Air departure tax in Scotland: an economic assessment

3. Logic Modelling Framework & Method

In detailing the potential outcomes and impacts of the introduction of ADT and with it potential tax reductions, a logic modelling approach was developed. A logic model is simply a mechanism by which the sequence of events between cause and effect is set out in logical and sequential steps. The set of intervention logic models are intended to identify and demonstrate how the reduction in APD would feed through to wider outcomes and impacts.

The development of the logic models was informed by a detailed literature review, which considered the approaches to modelling the outcomes and impacts of changes to APD found in other studies. The main finding of this review was that the bulk of previous research studies have assumed that any reduction in APD would be passed straight through to the end-user (i.e. the passenger). This, in our view, is unrealistic given the complexities of aviation economics and the number of potentially affected parties, and the approach adopted in the logic mapping below therefore considers potential variations in pass-through.

The structure recently put forward by the Tavistock Institute for the DfT ^[25] with respect to transport evaluations has been adopted here and this suggests the following logic chain as the basis for considering the impacts of a transport intervention: Context – Input – Output – Outcomes – Impact. This structure is typical of logic modelling approaches which are widely used across disciplines, and this is expanded on below:

• Context – issue addressed and the context in which it is taking place: the issues to be addressed are (i) the concerns surrounding the current APD tax regime and its impact on airlines, air travel, connectivity & the economy, and (ii) the recent devolution of power enabling Scottish Government to set the parameters of this tax.
• Input – What is invested, e.g. money, skills, people, activities: essentially an acceptance of reduced tax revenues from this source and the specification of the tax rate and bands / thresholds etc.
• Output – What has been produced? The primary output is reduced costs to the airlines as their tax liability per passenger will fall.
• Outcomes – Short and medium term results: this reduction in costs to the airlines may: (i) generate reductions in fares (depending on the degree to which the tax saving is passed on); (ii) lead to new or improved connections; and / or (iii) lead to changes in organisational structure within the airlines. Therefore, there are essentially two outcomes:
  • Changes in the organisation of airlines / airports in response to changes in costs, passenger numbers and new opportunities created [supply side response].
  • Consumer behaviour and hence spending patterns change in response to lower fares and new / improved connections [demand side response].
  • The supply side response of different airlines will vary both in terms of the degree of pass-through and the time horizon over which changes occur.
  • Impacts – Long term outcomes: these should reflect the long-term objectives of the policy, as reflected in the Bill documentation e.g.:
  • ‘ boost Scotland‘s air connectivity and economic competitiveness;
  • encouraging the establishment of new routes which will enhance business connectivity and tourism;
  • secure new routes and capacity;
  • encourages airlines to base more aircraft in Scotland’; and
  • also, contribute to SG overarching Purpose and other policy objectives (i.e. creating & safeguarding employment and generating net additional GVA).

It is notable that the policy objectives are formulated around greater connectivity for Scotland, rather than cheaper air travel. This underlines the importance of understanding and accounting for the potential range of supply side responses, rather than focusing solely on the possibility of cheaper fares.

The key issue for this study has been to work through this logic, considering in particular a range of possible outcomes and impacts via a framework of demand, supply and economic impact logic models. The process adopted is outlined in the figure below and expanded on in the sections which follow.

Figure 3.1: Overview of Logic Model Approach

The Inputs and Outputs in this case are essentially the resources required (i.e. the tax take forgone, or Inputs) to enable the level of tax to be reduced, i.e. the Output. The following sections consider the treatment of Outcomes and Impacts in more detail, together with the underlying transmission mechanisms through which the change in tax becomes an impact ‘on the ground’.

Supply Side Outcomes

The first substantive step in the developing the logic modelling approach overall is to consider the potential impact of the introduction of ADT and the expected reductions in taxation on the supply side (i.e. the aviation industry) as it is the first link in the chain that ultimately leads to the passenger.

The outcomes associated with the reduction in APD will be determined by the rates and bandings applied, and the extent to which the tax reduction is passed on to the end user in the form of reduced fares. This will determine the balance of the outcomes and impacts in terms of the supply side (i.e. new routes, higher frequencies etc.) and the demand side (i.e. increased passengers through lower prices / new routes).

Rates & Bandings

For the purposes of this research, and for modelling purposes, the scenarios introduced later in this report assume that the current APD tax bandings structure remains intact, with changes in the rates within the bands only. This is because modelling a new bandings regime would introduce a further range of scenarios which would detract from the analysis of a tax rate cut which is the central objective/scenario running through the analysis. This analysis does not preclude a change in bandings either from April 2018 or in the future.

Extent of Pass-Through & Impact on Services

The extent to which the APD-related reduction in fares is passed through to the end customer is determined by a wide range of factors, including but not limited to:

• The way in which the airlines currently impose this tax (and the transparency with which they do so) – some airlines, such as British Airways, clearly itemise the tax on bookings / invoices, whilst other airlines (particularly low cost carriers) do not explicitly recognise the APD burden within the fare.
• Whether the airline is expanding its network and seeking additional utilisation of its assets or contracting its network through reducing aircraft / routes.
• Airline cost structure and how they account for APD currently (i.e. as a central overhead or on a per passenger by route basis).
• The level of competition – this will be particularly pertinent for airports served by multiple airlines. It will be less significant where one airline tends to be dominant at a particular airport or cluster of airports.
• The extent of interlining – for several long-haul carriers, the objective of their short-haul connections is to feed passengers into their more lucrative long-haul route network.

The type of supply side response is likely to vary by airline type. Based on consultations with industry, we have set out below how we think the respective sector will respond:

• Domestic Scotland (e.g. Loganair): Pass-through where appropriate, but potentially also using APD reduction to support the viability of ‘thin’ routes (either altogether or in terms frequency).
• Network (e.g. British Airways, United Airlines, Emirates): More likely to pass on reduction to customer as 1) less likely to commence any new routes from Scotland (unless low cost); and 2) can potentially support primary objective of getting passengers onto their long-haul network / interlining.
• Regional (e.g. FlyBe / Eastern Airways): Likely to use APD reduction to either support ‘thin’ routes or support new route development, depending on the strategic objectives of the airline. The chosen strategy will influence the degree of pass-through.
• Low Cost (e.g. easyJet and Ryanair): With large interchangeable fleets and airport networks, the APD reduction is likely to be at least partially retained and used as the basis of further route development / asset utilisation.
• Chartered (e.g. Thomas Cook and Thomson): Given the highly competitive nature of the package holiday market, it is likely that the cost reduction will be passed on (either directly or indirectly) to the end customer. However, APD is likely to be a very small element of the total cost of package (and especially all-inclusive) holidays.

It is important to note that the timeline over which the supply side response will be manifested will vary. A number of low cost airlines, including Ryanair and easyJet, have already committed in-principle to new routes at the point at which ADT changes are introduced. With other market segments, particularly long-haul, the market response may be in the medium-term (e.g. 2-5 years), although some long-haul routes may come forward more quickly than they otherwise would have.

It is worth noting that the airports may also ‘skim’ a proportion of the APD reduction through increasing landing charges levied on airlines, so even where an airline offers full pass-through, this is potentially less than the actual amount of the tax reduction. This in itself is perhaps less of an issue if the airport uses the additional funding to support further expansion of its operation / route network. It only becomes an issue if the airport is engaging in a form of ‘ rent extraction’ ^[26] . In any case, it is not possible to systematically lay out this impact, as details of the contracts that airlines have with airports are commercially confidential, but it is an issue and risk worth highlighting as a further complication in the analysis.

In terms of estimating the supply side impacts and outcomes of the APD reduction, different scenarios of pass-through have been modelled (i.e. full, partial and zero pass-through). The analysis also picks up on the likely differential impact on short-haul and long-haul connections.

For each airline type under each scenario, the subsequent analysis has made an assumption on a supply side response, drawing on a combination of the industry consultation and desk-based review considering the following potential outcomes which are likely to emerge:

• Larger aircraft on existing routes - assuming aircraft can be accommodated within airport parameters (e.g. stand capacity, runway length & capacity etc.), larger aircraft may be deployed to meet an increase in demand.
  • This outcome is most likely to emerge where there is full or near to full pass-through of the APD reduction to the end user, generating additional demand from lower fares.
  • This is likely to be the least prominent outcome as a number of airlines, particularly the low cost carriers, operate a single aircraft type. This outcome may nonetheless occur with airlines with a larger or more diverse fleet.
• Higher frequencies on existing routes
  • This outcome is again most likely to emerge where there is full or near to full pass-through of the APD reduction to the end user, generating additional demand from lower fares.
  • Without an airline procuring additional aircraft or using available spare capacity, there will be an opportunity cost associated with increasing the frequency on any given route (due to aircraft being redeployed from elsewhere). This impact is therefore only likely to materialise where the yield from running additional services from Scottish airports exceeds the next best alternative.
  • This extent to which this outcome will materialise in the short-term is limited. However, a number of airlines, including easyJet and Ryanair, have substantial orders in place for new aircraft and thus enhanced frequencies on existing routes could be a more prominent outcome in the medium-term. This could also feed into the longer-term aspiration to have more aircraft based in Scotland.
• New direct routes - the development of new routes is the most likely supply side outcome associated with the reduction in APD. There are two derivatives of this outcome:
  • Where the APD reduction is largely passed through to the passenger, this will generate additional demand but would be largely cost neutral from the perspective of the airline. The demand (and resultant revenue) which would be generated would need to be capable of covering the cost of operating the route.
  • Where the APD reduction is largely retained by the airline, this may reduce the break-even point for any given route, thus making it more viable.

In terms of the above outcomes, the establishment of new routes is, from a Scottish Government policy perspective, likely to be one of the most desired outcomes emerging from the APD reduction.

Demand Side Outcomes

There are two potential demand responses which may emerge (neither of which are mutually exclusive):

• Additional demand generated as a result of lower prices on existing routes; and / or
• New demand generated as a result of new routes / connections / higher frequency etc.

This section considers how the demand side outcomes are represented within the overall logic framework.

As discussed in Section 2, a representation of the present day Scottish air market by market segment was developed, and an estimate of the value of APD raised, by segment, class and distance band was established. Our analysis suggests circa £251m of APD was collected on flights from Scotland in 2016.

Similarly, estimates of future Do Nothing demand (i.e. continuation of existing APD levels) were developed via a number of ‘Reference Cases’. These reference cases reflect a variety of potential baseline growth scenarios and represent the counterfactual, against which ADT intervention scenarios have been tested.

Estimating the ‘Do Something’ (tax reductions from APD levels) Demand

There are two elements to estimating future demand as a result of the introduction of ADT at a lower rate than the current APD: 1) demand generated as a result of reduced fares (i.e. demand side change); and 2) demand associated with new direct connections to / from Scotland (supply side change).

Reduced Fares & Elasticities

The introduction of APD could result in a reduction in air fares and the size of the saving will dictate the scale of the market response. As such, in order to understand how much of a saving will be made under future ADT scenarios, it is also important to incorporate average 2016 air fare data within the baseline. Fare data from the CAA Passenger Survey was used in the first instance to calculate average fares for each passenger category; however, only circa 30% of CAA Passenger survey responses included this information and so there remained a number of categories for which passenger data were missing. To address these gaps we:

• Generated factors based on the whole CAA Passenger Survey 2013 dataset (circa 55,000 responses), which illustrated the relationship between average single and return fares, and average Band 1 (Reduced Rate) and Band 2 (Standard Rate) fares. These factors were applied where average fare data was available for Band 1 but not Band 2 or vice versa, and similarly with single / return fares.
• Where gaps still remained, we identified the most common country destination of passengers falling within that passenger category and then the most commonly used airport within that country.
• In the first instance we searched RDC Aviation fare data for air fare estimates between the relevant airports and destinations, and then supplemented this data with that from an online fares search (Skyscanner), assuming travel in October 2017.

Fares from the 2013 CAA Passenger Survey were inflated to 2016 levels through application of the GDP Deflator series from WebTAG, the UK Government’s transport appraisal guidance. Fares obtained online were assumed to be representative of 2016 levels. Having established a robust fare for each market segment, the estimation of demand as a result of reduced fares is fairly straightforward, applying the relevant price elasticities of demand to the different market segments.

Academic and Government studies of price elasticity of demand within the air travel market have generated a wide variety of potential elasticities. Our core analysis has utilised elasticities from the DfT’s UK Aviation Forecasts 2013, and sensitivity testing has been undertaken applying two sets of alternative elasticities, from widely cited papers applied as a sensitivity as follows:

• InterVISTAS Consulting Inc. (2007). Estimating Air Travel Demand Elasticities, Final Report. International Air Transport Association ( IATA);
• Gillen, D., Morrison, W., and Stewart, C. (2003). Air Travel Demand Elasticities: Concepts, Issues and Measurement. Department of Finance, Government of Canada.

New Connections

As noted previously, some airlines are likely to treat a reduction in air passenger taxation as a reduction in their cost base and therefore:

• fly new routes which are currently not viable;
• reinstate routes which have been withdrawn; and / or
• convert seasonal routes to year-round routes.

The APEX ^[27] modelling tool has been used to test a sample of new routes across the different airports and market segments. These case studies have been used as context to develop a series of supply side response factors (by airline type and short haul / long haul) and, in turn, an estimate of passenger numbers which will emerge from this.

A supply side reaction is unlikely to be immediate, and consultations with airlines suggests that the lag time will vary by market segment. Our modelling has assumed that any supply side reaction would be staggered over three years, beginning in 2019. It is noted though that this will vary by airline and some may make an early symbolic gesture in recognition of the new policy.

Uplift in Passenger Numbers

The combined uplift in passenger numbers from (i) reduced fares and (ii) new connections is compared against the various Reference Case / Do Nothing forecasts and fed into the economic impact model.

Determining Economic Impacts

An economic impact model was developed, fed from the disaggregated forecasts of passenger numbers by market segment to derive the economic impacts of changes to ADT for each year between 2018 and 2027. These impacts are expressed in the form of jobs and GVA at the Scotland level.

There are four main elements to this model:

• direct operational impacts ((i) airport related, and also (ii) the basing of additional aircraft in Scotland);
  • additional airport and airline staff required to handle additional passenger numbers;
  • additional aircraft based in Scotland, with Scotland-based crews (also implies more flights);
  • increased density of aviation operation, creating the potential for spin-off employment in terms of e.g. maintenance.
• wider economic impacts (associated with (iii) tourism, and (iv) improvements in business performance / productivity etc.).

The objectives for the introduction of ADT set by the Scottish Government explicitly recognise the potential of any tax reduction to increase tourism in Scotland. The logic applied is that lower prices will stimulate visitor demand on existing routes and create the demand for new routes.

Similarly, there is a recognised relationship between business productivity and business air travel, as a proxy for increased business connectivity. This is used to develop productivity impacts of increased business travel brought about by ADT cuts and is discussed further in Chapter 6.

The approach taken to the estimation of each of these impacts is discussed in more detail in Chapter 6.