Author: admin

8. Data sources and quality

Data sources

Balance of payments statistics are compiled from a variety of sources, and produced using the national accounts sector and financial accounts (SFA) framework. Some of the main sources used include:

• overseas trade statistics (HM Revenue and Customs (HMRC))

• International Trade in Services Survey (ITIS) from the Office for National Statistics (ONS)

• International Passenger Survey (ONS); this was suspended between March 2020 and January 2021 because of the coronavirus (COVID-19) pandemic, and is currently undergoing transformation

• Foreign Direct Investment Survey (ONS and Bank of England (BoE))

• various financial inquiries (ONS and BoE)

• Ownership of UK Quoted Shares Survey (ONS)

Trade is measured through both exports and imports of goods and services. Data are supplied by over 30 sources, including several administrative sources, with HMRC being the largest for trade in goods. The ITIS, conducted by the ONS, is the largest single data source for trade in services.

The main source of information for UK foreign direct investment (FDI) statistics is our FDI Survey. Separate surveys are used to collect data on inward and outward FDI. This is combined with data from the BoE on the banking sector.

The statistics in this bulletin are compiled using the asset and liability measurement principle, which uses residency as the main distinction between outward and inward investments.

Changes affecting UK trade statistics

HMRC error affecting mineral fuels and oils

In this publication and our GDP quarterly national accounts, UK: July to September 2025 bulletin, we have corrected HMRC Overseas Trade (OTS) data, following the identification of an error. This correction is for exports of mineral fuels and oils and affects data back to March 2024. This release revises data back to Quarter 1 (Jan to Mar) 2024. Please see our UK trade: October 2025 bulletin for further information, including indicative impacts. In line with our National Accounts Revisions policy, we will use the corrected HMRC data in our UK trade: November 2025 release, scheduled for 15 January 2026.

Precious metals methods improvement

As previously communicated, when implementing improvements to recording trade in precious metals as part of our Methods improvements for Blue Book and Pink Book 2025, we removed the double counting of some precious metals bars and included previously under-recorded, non-monetary gold that is not in bar form. 

These improvements, however, were not fully applied to a small number of countries in 2024 and 2025 because of a processing error. For full details including indicative impacts previously published, please see our Balance of payments, UK: April to June 2025 bulletin and UK trade: October 2025 bulletin.  

We have now corrected the processing error and, in this release, we fully implement this improvement. Estimates of trade in goods for Quarter 1 2024 to Quarter 2 (Apr to June) 2025 were updated in our UK trade: August 2025 bulletin and GDP first quarterly estimate, UK: July to September 2025 bulletin. These corrected estimates were completed exceptionally outside of the usual National Accounts Revisions Policy to provide consistency between GDP and UK trade data. As a result of the corrections included in this publication, estimates are now consistent between UK trade, GDP, and balance of payments publications. 

Data collection changes

Since the UK left the EU on 31 January 2020, the arrangements for how the UK trades with the EU changed. HMRC implemented some data collection changes following Brexit, which affected statistics on UK trade in goods with the EU. We have made adjustments to our estimates of goods imports from the EU in 2021 and 2022 to account for these changes, however, a structural break remains in the full time series for goods imports from and exports to the EU from January 2021. 

We therefore advise caution when interpreting and drawing conclusions from these statistics. Our Impact of trade in goods data collection changes on UK trade statistics: summary of adjustments and the structural break from 2021 article provides more detail. 

International trade in services estimates

From September 2025 until early 2027, International trade in services survey (ITIS) data, which account for approximately 50% of total trade in services, are being processed once per quarter. During this period, the data are based on a survey response rate of approximately 60% to 70%. This enables more focus on improving processing systems and ensuring methods and quality in the future.

The ITIS data that currently inform trade in services estimates are based on benchmarked annual 2023 survey data and quarterly ITIS survey data for periods from Quarter 1 (Jan to Mar) 2024 onwards. We will incorporate benchmarked annual 2024 data in the future, in line with the National Accounts Revisions Policy.

The International Passenger Survey (IPS), which is the source of travel services estimates, accounting for approximately 8% of total trade, is being transformed under our travel and tourism project. The travel services estimates have been forecast since Quarter 3 2024, and will be forecast during travel and tourism transformation.

Financial sector statistics

Our Financial Services Survey (FSS) transformation will improve the quality of our financial sector statistics. During the period of transformation, starting from Quarter 1 2024, financial services statistics in this release are based on forecasts.

Quality and methodology

More quality and methodology information on strengths, limitations, appropriate uses, and how the data were created is available in our Balance of payments quality and methodology information (QMI).

We will continue to produce our UK balance of payments statistics in line with the UK Statistics Authority’s Code of Practice for Statistics, and in accordance with internationally agreed statistical guidance and standards. This is based on the International Monetary Fund’s Balance of Payments and International Investment Position Manual: Sixth Edition (BPM6) (PDF, 3.0MB), until those standards are updated.

Accredited official statistics

These accredited official statistics were independently reviewed by the Office for Statistics Regulation in December 2011. They comply with the standards of trustworthiness, quality and value in the Code of Practice for Statistics and should be labelled “accredited official statistics”.

Back to table of contents

12. Data sources and quality

The three approaches to measuring GDP

There are three approaches to measuring gross domestic product (GDP):

• the output approach

• the expenditure approach

• the income approach

The data and data quality are different for each approach, and this dictates the approach taken in balancing quarterly data. There are more data available on output in the UK in the short term, than in the other two approaches. To get the best estimate of GDP, our published figure, estimates from all three approaches are balanced to produce an average, except in the latest two quarters where the output data take the lead, because of the larger data content.

The three approaches to measuring GDP allow us to confront our data sources within the national accounts framework. Figure 3 shows that the three approaches to measuring GDP are closely aligned. However, there can still be uncertainty at the component level, at this stage in the production cycle for 2024 and 2025, until these data have been confronted through the supply and use tables (SUTs) framework. This uncertainty may be for various reasons and is discussed further later in this section.

Output approach

In the output approach, we do not currently have final estimates for intermediate consumption (the value of goods and services purchased to be used up in the production of goods and services). This is outlined in our Blue Book 2025: advanced aggregate estimates article. Initially, we use turnover and output as a proxy for changes in gross value added. We assume that the intermediate consumption ratio by industry, calculated in 2023, holds constant into 2024 onwards. More information on this is provided in Section 11: Data sources and quality of our GDP quarterly national accounts, UK: April to June 2024 bulletin.

Expenditure approach

In the expenditure approach, we currently have lower response rates for areas, such as the Living Costs and Food Survey, which is one of many data sources that inform our estimates of household consumption. We therefore rely on additional indicators, such as our Monthly Business Survey, to quality adjust some of our estimates in the short term.

Income approach

In the income approach, we do not have up-to-date quarterly information on the gross trading profits of businesses. These data are collected from HM Revenue and Customs (HMRC) and are available with a lag of approximately two years.

We rely on contextual data from other sources to inform these quarterly estimates, as outlined in our Profitability of UK companies quality and methodology information (QMI). There is currently more uncertainty around the compensation of employees figures in this release because of lower response rates in our Labour Force Survey (LFS), as described in our LFS: planned improvements and its reintroduction methodology. We have used additional information from our Earnings and employment Pay As You Earn Real Time Information, UK: January 2025 bulletin to help inform the estimates.

Reaching the GDP balance

Quarterly GDP is a balanced measure of the three approaches. The GDP monthly estimate focuses on gross value added (GVA) and output as a proxy for GDP. This results in data differences, in both levels and growth terms, between our quarterly bulletins (average GDP) and our GDP monthly estimate bulletins (output approach to GDP). Quarterly GDP is the lead measure of GDP because of its higher data content and inclusion of variables, that enable the conversion from a GVA concept to a GDP basis.

Information on the methods we use is in our Balancing the output, income and expenditure approaches to measuring GDP report.

Alignment adjustments, found in Table M of our GDP data tables, have a target limit of plus or minus £3,000 million on any quarter. However, in periods where the data sources are particularly difficult to balance, larger alignment adjustments are sometimes needed. This is explained in more detail in our Recent challenges of balancing the three approaches of GDP article. Our standard practice is to prefer that the alignment adjustment be out of tolerance rather than over-adjust individual GDP components to achieve a balance. This is most likely to occur in the latest quarter, where the constraints are larger, and where we must align to the output estimate for the change in GDP, and where the data content is at its lowest.

To achieve a balanced GDP dataset through alignment, we apply balancing adjustments to the components of GDP where data content is particularly weak in each quarter because of a higher level of forecast content. Table 7 shows the balancing adjustments applied to the GDP quarterly national accounts dataset.

Net trade

Since the UK left the EU on 31 January 2020, arrangements for how the UK trades with the EU changed. HMRC implemented some data collection changes following Brexit, which affected statistics on UK trade in goods with the EU. We have made adjustments to our estimates of goods imports from the EU in 2021 and 2022 to account for these changes. However, a structural break remains in the full time series for goods imports from, and exports to, the EU from January 2021.

We advise caution when interpreting and drawing conclusions from these statistics. More detail is provided in our Impact of trade in goods data collection changes on UK trade statistics: summary of adjustments and the structural break from 2021 article.

International Trade in Services estimates

From September 2025 until early 2027, International Trade in Services (ITIS) data (which account for approximately 50% of total Trade in Services) will be processed once each quarterly period. During this period, the data will be based on a robust survey response rate of between approximately 60% and 70%. This will enable more focus on improving processing systems and ensuring methods and quality in the future. Forecasted data for Quarter 3 (July to Sept) 2025 have now been replaced with ITIS-based estimates.

ITIS-based data in Trade in Services estimates at first quarterly estimate will be forecast until early 2027.

The International Passenger Survey (IPS), which is the source of travel services estimates (accounting for approximately 8% of total trade), is being transformed as part of our Improving our travel and tourism statistics project, and travel services estimates have been forecast since Quarter 3 (July to Sept) 2024. Estimates will be forecast during the period of the travel and tourism transformation.

Our Financial Services Survey (FSS) is undergoing transformation to improve the quality of our financial sector statistics. During the period of transformation, starting from Quarter 1 2024, financial services trade statistics in this publication are based on forecasts.

Restarting of Producer Prices publications

We restarted publication of our monthly business prices publications on 22 October 2025. Business prices data with corrected chain-linking methods and updated historical weights have been used in our monthly GDP datasets for Producer Price Indices (PPI), Import Price Indices (IPI), Export Price Indices (EPI), and Service Producer Price Indices (SPPI) in this release. The quarterly SPPI estimates are splined to months for use in monthly GDP calculations.

These updates to business prices data will be incorporated in GDP estimates in line with our National Accounts Revisions Policy, becoming fully integrated for the entire time series in our Blue Book 2026 publication.

Further information on the chain-linking error and the impact of methodological changes in the producer prices dataset are detailed in our Impact of correction to chain-linking methodology used in Producer Price Indices and Services Producer Price Indices: October 2025 article.

Strengths and limitations

The UK National Accounts are drawn together using data from many different sources. This ensures that they are comprehensive and provide different perspectives on the economy, for example, sales by retailers and purchases by households. Further information on measuring GDP can be found in our Guide to the UK National Accounts. More quality and methodology information is available in our GDP quality and methodology information (QMI).

Seasonal adjustment

The headline estimates of quarterly GDP are seasonally adjusted. Seasonal adjustment is the process of removing the variations associated with the time of year, or the arrangement of the calendar, from a data time series.

GDP estimates, as for many data time series, are difficult to analyse using raw data because seasonal effects dominate short-term movements. Identifying and removing the seasonal component leaves the trend and irregular components.

We use the X-13-ARIMA-SEATS approach to seasonal adjustment. Seasonal adjustment parameters are monitored closely and regularly reviewed. For more information, please see our seasonal adjustment methodology page.

In our quarterly GDP estimates, seasonal adjustment is applied at a low level, and the seasonally adjusted series are aggregated to create estimates by sector and total output. As part of our quality assurance approach, residual seasonality checks are regularly completed by our time series analysis team on both the directly seasonally adjusted series, and also the indirectly derived aggregate time series.

This topic is explored further in Section 5: Case study: quarterly GDP of our Assessing residual seasonality in published outputs article, updated on 30 September 2025.

Important quality information

There are common pitfalls in interpreting data series. These include:

• expectations of accuracy and reliability in early estimates are often too high

• revisions are an inevitable consequence of the trade-off between timeliness and accuracy

• early estimates are often based on incomplete data

Very few statistical revisions arise because of “errors” in the popular sense of the word. All estimates, by definition, are subject to statistical “error”.

Many different approaches can be used to summarise revisions. The section on Accuracy and reliability in our GDP QMI analyses the mean average revision and the mean absolute revision for GDP estimates over data publication iterations. For more information, please refer to our GDP revisions in Blue Book: 2025 article, published on 31 October 2025.

Accredited official statistics

These accredited official statistics were independently reviewed by theOffice for Statistics Regulation in October 2016. They comply with the standards of trustworthiness, quality and value in the Code of Practice for Statistics and should be labelled “accredited official statistics”.

Back to table of contents

by Dimitris Georgarakos, Geoff Kenny, Luc Laeven and Justus Meyer^[1]

In this article, we draw on rich micro-level data from the ECB’s Consumer Expectations Survey (CES), covering the 11 largest euro area countries. Our findings suggest that many consumers would be open to using central bank money in digital form – a digital equivalent of cash. Under normal conditions, this digital equivalent of cash would primarily be used, as intended, for transactions rather than as an investment or savings vehicle. Accordingly, the introduction of digital central bank money is estimated to lead to only a small reallocation of liquidity away from bank deposits. Different holding limits between €1,000 and €10,000 would have quite small and insignificant effects on the composition of households’ liquid asset holdings. We also find evidence that effective communication can play an important role in informing consumers about this new form of payment and encouraging adoption.

Several central banks, including the European Central Bank (ECB), are pursuing plans to potentially introduce a central bank digital currency (CBDC) as a form of central bank digital money alongside physical money. Such plans are being drawn up in the context of a fast‑changing payment landscape – marked by reduced usage of cash, rising geostrategic considerations, and the rise of fintech (Berg et al., 2024; and Cong and Mayer, 2025). At the same time, central banks are trying to avoid a potentially destabilising allocation of wealth into central bank digital money. In particular, it might result in financial disintermediation – a move away from traditional bank deposits into central bank digital money – and pose threats to financial stability (e.g. Andolfatto, 2021; Ahnert et al. 2023, 2024; Whited et al., 2023; Schilling et al., 2024; and Bidder et al., 2024). For this reason, in designing the digital euro, the Eurosystem has included several safeguards in the design of the digital euro design to avoid that it might result in financial disintermediation.^[2]

In Georgarakos et al. (2025), we attempt to shed light on the potential effects of the prospective introduction of a digital euro and the role it could play in households’ financial asset portfolios. First, we document how many consumers are aware of the ECB’s plans to potentially introduce a digital euro and whether they would be willing to adopt it. Second, we look at the risks of financial disintermediation during non-crisis times. We do this by showing how the availability of a digital euro would affect households’ portfolio allocation choices and how holding limits – i.e. the maximum amount of digital euro that could be held in a digital wallet at any point in time – would influence their choice to reallocate liquid wealth to a digital euro. Third, we demonstrate how effectively communicating the key features of the CBDC to consumers could increase the likelihood that they would adopt it. Throughout, we draw on rich micro-level data from the ECB’s Consumer Expectations Survey (CES), covering the 11 largest euro area countries.^[3]

Awareness and adoption: who would use a digital euro?

On 14 July 2021 the Governing Council of the ECB announced the launch of the investigation phase of the digital euro project. According to CES data collected back in 2021, only about 9% of respondents had heard about the digital euro (see Ehrmann et al., 2025). But this had increased to around 40% by March 2024, reflecting the ECB’s public communication efforts (Chart 1, panel a). In addition, the CES results show that about 45% of all surveyed consumers indicate they would be likely to adopt a digital euro and use it in their daily lives (Chart 1, panel b).^[4]

However, consumers’ propensity to adopt digital central bank money differs across use cases and is highest for online and in-person retail payments (Chart 1, panel b). Looking at different groups of consumers, we also document substantial differences in their willingness to use a digital euro. For example, younger, higher-income and more educated consumers are more likely to adopt digital central bank money. Among consumers aged 18-34, about 55% said that they would probably use a digital euro for any of the different use cases, and among consumers in the highest income quartile, 53% said they would likely use digital central bank money for any of the use cases. There are also some less pronounced differences across groups of consumers with different levels of education; 48% of consumers who have a bachelor’s degree or above would likely use a digital euro compared with 42% of consumers who do not have a bachelor’s degree. Overall, the differences in adoption propensity across demographic groups highlight how important it is to consider financial inclusion when designing and introducing digital central bank money.

Household portfolio choices if digital central bank money were available

Another frequently discussed question is how the availability of digital central bank money – an unremunerated, highly liquid, safe asset directly issued by the central bank – might affect consumers’ portfolio allocation. In this respect, earlier research by Keister and Sanches (2023) shows a liquidity-safety trade-off, with bank deposits yielding a return but involving default risk whereas central bank digital money, being risk-free, has the potential to crowd out deposits. Overall, while the introduction of central bank digital money could thus help improve liquidity allocation in the economy, if no safeguards are applied, there is also a risk that it might lead to a shrinkage in the deposit base, an increase in bank funding costs or, in extreme cases, disintermediate banks.

Therefore, a central question arises: how would households adapt their financial asset portfolios following the introduction of digital central bank money? In order to try and answer this question, we asked CES respondents to imagine they had received a windfall of €10,000 (e.g. a bonus, a gift or a bequest). We then asked consumers to report how they would allocate this unexpected gain in wealth across different assets, including liquid asset holdings like physical cash and the digital euro. This approach is in line with recent developments in survey methodology (Stantcheva, 2023). Previous studies have used similar scenario questions to examine how consumption responds to hypothetical positive wealth shocks (e.g. Shapiro and Slemrod, 2003; and Christelis et al., 2025).

Our results indicate that consumers would only allocate on average about 5% of the hypothetical windfall of €10,000 to a digital euro (Chart 2). This level of allocation does not significantly crowd out other asset classes. In contrast, consumers would allocate more than half of the windfall to current accounts or savings accounts (53%). They also choose to allocate 12% to physical cash which is the closest substitute to a digital euro in the sense that both cash and the digital euro are central bank money, with the difference being that cash cannot be used as a substitute when it comes to paying digitally. But the average portfolio share allocated to digital euro exceeds the share allocated to highly risky, alternative digital crypto-assets. These results suggest that, under normal conditions, digital central bank money would act primarily as a transactional asset rather than an investment or savings vehicle. So the introduction of a digital euro would result in only a small reallocation of liquidity away from retail bank deposits.

One important design feature of a digital euro, aimed at addressing any possible unintended consequences for bank disintermediation, concerns the maximum amount of digital euros that could be held in a digital wallet at any point in time. To assess how such holding limits could affect portfolio allocations, we randomly assigned different possible holding limit values to CES respondents, ranging from a limit of €1,000 to limits of €10,000. Subsequently, we asked respondents how much money they would allocate out of their current liquid resources to a digital euro, subject to the holding limit they had been randomly assigned. This way, we could estimate the effect of the holding limits on the fraction of consumers’ current total liquid resources (e.g. in current/savings bank accounts or in cash) that they would allocate to a digital euro. Our results suggest that the size of the holding limit made no noticeable difference to the allocation of current liquid resources to a digital euro. These findings are consistent with the view that, under normal conditions, CBDC adoption is unlikely to trigger large-scale financial disintermediation.

Targeted communication can increase adoption

While 45% of consumers reported in 2024 that they were willing to adopt a digital euro, other consumers reported that they were unlikely to adopt it. This raises the question as to whether central bank communication can contribute to broadening adoption and use. To assess this, we made use of a randomised control trial embedded in the March 2024 CES web survey round. It provided a random sub-group of respondents with an official ECB video, which gave information about the digital euro’s key features and potential uses. We could then estimate the effect of this video communication on the likelihood of consumers adopting a digital euro.

We found that the information in the video had sizeable and statistically significant positive effects on the propensity of adoption for each of the four possible uses. For example, it increased the propensity of adoption for offline retail payments by 13 percentage points. In fact, the video communication not only affected adoption propensities among consumers without any prior knowledge, but also among those who had already heard about the CBDC before watching the video. Nonetheless, there is also evidence of habit persistence in payment preferences. Some consumers, even after having watched the video, reported a low propensity of adoption. Additionally three months later, in June 2024, we asked all CES respondents whether they were aware of the digital euro. As in previous studies (e.g. Coibion et al., 2024), we found evidence that the effect of the information in the video seemed to fade quite quickly.

Conclusions

Understanding the possible future consumer response to the introduction of digital central bank money is an important research and policy priority. In our recent research, we provide several new insights. For example, our results suggest that many consumers would be open to using a digital euro. In addition, under normal conditions and as intended, a digital euro would primarily be used for transactions rather than as an investment or savings vehicle. Accordingly, the introduction of a digital euro is estimated to lead to only a small reallocation of liquidity away from bank deposits. Our findings also suggest that different holding limits from €1,000 up to €10,000 would have quite small and insignificant effects on the composition of households’ liquid asset holdings. And last but not least, we find evidence that effective communication can play an important role in informing consumers about this new form of payment – and encouraging them to adopt it.

References

Ahnert, T., Hoffmann, P., Leonello, A. and Porcellacchia, D. (2023), “Central Bank Digital Currency and Financial Stability”, Working Paper Series, No 2783, ECB.

Ahnert, T., Assenmacher, K., Hoffmann, P., Leonello, A., Monnet, C. and Porcellacchia, D. (2024), “The Economics of Central Bank Digital Currency”, International Journal of Central Banking, Vol. 20, No 4, pp. 221-274.

Andolfatto, D. (2021), “Assessing the impact of central bank digital currency on private banks”, The Economic Journal, Vol. 131, No 634, pp. 525-540.

Berg, T., Keil, J., Martini, F. and Puri, M. (2024), “CBDCs, payment firms, and geopolitics”, NBER Working Papers, No 32857.

Bidder, R., Jackson, T. and Rottner, M. (2024), “CBDC and banks: Disintermediating fast and slow”, Deutsche Bundesbank Discussion Paper, No 15/2024.

Christelis, D., Georgarakos, D., Jappelli, T. and Kenny, G. (2025), “Wealth shocks and portfolio choice”, Journal of Monetary Economics, Vol. 149, No 103632.

Cong, L.W. and Mayer, S. (2025), “Strategic digitization in currency and payment competition”, Journal of Financial Economics, Vol. 168, No 104055.

Coibion, O., Georgarakos, D., Gorodnichenko, Y., Kenny, G. and Weber, M. (2024), “The Effect of Macroeconomic Uncertainty on Household Spending”, American Economic Review, Vol. 114, No 3, pp. 645‑677.

Ehrmann, M., Georgarakos, D. and Kenny, G. (2025), “Credibility gains from central bank communication with the public”, European Economic Review, No 105069.

ECB (2021), “ECB Consumer Expectations Survey: an overview and first evaluation”, Occasional Paper Series, No 287, ECB, K. Bańkowska, Borlescu, A.M, Charalambakis, E., Dias Da Silva, A., Di Laurea, D., Dossche, M., Georgarakos, D., Honkkila, J., Kennedy, N., Kenny, G., Kolndrekaj, A., Meyer, J., Rusinova, D., Teppa, F. and Törmälehto, V.-M.

Georgarakos, D. and Kenny, G. (2022), “Household spending and fiscal support during the COVID‑19 pandemic: Insights from a new consumer survey”, Journal of Monetary Economics, Vol. 129, pp. 1-14.

Georgarakos, D., Kenny, G., Laeven, L., and Meyer, J. (2025), “Consumer attitudes toward a central bank digital currency”, Working Paper Series, No 3035., ECB

Keister, T. and Sanches, D. (2023), “Should central banks issue digital currency?”, The Review of Economic Studies, Vol. 90, No 1, pp. 404-431.

Shapiro, M.D. and Slemrod, J. (2003), “Consumer Response to Tax Rebates”, American Economic Review, Vol. 93, No 1, pp. 381-396.

Schilling, L., Fernández-Villaverde, J., and Uhlig, H. (2024), “Central bank digital currency: When price and bank stability collide”, Journal of Monetary Economics, Vol. 145, No 103554.

Stantcheva, S. (2023), “How to Run Surveys: A Guide to Creating Your Own Identifying Variation and Revealing the Invisible”, Annual Review of Economics, Vol. 15, pp. 205-234.

Whited, T.M., Wu, Y. and Xiao, K. (2023), “Will Central Bank Digital Currency Disintermediate Banks?”, IHS Working Paper Series, No 47, Institute for Advanced Studies.