A. introduction

1. The global financial crisis and the ensuing sovereign debt crisis in Europe have had a marked impact on financial integration and debt funding costs in Europe. The blow-out in credit spreads for the sovereign debt of several euro area member states came unexpected and sharply reversed a long period of convergence. Before the financial crisis, the euro area bond market was one of the most integrated financial market segments (ECB, 2009). Following the introduction of the euro, debt funding costs fell sharply for many member states to the much lower German level and differences in government bond yields never exceeded 50 basis points until August 2007. Not only the public sector benefited from the common currency but banks and other firms also saw a marked decline in debt costs. Once corrected for corporate- and sector-specific risk, the geographic location of a firm explained only a very small portion of the variance of corporate bond yields, typically no more than 2 percent (ECB, 2004).

2. Markets changed course after the onset of the financial crisis and a jump in global risk aversion. Some sovereigns benefited from a flight to safety, while others and risky corporate debt saw their spreads rise sharply. Initially, the common currency shielded euro area member states from liquidity shortages and sudden stops in debt flows. With augmented access to the ECB and swap lines agreed between the ECB and the U.S. Federal Reserve, banks could receive unlimited euro and dollar liquidity against a broad range of collateral, including of course, euro area government paper. However, when public debt dynamics deteriorated sharply and sustainability concerns arose, membership of a currency union turned into a double-edged sword for some. Lacking control over the currency in which their debt is issued, members with a surge in public debt suddenly appeared fiscally fragile (De Grauwe, 2011). Mody (2009) identifies the rescue of Bear Stearns in March 2008 as the turning point after which euro area countries became increasingly differentiated. Thereafter, sovereign spreads tended to rise when the prospects of the domestic financial sector worsened which was especially evident in Ireland. Following the failure of Lehman Brothers in October 2008, spreads also rose faster for countries with high ratios of public debt-to-GDP (Sgherri and Zoli, 2009), and blew out after the dismal Greek fiscal position was fully revealed.

3. The surge in credit spreads was not limited to sovereign debt markets. Pronounced global risk aversion also pushed the spread for riskier corporate debt to record levels. However, this spike was short lived and the pricing of average corporate risk has eased significantly over the past years. In contrast, sovereign spreads for several euro area member states continued to rise in 2010 and are close to their highest levels since introduction of the euro (Figure I.1). These developments were mirrored by private funding costs in those member states as sovereign spreads are often perceived as the benchmark for the pricing of other debt instruments (ECB, 2011).

Monthly 5-year Sovereign CDS Premia of Selected Euro Area Countries

Citation: IMF Staff Country Reports 2011, 186; 10.5089/9781462338542.002.A002

Source: DataStream

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Monthly 5-year Sovereign CDS Premia of Selected Euro Area Countries

Citation: IMF Staff Country Reports 2011, 186; 10.5089/9781462338542.002.A002

Source: DataStream

• Download Figure
• Download figure as PowerPoint slide

Monthly 5-year Sovereign CDS Premia of Selected Euro Area Countries

Citation: IMF Staff Country Reports 2011, 186; 10.5089/9781462338542.002.A002

Source: DataStream

• Download Figure
• Download figure as PowerPoint slide

4. This note sets out to explore the recent relationship between sovereign and private funding costs in the euro area which have displayed a strong correlation over the past year in several member states. The link between sovereign spreads and private funding costs is complex and causality can run in both directions. Moreover, both may at times reflect changes in overall economic conditions or a country’s outlook causing a spurious relationship between public and private funding costs. Generally, differences in individual euro area sovereign debt spreads reflect default and liquidity risk given the absence of currency risk. The liquidity premium for smaller member states has likely increased somewhat but the sharp differences in the price of public debt for Austria and Finland, on the one hand, and Greece, Ireland and Portugal, on the other hand suggest that perceived default risk accounts for the bulk of these countries’ debt spreads. There are several channels through which their high sovereign funding costs may affect private borrowing costs, and vice versa:

• A country with significantly higher sovereign funding costs may have to reduce its deficit sharply and would be expected to take fiscal consolidation measures, possibly including higher corporate taxes (Papademos, 2010). This would lower firms’ net profits, worsen their credit risk assessment and drive up their credit risk premia.

• Strict fiscal austerity may have a negative short-term impact on economic growth and weigh on firms’ profits as well, driving up credit risk premia.

• In the extreme event of a disorderly sovereign default, a country’s foreign debt financing often dries up completely at least for some time and, especially for exporters, effective tax rates can rise sharply as the government seeks to raise revenue in foreign exchange.

• Large systemically important firms including banks are likely to receive some state support if a crisis hits, establishing a direct link from sovereign to private funding costs.

• At the same time, the financial health of private firms may have implications for public funding costs if these firms are considered systemically important for the country’s economy, or the European financial system. The Irish case is a prominent example where private banks’ default risk spread to its sovereign after debt liabilities were first guaranteed and later transferred to the government.

B. Data and Estimation

5. The estimation first separates country effects in corporate CDS premia from sector-specific and individual credit risk. A second step relates these estimated country effects to changes in sovereign CDS premia and a country’s economic outlook, measured by the national stock index. Using monthly data comprising CDS for about 200 euro area firms, their credit ratings and industry classification, sovereign CDS and national stock market indexes, the following section presents evidence that there is a positive and significant link between sovereign and private funding costs.

Data description

6. The sample includes monthly data of CDS for about 200 euro area firms, their credit ratings and industry classification, sovereign CDS and national stock market indexes of 11 euro area countries starting in 2008.² Differences in funding costs are measured by 5-year CDS for senior debt denominated in euro. While CDS premia and funding spreads should theoretically move in parallel, Fontana (2010) and Fontana and Scheicher (2010) show that CDS premia for sovereign and corporate debt have deviated at times from debt spreads, defined as debt yield minus a benchmark rate, since the onset of the financial crisis. But the difference rarely exceeded 100 basis points. Most euro area firms issue debt infrequently and at different maturities. The use of 5-year CDS simplifies the analysis and does not require correcting for differences in term structure across corporate debt. The credit rating is the lower of either Moody’s or S&P’s rating if available and adjusted over time. Industry classification includes the following sectoral classification: banking, non-bank financials, manufacturing, energy, transport and telecommunications. Figures I.2 and I.3 present CDS premia for a selection of euro area banks and telecoms. Greek, Irish, and Portuguese firms clearly stand out with Spain marking the border between low and high premia. CDS premia for telecoms are markedly lower than the ones for sovereign and bank CDS premia proving that the sovereign does not necessarily set a lower bound for private credit costs.

Monthly 5-year CDS Premia of Large Banks in Selected Euro Area Countries

Citation: IMF Staff Country Reports 2011, 186; 10.5089/9781462338542.002.A002

Source: DataStream

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Monthly 5-year CDS Premia of Large Banks in Selected Euro Area Countries

Citation: IMF Staff Country Reports 2011, 186; 10.5089/9781462338542.002.A002

Source: DataStream

• Download Figure
• Download figure as PowerPoint slide

Monthly 5-year CDS Premia of Large Banks in Selected Euro Area Countries

Citation: IMF Staff Country Reports 2011, 186; 10.5089/9781462338542.002.A002

Source: DataStream

• Download Figure
• Download figure as PowerPoint slide

Monthly 5-year CDS Premia of Telecom Firms in Selected Euro Area Countries

Citation: IMF Staff Country Reports 2011, 186; 10.5089/9781462338542.002.A002

Source: DataStream

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Monthly 5-year CDS Premia of Telecom Firms in Selected Euro Area Countries

Citation: IMF Staff Country Reports 2011, 186; 10.5089/9781462338542.002.A002

Source: DataStream

• Download Figure
• Download figure as PowerPoint slide

Monthly 5-year CDS Premia of Telecom Firms in Selected Euro Area Countries

Citation: IMF Staff Country Reports 2011, 186; 10.5089/9781462338542.002.A002

Source: DataStream

• Download Figure
• Download figure as PowerPoint slide

Estimation of country effects

7. To explore the link between private and public borrowing costs, the estimation proceeds in two steps. First is the decomposition of the variation in corporate CDS premia into sector-, credit-, and country-specific effects. Second is the estimation of a relationship between these country-specific factors and sovereign CDS premia. Following broadly the methodology of Baele and others (2004), based on the Heston and Rouwenhorst (1994) approach for equity returns, corporate default spreads are modeled to consist of five components: a common factor (α), a credit risk factor (β), an industry, or sector-specific factor (γ), a country factor (δ), and a firm-specific disturbance (ε).

$\begin{array}{cc}CD{S}_{it}={\alpha }_t+{\beta }_{it}+{\gamma }_{it}+{\delta }_{it}+{\varepsilon }_{it}& \left(1\right)\end{array}$

A time series for common, credit, industry and country factors is estimated by running the following cross-sectional regression each month:

$\begin{array}{cc}CD{S}_i=\alpha +{\mathrm{\Sigma }}_{j=1}^J{\beta }_j{I}_{ij}^{Credit}+{\mathrm{\Sigma }}_{k=1}^K{\gamma }_k{I}_{ik}^{Industry}+{\mathrm{\Sigma }}_{l=1}^L{\delta }_i{I}_l^{Country}+{\varepsilon }_i& \left(2\right)\end{array}$

where (I) is a dummy variable that is one if the firm with CDS_i has credit rating (j), belongs to industry (k), or country (l). Since each firm belongs to at least one country, credit rating and industry, the estimation can only reveal cross-sectional differences between countries, industries and credit ratings and the following parameter restrictions are imposed:

$\begin{array}{cc}{\mathrm{\Sigma }}_{j=1}^J{\beta }_j=0;{\mathrm{\Sigma }}_{k=1}^K{\gamma }_k=0;{\mathrm{\Sigma }}_{i=0}^I{\delta }_i=0& \left(3\right)\end{array}$

8. Credit rating and country effects may not be fully independent. Sovereign credit worries often lead to across-the-board downgrading of the credit ratings of that country’s firms and especially banks. They may even spill over from one member state to others that are perceived to display similar characteristics (see, Arezki and others 2011). Therefore, it is possible that the above equation may feature a bias and likely underestimate country effects to some extent. Figure I.4 presents the evolution of country effects since early 2008.³ The variation in country effects has risen markedly since early 2008, especially for countries in the periphery: Greece, Ireland and Portugal stand out with Spain being the borderline case. Also, the statistical significance of country effects rises over time. Other than for banks, sector-specific effects are not significant and, as a result, sectoral differentiation is reduced to banks and non-banks with the former, as expected, featuring higher CDS premia. The credit rating factors (3 for investment grade credit and 1 for speculative/non-investment credit) are mostly significant and the factor for speculative credit nicely tracks the credit spread between average 5-year BBB euro corporate yields (IBOXX) and German Bunds.

Estimated Country Effects for Selected Euro Area Countries

Citation: IMF Staff Country Reports 2011, 186; 10.5089/9781462338542.002.A002

Source: Staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Estimated Country Effects for Selected Euro Area Countries

Citation: IMF Staff Country Reports 2011, 186; 10.5089/9781462338542.002.A002

Source: Staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Estimated Country Effects for Selected Euro Area Countries

Citation: IMF Staff Country Reports 2011, 186; 10.5089/9781462338542.002.A002

Source: Staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Estimation of the relationship between sovereign and corporate CDS

9. Country effects may reflect sovereign worries. They could also signal changes in a country’s economic outlook measured by changes in national stock market indexes, as discussed above, although most firms in the sample (with actively traded CDS) are relatively large and their operations internationally diversified. In a panel regression, the time series of country factors (δ) is regressed on sovereign CDS premia (x) and national stock market indexes (y) with various specifications of fixed/random cross-section and time effects.

$\begin{array}{cc}{\delta }_{lt}=\alpha +\mu x_{lt}+\sigma y_{lt}+{\pi }_l+w_t+{\varepsilon }_{lt}& \left(4\right)\end{array}$

10. For various specifications (with/without fixed and random cross-section/time effects), estimation of (4) reveals a significant and positive relationship between sovereign CDS premia and the country effects that were estimated in the previous section. Of a 100-basis point increase in sovereign spreads about 50–60 basis points are passed on to private firms (Table 1), noting that causality may not always work in only one direction. At the same time, the relationship between changes in national stock market indexes and country effects is small and not significant for most specifications. Sovereign CDS premia and national stock market indexes may not be fully independent but regressed on each other do not exhibit any significant relationship over the sample. To learn more about possible causality, correlations and Granger causality is explored at the country level for sovereign CDS premia and country effects. Contemporaneous correlation (no lags) dominates and Granger causality tests are largely inconclusive.

Table I.1.

Panel Estimates of the Effects of Sovereign CDS and Stock Price Developments on Country Effects

Source: Staff estimates. Sample (total pool) contains 142 monthly observations from 2008Q1-2011Q1 for the following countries: Austria, Belgium, Finland, France, Germany, Greece, Italy, Ireland, Netherlands, Portugal, Spain. Stock index remains insignificant if included as first differences instead of levels; ^***p<0.01, ^**p<0.05, ^*p<0.1.

Table I.1.

Panel Estimates of the Effects of Sovereign CDS and Stock Price Developments on Country Effects

	(1) Fixed effects Cross-country	(2) Fixed effects Time	(3) Fixed effects Cross-country+Time	(4) Random effects Cross-country	(5) Random effects Time	(6) Random effects Cross-country+Time
Variables
Sovereign CDS	0.512***	0.523***	0.581***	0.529***	0.537***	0.537***
	(0.034)	(0.047)	(0.054)	(0.035)	(0.037)	(0.037)
Stock Index (level)	0.000	1.306**	-0.567	-0.244	0.265	-0.266
	(0.002)	(0.601)	(0.916)	(0.325)	(0.353)	(0.362)
Constant	27.630**	125.600**	-19.630	10.691	47.180	8.347
	(19.134)	(48.717)	(72.252)	(29556)	(29.311)	(32.38)
R-squared	0.82	0.72	0.84	0.66	0.69	0.66
Prob(F-Statistic)	0.00	0.00	0.00	0.00	0.00	0.00

Source: Staff estimates. Sample (total pool) contains 142 monthly observations from 2008Q1-2011Q1 for the following countries: Austria, Belgium, Finland, France, Germany, Greece, Italy, Ireland, Netherlands, Portugal, Spain. Stock index remains insignificant if included as first differences instead of levels; ^***p<0.01, ^**p<0.05, ^*p<0.1.

View Table

Table I.1.

Panel Estimates of the Effects of Sovereign CDS and Stock Price Developments on Country Effects

	(1) Fixed effects Cross-country	(2) Fixed effects Time	(3) Fixed effects Cross-country+Time	(4) Random effects Cross-country	(5) Random effects Time	(6) Random effects Cross-country+Time
Variables
Sovereign CDS	0.512***	0.523***	0.581***	0.529***	0.537***	0.537***
	(0.034)	(0.047)	(0.054)	(0.035)	(0.037)	(0.037)
Stock Index (level)	0.000	1.306**	-0.567	-0.244	0.265	-0.266
	(0.002)	(0.601)	(0.916)	(0.325)	(0.353)	(0.362)
Constant	27.630**	125.600**	-19.630	10.691	47.180	8.347
	(19.134)	(48.717)	(72.252)	(29556)	(29.311)	(32.38)
R-squared	0.82	0.72	0.84	0.66	0.69	0.66
Prob(F-Statistic)	0.00	0.00	0.00	0.00	0.00	0.00

Source: Staff estimates. Sample (total pool) contains 142 monthly observations from 2008Q1-2011Q1 for the following countries: Austria, Belgium, Finland, France, Germany, Greece, Italy, Ireland, Netherlands, Portugal, Spain. Stock index remains insignificant if included as first differences instead of levels; ^***p<0.01, ^**p<0.05, ^*p<0.1.

C. Conclusion

11. The reemergence of perceived credit default risk for euro area sovereigns is a key legacy of the global financial crisis. This note has provided some evidence that the associated premia embedded in sovereign credit costs are closely correlated with private funding costs and likely in large part passed on to the firms located in these countries. This is the case not only for banks but also for other non-financial firms, including telecoms, utilities, and others.

12. First and foremost, strict fiscal consolidation is needed in countries with high public debt and funding costs. If sustained, however, significant differences in the cost of private capital across countries reflecting sovereign concerns should complicate monetary policy and may trigger a relocation of capital intensive activities to low risk/cost destinations. Large banks and other firms that rely on external debt financing may await a similar process of consolidation across Europe that national airlines experienced over the past decade, although for different reasons. However, member states may resist such developments which could trigger protectionism and fragmentation of markets with highly adverse economic consequences.

13. To avoid such setbacks in economic and financial integration of Europe’s economy the links between sovereigns and firms located in their jurisdiction will need to be loosened. How this could be achieved in practice is open to debate but further European integration along the following lines should help in breaking the financial links between firms and their national sovereigns and eliminate possible channels through which costs associated with sovereign default risk could affect private funding costs and vice versa:

• European instead of national support schemes for bank restructuring or resolution support should decouple banks from their national sovereign and significantly reduce differences in banks’ debt costs across Europe. Moreover, they could internalize the negative external effects of ailing banks on the entire euro area financial sector and facilitate resolution of cross-border institutions and cross-border M&As.

• Harmonizing regulation and taxation and working towards the introduction of a common corporate tax for large firms would help. This would eliminate the threat of large effective corporate tax increases in a country where the sovereign is under financial pressures, or even may default on its debt obligations.

• A strong role for European competition and supervisory authorities should accompany restructuring and any cross-border consolidation of banks and other firms that may occur to safeguard competition and financial stability.