Estimate of The Impact on Emissions of a Reduction in Air Passenger Duty in Scotland
Estimate of the Impact on Emissions of a Reduction in Air Passenger Duty in Scotland
1. Air Passenger Duty (APD) was introduced on UK flights in 1994 in a simple two-band structure. It was only in 2007 though that APD rates began to increase significantly with a four band APD structure introduced in 2009 in an attempt to take better account of the impact different flight lengths had on emissions. These bands were based on the distance between London and its equivalent capital in the destination country. Further APD rate increases were introduced in 2010, 2012, 2013 and 2014. A full breakdown of the history and rates of APD charged in the UK is shown in Annex A.
2. The Scottish Government made a commitment in Parliament to publish an estimate of the impact that a reduction in APD would have on emissions. This note fulfils that commitment by reporting the likely impact from a 50% reduction in APD alongside the methodology used to generate the estimate.
Data requirements, data availability and assumptions
3. Transport Scotland does not possess models capable of producing an estimate of the impact on emissions from altering APD. This calculation is built instead on published price elasticities and data available from a range of sources. These data include APD bands and rates, demand price elasticities, ticket prices by destinations, passenger numbers by broad destination and aviation emission estimates. Unfortunately these data are not all framed in terms of current APD bands so require some manipulation.
4. The realities of the aviation market also create a number of complications, particularly in relation to data on ticket types and prices. Air travel is not a single homogenous commodity but instead a series of interconnected overlapping markets and prices with different ticket classes, APD rates and travel purposes. Having multiple markets and products makes interpreting the available data more complex as well as introducing greater uncertainty into the analysis and results. This issue is discussed in more detail below.
Air Passenger Duty
5. Charged on most UK flights from 1994, since 2001 there has been a differentiation by price in respect of the class of ticket. If the entire flight is a single class, usually economy, then all tickets are charged APD at the reduced rate. Any tickets purchased for a higher cabin class - business or first are charged at an increased rate. A further higher APD class was introduced from 2013. The vast majority of tickets sold in the UK (90% or more) are charged at the reduced rate of APD.
Aviation price elasticities
6. Research suggests that the most inelastic, i.e. least price sensitive, trips are long haul business journeys followed by longer haul leisure trips. Short haul leisure travel - both domestic and international - tends to be the most elastic. Focusing in on the UK aviation market, DfT publishes estimates of price elasticity of demand for flights from the UK. These elasticities vary by travel purpose (broken down into business and leisure) and by destination (separated into short haul or long haul or level of economic development). The most recent DfT estimates were published as part of UK Aviation Forecasts in January 2013.
7. Table 1 below sets out the elasticity estimates used in this analysis. These elasticity values cannot be directly read from the 2013 DfT paper (the paper does not break down elasticities by APD band) so an explanation of how they were created from the DfT work is shown in Annex B.
Source: own estimates from DfT analysis
8. No regional breakdown of the UK aviation market is provided within the DfT analysis. The similarities between Scottish and UK travellers and the numbers of airlines that fly within and to UK destinations might strongly suggest however that the Scottish price elasticities of demand for the different journey types will be similar to those for the UK as a whole. Ticket prices to international destinations are also likely to be very similar from airports within the UK, with the possible exception of the more southerly European destinations where flying from Scotland takes considerably longer relative to the time taken from airports in southern England.
Ticket classification and travel purpose
9. The definition of 'business' differs between ticket category and trip purpose. A 'business' airline ticket is used to distinguish the on-board service rather than purpose of trip. Thus leisure travellers can and do fly on a 'business class' ticket. Equally, people travelling for business often qualify for the 'reduced' APD rate because there is only one class ticket on the flight (which attracts the reduced rate of APD) or their company chooses not to pay for a business or first class airline ticket. For this assessment the distinction made was by ticket class rather than the purpose of trip. This ticket class difference can be identified in the APD data published by HMRC.
UK passenger numbers and destinations
10. The Civil Aviation Authority (CAA) count of passengers arriving or departing the UK by airport and destination allows for a passenger breakdown for Scotland as well as for the UK as a whole, by APD Band. Table 2 sets out this disaggregated data.
|Band A - domestic (UK)||16.9||52.0|
|Band A - international||61.5||43.7|
|Total passengers (millions)||228||23.3|
Source: CAA and own calculations
11. The most obvious difference between the UK and Scottish data sets is the proportion of domestic flights under each heading. For the UK as a whole they account for one sixth of trips whereas in Scotland domestic trips account for over half of all trips. Otherwise, Band C and D accounts for 6.8% of all UK flights but only 0.2% from Scotland, all of which are in Band C.
'Typical' airline ticket prices, currently and in 2009
12. Due to the highly developed, competitive and atomised nature of the UK air market (multiple class of seats and prices, large number of airlines, airports and destinations along with seasonal variations in ticket prices) there is no readily available public information source providing 'average' current airline ticket prices by destination, or any equivalent price in previous years.
13. The ticket price estimates quoted in this analysis are therefore internal estimates based on popular destinations from the UK and Scotland using passenger destinations and numbers as a guide to this choice. Prices were estimated using an available price comparison site that recorded average monthly prices to the destination in question. Each monthly estimate provided was a point price figure and the yearly average provided is likely a straight un-weighted mean. This simple approach will hide the variance in ticket prices between airlines, destinations and time of year (modern airline ticket pricing systems allow prices to vary by ticket within a single flight giving potentially millions of data points for each route). Clearly using the available data without access to the primary information creates uncertainty in the analysis, but there are no readily available alternatives.
14. From these current price estimates an equivalent earlier estimate can be produced by deflating the current ticket prices using the UK sub category CPI index for air passenger transport. This is a single index so again it is necessary to apply a simplifying assumption that all ticket prices for all types of tickets and destinations have changed in line with this CPI estimate. According to the CPI series the cost of air travel has risen by just over 31% since 2009. The price estimates for both 2013 and 2009 are set out in the table below. Annex C contains a link to the websites used along with the sources for other key data.
|Band A domestic (London- Scotland)||£120||£91|
|Band A international (Malaga)||£150||£118|
|Band B (Newark, New York)||£400||£305|
|Band C (Mexico)||£600||£458|
|Band D (Australia)||£800||£611|
Source: own estimates based on CAA destination data, CPI and available price information
15. Each of these data sources has been developed independently and for different purposes. As such there is not a common factor linking them all together to enable direct comparisons. However APD financial returns to HMRC show the overwhelming majority of all UK seats (90% plus) are charged the reduced rate of APD and surveys from the CAA show that, particularly for airports outside the London hubs the vast majority of trips are for leisure purposes - either holidays or visiting friends and family. The skewed nature of this distribution (and hence ticket prices) enables the main focus of the impact analysis to fall on one cabin class of prices and one purpose for travel.
16. Aviation emissions are published as part of the UK emissions inventory. They are published in aggregate for the UK, and separately for Scotland. Aviation emissions can be disaggregated between domestic aviation (take off/landing, cruise and military) and international aviation (single category). This is a sufficient level of disaggregation for this analysis. Table 4 sets out the latest UK and Scottish aviation emissions information having stripped out military aviation emissions.
|Domestic Cruise||1.171||1.118||1.077||0.08 tCO2e|
|Domestic Take-off and Landing||0.455||0.434||0.407|
|Domestic Cruise||0.329||0.325||0.319||0.01 tCO2e|
|Domestic Take-off and Landing||0.106||0.103||0.101|
Source NAEI and own calculations
17. Splitting passenger numbers between domestic and international flights can then generate an emissions per passenger figure for domestic and international flights and this estimate is shown in the final column of Table 4. While this is a simple approximation of the impact on emissions following an incremental change in ticket prices it is a reasonable methodology to use when exploring aggregate impacts.
Estimated impact of 50% reduction in APD on Scottish passenger numbers and emissions
18. With all of the data now expressed in some form of APD banding it is possible to carry out a static analysis of the impact of reducing APD across each passenger category to generate an estimate of the likely change in passenger numbers by travel purpose and APD band. Bands B, C and D have been combined in the analysis and the ticket prices weighted by the proportion of passengers travelling to each band.
Methodology and results
19. Table 5 shows that applying a 50% reduction in APD in Scotland cuts the return ticket price by between 3.7% and 8.8% and increases passenger numbers by an estimated 742,000 - a rise of over 3%. Of these additional passengers, 724,000 or 98% are travelling on the reduced APD rate band and when combined with the information on journey purpose suggests that the majority of these additional passengers are travelling for leisure purposes. Internal UK travel makes up 446,000 or 60% of this total increase in passenger numbers.
20. Assuming a linear average relationship between passengers and emissions (total domestic/international emissions divided by total number of domestic/international passengers) suggests the reduction in APD could increase aggregate emissions by 34 KtCO2e (0.034 MtCO2e) over the course of a full year. This represents an increase in transport emissions of just over 0.25% over the current total. The vast majority of the increase in emissions comes from international passenger trips to or from Band A countries travelling at a reduced rate of APD.
Band A domestic
Band A domestic
Band A international
Band A international
Band B, C, D
Band B, C, D
Ticket Price Sensitivities
21. With the large number of assumptions necessary to create this estimate it is appropriate to carry out some sensitivity analysis around the ticket price charged. If ticket prices are 10% lower than the above price estimates the forecast increase in emissions relative to the situation without the reduction in APD rises to 38 KtCO2e. Ticket prices 10% above the levels suggested above reduces the forecast emissions increase relative to the base position to 31 KtCO2e.
Discussion: alternative assessments of the impact of APD on emissions
UK 2009 APD Impact Assessment
22. While not undertaken on the same proposition as is being analysed here, an APD Impact Assessment (IA) carried out by DfT analysts in 2009 can be used to 'benchmark' this assessment of the impact of the 50% reduction in APD in Scotland.
23. The DfT work assessed the impact of the change to the APD scheme announced in the 2009 UK budget statement. This change was a complex one in that the assessment had to consider the impact of moving from two to four APD bands as well as raising the rate of tax for each band in the following year. The IA estimated that the changes to the scheme and the increase in APD would reduce UK annual aviation emissions by 0.6MtCO2e in 2011-12, relative to where emissions would otherwise have been. No more detailed workings were published alongside this overall impact assessment, nor was the assessment broken down by the four countries of the UK.
24. A straight pro rata by population of this 2009 UK DfT estimate suggests the APD increase announced in 2009 reduced Scottish emissions by 0.05 MtCO2e. Basing the impact instead on the split in aviation emissions between Scotland and the UK, and adjusting the ticket prices shown for 2009 in Table 3 for the APD change, increases domestic ticket prices by 2% international ticket prices by 6% and reduces Scottish emissions by 0.026 MtCO2e.
25. The work in this paper suggests a 50% reduction in APD in Scotland reduces current domestic ticket price by 5%, international prices by 4% and together this generates an estimated increase in emissions of 0.034 MtCO2e. A slightly larger percentage price reduction produces a slightly larger increase in emissions.
26. Applying the static methodology developed in this paper to the 2009 UK APD change generates an emissions reduction estimate for the UK as a whole of 0.61 MtCO2e and a fall in passenger numbers of around 3.4 million or 1.7%. This result is very similar to the 0.6MtCO2e reduction estimate from the DfT model. This provides some confidence that the estimated impact generated in this paper of a 50% cut in APD in Scotland is reasonable.
27. On top of the increase in demand from Scottish residents in response to the fall in airline ticket prices there is also potential for passengers in the north of England who currently use Newcastle and Manchester (as well as some Scottish residents who do likewise) to switch to Scottish airports to take advantage of the lower air fares. This assumes of course that the airports provide the same routes and flight timings are similar.
28. Without a fully integrated passenger demand model it is not possible to completely account for the extent of passenger switching. An earlier HMRC paper did however look at the impact of an APD cut in Scotland and suggested that reducing APD in Scotland by 50% might lead to up to 300,000 passengers switching from airports in Northern England (principally Newcastle and Manchester) to Glasgow and Edinburgh.
29. Using this figure and a similar distribution of passenger destinations to the national Scottish picture suggests that this switching impact might add a further 0.017 MtCO2e to the overall impact calculated above. In reality this estimate is likely to be an underestimate because passengers flying domestically from the north to the south of the UK (e.g. Newcastle to Gatwick or Manchester to Heathrow) would be unlikely to travel too far in the opposite direction to a different airport to catch a cheaper flight as this increases their overall travel time. In reality proportionately more of the 'switchers' are likely to be travelling further afield than the Scottish average. As international aviation flights have a significantly higher per passenger emissions estimate than domestic travellers the emissions estimate from switching may be low. Using a split more in keeping with this outcome generates a higher estimate of the increase in emissions of 0.025MtCO2e.
30. This additional impact on Scottish emissions does of course rely on the price differential between the airports remaining. Were this price difference to erode (either through the cutting of ticket prices or a similar reduction in APD for northern English airports) then this additional emissions impact from switching would not materialise.
31. One other consequence of a reduction in APD might be the start-up of additional destinations from Scottish airports. This analysis takes no account of this possibility. Should there be an extension of the route network to destinations in Band C and beyond, particularly for charter flights, then this would add to the total impact on emissions.
32. Reducing APD on flights departing from Scotland is likely to lead to relatively lower airline ticket prices and a particularly marked increase in demand for short haul leisure seats. The impact is estimated to lead to an additional 742,000 seats being sold, most of which attract the reduced rate of APD. These extra trips are forecast to generate an increase in emissions of around 0.034 MtCO2e. While different air fares can and will affect passenger numbers and emissions, the sensitivity analysis carried out suggests the estimated impact on emissions will not change markedly either way from this point estimate.
33. Adding in the potential of passengers switching from northerly English airports to Scottish airports in response to lower air fares from Scottish airports, (estimated by HMRC to possibly be up to 300,000) raises the emissions impact estimate to between 0.05 MtCO2e to 0.06 MtCO2e.
34. The estimates generated in this paper should be seen in the context of the uncertainties within the available data. The data were produced for other purposes and adjustments and assumptions have had to be made. That said, benchmarking the results of this analysis against the earlier DfT analysis suggests that the estimate, including the switching effect, is of the right order of magnitude.
|Impact of 50% reduction in APD (central estimate)||742,000||0.034|
|Impact on central estimate if ticket price is 10% higher than central estimate||-67,000||-0.003|
|Impact on central estimate if ticket price is 10% lower than central estimate||+82,000||+0.004|
|Impact from switching||Up to 300,000||0.017 to 0.025|
|Total impact under central estimate and with switching||Up to 1,042,000||0.05 to 0.06|
Transport Analytical Services