2025 Joint ECB-Eurostat annual report on the consistency of National Accounts and Balance of Payments concerning quarterly financial and non-financial accounts

The methodological standards for compiling European Balance of Payments statistics (BOP) and National Accounts (NA)[1] were harmonised with the implementation of the 6th edition of the Balance of Payments and International Investment Position Manual (BPM6) and the European System of Accounts 2010 (ESA2010). Following BPM6 paragraph 1.6 and ESA2010 paragraph 18.05 the two manuals aim for conceptual consistency regarding the recording of transactions between an economy with the rest of the world.[2]

However, in practice there are still observed differences in the published figures, for reasons such as, no fully harmonised revision policies between the institutions responsible for the compilation of BOP and NA data, use of different data sources, different compilation methods, and non-identical interpretations of the two manuals.

Eurostat and the ECB started measuring the degree of actual consistency between BOP and NA data for the first time in September 2015, for reference year 2014, when Member States were legally obliged to compile data according to the standards of the BPM6 and the ESA2010.[3] Since then, this comparison has been carried out on a regular basis. Since January 2022 a single annual report covers both the financial accounts and the non-financial accounts, using a homogeneous set of indicators and referring to the same time span.[4]

As already highlighted in previous reports, substantial differences remain for a number of countries resulting in considerable inconsistencies at the EU level. This picture holds despite the implementation of the benchmark revisions in September 2024 by the vast majority of EU countries[5] which should have solved such discrepancies. Therefore, it is relevant to continue the regular analysis of the consistency between both domains as well as to strongly encourage the countries, especially those contributing the most, to work on the remaining inconsistencies – including for longest possible periods of back data – so that they are – as soon as possible – eliminated or at least significantly reduced.

With the benchmark revisions several countries achieved significant consistency measured in relative terms across the two domains, namely the following ones:

• For the current account, five countries (DK, ES, FI, LV and NL) currently display fully harmonised BOP and RoW figures where discrepancies are below 0.05% of the GDP and another nine Member States (AT, CY, DE, HR, HU, IT, LT, PT and SI) achieved a good harmonisation where discrepancies are between 0.05% and 0.5% of the GDP;

• On the financial account transactions side, 15 countries observe a consistent picture (i.e. discrepancies below 1% of the GDP): AT, CY, EE, ES, HR, HU, IT, LT, LV, NL, PL, PT, SE (only liabilities), SI (only liabilities) and SK; and,

• On the financial account positions side, 17 countries observe a consistent picture (i.e. discrepancies below 2% of the GDP): AT (only liabilities), BG (only assets), CY (only liabilities), CZ (only assets), DK (only liabilities), EE, ES, FI (only liabilities), HR, HU (only liabilities), IT, LT, LV, NL, PL, PT, SE (only liabilities), SI (only liabilities) and SK.

Nonetheless substantial absolute differences still remain for the following countries for selected items of the non-financial and financial accounts for which they are the biggest contributors to the current inconsistencies:

• On the current and capital account side, FR (goods, services and primary income), DE (secondary income and capital account), IE (services), LU (all items), MT (services and primary income) and SK (capital account);

•  On the financial account transactions side, BE (“other” residual), LU (equity and “other”), DE (equity and “other”), FR (“other) and IE (equity and “other”); and,

•  On the financial account positions side, DE (equity and “other”), FR (equity and “other”) and IE (“other”).

This document analyses the current differences between quarterly national BOP and NA financial and non-financial accounts figures and their evolution as compared to those observed one year ago before any benchmark revision.[1] Section II elaborates on the differences currently observed at the EU and national level regarding non-financial accounts, while Section III analyses those referring to the financial accounts. The detected differences between the statistical discrepancy in BOP (net errors and omissions, NEO) and the Rest of the World (vertical discrepancy, VD) in NA are studied in Section IV. Finally, Section V concludes.

[1] The term NA in the context of this note means national (sectoral) quarterly financial and non-financial accounts.

[2] Some methodological differences between the BPM6 and ESA2010 were detected and are discussed in the joint ECB-Eurostat methodological advice on consistency between BOP and ROW (NA). Annex I includes a list of the topics covered. The consistency between BOP and NA is one of the overall priorities in the ongoing update of the BPM and SNA/ESA Manuals.

[3] For NA data the first reporting under ESA2010 standards took place in 2014, with annual data for reference year 2013, and quarterly financial accounts to the ECB for reference quarter 2013Q4, while for BOP data transmission to the ECB (Eurostat) according to the BPM6 became mandatory in 2014 (2015) for reference period 2013Q1 (2014).

[4] The ECB and Eurostat share the outcomes of this report with their respective working groups (WG ES, WG FGS, BOPWG and NAWG) and also assess inconsistencies between BOP and NA in the context of MIP visits.

[5] The only exceptions concern Romania and Luxembourg, who plan to implement them in 2025 and 2027, respectively.

[6] The data vintages used in this analysis are those up to 2023Q2 (reference period of last year) and for the most recent vintage up to 2024Q2. For each vintage, the last twelve quarters were considered.

Table 1 indicates how average quarterly absolute inconsistencies between BOP and RoW on EU level for the current and capital account have changed for the vintages analysed. Overall inconsistencies decreased for the total current account and for transactions related to primary income but increased for all other sub items of the current account as well as for the capital account.  

For the credit side, total current account differences decreased significantly in the most recent vintage by EUR 13.2bn to EUR 27.5bn compared to data for the previous vintage. For the debit side of the current account, differences shrank by EUR 3.4bn to EUR 21.5bn.

The following developments are observed for the main non-financial account items:

• Primary income is the largest contributor to the absolute differences on the credit side and the third largest for the debit side regarding data based on the 2024Q2 vintage. However, due to a considerable decrease in inconsistencies in two countries (FR and MT), these differences decreased significantly when compared with the 2023Q2 data transmission by EUR 10.4bn for the credit side and EUR 5.5bn for the debit side;

• Services exhibit the second largest inconsistencies between BOP and RoW for data based on 2024Q2 vintage for the credit side and the highest for the debit side. Average differences grew considerably compared to the 2023Q2 vintage for the credit side (by EUR 4.8bn), and the debit side (by EUR 5.8bn);

• Average differences concerning transactions for goods increased on both sides (by EUR 2.6bn for credits and EUR 3.6bn for debits). They represent the third highest absolute difference the credit side and the second largest for the debit side of the current account for the 2024Q2 vintage;

• Secondary income remains the current account item that contributes the least to the differences between BOP and RoW. However, differences grew also for this sub-account for the credit and debit sides (by EUR 1.0bn and EUR 1.6bn, respectively); and,

• Finally, average differences for the capital account balance[1] worsened as well by EUR 1.0bn to EUR 6.2bn for the 2024Q2 vintage.

Chart 1.1 displays average inconsistencies for the credit side and Chart 1.2 for the debit side of the current account for each Member State in % of GDP[1], while Annexes 2 and 3 provide the details of these differences in absolute terms and broken down by item. The following developments are observed:

• In relative terms and by comparing BOP and RoW data based on the 2024Q2 vintage with the 2023Q2 vintage, the relative average differences in the current account improved for eleven Member States on the credit side and also for eleven on the debit side. The relative differences worsened for seven Member States on the credit side and for eight on the debit side (see Charts 1.1 and 1.2).

• Based on the 2024Q2 vintage, five Member States (DK, ES, FI, LV and NL) exhibit fully harmonised[2] BOP and RoW figures for the current account, compared to seven Member States (CY, ES, FI, IE, LT, LV and NL) in the previous year. In addition, another nine Member States (AT, CY, DE, HR, HU, IT, LT, PT and SI) achieved a good harmonisation[3] of their BOP and RoW values for the current account concerning the most recent vintage in comparison to just seven (AT, DE, DK, HU, IT, PT and SE) based on the 2023Q2 vintage.    

• For the 2024Q2 vintage, MT is still showing one of the highest relative average inconsistencies for the credit and debit sides of the current account. However, a big improvement in reducing these differences compared to the 2023Q2 vintage has been achieved.    

• In absolute terms, the largest BOP-RoW differences are displayed by FR in particular for primary income, goods and services, followed by IE for primary income and DE, that exhibits the highest observed differences for secondary income and the capital account. These three countries represent almost 78% of the EU total current and capital account differences in the 2024Q2 vintage.   

[1] Note that GDP level data was also revised in the benchmark revisions.

[2] BOP and RoW figures are considered to be fully harmonised if the relative difference is below 0.05% of GDP.

[3] A good harmonisation of BOP RoW figures is achieved if the relative difference is between 0.05% and 0.5% of GDP.

Relative average differences (BOP - RoW) of the sum of debits. 2020Q3-2023Q2 vis-à-vis 2021Q3-2024Q2. Percentages of the GDP

Note: * For 2020Q3-2023Q2 RoW data was not fully available for CZ, GR and RO and for 2021Q3-2024Q2 not for CZ, EE, GR, PL and RO.

Chart 1.1 displays average inconsistencies for the credit side and Chart 1.2 for the debit side of the current account for each Member State in % of GDP[1], while Annexes 2 and 3 provide the details of these differences in absolute terms and broken down by item. The following developments are observed:

• In relative terms and by comparing BOP and RoW data based on the 2024Q2 vintage with the 2023Q2 vintage, the relative average differences in the current account improved for eleven Member States on the credit side and also for eleven on the debit side. The relative differences worsened for seven Member States on the credit side and for eight on the debit side (see Charts 1.1 and 1.2).

• Based on the 2024Q2 vintage, five Member States (DK, ES, FI, LV and NL) exhibit fully harmonised[2] BOP and RoW figures for the current account, compared to seven Member States (CY, ES, FI, IE, LT, LV and NL) in the previous year. In addition, another nine Member States (AT, CY, DE, HR, HU, IT, LT, PT and SI) achieved a good harmonisation[3] of their BOP and RoW values for the current account concerning the most recent vintage in comparison to just seven (AT, DE, DK, HU, IT, PT and SE) based on the 2023Q2 vintage.    

• For the 2024Q2 vintage, MT is still showing one of the highest relative average inconsistencies for the credit and debit sides of the current account. However, a big improvement in reducing these differences compared to the 2023Q2 vintage has been achieved.    

• In absolute terms, the largest BOP-RoW differences are displayed by FR in particular for primary income, goods and services, followed by IE for primary income and DE, that exhibits the highest observed differences for secondary income and the capital account. These three countries represent almost 78% of the EU total current and capital account differences in the 2024Q2 vintage.   

[1] Note that GDP level data was also revised in the benchmark revisions.

[2] BOP and RoW figures are considered to be fully harmonised if the relative difference is below 0.05% of GDP.

[3] A good harmonisation of BOP RoW figures is achieved if the relative difference is between 0.05% and 0.5% of GDP.

Sum across countries of the average quarterly absolute differences (BOP - RoW) of transactions and positions. EUR billion

Notes: a. There are certain constraints on the BOP data availability for some countries regarding foreign direct investment (FDI) debt securities and loans data affecting both transactions and positions. If referring to FDI debt securities it is assumed that missing data are zero. The missing data and partial coverage for FDI loans is considered for the compilation of the average quarterly absolute differences either excluding the country from the analysis (missing data) or adjusting the time span for which data are available (partial coverage). This limitation in the 2020Q3-2023Q2 vintage refers to BG (missing data), DK (partial coverage for transactions) and MT (partial coverage) while in the 2021Q3-2024Q2 vintage applies only to BG (missing data). b. Financial derivatives are reported in net terms for transactions assets. c. As the indicators are defined using absolute values, the sum of the subcomponents might be different than the total assets (liabilities) transactions (positions). This applies to the "Other" aggregate item that is compiled residually. d. The BOP positions assets exclude gold bullion in order to allow for a consistent comparison with RoW position assets.

The inconsistencies for the financial accounts at EU level are summarised in Table 2, using the same methodology as for the non-financial accounts. The analysis provides the following main results:

• Overall inconsistencies for positions have notably decreased (by around 22% and 38% for asset and liabilities, respectively) in the 2024Q2 vintage compared to the 2023Q2 vintage, though remaining at a substantial level. Inconsistencies in transactions have remained rather stable.

• For transactions, a small worsening is observed on the asset side (EUR 6bn) and mostly driven by the “other” (mainly debt-related residual) category (EUR 24bn), while on the liability side inconsistencies decreased by EUR 9bn, due mostly to equity (EUR 8bn). In the “other” category, debt securities recorded reduced inconsistencies both for assets and liabilities.[1]

• For positions, an improvement is observed both for assets (EUR 485bn) and liabilities (EUR 601bn) which is in both cases mostly driven by financial derivatives (EUR 531bn and EUR 566bn, respectively) and by equity for liabilities (EUR 122bn). As for transactions, the “other” category shows declining inconsistencies for debt securities assets and liabilities.

The still sizeable EU BOP-RoW discrepancies for both transactions and positions despite the implementation of the benchmark revisions points to the need for a detailed analysis at the country level to gain a better understanding of the drivers of these developments.

[1] The “other”/residual item also includes other items (such as deposits and trade credits) which cannot be singled out in the analysis.

Relative average differences (BOP - RoW) of the sum of assets transactions. 2020Q3-2023Q2 vis-à-vis 2021Q3-2024Q2. Percentages of the GDP

Charts 2.1 (assets) and 2.2 (liabilities) summarise the analysis of the overall BOP-RoW differences at the country level for transactions in relative average terms (as % of GDP). This analysis is supplemented with more detailed information broken down by item in absolute average terms in Annex 4 (2024Q2-2023Q2 vintage changes), Annex 5 (2024Q2 vintage) and Annex 6 (2023Q2 vintage). The main observations are the following ones:

• In relative terms (see Charts 2.1 and 2.2), LU, MT, DK and IE show the largest inconsistencies for both assets and liabilities transactions in the 2024Q2 vintage. For LU and DK inconsistencies increased sharply in the latest vintage, while MT and IE recorded the strongest improvements. For 12 countries the consistency remained at a good level on the assets side (14 countries on the liabilities side), with average relative differences not exceeding 1% of GDP (see Charts 2.1 and 2.2). 

• In absolute terms (see Annexes 4 to 6), DE, IE, DK and FR record the largest outstanding BOP-RoW differences for both assets and liabilities transactions (accounting for 70% and 75% of total EU inconsistencies, respectively).

• The main contributors to the observed worsening in absolute terms for assets was DK and, to a lesser extent, LU. In both cases this reflected a worsening for “other” as well as financial derivatives in the case of DK.[1] The largest improvement was observed for IE, mostly due to equity.

• On the liability side, the main contributor to the improvement in absolute terms was IE, mostly linked to the “other” category. The largest worsening was observed for DK and DE, triggered in both cases by the “other” category.

[1] The differences observed for Denmark mostly reflect incorrect reporting of financial derivatives assets and liabilities before 2024Q1 reference period. This issue will be solved by the submission of yearly revisions in 2025.

Relative average differences (BOP - RoW) of the sum of liabilities transactions. 2020Q3-2023Q2 vis-à-vis 2021Q3-2024Q2.  Percentages of the GDP

Charts 2.1 (assets) and 2.2 (liabilities) summarise the analysis of the overall BOP-RoW differences at the country level for transactions in relative average terms (as % of GDP). This analysis is supplemented with more detailed information broken down by item in absolute average terms in Annex 4 (2024Q2-2023Q2 vintage changes), Annex 5 (2024Q2 vintage) and Annex 6 (2023Q2 vintage). The main observations are the following ones:

• In relative terms (see Charts 2.1 and 2.2), LU, MT, DK and IE show the largest inconsistencies for both assets and liabilities transactions in the 2024Q2 vintage. For LU and DK inconsistencies increased sharply in the latest vintage, while MT and IE recorded the strongest improvements. For 12 countries the consistency remained at a good level on the assets side (14 countries on the liabilities side), with average relative differences not exceeding 1% of GDP (see Charts 2.1 and 2.2). 

• In absolute terms (see Annexes 4 to 6), DE, IE, DK and FR record the largest outstanding BOP-RoW differences for both assets and liabilities transactions (accounting for 70% and 75% of total EU inconsistencies, respectively).

• The main contributors to the observed worsening in absolute terms for assets was DK and, to a lesser extent, LU. In both cases this reflected a worsening for “other” as well as financial derivatives in the case of DK.[2] The largest improvement was observed for IE, mostly due to equity.

• On the liability side, the main contributor to the improvement in absolute terms was IE, mostly linked to the “other” category. The largest worsening was observed for DK and DE, triggered in both cases by the “other” category.

[1] The differences observed for Denmark mostly reflect incorrect reporting of financial derivatives assets and liabilities before 2024Q1 reference period. This issue will be solved by the submission of yearly revisions in 2025.

Relative average differences (BOP - RoW) of the sum of assets positions. 2020Q3-2023Q2 vis-à-vis 2021Q3-2024Q2. Percentages of the GDP

 Source: ECB.

Charts 2.3 (assets) and 2.4 (liabilities) summarise the analysis of the overall BOP-RoW differences at the country level for positions in relative average terms (as % of the GDP). This analysis is supplemented with more detailed information broken down by item in absolute average terms in Annex 4 (2024Q2-2023Q2 vintage changes), Annex 5 (2024Q2 vintage) and Annex 6 (2023Q2 vintage). The main observations are:

• In relative terms (see Charts 2.3 and 2.4), IE, FR, LU and GR show the largest inconsistencies for asset and liability positions in the 2024Q2 vintage. For FR inconsistencies decreased sharply in the latest vintage, while IE recorded the largest deterioration observed. For 12 countries the consistency remained at a good level on the asset side (17 countries on the liabilities side), with average relative differences not exceeding 2% of GDP (see Charts 2.3 and 2.4). 

• In absolute terms (see Annexes 4 to 6), FR,[2] DE and IE recorded the largest BOP-RoW differences for both assets and liabilities positions (accounting for 91% and 90% of total EU inconsistencies, respectively).

• The main contributors to the observed improvement at the EU level in absolute terms for assets were FR and, to a lesser extent, IT and MT. For FR financial derivatives contributed the most while for IT and MT this reflected mainly a better consistency for “other”. The largest deteriorations were observed for DE and IE (due to the “other” category).

• On the liability side, the main contributors to the improvement in absolute terms were FR and, to a lesser extent, MT. For FR this primarily reflected a better picture for financial derivatives, while for MT this was due mostly to equity. The largest deterioration was observed for IE (driven by “other”).

[1] The BOP positions assets exclude gold bullion in line with both BPM6 and 2008 SNA and to allow for a consistent comparison with RoW positions. In the case of Austria and Hungary, the difference is exclusively due to the different recording convention for gold bullion. In the case of Italy, the 2020Q3-2023Q2 vintage includes gold bullion, which explains a large part of the differences observed. 

[2] The discrepancies regarding equity instruments reflect different valuation practices for unlisted equity: own funds at book value in BOP and market capitalization in RoW. 

Relative average differences (BOP - RoW) of the sum of liabilities positions. 2020Q3-2023Q2 vis-à-vis 2021Q3-2024Q2. Percentages of the GDP

Source: ECB.

Charts 2.3 (assets) and 2.4 (liabilities) summarise the analysis of the overall BOP-RoW differences at the country level for positions in relative average terms (as % of the GDP). This analysis is supplemented with more detailed information broken down by item in absolute average terms in Annex 4 (2024Q2-2023Q2 vintage changes), Annex 5 (2024Q2 vintage) and Annex 6 (2023Q2 vintage). The main observations are:

• In relative terms (see Charts 2.3 and 2.4), IE, FR, LU and GR show the largest inconsistencies for asset and liability positions in the 2024Q2 vintage. For FR inconsistencies decreased sharply in the latest vintage, while IE recorded the largest deterioration observed. For 12 countries the consistency remained at a good level on the asset side (17 countries on the liabilities side), with average relative differences not exceeding 2% of GDP (see Charts 2.3 and 2.4). 

• In absolute terms (see Annexes 4 to 6), FR,[1] DE and IE recorded the largest BOP-RoW differences for both assets and liabilities positions (accounting for 91% and 90% of total EU inconsistencies, respectively).

• The main contributors to the observed improvement at the EU level in absolute terms for assets were FR and, to a lesser extent, IT and MT. For FR financial derivatives contributed the most while for IT and MT this reflected mainly a better consistency for “other”. The largest deteriorations were observed for DE and IE (due to the “other” category).

• On the liability side, the main contributors to the improvement in absolute terms were FR and, to a lesser extent, MT. For FR this primarily reflected a better picture for financial derivatives, while for MT this was due mostly to equity. The largest deterioration was observed for IE (driven by “other”).

[1] The discrepancies regarding equity instruments reflect different valuation practices for unlisted equity: own funds at book value in BOP and market capitalization in RoW. 

Sum across countries of the average quarterly NEOs and RoW vertical differences and the differences (BOP-RoW) between these vertical differences. EUR billion

Source: ECB and Eurostat.

Notes: For 2020Q3-2023Q2 RoW data was not fully available for CZ, GR, and RO and for 2021Q3-2024Q2 not for CZ, GR, PL and RO; BOP data was used instead to build EU 27 aggregates.

1) Capital and current account net transactions minus RoW net lending/net borrowing (B.9).

2) BOP net financial transactions minus RoW net lending/net borrowing (B.9F).

Table 3 summarises how average quarterly vertical inconsistencies changed for the vintages analysed, not being computed in absolute terms to show the direction of the changes observed. The following developments are observed:

• The BOP-RoW difference between the discrepancy measures decreased to EUR -5.7bn from EUR 21.8bn, mainly due to improvements in the RoW VD and the BOP-RoW differences in financial transactions (net);

• The NEO in BOP worsened slightly to EUR -9.4bn from EUR -3.6 bn;

• The VD in the RoW accounts improved significantly to EUR -3.7bn from EUR -25.4bn;

• The BOP-RoW difference in non-financial transactions (net) improved to EUR 6.2bn from EUR 12.8 bn;

• The BOP-RoW difference in financial transactions (net) improved significantly to EUR 0.8bn from EUR 34.6bn.

The average vertical consistency measures for the EU aggregates are below 0.1% of GDP, however developments on the country level are more pronounced. The vertical discrepancy measures are presented on a four-quarter sum basis in line with the European guidance on vertical consistency.[1] The four-quarter sum basis allows the offsetting of short-term discrepancies and facilitates the comparison with annual financial accounts.

[1] Compilers are recommended to keep in each compilation round the four-quarter sum/annual vertical discrepancy for each sector (and sub-sector) below 1% of the four-quarter sum/annual GDP (‘target’). See: “Report on developing a common approach to improve vertical consistency” published by ECB and Eurostat. Several Member States had planned to implement elements of the recommendations at the time of the benchmark revision.

Relative average absolute differences (BOP – RoW) between NEO and RoW VDs (four-quarter sums) 2019Q3-2022Q2 vis-à-vis 2020Q3-2023Q2. Percentages of GDP (four-quarter sums).

Source: ECB and Eurostat.

Note: For 2020Q3-2023Q2 RoW data was not fully available for CZ, GR, and RO and for 2021Q3-2024Q2 not for CZ, EE, GR, PL and RO.

Chart 3.1 presents the country developments of the vertical consistency measure differences relative to country GDP. The main observations are:

• The differences increased between the two reference periods in nine countries. Decreases are observable in eight countries, with the largest improvements in MT and CY.

• While in a few countries NEOs and RoW VDs are identical (LV, NL) or nearly identical (AT, ES, HR, HU, LT, LU), for most of the countries they fluctuated between 0.3% and 2% of GDP.

• IE and MT exhibit differences which are larger than 8%; with IE registering the highest increase. 

The increase in the vertical consistency differences may be due to changes in the NEOs and changes in RoW VDs. The country developments are presented below.

Relative absolute NEOs (four-quarter sums). 2020Q3-2023Q2 vis-à-vis 2021Q3-2024Q2. Percentages of GDP (four-quarter sums). 

Source: ECB and Eurostat.

Chart 3.2 presents the country developments of the NEOs relative to country GDP. The main observations are:

• NEOs are below 1% of GDP in 11 countries (BE, CZ, DE, ES, FR, GR, IT, LU, NL, PL, and PT).

• Of the countries with NEOs still above this goal for vertical discrepancies, large improvements are observed for MT (to 1.5% from 7.9%), BG (to 2.5% from 4.1%) and FI (to 1.3% from 2.3%) and smaller improvements for AT, DK, HU, and SE; while

• NEOs worsened or were unchanged for CY, IE, SK, RO, LV, SI, and LT.

Relative absolute vertical differences (four-quarter sums). 2020Q3-2023Q2 vis-à-vis 2021Q3-2024Q2. Percentages of GDP(four-quarter sums)

Source: ECB and Eurostat

Note: For 2020Q3-2023Q2 RoW data was not fully available for CZ, GR, and RO and for 2021Q3-2024Q2 not for CZ, EE, GR, PL and RO 

Chart 3.3 presents the country developments of the RoW VDs relative to country GDP. The main observations are:

• RoW VDs are below 1% of GDP in eight countries (BE, ES, FR, HR, IT, LU, NL, and PT).

• Of the countries with RoW VDs still above this goal for vertical discrepancies, larger improvements are observed for MT (to 10.2% from 12.0%), BG (to 2.0% from 3.2%), SE (to 2.8% from 3.8%) and CY (to 2.2% from 3.0%),  and smaller improvements for  AT, DE, LT, HU, and SE; while

• RoW VDs worsened or were unchanged for IE, DK, SK, LV and FI.

• The improvement observed for MT in the RoW VD is significantly smaller than the one observed in the BOP NEO. This explains the still high RoW to BOP difference of the vertical discrepancies.  The worsening for IE in the RoW VD explains the significant worsening of the RoW to BOP comparison (see chart 3.1).

Several countries took advantage of the benchmark revision 2024 to implement revisions of the financial and non-financial accounts as well as of the BOP statistics which improved the consistency of the two. However, this document also clearly illustrates that in the 2024Q2 vintage substantial discrepancies in the EU for the non-financial and financial accounts still exist between BOP and RoW data. At the level of EU Member States, relative inconsistencies even increased compared to the 2023Q2 vintage in the non-financial accounts for seven countries on the credit side and for eight countries on the debit side, while in the financial accounts inconsistencies – both for transactions and positions – widened for 13 countries in assets and for 11 countries in liabilities. The NEO in BOP increased in 15 countries and the RoW vertical differences increased in nine countries. There have been, however, large country-specific changes in both directions with some countries showing improved consistency – in particular MT and FR for the non-financial accounts and FR for the financial accounts - but insufficient to contribute to an overall improved and consistent BOP-RoW picture at the EU level.   

The most significant inconsistencies in EU countries are as follows:

• In the non-financial accounts, and despite significant improvements, MT displays very large differences in relative terms for the credit and debit sides followed by LU. For IE differences between BOP and RoW measured in absolute as well as relative terms worsened the most. Although some improvements were made, FR still exhibits the highest differences in absolute terms. In relative terms, DE has achieved a good harmonisation of BOP and RoW data for the current account but displays considerable absolute differences for some current account sub items and for the capital account balance.

• In financial accounts transactions, LU, MT, DK and IE record significant inconsistencies in relative terms, while DE, IE, DK and FR contribute the most to the inconsistencies in absolute terms. Regarding positions, IE, FR and LU show the largest inconsistencies in relative terms, while FR, DE and IE contribute the most in absolute terms.

• The NEO in BOP is particularly large in relative terms for SE, BG and LV; for CY, IE, SK, RO,  LV, SI, and LT it increased and is above 1% GDP; FR, SE and DE contribute the most in absolute terms. The RoW VD is large in IE where it increased, followed by MT and SE where it decreased; for IE, DK, SK, LV and FI it increased and is above 1% GDP; while DE, IE and SE contribute the most in absolute terms.

The ECB and Eurostat will continue to provide methodological guidance for relevant topics[1] and the respective groups regularly discuss issues of common relevance.[2][3] Countries are reminded to implement the recommendation of the “Report on developing a common approach to improve vertical consistency” in the sector accounts as approved by the STC and DMES and published in March 2022. While the recommendations remain voluntary, national compilers are encouraged to take them into account in their compilation systems and in the context of the cooperation between financial and non-financial sector accountants. A summary table presenting the current reconciliation practices covering all EU countries and the euro area was published by the ECB and Eurostat in October 2023.[4]

In the context of the Macroeconomics Imbalance Procedure (MIP), quality assurance framework issues related to the consistency of the two statistics are addressed as needed. For this purpose, the visits include specific action points aimed to address the NA-BOP differences and vertical consistency.[5]

The harmonised benchmark revision in 2024 has unfortunately not improved the overall picture of the BOP-RoW consistency as initially expected. This suggests that certain structural issues remain unsolved, being especially relevant for those countries contributing the most to the differences in absolute terms. Therefore, institutions in charge of compiling NA and BOP statistics where differences remain high are strongly encouraged to establish dedicated, regular coordination touchpoints to progress towards the goal of consistency between the two domains as soon as possible, thereby ensuring the serviceability of these statistics.

[1] See the document entitled “Consistency between national b.o.p./i.i.p. and financial accounts: ECB/Eurostat methodological advice 2020 update” circulated in November 2020.

[2] Eurostat BOPWG, NAWG, ECB working groups FGS and ES, the EG AFA and the Eurostat Expert Group on Sector Accounts (EG SA).

[3] Moreover, the ECB and Eurostat are regularly monitoring national vertical discrepancies on an annual level and issues relevant recommendation in the context of the national accounts quality reports (see ECB and Eurostat).

[4] See ECB and Eurostat.

[5] The countries visited so far in the MIP quality assurance framework are the following: BE, CY, DE, FI, FR, GR, HR, IE, LU, NL, MT and PL (a visit of BG is scheduled for December 2024). While the visits focus on the financial accounts and balance of payments data underlying the MIP indicators, the consistency between the financial account in the balance of payment and the financial account in the national accounts is also discussed and the recommendation on the best use of sources and methods often address both domains.

• Treatment of off-market swaps

• Net versus gross recording of financial derivatives

• The treatment of closed-end/unincorporated investment fund units

• Compilation of FDI income (reinvested earnings) 

• Treatment of listing/de-listing of equity shares

• Classification of deposits/loan positions of deposit-taking corporations

• Distinction between debt-securities and loans

• Classification of trade credits accepted by factoring companies

• Instrument treatment of cash pledged as collateral

• Recording of timing differences in securities trading between change in ownership and payment date

• Recording of subsidies

• Investment fund shares

• Other accounts receivable/payable

• Euro banknotes and euro coins

• Social benefits and contributions

• Vertical consistency in sector accounts