Plan a sampling method for Air Passenger Duty
Find out how to use sampling to calculate Air Passenger Duty using a Special Accounting Scheme.
If you do not have enough information about your passengers to complete an Air Passenger Duty Return, you can use sampling to estimate the percentage of passengers who are:
- liable for each rate of duty
- exempt from duty
You then apply these percentages to the actual number of passengers to calculate the duty you need to pay.
You can only use sampling if HMRC has agreed that you can use a Special Accounting Scheme.
Checklist for statistical sampling
If you decide to use statistical sampling within your scheme, follow these steps.
You must explain these steps in your proposal to use a Special Accounting Scheme.
- Decide the general approach you intend to take, including the data you’ll use, when and how you’ll take samples, and if you’ll take samples for all flights or only some.
- Assess the effects of seasonal variations in terms of the percentages of passengers at each rate of duty.
- Establish rules for how you’ll group flights together.
- Consider methods of choosing which of the flights or passengers within each group will form the sample.
- Select a suitable sample size.
Decide the approach you take
You must plan an accounting scheme that will make the best use of the data available to you, and sampling that produces an accurate calculation for your duty liability.
You may have small under or over payments when you use sampling, but you can lessen the risk of this by:
- having a large sample size
- choosing a sampling method carefully
The data you use for sampling in your special scheme may come from several sources (like a computer system or a passenger survey). There are no restrictions, as long as you can show that the source is accurate and reliable.
Live or current sampling
Current sampling involves using data as soon as it’s available and calculating exemption rates as soon as the sample flights are made.
You can use passenger surveys to collect data if you use live sampling. But, using passenger data from existing systems may be a quicker way to collect data.
There may be more information available at the time when you use live sampling, increasing the accuracy of the calculation. But you’ll need to do sampling often, at intervals, to take account of seasonal variations.
Historical sampling
You can use past data to work out exemption rates for each flight in advance (normally at the start of the year). If you have data from different parts of the year, you can calculate exemption rates for each month or quarter, to take seasonal variations into account.
You can use these percentages at the end of every month to assess the amount of duty due.
The disadvantages of this method is that data might not be available which may cause a lower level of accuracy.
Decide how often to take samples
When you propose a Special Accounting Scheme you must specify when, and how often, you intend to take samples. You’ll need to consider:
- any seasonal variations, for instance, if exemption rates for flights change throughout the year
- desired level of accuracy, you’ll have to do more sampling if you want very accurate results
You may decide to:
- sample monthly, quarterly or biannually
- take more samples when there are more passengers (increasing the number of samples over the summer months)
- sample at fixed periods (for instance, quarterly) but doing additional samples at Easter and Christmas
Group your flights
You may wish to group flights together to reduce the amount of sampling you need to do.
A ‘flight group’ is a group of similar flights that are expected to have a similar percentage of passengers at each rate of duty.
This means a sample from some flights can be used to calculate the exemption rates for the whole group.
You may group flights in many ways. There are no rules as long as each group contains similar flights. The most common groupings are:
- seasonality
- airport of departure
- airport of destination
- type of flight (charter or schedule)
- type of plane
Other factors, such as time of day or time of week may also affect the types of passenger on the flights and might be appropriate.
Choose what to sample
You need to decide whether to sample whole flights or a percentage of passengers, and how you’ll choose your samples.
If you work out your liability by using actual passenger information, and flights within each group are of a similar size, it may be best to sample whole flights.
If you use systems, like a departure control system, then you may want to sample passengers rather than flights.
For instance, you could take a sample of 5% of the total number of passengers in a flight grouping.
If you determine liability by considering all individual flights in detail, you may want to take a 5% sample of passengers from every flight.
There are 2 possible ways for choosing the flights (or passengers) that will form each sample. You can either select:
- flights (or passengers) at random, we can advise you on the best way of doing this
- all flights for a specified period, for example, you might analyse a full week’s flights and use the results for the rest of the season
You must make sure that the sample you choose is representative of the full period.
Decide on the size of your sample
When considering accuracy levels and sample sizes, aim for the minimum requirements set out in the table.
| Number of passengers in the flight group | Minimum sample size guidelines |
|---|---|
| Below 5,000 | Take a 20% sample |
| Between 5,000 and 25,000 | Take a 10% sample |
| Above 25,000 | Take a 5% sample |
In terms of accuracy, these requirements are designed to estimate proportions of passengers at each rate of duty to within 3 percentage points. So if a sample suggested 20% of passengers were exempt, we would expect the true figure to be between 17% and 23%.
The precise level of sampling appropriate to each scheme will vary. Contact the helpline to discuss the application of our minimum requirements for your scheme.
Weighting your sample results
You may need to weight the average percentages for each rate and exemption within a flight group. You need to do this to account for the different numbers and types of passengers who travel:
- at different times of the day or week
- on different routes
The weighted average should take into account the share of relevance for each part, rather than treating each part equally.
Contact the Central Assurance Team if you need to discuss this further.
Conduct a preliminary sampling exercise
If you have never used a special scheme or you’re considering changing your sampling method, we recommend that you conduct a number of sampling exercises.
A preliminary sampling exercise will let you:
- clarify the amount and types of data available to you
- assess the similarity between flights and help choose your flight groups
- decide suitable sample sizes for the main scheme
Example of how to calculate the duty using sampling
In this example the airline only flies to destinations in band A. It shows how to use a sample to calculate the number of passengers:
- at each rate of duty
- exempt as they are on a connecting flight
Percentages should be rounded to 2 decimal places.
You should round the total numbers of passengers as your final figure for each rate must be divisible by the rate for a single passengers, in this case £13 or £26.
Chargeable passengers in sample
| Group 1 flights | Group 2 flights | |
|---|---|---|
| Total number of persons flown in sample | 1,500 | 2,000 |
| Less aircrew and other staff exemptions | 40 | 60 |
| Less deportees and passengers carried under a statutory obligation | Nil | 5 |
| Less infants | 10 | 20 |
| Chargeable passengers in sample | 1,450 | 1,915 |
Sampled passengers at each rate and on connecting flights
| Group 1 flights | Group 2 flights | |
|---|---|---|
| Total chargeable passengers in sample | 1,450 | 1,915 |
| Connecting | 290 passengers = 20% | 465 passengers = 24.28% |
| £26 rate | 1,160 passengers = 80% | 450 passengers = 23.50% |
| £13 rate | 0 passengers | 1,000 passengers = 52.22% |
Calculate total passengers at each rate and on connecting flights
| Group 1 flights | Group 2 flights | |
|---|---|---|
| Total chargeable passengers flown in period | 12,000 | 25,200 |
| Connecting exemptions | 20% = 2,400 passengers | 24.28% = 6,119 passengers |
| £26 rate | 80% = 9,600 passengers | 23.50% = 5,922 passengers |
| £13 rate | 0 passengers | 52.22% = 13,159 passengers |
Final calculation
| Group 1 flights | Group 2 flights | |
|---|---|---|
| Duty due at £26 rate | 9,600 x £26 = £249,600 | 5,922 x £26 = £153,972 |
| Duty due at £13 rate | 0 x £26 = £0 | 13,159 x £13 = £171,067 |
| Total Air Passenger Duty due | £249,600 | £325,039 |