Air departure tax in Scotland: an economic assessment

Appendix A – APEX Modelling

Introduction

The purpose of this Appendix is to set out the scope, methodology and headline results from additional route specific analysis, commissioned by Transport Scotland using RDC Aviation’s APEX ^[51] route economics evaluation software. This is designed to complement the broader market segment approach employed elsewhere in the report by considering at an individual route level, the potential impact of discounting ADT for airport pairs with an origin or destination at one of the main Scottish Airports (i.e. Edinburgh, Glasgow, Aberdeen, Prestwick, Dundee).

The work subjected selected routes to commercial sensitivity testing using the RDC APEX route evaluation software, at the two levels of discounting of ADT (50% and 100% of the current APD tariff) used elsewhere in the report, enabling comparison against a baseline where ADT continues to be collected in full once powers for collecting APD are devolved to the Scottish Government. The APEX software does not explicitly model APD but by varying airfares against constant variables such as aircraft seat capacity, load factors and frequency, the impact on profitability or passenger numbers of an ADT cut can be assessed. The methodology developed by Northpoint and the way it was deployed using APEX, was discussed with RDC Aviation; they also contributed to the analysis undertaken with the model and oversight of its outputs ^[52] .

Choice of Routes

The starting point for the APEX analysis, was to agree the routes to be examined based on a long list of options presented by the study team to Transport Scotland. Primary factors considered in the short-listing process that followed were the need to include:

• destinations from all six of the major mainland Scottish airports where enhanced connectivity is considered likely to be economically beneficial;
• a range of different types of service (i.e. long haul/short haul international and domestic, business focused/leisure orientated, and year round/seasonal);
• new links and existing services where frequency and/or competition is being added;
• airport pairs served by different types of carrier (i.e. network, regional, low cost, ultra low cost);
• routes that used to have flights but do not any longer, those that are currently marginal in terms of viability, and those that have not been tried before.

Routes from island and smaller Scottish mainland airports were discounted as they are for the most part internal to Scotland and therefore already have operating subsidy in the form of a Public Service Obligation ( PSO) or Aid of a Social Character that would exceed any benefits from changes to APD associated with the introduction of ADT in Scotland.

The table below shows the list of short-listed routes, those circled in red comprised four initial case studies undertaken to validate the methodology and outputs that were reported to Transport Scotland during one of the study progress meetings. Based on that initial presentation the analysis of the other 20 routes encompassed by the short-list were then progressed.

Table A.1: Routes used for APEX Analysis

Source: Consultants in conjunction with Transport Scotland.

Methodology

The route studies using APEX required the generation of route based demand forecasts for 2015, derived from the same 2013 CAA survey data used elsewhere in the report. This enabled an estimate to be derived of the total number of passengers, flying directly or indirectly (i.e. via an intermediate airport) to the chosen end-destinations, from relevant markets across Scotland (and the North East of England). In appropriate cases, a proportion of traffic currently using alternative airports capable of serving the end-destination market were assumed to switch in response to preferred airport pairs or increased frequency, was also included in the market assessment. So for example, the Aberdeen route to Doha is considered likely to attract a small proportion of Grampian or Highlands and Islands passengers who might otherwise have used existing services from Scottish Lowland airports to Dubai and Abu Dhabi. Data for a route to Heathrow (e.g. from Prestwick or Inverness), also takes into account traffic that may currently use Gatwick.

The next step was to grow the ‘estimated catchment market’ for each route by an assumed background CAGR ^[53] of 2% to derive a forecast baseline 2020 figure which was then factored to generate a more realistic ‘potential market size’ using a series of filters covering:

• Potential to attract onward connecting traffic beyond the end destination (i.e. interlining passengers)
• Scope to stimulate the existing market size by increasing frequency or adding a new route
• The extent of likely market penetration.

Each of these factors is assumption based, but the complicated commercial considerations that have been used to determine the appropriate scale at which each is applied to the routes under consideration are outlined in the footnotes to the table. The output from this process is a figure for estimated ‘Total Potential Market’ by route and this is a key data input to the later APEX modelling process.

The second block of analysis to inform inputs to the APEX modelling then identifies the most likely form of carrier to serve the type and size of route in question. A notional operator was then also chosen so that an aircraft type could be assigned and seat capacity potentially available therefore better understood. Frequency was then adjusted to generate a proportionate total annual seat capacity ^[54] and ultimately a projected load factor.

Finally, an initial indicative assessment of the likely level of impact that reducing ADT will have on individual route pairs was made at this stage prior to the APEX analysis

APEX Modelling

APEX primary value is as a route economics evaluation tool, which allows the potential commercial viability of new routes (or enhanced frequencies) on existing routes to be examined in the context of a wide range of input variables including:

• End destination;
• Airline;
• Aircraft type;
• Frequency;
• Fares; and
• Load factors.

For the purposes of this analysis, we used APEX to model the effect of changes to the level of ADT on a baseline profit/loss breakeven position on each of our case study routes. By running a series of sensitivity tests, the functionality of the model allowed us to examine the impact of changes to implied fare levels by subtracting a ‘blended rate’ for ADT (i.e. a proportionate mix of standard and business tariffs, based on the existing business/leisure split unique to each route) from the breakeven fare and then comparing outputs under each of the main discounting scenarios (i.e. 100% and 50%). The key output variables in the sensitivity tests are the change in the projected profit/loss position or the uplift in passenger numbers (and hence average load factors on the flights).

Results

The results from the Apex modelling were then compared with those from earlier market based route evaluations – essentially our hypothesis as to the likely impacts of changes to the level of ADT - to highlight where the different methodologies produce consistent assessments and those where the outcomes are divergent and why. These conclusions are summarised at an individual route level in the table below.

Table A.2: Summary of APEX Conclusions

A guide to how the evaluation descriptors used in that table have been derived is shown in the box below.

Negligible
Market Assessment: Very unlikely to encourage new route formation or added frequency.
Apex Analysis: No or very small changes to potential airline profitability or passenger volumes – unlikely to significantly affect airline decisions on potential new routes or encourage enhanced frequency on established ones.

Minor
Market Assessment: Underlying market/route conditions are such that ADT unlikely to have large enough impact to encourage new route formation/increased frequency; may help extant marginal routes survive or new routes already considered to have potential to come forward slightly earlier.
Apex Analysis: Changes to profitability and passenger volumes are typically less than 10% under the 50% ADT discount and 10-19% if it is removed altogether – a route or increased frequency would need to be on the edge of viability for this to level of impact to generate positive outcomes and this is unlikely in most cases.

Modest
Market Assessment: moderate changes to prices are considered likely to have a moderate positive impact on passenger demand, or help to change airline risk perceptions about the viability of ‘next in line’ routes and help secure earlier and possibly more substantive commitments to these.
Apex Analysis: In this category changes to profitability and/or passenger volumes are typically 15-20% under a 50% discount and 20-30% with a 100% discount, the affect being substantive enough on the outbound leg to make up for revenue losses associated with currency changes whilst still allowing some improvement to commercial prospects for new routes or enhanced frequencies.

Material
Market Assessment: Discounts offered by ADT are considered likely to have a noticeable positive impact on new route formation and encourage additional frequency
Apex Analysis: In this category changes to profitability and/or passenger volumes are typically 30-45% under a 100% discount and over 20% with a 50% discount, the affect being substantial enough on the outbound leg to exceed revenue losses associated with currency changes, whilst still allowing notable improvement to commercial prospects for new routes or enhanced frequencies.

Significant
Market Assessment: ADT is likely to transform the prospects for new routes or enhanced frequency of operations, encouraging substantial new levels of connectivity and direct frequency and opening up the prospects of links to wholly new market areas.
Apex Analysis: In this category changes to profitability and/or passenger volumes are typically over 45% under a 100% discount and plus 30% with a 50% discount, the affect being large enough on the outbound leg to materially exceed revenue losses associated with currency changes, whilst still allowing substantial improvement to commercial prospects for new routes or enhanced frequencies.

The more generic observations that we have drawn from the collective results of the individual route appraisals are then set out in our conclusions below.

Generic Conclusions

Taken together, the APEX modelling produced a third of outcomes that confirmed the initial market-based hypothesis, with an equal number one category higher and a category lower. The evaluations departed markedly in only two out of 24 cases, and in both cases for good reasons set out in the table below. These results provide some confidence that the combination of the two approaches, and especially the use of APEX, offer a useful insight into how changes to ADT may impact upon different types of route.

Table A.3: Summary of Route Evaluation Results

This suggests that both methodologies anticipate reductions in ADT having a noticeably positive impact on between 60-65% of the short-listed routes considered and little or no impact on the remainder.

Although the APEX modelling tends toward results that are more polarised than the more subjective market-based assessment, this can largely be explained by the fact the modelling looks at the effects on the leg affected by changes to ADT only, rather than on return trips. The latter would require input assumptions about approaches to air passenger taxation in different destination countries; since they already vary markedly and are changing constantly, it would be difficult to second guess appropriate levels three years ahead let alone five or more. Of more significance, is the relative high level consistency of expected outcome over the full suite of potential new routes (or routes where increased frequency is desirable) that we examined. This again offers some comfort that a policy of intervention based on reductions in ADT would generate materially beneficial outcomes overall, although at a route by route level there would be inconsistencies.

If we dig a little deeper into the results reported above, a number of other generic patterns can be highlighted:

a). Types of Carrier: Routes suitable for low cost, or ultra low cost, carriers are likely to see positive responses to changes in ADT. This is because their relatively lower fares and the price sensitivity of their majority leisure customer base will produce a much stronger price elasticity response compared with more expensive business orientated regional services using smaller aircraft with less beneficial operating economics occurring on thin routes. The latter kind of routes, exemplified by services from smaller airports like Inverness and Dundee, where the route economics are testing, may be better supported by other means (e.g. PSOs), which under current Treasury rules are already exempt from APD. In the case of network carriers the response will depend heavily on the extent of the proportion of business passengers and whether the service is in Band A (smaller effect) or Band B (larger impact). For example, a route from Edinburgh to the Far East (Band B), appears to benefit substantially from ADT reductions, whereas one from Glasgow to Istanbul (in Band B) does not, due to a higher proportion of business passengers in the former.

b). Long haul routes with above 20% of total passengers flying on business or operated by a long haul low cost carrier (e.g. Norwegian) appear to perform well). The very strong projected response of the Aberdeen to Doha route is because the average fares are relatively low,) as it is assumed an A321 will be used and this makes aircraft operating costs proportionately much lower than for the larger wide-bodied aircraft used on the EDI- HKG route. At the same time, the business element of the traffic forecast to use the route is anticipated to be high, and hence the relative quantum of benefit from discounting non-standard ADT tariffs, is likely to be substantial. The APEX analysis suggests that if the whole of any change to ADT were to be passed on, then there appears scope to increase frequency over the baseline assumptions, from 5 to 6 flights a week for EDI- HKG and 4 to 5 flights a week between Aberdeen and Doha.

c). Short haul services to international destinations typically appear to be less responsive to changes to ADT than long haul or domestic operations (see below); this reflects the lower rates of APD that apply. At a 100% level of discount the change to ADT may still be significant enough to make some routes more attractive, but most probably those where average fares are low, competition is slight and the carrier is a low cost operator. Thin routes, using regional jets subsidiaries of network carriers appear likely to respond very little to even the highest level of ADT adjustment.

d). Domestic services may be more sensitive where ADT would make up a bigger proportion of the average fare than on longer short haul international routes – in addition to which there is also the ‘double whammy affect’ of having ADT/ APD on both legs of the journey if it were also to be removed in say Wales and Northern Ireland (see the Cardiff service that was examined) or Heathrow with all its connecting passengers as an end destination.

In the case of Prestwick to Heathrow, the breakeven fare projected by APEX is quite modest, so we suspect that much of the reduction in ADT would be absorbed by a network carrier; similarly a low cost carrier, although perhaps to a slightly lesser extent. If all of the decrease were passed on to passengers then it would be possible to imagine an additional daily frequency being justified, but that might be likely to run into slot availability problems and hence the introduction of a larger aircraft like an A319/A320 could be a more realistic response.

Finally, of the airports we considered, Edinburgh, Prestwick and Dundee appear likely to derive the greatest benefit from the ADT scenarios examined, with Glasgow, Aberdeen and Inverness doing less well. Although this is primarily a function of the route short-listed, what we can say is that where markets are relatively thin and short haul and hence greater reliance is placed on regional operators, the response may be less than airports where markets are larger and yields for low cost and long haul carriers higher. The scale of and form of intervention may need to be adjusted accordingly.