Euro area government debt amounted to 90% of gross domestic product (GDP) at last count. The debt ratio has thus declined since 2020. However, it is still 4 percentage points higher than in the pre-crisis year of 2019. This article describes how the creditor structure of government debt has evolved in recent years. The focus is on the euro area as a whole and on the four large Member States – Germany, France, Italy and Spain.

The creditor structure of government debt has shifted strongly since 2015 as a result of the extensive asset purchase programmes conducted by the Eurosystem central banks. This reduced, in particular, the share of assets held by euro area commercial banks and by creditors outside the euro area. In response to high inflation, the euro area national central banks then tightened the monetary policy reins and are now shrinking their holdings of government bonds. Nevertheless, as at end-2023 they were still the largest creditor, accounting for 27%. The weight of creditors outside the euro area and especially of households has increased recently. 

In the four large Member States, the creditor structure has evolved very differently since the Eurosystem reduced the envelope of its asset purchase programmes. Foreign creditors, in particular, gained in importance in Germany and France. By contrast, the share of households in Italy and Spain especially rose. The interest rate reversal is weighing on the national central banks’ balance sheets, as, while deposit rates on bank deposits are increasing, interest income from the longer-dated bonds purchased has initially remained unchanged. Most recently, only Banca d’Italia transferred a relatively small amount to its central government.

National banks continue to be a significant creditor group. The Spanish banking system, and in particular the Italian banking system, are most highly exposed to their home country. The fact that capital does not have to be held against many government bonds makes it attractive for banks to hold them. In addition, large exposure limits do not apply; as a result, in some cases very large stocks of a sovereign’s bonds are being held. Should doubts regarding the sustainability of government finances arise, this can rapidly engender additional risks to financial stability. In order to mitigate these risks, it would make sense to curtail regulatory incentives by reducing existing regulatory privileges. That would also facilitate negotiations to complete the banking union by establishing a European deposit insurance scheme. To date, however, political majorities for deprivileging are nowhere in sight. If, even absent this, the European banking union is to be advanced with a single European deposit insurance scheme which envisages a redistribution of risks at the European level, it would need to be assured beforehand through some other method that proper account is taken of the various risks associated with government debt, depending on the desired design. 

Euro area countries’ debt amounted to 90% of GDP at the end of 2023. 1 At end-2020, the debt ratio hit a new all-time high of 99%, also in the light of the sharp slump in GDP (see Chart 4.1). Since then, though, absolute debt has risen further. However, the ratio has shrunk distinctly owing to the strong, particularly inflation-driven growth in nominal GDP. The debt ratio is still 4 percentage points higher than in the pre-crisis year of 2019. The European Commission expects the debt ratio to remain almost unchanged at this high level until 2025. 2

The debt ratios of the four largest Member States vary greatly in terms of level, but have evolved similarly. They have declined in recent years and are now above their level from before the outbreak of the COVID-19 pandemic (see Chart 4.1). As at the end of last year, Germany’s debt ratio stood at 64%, Spain’s at 108%, France’s at 111% and Italy’s at 138% of GDP.

In terms of maturity structure, medium to long-term liabilities are predominant in the euro area. 3 Their significance has increased slightly in recent years. Debt securities (especially bonds), which are generally marketable, make up the majority of debt instruments (see Chart 4.2). They have significantly greater weight than loans, which banks, in particular, issue to governments. State liabilities from currency and deposits on government accounts represent only a small share of government debt, at around 3%.

The debt ratios of the individual Member States vary widely. They range from 20% in Estonia to 163% in Greece (see Chart 4.3). Long-term debt securities predominate. Greece alone is the significant exception, as very large long-term assistance loans granted in connection with the 2010-2015 sovereign debt crisis are still outstanding. 4

The average residual maturity of government securities debt has been relatively high in recent years. However, owing to central banks’ extensive holdings, government finances are less well hedged against interest rate increases than would appear to be case just looking at residual maturities. Chart 4.4 shows the residual maturities for the euro area and the four large Member States. If average residual maturities are higher, interest payments are generally fixed for longer periods. 5 Interest rate changes then have only a very lagged impact on government finances. However, when taking a holistic view, the balance sheet link between central banks and government finances must be taken into account. As a result of the central bank’s holdings, the current short-term deposit rate, and no longer the bond rate, is effectively the relevant rate for this bond volume, irrespective of the residual maturity (see Chapter 2). The average interest rate fixation period adjusted for this is thus likely to be up to two years shorter than indicated by residual maturities.

Until the beginning of 2021, the structure 6 of euro area countries’ government debt holdings shifted significantly with the broad-based purchases of government bonds by euro area central banks. 7 At the beginning of 2021, the Eurosystem was the largest creditor, accounting for 25%. Although all other creditor sectors had seen their shares fall in the period beginning in 2015, the decline was particularly significant among foreign (non-euro area) creditors. Their share at the beginning of 2021 was a paltry 17% following as much as 26% in 2015. Looking at the domestic sectors, banks (monetary financial institutions excluding central banks) were the largest creditor group in 2021, accounting for 23%. They were followed by insurance corporations and pension funds (15%) as well as other financial institutions such as investment funds (7%). Looking at the euro area as a whole, households, international institutions and non-financial corporations played only a minor role.

From the beginning of 2021, euro area central banks responded to the rapid rise in the inflation rate. They tightened their monetary policy reins by initially purchasing smaller volumes of bonds and are now shrinking their holdings again. Since 2012, asset holdings under the securities market programme (SMP) have only been held to maturity, with reinvestments discontinued. Net purchases under the PSPP were discontinued in July 2022. From March 2023, holdings under the PSPP were then only partially reinvested, and no longer so since July 2023. Net purchases under the PEPP were discontinued with effect from April 2022. By contrast, principal payments from maturing securities purchased under the PEPP will be reinvested until the end of 2024. 8

Up until mid-2022, the share of government debt held by the Eurosystem nonetheless increased further. At the end of June 2022, the Eurosystem was by far the largest creditor group, accounting for 30%. In addition to debt held by foreign creditors, domestic banks have now recorded the most significant declines since the beginning of 2021 (see left-hand side of Chart 4.5). The shares of domestic insurance corporations and pension funds also declined slightly. The shares of other creditors remained virtually unchanged.

Since mid-2022, the share of government debt held by the Eurosystem has declined. At 27%, however, the Eurosystem was still by far the largest creditor group at the end of 2023. The shares of the other creditor groups assumed different trajectories over the same period (see right-hand side of Chart 4.5). Debt held by foreign creditors and, in particular, indebtedness to euro area households played a greater role. Euro area households increased their holdings of government bonds by more than two and a half times, yet, at 3%, they remain a relatively insignificant creditor group. Domestic banks reduced their share somewhat further.

In all four countries, the creditor structure exhibited a clear shift towards the Eurosystem central banks in the period up to mid-2022, owing to the extensive government bond purchases. There were differences in terms of the sectors from which borrowing was channelled away, though. Germany and France primarily saw a decline in indebtedness to non-resident creditors (hereinafter referred to as foreign creditors, including other euro area countries). To begin with, they still constituted the largest group of creditors in both countries. In addition, domestic banks also ceded some of their share, as did – in France – the comparatively weighty insurance corporations and pension funds. By contrast, Italian and Spanish government debt was mainly in the hands of domestic creditors from the outset. In Italy, the central bank’s purchases initially meant smaller shares for all other groups of creditors. In Spain, the proportion held by domestic banks, in particular, contracted (see Deutsche Bundesbank (2022)). Table 1 provides detailed information on the creditor structure for the period 2021-2023, for which consistent data are available. The table shows developments from 2021 to mid-2022 and from mid-2022 to the end of 2023. It shows holdings of the creditor groups both in relation to total debt and in relation to national GDP. The latter thus also reflects the different debt ratios of the countries. 9  

From the beginning of 2021, the further increase in the Eurosystem share in all four countries was accompanied by a decline in debt held by foreign creditors (see Chart 4.7). In Germany, France and Spain, however, domestic banks also ceded further shares, as did insurance and pension funds in France, Italy and Spain. Italy and to a somewhat lesser extent Spain increasingly made use of EU assistance loans.

In mid-2022, the share of government debt held by the central banks of the Eurosystem peaked in all countries. After that, it went down. The shares of the remaining creditor sectors followed divergent paths (see Chart 4.8). In Germany and France, foreign creditors were alone in regaining significant ground. In Italy and Spain, meanwhile, that group’s share changed only marginally. These two countries chiefly saw an increase in the weight of households: in Spain the share of households recorded a moderate rise, whilst in Italy the rise was strong and took the share to around 10%, making this the second largest domestic creditor group now. 10 This rendered the financial system less vulnerable. Italian banks have further scaled back their shares. That said, debt to domestic banks remained significantly higher in Italy and Spain, at 22% and 19% respectively, than in Germany and France (14% each). The Eurosystem’s share remained high at the end of 2023, despite waning. It held 29% of German, 23% of French, 27% of Italian and 30% of Spanish government debt.

Italy and Spain pay marked risk premia on their debt. These mainly flow to domestic creditors. The creditor structure can thus be used to determine the extent to which government interest expenditure remains within a country or goes beyond its borders. This also applies to the risk premia being paid on government debt. If interest payments are being made to domestic recipients, there will ultimately be no outflow of resources abroad. Interest payments remain as income in the national economic cycle (and can be taxed, for example, as investment income). In this sense, even comparatively high interest rates place less of a strain on macroeconomic developments than extensive borrowing from foreign creditors. 

The volume of securities held abroad (as measured by GDP) stands at 32% in Italy and 38% in Spain. Despite significantly higher levels of debt, these percentage shares are only moderately above the German figure of 27%, and are well below the figure for France (54%). In addition to the foreign creditors shown above, the numbers cited here also contain most of the government debt holdings at the Eurosystem central banks. This is because these holdings can be allocated, in each case, to the grouping of foreign creditors (including the rest of the euro area). From the national perspective, the majority of the ECB’s holdings are held by foreign creditors. This is due to the fact that they are subject to risk sharing and, therefore, the bulk of the interest paid on them goes to other countries. 11 By contrast, all interest payments on national central banks’ holdings remain within their respective home countries. 

Placed in relation to the respective national GDP, the holdings arising under the government bond purchase programmes, with interest flowing back to the Member State in question, amount to 21% in the case of Germany (Bundesbank holdings), 25% in the case of France and 29% each for Italy and Spain. 12 In addition, the Eurosystem central banks can (under the terms of the Agreement on Net Financial Assets (ANFA)) hold government bonds in portfolios for non-monetary policy purposes. The Bundesbank does not make use of this option. The relevant ANFA balance sheet items of the Banque de France total around 2% of national GDP, while those for the Banca d’Italia amount to 6% and those for the Banco de España 2%. 13 The share of government bonds issued by national central banks’ home countries within these items is not published, but is likely to be substantial in some cases. Table 2 provides an overview of Eurosystem securities holdings and risk sharing.

Sources: Eurostat, balance sheet data of the Banco de España, Banca d’Italia, Banque de France and Deutsche Bundesbank as well as Bundesbank calculations.

When it comes to the government bonds being held by their national central bank, Member States are effectively no longer paying the government bond interest rate and are instead paying the risk-free monetary policy rate on them. This is due to the decision that the vast majority of the large-scale bond purchases would not be subject to risk sharing. 14 The balance sheet link between central banks and government finances means that interest rates on bonds ultimately flow back to the Member States. The same goes for the risk premia that they include. As long as national central banks distribute the interest rate differential as profit to their respective home government (or deduct the interest rate burden from their profits), Member States pay the short-term deposit rate on the central bank’s holdings. 15 In this respect, the size of these holdings also renders public finances commensurately more sensitive to short-term changes in central bank interest rates. With regard to interest rate risk, this counterbalances the longer residual maturities described above. When central bank interest rates rise, central bank balance sheets feel the strain first. Public finances are then affected as central banks distribute lower profits or cease to distribute profits for a time.

The varying interest rates on holdings of government bonds issued by the home country are reflected in the national central banks’ balance sheets. The central banks of Germany, France and Spain did not record any profit in 2023 and, as such, did not make any tax payments or distribute any profit to their respective governments. In the case of Italy, by contrast, tax refunds were due. As a result, the central bank recorded a slight profit on its balance sheet and proceeded to make a small profit transfer. 16 All four central banks made heavy use of provisions and reserves and, in so doing, avoided reporting a loss on their balance sheet. However, there were differences when it came to how deeply the risk provisioning built up over previous years was tapped. Recourse was highest in Germany, at 0.5% of GDP, and lowest in Italy (including a tax refund), where it amounted to 0.3% of GDP. These differences reflect, not least, the differences in interest income from government bond holdings. 

In the euro area, the domestic banking system is the second-largest creditor sector to governments after central banks. The banking system is of particular importance in terms of financial stability. Problems with government finances and in the banking system can quickly become mutually reinforcing and pose a threat to financial stability. In view of this, changes in national banking sectors’ claims on their home countries are of particular interest.

In the euro area as a whole, the nexus between banks and their respective home countries has weakened moderately compared with the levels observed prior to the outbreak of the COVID-19 pandemic. At the end of 2023, banks still held claims on their respective Member State amounting to 14% of GDP; they stood at 16% at the end of 2019. Furthermore, banks still held claims on other euro area Member States and non-euro area countries (see Chart 4.10).

In relation to domestic banks’ own funds, claims on the home country also declined moderately. Compared with the end of 2019, this figure fell from 77% to 73% (see Chart 4.12). 17 Own funds are ultimately the buffer that banks can use to absorb any losses. Banking regulation stipulates that banks are required to maintain a corresponding capital buffer for risk-prone financial investment. Moreover, the potential volume of claims vis-à-vis individual creditors is limited (large exposure limits). Government bonds issued by EU Member States are given preferential regulatory treatment, though. Under European banking regulations, banks can apply a zero weight to these bonds in the capital backing requirements. In addition, government bonds then do not count against large exposure limits.

The claims of Spanish and especially of Italian banks on their home countries remain very high relative to GDP, even though they have decreased markedly in Italy compared to their pre-pandemic levels. Although the corresponding claims of Italian banks fell the most sharply (see Chart 4.11), they were still at the highest level in the euro area by far, at 29% of GDP. 18 Spain follows with a comparatively high value of 20%. 19 In France, the figure, almost unchanged at 15%, is just slightly above the euro area average. In Germany, it declined slightly to 9% of GDP by the end of 2023.

In Italy and Spain, banks’ claims on their respective home countries are particularly high, even relative to their own funds. 20 This ratio initially rose in all countries at the outbreak of the COVID-19 pandemic. In Germany and Italy, however, it had fallen again distinctly by the end of 2023 and is currently below the level recorded at the end of 2019. By contrast, it increased somewhat in France and even grew significantly in Spain (see Chart 4.12). At the end of 2023, Italy still recorded by far the highest level of all the euro area Member States, at 169%. In Spain, the total claims held by domestic banks on their home country now exceed own funds, at 112%. 21

Overall, it is clear that in some cases banks continue to be very closely interconnected with their home countries and exposed to the fiscal risks of those home countries. Euro area banks’ claims on their own Member States are mostly only moderately lower than before the outbreak of the COVID-19 pandemic. The Spanish banking system, and in particular the Italian banking system, are most highly exposed to their home country.

The fact that many government bonds do not need to be backed by capital makes them attractive to banks, but this is associated with risk. In addition, large exposure limits do not apply; as a result, in some cases very large stocks of a sovereign’s bonds are being held. Should doubts regarding the sustainability of domestic government finances arise, this can rapidly engender additional risks to financial stability. In order to mitigate these risks, it would make sense to curtail regulatory incentives by reducing existing privileges. That would also facilitate negotiations to complete the banking union by establishing a European deposit insurance scheme. To date, however, political majorities for deprivileging are nowhere in sight. If, even absent this, the European banking union is to be advanced with a single European deposit insurance scheme which envisages a redistribution of risks at the European level, it would need to be assured beforehand through some other method that proper account is taken of the various risks associated with government debt, depending on the desired design. 

This article draws on a range of data sources in order to paint as comprehensive a picture as possible. The stream of data has been consistent since spring 2021, while the data for the end of 2023 still contain estimates in some cases (see below). However, the data up until the end of 2020 are not always fully compatible and uniformly defined, resulting in ambiguities in some instances. 22 This applies, in particular, to the data prior to 2021 on the shares of Eurosystem central banks and non-resident creditors in the public debt of the Member States. This is due to the fact that the Eurosystem’s holdings are not available at face values or according to the statistical definition of the Maastricht debt level for the period prior to 2021. This affects the shares of non-resident creditors, which are calculated as a residual (see below).  

The data on the volume and structure of government debt stem from Eurostat’s Government Finance Statistics (GFS), indicating the debt of the general government sector (according to the European System of Accounts) at face values. 23 Debt levels for the last quarter of 2023 were not yet available for all Member States at the time this article was published. Where this is the case, estimates by the European Commission for this quarter are used. Hence, the resulting estimation errors are reflected in the creditor shares of non-residents calculated as residuals (see below). The effective residual maturities of government debt were estimated in two steps. In a first step, the PSPP and PEPP holdings were subtracted from government debt using the weighted residual maturities published by the ECB. Some deviations may occur as the holdings of these purchase programmes are defined differently from government debt in terms of their statistical characteristics (see below). In addition, they are measured at amortised cost rather than at their face value. In a second step, an effective residual maturity of zero was set for the holdings of the purchase programmes. This is because changes in the short-term deposit rate have a direct impact on a central bank’s income (and thus also on government finances via the balance sheet link; see Chapter 2  above). 

Data on the creditor structure of securities from the beginning of 2021 stem from the ECB’s Securities Holdings Statistics by Sector (SHSS). They comprise the bond holdings at face value of the aggregate sectors for the euro area as a whole and for individual Member States. The holdings of Eurosystem central banks taken from these statistics therefore only include bonds pertaining to the general government sector. In addition, the Eurosystem also purchased bonds from national development banks, other national public undertakings and supranational institutions. The latter are not usually included in national government debt. This means that the overall holdings of public bonds issued in Germany are considerably higher than the holdings reported here, which are attributable to Germany’s sovereign debt (as defined by the GFS). In particular, this concerns the bonds of KfW and other government development banks. Moreover, SHSS data do not distinguish whether Eurosystem central banks acquired the bond holdings under a purchase programme or for their own portfolios. 24  

These securities data are supplemented by data on the creditor structure of banks’ credit claims (nominal values) from the ECB’s balance sheet statistics. Added to this are data from the European Stability Mechanism (ESM) on the outstanding financial assistance provided under the European Financial Stability facility (EFSF) and the ESM (nominal values). Data from the European Commission on outstanding financial assistance from the European Financial Stabilisation Mechanism (EFSM), the balance of payments (BoP) assistance facility, the instrument for temporary support to mitigate unemployment risks in an emergency (SURE) and loans taken up under the NextGenerationEU(NGEU) European recovery plan were likewise included. No separate information on holders is available for credit claims vis-à-vis Member States not held by these institutions or for government liabilities from currency and deposits (2% and 3%, respectively, of total debt in the euro area). These categories are reported under “Other”.

The euro area creditor structure statistics used here do not explicitly record government bonds held abroad. This is because securities depositories from non-euro area countries are not subject to reporting requirements. 25 For this reason, the category of government debt held abroad was calculated as a residual figure in this article. This is the difference between the outstanding national debt securities according to the GFS and the holdings reported in the SHSS holder statistics. 26 The different statistical definitions of the data, especially in the period up to the end of 2020, affect this residual. There is one further aspect to bear in mind when interpreting and differentiating between the domestic country and foreign countries –  although the reported statistics distinguish between holdings of government bonds held domestically or abroad, this does not necessarily imply that the creditors themselves are domestic or foreign. Foreign creditors, too, can hold government bonds in domestic safe custody accounts, and vice versa.

Further restrictions relate to the national allocation of Eurosystem central banks’ bond holdings under risk-sharing. Publications from the ECB and the national central banks relevant to this breakdown together comprise all purchases of public bonds by Member States under the monetary policy asset purchase programmes. This also encompasses bonds issued by national development banks and other national public undertakings which are not included in government debt. In addition, there are holdings of bonds issued by supranational institutions. The holdings are stated at balance sheet values. However, in the SHSS holder statistics used for the creditor structure, they are recorded at face values.

For the depiction of the interconnectedness between governments and their national banking system, data on banks’ securities holdings and credit claims were taken from the ECB’s balance sheet statistics.

Banca d’Italia (2024), Annual Accounts – Ordinary Meeting of Shareholders 2023 – 130th Financial Year, 28 March 2024.

Banco de España (2024), Cuentas Anuales Del Banco De España 2023, 27 March 2024. 

Banque de France (2024), Rapport Annuel 2023, 15 March 2024.

Deutsche Bundesbank (2024a), Annual Report 2023, 23 February 2024.

Deutsche Bundesbank (2024b), Commentary: The Maastricht debt of the EU institutions and its fiscal implications for Germany, Monthly Report, April 2024.

Deutsche Bundesbank (2022), Government debt in the euro area: developments in creditor structure, Monthly Report, July 2022, pp. 77-89.

Deutsche Bundesbank (2021a), Government finances: Central bank bond purchases increase sensitivity to interest rate changes, Monthly Report, June 2021, pp. 39-45.

Deutsche Bundesbank (2021b), Planned refinement of the European resolution regime, Financial Stability Review 2021, pp. 72-74.

Deutsche Bundesbank (2018), Maastricht debt: methodological principles, compilation and development in Germany, Monthly Report, April 2018, pp. 57-78.

Deutsche Bundesbank (2016), The role and effects of the Agreement on Net Financial Assets (ANFA) in the context of implementing monetary policy, Monthly Report, March 2016, pp. 83-93.

European Central Bank (2023), Monetary policy decisions of 15 June 2023, Press release, Monetary policy decisions (europa.eu)

Federal Ministry of Finance (2023), Kreditaufnahmebericht des Bundes 2022, p. 72.

Ministero dell'Economia e delle Finanze (2024), Frequently asked Questions - BTP Valore

Official Journal of the European Union L 39 (2020a), Decision (EU) 2020/188 of the European Central Bank of 3 February 2020 on a secondary markets public sector asset purchase programme (ECB/2020/9) (recast), 12 February 2020, pp. 12-18.

Official Journal of the European Union L 91 (2020b), Decision (EU) 2020/440 of the European Central Bank of 24 March 2020 on a temporary pandemic emergency purchase programme (ECB/2020/17), 25 March 2020, pp. 1-4.