Determinants of EMU Government Bond Spreads
Berenice van Gessel * Master’s Thesis Msc BA – Finance
January 2012
Faculty of Economics and Business – University of Groningen
ABSTRACT
This paper designs a framework to empirically test the determinants of government bond yield spreads in the Euro area with a focus on developments during the sovereign crisis that started in 2009. Using panel data of sovereign spreads on nine EMU countries from November 2006
until May 2011, I test whether credit risk and liquidity risk have an increased effect on government bond yield spreads during the crisis. Results show that both credit and liquidity
risk are important in explaining the widening of spreads in the EMU. In times of financial crisis, liquidity risk appears to have a significantly less pronounced effect on sovereign
spreads, whereas credit risk becomes a more significant determinant of these spreads.
JEL Classification: E43, G01, G12, G15, H63
Keywords: Government bonds; credit risk; liquidity risk; financial crises
Supervisor: Dhr. L. Dam
I. Introduction
The objective of this paper is to test whether the determinants of European Monetary Union (EMU) government bond yield differentials have an increased effect during the recent
sovereign crisis. In general, there are three main determinants that cause spreads in sovereign yields: credit risk, liquidity risk and risk aversion (Manganelli and Wolswijk (2009)). Previous literature on Regime Switching Models (Hamilton (2005), Calvet and Fisher (2004), Hamilton 1988)) shows that in times of crisis, underlying economic parameters can behave differently compared to normal times.
In the aftermath of the global financial crisis, the Euro zone has witnessed a transformation from the financial crisis into a sovereign debt crisis. As the recent crisis unfolded, investors developed a strong preference for relatively safe government bonds compared to private debt or company stock. In particular, the US and Germany were singled out as harbors of safety. This behavior caused a higher degree of differentiation within the sovereign bond market. Where 10-year yield spreads to the German Bund averaged 18 basis points in the period from 1999 to mid-2007, they averaged 75 basis points since October 2008 and 380 basis points since May 2011. At the same time, differences between Euro area countries have become more pronounced, as the spreads of some countries, such as Ireland, Portugal, Italy and Greece, widened much more than those of other countries, like France and the Netherlands. The rising spreads put increased pressure on EMU countries that did not have a sound fiscal policy and eventually led to large political reforms. The prime ministers of Greece, Ireland, Portugal and Spain resigned and the markets even forced the prime minister of Italy, Silvio Berlusconi, to leave. The bond market proved to have the biggest disciplining power on fiscal policies of the EMU countries.
But why did the increased spreads in the sovereign bond market have so much disciplining power and what caused these spreads? And why did the central banks, the International Monetary Fund (IMF) and the European Central Bank (ECB) not succeed in resolving the crisis so far?
Union (EU) decided together with the International Monetary Fund (IMF) on a 110 billion emergency rescue package for Greece. Moreover, the EU and IMF created a European stabilization mechanism called European Financial Stability Facility (EFSF) worth 750 billion euro for countries that might face problem similar to Greece within the next three years. At the same time, the ECB has actively been buying long-term government bonds of Ireland, Greece, Spain and Portugal, thereby increasing the price of these bonds and reducing their yields. In the first two weeks of September 2011 alone, rescue packages were worth over $20 billion. However, these measures proved to have only a short-term effect on the yield spreads in the Euro zone.
Another approach to lower sovereign spreads that is considered seriously by the European Commission is to issue Euro denominated bonds. These bonds, guaranteed collectively by the governments of the Euro zone, would increase overall liquidity in the bond market and in turn could lead to lower yield spreads. However, a joint Euro bond issue also leads to the free riding problem. The governments of countries that already face high spreads such as Ireland, Greece, Portugal, Italy, and Spain will have fewer incentives to decrease the government deficit. As a result, the countries with sustainable fiscal policies and low spreads may have to bail out the governments of these countries in case of default.
To break down empirically the driving factors of yield differentials is challenging because of the frequency mismatch. Liquidity factors influence yields at high frequencies whereas credit risk characteristics are based on slow-moving factors such as public debt and current account deficits, which are observed at lower frequencies (Codogno et al. (2003)). I resolve this issue by using credit default swap (CDS) spreads as a proxy for credit risk to obtain a higher frequency measure for the credit risk component.
This paper is of particular relevance because it analyzes the effects of the recent sovereign crisis on the bond market of the European Monetary Union (EMU). Many authors have done research on this topic focusing on either liquidity risk (Bernoth et al. (2006), Beber et al. (2006)) or credit risk (Schuknecht et al. (2008), ECB (2009)). Furthermore, the sample of previous papers written on this topic did not include the most recent and intense phase of the sovereign crisis. In this paper I will combine these factors and show to what extent determinants of sovereign bond spreads have an effect on different Euro area countries and whether these effects have become more pronounced since the recent financial crisis.
The main research question of this paper will be:
Do credit risk and liquidity risk have an increased effect on EMU government bond spreads during the recent sovereign crisis?
II. Determinants of Yield Spreads
Yield differentials for the Euro area are shown in Figure 1. Table 1 shows that since November 2009, average spreads increased a little for Austria, Belgium, France and the Netherlands, and increased by a factor of eight in Greece, six in Portugal and five in Ireland.
A lot of research has been done on risk premia and credit spreads. Literature distinguishes in general between structural approaches derived from the Merton model (1974) that links short-term interest rates to corporate bond spreads and reduced form models (Jarrow and Turnbull (1995)). According to the structural approach, debt holders hold a put option on a company’s assets. Whenever the company’s enterprise value falls below the nominal value of the debt, the company defaults and the option will be exercised. In case of the reduced form model, the default is determined by exogenous factors such as country-specific and macroeconomic factors. Moreover, it uses a number of different variables as determinants of country risk. Country ratings are frequently used in order to determine country risk. However, the recent financial crisis has shown the flaws of the dependency on rating agencies, as they have been slow to adapt their ratings during the recent financial crisis.
Figure 1. Euro area sovereign spreads versus 10-year German bond yields
Table 1: Average government bond yield spreads
Austria Belgium France Greece Ireland Italy NL Portugal Spain Observations
Entire sample 33,47 47,99 24,30 287,82 166,25 80,87 22,89 131,38 75,75 1190
Pre-crisis 31,17 36,08 19,40 88,28 66,16 58,37 22,95 49,51 34,23 783
Crisis 37,90 70,91 33,73 671,71 358,81 124,15 22,78 288,87 155,65 407 Note: The yield spread is the spread between the yield of a EU government bond and the German government bond with a 10 year maturity. The
pre-crisis period is 1/11/2006-31/10/2009. The crisis period is 1/11/2009-24/05/2011. All statistics are in basis points
Source: Bloomberg
Differences in interest rates across countries can be explained by the variation in de risk premium that investors demand for holding these assets, which consists of both liquidity and credit risk factors. Duffie et al. (2003) and Longstaff et al. (2005) show that both credit and liquidity concerns are critical components of yield spreads. Furthermore, using data until 2004, Beber et al. (2006) find that, while credit risk matters for bond valuation under normal circumstances, liquidity becomes more important in times of financial distress.
I will elaborate on the determinants of bond yield spreads in the following paragraphs.
2.1 Regime Switching Models
Regime Switching Models are typically used to model and identify nonlinearities in time series from one period (or regime) to another. In many cases, the different regimes are not directly observable and one has to use the Markov switching model do indentify these regimes. This is not necessary in this paper since the different regimes (pre-crisis and crisis) are already known but it helps to explain why the underlying parameters of a process might behave differently over time.
2.2 Credit Risk
Credit risk refers to the risk that the country may not fully meet its obligations. It depends primary on a nation’s current and future stated and hidden debt, and debt sustainability. Debt sustainability depends on expected budget surpluses and deficits, as well as future economic activity and interest rates, which in turn are affected by domestic and international factors and policies (Codogno (2003)). With the introduction of the Euro, EMU member states lost the option of printing money to pay for their debts, hence credit risk may have become even more important as the exchange risk disappeared. According to Dötz et al. (2011), the long-term sustainability of a country’s public debt is a key determinant credit risk.
Brandner (2007) argues that differences in various country-specific macroeconomic fundamentals, such as a nation’s debt, can have an impact on yield spreads in the Euro area. However, economic theory provides no clear answer to the effects of fiscal policy on interest rates. Whereas the ECB (2006) holds that variations of fiscal fundamentals have large effects on yield spreads, the empirical evidence in Balassone, Franco, Giordano (2004) raises some doubts on such a strong relationship for the period since 1999. Furthermore, Mody (2009) finds no effect of fiscal variables on European bond yield spreads. In this paper, I use the debt-to-GDP ratio in order to explain the different effects of liquidity- and credit risk on sovereign spreads per country.
The recent financial crisis has shown the successfulness and the flaws of the Stability and Growth Pact (SGP). If markets would anticipate that highly indebted countries could turn to member states for a bail-out in case of distress, this would lower or eliminate risk premiums in the European Monetary Union (EMU). Thus, the existence of risk premiums in the EMU would prove the effectiveness of the "no bail-out clause" in the Maastricht Treaty. On the other hand, assuming that yield spreads reflect differences in credit standings, the SGP and the European fiscal framework are lacking to ensure that all member states have the same creditworthiness from the market point of view (Codogno et al. (2003)). Moreover, fiscal rules such as the SGP could have an impact on the market perception of default risk, and consequently on interest rates (see, for example, Poterba and Reuben (2001)).
in the Euro area. Since 2000, the credit derivatives market has experienced rapid growth, and among credit derivatives, the CDS has become the most widely traded instrument for transferring credit risk. In a standard CDS contract, the buyer insures itself against default of the underlying asset. It transfers the risk of default from the buyer to the seller in exchange for the payment of a regular fee, i.e. a premium.
Following Hull (2004), Zhang and Zhu (2005), and Deutsche Bundesbank (2011), I use the CDS premia of Euro zone countries versus the CDS premia of Germany as a direct measure of the CDS spread. The premium is calculated such that the CDS has zero value at the time of origination. The CDS spread is the insurance premium that the buyer pays to the seller each period until either default occurs or the contract matures (Hull 2004). In case of default, the seller is obligated to buy back the defaulted bond at its par value from the buyer (Longstaff (2005)).
It is important to emphasize that the nature of certain factors that drive default decisions, and the range of alternative restructuring and recovery plans considered in the event of default, typically differ between sovereign and corporate debt (Duffie (2003)). Sovereign default is, unlike default of a corporate or other entity, largely a political decision. Governments face a trade off between the cost of fulfilling their debt obligations against reputation costs (Eaton and Gersovitz (1981)), the costs of having assets abroad seized, and the costs of having international trade impeded (Bulow and Rogoff (1989b)). It is very uncommon for a sovereign to make an outright default. Rather, it may force a restructuring or renegotiation of its debt, as Greece has been force to do so in November and December 2011. The incentives of a corporate are normally less complicated: it defaults (in theory) when the company cannot meet its obligations to pay debt holders, or when equity holders find that the market value of their shares has dropped to zero.
the private to the public sector. Manganelli & Wolswijk (2009) show there is a clear correlation between a country’s debt developments and interest rate spreads; higher debt ratios lead to higher sovereign yield spreads. However, evidence also shows that debt is not a perfect proxy for a country’s fiscal soundness. For example, in 2007, Belgium has a high debt ratio but a relatively low yield spread against Germany. In addition, Austria, France and Portugal have approximately similar debt ratios, but the bond spread versus Germany of the latter country is much higher.
Based on previous literature I expect that credit risk and government bond spreads are positively related and that credit risk had a larger impact on government bond spreads during the recent sovereign crisis. Furthermore, I expect that the different debt-to-GDP ratios of the EMU countries can explain the differences in the effect of liquidity- and credit risk on the various sovereign spreads.
2.3 Liquidity Risk
Historically, empirical evidence in the literature suggests that liquidity risk exerts only a minor influence on bond yield spreads in the Euro area (Pagano and von Thadden (2004) and Brandner et al. (2007)) and that markets for sovereign bonds have different levels of liquidity. Liquidity may vary across countries depending on trading volumes, the amounts of bonds outstanding, the trading activity of market makers, and the efficiency of the secondary market (Codogno et al. (2003)). Bonds with a maturity of 10 years are highly standardized products, but outstanding amounts vary considerably depending on country specific fundamentals. Secondary market characteristics such as admission and trading rules or clearing and settlement procedures may equally be critical for liquidity, and especially the willingness of market makers to quote both bid and ask prices. The incentives to trade and invest in specific bonds may also depend on the availability of hedging and financing instruments, such as liquid and efficient future contracts (Codogno et al. (2003)). For instance, the well established futures market of German sovereigns is a key reason for its safe haven status.
execution. An investor willing to transact faces a trade-off: He may either wait to transact at a favourable price or insist on immediate execution at the current bid or ask price. The quoted ask (offer) price includes a premium for immediate buying, and the bid price similarly reflects a concession required for immediate sale. Thus, a natural measure of illiquidity is the spread between the bid and ask prices. The results of Amihud and Mendelson (1986) show that asset returns are an increasing and concave function of the spread.
Gomez-Puig (2006) and Manganelli and van Wolswijk (2009) find that there is an inverse relationship between liquidity and the yield spread between government bonds. Geyer, Kossmeyer and Pichler (2004) do not detect any impact of liquidity variables (issue size, bid-ask spread, yield differentials between on-the-run and off-the-run issued) on sovereign spreads in the Euro zone. Favero et al. (2009) provide both theoretical justification for, and empirical evidence of, a role of liquidity in the determination of sovereign risk spread and how it may interact with the aggregate risk factor. In a sample spanning 2002 and 2003, they confirm the role of the aggregate risk factor and demonstrate that liquidity is only significant when interacted with the aggregate risk factor. Thus, liquidity has a smaller effect on sovereign spreads in periods in which the level of risk is high. The total effect of liquidity risk on sovereign risk is thus negative in periods of high aggregate risk. This is explained by a reduced set of alternative investment opportunities limiting the willingness of investors to move away from bonds. Therefore, although in general investors value liquidity, they value it less when risk increases. In contrast, Beber et al. (2009) find that liquidity considerations are more important during episodes of market stress in a sample covering 2003 and 2004. In another study, Pastor and Staumbagh (2003) show that liquidity risk, defined as the possibility that securities will become illiquid exactly when investors want to close out their position, is the important factor priced in asset returns.
buying (selling) beyond the best bid and ask prices. The larger the order, the greater the price concession the seller usually makes to effect a sale, and the greater the price premium the buyer has to pay for an immediate purchase. Market-impact costs can be interpreted as the effective bid-ask spread for large orders, which is revealed through actual transaction prices. (3) Delay and search costs are incurred when a trader delays the execution of a transaction in an attempt to accomplish better trading terms-e.g., better prices than those quoted, or reduced market-impact cost. These costs include the actual cost of contacting potential trading partners and the risk borne by the investor while searching and delaying the execution.
In this paper I will focus on liquidity risk in the form of the bid-ask spread.
More studies support the theory that higher illiquidity entails higher expected return, after controlling for risk and other characteristics, using a number of alternative measures of liquidity. Brennan and Subrahmanyam (1996) use each stock’s price impact cost (per unit of order size) as well as the fixed costs, associated with the bid-ask spread, and estimated the effect of these measures of illiquidity on stock returns. They obtain a strong positive relationship between average stock return and both measures of illiquidity costs. Datar et al. (1998) measure liquidity using stock turnover (the ratio of trading volume to shares outstanding), and Brennan et al. (1998) use trading volume to measure stock liquidity. Both find that the higher the liquidity of a stock, the lower its return after controlling for risk and other characteristics. This relationship, which is consistent with the theory, is found to be very robust. Furthermore, differences in government bond market liquidity have also been found to be significant for many Euro area countries in a study by Bernoth et al. (2006).
Based on this literature I expect that liquidity risk and government bond spreads are positively related and that liquidity risk had a larger impact on government bond spreads during the recent sovereign crisis.
III. Data
between the start of November 2006 until mid May 2011 and obtained from Bloomberg. This period thus includes the period of the financial crisis starting in September 2008 and the sovereign crisis starting November 2009. Even though the Euro area was officially out of recession since early 2010, the turmoil in the Euro bond market continued throughout 2010 and 2011 so I decided to include these years in the crisis period as well. Because I expect that the financial crisis will have a lasting impact on the coefficients of the estimate, I will use a dummy variable for the period prior to and the period since the beginning of the recent crisis. Period 1 (the pre-crisis period) is from 1 November 2006 until 14 September 2008. The sample starts with 1 November 2006 due to the availability of data. Period 2 (crisis period) lasts 1 November 2009 until 24 May 2011. Even though the global financial crisis started with the collapse of Lehman Brothers in 2008, the sovereign crisis began with the escalation of the Greek debt crisis in November 2009. Table 4 in the Appendix also shows that the sovereign crisis period has a larger effect on the yield spreads than the global financial crisis period.
The following countries are included in the analysis: Austria, Ireland, Finland, Belgium, Spain, France, Greece, Italy, and Portugal (with Germany as a benchmark). The remaining EMU countries are not suitable to take into consideration either due to the unavailability of reliable data or because these countries only joined the Euro area in recent years and thus the exchange rate risk is an additional component which would have to be taken into account. In addition, for these missing countries there are no long enough time series on CDS spreads available so the data set consists of the ten countries mentioned above. I will use German government bonds as the appropriate benchmark in the Euro bond market.
German government bonds operated as a ‘flight-to-liquidity’ asset during the crisis (Arghyrou and Kntonikas (2011)). According to Dunne et al (2007) and Schuknecht et al (2009) the high liquidity of German Bunds has contributed significantly to the benchmark status of Bunds. This feature of the German bond becomes apparent in Figure 2, which plots the 10-year German government bond yield with the Chicago Board Options Exchange (CBOE) Volatility Index (VIX). This index is a measure of US implied stock market volatility1, often used as a general indicator of common international risk and obtained from
DataStream. In the second half of 2008, when the credit crisis reached its peak, the VIX increased sharply whereas the 10-year German Bund moved in the opposite direction. This occurred again during May 2010 and the second half of 2011 when the sovereign crisis escalated. These movements indicate that in an environment of increased uncertainty, investors ‘fled’ to the perceived safety of German bonds.
2 2,5 3 3,5 4 4,5 5 5,5 6 ja n '0 0 ju l ' 0 0 ja n '0 1 ju l ' 0 1 ja n '0 2 ju l ' 0 2 ja n '0 3 ju l ' 0 3 ja n '0 4 ju l ' 0 4 ja n '0 5 ju l ' 0 5 ja n '0 6 ju l ' 0 6 ja n '0 7 ju l ' 0 7 ja n '0 8 ju l ' 0 8 ja n '0 9 ju l ' 0 9 ja n '1 0 ju l ' 1 0 ja n '1 1 ju l ' 1 1 0 10 20 30 40 50 60 70 80 90 10y German Bund CBOE VIX
Figure 2. 10-year German government bond yield and the CBOE VIX
Source: DataStream
best measure of credit risk available at high frequency (e.g., Longstaff et al. (2005), Blanco, Brennan and Marsh (2005) or ECB (2009)). Alternative measures such as debt-to-GDP ratios are only available at yearly or quarterly frequency. As shown in Table 2, the CDS spreads increased two or threefold for Austria, Belgium, France, Italy and the Netherlands during the sovereign crisis. By contrast, Greece, Ireland, Portugal and Spain showed much higher multiples. Figure 5 in the Appendix plots the 10-year government bond yields spreads together with the corresponding CDS spreads. It proves that in most EMU countries the two variables tend to broadly co-move over time, especially during the credit and sovereign debt crisis. CDS spreads start to increase sharply in Portugal, Greece and Spain since November 2009.
Table 2: Average 10 year CDS spreads on government bonds
Austria Belgium France Greece Ireland Italy NL Portugal Spain Observations
Entire sample 58,96 63,74 41,42 278,41 167,49 95,99 35,45 133,95 101,60 1190 Pre-crisis 45,75 34,62 22,35 82,29 81,68 61,59 27,32 46,83 49,13 783 Crisis 84,39 119,77 78,11 655,72 332,57 162,18 51,08 301,55 202,56 407
Note: CDS spreads are measured as the 10 year CDS premium on government bonds in US$. The pre-crisis period is 1/11/2006-31/10/2009. The crisis period is 1/11/2009-24/05/2011. Total panel observations included: 10710.
Following Amihud and Mendelson (1986), Brandner et al. (2007) and De Nicolo and Ivaschenko (2008), I will measure liquidity risk by using bid-ask spreads as a fraction of its price: 100 * Pr ice Midquote Ask Bid isk LiquidityR (1)
slightly in Austria, Belgium, the Netherlands and Spain, and increased by a factor of ten for Greece, Ireland and Portugal.
Table 3: Average Bid-Ask spreads on government bonds
Austria Belgium France Greece Ireland Italy NL Portugal Spain Observations
Entire sample 7,49 9,22 4,09 33,93 35,31 9,81 7,24 36,98 8,54 1190
Pre-crisis 6,50 8,54 4,13 8,65 10,09 9,83 6,46 8,90 5,06 783
Crisis 9,40 10,53 4,01 82,50 83,77 9,78 8,74 90,93 15,21 407
Note: The Bid-Ask spread is calculated by the difference in the bid and ask price divided by the mid-quote price and multiplied by 100, denominated in US$. The pre-crisis period is 1/11/2006-31/10/2009. The crisis period is 1/11/2009-24/05/2011. Total panel observations
included: 10710. All statistics are in basis points.
Furthermore, I will use the debt-to-GDP ratios of the nine EMU countries to measure the additional effect of this ratio on the CDS- and Bid-Ask spreads in determining the sovereign yield spreads. According to Brandner (2007) and the ECB (2006) these country-specific macroeconomic fundamentals can have an impact on yield spreads in the Euro area. However, the amount and relationship can differ per country. The data on the debt-to-GDP ratios is quarterly from 1 November 2006 until 24 May 2011 and obtained from Eurostat.
Table 4: Average debt-to-GDP ratios
Austria Belgium France Greece Ireland Italy NL Portugal Spain Observations
Entire sample 67,7 93,1 73,4 121,3 55,8 112,2 55,5 78,5 47,1 19
Pre-crisis 65,4 90,0 68,0 110,9 37,1 108,2 51,4 71,0 40,1 12
Crisis 71,6 98,4 82,6 139,2 87,9 119,0 62,6 91,4 59,1 7
Note: The debt-to-GDP ratio measures the amount of national debt of a country as a percentage of its Gross Domestic Product (GDP). The pre-crisis period is 1/11/2006-31/10/2009. The crisis period is 1/11/2009-24/05/2011.
IV. Methodology
To study the influence of determinants on the sovereign yield spreads, an Ordinary Least Squares (OLS) regression is used. Firstly, I will estimate the influence of credit risk and liquidity risk for both period 1 and 2 on the Euro area as a whole, i.e. I measure the average effect of the CDS spread and the Bid-Ask spread on the yield spread as follows:
Where YieldSpreadit denotes the spread the 10 year government bond yield for country i to the German Bund at time t, c is a constant,CDSspread is the 10-year sovereign CDS spread it
for country i at time t , BidAskspreadit is the bid-ask spread for country i at time t . it is the stochastic error term for country i at time t.
Second, I estimate the effect of the crisis on the coefficients for the entire Eurozone:
it t it t it it it it crisis ad BidAskspre crisis CDSspread ad BidAskspre CDSspread c d YieldSprea * * 4 3 2 1 (3)
Where crisist is a dummy variable equal to one for the period of 1 November 2009 until 24 May 2011 and zero for November 2006 until 31 October 2009. In this way, I measure the additional effect that crisis may have on the determinants of the yield spread.
Furthermore, I measure the disaggregate effect of the bid/ask and CDS spread on the yield spread for every separate country, that is, I estimate an individual coefficient for every country. This is very similar to the Seemingly Unrelated Regression (SUR) approach, however I simply interact all independent variables with country dummy variables.
it i it i i it i
it c CDSspread Country BidAskspread Country
d
YieldSprea 1 * 2 * (4)
Where Countryi represents either Austria, Belgium, France, Greece, Ireland, Italy, the Netherlands, Portugal or Spain.
it t i it i t i it i i it i i it i it Crisis Country ad BidAskspre Crisis Country CDSspread Country ad BidAskspre Country CDSspread c d YieldSprea * * * * * * 4 3 2 1 (5)
Finally, I test whether the debt-to-GDP ratio has an additional effect on the determinants of the yield spread with the following equation:
it it it it it it it GDP D ad BidAskspre GDP D CDSspread ad BidAskspre CDSspread c d YieldSprea / * / * 4 3 2 1 (6) V. Empirical Results
Table 5: Average effect of credit risk and liquidity risk on yield spreads
Regression A B
CDS spread 0.010*** (0.00) 0.008 (0.00)
Bid-Ask spread 0.334*** (0.01) 0.855*** (0.09)
CDS during crisis 0.002*** (0.00)
Bid-Ask during crisis -0.651*** (0.09)
Constant -0.125*** (0.00) -0.003*** (0.00)
R² 0.969 0.969
Note: Regression A estimates YieldSpreadit = c + β1CDSspreadit + β2BidAskspreadit + εit, Regression B estimates YieldSpreadit = c + β1CDSspreadit + β2BidAskspreadit + β3CDSspreadit*crisist + β4BidAskspreadit*crisist + εit . Yield spread is the government bond spread of country I to the German bond. CDS spread is the CDS premium and Bid-Ask spread is the difference between the bid and the ask price divided by the mid-quote price of the government bond of country i. The crisis dummy is equal to 1 for the period 1 November
2009 until 24 May 2011 and zero for 1 November 2006 until 31 October 2009. Total panel observations included: 10710. *** indicates significance at a 1% level
Table 6 displays the contribution that both the credit risk component as the liquidity risk component provide to the bond yield spread for each country. A review of this table shows that the CDS spread has a positive and significant effect on the yields spreads on all nine Euro countries, which is consistent with the hypothesis that sovereign spreads and CDS spreads are positively related. The Bid-Ask spread has a positive and significant effect on the bond yield spread of Ireland, Italy, the Netherlands and Portugal. By contrast, the Bid-Ask spread has a significant negative effect on the yield spread for Austria, France and Greece. The results of these last three countries are inconsistent with my previous hypothesis namely, that liquidity risk and sovereign spreads are positively related, and suggests that when liquidity risk on these bonds become higher, investors demand lower compensation, i.e. lower interest rates.
Greece, Ireland, Portugal and Spain, but a slightly less significant effect for Belgium, France and the Netherlands. The CDS spreads in Austria and Italy do not have a significant effect increased or decreased effect during the crisis on their bond yield spreads. These results are consistent with Codogno (2003) and Brander (2007); countries with higher debt-to-GDP ratios and thus a higher probability of default, such as Greece, Ireland, Portugal and Spain, have higher credit risk than countries with normal government debt ratios and thus have increased government bond spreads. Moreover, the liquidity component shows more pronounced mixed results. The Bid-Ask spread becomes strongly and significant less important as a determinant of the sovereign yield spread for Austria, Greece, Portugal and Spain. This is explained by Favero et al. (2006) which show that a reduced set of alternative investment opportunities limits the willingness of investors to move away from bonds. Therefore, although in general investors value liquidity, they value it less when risk increases. On the other hand, liquidity risk becomes significantly more important for France, Ireland, and the Netherlands. This means that when liquidity risk increases, i.e. the liquidity of the bond decreases, the sovereign spread increases as well. Gomez-Puig (2006), Geyer et al. (2004) and Manganelli and van Wolswijk (2009) also find an inverse relationship between liquidity and the yield spread between government bonds. The liquidity risk factor does not have an additional effect during the sovereign crisis on the bond yield spreads in Belgium and Italy. These results are inconsistent with the hypotheses but consistent with previous literature. Codogno et al. (2003), Bernoth et al. (2004), Pagano and Von Thadden (2004) find that liquidity risk has a limited or declining effect on EMU spreads. By contrast, Beber et al. (2009), and Manganelli and Wolswijk (2009) show that this determinant has a more prominent effect on the bond yield spread, particularly during periods of financial crisis. In order to explain these mixed results, another variable, the debt-to-GDP ratio, is added to equation 2. This macro economic fundamental, will help to explain the different effects of credit risk and liquidity risk varies between for each country, both prior to and during the sovereign crisis.
Table 6: CDS spreads and Bid-Ask spreads regressed against the bond yield spreads separately for nine Euro countries
Austria Belgium France Greece Ireland Italy the Netherlands Portugal Spain
CDS spread 0.005*** 0.005*** 0.003*** 0.010*** 0.010*** 0.007*** 0.003*** 0.009*** 0.008*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Bid-Ask spread -0.628*** -0.019 -10.173*** -0.098*** 0.350*** 2.085*** 2.288 0.259*** 0.218 (0.17) (0.19) (1.57) (0.02) (0.01) (0.23) (0.32) (0.02) (0.16) Constant 0.094*** 0.141*** 0.520*** 0.031*** -0.102*** -0.074*** -0.037*** 0.026*** -0.076*** (0.01) (0.02) (0.07) (0.01) (0.01) (0.02) (0.02) (0.01) (0.01) R² 0.985
Note: The model estimated is YieldSpreadit = c + β1CDSspreadit *CountryI + β2BidAskspreadit*Countryi + εit, Yield spread is the government bond spread of country I to the German bond. CDS spread is the CDS premium and Bid-Ask spread is the difference between the bid and the ask price divided by the mid-quote price of the government bond of country i. Total panel observations included: 10710. Note that all
entries in this table are estimated jointly. *** indicates significance at a 1% level
Table 7: CDS spreads and Bid-Ask spreads regressed against the bond yield spreads for nine Euro countries and two periods separately
Austria Belgium France Greece Ireland Italy the Netherlands Portugal Spain
CDS spread 0.005*** 0.008*** 0.006*** 0.009*** 0.008*** 0.007*** 0.004*** 0.007*** 0.009*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Bid-Ask spread 0.292 0.368* -4.319*** 3.360*** -1.298*** 2.030*** 2.006*** 2.346*** 1.127*** (0.21) (0.20) (1.66) (0.28) (0.19) (0.40) (0.32) (0.21) (0.32) CDS during crisis 0.000 -0.003*** -0.004*** 0.002*** 0.001*** 0.000 -0.004*** 0.002*** 0.001*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Bid-Ask during crisis -0.717* 0.299 4.471*** -3.550*** 1.547*** 0.021 1.303** -2.156*** -1.283***
(0.37) (0.27) (1.04) (0.28) (0.19) (0.39) (0.54) (0.21) (0.36)
Constant 0.052*** 0.052*** 0.232*** -0.161*** 0.096*** -0.084** -0.015 -0.061*** -0.092***
(0.02) (0.02) (0.07) (0.02) (0.02) (0.04) (0.02) (0.01) (0.01)
R² 0.987
Note: The model estimated is YieldSpreadit = c + β1CDSspreadit *CountryI + β2BidAskspreadit*Countryi + β3CDSspread*Countryi*Crisist + β4BidAskspread*Countryi*Crisist + εit .Yield spread is the government bond spread of country I to the German bond. CDS spread is the CDS premium and Bid-Ask spread is the difference between the bid and the ask price divided by the mid-quote price of the government bond of
Table 8: Average effect of the debt-to-GDP ratios on the determinants of yield spreads CDS spread 0.004*** (0.00) Bid-Ask spread 1.361*** (0.05) CDS spread*D/GDP 0.004*** (0.00) Bid-Ask spread*D/GDP -0.01*** (0.00) Constant -0.589*** (0.00) R² 0.969
Note: The model estimated is YieldSpreadit = c + β1CDSspreadit + β2BidAskspreadit + β3CDSspreadit*D/GDPi + β4BidAskspreadit*D/GDPi + εi Yield spread is the government bond spread of country I to the German bond. CDS spread is the CDS premium and Bid-Ask spread is the difference between the bid and the ask price divided by the mid-quote price of the government bond of country i. Total panel observations included:
10710. *** indicates significance at a 1% level
Table 8 shows to what extent the debt-to-GDP ratio has an additional effect on the determinants of the yield spreads. It becomes clear that, on average, when the fiscal policy of a EMU country deteriorates, i.e. when the debt-to-GDP ratio increases, the effect of the CDS spread on the yield spread also increases. By contrast, the effect of the Bid-Ask spread on the yields spreads becomes weaker when the to-GDP ratio increases. For instance, the debt-to-GDP ratio in Greece increases significantly during the sovereign crisis period; hence the effect of the Bid-Ask spread on the yield spread during the crisis becomes less pronounced. In addition, the effect of the CDS spread becomes more important in determining the yield spread during the same period. Similar results apply to Austria, Portugal and Spain. By contrast, the debt-to-GDP ratio of Ireland increases significantly in the crisis period but the effect of the Bid-Ask spread on the bond yield spread also increases. This is inconsistent with expectations created in Table 8. The fact that Ireland was the first country within the EMU to be bailed out might be a possible explanation for market participants to continuously require a higher return when liquidity risk increases.
VI. Conclusion
This paper presents an extended approach for analysing recent movements in EMU sovereign bond spreads. Since the escalation of the Greek debt crisis, interest rate spreads across EMU countries widened sharply, thereby threatening the stability of the financial system and the broader economy. Liquidity risk and credit risk are the two primary determinants of the movements in government bond spreads.
I obtain a number of interesting and novel findings that can be summarized as follows: I find that on average both credit and liquidity risk are important in explaining the widening of spreads in the EMU. In times of financial crisis, liquidity risk appears to have a significantly less pronounced effect on the spreads. Credit risk, on the other hand, is an increased determinant of sovereign spreads during the crisis period. When estimating the effects of credit risk and liquidity risk before and during the crisis for each country separately, I find that credit risk has a significant positive effect on the yield spread for every country in the crisis period. The other determinant shows mixed results: increased liquidity risk in the pre-crisis period leads to increased spreads in Belgium, Greece, Italy, the Netherlands, Portugal and Spain; while it leads to decreased spreads in France and Ireland and has no effect in Austria. In addition, the CDS spread during the crisis has increased explanatory power on the yield spread of Greece, Ireland, Portugal and Spain, and decreased power on Belgian, French and Dutch bonds. Furthermore, the Bid-Ask spread has a decreased effect on bond spreads during the crisis in Austria, Greece, Portugal and Spain; an increased effect in France, Ireland and the Netherland and no effect in Belgium and Italy. Finally, when the debt-to-GDP ratio of EMU countries increases, the effect of the CDS spread on the yield spread also increases. By contrast, the effect of the Bid-Ask spread on the yields spreads becomes weaker when the debt-to-GDP ratio increases.
These findings support the effect of market discipline in the EMU bond market. Results show that credit risk and, to a lower extent liquidity risk, is priced in the EMU sovereign bond markets, which implies that investors still distinguish between credit quality and liquidity of bonds of the different countries. Market participants demand a higher interest rate if there is a higher risk of default caused by unsustainable public finances. The recent developments in the fiscal policies of EMU-periphery countries has shown that these spreads rose sufficiently high to encourage governments to make substantial changes to their debt levels. In other words, the spreads of these countries have risen to a level where countries face a bankruptcy if they do not maintain or regain confidence of markets that they are fully committed to a permanent improvement in macro-fundamentals.
previously overlooked. Countries did not only lose their monetary independence but also a central bank to back their government bonds. This makes the EMU governments vulnerable to self-fulfilling panic over default. It is difficult to determine how the cycle of high government debts, economic contraction, increased sovereign spreads and contagion can be broken without the promise of substantial ECB support for sovereign debts. Because the Euro and the bond market are to a great extent interconnected with countries’ economy, politics and the ECB, it is difficult fully grasp the determinants of sovereign spreads.
VI. References
Amihud, Y. and H. Mendelson, 1986, Asset pricing and the bid-ask spread, Journal of
Financial Economics, 17, 375 – 410
Amihud, Y. and H. Mendelson, 1991a, Liquidity, maturity and the yields on U.S. government securities, Journal of Finance, 46, 1411–1426
Amihud, Y. and H. Mendelson, 1991b, Liquidity, asset prices and financial policy, Financial
Analysts Journal, 47, 56–66
Arghyrou, M. and A. Kontonikas, 2001, The EMU sovereign-debt crisis: Fundamentals, expectations and contagion, European Economy Economic Papers, 436, 1 – 45
Attinasi, M.G., Checherita, C. and C. Nickel, 2009, What explains the surge in Euro area sovereign spreads during the financial crisis of 2007-09?, ECB Working Paper, 1131
Barrios, S., Iversen, P., M. Lewandowska and R. Setzer, 2009, Determinants of intraeuro area government bond spreads during the financial crisis, European Economy Economic Papers, 388, 1-28
Beber, A., M. Brandt, K. Kavajez, 2006, Flight-to-quality or flight-to-liquidity? Evidence from the Euro area bond market, National Centre of Competence in Research Financial
Valuation and Risk Management, 309
Bernoth, K., J. von Hagen, L. Schuknecht, 2004, Sovereign risk premia in the European government bond market, ECB working paper series, 369
BIS, 1999, Market liquidity: Research Findings and selected policy implications, Bank for
International Settlements, CGFS Publications 11
Blanco, R., S. Brennan, and I. Marsh, 2005, An empirical analysis of the dynamic relation between investment-grade bonds and credit default swaps, Journal of Finance, 60, 2255-2281 Brandner, P., H. Grech, and K. Kazemzadeh, 2007, Yield Differences in Euro Area
Government Bond Markets – A View from the Market, Austrian Ministry of Finance, Working Paper No. 7
Brennan, M., A. Subrahmanyam, 1996, Market microstructure and asset pricing: On the compensation for illiquidity in stocks, Journal of Financial Economics, 41, 441 – 464 Brooks, C., 2008, Introductory Econometrics for Finance, Cambridge University Press
Bulow, J., and K. Rogoff, 1989b, Sovereign debt: Is to forgive to forget?, American Economic
Review, 79, 43-50
Calvet, L., and Adlai Fisher, 2004, How to Forecast Long-Run Volatility: Regime-
Switching and the Estimation of Multifractal Processes, Journal of Financial Econometrics 2, 49-83
Codogno, L., Favero, C. and A. Missale, 2003. Yield Spreads on EMU Government Bonds, Economic Policy, 18, 503-532
Copeland, L. and S. Jones, 2001, Default Probabilities of European Sovereign Debt: Market-based estimates, Applied Economic Letters, 8, 321-324
Datar, V., N. Naik and R. Radcliffe, 1998, Liquidity and stock returns: an alternative test,
Journal of Financial Markets, 1, 203-219
De Nicolo, G. and I. Ivaschenko, 2009, Global liquidity, risk premia and growth Opportunities, CESifo Working paper, 2598
Dötz, N. and C. Fischer, 2001, What can EMU Countries’ Sovereign Bond Spreads Tell Us About Market Perceptions of Default Probabilities During the Recent Financial Crisis?,
Deutsche Bundesbank Working Paper, 69
Deutsche Bundesbank, 2011, Sovereign yield spreads in de Euro area, Monthly Report June 2011
Duffie D., L. Pedersen, K. Singleton, 2003, Modelling sovereign yield spreads: a case study of Russian debt, Journal of Finance, 58, 119-159
Dunne, P., M. Moore and R. Portes, 2007, Benchmark status in fixed-income asset markets,
Journal of Business Finance and Accounting, 34, 1615-1634
Eaton, Jonathan, and Mark Gersovitz, 1981, Debt with potential repudiation: Theoretical and empirical analysis, Review of Economic Studies, 48, 289-309
ECB (2008), ‘Recent widening in Euro area sovereign bond yield spreads’ , ECB monthly
report
Flavin, T J, M G Limosani, 2007, Fiscal, monetary policy and the conditional risk premium in short-term interest rate differentials: an application of Tobin’s portfolio theory, International
Review of Economics and Finance, 16, 101–112
Gerlach, S., A. Schulz and G. Wolff, 2010, Banking and sovereign risk in the Euro area,
Deutsche Bundesbank Research Centre, 09/2010
Geyer, A., S. Kossmeier, and S. Pichler, 2004, Measuring Systematic Risk in EMU Government Yield Spreads, Review of Finance, 8, 171–197.
Hamilton, J., 1988, Rational-Expectations Econometric Analysis of Changes in
Regime: An Investigation of the Term Structure of Interest Rates, Journal of Economic
Dynamics and Control, 12, 385-423.
Hamilton, J., 1989, A New Approach to the Economic Analysis of Nonstationary Time Series and the Business Cycle, Econometrica, 57, 357-384.
Hamilton, J., 2005, Regime Switching Models, Palgrave Dictionary of Economics, 1-15 Haugh, D., P. Ollivaud and D. Turner, 2009, What Drives Sovereign Risk Premiums? An Analysis of Recent Evidence from the Euro Area, OECD Economics Department Working
Papers, 718
Hull, J., M. Predescu and A. White, 2004, The relationship between credit default swap spreads, bond yields and credit rating announcements, Journal of Banking and Finance, 28, 2789–2811
Jarrow, R., S. Turnbull, 1995, Pricing derivatives on financial securities subject to credit risk,
Journal of Finance, 50, 53-85
Longstaff, F., S. Mithal, E. Neis, 2005, Corporate yield spreads: Default risk or liquidity?,
Journal of Finance, 55, 2213-2253
Manganelli, S, G. Wolswijk, 2009, What drives spreads in the euro area bond market?,
Economic Policy, 24, 191-240
Merton, R., 1974, On the pricing of corporate debt: the risk structure of interest rates, Journal
of Finance, 29, 449-470
Min, H., 1998, Determinants of Emerging Market Bond Spreads - Do Economic Fundamentals Matter?, The World Bank, Policy Research Paper No. 1899.
Mody, A., 2009, From Bear Stearns to Anglo Irish: how euro zone sovereign spreads related to financial sector vulnerability, IMF Working Paper, WP/09/108.
Pastor, L., and R. Stambaugh, 2003, Liquidity risk and expected stock returns, Journal of
Policitcal Economy, 111, 642-685
Poterba, J.M. and K.S. Reuben, 2001, Fiscal news, state budget rules, and tax-exempt bond yields, Journal of Urban Economics, 50, 537–62
Sgherri, S. and E. Zoli, 2009, Sovereign Risks in the Euro Area: Are Markets concerned about solvency?, IMF Working Paper, 222
Zhang, B., H. Zhou and H. Zhu, 2005, Explaining Credit Default Swap Spreads with Equity Volatility and Jump Risks of Individual Firms, Bank of International Settlements Working
Appendix
Table 1: Descriptive Statistics of levels of Bond Yield Spreads
The spread between the yield of an EU government bond and the German government bond with a 10 year maturity. The sample periods are 1 November 2006 until 31 October 2009 (Period I) and 1
November 2009 until 24 May 2011 (Period II). All statistics are in basis points. Period I
Mean Median Maximum Minimum Std. Dev. Observations
Austria 31,17 18,10 133,50 2,80 31,59 783 Belgium 36,08 32,60 123,60 3,50 29,48 783 France 19,40 14,20 62,70 1,00 15,64 783 Greece 88,28 51,40 298,50 16,10 77,16 783 Ireland 66,16 28,30 262,40 -3,70 80,46 783 Italy 58,37 41,70 155,90 16,00 39,10 783 Netherlands 22,95 16,40 82,40 1,70 20,33 783 Portugal 49,51 35,60 161,80 10,90 38,03 783 Spain 34,23 21,60 123,30 0,00 30,88 783 Period II
Mean Median Maximum Minimum Std. Dev. Observations
Austria 37,90 37,40 83,70 17,10 8,94 407 Belgium 70,91 73,10 136,10 27,60 26,06 407 France 33,73 33,60 54,10 18,10 7,15 407 Greece 671,71 765,10 1401,10 137,30 304,38 407 Ireland 358,81 301,80 745,70 133,40 191,81 407 Italy 124,15 142,20 198,40 55,40 38,00 407 Netherlands 22,78 22,20 37,80 12,00 4,91 407 Portugal 288,87 305,60 697,80 47,70 162,88 407 Spain 155,65 176,60 290,70 46,50 66,61 407
Table 2: Descriptive Statistics of CDS Spreads
The spread is measured as the 10 year CDS premium on government bonds in US$. The sample periods are 1 November 2006 until 31 October 2009 (Period I) and 1 November 2009 until 24 May 2011 (Period
II).
Period I
Mean Median Maximum Minimum Std. Dev. Observations
Period II
Mean Median Maximum Minimum Std. Dev. Obsevations
Austria 84,39 85,07 115,82 52,01 12,66 407 Belgium 119,77 126,35 250,06 45,01 48,30 407 France 78,11 84,26 121,20 28,49 24,85 407 Greece 655,72 710,64 1445,16 153,10 291,52 407 Ireland 332,57 261,20 632,30 116,98 169,24 407 Italy 162,18 168,63 279,13 84,49 42,13 407 Netherlands 51,08 51,20 73,89 30,64 10,55 407 Portugal 301,55 297,12 600,05 68,59 146,48 407 Spain 202,56 216,23 364,42 83,63 68,27 407
Table 3: Descriptive Statistics of Bid-Ask Spreads
The Bid-Ask spread is calculated by the difference in the bid and ask price divided by the mid qoute price and multiplied by 100, denominated in US$. The sample periods are 1 November 2006 until 12
September 2008 (Period I) and 15 September until 24 May 2011 (Period II). Period I
Mean Median Maximum Minimum Std. Dev. Observations
Austria 6,50 4,99 12,68 0,71 3,18 783 Belgium 8,54 8,89 12,98 2,13 2,95 783 France 4,13 4,23 6,54 3,49 0,31 783 Greece 8,65 8,03 14,98 6,02 2,50 783 Ireland 10,09 9,99 15,00 3,98 3,09 783 Italy 9,83 10,01 10,66 7,32 0,71 783 Netherlands 6,46 6,11 12,16 1,89 3,57 783 Portugal 8,90 5,66 24,04 2,87 6,00 783 Spain 5,06 4,78 11,67 2,05 2,18 783 Period II
Mean Median Maximum Minimum Std. Dev. Observations
Table 4: Crisis period regressed against the bond yield spread for the nine Euro countries
Crisis starting 15 September 2008 Crisis starting 1 November 2009
Coefficient Standard Error Coefficient Standard Error
CRISIS 1.339*** 0.029 CRISIS2 1.509*** 0.03
R² 0.411 R² 0.441
Note: the model estimated is Yieldspreadit= crisis, and Yieldspreadit = crisis2. Crisis runs from 15/09/2008-24/05/2011 and denotes the global financial crisis period. Crisis2 runs from
1/11/2009-24/05/2011 and denotes the sovereign crisis. *** indicates significance at a 1% level.
Figure 2. Spreads of 10-year government bonds versus German Bund for Low Risk countries
Source: Bloomberg
Figure 3. Bid-Ask spreads
Figure 4. 10-year CDS spreads
Source: Bloomberg
Figure 8. Debt/GDP ratios for nine Euro countries
Figure 5. 10-year government bond yield spreads and 10-year CDS spreads
Note: The left vertical axis shows the CDS spread in US$ and the right vertical axis shows the yield spread in basis points
Figure 6. Yield spread and Bid-Ask Figure 7. Yield spread and CDS
