Growth-indexed Securities as a Sovereign Financing Tool for the Eurozone

Index

Abstract English

Abstract German

Index for Illustrations and Tables

List of Abbreviations

Executive Summary

1 Introduction

2 The EMU and the aftermath of the global financial crisis

3 The need for cyclical stabilization in the EMU

4 Sovereign debt structure for crisis prevention
4.1 Public debt structures in emerging and advanced economies
4.2 Sovereign versus corporate debt structures
4.3 Instruments to render debt structures less crisis prone
4.3.1 Explicit seniority in sovereign debt
4.3.2 Real indexation of public debt

5 GDP-linked bonds
5.1 Variants of growth-linked bonds
5.2 Pricing of GDP-linked bonds
5.3 Benefits of GDP-indexed securities
5.3.1 Benefits for issuers
5.3.2 Benefits for investors
5.3.3 Benefits for the global economy and the global financial system
5.4 Concerns, issues and obstacles
5.4.1 Moral hazard
5.4.2 Accuracy and timeliness of GDP data
5.4.3 Uncertainty about sufficient liquidity
5.4.4 Pricing difficulties
5.4.5 Long-term benefits versus short-term costs

6 GDP-indexed bonds as a stabilization tool for the EMU
6.1.1 Simulation of the diversification effects
6.1.2 Simulation of the stabilization effects
6.1.3 Results
6.1.4 Sensitivity analysis

7 Summary and conclusion

8 Bibliography

9 Appendix

Abstract English

The euro area members have delegated monetary policy to a communitarian central bank, while fiscal and macroeconomic policy strictly remain under national responsibil­ity. In a currency union, the governments issue debt in a currency whose sup­ply they don’t control. As a result, vulnerability to changing market sen­timent and the inability to use counter-cyclical fiscal policies for economic stabili­zation tends to produce more pronounced booms and busts.

GDP-indexed securities have been proposed in the academic litera­ture as an instrument to stabilize sovereign debt dynamics across the Eurozone. This Master’s Thesis evalu­ates whether the introduction of GDP-indexed sovereign bonds would provide substan­tial stabilization to the Eurozone in the event of a macroeconomic shock. The results confirm that GDP-indexed sovereign debt does have substantial stabi­lizing effects on the debt-to-GDP ratios of its issuers. However, the stabilizing capacity of GDP-indexed bonds seems insufficient to protect the Eurozone’s most vulnerable members from li­quidity crises. Nevertheless, GDP-indexed instruments may well func­tion as a contrib­utor to the stability of the Eurozone, along with existing instruments, and potentially with a still-to-be-introduced EMU-wide risk-sharing mechanism.

Abstract German

Die Euroländer haben die Gestaltung der gemeinsamen Geldpolitik einer kommuni­tären Zentralbank übertragen, während Fiskal- und Wirtschaftspolitik streng in die Ver­ant­­wortung der nationalen Regierungen verblieben.

In einer Währungsunion werden Staatsschulden in einer Währung finanziert, die von ei­ner gemeinschaftlichen Zentralbank kontrolliert wird. Dadurch wird die Nutzung einer an­ti­zyklischen Fiskalpolitik zur Stabilisierung der Wirtschaft behindert, was zu einer stär­keren Ausprägung von Aufschwüngen und Krisen in den Mitgliedsstaaten einer Währungs­­uni­on führen kann.

In der akademischen Literatur wurden BIP-indexierte Wertpapiere als Instrument zur Stabilisierung der Schuldendynamik in der Eurozone vorgeschlagen. Die Masterarbeit untersucht, ob die Einführung von BIP-indexierten Staatsanleihen einen substantiellen Beitrag zur Sta­bilisierung der Eurozone im Falle eines makroökonomischen Schocks leisten könnte. Die Ergebnisse bestätigen, dass BIP-indexierte Staatsanleihen eine er­hebliche stabili­sierende Wirkungen auf die Schuldenquoten ihrer Emittenten ausüben. Allerdings scheint die Stabili­sierungskapazität solcher Instrumente alleine nicht auszu­rei­chen, um die Mitglieder der Eurozone nachhaltig vor einer Liquiditätskrise zu schüt­zen. Den­noch können BIP-indexierte Wertpapiere einen Beitrag zur Stabilität der Euro­zone leis­ten, allerding nur im Zusammenspiel mit den bestehenden Instrumenten und möglicherweise mit einem EWU-weiten zyklischen Risikoversicherungsmechanis­mus.

Index for Illustrations and Tables

Figure 1: First-in-time seniority debt versus conventional debt: marginal borrowing cost

Figure 2: Selected Eurozone countries – actual nominal GDP growth and stock market performance (selected years)

Figure 3: Selected Eurozone countries – Standard Deviation of 2001 to 2013 growth rates

Figure 4: Southern Eurozone members – nominal GDP baseline and actual nominal GDP (normalized to 100)

Figure 5: Ten-year yields on sovereign bonds, January 1993 to February 2012

Figure 6: Yields on 10-year sovereign bonds, October 2009 to June 2012

Figure 7: Ireland and Slovenia – nominal GDP baseline and actual nominal GDP (normalized to 100)

Figure 8: Northern and central Eurozone members – nominal GDP baseline and actual nominal GDP (normalized to 100)

Figure 9: Debt-to-GDP ratios in 2007 and 2014 assuming 0 percent and 50 percent of total debt being financed through indexed borrowing

Figure 10: Selected Eurozone countries – net lending (+) or net borrowing (-) excluding interest (in billion euros)

Figure 11: Change in debt-to-GDP ratios in 2014 resulting from the issuance of 50 percent indexed debt in 2007 (in percentage points)

Figure 12: Increases in debt-to-GDP ratios between 2007 and 2014 - drivers

Figure 13: Relation between the severity of the output shock and accumulated primary balances between 2007 and 2014

Figure 14: Relation between the increase in debt-to-GDP ratios from 2007 to 2014 and average interest on newly issued 10-year bonds in the period under review

Figure 15: Sensitivity analysis - risk premium

Figure 16: Sensitivity analysis - percentage of indexed bonds

Figure 17: Sensitivity analysis - debt-to-GDP ratios in 2007 and 2014 - comparison of continued indexed debt issues after 2007 versus simulation (in percentage points)

Figure 18: Sensitivity analysis - change in debt-to-GDP ratios in 2014 - comparison of continued indexed debt issues after 2007 versus simulation (in percentage points)

List of Abbreviations

illustration not visible in this excerpt

Executive Summary

The 2008 financial crisis triggered the most severe global recession since the Second World War, leading economies across the globe to the brink of disaster, and still raising doubts upon the sustainability of the European Economic and Monetary Union (EMU) in its present design.

The Eurozone members have delegated monetary policy to a communitarian central bank, while all aspects of fiscal and macroeconomic policy strictly remain under na­tional responsibility. However, the 2008 crisis clearly illustrated that the Eurozone lacks adequately developed automatic stabilizers, and that the employment of classic counter-cyclical fiscal policy at the part of the euro area governments seems insuffi­cient to counter idiosyncratic shocks. In a currency union, the individual governments have to issue debt in a currency whose supply they don’t control, a fact that seems to funda­mentally impede their ability to fi­nance sovereign deficits. As a re­sult, the members of a currency union face is­sues typical for emerging econo­mies, where a pro­nounced vulner­ability to changing market sentiment and the inability to use Keynesian fiscal policies for economic stabili­zation tends to produce more dis­tinct booms and busts.

After the 2008 crisis had exposed the flaws of the Eurozone, the Union reworked its governance framework, also introducing EMU-wide financial backstops such as the Eu­ropean Stability Mechanism (ESM) and the Banking Union. However, these reforms were accompanied by a reappearing debate about their sufficiency in order to effectively contain future output shocks. In the course of this debate, a variety of communitarian risk-sharing mechanisms have been proposed. These instruments would transfer a sub­stantial part of cyclical stabili­zation to the European level, and thus allow the euro area mem­bers to focus their fiscal policies on structural aspects.

Further authors propose instruments that can render public debt structures less crisis prone, such as the introduction of seniority in sovereign debt issued by EMU countries, or the Eurozone-wide introduction of GDP-indexed debt. The latter would stabilize sov­er­eign debt dynamics, reduce the likelihood of liquidity crises in the event of an eco­nomic downturn, and moreover enable sovereigns to smooth national economic output through counter-cyclical fiscal policies, rather than being forced into damaging pro-cy­ al measures.

The investors community would benefit from a broad introduction of growth-linked se­curities in the sense that such instruments would allow their holders to take well diversi­fied equity-like stakes in the future growth prospects of individual countries. In addi­tion, many authors view GDP-indexed bonds as a public good that generate systemic ben­efits by reducing the likelihood of debt crisis and sovereign default in general. How­ever, the above benefits come with a range of concerns, such as the possibility of moral haz­ard, issues related to the accuracy and timeliness of GDP data, pricing difficulties, and uncertainty about sufficient liquidity at the introduction of such instruments.

This Master’s Thesis evaluates whether the introduction of GDP-indexed sovereign debt would provide substantial stabilization to the Eurozone in the event of a macroeconomic shock. To this end, a simulation is performed that quantitatively assesses the stabili­za­tion effect that might have resulted from the EMU-wide adoption of GDP-in­dexed sov­ereign debt prior to the 2008 crisis.

The results confirm that GDP-indexed public debt does have substantial stabilizing ef­fects on the debt-to-GDP ratios of its issuers. Under the assumption that the Eurozone governments had raised 50 percent of their sovereign debt stock in the form of GDP-in­dexed bonds, particularly the southern Eurozone periphery shows substantially im­proved debt-to-GDP ratios, ranging from 5 percentage points in Italy and Spain, 6 per­centage points in Portugal, 7 percentage points in Cyprus to 27 percentage points in Greece.

However, at least for the assumptions made in this simulation, the stabilizing capacity of GDP-indexed securities seems insufficient to prevent the Eurozone’s most vulnerable members from plunging into liquidity crises. Nevertheless, GDP-indexed instruments may well function as a contributor to the stability of the Eurozone, along with existing instruments like rigorous macroeconomic and fiscal governance, as well as with finan­cial backstops like the ESM and the Banking Union, and potentially with a still-to-be-introduced EMU-wide risk-sharing mechanism.

1 Introduction

Triggered by the US subprime mortgage crisis, the 2008 financial crisis induced the most severe global recession since the Second World War. In addition to leading the global economy to the verge of catastrophe, the crisis also casts doubts over the func­tioning of the Economic and Monetary Union (EMU) and its com­mon currency, the euro.

The Eurozone members have delegated all aspects of monetary policy to a com­muni­tarian institution, the European Central Bank (ECB), while the areas of fiscal and mac­roeconomic policy strictly remain under national responsibility, only controlled by a set of basic fiscal governance rules formal­ized in the Stability and Growth Pact (SGP). Thus, and in contrast to most currency unions, the stability of the Eurozone fully de­pends on its capacity to cope with asymmetric shocks through the workings of auto­matic stabi­lizers and through counter-cyclical fiscal policy. However, evidence suggests that the Eurozone lacks adequately developed automatic stabilizers that could compen­sate for the loss of national monetary pol­icy as an instrument to counter idiosyncratic shocks. Moreover, the incongruence between the political and the economic cycles made most Eurozone members struggle with the implementation of classic Keynesian fiscal policies, particularly in times of favorable economic conditions.

In addition to the above political obstacles, the entry into a currency union seems to fundamentally re­strict the ability of governments to finance sovereign deficits. Govern­ments have to issue debt in a currency whose supply they don’t control in a di­rect man­ner. This mechanism increases also the Eurozone’s vulnerability to changing mar­ket sentiment and makes its members face issues that are typical for emerging econo­mies, where the inability to use Keynesian fiscal policies for eco­nomic stabiliza­tion tends to produce more pronounced booms and busts.

After the 2008 cri­sis had exposed the flaws of the Eurozone, the Union sub­stan­tially re­formed its fiscal and economic governance framework and introduced the Eu­ropean Stability Mechanism (ESM) and the Banking Union in order to prevent future asymmet­ric shocks from spreading across its member states. These reforms were accompanied by a reappearing debate on the necessity and feasibility of an EMU risk-sharing mecha­nism that would lift a substantial part of cyclical stabili­zation to the European level, and so allow the EMU mem­bers to focus their fiscal policies on structural aspects. In the course of this debate, a wide variety of differ­ent risk sharing mechanisms have been proposed, for instance the institution of a substantial budget at EMU level, an Eurozone wide un­em­ploy­ment insurance, or a cyclical insurance scheme that would man­age trans­fers between the member states depending on their individual busi­ness cy­cle positions.^[1]

Other authors propose to address the specific financial vulnerabilities that the members of a cur­rency union typically face through the use of instruments that can render public debt structures less crisis prone. Such in­novation could include contractual aspects, as for in­stance the intro­duction of seniority in sovereign debt issued by EMU countries, or the Euro­zone-wide introduction of GDP-indexed sovereign bonds.^[2]

The authors argue that GDP linked debt would stabilize sovereign debt dynamics, raise the level of sustainable debt, and therefore reduce the likelihood of liquidity crises or sovereign default in the event of an economic downturn. Their capacity to function as “automatic stabilizers” would furthermore enable sovereigns to smooth national eco­nomic output through counter-cyclical fiscal policies, rather than being forced into dam­aging pro-cyclical measures. Some variants of GDP-linked bonds exert their stabilizing properties through a reduction of the issuer’s total debt rather than through lower inter­est payments, which might spe­cifically help the euro-area members in stabilizing their debt-to-GDP ratio at the level re­quired by the SGP.

The investors community would benefit from a broad introduction of growth-linked se­curi­ties in the sense that such instruments would allow their holders to take well diversi­fied equity-like stakes in the future growth prospects of individual countries. Finally, many authors view GDP-indexed bonds as a public good, gen­erating systemic benefits by reducing the likelihood of debt crisis and sovereign default in gen­eral.

However, the above benefits come with a range of concerns, such as the possibil­ity of moral hazard, issues related to the accuracy and timeliness of GDP data, pricing diffi­culties, and uncer­tainty about sufficient liquidity at the introduction of such instru­ments.

This Master’s Thesis evaluates the most common variants of GDP-linked debt in­stru­ments for their capacities to foster cyclical stabilization in the Eurozone.

The Thesis is structured as follows: Chapter 2 examines the impacts of the 2008 eco­nomic and finan­cial crisis on the economies that comprise the Eurozone, the institu­tional and procedural weaknesses exposed by the crisis, and the resultant re­form of the EMU’s fiscal and eco­nomic governance framework. Chapter 3 con­denses the arguments on whether the proper functioning of the European single-currency area requires a fi­nancial stabiliza­tion mechanism at the federal level, and provides a brief overview of the various risk-sharing tools proposed in the aca­demic literature. Both chapter 2 and 3 represent summaries of an earlier work of the author.^[3]

Chapter 4 investigates how sovereign debt structures impact financial stability. Chapter 5 then analyses the literature on growth-linked debt instruments, lays out the most common variants, their benefits, as well as concerns, issues and obsta­cles. In chapter 6, a model is devel­oped and applied in order to evaluate the size of the stabilization effect that GDP-linked sovereign debt would have generated during the 2008 crisis. A summary and conclusion is then presented in Chapter 7.

2 The EMU and the aftermath of the global financial crisis

The global eco­nomic and finan­cial crisis of 2008 materialized in the European single-currency area as three distinct crises that are interconnected with each other in several different ways. At the onset of the 2008 crisis, the euro area has been hit by a severe banking crisis. The subsequent bailout of national banking sectors left the countries of the southern Eurozone periphery with dangerously elevated levels of sovereign debt, and in some cases, led to sovereign debt crises, with rising in­terest rates challenging the capacity of the respective governments to fund their sovereign debt. As a re­sult of ex­tensive cross-border holdings of public bonds, the potential default of some sover­eigns threatened the solvency of the entire banking system of the currency area. Fiscal auster­ity measures forced upon the distressed euro area members by the capital mar­kets fur­ther dampened GDP growth and thereby exacerbated the debt crisis. In addi­tion, the Eu­rozone increasingly suf­fered from a macroeconomic crisis, as converging interest rates had initiated a credit driven boom in the southern periphery that had led to declin­ing rel­a­tive competitiveness and substantial current account deficits, which in turn fur­ther weakened the economic environment of these countries. In a nutshell, the 2008 eco­nomic and finan­cial crisis left the Eurozone with high, in some cases even worryingly elevated debt-to-GDP ratios and substantial mac­roeco­nomic imbal­ances.^[4]

However, austerity measures in combination with the ongoing decontamination of bank balance sheets will likely dampen eco­nomic growth across euro area in the up-coming decade and significantly impede its member’s capacity to lower public debt to more sustainable lev­els. Moreover, the distribution of GDP growth will likely continue to re­main unbalanced, as in the Eurozone periphery, fiscal auster­ity will be accompanied by delever­aging of the highly indebted private sector, which will pose a further drag on consumer demand and consumer price inflation. However, low relative inflation will impede the adjustment of relative prices as a means to re-establish competitiveness against the core euro area countries.^[5]

Also the implementation of structural reforms in the euro area periphery, though needed to restore competitiveness and growth in the long run, will come at the cost of consider­able temporary unemployment, while fiscal relaxation as a means to stimulate the econ­omy is no longer available due to he pressure of financial markets. Consequently, the 2008 crisis left the Eurozone periphery with prospects for prolonged slow growth, per­sistent unemployment, and no national policy lever left to combat this phenomenon in an effective way.^[6]

The 2008 crisis had exposed the fundamental shortcomings of the EMU govern­ance framework: inadequate fiscal and macroe­conomic surveillance and supervi­sion, as well as ineffec­tive mechanisms in order to enforce fiscal and macroe­co­nomic discipline at the part of the individual member states.

By mid of 2010, after the euro area members had granted bilateral loans in order to rescue Greece from immediate bankruptcy, the European Financial Stability Mechanism (EFSM) and the European Financial Stability Facility (EFSF) were established with the aim of providing liquidity assistance to distressed Eurozone countries. Two years later, the permanent ESM replaced the functions of these provisional bailout funds.

In addition, as the original Stability and Growth Pact (SGP)^[7] failed to secure fiscal disci­pline in the Eurozone, the EMU leaders implemented the most wide-ranging set of gov­ernance reforms since the establishment of the common currency in sev­eral consecutive steps, starting in June 2010 with the implementation of the Eu­ro­pean Semester, fol­lowed in November 2011 by the so-called Six-Pack^[8] and in May 2013 by the so-called Two-Pack.^[9] While the European Semester predomi­nantly inte­grates and synchronizes ex­isting fiscal and macroeconomic governance proce­dures across the Union, the Six-Pack adds substantive provisions to the fiscal gov­ernance framework of the euro area, amending the preven­tive a well as the cor­rective arm of the SGP. The Two-Pack com­ple­ments the Six-Pack through further detailed regulations on fiscal and macroeco­nomic surveillance. In addition, the Euro Plus Pact and the Fiscal Compact, constituted by a set of intergovern­mental treaties between the euro area members, extend the scope and depth of the Euro­pean governance system. Finally, the Banking Union lifts the re­spon­sibility for surveillance and resolution of systemically relevant European banks to the Eu­ro­pean level, thereby trying to prevent future asymmetric shocks from spreading across the Union’s member states.

The author summarizes the academic literature on the effects of the 2008 crisis on the EMU and the subsequent reform efforts in an earlier work.^[10]

3 The need for cyclical stabilization in the EMU

The various gov­ernance reforms that have been introduced in the EMU since 2010 aim at strengthening fiscal and macroeconomic disci­pline across the Union’s member states. However, the EMU leaders failed to address a fundamental design problem of the Euro­pean currency area: The Eurozone members have delegated all aspects of monetary po­licy to a communitarian national bank, while, in contrast to most currency unions, fis­cal and macroeconomic policy strictly remained in national hands. Lacking national mone­tary policy as an instrument to combat asymmetric output shocks, the stability of the Eu­ro­zone therefore entirely depends on the workings of automatic stabilizers and classic Keynesian fiscal policies.

However, evidence suggests that euro area lacks adequately developed automatic stabi­lizers that could compensate for the loss of national monetary policy. Moreo­ver, the en­try into a currency union seems to restrict in a fundamental way the ability of governments to finance their public deficits. Governments have to issue debt in euros, a cur­rency whose supply they do not control. Distrust of financial agents might push the mem­bers of a currency union into a bad equilibrium, where rising bond yields and the result­ing financial distress will progressively amplify each other. This phenomenon can lead into severe liquidity crises or even trigger sovereign default (“self-fulfilling sol­vency crisis”). Consequently, the members of a currency union face issues typical for emerg­ing economies, where high vulnera­bility to market sentiments prevents the effec­tive deployment of Keynesian fiscal policies for economic stabilization, and therefore pro­duces more pronounced booms and busts.

Supporters of efficient market theory argue that this increased vulnerability will help to avert unsustainable fiscal conduct on the part of Eurozone members. How­ever, evidence from the 2008 crisis suggests that financial markets can be driven by extreme euphoria, where economic agents systematically misjudge financial risks, or by irrational fear in times of impending disaster.

De Grauwe (2011) argues in this context that particularly the residents of Euro­pean countries view the calming effects of automatic stabilizers and counter-cycli­cal fiscal policies on booms and busts as important social achievements, and con­sequently, weak­ening those instruments might substantially reduce public support for the European sin­gle-cur­rency area in the long run. This is particularly relevant for the Eurozone’s south­ern periphery, where this increased vulnerability against market sentiments might im­pede the govern­ment’s efforts to re­store competitive­ness after the 2008 crisis. In a sin­gle currency area, these countries will have to reduce their internal price levels thorough deflationary mac­roeconomic policies, which will dampen growth or might even lead into recession. As a result, auto­matic sta­bilizers will inflate debt to GDP ratios, which will fuel concerns about debt sustainability and might push the respective countries into bad equilibria as a result of their efforts to restore competitiveness.^[11]

The union-wide obligation to fiscal austerity, which was brought about by the Eu­ropean governance reforms, further adds to the challenges for southern European economies by reducing demand growth and inflation also at their main trading partners.

Moreover, the design of the Union seems to obstruct also the execution of effi­cient pol­icy re­sponses to more synchronous crises. Lacking an EMU-wide fiscal coordination mecha­nism, the classical Keynesian policy lever will likely deliver suboptimal results in the Eurozone, mainly re­sulting from the fact that the indi­vidual member governments have strong in­centives to “free ride” on the stimuli of their main trading partners. There­fore, Pisani-Ferry et al (2013) argue that the ef­fects of individual policy re­sponses will likely be substantially weaker than a cen­trally coordinated stimulus, a fact that attributes even greater importance to mon­etary policy as the main instrument to address synchro­nous down­turns in the euro area.^[12]

However, evidence from the current crisis suggests that monetary policy can be an in­sufficient tool in stabilizing the euro area economies during a very severe downturn.^[13]

In summary, the evidence clearly points towards the fact that the members of the EMU, a currency union without any substantial fiscal integration, face considera­ble exposure to the adverse effects of area-wide and asymmetric shocks. Feld and Osterloh (2013) find in this context that in other monetary unions (US states, German Laender, Canadian prov­inces, Swedish regions), the fiscal channel sub­stantially contributes to economic sta­bilization (be­tween 13 and 54 percent).^[14]

Nevertheless, the debate on whether some sort of fed­eral stabilization mechanism is re­quired to sustain the proper functioning of Eurozone over the long run re­mained highly controversial: While some authors argue that the euro area needs to continue its devel­opment towards a deeper political union, endowed with a sub­stantial budget at the fed­eral level, others claim that the institution of further fiscal transfers between the individ­ual countries would threaten the political backing for the Union in most member states. The latter therefore tend to propose to further tighten fiscal and macroeconomic govern­ance across the euro area. Finally, a third school recom­mends sustaining or even loos­ening the Eurozone’s current institu­tional framework. The EU leaders themselves joined this debate, supporting the institution of a substantial fiscal capacity at the federal level as part of a compre­hensive reform of the Union’s procedural and institutional frame­work.^[15]

4 Sovereign debt structure for crisis prevention

Already prior to the 2008 financial and economic crisis, the fact that the structure of public borrowing can exert significant influence on the economic perfor­mance of highly indebted countries, impacting both the frequency of debt crises and the disruption those crises may cause when they strike, was widely recognized in the academic literature. However, the academic debate on government debt structures mostly focused on finan­cial innovations that may facilitate the resolution of debt crises, rather than on their pre­vention.^[16]

Addressing possible inefficiencies in existing debt structures through financial in­nova­tion requires some understanding of their underlying causes. Borensztein et al (2004) summa­rize the two dominating views in the respective academic debate as follows.^[17]

The first view assumes that the current debt instruments and structures can be ex­plained by historical circumstances and have persisted due to inertia. Conse­quently, financial innovation could increase the efficiency of public debt struc­tures. However, the follow­ing reasons make any change to the status quo difficult to achieve:

- The success of financial innovation often depends on its simultaneous imple­menta­tion by many contracting parties. Considerable price uncer­tainties and concerns about limited liquidity of the new instrument typi­cally make indi­vidual investors demand a ‘novelty’ premium. Further­more, the issuer of a new finan­cial instrument will disregard potential ben­efits for other market partici­pants, particu­larly long-term social benefits that would result from the institution of a new asset class.
- As replication of new financial instruments is relatively easy, a private finan­cial in­stitution would incur substantial development effort, but would soon loose its monopoly position over the provision of the instrument.
- Liquid secondary markets for financial instruments, which enable inves­tors to effi­ciently diversify their portfolios, require a set of verifiable and broadly ac­cepted standards (for instance for the event that trigger cash payments to the in­vestors).
- Finally, the financial markets might misread the introduction of a new finan­cial in­strument as a sign of weakness or lack of commitment to sound financial poli­cies on the part of the debtor (signaling).

The alternative view assumes that prevailing market practices are an efficient re­sponse to the possibility of moral hazard on the part of the governments. The high cost of debt crises may be the only way to discipline borrowing governments and discourage de­faults. Thus, the seemingly inefficient debt structures, which prevail in developed econ­omies, are needed in order to reduce moral hazard on the part of policy-makers. Simi­larly, the typically crisis-prone debt structures of developing markets, often biased to­wards foreign-currency-denominated debt or towards short-term debt, are viewed as a symptom rather than a cause of inadequate mon­etary and financial policies pursued in the respective countries. Consequently, this view considers any attempt to reform the international fi­nancial architecture with­out addressing fundamental distortions as inade­quate or even counterproductive.

4.1 Public debt structures in emerging and advanced economies

Borensztein et al (2004) review public debt structures of emerging and advanced econ­omies over the past decades and find that those mostly differ with regards to the fol­lowing three aspects:^[18]

- Average debt-to-GDP ratios of emerging market countries tend to lie well be­low those of advanced economies.
- Compared to mature economies, emerging market countries rely to a far greater ex­tent on debt denominated in a foreign currency and issued under a for­eign ju­ris­diction.
- While the structures of external debt of emerging and advanced economies are rel­atively homogeneous (typically fixed-rate bonds with long maturi­ties), the structures of their domestic debt widely differ in maturity, cur­rency denomi­na­tion and in the prevalence of inflation-indexed securities. In contrast to mature economies, emerging market countries have to meet their public financing needs mostly via external markets, while their debt issued domestically mainly takes the form of foreign-currency denomi­nated (or indexed) debt, as well as short-term and inflation-indexed debt.

However, this dependence on short-term (or debt indexed to short-term domestic inter­est rates) and foreign-currency debt increases a country’s vulnerability to changing mar­ket sentiments, where concerns about the solvency of a country can trigger a vicious cir­cle in a the sense that increasing doubts about debt sustaina­bility in turn reinforce rising vulner­ability. Consequently, these risky forms of debt can amplify the economic cycle, in­crease the probability of crises, and render crises more difficult to manage.^[19]

Various studies have confirmed the close relationship between short-term debt and ex­posure to international financial crises.^[20]

Rodrik and Velasco (1999) highlight in this context the role of capital flows. The au­thors show that the ratio of short-term debt exposure to reserves is a robust pre­dictor of debt crises, which substantially increases in severity in case capital flows reverse.^[21]

Borensztein et al survey the respective literature in depth and argue that this em­pirical association may result from increased vulnerabil­ity to debt rollover crises, where exter­nal factors, such as bad news related to the sol­vency of a country, im­pact the willingness at the part of the creditors to roll over their claims. Conse­quently, a high fraction of short-term debt will put the respective govern­ment at the mercy of self-fulfilling credi­tor panics. Equally important, short-term or floating-rate debt can get a country trapped in a bad equilibrium, as changes in the country’s creditworthiness will quickly affect its interest bill. High interest pay­ments will in turn increase the probability of the respec­tive country defaulting on its obliga­tions, which in turn will make investors ask for a higher default risk premium. As a re­sult, in emerging economies, economic downturns are typically coupled with increasing interest rates, which consequently further reduce the scope for countercyclical fiscal policies.^[22]

Also the exposure resulting from significant levels of debt denominated in (or in­dexed to) a foreign currency has also been evident in several recent crises, where local cur­rency depreciation often led to sharp increases in government debt-to-GDP ratios, am­plifying the effects of the initial downturn and reducing the ability of the government to mitigate the disruption in the private and public sector.^[23]

Borensztein et al attribute the risky debt structures prevailing in most emerging econo­mies mainly to the lack of credibility that the debtor’s monetary and fiscal policies en­joy among potential creditors, and to a lesser extent to factors such as the nature of do­mestic investors, specific characteristics of domes­tic financial reg­ulation and the debtor’s economic size. Lack of credibility makes creditors antici­pate that the govern­ment will either directly default on its obligations, or expro­priate its creditors in an indi­rect way through inflation (in case debt is denomi­nated in local currency). This phe­nomenon is of particular sig­nificance at the on­set of a crisis, as governments pres­sured by financial markets tend to shift the composition of their debt toward instruments that are more prone to crises, short-term and foreign-currency denominated debt. The is­su­ance of short-term debt di­lutes the out-standing long-term debt and is therefore usu­ally preferred by inves­tors in an environment of lacking trust in debtor’s monetary and fiscal policies. Moreover, in­vestors may expect short-term debt to exercise a discipli­nary ef­fect, as governments de­viating from desirable policies might face higher interest rates or even a rollover crisis. Consequently, short-term debt may be viewed as a symptom, ra­ther than the cause of a looming crisis.^[24]

Borensztein et al furthermore argue that episodes of hyperinflation can lead to a sub­stantial change in the structure of governments liabilities towards foreign-cur­rency de­nominated or indexed debt, a phenomenon that often persist for decades after macroe­conomic and fiscal stabilization efforts have succeeded.^[25]

Among the minor factors impacting debt structures in emerging economies, the authors stress the stabilization effects of a large base of domestic investors, and particularly of prefunded (private or public) pension funds, which tend to accu­mulate significant do­mestic savings that are available for investment in domestic government debt.^[26]

4.2 Sovereign versus corporate debt structures

Borensztein et al (2004) also analyze the debt structures of sovereigns in contrast to those of corporations and find three main points:^[27]

- First and most importantly, corporations finance their assets to a considera­ble ex­tent through equity and equity-like instruments, such as convertible bonds, while sovereigns rely exclusively on debt.
- Second, while corporations heavily use secured debt, as well as different seniori­ties within the remaining uncollateralized debt, sovereigns typically issue ra­ther little collateralized debt (Borensztein et al quote a share below 10 percent of to­tal debt), and moreover, do not prioritize their unsecured debt. Secured corpo­rate debt, which typically makes up a significant por­tion of the corporate debt stock in developed economies, can be classified according to its priority in the event of a bankruptcy, as secured debt, ordi­nary unsecured debt, subordinated debt, pre­ferred stock and common stock. While secured debt pro­vides the debtor with the title to a specific asset owned by the respective company, the remaining liability classes are serviced only in case the higher-ranking class has been re­paid in full. In contrast, secured sovereign claims are typically collateralized by future cash receipts, which have to occur outside the direct control of the re­spec­tive gov­ernment, rather than by domestic tax revenues or assets.
- Third, corporate bond contracts often contain covenants, which restrict fu­ture fi­nancing decisions of the respective company, for example limiting total debt levels or the issuance of debt of the same or a higher seniority. Clauses of this type, which aim at protecting private creditors from severe dilution of their claims, are gener­ally lacking in securities issued by sover­eigns. However, Borensztein et al argue that the conditions that are typi­cally attached to financing provided by international financial institutions, such as limits on the fiscal deficit or on total external debt, could be inter­preted as analogous to covenants.^[28]

4.3 Instruments to render debt structures less crisis prone

Borensztein et al argue that contractual innovations, particularly the introduction of seniority in sovereign debt, could potentially help to promote safer debt struc­tures, discourage over-borrowing, and lower interest costs for moderately indebted countries.^[29]

4.3.1 Explicit seniority in sovereign debt

Senior debt is widely used at the corporate level, where the presence of this in­strument can serve as a disciplining device in order to prevent a firm’s manage­ment from over-borrowing. With the issuance of new debt, the new capital re­ceives a share of the debt recovery value that corresponds to its share in the total outstanding debt. Therefore, new debt leads to a dilution of the claims held by the initial creditors, and consequently, poses a capital loss upon them. In anticipation of such an event, investors will either demand higher interest rates in the first place, or refuse to lend altogether, which makes the possibility of debt dilution eventually backfire on the debtors. However, if the origi­nal investors were senior in a sense that, in the event of bankruptcy, the claims of all creditors were served in the order in which the debt was issued, the dilution problem could be elimi­nated and creditors were less anxious about subsequent debt issues.

Borensztein et al claim that this instrument would lower incentives for the gov­ernments to issue debt in a form that is difficult to dilute, such as short-term for­eign-currency debt, and could therefore benefit public debt structures of emerging markets. At moder­ate levels of debt, borrowing cost would decrease, as creditors would not have to de­mand compensation for potential debt dilution, while high levels of debt would auto­matically tighten a gov­ernment’s access to the financial markets and consequently re­duce incentives to over-borrow. These benefits are most obvious when governments are biased toward excessive borrowing.^[30]

However, highly indebted countries might suffer from a structured seniority re­gime, particularly in the event of large adverse shocks, which would increase their financing needs significantly. New debt would rank junior to all earlier claims, and therefore the new creditors would demand higher compensation, or could even refuse to lend at all if the risk of default were sufficiently high.

Borensztein et al illustrate this effect in Figure 1. and represent the minimum re­cov­ery value of a particular country’s debt, as well as its maximum sustainable level of debt. Under a seniority debt regime, the investors would consider an extra debt unit issued at debt levels lower than basically risk free, while beyond , the cost for new debt rises first slowly and then at an exponential rate, reaching in­finity as the actual debt level approaches the maximum sustainable debt level . Un­der a conventional debt struc­ture, the prospect of debt dilution makes investors assign substantial risk already to the very first debt unit issued. However, the country will be able to borrow even beyond , as new debt continues to enjoy a share in the debt recovery value and investors ex­pect a positive return even in the event of certain default.

illustration not visible in this excerpt

Figure 1: First-in-time seniority debt versus conventional debt: marginal borrowing cost^[31]

In summary, introducing seniority debt in public borrowing might create welfare in terms of discouraging over-borrowing, lowering borrowing costs at moderate debt lev­els, and overall, reducing incentives to adopt risky debt structures. How­ever, at high levels of public debt, government would face more difficulties to at­tain financing than they do under a conventional debt regime. This is of particular importance in case a ran­dom external shock is pushing a country’s debt over a certain limit where it cannot ob­tain any further financing. This effect will at least substantially impede a country’s ca­pacity to combat the effects of cyclical shocks, severely restricting emergency financ­ing, or even push countries, whose financial policies are fundamentally sound, into bankruptcy. Consequently, seniority debt might be of limited help in order to stabilize the Eurozone members against asymmetric shocks.

The adoption of a seniority debt structure by governments requires either new legisla­tion at the international or domestic level, or contractual provisions that ex­plicitly pro­tect creditors from dilution by future debt issues. Such provisions might resemble the creditor protection clauses included in corporate bond cove­nants and would be enforced by the courts of the country issuing the debt.

Borensztein et al also highlight alternative ways that could mitigate debt dilution with­out incurring the legal complexity of full-fledged seniority schemes, and pro­pose further analysis in this area. Debt contracts could provide creditors with the option to influence subsequent borrowing at the side of the debtors, ask for re­payment in case certain debt limits are exceeded, or allow for a renegotiation of credit terms in the event of new bor­rowing. Furthermore, the authors mention collateralized debt, which shares some im­portant fea­tures with senior debt, but re­quires a country to hold assets or large cash flows in other ju­risdictions.^[32]

4.3.2 Real indexation of public debt

The sustainability of a country’s debt heavily depends on the development of real varia­bles, such as exports, commodity prices, or GDP. Adverse shocks that sub­stantially af­fect these variables can severely distress public finances and ulti­mately lead to debt cri­ses. For emerging countries, these shocks are typically prompted by unfavorable devel­opment of commodity prices, natural disasters, or declines in imports at the part of trading part­ners. For mature economies, the pos­sible triggers of output shocks are more complex and diverse.^[33]

Already in the aftermath of the debt crisis of the 1980s, several authors suggested that governments should link their debt service obligations to certain macroeco­nomic varia­bles, which would foster risk sharing between sovereigns and their creditors and reduce the incidence of costly public defaults.

Krugman (1988), for example examines the tradeoffs, which a creditor of a gov­ernment whose public debt has breached the maximum sus­tainable level and therefore has lost access to the credit markets, is facing. Providing in­cremental fi­nancing to the sovereign further increases the creditors financial exposure in ex­change for an option value: the country may default again after a while, or may regain the abil­ity to repay its entire debt. However, the resulting debt overhang heavily distorts the in­centives at the side of the government, since good economi­cal performance will mostly benefit the creditor, rather than the country itself. Krugman argues that these moral haz­ard complications can be avoided if incre­mental financing is made contingent on “states of nature” that lie beyond the con­trol of the governments.^[34]

Robert Shiller (1993, 2009) was the first economist to focus on the benefits of growth linked sover­eign debt instruments, proposing the introduction of so called “Trills”, secu­rities, which would grant their holders a long-term, or even perpetual claim on one tril­lionth of a country’s real GDP. Trills would allow public and pri­vate investors to hold a direct position in a country’s future growth prospects and therefore promote risk sharing between in­vestors and gov­ernments. In addition to their stabilizing influence on public finances (as the Trill’s coupon payments would drop with declining tax revenues), Trills would pro­mote international risk diversification and hedging at the side of private and institutional in­vestors.^[35]

Barro (1995) argues that the distortive effect, which real-world taxes pose on the eco­nomic agent’s decisions typically motivates governments to arrange their debt issues in a way that smooth­ens taxes through time. Consequently, linking sover­eign securities to public expendi­tures would be the optimal debt management strategy from a sovereign’s perspective. However, in light of the obvious moral hazard risks associated with such securities, Barro proposes the issuance of GDP linked bonds, which would produce similar tax smoothing effects while being much harder to manipulate, as a second best alternative.^[36]

The various authors discriminate the following types of metrics:

- Variables that cannot be controlled by the country’s government, such as com­mo­d­i­ty prices or natural disasters.
- Variables that directly represent a country’s economic activity. These measures are at least partly within the control of the respective govern­ment, and are typi­cally calculated and published by the country’s statistical agencies, such as ex­ports, public ex­pen­di­tu­res, industrial production, elec­tricity consumption or GDP.

The relative benefits of indexing public debt to one of the above variables largely de­pend on country specific characteristics, such as the sources and nature of po­tential ad­verse shocks, the availability and reliability of respective statistical data, as well as the perceived credibility of the country’s public authorities.

Generally speaking, indexation to variables that directly measure economic activ­ity, such as GDP, would likely provide greater insurance against debt crises trig­gered by adverse shocks. However, the use of these variables might raise concerns at the side of investors about incentives for gov­ernments to corrupt GDP data or even pursue less-growth-oriented policies.^[37]

5 GDP-linked bonds

Borensztein et al (2004) highlight three major benefits to sovereigns issuing at least part of their debt in the form of GDP-linked securities:^[38]

First, growth-linked bonds stabilize sovereign debt dynamics, raise the level of sus­tain­able debt and therefore reduce the likelihood of a country defaulting on its debt during an economic downturn. The stabilizing effect that some variants of GDP-linked bonds exert on variables like the debt-to-GDP ratio also benefits economies that have fallen into a path of per­sistent weak growth and would oth­erwise be pushed towards debt cri­ses. More gener­ally, the authors argue that real indexation may be instrumental to coun­tries, which are obliged by national legal constraints or international agreements such as the Stability and Growth Pact, to stabilize the debt-to-GDP ratio at a certain level.

Thereby, growth indexation also helps evade the disruptions and significant deadweight cost, typically arising from debt crisis and de­fault, to both govern­ments and creditors. De Paoli et al (2006) claim in this context that sovereign de­fault is often ac­companied by banking sector and cur­rency crisis, which consider­ably aggravates the di­rect cost of default. Across their sample of countries, the authors find that the output loss following a sovereign default typically exceeds 15 percent of GDP.^[39]

These contrac­tions often occur already prior to the formal default, which suggests that already the antici­pation of a debt crisis can trigger substantial cost. Further­more, despite of having im­posed hair­cuts upon their creditors, debtor countries frequently exit debt crises with in­creased debt-to-GDP ratios.^[40]

Second, Borensztein et al argue that GDP-linked bonds act as “automatic-stabi­lizers,” enabling sovereigns to smooth national economic output through counter-cyclical fiscal policies, rather than being forced to undertake damaging pro-cycli­cal measures. In other words, lower interest pay­ments in times of output contrac­tion will create budgetary room for Keynesian fiscal po­li­cies, while during eco­nomic booms, the higher debt-ser­vicing burden will force govern­ments to cut public expenses. GDP-linked securities therefore form a natural break on business cycle booms, reducing the likelihood that a government will over­spend and over­heat the economy.

Third, this stabilizing effect on national budgets also enables a more stable and predict­able tax path over the economic cycle, avoiding unexpected and erratic changes, which reduces uncertainty at the side of the economic agents re­lated to consumption and in­vestment decision.

Borensztein et al also highlight the benefits that GDP-linked bonds pose for in­vestors, who might want to hold a well diversified position in the future growth prospects of in­dividual countries.^[41]

5.1 Variants of growth-linked bonds

Several different detailed concepts of growth-linked securities have been proposed in the academic literature. Griffith-Jones and Hertova (2013) differentiate the fol­lowing three main variants:^[42]

Robert Shiller proposed already in 1993 a new financial instrument aimed at providing investors with a broader range of investment options, later referred to as a “Shiller secu­rity”. This security would pay the investor on a yearly basis a per­manent fraction, for in­stance one-tril­lionth (then called a “trill”), of the issuing country’s nominal GDP.^[43]

The second variant of growth-linked securities can be attributed to Borensztein and Mauro (2002). The author’s security would closely resemble regular, plain vanilla bonds in many aspects, but pay an interest rate that fluctuates with the real GDP growth rate of the is­suing country. Such a bond would fix “baseline” interest and growth rates and pay, for exam­ple, one percent of incremental interest for each per­cent of growth above the baseline rate. While Borensztein and Mauro propose a quite simple formula in order to reflect indexation in their security, other authors have built complex asymmetries or caps into their proposed payout schemes.^[44]

A third variant suggests coupon payments similar to a Borensztein/Mauro secu­rity, but ap­plies indexation to the principal of the security rather than to the cou­pon. In other words, the difference between the payments of an indexed security and the pay­ments of an otherwise identical conventional bond would be added or sub­tracted from the princi­pal, and consequently from the country’s debt.

The different designs of growth-linked securities have distinct implications on their practical workings as a public debt instrument. The Shiller security is the only variant that indexes payments to the issuer’s nominal growth and therefore also accounts for in­flation. Furthermore, changes in the issuer’s growth rate have different effects on cou­pon payments and principal values of the different security variants:

- In the short run, the issuer’s growth rate has only limited effects on the cou­pon pay­ments of a Shiller security, while GDP growth increases the nomi­nal value of the security and consequently the servicing payments in the long run.
- The issuer of a Borensztein/Mauro bond would benefit from a GDP growth rate below the agreed baseline in the short run through lower cou­pon payments, while the nominal debt value would remain unchanged. This would create space for counter-cyclical fiscal policies in the debtor’s public budget, but would not necessarily help the debtor’s debt-to-GDP ratio.
- In contrast, under the third variant, a growth blow to the baseline would not af­fect cou­pon payments, but decrease the value of the coun­try’s debt in the long run and therefore automatically exert a positive impact on debt sustainability measures like the debt-to-GDP ratio.

However, the fiscal stabilization benefits of the above variants depend on the mac­roeco­nomic scenario considered. The Borensztein/Mauro security would enable the issuer to pursue Keynesian policies in times of sluggish growth and might therefore be the most effective variant in supporting fiscal stabilization during normal business cycles. How­ever, in a deep and pertaining recession, a rising debt-to-GDP ratio might give rise to concerns about debt sustainability at the side of the investors, and consequently increase cost to roll over the existing debt, which might ultimately lead to default. The third vari­ant might provide here some automatic protective mechanism, which keeps the debt-to-GDP ratio at a save level, thereby ensuring continuing access to the financial markets. Also in an envi­ronment where legal constraints, like those introduced by the Stability and Growth Pact, prohibit debt-to-GDP ratio to cross certain levels, the third variant might be favored.

However, despite differences in payment schemes, set up in a similar way, both vari­ants will provide in the end equal financial benefits to the debtors.

5.2 Pricing of GDP-linked bonds

Borensztein and Mauro (2002) evaluate how markets would price Argentine GDP-in­dexed bonds. Using a simple Capital Asset Pricing Model (CAPM), the authors show that investors, who hold a well-diversified portfolio of US equities, would demand a risk premium of approximately 1 percentage point in excess of the yield of a plain Ar­gen­tine vanilla bond. The CAPM implies that investors only require compensation for the systematic risk component, as they can diversify away any unsystematic por­tion of risk. As GDP growth is typically fairly uncorrelated across countries, the sys­tematic portion of risk contained in an individual coun­try’s income growth rate is rather small. Moreo­ver, converting a large portion of Argentine’s debt into GDP-linked securities would likely decrease the country’s overall default risk and consequently further lower the payoffs of both GDP-in­dexed and conventional bonds.^[45]

Using a similar CAPM based pricing-model, Kamstra and Shiller (2009) found a re­quired risk pre­mium of 1.5 percent points for hypothetical “Trills” is­sued by the US government.^[46]

Chamon and Mauro (2005) confirmed the above assumption, introducing a Monte Carlo framework to predict the evolution of debt-to-GDP ratios under vanilla and GDP-in­dexed debt for selected emerging markets countries. They find that growth indexation can lower a country’s default frequency by one quarter to one third of its initial level. Furthermore, the issuance of GDP-indexed debt can lower the sen­sitivity of bond yields to “surprises” in the GDP growth rate by reducing the effect of those surprises on default probabilities.^[47]

While Chamon and Mauro use historical market prices of conventional bonds in order to estimate default risk, Ruban et al (2008) develop a structural model of sovereign default based upon a country’s capacity to serve its debt, which, pro­duces default profiles and prices for GDP linked bonds. The authors examine the resilience of several contract de­signs against output growth and exchange rate shocks and find that, in the event of a se­vere currency crisis, securities that are in­dexed to the US dollar value of a country’s nominal GDP have substantially lower default rates than conven­tional bonds.^[48]

However, despite the fact that the above authors base their pricing analyses upon in­creasingly complex and comprehensive models, they disregard many as­pects that are typical for the initial broad introduction of financial instruments. Uncer­tainties related to writing a new type of contract, marketing and pricing such a contract, and furthermore the risk of a thin secondary market will likely result in additional novelty and liquidity premia being demanded by potential investors.^[49]

5.3 Benefits of GDP-indexed securities

Griffith-Jones and Sharma (2006) divide the benefits of GDP-indexed bonds into gains for the countries issuing the respective securities, gains for investors and fi­nally, bene­fits that relate to the global economy and the global financial system.^[50]

5.3.1 Benefits for issuers

Griffith-Jones and Sharma argue that particularly borrowers from emerging econ­omies benefit from the workings of GDP-indexed bonds, which limit the pro-cy­clicality of fis­cal pressures by imposing lower interest payments at times of slower growth, and vice versa. This runs counter to the debt dynamics characteristic for emerging economies, which are often pushed into severe fiscal austerity during periods of slow growth in or­der to sustain access to international financial mar­kets.

Furthermore, while all countries benefit in some way from GDP-linked debt, emerging market economies that experience volatile growth and high levels of in­debtedness, are likely to benefit dispro­portionately from reduced vulnerability to debt cri­ses.^[51]

In this context, Borensztein et al (2004) analyze the “Tequila crisis” in 1995 and show that the respective gains for emerging economies can be substantial. The authors find that if Mexico had financed half of its debt through GDP-indexed bonds, the country would have avoided about 1.6 percent of GDP in interest pay­ments. These resources could have been spent on mitigating or even avoiding some of the worst effects of the crisis.^[52]

Also theoretical analyses have confirmed the benefits of using GDP-linked debt. Barr et al (2014), for example, calculate of the maximum level of sustainable debt of a ‘repre­sentative’ sovereign under different scenarios, assuming that all debt is either issued conventionally, or in the form of GDP-linked securities. The authors find that GDP-linked bonds can raise a sovereign’s sustainable debt limit by be­tween 45 and 100 per­cent of GDP, and reduce the probability of sovereign default accordingly. Barr et al also investigate potential welfare implications for their ‘representative’ issuer of GDP-linked bonds and find, apart from substantial sta­bilizing impacts on fiscal policy, net welfare benefits for the taxpayers that equal between 1 and 9 percent of GDP. In principle, po­tential investors in GDP-linked securities will demand a premium for providing GDP-volatility insurance, which in turn will lower the welfare of taxpayers. However, as a re­sult of the higher debt sustainability limit brought about by GDP-linked securities, an increasing debt-to-GDP ratio raises the default premium demanded on conventional bonds faster than that the default premium on GDP-linked bonds. Consequently, at higher debt-to-GDP ratios, the total expenses of servicing conventional debt will ap­proach those of GDP-linked securities. Finally, the cost of sovereign default will tip the welfare balance towards GDP-linked securities in Barr’s model.^[53]

In contrast to companies, sovereign nations cannot be liquidated in the event of default. Moreover, there are no national or international institutions that could be called upon in order to enforce sovereign debt contracts. Consequently, govern­ments face substantial incentives to default on their debt, or aim to restructure their liabilities at regular inter­vals. De Paoli et al (2006) summarize in this context the potential costs of sovereign de­fault, which seem to overcompensate the re­spective benefits, and therefore incentivize debtors to honor their obligations:^[54]

- The authors claim that the empirical evidence suggests that a history of de­fault sub­stantially increases the cost of external financing, but does not necessarily close off a debtor’s access to the financial markets. Particularly ‘serial default­ers’ face higher risk premia on their securities than ‘non-de­faulters’ of compara­ble financial strength.
- The tighter terms and conditions on borrowing alone, though, seem insuffi­cient to discourage governments from defaulting on their debt. A number of studies suggest that the avoidance of broader losses to the economy, which typically ac­company sov­ereign default, might be a more important incentive for debt re­payment. Those losses typically result from a weak­ening of domestic banks that serve as major creditors of governments, which in turn has constraining effects on liquidity and credit provided to the economy, and finally can result in severe GDP contrac­tion and a run on both domestic currency and the banks (‘triple’ cri­sis: sovereign, banking and currency).

Levy-Yeyati and Panizza (2006) examine the above in detail and finds that GDP con­tractions actually precede sovereign defaults by several months. Immediately after the default, growth typically sets in again, which suggests that output con­traction results from the anticipa­tion of a default, rather than the default itself.^[55]

However, there is little evidence in the literature that quantifies the actual costs as­soci­ated with these sovereign crises or with the different types of crises resolution. Never­theless, given that the costs of sovereign default appear to be high, GDP-in­dexed debt can be an appropriate measure to reduce the risk of defaulting in the first place.^[56]

While the benefits for emerging economies are obvious, GDP-indexed bonds may also be attractive for highly industrialized econo­mies, particularly for those, which make up the Eurozone, where the commonly agreed provisions of the SGP se­verely constrain the mem­ber’s fiscal room to maneuver, and tend to render their fiscal policies pro-cycli­cal.^[57]

5.3.2 Benefits for investors

Griffith-Jones and Sharma summarize the benefits that private and institu­tional inves­tors would draw from the introduction of growth-linked debt instruments as follows. Overall, such securities would allow investors to hold equity-like stakes in the future growth prospects of individual countries. Although a similar expo­sure can be achieved through investments in the stock markets of the individual countries, growth-linked se­curities might often be more representative of the economies, and more­over generate considerably higher diversification than stock market indices. Owed to the fact that growth rates across emerging market coun­tries are fairly uncorrelated, a port­folio of GDP-linked bonds issued by several different economies would provide even further di­versification benefits.

Kamal and Lashgari (2012) simulate the payoff of hypothetical GDP bonds in compari­son wit conventional high quality bonds issued by the US government between 1947 and 2010 and find that GDP-linked instruments compare very well versus both short and medium term bonds, and even out-perform long term corpo­rate bonds.^[58]

While the above applies to the investment community as a whole, the various cat­egories of investors might have individual preferences. For instance, growth in­dexed bonds could be attractive for pension funds, particularly for those located in countries like It­aly, where private pension funds benchmark their returns against the public pension system, which is indexed to domestic GDP growth.^[59]

5.3.3 Benefits for the global economy and the global financial system

GDP-indexed securities can be viewed as a public good in the sense that they are likely to generate sys­temic benefits over and above those that accrue at the level of the issu­ers, as well as at the part of indi­vidual private and institutional in­ves­tors. By reducing the likeli­hood of debt crisis and sovereign default in general, such in­struments do not only bene­fit their holders, but also all other creditors of a particular country, including the holders of plain vanilla bonds. These benefits are not restricted to governments and their lend­ers, but extend to various parties that may suffer from contagion effects in the event of financial crisis, and last but not least to multilat­eral institutions like the Inter­national Monetary Fund (IMF) or the ESM that may have to finance bailout packages.

Griffith-Jones and Sharma argue in this context that the above externalities ex­plain why growth-indexed securities have not yet proliferated widely through the power of finan­cial mar­kets, but their widespread introduction would require con­certed and continuous support by the international community.^[60]

5.4 Concerns, issues and obstacles

While growth-indexed instruments generate a variety of benefits for both their is­suers and their buyers, and moreover benefit the global financial system as a whole, the fol­lowing section summarizes the main concerns related to the intro­duction of such in­struments as highlighted in the literature.

5.4.1 Moral hazard

Moral hazard in the context of growth-indexed debt mainly relates to the notion that debt repayments, which are tied to GDP growth might reduce the incentives at the side of debtors to pursue policies that maximize economic growth. However, lower GDP growth would come at a cost for governments, for instance resulting from lower tax in­come, which would likely outweigh potential benefits in interest expenditures. Con­se­quently, it is difficult to imagine that the implementation of policies that deliberately limit growth would be of any practical relevance. This argument plays an even lesser role in developed economies like the Eurozone countries, where a strong institutional framework, and a well established systems of public accountability ensure support of sound eco­nomic and fiscal policies.

Nevertheless, the possibility of moral hazard makes some authors question whether GDP growth is the most suitable metric to index government debt in­stru­ments against. GDP is considered the most comprehen­sive measure of a country’s national income, and con­sequently, GDP-indexed debt would likely produce the maximum stabilization ef­fect. Alternatively, commodity-linked bonds could sub­stantially lower the risk of moral hazard, as sovereigns usually cannot exert any control over commodity prices. How­ever, the stabilization effects of such instru­ments are most pronounced for countries whose economies are substantially de­pend on a specific commodity, mainly emerging markets, while most devel­oped countries have well diversified economies and therefore lack a “natural” com­modity price to link debt service payments against. Therefore, us­ing commodity prices as a metric to index debt instruments against would require the institution of a considerable number of different variables, with negative impacts on the ad­ministration, as well as on the liquidity of the respective bonds.

5.4.2 Accuracy and timeliness of GDP data

An additional moral hazard risk could result from the fact that GDP-indexed debt in­struments incentivize the underreporting of GDP data at the side of the debtor in order to reduce debt-service payments. However, Griffith-Jones and Sharma (2006) assess these incentives to be rather weak. The authors argue that high GDP growth indicates political success and therefore helps governments to get re-elected. Furthermore, low growth tends to discourage both domestic and foreign investment and in addition, sub­stantially in­creases the cost of issuing new debt in the future.

While deliberate tampering of GDP statistics seems unlikely, inaccuracies in the calcu­lation of GDP growth may be a more relevant concern at the side of potential investors. Griffith-Jones and Sharma (2006) argue that national income account­ing has become a well-standardized procedure in recent years and many borrow­ing countries overcame similar concerns at the wide scale introduction of infla­tion-indexed bonds. Conse­quently, the clear definition of relevant variables and calculation methods in the bond contracts could partly mitigate this issue and in­crease confidence on the side of in­ves­tors. The authors furthermore propose the establishment of an independent inter­na­tional in­stitu­tion with the aim of providing tech­nical assistance in order to address de­fi­ciencies in the workings of national statistical agencies, and finally to verify and cer­tify the accu­racy of respec­tive na­tional calcula­tions.^[61]

Revisions of annual GDP statistics and changes of calculation methodologies are further impediments to the introduction of growth linked debt instruments. Also for advanced economies, the impacts triggered by revisions of complex variables like GDP can be substantial. From the viewpoint of investors, these data revisions might pose an unwel­come source of uncertainty, and in addition might be used by debtor countries to reduce debt service payments in an opportunistic way. How­ever, over the long run the impacts of yearly GDP revisions might likely cancel each other out. Furthermore, clearly speci­fying methods for dealing with revisions and methodological changes ex ante in the re­spec­tive bond contracts might help to mitigate remaining concerns about the integrity of underlying GDP statistics. Borensztein and Mauro (2004) suggest here for instance to establish in the bond contract that coupon payments would be based on GDP as calcu­lated on a set date. In any case, the existence of sizable markets for inflation-indexed bonds in many countries suggests that data integrity issues are not an insuperable con­cern at the side of potential investors.^[62]

While the accuracy of GDP data might be a concern for potential buyers of GDP bonds issued by developing economies, the EMU produces comprehensive and reliable statis­tics on GDP and its components, so that the Eurozone members may find it easier to sell respective bonds to the international investors community.

Sizable lags in the provision of GDP data may raise concerns at the part of the is­suers, rather than the buyers. The proper functioning of GDP-indexed securities as automatic stabilizers depends on the extent to which the actual coupon payments correlate with the economic cycle of the borrower. However, the high auto-corre­lation of GDP series pub­lished on a quarterly basis might mitigate this impact.

5.4.3 Uncertainty about sufficient liquidity

Uncertainty about the future liquidity of GDP-indexed securities constitutes an­other fundamen­tal concern for investors. Attaining sufficient liquidity is further­more vital for the issu­ers of indexed debt themselves, as the prospect of high li­quidity reduces the re­quired risk premium, as well as the “novelty premium” that the initial issuers of a finan­cial in­strument typically have to face. Consequently, the creation of a market for new financial in­struments such as GDP-indexed secu­rities requires the issuance of a certain initial debt vol­ume in order to en­sure a critical mass of transactions, for instance in the course of a major debt restructur­ing, rather than a the adoption of a gradual ap­proach.^[63]

Griffith-Jones and Sharma (2006) propose in this context the simultaneous and con­certed issuance of GDP-indexed bonds by several governments, coordinated and sup­ported by international or­ganizations. According to the authors, the current abundance of liquidity in the global financial markets, along with recent financial innovations such as the Economic Deriv­atives market, would further reduce the “novelty premium” de­manded by the inves­tors.^[64]

5.4.4 Pricing difficulties

Further concerns stem from the fact that GDP-indexed bonds are more difficult to price compared to plain vanilla bonds, which would make potential investors de­mand addi­tional compensation, the “novelty premium”, which potentially lies be­yond the mark-up the issuers would be willing to pay. Griffith-Jones and Sharma (2006) relate this phe­nomenon primarily to the lim­ited availability and quality of market-based GDP growth projec­tions. However, the authors argue that the devel­opment of a global GDP-linked bond market should cata­lyze respective improve­ments, and point out that also the intro­duction of in­fla­tion-in­dexed securities in the US in the late 1990s has initially been ac­companied by heavy skepti­cism at the side of many market participants. Griffith-Jones and Sharma further­more ar­gue that structural simplicity, as well the establishment of “comparables” through a range of GDP-linked bonds that have similar features and payment standards, is­sued by different mature and emerging econ­omies, would support the price dis­covery process and there­by help overcome pricing difficulties. Strict stand­ardiza­tion is moreover considered inevitable for the creation of a liquid secondary mar­ket.

Also multilateral financial institutions could play a certain role here, assisting in the standardization of GDP-linked bond contracts as well as in the de­velopment of pricing models for GDP-indexed instruments, and furthermore these institutions could execute financial derivative transactions that would facilitate price creation. A swap transaction involving a small quantity of a nominal bond against a compa­rable GDP-indexed bond would for instance provide the potential issuer of the in­dexed bond with a first pricing benchmark.^[65]

Furthermore, potential investors into GDP-indexed bonds will demand a premium as a compensation for the higher variability of interest payments compared to fixed-rate se­curities. However, the size of this mark-up should decrease with the investors’ ability to elimi­nate a portion of this risk through diversification. Conse­quently, the availability of bonds tied to growth in a variety of developed and emerging countries could yield con­siderable diversification benefits and therefore lower the premium required as a re­sult of the variability of interest payments on GDP-linked bonds.

5.4.5 Long-term benefits versus short-term costs

The fact that the benefits of GDP-indexed securities impact the financial position of a debtor in the long run, while the respective incremental premia hit public fi­nances im­mediately, runs contrary to the short political horizons prevalent in most mature and de­vel­oping economies. As the build up of unsustainable debt positions typically takes sev­eral years, some governments could be reluctant to pay a pre­mium to issue indexed debt that might benefit their successors several years down the road.

6 GDP-indexed bonds as a stabilization tool for the EMU

In the below section, a simulation is performed that quantitatively assesses the stabilization effect that might have resulted from the EMU-wide adoption of GDP-indexed sovereign debt prior to the 2008 crisis.

The introduction of indexed debt across the members of the EMU obviously re­quires broad acceptance of such securities at the part of private and institu­tional in­vestors. The academic literature claims that the potential appeal of GDP-indexed public debt to the investor’s community might stem from the fact that such in­struments provide superior diversification benefits compared to classical stock market indices, a characteristic that will be empirically assessed prior to the sim­ulation.

6.1.1 Simulation of the diversification effects

Griffith-Jones and Sharma (2006) claim that investors would favor growth-linked securi­ties over stock market indices in order to hold equity-like stakes in the economies of individual countries. The authors argue that this preference results from the fact that GDP-indexed securities more closely represent the economies of their issuers, and moreover, growth-linked instruments might generate considera­bly higher di­versification than stock market indices. Moreover, owed to the fact that GDP growth rates individual countries are fairly uncorrelated, a portfolio of GDP-linked bonds issued by several different economies would provide even further diversification benefits.^[66]

The above benefits seem vital for the issuers of indexed debt themselves, as sub­stantial diversification effects increase the attractiveness of GDP-linked securi­ties for the international investors community. The resulting prospect of high liquidity will in turn reduce the required risk premia, as well as the “novelty premium” that the initial issuers of a financial instrument typically have to face.^[67]

Below, an analysis is provided that empirically assesses whether the assumptions of Griffith-Jones and Sharma hold true for the scenario of a concerted introduction of GDP-indexed government debt across the euro area. Thereby, only the coun­tries that have joined the Eurozone before 2007, excluding Luxembourg and Malta are included in the analysis.^[68]

The analysis compares the variability of the GDP growth rates of the above coun­tries between the years 2001 and 2013 with the variability of stock market returns for the same period. Stock GDP growth rates for the individual euro area countries have been obtained from the IMF’s World Economic Outlook Database of April 2015, while Stock market index data stem from the YAHOO online finance data­base.^[69]

In order to obtain a GDP growth rate for total of all countries subjected to the analysis, the GDP growth rates of the individual countries have been weighted using their absolute nominal GDP in the year 2000 as a weighting factor. This weighting factor has not been adjusted for the individual years under examination.

Market performance is analyzed using 4 individual stock mar­kets indices as ex­ample: Austria is represented by the ATX, Germany by the DAX, France by the CAC 40 and Greece by the ATHEX. In addition, the MSCI Eurozone has been used in order to approximate the growth of stock markets across the euro area.

The variability of both, GDP growth rates and stock market performance is meas­ured via the Standard Deviation according to the following formula:

Thereby, are the individual value of the dataset, is the arithmetic mean of the data and N the total number of data points.

Figure 2 below shows GDP growth rates and stock market performance for se­lected years, as well as the Standard Deviation for the entire evaluation period, while Figure 3 contrasts the Standard Deviations for GDP growth and stock mar­ket performance of 4 exemplary markets.

illustration not visible in this excerpt

Figure 2: Selected Eurozone countries – actual nominal GDP growth and stock market performance (selected years)

illustration not visible in this excerpt

Figure 3: Selected Eurozone countries – Standard Deviation of 2001 to 2013 growth rates

The Standard Deviation for GDP growth rates of the Eurozone as represented here by 13 selected countries lies below 2%, a fraction of the stock market’s variability that reaches almost 26%. Nevertheless, both the Standard Deviations for the euro area’s GDP growth and its market performance show substantial diversification effects compared to the Standard Deviations of the individual countries. In fact, the variability of each individual country is higher than the variability of the euro area total. Consequently, we can assume that holding a well-diversified portfolio of GDP-indexed securities is as crucial for investors, as it is for stocks. Therefore, a concerted introduction of such instruments across the euro area can be expected to reduce the risk premia demanded by investors, at least compared to a stand-alone issue by an individual country. Furthermore, although a portfolio of GDP-indexed bonds issued by Eurozone members may already generate substantial di­versification, the respective benefits might be further amplified by the addition of indexed securities from emerging economies whose growth dynamics are im­pacted by a variety of different factors, for instance by commodity prices. This suggests that the worldwide introduction of growth-linked securities might further reduce their cost.

6.1.2 Simulation of the stabilization effects

The simulation laid out in this section models the hypothetical evolution of debt-to-GDP ratios for all members of the Eurozone under the assumption that GDP-indexed debt schemes had been in place at the end of 2007.

Obviously, the underlying model is simplified to a large degree and takes a purely mechanical approach, deliberately ignoring economic and political developments that may have resulted from changed debt dynamics. Likewise, the simulation ig­nores any cost related to the broad implementation of GDP-indexed bonds in the Eurozone, any impact on the respective issuer’s overall credit ratings that also might lower interest rates on the individual countries conventional debt, or any other potential externalities.

Overall, the study models the hypothetical effect that GDP-indexed securities would have exerted on the evolution of debt-to-GDP ratios in the euro area from 2007 to 2014. A scenario where all euro area members have issued a certain amount of indexed bonds at the onset of the 2008 crisis is compared to a “base­line” where all debt has been fi­nanced through conven­tional securities. As above, the analysis only includes countries that adopted the Euro prior to 2007 (excluding Luxem­bourg and Malta).

For the “indexed” scenario, it is assumed that by the end of 2007, the starting point of the analysis, the countries of the Euro­zone have is­sued 50 percent of their entire sover­eign debt stock in the form of GDP-in­dexed bonds with a maturity of 10 years, while the remaining debt has been fi­nanced through con­ventional secu­rities of equal maturity. All straight debt is assumed to carry fixed interest rates that equal the average market yield of 10-year sovereign bonds in 2007, while GDP-indexed securities pay an additional risk premium of 1 percent in order to compensate investors for increased variance of re­turns. The size of the GDP risk pre­mium follows the analysis issued by Borensztein and Mauro in 2012, who sup­pose that the holders of GDP-indexed bonds will be able to elimi­nate unsystematic risk through diversification. Conse­quently, the investors will only demand finan­cial com­pensation for the system­atic risk component.^[70]

All Eurozone countries are assumed to have issued GDP-indexed debt of a similar vari­ant (see chapter 5.1), where the face value of the secu­rity changes as nominal GDP deviates from the baseline that has been contracted between the issuing country and the investors. Thus, an unexpected de­cline in nominal output does not af­fect a country’s ongoing in­terest obligations, but decreases the face value of its debt stock and therefore automati­cally exerts a positive impact on its debt-to-GDP ratio.

In the years following 2007, the individual Eurozone members have to refinance capital needs re­sulting from primary government deficits through sovereign debt at the market interest rates of the respective year. However, due to the presumed aversion on the side of investors against buying indexed securities at the onset of an economic downturn, any borrowing from 2008 onwards is assumed to be non-indexed.

The face values of the indexed bonds result from the following formula:

While Dt is the value of the outstanding debt at the time t, D0 denotes the original face value of the security, and consequently also the value the issuer has to re­pay in case GDP growth equals the agreed baseline at maturity. stands for nomi­nal GDP at the time t and represents the baseline level of nominal GDP at that time.

The GDP baseline projections for the individual euro area countries have been obtained from the IMF’s World Economic Outlook Database of April 2008.^[71]

Also actual nominal GDP data are taken from the above IMF database, while av­erage interest rates on issues of conventional government bonds have been at­tained from the Statistical Data Warehouse of the ECB.^[72]

Finally, government debt levels of the individual Eurozone members at the end of 2007, as well as primary budget balances for all subsequent years are from the Eu­ropean Commission’s Annual Macro-Economic database (AMECO).^[73]

Starting from the debt-to-GDP ratios of the individual euro area members at the end of 2007, the study models the evolution of the individual sovereign debt val­ues between 2008 and 2014. For each year, the total of a country’s interest ex­pense and primary budget balance is assumed to be used in order to repay conven­tional sovereign debt or, in case negative, to trig­ger additional borrowing via con­ven­tional 10-year bonds at the average market in­terest rate. The debt-to-GDP ra­tios generated by the above “indexed” scenario are then compared to a “non-in­dexed” baseline. Finally, the results of the sim­ulation are complemented by sensi­tivity analysis.

6.1.3 Results

Table 1 in the Appendix shows that the actual nominal GDP numbers for the indi­vidual euro area members substantially deviate from the nominal GDP projections that have been published by the IMF in April 2008. The below figures show the respective data in the form of a line graphs (please see the source data in the Ap­pendix, Tables 1 and 2).

Already in the year 2008, the nominal GDP of all countries falls sharply below the baseline forecast, opening a gap between projected and actual data that none of the Eurozone members is able to close between the years 2009 and 2013. In addition, most countries seem to experience another blow to their nominal output, or at least a slow down in GDP growth in the year 2011. As a consequence, we can expect indexed debt that has been issued before 2008 to decrease in value substantially, generating substantial benefits on the part of all issuers.

However, following the initial recession in 2008, the nominal output graphs of the individual euro area members show quite different patterns of recovery. Figure 4 below compares nominal GDP baseline and actual nominal GDP for the southern members of the Eurozone, namely Greece, Cyprus, Italy, Portugal and Spain. While the 2008 crisis pushes Greece into a recession that persists throughout the entire evaluation period, Cyprus, Italy and Portugal show some signs of recovery in 2009 that, however, lasts only for merely 2 years and is followed by another re­cession. The GDP projection of Spain, on the other hand, in some sense resembles the Greek data, with GDP trending downwards throughout the entire evaluation period, although the decline seems milder compared to Greece.­

illustration not visible in this excerpt

Figure 4: Southern Eurozone members – nominal GDP baseline and actual nominal GDP (normalized to 100)

The 2008 crisis also put an abrupt end to the credit boom that led to the emergence of sizable construction sectors, particularly in Spain, or stimulated private and public consumption in Greece and Portugal. As a consequence, declining property prices increasingly distressed the banking sectors of those countries. However, until the end of 2009, relatively low debt-to-GDP levels still nurtured confidence on the part of private and institutional investors in the ability of all Eurozone members to sta­bilize their banking systems, which was viewed as a national re­sponsibility at this time. Figure 5 shows the development of yields on ten-year sover­eign bonds between the years 1993 and 2012: Interest rate spreads that had converged dramatically with the adoption of the common currency remained rela­tively narrow during the early years of the crisis.

illustration not visible in this excerpt

Figure 5: Ten-year yields on sovereign bonds, January 1993 to Feb­ruary 2012^[74]

However, by end of 2009, the newly elected government of Greece revised its cur­rent and past year’s budget deficits, which in turn triggered a substantial increase of sovereign bond yields across the southern Eurozone periphery, indicating that investors started to doubt the ability of these countries to service their sovereign debt. As a result, Greece asked for offi­cial assis­tance in May 2010, followed by Ireland in No­vember 2010, Portu­gal in May 2011, and then finally Spain and Cy­prus in June 2012.

Figure 6 shows the respective development of yields on ten-year sovereign bonds between 2009 and 2012. The number suggest that severely higher cost of govern­ment financing, along with further contraction of credit to the private sector caused by increasing pressure on the banks balance sheets, have at least substan­tially contributed to the second output contraction around 2011 in the southern euro area periphery.

illustration not visible in this excerpt

Figure 6: Yields on 10-year sovereign bonds, October 2009 to June 2012^[75]

The data for Ireland and Slovenia, on the other hand, show sharp decline of nomi­nal output in 2008, resembling the GDP development of Greece, Cyprus, Italy, Portugal and Spain. However, the “confidence fueled” second hit in 2011 is less pronounced and moreover followed by some signs of recovery.

illustration not visible in this excerpt

Figure 7: Ireland and Slovenia – nominal GDP baseline and actual nominal GDP (normalized to 100)

Finally, Figure 8 compares nominal GDP baseline and actual nominal GDP of the remaining euro area members, Austria, Belgium, Finland, France, Germany and the Netherlands, all of which are located in northern or central Europe. In contrast to the southern periphery, nominal output contracts here sharply in 2008, but then returns to a growth path immediately thereafter. Some markets, namely Germany and Belgium, even manage to partly close the output gap versus the baseline in the years 2009 and 2010. However, the shock of 2011 also leads to a slow down of nominal GDP growth also in northern and central Europe.

illustration not visible in this excerpt

Figure 8: Northern and central Eurozone members – nominal GDP baseline and actual nominal GDP (normalized to 100)

The results of the simulation as shown in Figure 9 confirm the above analysis of the individual Eurozone country’s nominal output development.

illustration not visible in this excerpt

Figure 9: Debt-to-GDP ratios in 2007 and 2014 assuming 0 percent and 50 percent of to­tal debt being financed through indexed borrowing

In the years fol­low­ing the 2008 crisis, all countries experience substantial surges in their debt-to GDP ra­tios, mainly resulting from declining nominal output and their primary balances turning red. In this context, Figure 10 lays out the public sec­tor primary balances for se­lected euro area members during the crisis years. While for the year 2007 only Greece shows a negative primary balance, by 2009 basically all countries run sub­stantial primary deficits.

illustration not visible in this excerpt

Figure 10: Selected Eurozone countries – net lending (+) or net borrowing (-) excluding in­terest (in billion euros)

However, all Eurozone members except Germany would benefit from the sub­stantial devaluation of their in­dexed debt.

Building on the above, Figure 11 shows the decrease of the individual countries debt-to-GDP ratios resulting from the introduc­tion of 50 percent indexed debt. While the northern and cen­tral European euro area members seem to benefit only marginally (Austria, Belgium and Finland 1 percentage point, France 2 percentage points and the Netherlands 2 percentage points), the southern Eurozone periphery countries that suf­fered the most from the distortions triggered by the crisis, show substantial benefits, ranging between 5 percentage points in Italy and Spain, 6 percentage points in Portugal, 7 percentage points in Cyprus and fi­nally 27 per­centage points in Greece. In addition, Ireland and Slovenia show some medium size benefits, with 4 and 3 percentage points respectively. Germany seems to be the only country that would have suffered from having issued indexed debt.

illustration not visible in this excerpt

Figure 11: Change in debt-to-GDP ratios in 2014 resulting from the issuance of 50 per­cent indexed debt in 2007 (in percentage points)

However, given the assumptions of the simulation, the devaluation of GDP-in­dexed debt seems insufficient to prevent substantial surges in the debt-to-GDP ra­tios of the Euro­zone members during the post-crisis years. The below Figure 12 splits the changes in debt-to-GDP ratios between 2007 and 2014 into their indi­vidual drivers, namely into the de­valuation impact on GDP-in­dexed debt that would have resulted from an unexpected slowing of out­put growth, into primary balances and interest expenses, and finally into the impact that output growth would have exerted on the debt-to-GDP ratio itself.

illustration not visible in this excerpt

Figure 12: Increases in debt-to-GDP ratios between 2007 and 2014 - drivers

The devaluation of indexed debt exerts a dampening effect on debt-to-GDP ratios across the Euro­zone members. However, the size of this effect differs substan­tially across the countries examined, remaining below 5 percentage points in countries where ac­tual out­put in 2014 fell less than 20 percent below the original base­line (Austria, Belgium, Finland, France, Germany, Ireland and the Nether­lands). A GDP shortfall of 20 to 40 percent versus the baseline triggers a moderate damp­ening effect between 6 to 10 percentage points (Cyprus, Italy, Portugal and Spain). Greece, however, stands out here, with an actual nominal GDP in 2014 that reaches less than half the size originally forecasted by the IMF in April 2008. The result­ant devaluation of indexed debt dampens the country’s debt-to-GDP ra­tio by 27 percentage points.

Between 2007 and 2014 all countries except Germany and Italy ran accumulated primary deficits. Figure 13 shows that the size of these deficits correlates closely with the re­spective country’s GDP shortfall versus the original baseline (used as an es­timate for the severity of the output shock incurred). Consequently, we can as­sume that those defi­cits result from the workings of automatic stabilizers and ad­ditional efforts to support con­sumption through Keynesian measures, rather than from irresponsible fiscal conduct. In the countries hit most harshly by the 2008 crisis, the upward force that accumulated primary deficits exert on debt-to-GDP ratios, by far outweighs the dampening effect of GDP-indexed debt. For in­stance, the average primary deficit in Spain, Portugal, Ireland and Greece is 4 times the size of the average debt devaluation effect that, given the as­sumptions of the sim­ulation, would have been generated by the introduction of GDP-in­dexed securities in these countries.

illustration not visible in this excerpt

Figure 13: Relation between the severity of the output shock and accumulated primary balances between 2007 and 2014

Interest expenses incurred by the Eurozone countries between 2007 and 2014 fol­low a similar dynamic Figure 14 shows that the accumulated interest expenses are the most significant driver of the increases in debt-to-GDP ratios during the sim­ulated crisis. The simulation assumes that government budget deficits between 2007 and 2014 are financed through new conventional 10-year debt at the average mar­ket inter­est rates of the respective year.^[76] Consequently, in some Eurozone mem­bers, high and persistent primary budget deficits during the crisis years and more importantly, ever increasing cost of new financing drive gov­ernment debt-to-GDP ra­tios to dangerously elevated, and in the case of Greece to unsustainable levels. Correspondingly, Figure 14 shows that interest pay­ments on new debt are a mayor driver of the increase of debt-to-GDP ratios in all southern European euro area members and Ireland.

illustration not visible in this excerpt

Figure 14: Relation between the increase in debt-to-GDP ratios from 2007 to 2014 and average interest on newly issued 10-year bonds in the period under review

Finally, the drivers presented in Figure 12 also indicate that output growth plays a major role in dampening the growth of the simulated debt-to-GDP ratios during the simulated crisis. How­ever, the size of this effect differs substantially across the countries ex­amined. While GDP growth substantially reduces debt-to-GDP ratios in most Eu­rozone members, Greece, Ireland, Portugal and Spain experience a sub­stantial de­cline in economic out­puts that contributes to, rather than dampens, the surge in debt-to-GDP ratios.

6.1.4 Sensitivity analysis

The above simulation is in part based on fairly arbitrary assumptions, for instance for the risk pre­mium that will be demanded by investors as a compensation for GDP volatil­ity, or for the percentage of the euro area member’s public debt stock that is assumed to be financed via GDP-indexed securities by the end of 2007. In addition, the simulation assumes that at the onset of a crisis, the Eurozone mem­bers will stop issuing GDP-indexed debt and instead cover additional financing needs exclusively through conventional 10-year bonds. While this assumption seems sensible in principle, one might as well assume that increasing concerns on the part of the investors would have made risk premia for growth indexed securi­ties diverge in the same way as this could be observed for plain vanilla bonds is­sued by the individual Eurozone governments.

The following section, the sensitivity of the simulation towards changes in its major assumption is tested.

Risk pre­mium

Figure 15 shows how the decline in GDP-to-debt ratios, discussed under section 6.2.1 Results, would change depending on different assumptions for the risk pre­mium demanded by investors as a compensation for the higher volatility of GDP-indexed securities.

In the academic literature, the premium that investors, who hold a well-diversified portfolio of US equities would demand, is estimated between 1 and 1.5 percent.^[77]

However, the effect of growth-indexed securities on GDP-to-debt ratios appears to be fairly sensitive towards changes in the assumed risk premium. A doubling of the assumption from 1 to 2 percent would result in debt-to-GDP ratios increasing versus the “non-indexed” scenario in 3 countries (Austria, Belgium and Ger­many), while the remaining countries would still marginally (Finland, France, It­aly and the Netherlands), or substantially (Cyprus, Greece, Ireland, Portugal, Slo­venia and Spain) benefit.

A premium of 3 percent would make another 5 Eurozone members incur negative impacts (Finland, France, Italy, Netherlands and Portugal), so that only Cyprus, Greece, Ireland Slovenia and Spain would be left as beneficiaries. This aspect is material therefore requires further investigation.

illustration not visible in this excerpt

Figure 15: Sensitivity analysis - risk premium

Share of GDP-indexed securities

The simulation implies that all Eurozone members enter the 2008 crisis with 50 per­cent of their debt stock financed through GDP-indexed bonds, which is a fairly arbitrary assumption. Figure 16 models the effect, which different percent­ages of in­dexed debt would exert on GDP-to-debt ratios. The results suggest that the size of debt reduction linearly increases with higher percentages of debt issued in the form of indexed bonds.

However, in contrast to changes in the assumed risk premium, a higher percent­age of indexed debt would impact all bond issuers in the same way, simply increasing their benefit or cost (in case of Germany).

illustration not visible in this excerpt

Figure 16: Sensitivity analysis - percentage of indexed bonds

Issue of GDP-indexed during the crisis years

The simulation has been built upon the assumption that the euro area members will stock up on GDP-indexed securities in “good times” when abundant liquidity in the global financial markets makes investors demand low risk premia, while during times of crisis, financing needs will be met through conventional bonds. However, indexed bond issues in a downturn would probably require the individ­ual Eurozone governments to accept risk premia that substantially differ by coun­try. In other words, the risk premia would diverge in the same way as this could be observed for plain vanilla bonds during the 2008 crisis.

The below sensitivity analysis evaluates how issuing 50 percent of all new debt in the form of GDP-in­dexed bonds during the crisis years would have changed the results of the simulation.

For this propose, the GDP-to-debt ratios individual countries are assumed to drive the risk premia of growth-indexed debt as follows: and are the risk premia and debt-to-GDP rations of the individual coun­tries, while represents the base risk premium that investors demand, even in case the individual economy is on the upswing. In case the formula calculates a value of that lies below for a specific country, the latter number will be used in the sensitivity analysis.

The above formula is applied to new GDP-indexed debt that a country has to issue in order to meet its financing needs between 2008 and 2014, respectively 50 per­cent of all new bond issues during the crisis years. However, also new GDP-in­dexed debt would devalue as soon as the issuers actual GDP starts trending below the baseline forecast at the time of the debt issue, and vice versa. For simplicity reasons, this impact is ignored here.

illustration not visible in this excerpt

Figure 17 and below shows the results of the sensitivity analysis. Overall, debt-to-GDP ratios for basically all countries except Finland have worsened substantially.

illustration not visible in this excerpt

Figure 17: Sensitivity analysis - debt-to-GDP ratios in 2007 and 2014 - comparison of continued indexed debt issues after 2007 versus simulation (in percentage points)

Figure 18 highlights the impacts of issuing indexed during the crisis versus the sim­ulation by country. The debt-to-GDP ratios of all countries worsened, with impacts below 6 percentage points in the northern and central Europe (Austria, Belgium, Cyprus, Finland, Germany, the Netherlands), as well as in Cyprus and Slovenia, followed by France and Italy still below 10 percentage points. Greece, Ireland, Portugal and Spain suffer impacts between 12 and 52 percent points.

illustration not visible in this excerpt

Figure 18: Sensitivity analysis - change in debt-to-GDP ratios in 2014 - comparison of continued indexed debt issues after 2007 versus simulation (in percentage points)

The impacts can be attributed on the one hand to the fact that some countries en­tered the 2008 crisis already with fairly elevated debt-to-GDP ratios (for instance Belgium 87 percent, Greece 103 percent, Italy 100 percent and Portugal 68 per­cent). Moreover, declining output in combination wit the workings of automatic stabilizers put upward pressure on the debt-to-GDP ratios of most Eurozone members, thereby substantially increasing the risk premia that investors demanded for new GDP-indexed debt issues. In addition, negative public sector primary bal­ances make some countries face heavy financing needs between the years 2008 and 2014 (see Figure 10, for instance France more than 350 million euros, Greece more than 75 million, and Spain 430 million), all of which need to be financed at these elevated rates. However, the results confirm the high sensitivity of the sim­ulation towards changes in the assumed risk premium, a topic that definitely needs further investigation.

7 Summary and conclusion

The 2008 financial crisis triggered the most severe global recession since the Se­cond World War, leading economies across the globe to the brink of disaster, and still raising doubts upon the sustainability of the EMU in its present design.

The Eurozone members have, the European delegated all aspects of monetary policy to a communitarian central bank, while fiscal and macroeconomic policies strictly remain under national responsibility, only controlled by a basic fiscal gov­ernance framework.

In contrast to most currency unions, the stability of the Eurozone fully depends on its capacity to cope with asymmetric shocks through the workings of automatic stabilizers and through national counter-cyclical fiscal policy. However, the 2008 crisis clearly illustrated that the Eurozone lacks adequately developed automatic stabilizers. Moreover, in a currency union, the individual governments issue debt in a currency whose supply they don’t control, a fact that seems to fundamentally impede their ability to finance sovereign deficits. As a result, vulnerability to changing market sentiment and the inability to use counter-cyclical fiscal policies for economic stabilization tends to produce more pronounced booms and busts.

After the 2008 crisis, the Union substantially reformed its governance framework, also introducing new EMU-wide financial backstops such as the ESM and the Banking Union, both aimed at containing the effects of future financial crises in the euro area. However, the debate about the sufficiency of those reforms in order to effectively control future output shocks has been continuing up to the present day.

In the course of this debate, a variety of communitarian risk-sharing mechanisms have been proposed. Such in­struments would transfer a substantial part of cyclical stabilization to the Euro­pean level, and thus allow the euro area members to focus their fiscal policies on structural aspects. Other proposals aim at rendering public debt structures less cri­sis prone, such as the introduction of seniority in sovereign debt issued by EMU countries, or the Eurozone-wide introduction of GDP-in­dexed debt. The latter would stabilize sovereign debt dynamics ad thus foster counter-cyclical fiscal policies in the event of an economic downturn, rather than forcing governments into taking damaging pro-cyclical measures.

A broad introduction of growth-linked securities would allow investors to take well diversified equity-like stakes in the future growth prospects of individual countries. In addition, GDP-indexed may generate systemic benefits for all eco­nomic agents by reducing the likelihood of debt crisis and sover­eign default in general. However, the above benefits come with a range of con­cerns, such as the possibility of moral hazard, issues related to the accuracy and timeliness of GDP data, pricing difficulties, and uncertainty about sufficient li­quidity at the introduc­tion of such instruments.

This Master’s Thesis evaluates whether the introduction of GDP-indexed sover­eign debt would provide substantial stabilization to the Eurozone in the event of a macroeconomic shock. To this end, a simulation is performed that quantitatively assesses the stabilization effect that might have resulted from the EMU-wide adoption of GDP-indexed sovereign debt prior to the 2008 crisis. The simulation implies that all Eurozone members enter the 2008 crisis with 50 percent of their debt stock financed through GDP-indexed bonds. The results confirm the stabi­lizing effect that indexed public securities exert on the debt-to-GDP ratios of their issuers. First and foremost, the countries that suf­fered the most from the 2008 cri­sis show substantial benefits, ranging from 5 per­centage points in Italy and Spain, 6 percentage points in Portugal, 7 percentage points in Cyprus to 27 percentage points in Greece. In addition, the results of a sensitivity analysis suggest that the size of debt reduction linearly increases with higher percentages of indexed debt. A share of 80 percent indexed bonds would, for instance, increase the dampening effect on debt-to-GDP ratios to 8 percentage points in Italy and Spain, 10 percent­age points in Portugal, 11 percentage points in Cyprus and 43 percentage points in Greece. However, the above effects seem too low in order to offset the adversities of the crisis in full. For Greece, the model projects a surge of the countries debt-to-GDP ratio from the pre-crisis level of 103 percent to 260 percent in 2014. The introduction of 50 or even 80 percent indexed debt can mitigate this development by 27 or 43 percent points respectively, which seems insufficient to prevent doubts in the Greece’s creditworthiness at the side of investors in the country’s sovereign debt. As a result, the increasing interest rate for new financing are likely to exert further pressure on Greece’s debt-to-GDP ra­tio, which might finally push the country into a severe liquidity crisis or even into public default.

Also for the remaining countries of the southern Eurozone periphery (Italy, Spain, Portugal and Cyprus), the model produces results that resemble those of Greece, although the impacts are less pronounced.

In a nutshell, the results suggest that GDP-indexed public debt does have substan­tial stabilizing effects on the debt-to-GDP ratios of its issuers. This stabilizing property stems from the fact that GDP-indexed bonds decrease in value as soon as the actual nominal GDP of the issuer dives below the baseline forecast prevalent at the time of the initial bond issue. Therefore, such securities can exert a damp­ening effect on the issuer’s debt-to-GDP ratios in the event of output shocks. The effect correlates with the size of the gap between projected and actual GDP, re­spectively with the size of the output shock hitting the issuer.

At least for assumptions made in this simulation, the effects generated by the in­troduction of GDP-indexed securities seem insufficient in order to prevent the Eu­rozone’s most vulnerable members from liquidity crises. However, such instru­ments may well function as a contributor to the stability of the Eurozone, along with existing instruments like rigorous macroeconomic and fiscal governance, as well as with financial backstops like the ESM and the Banking Union, and poten­tially with a still-to-be-introduced EMU-wide risk-sharing mechanism.

8 Bibliography

Barr, D.; Bush, O. and Pienkowski, A. (2014): GDP-linked bonds and sovereign default. Bank of England Working Paper No. 484, January 2014.

Borensztein, E. and Mauro, P. (2002): Reviving the Case for GDP-Indexed Bonds. IMF Occasional Paper 02/10, September 2002.

Borensztein, E.; Chamon, M.; Jeanne, O.; Mauro, P. and Zettelmeyer, J. (2004): Sovereign Debt Structure for Crisis Prevention. IMF Policy Discussion Paper No. 237, 2004.

Chamon, M. and Mauro, P. (2005): Pricing Growth-Indexed Bonds. IMF Working Paper 05/216, November 2005.

De Grauwe, P. (2011): The Governance of a Fragile Eurozone, Australian Eco­nomic Review 45/3, 255-268.

De Paoli, B.; Hoggarth, G. and Saporta, V. (2006): Costs of sovereign default. Bank of England Financial Stability Paper No. 1, July 2006.

Dullien, S. (2008): Eine Arbeitslosenversicherung für die Eurozone: Ein Vorschlag zur Stabilisierung divergierender Wirtschaftsentwicklungen in der Eu­ropäischen Währungsunion. Stiftung Wissenschaft and Politik Berlin S01, Feb­ruar 2008.

Enderlein, H; Guttenberg, L. and Spiess, J. (2013): Blueprint for a Cyclical Shock Insurance for the Euro Area. Notre Europe Jaques Delors Institute Studies & Re­ports 100, September 2013.

EUCO (2012): Towards a genuine economic and monetary union: Report by Presi­dent of the European Council Herman Van Rompuy, in collaboration with José Manuel Barroso, Jean-Claude Juncker and Mario Draghi. EUCO 120/2012. Brussels: 26 June 2012.

COM (2012): A blueprint for a deep and genuine economic and monetary union: Launching a European Debate. COM 777/2012. Brussels: 30 November 2012.

Feld, L.P. and Osterloh, S. (2013): Is a fiscal capacity really necessary to com­plete EMU? Paper presented at the workshop „How to build a genuine economic and Monetary Union“ on 30 May 2013.

De Paoli, B.; Hoggarth, G. and Saporta, V (2006): Cost of Sovereign Default. Bank of England, Financial Stability Paper No. 1.

Griffith-Jones, S. and Hertova, D. (2013): Growth-Linked Bonds. CESifo DICE Report 3, September 2013.

Griffith-Jones, S. and Sharma, K. (2006): GDP-Indexed Bonds: Making It Hap­pen. DESA Working Paper No. 21, April 2006.

Kamstra, M. and Shiller, R. J. (2009): The Case for Trills: Giving the People and their Pension Funds a Stake in the Wealth of the Nation. Cowles Foundation Dis­cussion Paper No. 1717, August 2009.

Kamal, L. and Lashgari, M. (2012): Comparing GDP-Indexed Bonds to Standard Government Bonds, Journal of Applied Business and Economics 13(2), 116-128

Kitzmueller, C. (2014): A Euro Area wide Unemployment Insurance to Improve Macroeconomic Stability. Munich: GRIN-Verlag. Available online: http://www.grin.com/en/e-book/287605/a-euro-area-wide-un­employment-insurance-to-improve-macroeconomic-stability

Krugman, P. (1988): Financing vs. Forgiving A Debt Overhang. NBER Working Paper No. 2486, January 1988.

Lane P. (2012): The European Sovereign Debt Crisis, Journal of Economic Per­spectives 26/3, 49–68.

Levy-Yeyati, E. and Panizza, U. (2011): The Elusive Cost of Sovereign Defaults, Journal of Development Economics 94/1, 95-105.

Pisani-Ferry, J.; Vihriala, E. and Wolff, G. (2013): Options for a Euro-Area Fiscal Capacity. Bruegel Policy Contribution 2013/1, January 2013.

Rodrik, D. and Velasco, A. (1999): Short-Term Capital Flows. NBER Working Paper No. 7364, September 1999.

Ruban, O.; Poon, S. and Vonatsos, K. (2008): GDP Linked Bonds: Contract De­sign and Pricing. Available online at SSRN: http://ssrn.com/abstract=966436.

Rodrik, D. and Velasco, A. (1999): Short-term capital flows. NBWE Working Pa­per 7364, September 1999.

Shambaugh, J.C. (2012): The Euro’s Three Crises. Brookings Papers on Eco­nomic Activity 44/1, 157-231.

Shiller, R. J. (1993): Macro Markets - Creating Institutions for Managing Soci­ety’s Largest Economic Risks. Oxford: Clarendon Press.

Wolff, G.B. (2012): A Budget Capacity for Europe’s Monetary Union. Bruegel Policy Contribution, 2012/22, December 2012.

9 Appendix

illustration not visible in this excerpt

Table 1: Selected Eurozone countries – nominal GDP baseline and actual nominal GDP (in billion euros)

illustration not visible in this excerpt

Table 2: Selected Eurozone countries – STANDARDIZED nominal GDP baseline and ac­tual nominal GDP (in billion euros)

illustration not visible in this excerpt

Table 3: Selected Eurozone countries – actual nominal GDP growth and stock market performance

[...]

---

^[1] In this context, the most widely discussed mechanisms are Dullien’s (2008) Eurozone-wide unemployment insurance scheme and the cyclical shock insurance scheme proposed by Enderlein et al (2013).

^[2] F.e. Borensztein et al (2004) and Griffith-Jones and Hertova (2013).

^[3] cf Kitzmueller 2014: 27-59

^[4] cf Shambaugh 2012: 159

^[5] cf Lane 2012: 61

^[6] cf Shambaugh 2012: 187

^[7] The SGP was formally implemented by way of three legal documents and entered into force on 1 January 1999: European Council Resolution of 17 June 1997 on the stability and growth pact, the Council Regulation (EC) No 1467/97 of 7 July 1997 on speeding up and clarifying the implementation of the excessive deficit procedure. Council Regulation (EC) No 1466/97 of 7 July 1997 on the strengthening of the surveillance of budgetary positions and the surveillance and coordination of economic policies.

^[8] The Six-Pack was implemented via six distinct legal changes: Regulation (EU) No 1175/2011 of the European Parliament and of the Council of 16 November 2011 amending Council Regulation (EC) No 1466/97 on the strengthening of the surveillance of budgetary positions and the surveillance and coordination of economic policies. Council Regulation (EU) No 1177/2011 of 8 November 2011 amending Regulation (EC) No 1467/97 on speeding up and clarifying the implementation of the excessive deficit procedure. Council Directive 2011/85/EU of 8 November 2011 on requirements for budgetary frameworks of the Member States. Regulation (EU) No 1173/2011 of the European Parliament and of the Council of 16 November 2011 on the effective enforcement of budgetary surveillance in the euro area. Regulation (EU) No 1176/2011 of the European Parliament and the Council of 16 November 2011 on the prevention and correction of macroeconomic imbalances. Regulation No 1174/2011 of the European Parliament and the Council of 16 November 2011 on enforcement measures to correct excessive macroeconomic imbalances.

^[9] The Two-Pack was set out in two additional EU regulations: Regulation (EU) No 472/2013 of the European Parliament and of the Council of 21 May 2013 on the strengthening of economic and budgetary surveillance of member states in the euro area experiencing or threatened with serious difficulties with respect to their financial stability. Regulation (EU) No 473/2013 of the European Parliament and of the Council of 21 May 2013 on common provisions for monitoring and assessing draft budgetary plans and ensuring the correction of excessive deficit of the member states in the euro area.

^[10] cf Kitzmueller 2014: 27-49

^[11] cf De Grauwe 2011: 266

^[12] cf Pisani-Ferry et al 2013: 4

^[13] cf Wolff 2012: 5

^[14] cf Feld/Osterloh 2013: 6

^[15] See the “Four Presidents” in their report on a genuine EMU (EUCO 2012) and the Commission’s blueprint for a deep and genuine economic and monetary union (COM 2012).

^[16] cf Borensztein et al 2004: 1

^[17] cf Borensztein et al 2004: 1-2

^[18] cf Borensztein et al 2004: 7-11

^[19] cf Borensztein et al 2004: 3

^[20] See for example Rodrik/Velasco 1999: 23-28

^[21] Rodrik/Velasco 1999: 22-23

^[22] cf Borensztein et al 2004: 14

^[23] cf Borensztein et al 2004: 14-15

^[24] cf Borensztein et al 2004: 15-16

^[25] cf Borensztein et al 2004: 16

^[26] cf Borensztein et al 2004: 18

^[27] cf Borensztein et al 2004: 11-13

^[28] cf Borensztein et al 2004: 13

^[29] cf Borensztein et al 2004: 23

^[30] cf Borensztein et al 2004: 24

^[31] cf Borensztein et al 2004: 27

^[32] cf Borensztein et al 2004: 27

^[33] cf Borensztein et al 2004: 29

^[34] Krugman 1988: 1

^[35] cf Shiller 2009: 20

^[36] cf Barro 1995: 1-3

^[37] cf Borensztein et al 2004: 29

^[38] cf Borensztein et al 2004: 29-30

^[39] cf De Paoli et al 2006: 15-17

^[40] cf Levy Yeyati/Panizza 2011: 14

^[41] cf Borensztein et al 2004: 31

^[42] cf Griffith-Jones/Hertova 2013: 34

^[43] See Kamstra/Shiller 2009 and Shiller 1993

^[44] See Borensztein/Mauro 2002

^[45] cf Borensztein/Mauro 2002: 8-11

^[46] cf Kamstra/Shiller 2009: 14

^[47] cf Chamon/Mauro 2005: 22

^[48] cf Ruban et al 2008: 35-36

^[49] cf Chamon and Mauro 2005: 3

^[50] cf Griffith-Jones/Sharma 2006: 2-3

^[51] cf Barr et al 2014: 30

^[52] cf Borensztein et al 2004: 39

^[53] cf Barr et al 2014: 30-31

^[54] cf De Paoli et al 2006: 2-6

^[55] cf Levy-Yeyati/Panizza 2006: 14

^[56] cf De Paoli et al 2006: 15

^[57] cf Griffith-Jones/Sharma 2006: 2

^[58] cf Kamal/Lashgari 2012: 126

^[59] cf Griffith-Jones/Sharma 2006: 3

^[60] cf Griffith-Jones/Sharma 2006: 3

^[61] cf Griffith-Jones/Sharma 2006: 8

^[62] cf Borensztein/Mauro 2004: 51 and 53

^[63] cf Borensztein/Mauro 2004: 48

^[64] cf Griffith-Jones/Sharma 2006: 9

^[65] cf Griffith-Jones/Sharma 2006: 9-10

^[66] cf Griffith-Jones/Sharma 2006: 2

^[67] cf Borensztein/Mauro 2004: 48

^[68] Austria, Belgium, Cyprus, Finland, France, Germany, Greece, Ireland, Italy, Netherlands, Portugal, Slovenia, and Spain

^[69] See online: http://finance.yahoo.com/market-overview/ and http://www.imf.org/external/pubs/ft/weo/2015/02/weodata/weoselgr.aspx

^[70] cf Borensztein/Mauro 2002: 8-11

^[71] See online: https://www.imf.org/external/pubs/ft/weo/2008/01/weodata/index.aspx

^[72] See online: http://sdw.ecb.europa.eu/home.do

^[73] See online: http://ec.europa.eu/economy_finance/ameco/user/serie/SelectSerie.cfm

^[74] cf Shambaugh 2012: 167

^[75] cf Lane 2012: 57

^[76] Please note that the simulation disregards the financial rescue programs provided by European funds to several distressed Eurozone countries, and moreover does not factor in the haircut that Greece forced upon its creditors.

^[77] See for instance Borensztein and Mauro (2002) or Kamstra and Shiller (2009)