting with multi-stage models: Predicting customer sales and profit in the retail industry", Journal of Targeting, Measurement and Analysis for Marketing McGill, J. I, Van Ryzin G. J., (1999), "Revenue Management: Research Overview and Prospects" Michael, Z. F. L. (2006). "A model of pricing perishable inventories using two restrictions, with an application to airline pricing." Journal of Revenue and Pricing Management Mike, M. (2003). "The issues and challenges of a pricing system implementation." Journal of Revenue and Pricing Management Morrison, S. A. and C. Winston (1990). "The Dynamics Of Airline Pricing And Competition." The American Economic Review Pompl, W., Schuckert, M., Möller, C. (2004) "The Future of Small and Medium Sized Airlines in Europe", in: AIEST (Hrsg.), The future of SME's in Tourism, St. Gallen 2004 Reeves, M., Deimler. M. S. (2009), "Strategies for winning in the current and post-recession environment." Strategy & Leadership Richard, R. and V. Ben (2005). "Airline pricing and revenue management: A future outlook." Journal of Revenue and Pricing Management Shields, J. F., Shields, M. D. (2005) "Revenue drivers: Reviewing and extending the accounting literature", Advances in Management Accounting, Emerald Group Publishing Limited Theodore, C. B. (2000). "Airline fare product design in the context of yield management." International Journal of Services Technology and Management Theodore, C. B. and P. P. Belobaba (1999). "Airline pricing and fare product differentiation: A new theoretical framework." The Journal of the Operational Research Society Whinston, M. D. and S. C. Collins (1992). "Entry and Competitive Structure in Deregulated Airline Markets: An Event Study Analysis of People Express." The Rand Journal of Economics William N.Evans, I. K. K. (1994). "Living by the "Golden Rule": multimarket contact in the U.S. airline industry." Jornal of Political Economy Yeoman, I. (2008). "Book Review; The Future of Pricing: How Airlines Pricing Has Inspired a Revolution." Journal of Revenue and Pricing Management 22