2. SESAR AND AIRLINES

Why is the focus here on airlines and SESAR? The reasons are that the core of ATM system design is to meet airline/passenger needs, that airlines make the bulk of the direct costs and contributions to air traffic control (ATC) route charges, and that airline activity generates substantial society/GDP gains. Changes to the design of the ATM system have to meet commercial aviation's needs – and airlines operate in a challenging commercial market. However, there are massive implications for military and business aviation (SDG, 2005). Progress with SESAR will only happen if all the main stakeholders agree to invest money in its core projects.

Features of SESAR include (SESAR Consortium, 2007):

• • Gate-to-gate system integration

• • Change from reactive ATM to anticipatory ATM

• • Network of ground-to-air data links to enable accurate ‘trajectory’ information exchanges

• • System-Wide Information Management & Interoperability

• • Exploit satellite navigation/communications technology

• • Co-operative:

  • ∘ 4D trajectory planning & support tools

  • ∘ New roles & task distribution for pilots & controllers

  • ∘ Airborne separation assistance

  • ∘ Collaborative Decision-making (ATM/Airlines/Airport)

Almost all of the technologies encompassed in SESAR have long histories of successful research, often by programmes covering several countries. Thus, many of the issues are about which R&D should lead to implemented operational systems. SESAR's objectives appear perfectly feasible, but the big questions are about how to achieve this through optimal-chosen technical investment paths. However, note that twenty years ago, the fourth and fifth in the list above would not be there – intranet-type systems are now a backbone for everything, and satellites are.

The tangible – ‘hard cash’ – financial benefits to airlines from SESAR potentially arise from:

• • Reduced fuel usage: aircraft can fly more direct, better flight-profile, more fuel-efficient routes.

• • Reduced Flight Delays: reduce short-term imbalances between demand and capacity.

• • Reduced ATM costs: improve productivity of ATM system capacity, e.g. by minimising growth in the required number of controller working positions, and hence reducing ANSP charges for flight.

• • Remove barriers to peak-hour flights: increased airport traffic is not constrained by airspace capacity.

There are four ‘SESAR Performance Objectives’ (and many other performance criteria):

• • Designed for more capacity: +73% in 2020 (compared to the 2005 situation) … and enabling three times in the longer term.

• • Improved safety: three times for 2020 … 10 times in the longer term.

• • 10% less environmental impact/flight due to ATM.

• • 50% less [direct] ATM cost/flight.

These are very demanding goals.

SESAR is already an enormously complex exercise. The SESAR Consortium produces a large number of substantial documents. It would be unusual for airline decision-makers – the people who actually authorise investment spending – to have a detailed technical knowledge of all the concepts potentially incorporated in SESAR. Airlines are concerned with successfully ‘packaging’ assets and services to make their living, rather than developing technology or software. Thus, airlines buy COTS technology and services – most obviously, aircraft, outsource catering, and use commercial computer reservation systems. To an airline, ATM is part of the business in much the same way that a gas cooker is to a restaurant owner. Does it do a necessary job? Is it good value? The kinds of questions an airline decision-maker asks about ATM spending and SESAR are business-focused:

• • What new kit spends are needed – and how do we know it will work?

• • What are the cash benefits from this spend on new kit?

• • What commercial/business environment risks do we need to mitigate?

• • Are we dependent on new ground ATC equipment to be in place?

• • Are the ANSPs going to charge us a great deal more for the necessary ground investments, training and operational staff?

Questions about potential cash benefits raise a larger set of questions about investment and uncertainty. Airlines will want to make decisions based on good information. Few of SESAR's ideas are very novel: they have already been the subject of research and development, in some cases over many years. For example, the PHARE programme of the early 1990s demonstrated in near-operational trials that aircraft could be flown on fuel-efficient 4D trajectories whilst subject to safety constraints to prevent flight conflicts (e.g. see PHARE (2008)). Further examples are discussed below.

3. AIRLINE INDUSTRY ISSUES

Airline profitability in recent decades has been subject to major problems. But now there are the effects of an oil price shock, with prices progressively rising since about 2001, and the ‘knock on’ effects of the 2007/2008 world financial and economic crisis. Airline financial decision-making is very different from that of ATC suppliers. The product is a commodity. There is a great deal of competition for customers, many of whom are price sensitive. Airlines have huge fixed costs.

The yearly figures in Figure 1 show how bad the 2001–2008 oil price shock has been. People now hope that oil prices have reached their peak, and the price will adjust down to markedly less than $100 a barrel over the longer term – although worldwide GDP falls will tend to bring the price down in the near (?) future. But jet kerosene costs now form a larger part of operating costs. Airline fuel costs will also increase when the EU's Emissions Trading Scheme (ETS) comes into operation in 2012 (EC, 2008).

Figure 1. Crude Oil Price History – and Shocks (2007$: EIA, 2008 plus notional $135/barrel for 2008).

Airlines as a whole do not consistently make profits. Figure 2 (derived from Jiang and Hansman, Reference Jiang and Hansman2006) shows an empirical best-fit to the profitability of the world airline industry, using profit data for 1978–2003. The fit is a sinusoidal curve, with an ~11-year cycle, overlaid by exponential growth. The airline profit system is not stable, although it might have been before USA airline deregulation in the late 1970s. In fact, recent data on profitability mean that reality is far less attractive than even this historically based fit would imply. Worldwide financial problems since 2007 have affected the world economy, and there has been a shock of huge increases in fuel prices since 2001. Projected $20Bn profits for 2008 exist only in the airlines' dreams.

Figure 2. Projected (2004) World Airline Profitability.

An unstable growth path is usually an indicator of problems with system lags. Jiang and Hansman developed a system dynamics model. They included two hypotheses: lag in capacity response and lag in cost adjustment. For example, their model of capacity response indicated that the system stability depends on the delay between aircraft orders and deliveries and the aggressiveness in fleet ordering. They judge that their coupled model with these two factors matched the industry dynamics reasonably well. Jiang and Hansman comment: “It is clear that future industry growth will be limited at some point, probably by capital investment as the industry becomes less-appealing to investors due to losses in the down cycle, and/or by capacity and traffic demand as the system reaches the limit of the national aerospace system in the up cycle.” An unstable financial system with periodic exponential oscillations is not an attractive proposition for most investors. Airlines are unusual firms because of the history of state ownership and continued strategic holdings by major investors. For example Turner and Morrell (Reference Turner and Morrell2003) present data on the concentration of ownership for some major airlines, which is in turn associated with very low volumes of trading of shares on stock markets.

The difficulty of surviving in the airline business is apparent from the fact that Wikipedia has a ‘list of defunct airlines’. Figure 3 presents the statistical data on defunct UK airlines' duration in years. This can only be a rough indication of airline lifetimes, given that ‘defunct’ does not have a precise definition. The evidence is that large proportion of airlines fail in some fashion during their first few years. The median lifetime of defunct UK airlines to mid-2008 (i.e. before the late-2008 combination of oil and financial problems for the airlines) was about nine years. Some defunct UK airlines operated quite large fleets of aircraft, e.g. Air Europe. Airlines become defunct for a variety of reasons, but common phrases used include short term cash flow problems', ‘a major, unforeseen downturn in traffic as a result of recessionary economic conditions’, undercapitalisation, unsound financial structure, financially overextended, high-risk strategy. Cash flow problems, i.e. not be able to pay bills for fuel, wages, etc, are always a ‘final cause’, rather in the same way that heart failure is the final cause of death.

Figure 3. Defunct Post-war UK Airlines (Mid-year 2008 data).

By mid-2008, fewer than 20 UK independent companies currently operate passenger services (from Civil Aviation Authority – CAA – and website searches). They roughly divide into traditional operators, charter firm companies, and low-cost operators. These three groups have very different business models. Successful traditional airlines tend to have valuable strategic operational assets, such as route licences for North Atlantic flights, slots at Heathrow airport or regional niche businesses. The median age of the ~25 currently operating UK passenger airlines, including subsidiary companies, is about 20 years. Just five of these airlines were more than 30 years old. Worldwide, even the largest airlines have difficulty in maintaining financial viability for long periods. The USA probably provides the most dramatic evidence. Its various bankruptcy laws covering airlines provide for liquidation, but more frequently for reorganization (chapter 11 protection). In 2005, four of the top seven carriers in the USA were under bankruptcy protection.

The oil price shock and the ETS, coupled with the 2008 financial/economic crisis and low GDP growth, means that fuel economy is now a strategic high priority and that the likelihood of a high growth scenario for air travel is lower. Low growth and higher costs tend to reduce cash flow benefits from new technology. This implies the need to look very carefully at SESAR's flexibility and the phasing of its project components.

4. CORPORATE FINANCE APPROACH TO ATM/SESAR

Which R&D & implementation projects should decision-makers chose? Which system functions need replacement, overlaid subsystems or must be entirely new? The approach here is to use some simple corporate finance ideas:

• • Examine different viewpoints & business environments of ANSPs & individual airlines.

• • Focus on airspace capacity.

• • Focus on how choice of R&D & implementation projects makes business sense.

Businesses make choices based on the information available to them about costs and benefits, in the context of general strategies. ANSPs mainly exist to serve airline needs based on large/medium airports: without airlines, the ATM system serving military and business/general aviation would surely be very different. ANSPs provide today's ATC services but they also invest in new ground equipment to meet tomorrow's needs. They are essentially acting as ‘ATM agents’ for the airlines in procuring the necessary ground ATM systems. ANSPs need to convince airline customers that capital expenditure plans are necessary to meet future traffic growth cost-effectively. Capital investments do not automatically reduce unit costs to customers: the rationale for an investment is that it is a better bargain for the future than other options. The impact on future prices to ATM system users depends on factors such as operational efficiencies and demand growth.

ANSPs are not normal commercial (profit maximising and economically unregulated) companies, i.e. equivalent to individual airlines. ANSPs are not there to make big profits for the state or their owners. This would lead to investment problems (e.g. see Brooker, Reference Brooker2004) – a reason why the UK Government set up NATS as a Public-Private Partnership.

5. CORPORATE FINANCE TOOLS FOR COST BENEFIT ANALYSIS

Table 1 summarises some appropriate corporate finance tools for cost benefit analysis (CBA) in the widest sense (Brealey et al, Reference Brealey, Myers and Allen2007).

Table 1. Investment Decisions: Corporate Finance Tools.

Modern large-scale businesses use Net Present Value (NPV) calculations to assist in these processes (Internal Rate of Return, IRR, closely related to NPV, is also widely used). NPV focuses on the present and future pluses and minuses of flows of hard cash (in constant price levels) in future years: money flows. The general calculation for a NPV is (the individual terms are ‘Present Values’ – PV_i):

(1)

• B_i, D_i, C_i=Benefits, Disbenefits, Costs in year i

• r%=Discount rate

• Summation over years 0 to n

When a business faces choices, the best investment according to this technique is the one that produces the highest positive NPV. The sum ranges over n+1 years – from 0 to n: this is to cover the case when there is an immediate investment – year 0. A cost would be an actual expenditure of cash to secure the investment (e.g. on maintaining equipment) while a disbenefit would be an estimated operational cost arising out of the investment (e.g. increased fuel usage).

It would be fanciful to forecast cash flows for an indefinite number of future years. Usually, there is a cash flow NPV planning – ‘horizon’ – period of several years, supplemented by a ‘Terminal Value’. The latter is an approximate value of all PV-ed cash flows after the horizon year. A rough estimate has to suffice because cash flows many years into the future become progressively more uncertain – although ATM NPV calculations often attempt very long-term estimates by extrapolating specific models of traffic and capacity. The terminal value calculation can use various assumptions about growth rate, e.g. using a constant growth rate [Gordon] model (Table 2). The Gordon model equation makes it apparent that the assumed net benefit growth rate (g) post the horizon is a critical parameter.

Table 2. Example of Airline ‘Reality’ Cash Flows, NPV, Payback Period.

Payback Period=3 years (ie sum of CF_is for years 1 to 3 equals the £180K initial investment)

PV_i definition in equation (1)

Assumed hurdle rate r−20%; assumed long-term growth g of benefits – 5%

NPV to five years=sum of PV₀ (i.e. −£180K) plus PV₁ to PV₅)

Terminal value (Gordon model)−PV₅ x (1+g) / (r−g) [a rough estimate at best]

NPV including terminal value=£271·8K

A firm has two crucial decisions when using NPV: deciding on the hurdle rate r% and the investment horizon n. Corporate finance texts usually recommend that the hurdle rate be set at the Weighted Average Costs of Capital (WACC):

• R_d=Company borrowing rate

• R_e=Shareholders' expected return on equity

• DE=Debt

• EQ=Equity

The company borrowing rate R_d – cost of debt – is a function of factors such as the financial condition of the business and the perceived quality of its managers. For example, contractors working for the UK CAA's regulatory assessment of the UK's en route ATC suppler NATS (PwC, 2004) estimated NATS cost of debt as 6·13% in real terms, 1·2% above a risk-free rate on government borrowings.

The estimation of the equity return figure R_e is complex. The value is a function of the risks that investors associate with the company's normal decisions. This most often makes use of the Capital Asset Pricing Model (CAPM):

• R_f=Risk-free Rate

• β=Equity Beta

• ERP=Equity Risk Premium (R_m – R_f)

Here R_m is the (expected future) return on shares in the stock market and Beta measures the volatility of the company's shares compared with the market generally. A high Beta – i.e. β>1 – is a higher risk stock, a low Beta – i.e. β<1 – is a lower risk stock.

Figure 4 is the well-known graph from corporate finance textbooks (e.g. Brealey et al, Reference Brealey, Myers and Allen2007). It shows the rate of return required on investments versus Beta. Beta measures the volatility in a firm's return on shares associated with general movements in the stock market and the economy. Airlines have effective Betas greater than 1, ANSPs probably no higher than 1. This simply means that airlines require higher rates of return on investments than ANSPs (PwC, 2004; Turner and Morrell, Reference Turner and Morrell2003).

Figure 4. Capital Asset Pricing Model.

Beta essentially reflects the variability in a firm's return on shares associated with general movements in the stock market and the economy. These systematic risks would include the consequences of interest rate changes, variations in the rate of inflation, political risk/legislative risk, and the general financial, economic or banking system. The intention is that hurdle rates correspond to the systematic risk of the type of activity under consideration, and would be the same as a firm's general hurdle rate as calculated by the CAPM, if the project is typical of the firm's activities. A project's cost of capital depends on the nature of the project, not the risk of the firm. The WACC is the correct discount rate for projects that have the same risk as the company's existing business. There are also firm-specific risks associated with the company's activities, for example Project Risk (uncertainty associated with a particular project) and Sector/Industry Risk (uncertainty associated with the performance of an industry sector.

Again, using the example of NATS, PwC had to carry out several calculation steps to estimate NATS' Beta – the CAA adopted a figure of 6·75%. NATS's WACC is low in comparison with firms in general, because ATC is a necessary function in an industry that forecasters believe will grow steadily over the long term, and NATS provides a monopoly en route ATC service in the UK.

How does the financial background to the air transport business sketched here affect Beta values, WACCs and NPVs for airline projects? The answer is that it has huge effects – and airlines seldom seem to follow standard assessment practices. Estimated airline Beta values do not conform to finance theory ideas that the cyclical airline industry is traditionally more risky than the market as a whole. It would be expected that β-values should be greater than 1·00, typically in the range 1·2–1·4, but the evidence, using a variety of estimation techniques, is for a value of 1 or less (Turner and Morrell, Reference Turner and Morrell2003). One explanation offered is that the shares of the airlines are not traded enough to respond sufficiently to changes in the market, and hence airline returns may not be as sensitive to shocks in the market. Another possibility is that the airline data was analysed when the market was more volatile, because of the increasing dominance of IT and telecommunications stocks.

If the ‘true’, i.e. in normal well-traded markets, Beta value for airlines were about 1·4, then the WACC would depend on a variety of factors, e.g. debt gearing. For a low gearing, the WACC might be up to 15%, depending on airline managers' views about the Equity Risk Premium. So do airlines use a WACC of 15% in NPV calculations? The answer is very probably no. If they use NPV, then they probably use a higher hurdle value – but they may well not use NPV. Empirical evidence for WACC ‘mark-ups’ on the Beta and debt calculated WACC is found in Meier and Tarhan (Reference Meier and Tarhan2007), which shows that the WACC used by financial managers exceeds the computed WACC by 5·3% to 7·5%, depending on the equity premium assumption. This would imply an airline WACC of about 20%. This premium is a way of adjusting decisions for project risks.

Gibson and Morrell (Reference Gibson and Morrell2005) examined the variations in airline financial assessment practices. Many airlines use Payback Period (PBK) rather than NPV. Payback period is the most intuitive of all measures, being the number of periods (months or years) required to recover the initial investment. PBK is a cash-based measure, but ignores the time value of money captured by the discount rate in NPV/IRR. Table 2 illustrates the nature of a typical flow of cash with an ATM investment, which generally have a lifetime of some decades. The firm invests a large sum of money and then a series of smaller cash flows come into the firm. The payback period is three years, the NPV for a 5-year time horizon is £19K, and the indefinite NPV over a large number of years (i.e. if the airline were to include a terminal value) is £272K. Marais and Weigel (Reference Marais and Weigel2006) note that commercial airline boards typically require a positive return on investment within eighteen months of investing – so in payback terms this would not be a worthwhile investment. On an NPV assessment, it would be a very good investment if the airline could be confident of long-term survival, but not exciting if the horizon was five years. But is this an appropriate judgement for this highly cyclical and cash-constrained industry? Airline finance directors are very important people, and one of the lessons they learn is to check all decisions that might lead to the company running out of cash. So terminal values are probably not that interesting to most airlines.

Financial decision-making textbooks and guidance material generally do not recommend adjusting the hurdle rate to take account of project risks. For example, IFAC (2007) recommends “calculating the probability-weighted expected value of cash flows of an investment. This is done by (a) developing several scenarios, and (b) assigning them probabilities of realization (including a probability of a project failure if applicable).” But the future is not wholly objectively assessable: forecasts and quantitative estimates rely on some things being unchanged and others changing in line with existing trends. Good projections of costs, the operating environment and air travel demand several years hence will include large elements of judgement and luck.

To illustrate the problems of the decision-maker, Table 3 shows the cash flow data in Table 2 again, but this time with a ‘correct’ hurdle rate of 15%. The base case cash flow shows a larger NPV for the five-year horizon, simply because the discount rate is lower. The decision-maker carries out two sensitivity analyses, one assuming that all the cash inflows slip by a year, and the second assuming that the cash flows reduce by 20%. These are both reasonable kinds of numbers: often projects take longer to achieve their benefits: often the benefits have been over-estimated, perhaps simply because of reduced market demand for airline flights. The consequence in both cases is that the NPV drops to zero or less: what does the ‘medium-term horizon’ decision-maker do? Note that the terminal value is very large for all the sensitivity cases – but can the airline be confident about existing beyond the planning horizon?

Table 3. Example of Airline Putative ‘Best Practice’ Cash Flows, NPV, Payback Period.

General definitions and caveats as in Table 2.

Assumed hurdle rate r – 15%, assumed long-term growth g of benefits – 5%.

An important aspect here is the kinds of technological changes envisaged in SESAR involving both ANSPs and airlines carrying out project developments. Airborne and ground-based investments are interdependent and need some degree of synchronization: so actually using the new kit on aircraft relies on successful ANSP hardware, software investments and appropriate controller training. The core problem is that such investments are essentially IT developments. IT projects tend to involve significant technical uncertainties (Tallon et al, Reference Tallon, Kauffman, Lucas, Whinston and Zhu2002). Large IT projects tend to take longer than estimated and cost much more. UK government guidance on ‘optimism bias’ in IT system development projects note overruns by 10% to 54% and overspends from 10%–200% (Mott MacDonald, Reference Mott2002).

A concern for airlines about spending money on aircraft fits is that some industry members may equip, but then the ANSP does not provide the requisite ground infrastructure or mandate for all users. Lester and Hansman (Reference Lester and Hansman2007) give some examples of USA problems:

• • Mode-S: the FAA tried to mandate Mode-S for all new transponders before building the Mode-S ground stations, but general aviation pressure forced the final rule mandating Mode-S and TCAS for aircraft with 10 or more seats.

• • Microwave Landing System (MLS): the airlines never had demonstrations of its advanced capabilities: they resisted equipping with airborne MLS receivers – and then GPS largely surpassed MLS technologically.

• • Controller Pilot Data Link Communications (CPDLC): American Airlines was an early adopter of the CPDLC technology, but – following 9/11 – the FAA did not continue with the deployment of the CPDLC ground infrastructure beyond its Miami Centre trial site, essentially making American Airline's investment worthless.

6. REAL OPTIONS

How could the concept of ‘Real Options’ help SESAR decision-making? Real options are a development of NPV techniques (e.g. Brealey et al, Reference Brealey, Myers and Allen2007). The aim is to fill the gap between finance and strategic thinking, so that management takes decisions to create highest value. Formally, a real option is the right but not the obligation to take an action in the future. There are at least five main types of real options described in the literature, some of great complexity: waiting-to-invest option, growth option, flexibility option, exit option and learning options. Real options take account of managerial decisions on investment with widespread uncertainty, where a degree of flexibility allows managers to make changes to the project. Simple NPV analysis ignores this project flexibility. A real option can capture a project's potential value by taking into account the value of being able to change it in response to new information, changing technical situations, market conditions, etc.

The simplest example is that a firm making a decision has the real option to decide to defer a project until better information becomes available about the likely cash flows. For example, if the best current estimate of a project's cash flows is £200K per year, then an NPV will be positive if the investment required is less than the discounted future flows. If this turns out to be a small number, then the firm may well not invest, given the kinds of discussion earlier here about sensitivity analyses. But suppose that in a year's time more information about markets and sales will be available, so that it will be clear if the annual cash flow is either £300K or £100K? After taking account of the different investment timing, it would be a good decision to opt to invest if the higher cash flow is available, and a – very – good decision not to take the investment option in the context of the lower figure.

Option valuation allows the flexibility of making decisions in the future that are contingent on the arrival of information. Information is seldom free – investing in an R&D project creates an option in investing in forthcoming development phases. Firms would have to take into account the likely quality of future information and the competitive environment (e.g. deferring decisions to invest could mean that rival firms which commit now would gain cash flows and dominant positions in the markets). Strategic issues that need examination include:

• • What are the project goals?

• • What are the uncertainties faced by management, and given these uncertainties, the most valuable options.

• • Can the costs of (additional) flexibility be justified by the added value when comparing the flexible alternative to the alternative without flexibility?

The literature on real options is huge. Dixit & Pindyck (Reference Dixit and Pindyck1994), and Copeland & Tufano (Reference Copeland and Tufano2004) are examples of business analysis approaches. Steffens & Douglas (Reference Steffens and Douglas2007) compares various NPV and real option methods, and has an extensive bibliography. ATM studies using real options are not yet common: MIT and the MITRE Corporation in the USA have sponsored some work, e.g. Steinbach & Giles (Reference Steinbach and Giles2005), and Rivey (Reference Rivey2007).

de Neufville et al (Reference de Neufville, Hodota, Sussman and Scholtes2008) discusses real options analysis for complex, large-scale, long-term infrastructure systems that are close to the SESAR issues examined here. The focus is the great uncertainties associated with large-scale systems, a root cause for unsatisfactory generation of value. de Neufville et al distinguishes between real options “on” projects, focused on accelerating or deferring projects (discussed above), and real options “in” engineering systems, focused on optimizing the technical design. The latter is about the designers taking special steps to provide flexibility in the system, recognising the additional costs involved in providing this flexibility. It takes decades to design and develop large technical systems, so there is the possibility of major changes in technology, economic situation, etc. Thus, the need is for concepts and procedures that enable decision-makers to anticipated possible uncertainties, and hence deal with them efficiently as they arise. This implies the need to develop the flexibility to react to events, to take advantage of new opportunities, and to exit from unproductive pathways. de Neufville et al therefore argues that designers of systems need the flexibility to alter development trajectories as needed, hence increasing expected value from investments. If developed wisely, flexible designs provide more value, because they systematically avoid bad outcomes and exploit opportunities.

7. ATM/SESAR AND PROJECT INVESTMENTS

The crucial financial logic for ATM and SESAR investment is:

• • Need to identify the optimum portfolio – projects sequence that makes financial sense to each of the stakeholders.

• • SESAR analyses of costs and benefits generally not estimates of the achievable project cash flows but illustrative calculations.

• • ANSPs' aim is a long-term cost-effective ATM system: must take into account NPV, and Terminal Value and very long-term Real Options.

• • But airline's aim is that the business makes cash (and does not go bust), so focus is on commercial NPV with a ~five-year horizon, plus a recognition of Real Options, e.g. reinvestment phasing.

The aim is to generate an ‘optimum necessary’ project portfolio, a sequence of projects that make financial sense to all stakeholders.

SESAR's central airline business question is to reconcile a long-term ATM vision with a financially-orientated airline stakeholder, which needs to make project investment decisions on a ‘pick-and-mix’ basis according to projects' cash paybacks. The nature of SESAR will be the result of the combination of a great number of individual project investment decisions to implement R&D products. Figure 5 shows the ATM and SESAR decision feedback loops. Airlines will want to have explicit choices and to ‘pick and mix’ projects whenever feasible if these choices improve projected cash flow benefits. They will want to optimise their project investment portfolio. For example, ANSPs might ask how to carry out SESAR projects better, but individual airlines will ask: “Is this project necessary? Does it need doing now?”

Figure 5. SESAR Strategy and Investment Feedback Loops.

It is not short-sighted of airlines to be careful in signing up to major project investments, given the track record of unstable cash flows. SESAR analyses of costs and benefits use the language of NPVs, but these are generally not estimates of the achieved cash flows from the implementation of specific projects but rather illustrative calculations based on hypothesized characteristics of possible future systems.

Figure 6 illustrates the project-focused financial decision-making for airlines and ANSPs. It shows a highly simplified four-stage process to get to the SESAR endpoint. Each stage, labelled 1 to 4, has several projects, labelled A to C. The labels in the project boxes indicate differing degrees of emphasis, thus the stage 1 boxes are in bold and the stage 4 boxes are in smaller font. This indicates the amount of financial certainty for the projects. The first stage projects are one for which there are actual contracts; the next stage is projects that are in financial budgets for the coming years; then in stage 3 the projects are specified in a plan but not firmly budgeted; and the final stage contains broadly-specified projects which complete the strategic intent. The different colours show where projects are ANSP spend (brown shading) and a combination of ANSP and airline spend (yellow shading). To get to the strategy-level and SESAR projects, ANSPs/airlines must implement/financially commit to the previous stages.

Figure 6. Schematic: Phased Investment Decisions.

The two shadings in the project boxes distinguish between projects carried out almost entirely by ANSPs, e.g. changes to ground systems, software, communications etc, and those that require both ANSPs and airlines to invest, e.g. both in ATC centres and through new kit on board aircraft. The important message from Figure 6 is that the Contracts and Budget level projects are the building blocks for SESAR, so the focus later here is on those kinds of pre-SESAR projects. Project linkages matter tremendously. If individual airlines do not decide to put their funding into a project then it and its potential successors disappear. To take a historical example, if airlines had refused to spend money on secondary radar surveillance transponders, none of today's radar and conflict alert systems would exist.

Synchronicity of a project's ground and airborne investment is nice if it happens, but the airlines will ask if it makes business sense. SESAR appears to assume a high degree of synchronicity (Marais & Weigel, Reference Marais and Weigel2006) for projects' ground and air implementation. But the financial decision-making will focus on NPVs/Terminal Value and Real Options. This will tend to mean that substantial airborne investments will take place later than ground investment spending, i.e. when there is assurance both that ground systems are operational and benefit generation is assured (unless there is external support from other sources, e.g. government). The assurance element of SESAR projects is vital. Airlines will need demonstrations that the project systems are technically feasible and that the estimated benefits are robust against a reasonable range of realistic ground implementation scenarios and business environments. Competent airline managers will want good information on the reliability and uncertainties of project benefit estimates.

8. SESAR AND CBA

As noted, almost all of the technologies encompassed in SESAR have long histories of successful research, often by programmes covering several countries. These technologies are generally compatible with existing ATM systems, thanks to good technical specifications many years ago (e.g. for Mode S in the case of ADS-B). Is there a major research problem? Are there important issues about which research should be developed into near-real systems and then implemented operationally, e.g. see Arthur D. Little Limited (2000)? The question examined here is how to determine optimal-chosen technical investment paths. ‘Optimal’ here mainly refers to spending money: aviation's decision makers are primarily concerned with ‘dollars and cents’ – corporate finance in more formal language.

Airlines will have to pay for most of SESAR, both through ATM user charges and airborne equipment costs. Their enthusiasm for strategic ATM system changes will be in jeopardy if these investments do not translate into a ‘good business deal’. Airlines do not want to commit to developing an ‘ultra-modern system’ per se, but rather to one that makes business-sensible investments in new technology, which is indispensable for achieving projected capacity requirements.

A notable ‘SESAR Deliverable’ was the ‘ATM Target Concept’ (SESAR Consortium, 2007). Its authors make it very clear that it is a vision not a plan, and certainly not a final blueprint of the future system. The Concept paper includes a financial section, which concludes with an outline cost benefit analysis (CBA). The Concept CBA is “‘what-if’ scenarios, with only trend and rough order of magnitude results”. Significant points are:

SDG (2005) and Brooker (Reference Brooker2008b) discuss benefits to society from SESAR, i.e. which could be seen as potentially justifying some degree of public funding. These assessments are complex, but the most important benefit contributions to society arise from passenger time-savings. These gains should generate worthwhile increases in Gross Domestic Product, so governments could be sympathetic to such investments.

In the present context, the most important features of the Concept CBA exercise are the discount rate used and time horizon for NPV calculations. The crucial document is SESAR Consortium (2008b):

Given the discussions in earlier sections here, this does not seem to be a judicious choice: rational business assessments by airlines generally use a markedly higher figure, 15% or higher. A further problem is the use of a very long time horizon for NPV assessment: the Start and Final Years adopted are 2007 and 2025 respectively, whereas airlines would generally consider a five-year time horizon to be a long one for cash flow examination. These two elements in the SESAR CBAs – low discount rate and very long time horizon – mean that the NPV calculations produce markedly higher positive figures than a typical airline calculation. Note the crucial distinction between the way a commercial airline views NPV assumptions with the way that commercial airlines, i.e. en masse – with ANSPs as their agent, might view financial decisions, particularly in the current non-regulated environment.

The most important aspect of the CBA for SESAR is that it is not a cost and benefit analysis. Normally, such an analysis would be analytic, in that the nature of the benefits that would flow from a specific investment would be set out and their financial implications estimated and summed. In contrast, the SESAR NPV estimates are synthetic, i.e. they are an artificial construct assuming future capabilities and performance from general considerations. SESAR Consortium (2008b) makes this clear:

9. PRE-SESAR PROJECTS

Figure 7 is a simplified version of a slide used in a very clear Eurocontrol presentation by Redeborn (Reference Redeborn2005). Before the ellipse indicating SESAR there are four examples of SESAR ‘precursor’ activities. Table 4 is a short technical description of what these abbreviations mean. DMEAN is essentially ‘best practice’ improvements, without significant changes to the existing concept of operation. LINK2000+ is a ‘critical step toward wide implementation of data-link technologies and applications’. CASCADE: ‘ADS-B is recognised as an essential element in SESAR’. FASTI ‘introduces improvements on controller tools, data interchange and integrity’.

Figure 7. ATM and SESAR Path.

Table 4. Pre-SESAR Projects.

These pre-SESAR projects deliver benefits in their own right and are strategic components of a SESAR concept for 2025 and beyond. Their value to airlines depends both on what they provide and as being components of the ATM system to which they lead. Table 4 also includes aFDPS: there are currently eleven European aFDPS initiatives covering 17 ANSPs, seen as critical to the ATC system and its links to advanced ATC functions (SESAR Consortium, 2008c). aFDPS is just one example of a project that modernises infrastructure to enable the implementation of longer-term SESAR concepts. There are other important projects in this category, for example System Wide Information Management (SWIM), which is essentially a net-centric system, providing the infrastructure and services to deliver network-enabled information access to a multitude of ATM system users. SWIM offers substantial system architecture benefits by reducing the number and types of interfaces and legacy sub-systems.

Given that SESAR is to be a ‘new ATM paradigm’, it is not obvious why DMEAN is ‘part of SESAR’, although it is discussed in SESAR documents as being in SESAR's first phase. Thus, in SESAR Consortium (2008d), the SESAR Implementation Package 1 has:

DMEAN includes elements such as capacity planning, improved Flexible Use of Airspace, airspace design improvements, data sharing, ASM/ATFCM (=air traffic flow improvements) processes, traffic forecast, operational air traffic harmonisation, airport integration into the network. The estimate is that DMEAN plus ATFCM produce an airspace capacity increase of 24%–32% (SESAR Consortium, 2008d – Table 2). Rather than being part of a new paradigm, these are essentially ‘bread and butter’ European tasks to increase the capacity of the current system through best practice. They are not significant changes to the existing concept of ATM operations. Many of them represent continuing work from earlier Eurocontrol programmes rather than new efforts. So, are these short-term improvements benefits from SESAR? – “Much of the benefit was achieved through management action using existing technology” (SESAR Consortium (2008c). This is not an academic point: the SESAR NPV calculations have to be made against a base case – ‘Business as Usual’ in SESAR jargon – so it is important to determine if that base case is the current operation or ‘normal progress’ from the past.

Appendix A examines the LINK2000+, CASCADE, FASTI and aFDPS cases. It makes it clear that there are varying amounts of CBA work done on these projects – some with good quality data and methodologies in authoritative national and international sources. But there has to be a warning here that reality can have markedly higher cost timescales than planned for (e.g. Mott MacDonald, Reference Mott2002).

Table 5 shows how the projects measure up against the financial criteria, based on these ‘official’ estimates. DMEAN has few strategic benefits because it is ‘best practice’. CASCADE is the most uncertain in term of estimated benefits for airspace benefits: is ADS-B a better investment than ‘multilateration’ for surveillance systems replacement? [Multilateration is locating an aircraft's position by computing the time difference signals arriving at multi-receivers – e.g. see Dow (Reference Dow2007). There do not appear to be quantified estimates of the merits of proposed later ADS-B stages at a European level (compare with USA, e.g. Scovel, Reference Scovel2007; Lester and Hansman Reference Lester and Hansman2007; Marais and Weigel, Reference Marais and Weigel2006). It is hard to find aFDPS benefit figures (SESAR Consortium, 2008c), as the aFDPS decision is strategic transformational IT. To simplify Table 5, the assumed Terminal Value of projects for airlines is taken as zero, i.e. their benefits are through a quick payback and/or the value of real options. Again, ANSPs are spending the ‘ATM system’ money that the airlines trust them to use wisely, while the airlines are buying new aircraft kit directly.

Table 5. Pre-SESAR Project Valuations.

Existing Eurocontrol/ANSP ‘best published’ estimates; en route airspace gains: ✓=estimated, ?=not known, ✗=unlikely. There can be arguments about valuations, eg see Appendix A.

Sources: DMEAN – SESAR Consortium (2008d), LINK 2000+ – Booker (Reference Booker2007), CASCADE (Dow, Reference Dow2007), FASTI – Brain (Reference Brain2008), aFDPS – SESAR Consortium (2008c)

In summary:

• • DMEAN, LINK 2000+, CASCADE, FASTI, aFDPS have long R&D histories.

• • Merits depend on different combinations of NPV, terminal value (dependent on growth), and real option value – data and calculations from authoritative national and international sources, but open to debate.

• • CASCADE needs more ‘hard CBA’ evidence: multilateration appears to be a better investment than surveillance systems replacement by ADS-B. What are the CBA merits of later ADS-B stages?

• • aFDPS investment decision is as the IT software platform, for implementing value-generating applications and reducing the costs of fragmentation.

• • Projects assessment has to be against good estimates of software development costs/timescales – in practice often much higher than anticipated in plans.

10. ECONOMIC CONTEXT FOR ATM/SESAR AND AIRPORTS

What is the economic context for ATM/SESAR? There are several facts and issues to consider, e.g.:

• • Airline profitability in recent decades subject to large – and increasing – cyclical oscillations.

• • Airline decision-making has to face the chronic oil price shock and the 2008 financial/economic crisis.

• • Airline costs to increase with the EU's Aviation ETS from 2012.

• • SESAR's goals now appropriate, given present and immediate future economic position and strategic views of the oil price?

• • Fuel economy issues and the 2008 financial crisis imply the likelihood of high growth scenario is markedly lower.

• • One effect of low growth rates is to reduce investment benefits cash flows – so need to look carefully at SESAR's flexibility, e.g. phasing of its project components.

On simple calculations, the European ETS costs are probably a smaller effect than the oil price – but costs that will still be unwelcome for airlines. A low growth scenario is more likely than before: such a scenario damages the gains shown in cost benefit analyses. So, how could SESAR adapt to this through project flexibility and phasing?

SESAR currently focuses on high growth scenarios and peak hour loading. A Eurocontrol ‘Low Growth scenario’ (Eurocontrol, 2004) – still 2·5% per annum – makes the airline/ANSP decision-makers look very carefully at project financial benefits and phasing. Under-investment is a bad thing over the long term, but financially-constrained airlines will not like actual over-investment.

Is the right goal a challenging ‘peak hour loading’ scenario, which might well have very low probability of occurrence? Financial decision-makers have to assess the range of strategies in terms of the costs involved for the stakeholders and the gains/losses they would get from over- or under-investment. A Low Growth scenario has major implications for the current implementation projects and SESAR R&D/project portfolios: it suggests deferring projects in the portfolio that deliver capacity above what is projected as needed or do not have big real options values; and it makes the case for matching spending to the available investment budget.

An underlying – and largely implicit – SESAR assumption is large growth in new airports/runways. It is necessary to explain some aspects of Peak Hour Airspace demand estimation. Figure 8 (left hand side) illustrates the summed diurnal demand from a congested – i.e. ‘full’ at peak hours – a group of nearby airports producing a typical diurnal demand pattern, with a flat top roughly the sum of the hourly runway capacities. Adding one or more new runways produces Figure 8's top right hand side, with a new airspace demand peak corresponding to the sum of the new set of runway capacities. If there are no new runways, then the result is the diagram a the bottom right, in which the extra traffic demand spreads across what were previously shoulder hours, so that the peak hour demand is unchanged. This is very simplified as, in reality, airport slots would get more valuable, and average aircraft sizes and load factors would tend to increase. But the point is that the need for extra peak hour capacity depends on new airports/runways coming into service.

Figure 8. Airport Diurnal demand patterns for congested group of airports.

Eurocontrol's most recent Long Term Traffic Forecast notes (Eurocontrol, 2006):

Note, with the benefit of hindsight, the phrase ‘slower growth in oil prices’. In 2004, Eurocontrol had produced its Challenges to Growth Report (Eurocontrol, 2004). This examined four scenarios for air transport growth; one of them, ‘D’, assumed high oil prices, although this did not quote the price of oil. It concluded:

The other scenarios had markedly higher growth rates.

A recent CAA paper (CAA, 2008) noted the importance of the oil price in air transport demand. From the paper, noting particularly the comment about the 1970s position:

CAA (2005) discusses the various elasticities of demand for air travel, which also includes a substantial bibliography on the topic. High oil prices also have a marked impact on a country's inflation and GDP growth, and hence in individuals' personal income. CAA (2005) suggests that demand for air travel is income elastic, i.e. income change causes a more than proportionate demand change for leisure air travel.

Eurocontrol (2004) estimated the extent of constrained demand:

Will all the potential new airports/runways materialize?

The Challenges to Growth report headlines “The need to plan for the most challenging scenario”. Is this the right goal? What are its merits if the evidence is that this challenging scenario has a low probability of occurrence? To reiterate, from a financial decision-making viewpoint, the need is for assessments of the range of strategies in terms of the costs to stakeholders and the benefits/disbenefits they would get from over- or under-investment.

11. SESAR PLANNING – LOW GROWTH SCENARIO

Thus, it is vital to examine the context for SESAR planning. Planning for airspace capacity growth has to take place against projections of the future aviation and general business environment. Figure 9 shows some linkages. The point about new airports/runways coming into service has already been made. A reducing oil price is not a good economic symptom if it simply reflects slow or negative GDP growth.

Figure 9. Selected Factors in Predicting Peak Hour Airspace Demand.

It is essential to work through the logical sequence in Figure 9 from an assumption of a Low Growth scenario to the impact on peak hour airspace demand. Without new airport capacity, the discussions on large increases in airspace capacity are academic, because demand will be constrained by airport capacity. To spell out the logic:

• • Peak demand is constrained by the airport capacities in use.

• • Therefore planning for extra airspace capacity requires good projections on the number of new runways.

• • Eurocontrol's 2004 Challenges to Growth study (Eurocontrol, 2004) attempted estimates of constrained demand if major new commercial airports were not developed – will they be?

• • Low GDP growth and long-term high oil prices generate a Low Growth scenario.

The simple conclusion is that setting the phasing of new SESAR related projects to deliver airspace capacity gains must take account of fluctuating demand, likely GDP growth, oil prices, and the realities of the phasing of new airport capacity.

What are the quantitative implications of a Low Growth scenario for SESAR requirements? The easiest way to see this is to present some crude sums about the airspace capacity gains from some of the pre-SESAR projects. The increases expected from three of the projects examined above are substantial, noting that these all derive from official State and Eurocontrol CBAs, not a reworking of the calculations:

• • DMEAN plus air traffic flow measures together produce an airspace capacity increase of 24%–32% (SESAR Consortium, 2008d).

• • CPDLC at 100% fit: estimated at 14% improvement (Booker, Reference Booker2007).

• • FASTI generates gains of up to 15% (Brain, Reference Brain2008).

• • In combination, the sector capacity gain would be between 63%–73%, i.e. about the amount required for the Low Growth scenario above.

Thus, taken as a whole, a rough figure across Europe, these pre-SESAR projects deliver the bulk of the capacity needed on a Eurocontrol Low Growth scenario to 2025.

A Low Growth scenario has major implications for the portfolio of SESAR projects. Any projects in the portfolio that deliver capacity above what is projected to be needed or which do not have sizable real options values would be deferred, i.e. moved up the financial priority stages in Figure 6. A Low Growth scenario would change the available investment budget over the coming years, putting constraints on the number of contracted and budgeted projects.

This second impact on the R&D/project portfolio needs some explanation. An analogy is the UK water industry, privatised for several years (Newbery, Reference Newbery2005). In the UK, privatised utilities have to be profitable and self-financing. They are subject to price regulation, which takes the form of ‘RPI-X’ regulation, where charges to customers cannot increase by more than the ‘rate of inflation’ (RPI) less an efficiency factor (X) in each year. But the water industry needed a huge capital investment programme to remedy past under-investment and to bring the quality of water/wastewater up to EU standards. Therefore, water prices to customers were increased annually by several per cent more than an RPI-X figure.

This is crucial: revenue must finance both operating expenditure and the capital investment programmes, and there are inherent limits on efficiency savings in operating expenditure to maintain the quality of service. Revenue also needs to be able to finance previous capital investment through the return the regulated company earns on its regulatory capital value. A continuing high level of capital investment therefore generally requires the amounts paid by customers to increase if the necessary finance is to be raised. The actual effects on prices to customers depend on the size of the asset base and the growth in demand. For the ANSP/'airlines agents' investments, the airlines have to scrutinise the portfolio of budgeted/planned projects. The regulator wants assurance that budgeted/planned projects would provide better capacity performance, or be a direct enabler of projects with a certainty of providing enhanced performance; and that the evolution of demand in terms of scale and pattern is robust; and that the technologies to be implemented are known and stable. User charges should increase from investment only if that extra capacity is added cost-effectively.

Thus, high NPV/real options projects, with the new lower growth traffic projections, would go ahead. The remainder would move up to the planning/strategy stages from contracted / budgeted, or be deleted from the present portfolio. Projects dependent on individual airline investment with increased payback periods would be at particular risk. The project portfolio would be the minimum capital investment necessary to deliver the forecast required capacity over the period. If, in future years, a high growth scenario were to develop, then deferred/deleted projects would be re-appraised.