Improving our travel and tourism statistics: October 2026 update

• This article outlines our progress in developing new methods and the data underpinning our travel and tourism statistics; this also accompanies the release of our updated 2025 estimates.

• We have consistently published updated estimates for the period between Quarter 3 (July to Sept) 2024 and Quarter 4 (Oct to Dec) 2025; this means our users now have six quarters of comparable estimates under the new International Passenger Survey (IPS) design, as outlined in our Improving our travel and tourism statistics: September 2026 update article.

• This new consistency in methodology allows users to make comparisons between these six quarters; however, we do not recommend comparing 2024 annual estimates with 2025 annual estimates, as the new IPS design was implemented part-way through 2024.

• This article also presents findings from an interim assessment into the potential presence of discontinuity in headline estimates, caused by the change in IPS design in Quarter 3 2024; we require further data collection and method development before we can confidently estimate the size of the discontinuity between estimates under the old and new designs.

• We are now focusing on additional methodology areas for further development and investigation, relating to weighting, sampling, processing, geography coverage, and the way we measure UK residents' spend abroad; in the meantime, we will publish future travel and tourism estimates consistent with those published for Quarter 3 2024 to Quarter 4 2025.  

• We invite user feedback on our published estimates and developing methods.

Overview

In 2022, we conducted a comprehensive review of travel and tourism statistics to investigate and recommend how these statistics should be transformed to meet user needs, improve quality, and establish a plan to deliver this change. We have made progressive methodological improvements since the implementation of the new International Passenger Survey (IPS) design, in July 2024.

We have detailed these improvements in previous releases of our Improving our travel and tourism statistics series. While we have made substantial progress since the initial consultation, there are still areas that require ongoing improvement. The purpose of this article is to provide updates on our current progress and our roadmap for future developments over the coming year.

In our Improving travel and tourism statistics: September 2026 update article, we outlined the latest data collection changes and methodological improvements implemented for Quarter 3 (July to Sept) and Quarter 4 (Oct to Dec) 2024. In this release, we have published estimates for Quarter 1 (Jan to Mar) to Quarter 4 2025 on the same basis, so that that there are now six quarters of comparable estimates under the new International Passenger Survey (IPS) design.

This section summarises our progress on the main areas of progress.

Improving our coverage of travel and tourism statistics

The methodological approach to IPS sampling changed notably when the new survey design was introduced in July 2024. Sampling at airports was moved from a terminal-basis to a route-basis, an approach that aligns with producing more precise statistics at individual country level. 

We have introduced pseudo records in the data for Dover freight and Eurotunnel freight to reduce bias from not sampling enough freight traffic. These represent passengers who can no longer be sampled under the new method (such as those from Dover) because of health and safety policies, or those that we do not have a large enough sample for (such as those from Eurotunnel).

UK-level estimates

From 1 April 2024, IPS data collection in Northern Ireland (NI) stopped. Data on international tourism to and from NI are now provided by the Northern Ireland Statistics and Research Agency (NISRA). Our Travel trends: 2024 article and our Travel trends: 2025 article include UK-level world totals in their headline estimates only. The remaining estimates and breakdowns cover Great Britain-level geographies only, and our quarterly tables continue to cover the Great Britain-level only.

We have previously published updates on the challenges in incorporating NI data into quarterly travel and tourism estimates produced under the new IPS design. NISRA uses four different data sources to compile international travel and tourism statistics for NI. NISRA only publishes tourism data once all four data sources for the reference period have been collected and aggregated. NI quarterly data are not available until five months after the end of the reference period, which is too late for these data to be incorporated into our quarterly publications.  

We conducted a user needs review over spring 2026, which showed that while most users can use the data that we are currently providing, they would favour having UK-level estimates. This need was strongest for UK-level annual estimates, compared with quarterly estimates.

Data processing and weighting

As outlined in our Improving our travel and tourism statistics: September 2026 update article, we have revised our approach to calibration and now use more detailed country-level data introduced through a second partition of the population totals used in weighting. This new methodology has been applied from Quarter 3 2024 onwards. This is one of the improvements that is now consistent across the six comparable quarters and allows for more precise estimates at the country level.

We are considering how we include and estimate spending by interliners in our statistics. Decisions around this will be led by user needs, and we are interested in better understanding how our users want interliner passengers' trips through the UK to be represented in our statistics. 

Post-departure spending on trips abroad

We previously reported that some Great Britain Tourism Survey (GBTS) respondents were including their pre-departure accommodation spend in their reported post-departure spend. Qualitative research is currently underway to explore options for improving collection of post-departure spending in the GBTS. This includes improving estimates for the "fly or cruise stay on board" category. 

In the meantime, we are continuing to remove reported post-departure accommodation spending before fitting the GBTS model. As a result, estimates of overseas spending by Great Britain (GB) residents continue to rely on assumptions and data sources with known limitations. For this reason,it is likely that estimates of overseas spending by GB residents from Quarter 3 2024 will be underestimated.

These limitations also affect how we measure and communicate uncertainty in overseas spending estimates. Uncertainty from the GBTS model is also used in the calculation of standard errors and confidence intervals for overseas spending by GB residents' estimates. While this work is ongoing, we have paused publication of confidence intervals for GB residents' overseas spending in our quarterly and annual datasets. We also remind users to exercise caution when using these estimates

The changes outlined in Section 3: Progress updates have improved the overall quality of these statistics so far, although we have further developments to complete.

We are committed to transparency in our proposed timelines and have an intended roadmap for improvement. Many of the quality improvements have dependencies on others and progress is not always linear. Some developments are likely to be complete within the next twelve months, while others are longer-term.

The following timelines represent our current aims and are subject to change.

In the next 12 months

• We will be conducting further work to make our data processing pipelines more reproducible, with less reliance on manual processing by early 2027.

• Following user feedback, we will be exploring how Northern Ireland (NI) data can be further disaggregated in our annual publications with a view to producing more UK-level breakdowns in our Travel Trends: 2026 bulletin, publishing by autumn 2027.

• We will continue to publish quarterly estimates from Great Britain (GB) only, with the timeliness of the data delivery from the Northern Ireland Statistics and Research Agency (NISRA) likely to improve and fit better with our publishing tables; we will review this decision at the end of 2027.

• We are making further improvements to data on freight traffic at Dover Seaport and Cheriton Eurotunnel; we increased the sample size at Cheriton Eurotunnel from Quarter 1 (Jan to Mar) 2026, and we are investigating whether we now have sufficient data on freight passengers to replace Eurotunnel pseudo records.

• We are also seeking alternative data sources from other government departments on passengers traveling through Dover Seaport; this is intended to improve the accuracy of pseudo records for Dover freight, which currently relies on information provided on the Port of Dover website, only available up to 2024.

• When we moved to the new International Passenger Survey (IPS) design, the existing imputation system was uplifted into the new processing pipeline, but a full review of the methods was not carried out; now that we have implemented most refinements to our sampling and weighting approach, we are taking the opportunity to begin a review of the imputation methods, expected to conclude in late 2027, which will inform whether any changes to the method are required.

• We also plan to review the apportionment method for our lower-level regional breakdowns and improve measures of uncertainty for these estimates; this is now a lower priority workstream as we focus on tasks of greater precedence, but we plan to complete the initial assessment by September 2027.

After 2027

Once we have identified a suitable solution for accurately collecting data on post-departure spend for GB residents, we will develop and test new models for estimating post-departure spend for IPS respondents. This will include improving treatment of unusually high spending values (outliers). To maintain comparability of estimates of GB spend abroad, we are considering a phased implementation to allow the current and revised methods to run in parallel for a period of time.

We will continue to produce estimates of spending for GB residents abroad using the current approach in the meantime, which will enable us to assess the impact of the revised approach, including any discontinuities introduced in the GB resident overseas spend series (and the magnitude of these). The outcome of this will determine whether adjustments to estimates from Quarter 3 (July to Sept) 2024 onwards are required to maintain a comparable time series before introducing changes to the methodology.

This work is estimated to take 12 months, so the earliest we expect to be able to produce improved estimates of spend abroad for GB residents using new methods is summer 2028. As part of this, we will also review how uncertainty from the Great Britain Tourism Survey (GBTS) model is reflected in the standard errors and confidence intervals for overseas spending by GB residents' estimates, so that we can start publishing confidence intervals for GB residents spending abroad again.

Longer-term work includes investigating the feasibility of including Civil Aviation Authority (CAA) departures survey data in travel and tourism statistics. The CAA survey asks some questions of departing passengers that are also asked on the IPS, potentially allowing us to make use of a larger, aggregate sample size. However there may be differences in data quality between the two surveys, and each of the surveys have their own, distinct sample design, meaning that combining the data is not a simple task. We will need to assess and compare CAA and IPS data, and research and develop weighting and estimation methods for using CAA data in our travel and tourism estimates.

One of our aims over the coming years is to improve the timeliness of our publication schedule. Because of methodological improvements that have taken time to implement and associated issues that have arisen from it, some of our releases have been delayed. Most recently, our need to make a correction to our Travel trends: 2024 article meant that we had to postpone the publication of our Travel trends: 2025 article, which has consequences for subsequent releases. We are developing a schedule that allows us to prioritise the necessary quality improvements while trying to reduce the time gap between the publications.

Discontinuities are changes in a time series that are caused by changes in data collection or statistical methodology, making comparisons over time difficult. We are investigating whether our new methods for collecting and processing travel and tourism statistics have introduced discontinuities.

In this section, we present provisional findings on the potential magnitude of discontinuity in headline travel and tourism estimates between the old and new International Passenger Survey (IPS) designs. Our analysis uses the first six quarters of data available under the new design, from Quarter 3 (July to Sept) 2024 to Quarter 4 (Oct to Dec) 2025. A longer time series is needed before discontinuity can be assessed with greater confidence.

While survey methods are still under final refinement, discontinuity estimates remain liable to change.

Methods

Our preliminary analysis uses quarterly publication tables covering visits and nights by area of residence or visit for inbound and outbound travel. It also uses spend by area of residence for inbound travel only.

To assess whether the change in methodology introduced a discontinuity, we used State Space Models (SSM) to estimate the historical trend and seasonal pattern of each series. This is a powerful modelling technique that accounts for differences in the precision of our quarterly estimates. The models are used to project the estimates that might have been expected after Quarter 3 2024 if the historical pattern had continued unchanged, which provides a reference from which to identify and investigate differences.

We compared modelled expected values with estimates produced under the new methodology. Persistent differences provide evidence of potential discontinuity, while the size and uncertainty of the estimated difference help assess whether there is evidence of a meaningful level shift.

Figure 1: Conceptual illustration of potential level shifts from Quarter 3 (July to Sept) 2024

Notes:

1. The period from 2020 to 2022 was not included in the modelling because of data collection difficulties during the coronavirus (COVID-19) pandemic. The State Space Model (SSM) can account for missing periods without treating modelled or interpolated values as observed data. 
2. The green line shows the observed series, while the blue line shows the expected level derived from the historical data (SSM). The vertical dashed line marks the introduction of the new methodology, in Quarter 3 2024. The difference between the observed and expected values after this point represents any level shift. 
3. These data are illustrative and do not represent true International Passenger Survey (IPS) estimates.

Results and recommendations

At this early stage, most estimated shifts are negative (Tables 1 and 2), suggesting that many estimates produced under the new methodology may be lower than expected if previous trends and seasonal patterns had continued. For example, the estimated relative level shift of negative 3.4% for spending by European residents visiting Great Britain (Table 2) would mean that estimates are, on average, around 3.4% lower than modelled expected figures after accounting for trend and seasonality.

For high-level inbound and outbound series, only a limited number of estimated shifts are currently statistically significant (Tables 1 and 2). This does not mean that there is no discontinuity. Rather, it means that the initial analysis remains inconclusive because it is based on limited time periods.

The ability to detect a level shift varies across series. High-level estimates based on larger samples, such as world totals, are generally more precise. For more disaggregated series based on smaller samples, greater volatility can make level shifts harder to estimate precisely and less likely to be statistically significant.

Uncertainty is expected to reduce as more data are added, improving our ability to detect discontinuities where they exist. At this stage, statistical significance alone cannot be used to formally declare a discontinuity or assess its practical importance.

Therefore, we are considering the estimated magnitude of each shift, its uncertainty, and the underlying variability of each series. As more data become available, we will be better able to assess whether apparent shifts are stable and materially important, and we will provide a further update as this work progresses.

How these figures can be used

These findings provide an initial indication of the direction and possible magnitude of discontinuities following methodological changes. They show progress in our investigation into comparability over time.

The results remain preliminary and are likely to change as more data are collected under the new design, and as the time series modelling framework develops. They do not represent final discontinuity estimates or an agreed method for adjusting historical series.

The magnitude of discontinuity also varies across the series analysed. For example, visits from "other" countries show a higher interim impact figure, reflecting the introduction of more relevant country-specific inputs under the new route-based design.

Therefore, users should not apply the estimated level shifts to create their own adjusted historical time series. Although the modelling approach provides a suitable framework for investigating discontinuities, a longer time series under the new methodology is needed before adjustments can be made robustly.

Next steps

1. We are continuing to develop the method for assessing discontinuity, including adding external data sources into the model to help differentiate between discontinuity driven by methodological changes and genuine changes in travel behaviour; we hope to have a final method, which has undergone methodological assurance, by summer 2027.

2. We are reliant on sufficient data collected under the new design to reliably detect the magnitude of discontinuity; this is the greatest area of uncertainty, and we are working with methodological experts to understand how much data is required under the new design.

3. As we continue to refine our methods for producing travel and tourism statistics, we expect the level of discontinuity between estimates produced under the old and new IPS designs to change; our aim is that by autumn 2027 most of the refinements to the sampling and weighting methodology will be implemented, and estimates will begin to stabilise.

For these reasons, we are still exploring when it will be feasible to start adjusting the historical back series to reintroduce a comparable timeseries. We are considering a phased approach, with estimates of overseas residents' visits and spending visits likely being adjusted first, followed by estimates of visits and spending abroad. This reflects the different timelines for the development of these series.

In the meantime, we will continue to keep users updated on our progress in assessing discontinuity and our timelines for reinstating comparability of historical estimates, as they become more certain.

Interliner

Passengers whose main purpose for being at an airport is to change flights.

Discontinuity

Changes in time series resulting from a sudden change in the average level of the series or a complete shift in periodic effects, such as seasonal patterns.

Overseas residents

An overseas resident is a person who permanently resides outside the UK.

Pseudo records

Pseudo records allow us to include estimates for missing groups by adding equivalent artificial records to the sample file where we have gaps.

Visits

The figures for visits relate to the number of completed visits, not the number of visitors. Anyone entering or leaving more than once in the same period would be counted on each visit.

Visits by overseas residents to Great Britain or the UK

A visit by an overseas resident is a visit for a period of less than 12 months. Great Britain or UK citizens residing overseas for 12 months or more coming home on leave are included in this category.

Visits outside the UK by residents of Great Britain or the UK

Visits for a period of less than 12 months by people permanently residing in Great Britain or the UK, who may be of foreign nationality.

Weighting

A method used in an attempt to make the responding sample of passengers representative of all passenger trips. The final weight includes components that represent the sample design, non-response, and calibration to known passenger totals to improve consistency.