How well has the ECB’s policy suited the needs of the EMU and its members during the financial crisis, the European sovereign debt crisis and their aftermath according to the Taylor rule?

How well has the ECB’s policy suited the needs of the

EMU and its members during the financial crisis, the

European sovereign debt crisis and their aftermath

according to the Taylor rule?

Abstract

This thesis analyses how well the interest rate policy of the ECB has suited the needs of the individual members of the EMU and the area as a whole during the financial crisis, the European sovereign debt crisis and their aftermath. Optimal interest rates are calculated with the Taylor rule and compared to the EONIA. The finding is that the interest rate policy has suited the EMU as a whole, but has not been a good fit for countries individually.

Bachelor thesis Economics & Finance Author: Mariëlle Dreuning

Student number: 10559035

Supervisor: C.G.F. van der Kwaak Date: 29-06-2016

2 Statement of Originality

This document is written by Student Mariëlle Dreuning who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those

mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents

3 Contents

I. Introduction 4

II. Literature review 5

III. Methodology 8 Data 8 Method 9 IV. Results 11 EMU 11 Regression 13 Individual 13 V. Discussion 17 VI. Conclusion 19 References 20

4 I. Introduction

The establishment of the Economic and Monetary Union in 1999 gave the European Central Bank the task to decide on a single monetary policy for a variety of countries. It entails that only one short-term interest rate applies to all members of the union. As an independent central bank, the ECB strives for this rate to be suitable for all

members. This has been a great challenge from the beginning since every country has its own characteristics and economic structure. The capacity to cope with negative macroeconomic shocks has been viewed as one of the largest challenges since the beginning of the eurozone. By removing the possibility of national currency devaluation, a traditional way to adjust between economies was placed out of order.

The emergence of the financial crisis in 2008 has provided the possibility to put this to the test. The collapse of Lehman Brothers is often appointed as the moment after which the crisis entered a more acute phase. In reaction to this, the ECB cut interest rates aggressively, eventually reaching a historical low level. Many unconventional measures have been taken, of which a massive expansion of the bank’s balance sheet and the attempt to influence other interest rates than the usual short-term official rates are the more important ones (Joyce, Miles, Scott & Vayanos, 2012). That the measures were unconventional contributes to the importance of evaluating the monetary policy of the ECB during the financial crisis.

The bank did not get an opportunity to catch its breath, as it quickly became clear that Europe was experiencing a sovereign debt crisis. In late 2009, several countries reported larger deficit to GDP ratios than expected. This translated in increasing yields on sovereign bonds for those countries, making it more difficult to fund the deficits (Lane, 2012). The problem grew until several bailout programs were set up in 2010 and 2011. These programs provided funding for countries in need on the condition that fiscal austerity packages and structural reforms were implemented. The necessary measures slowed down the economy, a development to which the ECB reacted with lower interest rates. All the turbulence created by the crises put the ECB in the spotlights, as economists questioned if the central bank of such a

diversified monetary union would be able to conduct a policy that fits all members. This thesis intends to provide an answer to this, as it researches how well the interest rate policy of the European Central Bank has suited the needs of the

5

during the financial crisis, the European sovereign debt crisis and their aftermath. This evaluation is based on the Taylor rule, the monetary policy rule proposed by John B. Taylor in 1993. The simple equation has been put to the test many times and still proves to be a sound tool to assess monetary policy. The rest of this thesis has the following structure: section 2 provides relevant literature about the Taylor rule and monetary policy in the EMU, section 3 describes how the Taylor rule is used to

evaluate the ECB, section 4 presents the results of the research, section 5 discusses these results and section 6 gives concluding remarks.

II. Literature review

This literature review is composed of three parts. First the Taylor rule itself is reviewed, in particular its applicability to the EMU. Second, attention is given to the ability of a single policy to suit the whole EMU area, zooming in on the divergences within the union. Finally, an inquiry is made about the environment for monetary policy during the financial crisis and the European sovereign debt crisis.

Taylor (1993) has proposed a simple monetary policy rule where the optimal nominal interest rate is contingent on real GDP and inflation. According to this rule, the interest rate should rise when real GDP rises above trend GDP or when the inflation rate rises above its target. Next to these output and inflation gaps, which are assigned a weight of 0,5, the equation contains the current inflation and the

equilibrium real interest rate. This means that the nominal interest rate reacts to inflation two times. The Taylor rule has experienced great popularity since its introduction. In part this popularity can be attributed to the simplicity of the rule and the explicit linking between a current policy rate to current economic conditions. Even more important, Taylor (1993) found that his rule fits the path of the federal funds rate from 1987 till 1992 remarkably well.

Later studies on the US and on European countries produce the same

outcome. Gerlach and Schnabel (2000) have contributed to the research by looking at the behavior of interest rates in the EMU countries in the period prior to the

establishment of the union. Their results show that the Taylor rule tracks the interest rate setting of the central banks closely, which leads them to the suggestion that the use of the Taylor rule by the ECB would lead to continuity of monetary policy in the newborn euro area. Based on their data the hypothesis that the weights on the

6

inflation and output gap should both be 0,5 cannot be rejected. Also focusing on the Euro area, Peersman and Smets (1999) study the usefulness of the Taylor as a benchmark for monetary policy. They use aggregate data of five EMU countries to estimate the monetary policy transmission process in these countries and conclude that a simple Taylor rule would perform quite well in stabilizing output and inflation. The performance of the rule seems to be robust to small changes in the parameters of their model, meaning that the uncertainty surrounding estimations of output gaps does not seem to impede with the reliability of the rule. This is in consensus with later research on the subject by Smets (2002). The problem brought to the attention in that paper however, is that the uncertainty surrounding output gaps does lead central banks to place a lower weight on the variable than what would be efficiently optimal. This is considered a serious drawback of the rule.

All in all it can be concluded that the Taylor rule is applicable to the EMU and can be used to evaluate the monetary policy of the ECB. This is done by Moons and Van Poeck (2008) as they question whether the interest rate setting of the ECB is in accordance with the needs of the individual countries and the EMU area as a whole, using the Taylor rule as a benchmark. They calculate the Taylor-based desired rates for all members and the whole union and compare these to the EONIA, which closely tracks the ECB policy rate. Their findings indicate that the interest rate policy of the ECB did not fit all countries equally well. Also, they find no tendency of improvement over the period 1999 to 2003 for these countries or for the area as a whole.

This raises doubt about the possibility of a single monetary policy in the eurozone. With the area being a collection of different countries, the same policy could have different reactions. The paper of Huchet (2003) shines a light on this subject. She studies the effects of the common monetary policy of the ECB on the activity of European countries. Huchet (2003) finds that changes in the interest rate have asymmetric effects from one country to another. Some countries react stronger to a rise in the interest rate, whereas others are more sensitive to a drop in the

interest rate. This is problematic because a common monetary change could become an asymmetric shock, which would lead to an uneven distribution of output across the union. The asymmetries are attributed to structural national differences. Many

research on the latter has been done, for one by Crespo-Cuaresma and Fernández-Amador (2013), who focus on business cycle convergence in the EMU. Their study shows a period of significant convergence in the nineties. This finished at the start of

7

the EMU, when a divergence period began. The inclusion of new members since then did not have a significant impact on the level of divergence, so the size of the EMU has no influence. The authors emphasize the role of cyclical homogeneity for the conduct of a common monetary policy in a currency area. Homogeneity enables the single policy to be optimal for al member countries.

In the same line of thinking, Arestis and Sawyer (2011) comment that the convergence criteria of joining the EMU were not sufficient. The criteria do not

mention business cycle convergence or unemployment levels. A high unemployment country is constrained in leveling with low unemployment countries once in a fixed exchange rate regime, because there is no lower interest rate available. Therefore, a lack of concern for unemployment levels impedes convergence. The authors also criticize that while there were reasonable requirements set for the inflation rate, no attention was paid to inflationary patterns of the prospective members. Whether there was a tendency for one country to inflate faster than another for example was

ignored. Due to this, acceptable inflation rates when the EMU was established did not deter divergence.

Estrada, Galí and López-Galido (2013) also study patterns of convergence and divergence in the Euro Area, with a focus on the recent financial crisis. They see a reduction in the dispersion of the unemployment rate between the countries in the period preceding the crisis. From the onset of the crisis however they see a

substantial divergence in these numbers. While dispersion of unemployment rates is a common phenomenon during recessions, the increase has been much larger across Euro members than across non-Euro members. This would make conducting a single policy during the crisis that fits all countries even more difficult. Therefore, the ECB could be induced to deviate from the Taylor rule.

A possible reason for this is given by Gerlach and Lewis (2014), who study if the ECB’s interest rate setting during the financial crisis was different than before. They conclude that while the rapid worsening of economic conditions in the fall of 2008 suggests a rapid relaxation of monetary policy, the ECB in fact cut interest rates even more rapidly. The deviation of what could be expected from pre-crisis policy is attributed to the presence of the zero lower bound on interest rates. Theoretical literature on this subject implies that the anticipation of hitting the zero lower bound in the future might induce a central bank to cut interest rates more aggressively today. The rationale for this is that when the optimal interest rate falls below zero, the actual

8

policy rate at that point will be too high and so will long term interest rates. To prevent this from happening, the policy rates should be cut rapidly, to lower long term rates and stimulate the economy. Jung, Teranishi and Watanabe (2005) find that when a country has eventually reached the zero lower bound, the best option is to commit to continuing the low interest rate, even if it would not be optimal under normal

circumstances. This lowers longer term interest rates, which according to them is “as if the central bank “borrows future monetary easing in the periods when current monetary easing is already exhausted.” (Jung, Teranishi & Watanabe, 2005, p.825). It needs to be kept in mind that due to the zero lower bound the ECB may had to act in a way that normally would not be optimal for the EMU, according to monetary policy rules. To summarize, the environment in which the ECB had to operate was different during the financial crisis than in normal circumstances. The dispersion in unemployment rates across countries became larger and the bank had to account for the presence of the zero lower bound.

The effects across the euro area were not only different during the financial crisis, but also afterwards. International financial flows had dried up after 2008, something that disproportionately affected countries who are most dependent on external funding. These countries were confronted with a decline in economic activity, leading to a large rise in the deficit to GDP ratios, according to Lane (2012). It

created greater imbalances across the EMU, making it more difficult to find a single interest rate that is suited for all countries in the union. Where one country could facilitate a higher interest rate, this would be disastrous for countries with high deficits.

III. Methodology

Data

For evaluating the interest rate policy of the ECB during the financial crisis, the European sovereign debt crises and their aftermath, quarterly data is needed on several variables for the period 2008 to 2015 for all EMU members as of January 2008. This means that the countries involved in this thesis are Austria, Belgium, Cyprus, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Malta, the Netherlands, Portugal, Slovenia and Spain. Output gaps and inflation rates are obtained from the OECD, except for Cyprus and Malta. The data for these countries

9

is gathered from the IMF. Yearly output gaps are transformed to quarterly rolling averages, calculated at an annual rate. The inflation included is core inflation, as policies based on this exhibit better stabilization properties according to Bodenstein, Erceg and Guerrieri (2008). This entails the Harmonized Index of Consumer Prices excluding energy, food, alcohol and tobacco. Quarterly inflation at an annual rate is used. For the rate that the ECB actually sets the Euro Overnight Index Average (EONIA) is chosen. Data on the EONIA is accrued from Eurostat, as is data on quarterly real GDP. For the latter chain linked volumes are used.

Method

The method used in this thesis is based on the research by Moons and Van Poeck (2008). To calculate the optimal nominal interest rate, the following rule proposed by Taylor (1993) is used:

𝑟_𝑖𝑡∗ = 𝜌 + 𝜋𝑖𝑡+ 𝜃(𝜋𝑖𝑡− 𝜋∗) + (1 − 𝜃)𝑦𝑖𝑡

Where 𝑟_𝑖𝑡∗ is the nominal desired short term interest rate, 𝜌 is the equilibrium real interest rate, 𝜋_𝑖𝑡 is the actual inflation, 𝜋∗ is the target inflation and 𝑦_𝑖𝑡 is the output gap, defined for country 𝑖 and quarter 𝑡. 𝜃 and 1 − 𝜃 are the weights given to the deviation from the inflation target and to the output gap. An inflation target of 2% is used, as the ECB states as its target (2016). The equilibrium real interest rate is set at 2%, following the work of Taylor (1993). Using those rates and assigning weights of 0,5 to 𝜃 and 1 − 𝜃, as is validated by Gerlach and Schnabel (2000), the equation can be reorganized as:

𝑟_𝑖𝑡∗ = 1 + 1,5𝜋_𝑖𝑡 + 0,5𝑦_𝑖𝑡

To derive the Taylor-based desired rate for the EMU area as a whole a weighted average of the Taylor-based desired rates of the individual member countries is calculated, giving each country a weight equal to the share of its output in the output of the EMU area:

𝑟_{𝐸𝑀𝑈,𝑡}∗ = ∑ 𝑎_𝑖𝑡𝑟_𝑖𝑡∗

𝑛 𝑖=1

Where 𝑎_𝑖𝑡 is the share of country 𝑖’s output in total EMU area output in quarter 𝑡 and 𝑛 is the number of EMU members as of January 2008. This weighted average is consistent with a non-discriminatory policy, one that would be expected from the

10

independence of the ECB. The following regression equation is used to compare the Taylor-based desired rate to the actual rate set by the ECB:

𝐸𝑂𝑁𝐼𝐴_𝑡 = 𝛽₀+ 𝛽₁𝑟_{𝐸𝑀𝑈,𝑡}∗ + 𝜀_𝑡

T-tests are used to examine if 𝛽1 is significantly different from 1 and if 𝛽0 is

significantly different from 0.

To take a look at each country individually, the root mean squared interest rate gap is calculated:

𝑅𝑀𝑆𝐼𝐺𝑖 = √

∑𝑇 (𝑟_𝑖𝑡∗ − 𝐸𝑂𝑁𝐼𝐴_𝑡)2 𝑡=1

𝑇

Where T is the number of quarters. Positive and negative deviations cannot cancel each other out in the summation since every deviation is squared first. This makes the RMSIG a measure of the total deviation of the Taylor-based desired rate from the EONIA for a given country. It therefore means that a higher RMSIG indicates that the interest rate setting of the ECB is less suited for that country. The RMSIG is

compared with the mean interest rate gap, defined as: 𝑀𝐼𝐺_𝑖 = ∑ (𝑟𝑖𝑡 ∗ _{− 𝐸𝑂𝑁𝐼𝐴} 𝑡) 𝑇 𝑡=1 𝑇

When the MIG is positive, the EONIA has been too low on average for the country for which it is calculated. Vice versa, a negative MIG tells that the EONIA has been too high on average. The RMSIG and MIG are compared to see if there is a relation between the two, for example that countries with a high RMSIG generally have a high MIG. In this case, it can be stated the EONIA was generally too low for countries for which the interest rate setting of the ECB was least suited. For the comparison the Spearman rank correlation coefficient (𝜌_𝑠) is calculated by ranking the RMSIG and MIG and measuring the correlation between the rank numbers. A t-test is used to see if 𝜌𝑠 differs significantly from 0, which would indicate that there is a relation between

RMSIG and MIG. If the coefficient is significantly larger than 0, the EONIA was

generally too low for countries for which the interest rate setting of the ECB was least suited. If the coefficient is significantly smaller than 0, the EONIA was generally too high for countries for which the interest rate setting of the ECB was least suited.

Last, the development of root mean squared interest rate gaps in the EMU is taken into consideration. This is done by calculating the RMSIG of the EMU for every quarter:

11

𝑅𝑀𝑆𝐼𝐺𝐸𝑀𝑈,𝑡 = √

∑𝑛𝑖=1 (𝑟_𝑖𝑡∗ − 𝐸𝑂𝑁𝐼𝐴𝑡)2

𝑛

When this RMSIG rises, the interest rate setting of the ECB becomes less suited for the EMU as a whole.

IV. Results

EMU

Graph 1 shows the calculated Taylor-based desired rate for the Economic and Monetary Union and the EONIA to compare.

Graph 1: Taylor rule vs. EONIA

The onset of the financial crisis is clearly visible as the Taylor-based desired rate drops fast from the third quarter of 2008. Such a drop indicates that there is an economic downturn that needs an expansionary monetary policy, which translates to lower interest rates. This economic downturn is pointed out by the fast decrease in output gaps, quickly becoming negative, as depicted in graph 2 for the four biggest economies of the eurozone. A negative output gap means that actual output is less than potential output, so the economy is underperforming given its resources. At first glance it seems that the EONIA reacts as it should according to the Taylor rule. The policy rate roughly follows the same trend as the Taylor-based desired rate of the

-1 0 1 2 3 4 5 6 EONIA EMU

12

EMU. It is notable however that the latter is systematically higher than the EONIA. An explanation for this is that the EONIA closely tracks the deposit rate, while the policy rate on which the ECB focusses is actually the main refinancing rate. This rate moves together with the deposit rate, but on a higher level.

Graph 2: Output gaps

From the second quarter of 2010 the dispersion between the EONIA and the Taylor-based desired rate becomes larger. It can be assumed that this would also be the case if the chosen policy rate was the main refinancing rate, since this rate moves together with the EONIA. This means that the monetary policy of the ECB became less optimal. An explanation for this could be that the policy rate was close to the zero lower bound at the time. As noted earlier, literature on the zero lower bound advices to commit to low interest rates in the presence of this bound. From the end of 2011 the economic conditions again ask for a lower policy rate according to the Taylor rule. The financial crisis at this point has transformed into the European sovereign debt crisis, requesting relaxation of monetary policy. Unfortunately, the EONIA has little leeway here, as it is already almost zero. Interesting is that when the ECB eventually sets a negative policy rate, in the fourth quarter of 2014, the direction of the Taylor-based desired rate is actually upwards. This starts another trend of divergence. -10 -8 -6 -4 -2 0 2 4 6 20 08Q1 20 08Q3 20 09Q1 20 09Q3 20 10Q1 20 10Q3 20 11Q1 20 11Q3 20 12Q1 20 12Q3 20 13Q1 20 13Q3 20 14Q1 20 14Q3 20 15Q1 20 15Q3 France Germany Italy Spain

13

Regression

The regression of the EONIA on the Taylor-based desired rate for the EMU gives the following regression equation:

𝐸𝑂𝑁𝐼𝐴 = −0,8058836 + 0,7949667 𝑟_𝐸𝑀𝑈∗

(−4,742) (−2,929)

The coefficient on 𝑟_𝐸𝑀𝑈∗ indicates that the Taylor rule tracks the actual policy rate rather well. However, the t-value between brackets states that the coefficient is significantly different from 1, indicating that the ECB could still improve its interest rate setting according to the Taylor rule. Attention also needs to be directed to the constant in the regression equation. The t-value of this negative number indicates that it differs very significantly from 0, stating that the EONIA is systematically too low for the EMU area. This is an expected result when it is taken into account that the EONIA is lower than the correct policy rate to compare the Taylor rule to. It is

interesting to compare the result to the findings of Moons and Van Poeck (2008), who run the same regression for the period 1999 till 2003. They find a constant of 1,11 and a coefficient on 𝑟_𝐸𝑀𝑈∗ of 0,68. The latter suggests that the Taylor rule tracks the movements in the policy rate of the ECB better during the financial crisis and its aftermath than during the first years of the monetary union, because the coefficient here is closer to 1.

Individual

Graph 3 shows the calculated root mean squared interest rate gaps for the EMU members. The three countries with the lowest RMSIG and the three countries with the highest RMSIG are chosen to take a closer look at. The graph reveals that the countries with the lowest gaps are France, Italy and Portugal. This means that the interest rate setting of the ECB has suited these countries best.

14 Graph 3: Root mean squared interest rate gaps

Graph 4: Taylor-based desired rates vs. EONIA

As can be seen in graph 4, the Taylor based desired rate for Portugal drops fast due to the financial crisis and drops fast again with the start of 2011. The latter can be attributed to the measures that needed to be taken in the first half of this year to improve the financial situation of the government by decreasing its deficit. The measures included spending cuts and tax increases, putting pressure on the economy (Wise, 2011). This asks for lower interest rates. With Portugal and Italy being the most indebted countries after Greece, it is not surprising Italy follows

0 1 2 3 4 5 6 7 8 RMSIG -4 -3 -2 -1 0 1 2 3 4 5 6 7 EONIA France Italy Portugal

15

Portugal not much later in the attempt to reduce its deficits with the implementation of an austerity package. For these countries the historically low EONIA has been

favorable. Besides that, graph 4 makes clear that every country has a unique pattern of optimal interest rates.

Graph 5: Mean interest rate gaps

The root mean squared interest rate gaps are the highest for Greece, Ireland and Luxembourg. A look at the mean interest rate gaps, as presented in graph 5, shows that for Greece and Ireland the EONIA has been too high, where it has been too low for Luxembourg. For these countries the monetary policy of the ECB is least suited, but graph 6 points out that the ways to improve this are often the complete opposite from each other. Where for Luxembourg it would mean that the EONIA should be higher in every quarter, this would be disastrous for Greece from 2011 on. Where Greece suffers from huge negative output gaps and low inflation or even deflation, Luxembourg has relatively high inflation and low output gaps. From the second quarter of 2009 Ireland has large negative output gaps and very volatile inflation. The first would indicate low interest rates and the second is rather inconclusive, so the country would need a very tailored policy for it to be optimal.

-3,0000 -2,0000 -1,0000 0,0000 1,0000 2,0000 3,0000 4,0000

MIG

16 Graph 6: Taylor-based desired rates vs. EONIA

As graph 5 makes clear, the EONIA has on average been too low for most countries, as twelve out of fifteen mean interest rate gaps are positive. To see whether there is a connection between the absolute size of this deviation and the EONIA being too low, the Spearman rank correlation coefficient is calculated from the data on RMSIG and MIG. This coefficient, Spearman’s rho, is 0,0607. A t-test shows that this

coefficient is not significantly different from zero. That means that there is no positive relation between the RMSIG and MIG, so there is no relation between the absolute size of the interest rate gaps and the gaps being high on average. In other words, it cannot be stated that for the countries for which the policy rate is least suited, the rate is too low. The extraordinary debt situation of Greece, which asks for extremely negative interest rates, may color this picture. For that reason, Spearman’s rho is also calculated for the EMU without Greece. This gives a coefficient of 0,3055, which is notably higher, but still not significantly different from zero.

-15,0000 -10,0000 -5,0000 0,0000 5,0000 10,0000 15,0000 EONIA Greece Ireland Luxembourg

Table 1: Spearman’s rank correlation

All countries Without Greece

Spearman’s rho 0,0607 0,3055

17

For the period 1999 until 2003, Moons and Van Poeck (2008) found a Spearman’s rho of 0,952. This number indicates that when absolute interest rate gaps are high, they are often very positive. It therefore suggests that during the first years of the monetary union the interest rate was generally too low for countries for which this rate was least suited.

Graph 7: Root mean squared interest rate gap

A look at the development of the root mean squared interest rate gap of the EMU over time in graph 7 shows that countries experienced larger gaps during the financial crisis. This corresponds to the findings of Estrada, Galí and López-Galido (2013) described earlier. As the dispersion in unemployment rates became higher, it became more difficult for the ECB to find a monetary policy that is suited for all member countries. When the financial crisis rolled over in the sovereign debt crisis, the aggregate interest rate gap again starts to rise. This is an expected result, as this crisis created great imbalances within the area.

V. Discussion

Methodological issues

In this thesis the EONIA is chosen as the rate that the ECB sets, following the work of Moons and Van Poeck (2008). The EONIA closely tracks the deposit rate, while the key interest rate of the ECB is actually the main refinancing rate. During the period on which Moons and Van Poeck conduct their research however, the spread between

0 0,5 1 1,5 2 2,5 3 3,5 4 4,5

RSMIG EMU

18

these rates was low, making the EONIA a close substitute for the main refinancing rate. This is not the case after October 2008, when the introduction of the fixed rate full allotment caused the spread to increase. Therefore, the EONIA is not the

appropriate rate to compare the Taylor rule to. It does not completely invalidates the results in this thesis as the EONIA and main refinancing rate do follow the same trend. Yet, to improve the analysis of the monetary policy of the ECB, further research should be done using the main refinancing rate.

Results

The regression results indicate that the interest rate setting of the ECB has suited the union as a whole quite good. The coefficient of 0,7949667 on the Taylor-based desired rate for the EMU is close to 1, meaning that the EONIA moves in accordance with the Taylor rule. This coefficient is significantly different than 1 however,

indicating that the interest rate setting could still be improved. -0,8058836, the

constant in the regression equation, differs very significantly from zero. It means that EONIA is systematically set too low compared to the Taylor rule. No conclusions about the performance of the ECB can be made from this finding however, because the EONIA is not the appropriate rate to represent the interest rate policy of the ECB. Especially when it is kept in mind that the correct policy rate would be higher, it can be stated that the ECB has performed rather well during the crises and their

aftermath for the EMU as whole.

The differing root mean squared interest rate gaps across the union make clear that the interest rate policy did not fit all countries equally well. Graphing the individual Taylor-based desired rates shows that each country has a different pattern of optimal interest rates, reflecting differing needs. This corresponds with the notion that there are great economic differences within the EMU. For some countries the EONIA has been too low on average whereas for others it has been too high, as pointed out by the mean interest rate gaps, making it impossible to improve the interest rate setting for all countries. The deviation from what is optimal becomes larger during the financial crisis. This is expected, as the dispersion between

countries grew, making a single interest rate less suited. A comparable development induced the interest rate gaps to start rising again when the European sovereign debt crisis unfolded. The analysis leads to the finding that the differences within the EMU

19

during the period that is researched are too persistent to conduct a single interest rate policy that suits the needs of countries on an individual level.

VI. Conclusion

This thesis has researched how well the interest rate policy of the European Central Bank has suited the needs of the individual members of the Economic and Monetary Union and the area as a whole during the financial crisis, the European sovereign debt crisis and their aftermath using the Taylor rule. The interest rate setting of the ECB was a good fit from an area wide perspective. This does not hold for countries individually, as their needs differ too much for a single policy to suit them all.

20 References

Arestis, P. & Sawyer, M. (2011). The Design Faults of the Economic and Monetary Union. Journal of Contemporary European Studies, 19:1. 21-32.

DOI:10.1080/14782804.2011.554191

Bodenstein, M., Erceg, C., Guerrieri, L. (2008). Optimal monetary policy with distinct core and headline inflation rates. Journal of Monetary Economics, 55. S18-S33. doi:10.1016/j.jmoneco.2008.07.010

Crespo-Cuaresma, J. & Fernández-Amador, O. (2013). Business cycle convergence in the EMU: a first look at the second moment. Journal of Macroeconomics, 37. 265-284. doi:10.1016/j.jmacro.2013.02.001

ECB (2016). https://www.ecb.europa.eu/mopo/html/index.en.html

Estrada, A., Galí, J. & López-Salido, D. (2013). Patterns of Convergence and Divergence in the Euro Area. IMF Economic Review, 61:4. 601-630. doi:10.1057/imfer.2013.22

Gerlach, S. & Lewis, J. (2014). Zero lower bound, ECB interest rate policy and the financial crisis. Empirical Economis, 46:3. 865-886. doi:10.1007/s00181-013-0713-6

Gerlach, S. & Schnabel, G. (2000). The Taylor rule and interest rates in the EMU area. Economic Letters, 67:2. 165-171. doi:10.1016/S0165-1765(99)00263-3

Huchet, M. (2003). Does single monetary policy have asymmetric real effects in EMU? Journal of Policy Modeling, 25:2. 151-178. doi:10.1016/S0161-8938(02)00204-1

Joyce, M., Miles, D., Scott, A. & Vayanos, D. (2012). Quantitative Easing and

Unconventional Monetary Policy – an Introduction. The Economi Journal, 112. F271-F288. DOI: 10.1111/j.1468-0297.2012.02551.x

21

Jung, T., Teranishi, Y. & Watanabe, T. (2005). Optimal Monetary Policy at the Zero- Interest-Rate Bound. Journal of Money, Credit and Banking, 37:5. 813-835.

Lane, P. (2006). The Real Effects of European Monetary Union. Journal of Economic Perspectives, 20:4. 47–66

Lane, P. (2012). The European Sovereign Debt Crisis. Journal of Economic Perspectives, 26:3. 49-69. DOI: 10.1257/jep.26.3.49

Moons, C. & van Poeck, A. (2008). Does one size fit all? A Taylor-rule based analysis of monetary policy for current and future EMU members. Applied Economics, 40:2. 193-199. DOI: 10.1080/00036840600749763

Peersman, G. & Smets, F. (1999). The Taylor Rule: A Useful Monetary Policy Benchmark for the Euro Area? Internatonal Finance, 2:1. 85-116. DOI: 10.1111/1468-2362.00020

Smets, F. (2002). Output gap uncertainty: Does it matter for the Taylor rule? Empirical Economics, 27. 113-129.

Taylor, J. (1993). Discretion versus policy rules in practice. Carnegie-Rochester Conference Series on Public Policy, 39. 195-214. doi:10.1016/0167-2231(93)90009-L

Wise, P. (October 13, 2011). Portugal announces more austerity measures. Financial Times. Retrieved from www.ft.com