Open Economies Review

, Volume 25, Issue 1, pp 71–91 | Cite as

The Tragedy of the Commons and Inflation Bias in the Euro Area

  • Valeriya Dinger
  • Sven Steinkamp
  • Frank Westermann
Research Article


Central bank credit has expanded dramatically in some of the Euro Area member countries since the beginning of the financial crisis. This paper makes two contributions to understand this stylized fact. First, we discuss a simple model of monetary policy that includes (i) a credit channel and (ii) a common pool problem in a monetary union. We illustrate that the interaction of the two elements leads to an inflation bias that is independent of the standard time-inconsistency bias. Secondly, we present an institutional analysis that is consistent with the view of fragmented monetary policy and empirical evidence that illustrates the heterogeneity of central bank credit expansion.


Tragedy of the Commons Inflation bias Credit channel TARGET2 Euro area 

JEL Classifications

E52 E58 H41 

1 Motivation

Central bank credit has expanded on an order of magnitude that is unprecedented in the post-war history of Europe. In countries with a negative output shock–Greece, Italy, Ireland, Portugal and Spain–it has increased by more than 1,000 % from 2007 to 2012. In this paper, we review the institutional arrangements in the euro area and present a theoretical model, as well as empirical evidence, that help to understand this stylized fact. We argue that the expansion was facilitated by a fragmentation of the ECB’s monetary policy that is implemented by the national central banks. This fragmentation gives rise to classical tragedy of commons problems and generates a positive inflation bias.

The institutional changes that set the stage for a tragedy of the commons-dilemma have started in 2007, when the ECB introduced the TARGET2 clearing system.1 This clearing system linked the money markets across Europe, creating a common pool of money demand to which all central banks had access.2 Furthermore, the ECB, which in principle controls monetary policy, announced a “full allotment” policy. Under this policy, private banks had unlimited access to central bank credit, provided that they were able to pledge collateral that was acceptable to the ECB. The national central banks (NCBs) finally gained control over monetary policy in their jurisdiction, when the ECB relaxed their collateral standards. An important institutional aspect is that the NCB’s assess the quality of this collateral as well as the solvency of their counterparties. It is thus the implementation of monetary policy, not the decision making that is the source of the common pool problem.3

To motivate why such an institutional setting leads to an additional inflation bias in monetary policy, we set up a simple partial-equilibrium model that starts with a standard central bank utility function, including the output gap and inflation. We add two non-standard elements to this model: (i) A credit channel of monetary policy. We assume that the national central banks can have a direct impact on the domestic economy by extending central bank credit to relax credit constraints. (ii) We assume that there is a common pool problem in a monetary union of the following type: Central banks can extend credit to private banks in their jurisdiction. However, the associated increase in money supply is not confined to lead to inflation in this particular country. There exists a single money market in the currency union, thus prices will increase in all countries. By extending central bank credit the NCB’s reap the full marginal benefit of their policy, but face only the average cost in terms of inflation.4

We show in the model section that there is an inflation bias resulting from this common pool problem that is independent of the standard time-inconsistency bias.5 Our model builds on an earlier literature on inflation bias in currency unions, in particular Aizenman (1992) and Casella (1992). In their papers, the inflation bias arises from externalities among jurisdictions that are competing for seigniorage and the inflation tax. On an abstract level, our model can be viewed as a simple partial-equilibrium representation of these ideas. Furthermore, it takes special features of the current euro crisis into account. This includes the central banks objective of achieving financial stability–i.e. avoiding the cost of bank closures–and the credit crunch suffered in several European countries after the 2007/8 financial crisis.

In the empirical section of the paper, we illustrate key stylized facts about central bank credit expansion. Among these, three observations stand out: First, countries that recently experienced a negative output shock have increased central bank credit substantially. Furthermore, this central bank credit did not stay within the national borders of the respective country. It has been wired to other countries via the TARGET2 clearing system to buy goods and assets abroad and to repay international loans, thus increasing the money supply throughout the whole monetary union. The absolute values of TARGET2 liabilities are nearly identical to the increase in central bank credit.6 Secondly, we illustrate that central bank lending already constitutes a substantial share of total lending in the countries in crisis. In the GIIPS countries (Greece, Ireland, Italy, Portugal and Spain), it has increased from 1.6 % in 2007 to 16.4 % in 2012. Third, consistent with our common pool argument, money M1 and prices have increased only moderately in the countries that experienced the negative shock. This could of course also be due to deflationary pressure in these countries.

The findings of our model reflect the option of individual countries to push their fiscal limit (Davig et al. 2012) without immediately bearing the political costs of domestic inflation. In section 5, we provide an extended review of related common pool problems in fiscal policy. They also raise important questions about the long-term consequences of non-standard monetary policy measures, such as a reduction of collateral standards and full allotment policies. Lastly, the inflationary bias documented in the model is a potential explanation of why, despite strong deflationary pressures currently at force in some euro area countries, observed inflation rates are still positive.

In section 2 of the paper, we discuss in more detail the institutional arrangements. In section 3, we present the theoretical model. Section 4 illustrates some stylized facts. In section 5, we relate our findings to previous studies in the literature. Section 6 concludes by pointing to concrete policy proposals that help to internalize existing externalities.

2 Institutional Framework of the Eurosystem

The tragedy of the commons (or common-pool problem) describes a situation when multiple players operate in a framework where it is possible that each individual player extracts some benefits from exploiting a common pool of resources while paying only a portion of the costs. The key institutional questions that we review in this section are: What is the common pool that is overly used, and why do national central banks have access to this common pool?

The common pool is the euro-area wide money demand.7 It has been created in several steps. First, the introduction of the euro itself eliminated exchange rate risks among member countries and created an integrated capital market. The ECB in Frankfurt centrally sets the interest rate for all countries of the monetary union. However, national central banks continued to exist. The allocation of money creation by the different national central banks, until 2007, was mainly driven by the demand of private banks for central bank credit. This demand, however, was limited by the money demand in the respective country. This was changed when the TARGET2 clearing system was introduced that contained a so-called “real-time settlement system”. In this settlement system, national central banks have the task to execute transfer payments, even before the incoming funds have arrived.

In principle this leads to imbalances, which last only for a few seconds. However, in the financial crisis, these balances have become more and more persistent, accumulating to substantial levels of claims and liabilities between the central banks of the Eurosystem (see Sinn and Wollmershäuser (2012) and Garber (1999) for further details). With regard to the common pool problem, this meant the following: the national central banks where no longer limited by the demand for money in their own country, but instead they could service the entire euro-area wide money demand. While the demand for money used for domestic purposes was small, the demand for money for the purpose of international transactions was very large.

2.1 Decision on Eligible Counterparties

The second institutional question is how the national central banks gained access to this common pool. The classical perception of monetary policy operations within the EMU excludes this possibility. The ECB is supposed to provide uniform monetary policy by setting uniform interest rates and equal conditions for the access of banks from all participating countries to central bank funding. If this is indeed the case–as it was broadly the case until the end of 2007–exploiting the common pools should be hardly possible for any individual member of the union.

However, since 2007, individual countries gradually gained control over credit extension so that - even though the ECB still determines a single interest rate for central bank credit–substantial heterogeneity across member countries exists as to the quality of financial institutions and to the collateral to which this interest rate applies. To start with, the national regulators de facto decide on the list of banks eligible to receive central bank’s funds.8 This is the case since the ECB declares all solvent banks eligible for central bank refinancing, but the definition by which banks are solvent or not is still made by the national financial regulation authorities.9 The ECB clarified in the debate about the solvency of the Cypriot Laiki bank that the NCBs are solely responsible for solvency supervision of their counterparties as well.10 By not closing down virtually insolvent banks national authorities achieve at least two targets. First, they avoid both the political and economic costs of the liquidation of these banks. Second, they exploit the advantage of the fact that these banks’ remaining assets–instead of being liquidated under the currently adverse conditions–can be used as collateral to generate increased central bank credit to the domestic banking sector. This is why the right of national regulators to decide on the solvency of the banks is an important determinant of the amount of central bank’s credit to commercial banks and a key component of the tragedy-of-the-commons problem.

2.2 Collateral Standards

Another important determinant of this amount is the gradual reduction of collateral standards by the ECB starting in October 2008. In fact, from the end of 2011, the European Central Bank has virtually given up control over the eligible collateral on the central banks refinancing operations. This fact is nicely illustrated by a Governing Council decision announced by the ECB on February 9th, 2012 which allows “specific national eligibility criteria and risk control measures for the temporary acceptance of additional credit claims as collateral in Eurosystem credit operations” (italics added).

Again, not only the definition of eligible collateral has been delegated to the NCBs. They are also responsible for the collection of necessary information on the quality of potential collateral and the concluding eligibility assessment.11

This decision puts the national central banks in charge of the decision about which assets can serve as eligible collateral and which cannot. The fact that the ECB abandons control over collateral quality is also illustrated by a most recent decision of the ECB announced on September 6th, 2012 which suspends the application of the minimum credit rating threshold in the collateral eligibility requirements.

The expansion of eligible collateral categories affects not only the volume of central bank’s funding. In combination with the relatively rough grid of asset categories used for the determination of the collateral haircuts it also gives rise to substantial variation of the costs of central bank funding for banks from different countries. Given almost uniform haircuts for each asset category banks with more risky assets enjoy an advantage in the de facto costs of funding relative to banks with safer assets from the same haircut category.

In sum, the volume of refinancing can significantly differ across countries. In the beginning of the crisis, this was due to a built-in flexibility under existing rules. But since the ECB’s reduction of collateral standards, this expansion has been increasingly a result of the national central bank’s policies. 54 % of the total expansion occurred after December 8th, when the most significant drop in collateral standards was announced. The former ECB Chief Economist, Jürgen Stark, recently summarized these developments in the statement The ECB is about to lose its ability to perform uniform monetary policy.12

3 A Model of Monetary Policy with a Credit Channel and a Common Pool Problem

In this section, we illustrate the effect of a common pool problem in a simple model of monetary policy. We start from a standard loss function for the central bank (see e.g. Walsh (2010) and Barro and Gordon (1983)) and add two new elements. First, we assume that there is a direct effect of central bank credit on output. This assumption can be motivated by the literature on the credit channel of monetary policy (Kashyap, Stein and Wilcox (1993) and Tornell and Westermann (2005)). In the framework of the European problem at hand one can also think of the positive effects in terms of avoiding the costs of liquidating financial institutions. These costs will be represented by increased financial uncertainty and instability, as well as reduced aggregate investment due to limited access to credit.

Secondly, we assume that there is a common pool problem in the currency union as discussed above. Each individual central bank can extend credit to its banks. By doing so the central bank reaps the full benefits from this credit extension, but only bears the average loss in the form of the average inflation in the euro area.13

4 A Single Country

Let us start with a single country as a point of reference, where only the credit channel is added to the standard model of the central bank’s optimization problem. The notation of the model is as follows. The utility function of a national central bank is given by U(y,π), where y denotes output and π denotes inflation. The utility function is taken from a standard textbook. The central bank gets positive utility from closing the gap between output y and the exogenously given potential output \( \overline{y_n} \). Furthermore, there is a quadratic loss from inflation. λ denotes the weight attached by the central bank to closing the output gap.
$$ U\left(y,\pi \right)=\lambda \left(y-\overline{y_n}\right)-\frac{1}{2}{\pi}^2 $$
The output function consists of two components. First, the potential output, \( \overline{y_n} \), and a cyclical component, yc, that depends on the change of central bank credit in the economy, which in turn depends on the central banks supply of credit to commercial banks, Δd. This later term is kept very general and could take various functional forms. For the moment we only assume that the effect of central bank’s credit on output is positive \( \frac{\partial y}{\partial \Delta d}>0 \).14
$$ y=\overline{y_n}+{y}_c\left(\Delta d\right) $$

In order to link this credit channel to money supply and inflation, we furthermore assume that the change in the monetary base is equal to the change in the amount of central bank’s credit provided to the commercial banks: Δd = Δm where m denotes the monetary base (we thus abstract from other forms of monetary expansions that would for instance follow from explicit bond purchases of the central bank, as well as minimum reserve holdings).15

We further assume that inflation is a function of monetary base changes,  πm), where \( \frac{\partial \pi }{\partial \Delta m}>0 \). The value of this partial derivative depends on the money multiplier and in particular on the commercial banks’ reserves with the central bank (e.g. the higher the propensity of commercial banks to distribute the funding received from the central bank to the non-financial sector, the higher the multiplier). Since these are not at the core of the analysis presented here, we assume for simplicity that \( \frac{\partial \pi }{\partial \Delta m}=1. \)

Lemma 1

In a single country there is an inflation bias of \( \lambda \frac{\partial y}{\partial \Delta m} \).


\( \mathrm{ARG}\;{\mathrm{MAX}}_{\varDelta m}U=\lambda {y}_c\left(\Delta \mathrm{m}\right)-\frac{1}{2}{\left(\Delta m\right)}^2;\frac{\partial U}{\partial \Delta m}=\lambda \frac{\partial y}{\partial \Delta m}-\Delta m=0;\varDelta {m}^{\ast }=\lambda \frac{\partial y}{\partial \varDelta m}. \)

The intuition for this result directly follows from the utility function of the central bank. Since inflation enters as a quadratic term, most functional forms for ycd) will generate a positive equilibrium inflation. For instance if ycd) is linear, the additional inflation would simply be a constant added to the term λ.

4.1 A Currency Union

In this section we now extend the model to a currency union, with i = 1, …, n countries. In this currency union, each country has its own central bank utility function Ui(yi,πi). It also has a country specific output function yi and the currency unions inflation rate, πi. The monetary base in each country is denoted by mi. Again, changes in the monetary base are equal to the changes in credit provided by the central banks to the domestic banking system, di. In the utility and output functions we furthermore assume that the potential output is the same for all countries:
$$ {U}_i\left({y}_i,{\pi}_i\right)=\lambda \left({y}_i-\overline{y_n}\right)-\frac{1}{2}{\pi}_i^2 $$
$$ {y}_i=\overline{y_n}+{y}_c\left(\Delta {d}_i\right) $$
with \( \frac{\partial {y}_i}{\partial {\Delta \mathrm{d}}_{\mathrm{i}}}>0 \). The common pool problem in this setup follows from the fact that in an integrated capital market of a currency union the inflation rate is a positive function of each country’s change in monetary base:16
$$ \pi =\frac{1}{n}{\displaystyle {\sum}_{i=1}^n\Delta {m}_i} $$

This is the case because changes in the monetary base do not stay in the country of origin but can spread across the currency union, i.e. inflation arising from expansionary policies is shared across member countries. Each country thus has the full marginal benefit of central bank credit extension that enters its output function, but faces only the average marginal cost that derives from inflation.

Proposition 1

In a currency union with a common pool problem, the inflation bias is larger than in a single country.


\( \mathrm{ARG}\ {\mathrm{MAX}}_{\Delta {m}_i}{U}_i=\lambda {y}_i\left(\Delta {m}_i\right)-\frac{1}{2}{\left(\frac{1}{n}{\displaystyle {\sum}_{i=1}^n\Delta {m}_i}\right)}^2; \) In symmetric equilibrium: \( \frac{\partial {U}_i}{\partial \Delta {m}_i}=\lambda \frac{\partial {y}_i}{\partial \Delta {m}_i}-\frac{1}{n}\Delta {m}_i=0;\Delta {m}_i^{\ast }= n\lambda \frac{\partial {y}_i}{\partial \Delta {m}_i}; n\lambda \frac{\partial y}{\partial \Delta m}-\lambda \frac{\partial y}{\partial \Delta m}=\frac{\partial y}{\partial \Delta m}\lambda \left(n-1\right)>0. \)

The intuition for our proposition follows from the tragedy of the commons. Each country has an incentive to exploit the credit channel effect of its central bank credit extension, but it only bears the average cost of such action. The temptation to extend central bank credit and to contribute to the average inflation rate is, therefore, extraordinarily high. In the presence of deflationary pressures the inflationary bias documented here can play a mitigating role and explain why we actually observe less deflation in countries with substantial negative output shocks than one might have otherwise expected.

5 Stylized Facts

The simple model outlined above helps to explain the pattern of central bank credit expansion in the euro area since the beginning of the 2007/8 financial crisis. In this section, we document the recent developments by pointing out three stylized facts that are consistent with the view that national central banks have indeed gained room to implement country specific monetary policy.17

Stylized Fact 1

Countries with a negative output shock have expanded central bank credit substantially.

Most of the world’s major economies are engaged in expansionary monetary policy; the ECB’s situation, however, differs considerably from other central banks. While the increase of the Eurosystem’s aggregate balance sheet from the beginning of 2007 is comparable with other central banks (see Fig. 6 of Appendix C), a remarkable feature of the European financial crisis is that national central banks of the Eurosystem expanded credit to domestic banks in an asymmetric pattern. While central banks in countries with negative output shocks have expanded credit to domestic commercial banks (in particular Greece, Ireland, Portugal and Spain), countries which did not experience such a negative shock kept their central bank’s credit relatively constant or even reduced it. Figure 1 shows that the credit expansion in the crisis countries is unprecedented in post-war history. On average, it amounts to a more than 1,000 % increase in less than 5 years, or more than 800bn Euros in absolute terms.18,19 This phenomenon reflects in part the fact that in these countries the central bank was taking over the liquidity insurance role earlier provided by private interbank markets. In the presence of sharp reversals in private capital flows, they have replaced private capital by central bank credit.
Fig. 1

Central bank credit and TARGET2 balance

The dashed series in Fig. 1 illustrates that the money extended to the domestic banking system was not used to purchase goods or assets domestically, but rather financed international transactions. The TARGET2 balances measure the international balance of payments within the European Monetary Union (See Sinn and Wollmershäuser (2012)).20 The fact that central bank’s credit is used to facilitate transactions outside the individual country, but within the EMU, illustrates the spread of inflationary pressures generated by individual countries expansionary policy to other members of the Union.

Stylized Fact 2

Private bank lending has increasingly been funded by central bank credit.

Figure 2 shows that the commercial banks have funded a substantial share of their lending to the non-financial sector by credit from the central bank. The average share of total lending to the non-financial sector funded by central bank credit went up from around 1.6 % in 2007 to 16.4 % in the third quarter of 2012. The dynamics of central bank’s credit in the individual countries is illustrated in Fig. 5 of the appendix. It shows that the issue is particularly severe in Greece and Ireland. This fact is indicative for the potential benefit of central bank credit for the national economies. Furthermore, it is in line with recent results of Jiménez et al. (2012), who found out that the credit channel of monetary policy is particularly strong in environments of weak bank balance sheets.
Fig. 2

Ratio of central bank credit to private bank lending

The central banks thus acted to prevent a credit crunch that would have occurred without the intervention of the central banks, with potentially severe negative consequences for the real economy.

Stylized Fact 3

Central bank credit correlates positively with unemployment, but not with inflation.

Figure 3 shows that central bank credit reacts to negative output shock as approximated by the level of unemployment. It does not seem to be related to inflation. This fact is consistent with our argument that the policy has been to exploit the common pool of euro-wide money demand for the stabilization of domestic financial systems and avoiding the short-term costs of the liquidation of financial institutions without internalizing any potential longer-term inflation biases. Figure 4 in Appendix B of the paper shows that the same pattern also holds at the country-level.
Fig. 3

Central bank credit, unemployment and inflation

Finally, the monetary expansion in the countries in crisis was feasible without any substantial impact on domestic inflation.21 As the additional money was used primarily to purchase goods and assets abroad, domestic prices were little affected. In Fig. 3, both unemployment and domestic prices are displayed on the same (left) scale, starting at 100. While unemployment increased by a factor of 2.5 on average, domestic prices had a cumulative increase of only about 12 % during the same period. In fact, as the crisis has not reached the largest countries in the euro area, the central bank expansion has left the total monetary base relatively unchanged until the end of 2011. And, the aggregate increase is still not remarkably large when compared to the US Federal Reserve Bank or the Bank of England. The future inflationary effect of the Euro Area’s expansionary behavior will therefore crucially depend on whether the ECB gets the tragedy-of-the-commons problem under control.

6 Related Literature

A natural area of application of the common-pool problem approach in macroeconomics has been the use of common-tax-pools in fiscal policy. Weingast, Shepsle and Johnson (1981) apply a version for the common-pool problem to the inefficiency of regional fiscal distributions by arguing that whenever central parliaments decide on the funding of regional projects an overspending bias will arise. This is the case since the members of parliaments are elected from certain regional areas and as such overvalue the interest of these regions. From the regional perspective the benefits of a locally executed project, which is funded by the central budget, will always be overvalued since the full marginal benefit of the project for the region is not weighed against the full marginal costs of the project, but rather against the costs covered by the region, which is only a negligible portion of the full costs.

Alesina and Drazen (1991) take the interpretation of the common pool problem of fiscal policy further. They argue that the common-pool problem is not only a potential source of fiscal instability, but it is also at the core of delayed fiscal stabilization. This conclusion is based on applying the common-pool problem in a political economy framework with heterogeneous groups, where delayed fiscal stabilization is modeled as the result of disagreements upon the distribution of the costs and benefits of stabilization.

The common-pool problem issues of fiscal policy have also been extensively studied with regard to the functioning of fiscal unions (see Knight (2003) and Alesina and Perotti (1999) for detailed discussions of studies documenting the common-tax pool problem and proposed solutions). With this regard the common–pool problem has typically been employed in the analysis of the funding of infrastructure or other locally used public goods where a large portion of the benefits stays within one member of the union, while the costs are symmetrically distributed across all participants (Basley and Coate (2003)). The objects of the analysis in this strand of the literature have been examples for public overspending from around the world ranging from the US (Feldstein and Vaillant (1998)) and the European Union (e.g. Hallerberg and von Hagen (1999)) to developing countries (e.g. Kletzer and Singh (1997) and Hausmann and Purfield (2004)).

In the years following the decision for the establishment of the European Monetary Union a broad strand of the literature has been focused on the analysis of fiscal common-pools problems in the European Union. One strand of this literature focused on the interaction between the loss of monetary and fiscal stabilizers in the framework of the monetary union (Galí and Perotti 2003). Another strand of this literature has been concerned with analyzing whether the existence of common currency will generate bail-out expectations for countries in fiscal distress and thus aggravate the standard common-pool problem (von Hagen and Eichengreen (1996), Chari and Kehoe (2008), Krogstrup and Wyplosz (2010)). These studies recognize that if the fiscal issues arising from the common-pool problem are not successfully solved these would generate incentives to jeopardize monetary policy stability, since in this case the ECB is forced to accommodate the lax fiscal policy and engage in bail-outs as we observe now. However, this literature has so far assumed that the ECB will keep its ability and willingness to perform uniform monetary policy, so that even if a bail-out is decided the limits and the conditions of the bail-out will be determined by the ECB. As we discussed above the recent undermining of the ECB’s institutional setting has raised substantial concerns about whether this is indeed the case. In particular, individual member countries have been given the opportunity to modify the stance of the Union’s monetary policy which in turn has created incentives to apply monetary policy tools such as central bank’s credit to commercial banks in a way that is prone to the emergence of common-pool type externalities.

This common-pool distortion of the incentives to create inflation has so far only been discussed by Aizenman (1992), Casella (1992) and von Hagen and Süppel (1994). Similar to the set-up analyzed in our paper Aizenman shows that the inflationary bias will be high if the optimal inflation rate is set by several decision makers rather than by a centralized decision maker. We generally come to a similar conclusion with regard to inflationary biases as Aizenman (1992). However, our approach differs from his in that we focus on the trade-off between the benefits of credit expansion and costs of inflation, while Aizenman (1992) is concerned with the optimality of inflation tax from a Laffer curve perspective. Casella (1992) and von Hagen and Süppel (1994) discuss how the inflationary bias depends on the decision structure of the monetary union’s central bank. While in our framework we allow for completely decentralized monetary policy within the union, these authors assume common monetary policy and discuss, which is the optimal design of the decision about these common monetary policy rules and what are the incentives of countries with small impact on the monetary decision making process to participate in the union.

The possibility of exploiting the ECB monetary policy tools as a common-pool has only been analyzed by few works, none of which explicitly covers inflation biases. In a policy paper Tornell and Westermann (2012a, 2012b) suggest that the implicit bank bail-outs given by the lax monetary policy and emergency funding in the euro area are an example for the common-pool problem, since each of the regional banking supervisor can (at least in the short-run) achieve the benefit of not having to bear the costs of bank liquidations, while it’s paying only a minor share of the costs generated by the explosion of central bank’s credit to the banking sector. Further, Tornell (2012) discusses the emergence of overwhelming TARGET2 balances as an expression of the common-pool problem. He presents a formal general-equilibrium model on the political economy of TARGET2 balances.

Buiter (2012) presents a detailed review of the current ECB institutional framework and draws parallels between the ruble union and the current ECB policy. Indeed, Lipton and Sachs (1992) and Eichengreen et al. (1993) documented a similar free-rider problem faced by the Russian central bank during the disintegration of the Soviet Ruble Zone, 1991–1995. While the Central Bank of Russia maintained control over physical money creation, each central bank of the ruble zone had the authority to issue ruble credits. Lipton and Sachs conclude:

“[…] each of the fifteen independent states had a central bank with the ability to exercise an important degree of control over monetary policy. […] In our view, there is no realistic possibility of controlling credit in a system in which several independent central banks each have the independent authority to issue credit. The reason is simple. Pressure is overwhelming in each of the states to “free ride” by issuing ruble credits at the expense of the rest of the system. It is a nearly self-evident proposition that a single currency area should have a single bank of issue.” (author’s emphasis)

Our paper is also related to the recent literature on the potential risks stemming from the expansion of the central bank balance sheet. Williamson (2013) argues that while the Fed’s balance sheet has increased by a factor of almost 4 between 2007 and 2013 no direct inflation consequences follow since the expansion simply counteracts misintermediation by the private sector. In his argument the price level in a monetary system close to the zero bound is determined by the demand and supply of all assets that can be intermediated and transformed into assets that are used in exchange. Given substitutability between private and public intermediation and maturity transformation of these assets, the size of the central bank balance sheet is irrelevant for inflation. Our argument essentially differs from Williamson’s in that we describe a segmentation of public and private asset markets along the borders of the Euro Area member countries that makes the size of central bank credit relevant.

Finally, our paper is reminiscent of the competition among bank regulators that has been modeled by Sinn (2003). While Sinn models the race-to-the-bottom with regard to capital-ratios, illustrating that national regulators neglect the external effect on other countries, his model could be extended to the decision of whether or not a bank is classified “solvent”, or the quality of collateral acceptable for central bank refinancing.

7 Conclusions

In this paper we show how a tragedy-of-commons, stemming from the institutional shortcomings of the European Monetary Union, gives rise to excess central bank credit and an increased inflation bias.

We present a brief description of the institutional features of the EMU. We then discuss the inflation effects in the framework of a simple model of monetary policy where the central bank minimizes a loss function with two arguments: the output gap and inflation. We add two new elements two this model. First, we directly include a credit channel effect: the central bank can generate positive output changes by expanding credit to domestic banks. Second, we model the common-pool problem by assuming that while the positive credit channel effects are fully appropriated at “home” the inflation biases generated by the credit extensions and the increased money supply are shared across all union members.

We support the tragedy-of-commons argument and the implications of the theoretical model by presenting an empirical examination of the dynamics of central bank’s credit, monetary aggregates and unemployment in the EMU. Plotting the dynamics of these variables we graphically show the abrupt expansion of central bank credit in some EMU countries, which correlates with unemployment, has not increased money supply and inflation in these countries.

We do not argue that a period of high inflation–or even hyperinflation–is inevitably around the corner, but rather that a decentralized monetary policy in a currency union gives rise to expansionary pressure and thus may lead to a higher inflation compared to a situation with only one central authority being able to create money.22 At the moment, a fragmented monetary policy and the resulting export of inflationary pressure may even be beneficial for the euro area as a whole by supporting price convergence and thus reducing internal balance-of-payments imbalances. Whether the currently observed expansionary bias will result in high inflation levels will crucially depend on whether the ECB will be able to implement an exit-strategy when the deflationary pressure loses momentum. In principle the ECB could centralize monetary policy at any time. It is, however, debatable on whether it can successfully implement a contractionary policy in spite of the disparities in member countries banking systems.

If a price stability target should be achieved in the long-run, the monetary policy common-pool problems presented in this paper illustrate the need for an institutional reform of the Union. Applying the implications of the literature focused on fiscal common pools problems to the area of monetary policy common-pool problems discussed in this paper would suggest that creating a stronger institutional framework, which is able to endogenize the externalities of excessive monetary policy expansion, is essential. In particular, a centralized decision making process about the key features of monetary policy and a uniform implementation of this policy can help mitigate the problems arising from the common-pool incentives to access central bank’s credit. These should include both a centralized decision making about the solvency of banks, as well as going back to a uniform catalogue of eligible collateral. The joint Euro Area single supervisory mechanism in its currently scheduled form will not be sufficient both because it only covers a small share of banking institutions eligible to central bank credit and because it does not require the uniform treatment of collateral.


  1. 1.

    The tragedy of the commons interpretation was also given in Tornell and Westermann (2012a, 2012b) and Tornell (2012).

  2. 2.

    See Sinn and Wollmershäuser (2012) and Garber (1999), as well as Section 2 of this paper for more institutional details and the economic interpretation of TARGET2 balances.

  3. 3.

    See von Hagen and Süppel (1994) for an analysis of a common pool problem in the central bank decision making.

  4. 4.

    More broadly, the cost could also include potential losses of extending credit to illiquid banks. These would be shared by the capital key in the Eurosystem. In our model, we abstract from this possibility.

  5. 5.

    See e.g. Barro and Gordon (1983) and Walsh (2010).

  6. 6.

    See also Neumann (2012).

  7. 7.

    See also Tornell (2012) and Tornell and Westermann (2012a).

  8. 8.

    For the purpose of the subsequent analysis we can view central banks and national regulators as one entity. In the policy conclusions we highlight the need for both, a common regulation and a uniform catalogue of eligible collateral.

  9. 9.

    See Tornell and Westermann (2012a).

  10. 10.

    “Die zyprische Notenbank ist dafür verantwortlich, die Solvenz ihrer Banken zu beurteilen”, ECB spokeswoman, June 26th, 2013, Süddeutsche Zeitung.

  11. 11.

    See e.g. European Parliament (2013), Question P-004750/2013. See also ECB website: “Prior to the publication […] in the list of eligible marketable assets, national central banks (NCBs) proactively assess the eligibility of the marketable assets. The NCB of the country where the asset is admitted to trading on a regulated market or traded on a non-regulated market is responsible for the assessment of the eligibility of the marketable asset” (

  12. 12.

    Frankfurter Allgemeine Zeitung, October 24, 2012.

  13. 13.

    Also in the case of losses, if counterparties become insolvent, it only shares the average of these costs.

  14. 14.

    The credit channel effect is assumed to be the same for all countries. The results of Jiménez et al. (2012), however, suggest that the credit channel of monetary policy may be particularly strong in environments with weak bank balance sheets such as in some of the euro area member countries in crisis.

  15. 15.

    Refinancing credits in the European System of Central Banks usually make up the largest part of the monetary base, e.g. in the end of 2012 central bank credit accounted for 69.2 % of the total monetary base.

  16. 16.

    This assumption can also be motivated by the law of one price. Money printed in one country can be used to purchase goods in any other member country, thus in an arbitrage-free world the price level will be the same for all countries. It is, however, sufficient to assume that domestic inflation costs do not rise proportionally to domestic credit for the inflation bias to occur.

  17. 17.

    See also Tornell and Westermann (2012a) for an overview of some of these patterns.

  18. 18.

    See Sachs, Tornell and Velasco (1996) for the analysis of a similar pattern in Mexico 1994/5.

  19. 19.

    Some expansion of central banks credit was also observed in countries without a negative output shock, however, in these countries the change in central banks credit was offset by an increase of commercial banks reserves with the central bank of a similar magnitude (See Fig. 5 of Appendix B).

  20. 20.

    See also Alessandrini et al. (2013) who discuss how TARGET2-balances have contributed to macroeconomic imbalances in the Euro Area. The authors argue that limits on TARGET2-balances would be hard to implement, as they would increase the risk of a speculative attack within the Euro Area.

  21. 21.

    Also in other countries inflation is still moderate at this point. When we discuss the inflation bias and the “costs in terms of inflation”, we also mean the risks of future inflation that NCBs are willing to accept. The banks present holdings of excess deposits have prevented a larger inflation for the euro area as a whole so far, but certainly bear the risk that this inflation will come at a later stage, once excess deposits are withdrawn from the central banks.

  22. 22.

    A condition called a “key feature of a unified currency” by Friedmann (1992).


  1. Aizenman J (1992) Competitive externalities and the optimal seigniorage. J Money, Credit, Bank 24(1):61–71CrossRefGoogle Scholar
  2. Alesina A, Drazen A (1991) Why are stabilizations delayed? Am Econ Rev 82:1170–1188Google Scholar
  3. Alesina A, Perotti R (1999) Budget deficits and budget institutions. In: Fiscal Institutions and Fiscal Performance, Poterba JM, von Hagen J (eds), University of Chicago PressGoogle Scholar
  4. Alessandrini P, Fratianni M, Hallett AH, Presbitero AF (2013) External imbalances and fiscal fragility in the euro area. Open Economies ReviewGoogle Scholar
  5. Barro RJ, Gordon DB (1983) Rules, discretion and reputation in a model of monetary policy. J Monet Econ 12:101–121CrossRefGoogle Scholar
  6. Basley T, Coate S (2003) Centralized vs. decentralized provision of local public goods: A political economy analysis. J Public Econ 87:2611–2637CrossRefGoogle Scholar
  7. Buiter W (2012) Is the Eurozone at risk of turning into the rouble zone? Citi Economics, Global Economics View, LondonGoogle Scholar
  8. Casella A (1992) Participation in a currency union. Am Econ Rev 82(4):847–863Google Scholar
  9. Chari VV, Kehoe PJ (2008) Time inconsistency and free-riding in a monetary union. J Money, Credit, Bank 40:1329–1356CrossRefGoogle Scholar
  10. Davig T, Leeper E, Walker T (2012) Inflation and the fiscal limit. Eur Econ Rev 55(1):31–47CrossRefGoogle Scholar
  11. Eichengreen B, Grilli V, Fischer S (1993) A payments mechanism for the former soviet union: Is the EPU a relevant precedent? Econ Policy 8(17):309–353CrossRefGoogle Scholar
  12. Feldstein M, Vaillant M (1998) Can state taxes redistribute income? J Public Econ 68(3):369–396CrossRefGoogle Scholar
  13. Friedmann M (1992) Money mischief: Episodes in monetary history. Harcourt Brace Jovanovich, New YorkGoogle Scholar
  14. Galí J, Perotti R (2003) Fiscal policy and monetary integration in Europe. Econ Policy 18(37):533–572CrossRefGoogle Scholar
  15. Garber PM (1999) The target mechanism: Will it propagate or stifle a stage III crisis? Carn-Roch Conf Ser Public Policy 51(1):1999Google Scholar
  16. Hallerberg M, von Hagen J (1999) Electoral institutions, cabinet negotiations, and budget deficits within the european union. In: Fiscal institutions and fiscal performance, Poterba J, von Hagen J (eds.) University of Chicago Press, 209–232.Google Scholar
  17. Hausmann R, Purfield CM (2004) The challenge of fiscal adjustment in a democracy: The Case of India. IMF Working Paper 04/168Google Scholar
  18. Jiménez G, Ongena S, Peydró J-L, Saurina J (2012) Credit supply and monetary policy: Identifying the bank balance-sheet channel with loan applications. Am Econ Rev 102(5):2301–2326CrossRefGoogle Scholar
  19. Kashyap AK, Stein JC, Wilcox DW (1993) Monetary policy and credit conditions: Evidence from the composition of external finance. Am Econ Rev 83:78–98Google Scholar
  20. Kletzer KM, Singh N (1997) The political economy of Indian fiscal federalism. In: Fiscal Policies in India, Mundle S (ed.), Oxford University Press.Google Scholar
  21. Knight B (2003) Common tax pool problems in federal systems. In: Democratic Constitutional design and policy, Congelton RD, Swedenborg B (eds.) MIT Press, 229–247Google Scholar
  22. Krogstrup S, Wyplosz C (2010) A common pool theory of supranational deficit ceilings. Eur Econ Rev 54(2):269–278CrossRefGoogle Scholar
  23. Lipton D, Sachs JD (1992) Prospects for Russia’s economic reforms. Brook Pap Econ Act 1992(2):213–283CrossRefGoogle Scholar
  24. Neumann MJM (2012) The refinancing of banks drives target debt. CESifo Forum, 29–32Google Scholar
  25. Sachs J, Tornell A, Velasco A (1996) The collapse of the Mexican Peso: What have we learned? Econ Policy 11(22):13–63CrossRefGoogle Scholar
  26. Sinn H-W (2003) Risktaking, limited liability, and the competition of bank regulators. FinanzArchiv/Public Finance Analysis 59(3):305–329CrossRefGoogle Scholar
  27. Sinn H-W, Wollmershäuser T (2012) Target loans, current account balances and capital flows. The ECB’s rescue facility. Int Tax Public Financ 19(4):468–508CrossRefGoogle Scholar
  28. Steinkamp S, Westermann F (2012) On creditor seniority and sovereign bond prices in europe. CESifo Woking Paper No. 3944.Google Scholar
  29. Tornell A (2012) The dynamic tragedy-of-the-commons in the Eurozone, the ECB and Target2 imbalances. UCLA mimeoGoogle Scholar
  30. Tornell A, Westermann F (2005) Boom-bust cycles and financial liberalization, MIT Press.Google Scholar
  31. Tornell A, Westermann F (2012a) The tragedy-of-the-commons at the European Central bank and the next rescue operation. VOXeu.orgGoogle Scholar
  32. Tornell A, Westermann F (2012b) Europe needs a federal reserve, The New York Times.Google Scholar
  33. Von Hagen J, Eichengreen B (1996) Federalism, Fiscal Restraints and European Monetary Union. Am Econ Rev 82(2):134–138Google Scholar
  34. Von Hagen J, Süppel R (1994) Central bank constitutions for federal monetary unions. Eur Econ Rev 38(3–4):774–782CrossRefGoogle Scholar
  35. Walsh CE (2010) Monetary theory and policy. 3. ed., MIT Press.Google Scholar
  36. Weingast B, Shepsle K, Johnson C (1981) The political economy of benefits and costs: A neoclassical approach to distributive politics. J Polit Econ 96:132–163CrossRefGoogle Scholar
  37. Williamson SD (2013) The balance sheet and the future of fed policy. Open Economies Review.Google Scholar

Copyright information

© Springer Science+Business Media New York 2014

Authors and Affiliations

  • Valeriya Dinger
    • 1
  • Sven Steinkamp
    • 1
  • Frank Westermann
    • 1
  1. 1.Institute of Empirical Economic ResearchOsnabrueck UniversityOsnabrückGermany

Personalised recommendations