! ! ! ! ! ! ! !
!
Analysis( of( the( existence( of( political( pressure( of( the(
strong(member(states(in(the(Governing(Council(of(the(
European(Central(Bank(
Philip Botros 10358196
! !
Faculty: Economics and Business Specialization: Economics & Finance Supervisor: Egle Jakucionyte
ABSTRACT
This paper explores the presence of political pressure in the Governing Council of the European Central Bank. Most literature agrees that regional influences play a significant part in the decision-making process of central bank members. This regional bias could make central bank members susceptible to political pressure from strong member states. The ‘country, one-vote’ system employed in the Governing Council until recently also made exerting political pressure rewarding. In this paper, political pressure is defined as the overweighting of the economic preferences of France and Germany. The Taylor rule is used to measure the weighting of respective economic preferences for the group of France and Germany and a group containing the remaining Eurozone countries. Over the whole period, weak signs of political pressure from the large member states were observable, especially during the period before the crisis. Where the larger member states were expected to get more accommodated during the crisis, as the result of increased political pressure following larger economic differences, the opposite was found. The interest rates during the crisis period seemed to favor the weaker member states.
This document is written by Student Philip Botros 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.
I.! INTRODUCTION
This quote by former ECB president Wim Duisenberg (2002) captures the essence of the goal of the decision-making process of the Governing Council: “The members of the Governing Council consider the interests of the euro area as a whole; they do not represent their respective countries”. While the members of the Governing Council should put the interests of the euro area above the interests of their own country, questions are raised about the true objectivity of members in any monetary policy committee dealing with regional representatives. This question is especially valid in the case of a central bank representing a monetary union as the European Central Bank (ECB), where the Governors of the National Central Banks represent their home countries instead of their region and are also appointed by their respective national governments.
Hayo & Meon (2013) recently found that, instead of a euro-wide perspective, a nationalistic view is at the core of the ECB decisions regarding the interest rate policy. While it is uncertain if this is detrimental for the welfare of the Eurozone countries, it is directly contradicting the intended Euro-wide perspective policy as mandated in the ECB statutes.
The Governing Council consists of two parts; the executive board, who are appointed by the European Council and the governors of the National Central banks, appointed by their respective governments. Up to 2015 the Governing Council employed the ‘country, one-vote’ principle where every member had one vote. Decisions in the Governing Council are made according to a majority voting system where the president gets the decisive vote in case of a tie.
Since 1 January 2015, the Governors of the National Central Banks are divided into two tiers. This division is based on the strength of the economy, size of the financial sector and capital share in the ECB of the respective state.
The ‘one-country, one-vote’ principle employed by the Governing Council until recently should, in theory, have given a disproportionally large vote to smaller countries, as Malta receives the same voting rights as Germany (Berger & Mueller, 2007). But this feature also made the system susceptible to political pressure, as it increased the incentives and rewards for larger economies to exert this political pressure.
Especially Germany and France could have the incentives to push through lower interest rates because of their relatively low inflation. The historical inflation aversion of the Bundesbank also suggests a higher preference for Germany to have inflation movements at the core of the monetary policy relative to output movements (Kahn & Benolkin, 2007). Considering the fact that the two of them combined have the same GDP as the rest of the Euro countries together, they also have the political power to accomplish this. Early studies by Kool (2005) and Von Hagen & Bruckner (2002) found that these conditions have been translated in the overweighing of Germany and France’s economic preferences.
This paper aims to examine political pressure by analyzing the weights of the economic conditions of the two biggest countries of the Eurozone, France and Germany, on the interest rate setting behavior of the ECB using the Taylor rule. By comparing the fit of the data of the pre-crisis and post-crisis period for France and Germany, inferences can be made about the presence of higher political pressure during larger economic differences.
This paper has the following structure: the existing literature is presented and discussed in Section 2. In Section 3 the various forms of the Taylor rule and methodology of this paper are explained. The results and interpretation of these results are presented in Section 4 and the conclusion is presented in Section 5.
II.! LITERATURE REVIEW
The objectivity of central bank members has been the subject of extensive research in the past decades. Confirming these concerns, Gildea (1992) and Meade & Sheets (2005) have shown that regional aggregates have had a statistically significant effect on the voting pattern of the Federal Reserve policymakers. At the Bundesbank the same principle of regional bias applied (Berger & de Haan, 2002). It is certainly realistic to assume that this is also the case at the ECB where members represent their countries instead of their region, which would make them even more sensitive to pursuing their countries’ objectives rather than the goals of the organization.
To investigate the objectivity of the ECB, Meade & Sheets (2002) performed an experiment where they assumed that each member of the Governing Council casts his or her vote on the basis of the differential between their national inflation rate and the euro-area average. Based on a simple majority voting role they found that in nearly every case the majority of ECB members voted for a change in policy that was optimal when considering their respective national aggregates. This resulted in their conclusion that ECB monetary policy decisions are not inconsistent with the regional bias hypothesis.
In line with this criticism, Hayo & Meon (2013) more recently constructed and tested various scenarios with different assumptions regarding the voting behavior of members in the Governing Council. They concluded that individual members of the Governing Council follow national objectives over long-term ECB aggregates. This also included the members of the Executive board, who are appointed by the European Council and do not represent their national governments. Furthermore, they found that decisions about interest rates are better explained using bargaining scenarios where every country gets weighted by their relative GDP in relation with the whole Eurozone as a measure of bargaining power than explained by majority voting as mandated in the ECB statutes. The scenario where the whole Governing Council was assumed to have a euro-wide perspective at the core of their voting behavior was, after the scenario where the president of the ECB had full dominance over the interest rate policy, the worst performing scenario in explaining the interest rate decisions made by the ECB over the years prior to the crisis.
It is interesting to see if this regional bias also translates in susceptibility to political pressure. Dixit (2001) already explained this phenomenon where a country can attempt to influence outcomes indirectly by using political bargaining. There have been numerous instances of this in the history of the European Union and according to him it is not implausible to assume that this bargaining is also present at the monetary policy decision-making process of the ECB. A decision that appears to be the result of a consensus in the Governing Council may in reality be the mere ratification of previous political bargaining. This political pressure on the ECB would be more likely during a period where the difference between the actual monetary policy and the monetary policy preferred by individual member countries is larger (Sturm & Wollmershäuser, 2008).
While it is very complicated to verify the concept of ‘political pressure’ empirically, studies have been performed trying to explain the interest rate behavior of the ECB while looking at national aggregates. Kool (2005) found evidence that prior to 2001, the actual interest rate of the ECB approximately matches the rate Germany, France and Austria
preferred. Since 2001 the ECB interest rates have been too low from the perspective of almost all euro area countries except for Germany. He rejects a claim that small peripheral countries have a more than proportional say in ECB policymaking through the voting schedule and claims that Germany is “in the driving seat altogether”.
Obtaining the same results but using a different explanation, Von Hagen & Bruckner (2002) found that a Taylor-Rule constructed for Germany and France only did more to explain the actual interest rate than a Taylor rule consisting of all Euro countries where the median Taylor rule was chosen as an appropriate interest rate rule. Their explanation for these results was the possibility that other members of the Governing Council acknowledge the economic importance of these two countries rather than this result being the consequence of political bargaining.
There is no clear answer found in the literature about the consequences of deviating from a euro-wide perspective. According to Haskoy, De Grauwe & Dewachter (2002) a regime in which all members of the Governing Council take a nationalistic view is detrimental for welfare in the Eurozone in comparison to a euro-wide perspective. A regime as the European Monetary System, the predecessor of the ECB, where Germany sets monetary policy, also compares unfavorably compared to a euro-wide perspective. Arnold (2005) however states that an optimal monetary policy in the Eurozone may imply that the ECB overweighs output and inflation developments in the larger member states and discounts what happens in small Eurozone countries. For legal and political reasons discussed in the introduction, the ECB may not want to admit that it pays disproportionate attention to the economic circumstances in larger Eurozone countries.
I want to expand on the studies of Von Hagen & Bruckner (2002) and Kool (2005) by examining the weighting of the preferences of the strong member states.The sample size was however quite small in the abovementioned studies due to the relative recent founding of the ECB and the crisis period was not included. Aside from increasing the sample size and evaluating the interest rate setting during the crisis period I expand these studies by verifying if larger economic differences have led to a higher presence of political pressure in the ECB as predicted by the theory. During the Euro crisis the Eurozone was hit by asymmetric shocks affecting especially the countries in Southern Europe. This led to large economic differences between member states in the Eurozone and the crisis is in turn a good period to use for testing the hypotheses of increased political pressure following large economic differences.
III.! EMPIRICAL MODEL AND DATA
To answer the research question, I employ the Taylor rule as the key model. The Taylor rule has been the benchmark method for modelling the reaction function of central banks since it was introduced by John Taylor in 1993. For the ECB in particular, Belke & Polleit (2007) concluded that the interest rate setting behavior of the ECB can be characterized reasonably well by some form of Taylor rule. The original Taylor rule is described by the following equation:
!"= $∗+ '"+ ( '"− '∗ + *(,"− ,∗)
Where !" is the nominal policy rate at time t, $∗is the long-run or equilibrium real rate of
interest, '" is the inflation rate at time t and '∗is the inflation target of the central bank. The
current period inflation gap is thus specified by '"− '∗ and (,"− ,∗) is the current period
output gap, with ," as the actual output at time t and ,∗as the potential output. The coefficients
(,.* characterize the reaction of the central bank on the interest rates following output and inflation deviations. John Taylor assumed that the Federal Reserve set coefficients for both the inflation gap and the output gap to 0.5, implying that a rise of the inflation rate of 1% should be accompanied by a rise of the interest rate of 1.5% (the inflation level is also in the equation). This particular aspect of the Taylor rule, where central banks are prompted to raise the interest rate by more than 1% following a rise in inflation levels of 1%, is called the Taylor principle.
Table 1: Taylor rule estimates of the ECB
Study Time Period Constant Inflation Output
Belke & Polleit 1999-2005Q2 0.02 0.49 1.94
Sturm & Wollmershauser 1999-2006 0.50 1.63 1.56
Belke & Klose 1999-2007Q2 0.06 0.80 0.80
Gorter, Jakobs, de Haan 1997-2006 1.60 1.39 1.52
A lot of studies have tried to improve the explanation power of the Taylor rule by adding new terms. Peersman & Smets (1999) and Gerdesmeier & Roffia (2003) found that including interest rate smoothing produced significantly better results. Interest rate smoothing is the practice used by central banks where they gradually adjust the interest rate levels to the optimal
level to keep interest rate volatility at a lower level. Another modification that is frequently used is the forward-looking Taylor rule. Instead of using actual data, expected inflation and output gap data is used. Evidence from Clarida et al. (1998) suggests that central banks are forward-looking, they respond to anticipated inflation instead of lagged inflation.
For estimation purposes, the real interest rate ($∗).on the right hand side of the equation is
converted into a nominal interest rate using the Fisher equation (! = $ + '!): ( !"= $*+ ' "+ ( '"0'* + * ,"0,* ! !"= (!*0'*) + ' "+ ( '"0'* + *(,"0,*)! !"= !* + (1 + () ' "0'* + *(,"0,*)..! !"= $*+ '*+ (1 + () ' "0'* + *(,"0,*)..! Where 2 = $*+ '*
Now the real equilibrium interest rate is obtained by subtracting the inflation target of 2% from the constant. To control for possible heteroskedasticity and autocorrelation, as previously found by Gerdesmeier and Roffia (2003) when examining the interest rate policy of the ECB, Newey-West standard errors are used. While OLS is still unbiased in the presence of heteroskedasticity and autocorrelation, it is inefficient because the variances are biased, resulting in invalid test scores. The Newey-West standard errors are consistent in the presence of both heteroskedasticity and autocorrelation, solving the aforementioned problem.
To estimate the reaction function of the ECB, data from the entire Eurozone is needed. To be able to split the Eurozone in two groups, data from all separate member states is also required. The quarterly real GDP data is retrieved from the Statistical Data Warehouse of the ECB. As not all quarterly real GDP data from Luxembourg and Ireland is available at the ECB warehouse, this data is acquired from the database of the Organisation for Economic Co-operation and Development (OECD). To estimate the potential GDP, the Hodrick-Prescott filter (HP filter) is applied to the real GDP as proposed by Giorno et al. (1995). The HP filter is used to decompose a time series in a structural component and a cyclical component. A smoothness factor of 1600 is used, as recommended when applied to quarterly data.
A problem with the HP filter is the end-point problem: if the beginning- and end-points of the data set do not reflect similar points in the cycle, then the trend will be pulled upwards or downwards towards the path of actual output for the first and last few observations. To address the end-point problem, output data extending the estimated time period is commonly
used. By using data starting 16 quarters before respective entrance to the Eurozone this problem is mitigated at the start of the time period. As there are no quarterly forecasts available for every Eurozone country, the end of the sample can only be extended by real GDP data of the first quarter of 2015. In order to produce more accurate estimates, data from all individual countries is used to estimate respective potential output. These output gaps are then weighted by weighing the real GDP of each country relative to the real GDP of the Eurozone to obtain the output gap of the whole Eurozone.
Graph 1: Output gap in the Eurozone.
The Harmonised Index of Consumer Prices (HICP) is used as the inflation rate, as defined by the ECB as the measure used for inflation when conducting monetary policy. Official policy stance of the ECB is an inflation rate close to, but below, 2% over the medium term. Based on this an inflation target of 2% is used. Both headline inflation, which is the total inflation and the official ECB inflation aggregate, as well as core inflation, are used. Core inflation excludes the volatile energy and unprocessed food prices. Data on both inflation measures are retrieved from the ECB Statistical Data Warehouse.
+4.00% +3.00% +2.00% +1.00% 0.00% 1.00% 2.00% 3.00% 4.00% 1999Q 1 1999Q 3 2000Q 1 2000Q 3 2001Q 1 2001Q 3 2002Q 1 2002Q 3 2003Q 1 2003Q 3 2004Q 1 2004Q 3 2005Q 1 2005Q 3 2006Q 1 2006Q 3 2007Q 1 2007Q 3 2008Q 1 2008Q 3 2009Q 1 2009Q 3 2010Q 1 2010Q 3 2011Q 1 2011Q 3 2012Q 1 2012Q 3 2013Q 1 2013Q 3 2014Q 1 2014Q 3 Ou tp ut !G ap Period
Graph 2: Headline and core inflation in the Eurozone.
Average headline inflation over the ECB time span is 1.92%, in line with the target of below but close to 2%. The core inflation rate indeed exhibits a lower level of volatility.
To examine the reaction of the ECB to economic developments in the two groups the Eurozone countries are split into two groups, one group containing the strong member states, France and Germany, and the other group containing the remaining Eurozone members. Respective national data weighted by real GDP is used to measure the output gap and inflation gap of both groups.
Graph 3: Output gaps.
The respective output gaps of the two groups aligned closely until the crisis, where after they started to divert. +1.00% 0.00% 1.00% 2.00% 3.00% 4.00% 5.00% 1999Q 1 1999Q 3 2000Q 1 2000Q 3 2001Q 1 2001Q 3 2002Q 1 2002Q 3 2003Q 1 2003Q 3 2004Q 1 2004Q 3 2005Q 1 2005Q 3 2006Q 1 2006Q 3 2007Q 1 2007Q 3 2008Q 1 2008Q 3 2009Q 1 2009Q 3 2010Q 1 2010Q 3 2011Q 1 2011Q 3 2012Q 1 2012Q 3 2013Q 1 2013Q 3 2014Q 1 2014Q 3 In fla tio n! ra te Period Headline!inflation Core!inflation -4.00% -3.00% -2.00% -1.00% 0.00% 1.00% 2.00% 3.00% 4.00% 1999Q 1 1999Q 3 2000Q 1 2000Q 3 2001Q 1 2001Q 3 2002Q 1 2002Q 3 2003Q 1 2003Q 3 2004Q 1 2004Q 3 2005Q 1 2005Q 3 2006Q 1 2006Q 3 2007Q 1 2007Q 3 2008Q 1 2008Q 3 2009Q 1 2009Q 3 2010Q 1 2010Q 3 2011Q 1 2011Q 3 2012Q 1 2012Q 3 2013Q 1 2013Q 3 2014Q 1 2014Q 3 O ut put G ap Period
Graph 4: Core inflation rates.
Inflation rates of France and Germany were systematically lower than the inflation rates of the rest of the Eurozone until the last quarter of 2013, after which their inflation levels rose above the rest of the Eurozone. It is interesting to see in which way the ECB reacted to this development.
Finally, the Euro OverNight Index Average (EONIA) is used as the chosen interest rate over time by the ECB. This is the weighted average of Euro Interbank Offer Rates for inter-bank loans. Because this is an overnight rate, the ECB can control it daily through their standing facilities and consequently the EONIA is a good reflection of the interest rates chosen by the ECB. Data from the EONIA is also from the ECB Statistical Data Warehouse.
Graph 5: Short-term interest rate EONIA.
0.00% 1.00% 2.00% 3.00% 4.00% 1999Q 1 1999Q 3 2000Q 1 2000Q 3 2001Q 1 2001Q 3 2002Q 1 2002Q 3 2003Q 1 2003Q 3 2004Q 1 2004Q 3 2005Q 1 2005Q 3 2006Q 1 2006Q 3 2007Q 1 2007Q 3 2008Q 1 2008Q 3 2009Q 1 2009Q 3 2010Q 1 2010Q 3 2011Q 1 2011Q 3 2012Q 1 2012Q 3 2013Q 1 2013Q 3 2014Q 1 2014Q 3 Inf la ti on ra te Period
France & Germany Rest of the Eurozone
+1.00% 0.00% 1.00% 2.00% 3.00% 4.00% 5.00% 6.00% 1999Q 1 1999Q 3 2000Q 1 2000Q 3 2001Q 1 2001Q 3 2002Q 1 2002Q 3 2003Q 1 2003Q 3 2004Q 1 2004Q 3 2005Q 1 2005Q 3 2006Q 1 2006Q 3 2007Q 1 2007Q 3 2008Q 1 2008Q 3 2009Q 1 2009Q 3 2010Q 1 2010Q 3 2011Q 1 2011Q 3 2012Q 1 2012Q 3 2013Q 1 2013Q 3 2014Q 1 2014Q 3 In te re st !ra te Period
The sample period is from 1999 to 2014, this period is split in a pre-crisis period and a post-crisis period. The pre-post-crisis period contains data from 1999-2008Q3 and the post-post-crisis period from 2008Q4-2014. Data from 2015 is not included because the recent change in voting rules could influence the results.
IV.! RESULTS
First I will discuss the results obtained by using the Taylor rule with estimated coefficients. Following this, the results of the method using the Taylor rule with the original coefficients are discussed.
i.! Taylor rule with estimated coefficients
To!see!which!Taylor!rule!has!the!best!fit!for!this!study,!four!separate!regressions!are! performed.!!
Table 2: Regression Eurozone Taylor rule 1999-2014.
Without crisis dummy With crisis dummy Headline inflation Core inflation Headline inflation Core inflation Inflation Gap 0.44* (0.241) 0.96*** (0.292) 0.24 (0.289) -0.33 (0.210) Output Gap 0.50*** (0.116) 0.55*** (0.095) 0.67*** (0.108) 0.63*** (0.062)
Inflation gap crisis 0.51
(0.446)
2.88*** (0.223)
Output gap crisis -0.59*
(0.231) -0.63*** (0.122) Constant 2.14*** (0.156) 2.46*** (0.135) 2.10*** (0.187) 2.58*** (0.135) Number of observations 64 64 64 64 R-squared 0.3660 0.4257 0.4011 0.7634
Regression coefficients with the EONIA as the dependent variable, standard errors in brackets. The *, **, *** indicate a significance level of 10%, 5% and 1%
The first two regressions use data from the whole sample period without adjusting for the crisis. Two different inflation measures are however used in the two regressions, headline and core inflation. It is certainly not unthinkable that the crisis, where stimulating the economy could get a high priority next to the original goal of maintaining price stability, has had an effect on the way the ECB conducted their monetary policy. This is why the crisis is controlled for in the last two regressions, this is done by multiplying the inflation and output gap data by a dummy variable that takes on the value of 1 during the crisis period and 0 otherwise.
When using headline inflation as the inflation measure, a higher significance of the output gap compared to the inflation gap of the regressions is noticeable. Since the main goal of the ECB is price stability, this result is contrary to the theory. On the other hand, using core inflation as the inflation measure produces an inflation coefficient for the Eurozone that is almost equal to 1 and significant at the 1% level, more in line with the goals of the ECB. To examine which set of data better fits the statistical model, we look at the R2. Using core inflation also produces higher R2 results than the regressions where headline inflation is used. This result is interesting, as the headline inflation is the official aggregate used by the ECB. Possible explanation for this is the practice of interest rate smoothing which is not incorporated in the Taylor rule I have used. The core inflation is less volatile than the headline inflation and the EONIA was also relatively stable, quite possibly due to interest rate smoothing. This could cause the better fit of the core inflation.
Including the dummy variable to control for the crisis improves the R2 of both regressions, albeit only a small improvement is observable when headline inflation is used. Only the output dummy is significant using headline inflation, whereas the inflation dummy is not significant. On the other hand, both dummies produce significant coefficients using core inflation. As core inflation produces a higher level of explanation power and higher significant inflation coefficients, core inflation is chosen as the measure of inflation. Furthermore, controlling for the crisis produces a higher R2 score and two significant dummies for the core inflation regression, therefore resulting in the addition of the dummies. The original Taylor rule with core inflation and dummy variables controlling for the crisis is thus chosen as the model for the remainder of this study.
The constant of the regression has a value of 2.58, by subtracting the inflation target of 2% we obtain the equilibrium real interest rate. Subsequently, the equilibrium real interest rate over the whole period was 0.58%, this would be a very low value of the equilibrium real interest rate under normal circumstances. The inclusion of a crisis period however, where the central bank has to lower the interest rates to stimulate the economy, makes this low equilibrium
interest rate understandable. While the output gap was at the core of the interest rate setting behavior in the first period, there is a reverse trend observable in the second period to a focus on the inflation gap. The inflation coefficient is not significant over the pre-crisis period, but the crisis dummy is significant at the 1% level and obtains a coefficient of 2.88. Whereas the output gap has a coefficient of 0.63, the output dummy has a negative coefficient of -0.63, balancing it to 0. This indicates that output deviations had no effect on the interest rates during the crisis.
As the practice of modelling the reaction function of the ECB is a very complicated objective, there are several things to account for. To test if the model is correctly specified, the Ramsey RESET test is performed. The Ramsey RESET test tests whether non-linear combinations of the estimated values help explain the regressor, the interest rate in this case. If non-linear combinations do explain the regressor, then this is seen as evidence of omitted variable bias. The test produced a test score of F=15.52, which is significant at the 1% level as evidence that there is omitted variable bias. Because the Taylor rule is a fairly simple model which is used to model a complex mechanism, omitted variable bias is expected. The model also suffers from endogeneity; nominal interest rates are based on the real interest rate plus the inflation expectations. Because the inflation expectations are partly based on the present inflation rate the inflation also influences the interest rates in this way. This should be taken into account when considering the estimates. The Durbin-Watson test is performed to test for autocorrelation, with a d-statistic of 0.19 there is positive autocorrelation found. Finally, the White test is performed to test for heteroskedasticity, resulting in the rejection of homoscedasticity at the 5% significance level. By using the Newey-West standard errors however, heteroskedasticity and autocorrelation are already accounted for. To examine how the ECB reacted to the respective developments in both groups, coefficients over the whole period are estimated.
Table 3: Regression original Taylor rule using core inflation 1999-2014. France & Germany Rest of the Eurozone Inflation gap -0.49*** (0.161) 0.65*** (0.246) Output gap 0.65*** (0.051) 0.61*** (0.089) Inflation gap crisis 2.85***
(0.219)
1.18*** (0.367) Output gap crisis -0.78***
(0.105) -0.19 (0.179) Constant 2.45*** (0.144) 2.35*** (0.163) Number of observations 64 64 R-squared 0.8394 0.6660
Regression coefficients with the EONIA as the dependent variable, standard errors in brackets. The *, **, *** indicate a significance level of 10%, 5% and 1%
To test whether the coefficients on the different data sets are different, the Chow test is performed. The Chow tests whether the data from two groups, in this case France and Germany and the rest of the Eurozone, can be pooled by checking for structural breaks. Obtaining a test score of F=15.65 and thus rejecting the hypothesis of equal coefficients at a significance level of 1%. This indicates that the ECB did have different reactions to the respective economic conditions in both groups. More specifically, inflation and output movements of France and Germany better fit the interest rates set over the whole sample period, indicated by a higher R2. Significant negative coefficients for the inflation gap pre-crisis and the output gap during the crisis are however questionable. To obtain more information about the reaction of the ECB, another regression is done by splitting the sample in two periods, a pre-crisis period and a post-crisis period.
Table 4: Regression original Taylor rule using core inflation. 1999-2008Q3 2008Q4-2014 France & Germany Rest of the Eurozone France & Germany Rest of the Eurozone Inflation Gap -0.25* (0.134) 0.25 (0.163) 0.89** (0.411) 0.70*** (0.191) Output Gap 0.61*** (0.049) 0.54*** (0.071) -0.08 (0.096) 0.09 (0.109) Constant 2.74*** (0.088) 2.89*** (0.091) 1.17*** (0.329) 0.93*** (0.158) R-squared 0.7470 0.6627 0.1785 0.3856 Number Of observations 39 39 25 25
Regression coefficients with the EONIA as the dependent variable, standard errors in brackets. The *, **, *** indicate a significance level of 10%, 5% and 1%
There are vast differences observable between the reaction of the ECB on the respective inflation and output movements in both regressions. The reaction to the output gap in both groups is comparable over both periods but the difference arises with regard to inflation movements. Inflation movements in the rest of the Eurozone were insignificant in the pre-crisis period and the ECB decreased the interest rates following inflation increases in France and Germany. Especially with the expectation that Germany is inflation averse, this negative coefficient seems questionable. The data of this group does a better job in explaining the interest rate setting in the relatively normal, pre-crisis period though.
This!is different from what I observed in the estimates for the post-crisis period. Output
and inflation movements from both groups explain less of the chosen interest rates compared to the pre-crisis period. Especially the group of France & Germany, produces a low R2 value of 0.1785 compared to a R2 of 0.7470 for the pre-crisis period. The R2 of the rest of the Eurozone shows a more moderate decline, from 0.6627 to 0.3856. The group containing countries with weaker economic conditions seem to get more accommodated since the crisis according to this method.
ii.! Taylor rule with original coefficients
To take another look at the policy of the ECB, the original coefficients assumed by Taylor, 1.5 for the inflation gap and 0.5 for the output gap, are used to construct the interest rate path of both groups.
Graph 5: Taylor rates with original coefficients plotted against the EONIA
Only at the initial period of the ECB the interest rate is above the predicted rates of both groups. This is most likely due to the aim of the ECB to gain a credible reputation as an inflation fighting central bank (Kahn & Benolkin, 2012). Even while the inflation level was around the inflation target, the interest rate increased during this period. Interest rate levels then gradually adjusted to normal levels in 2002. As this behavior is not incorporated in the Taylor rule, deviations from the real chosen interest rate arise. This could also be a possible explanation for the negative value of France and Germany and insignificant value of the whole Eurozone of the inflation gap. From the point of view of the original Taylor rule, the interest rate since 2011 is too low for both groups. France and Germany preferred a lower interest rate over the whole period until the last quarter of 2013. As the EONIA is structurally below the preferred rate of the rest of the Eurozone since the end of 2001 the economic preferences of the strong member states seem to have been overweighed and this can be interpreted as evidence of political pressure in the Governing Council.
-1.00% 0.00% 1.00% 2.00% 3.00% 4.00% 5.00% 6.00% 7.00% 1999Q 1 1999Q 3 2000Q 1 2000Q 3 2001Q 1 2001Q 3 2002Q 1 2002Q 3 2003Q 1 2003Q 3 2004Q 1 2004Q 3 2005Q 1 2005Q 3 2006Q 1 2006Q 3 2007Q 1 2007Q 3 2008Q 1 2008Q 3 2009Q 1 2009Q 3 2010Q 1 2010Q 3 2011Q 1 2011Q 3 2012Q 1 2012Q 3 2013Q 1 2013Q 3 2014Q 1 2014Q 3 Int er es t r at e Period
V.! CONCLUSION
The aim of this study was to examine the presence of political pressure in the Governing Council. The presence of political pressure was defined as the overweighting of the economic preferences of the two biggest member states, France and Germany, on the interest rate setting behavior of the ECB. By comparing the explanation power of the regressions of the pre-crisis and post-crisis period for France and Germany, inferences could also be made about the presence of higher political pressure during the crisis as suggested by Sturm and Wollmershäuser (2008). The method chosen to accomplish this was making use of the original Taylor rule with core inflation. Taking into account that the ECB uses headline inflation as the signal for medium-term inflation when conducting monetary policy, the higher explanation power of the regressions using core inflation, was noteworthy.
The results in the pre-crisis period between the two groups exhibited the characteristics of a slightly better fit for the stronger member states. This weighting scheme changed however in favor of the group without France and Germany during the post-crisis. Where increased political pressure was expected during the crisis, the results showed the opposite trend. The interest rates were accommodating the weaker member states during the crisis instead of the two most powerful countries of the Eurozone.
Another method was used to examine the monetary policy of the ECB. By using the coefficients originally proposed by John Taylor, the Taylor rates were calculated and plotted against the EONIA. Since the end of 2001 the interest rates were structurally too low from the perspective of the weaker member states, supporting the claim that the interest rates seemed to favor the group of France and Germany. No evidence is found of increased political pressure during the crisis period, on the contrary, interest rates seem too diverge from the Taylor rates of France and Germany. Both methods seem to support the claim that over the whole period the group with France and Germany produced better results, here implying the presence of political pressure, while during the crisis period, the weaker member states have been accommodated.
The main limitation of this study is the use of a fairly simple model for modelling a very complex process. Only output and inflation data are used to explain the interest rate setting procedure, while in reality this is a more refined process. This causes the estimates to be susceptible to bias. Further research can be aimed at analyzing the impact of the recent change in voting rules on the interest rates. As the ‘one-country, one-vote’ principle, that was employed during the period of 1999-2014, is now replaced by a voting system where the five
member states from Tier 1 receive relatively more votes. Theoretically, the change in voting rules should have no impact on the policy of the ECB as all members are assumed to have a Euro-wide perspective. Interesting question is however if the higher voting power of the Tier 1 member states has resulted in interest rates more fitting their economic preferences compared to the ‘one-country, one-vote’ voting system.
!
VI.! REFERENCES
Aksoy, Y. and De Grauwe, P. and Dewachter, H. (2002). “Do asymmetries matter for European monetary policy?” European Economic Review, 46, pp. 443–69.
Arnold, I. (2005). “Optimal regional biases in ECB interest rate setting.” NRG working paper,
05(01)
Belke, A. and Klose, J. (2009). “Does the ECB rely on a Taylor rule? Comparing ex-post with real time data.” ROME Discussion Paper Series, 09(05)
Belke, A. and Polleit, T. (2007). “How the ECB and the US Fed set interest rates.” Applied
Economics 39: pp. 2197-2209
Berger, H. and De Haan, J. (2002). “Are small countries too powerful within the ECB?”
Atlantic Economic Journal, 30(3), pp. 263-282
Berger, H. and Mueller, T. (2007). “How should large and small countries be represented in a currency union?” Public Choice, 132, pp. 471-484
Dixit, A. (2001). “Games of monetary and fiscal interactions in the EMU.” European
Economic Review, 45, pp. 589-613
Duisenberg, W. (2002), “Some remarks on the euro in a US context, Speech by Dr. Willem F. Duisenberg, President of the European Central Bank, at a breakfast meeting of the Council on Foreign Relations”, New York, 19 April, www.ecb.int.
Gerdesmeier, D. and Roffia, B. (2003). “Empirical Estimates of Reaction Functions for the Euro Area.” ECB Working Paper 206
Gildea, J.A. (1992). “The Regional Representation of Federal Reserve Bank Presidents.”
Journal of Money, Credit and Banking, 24(2), pp. 215-225
Giorno, C. and Richardson, P. and Roseveare, D. and Van den Noord, P. (1995). “Potential Output, Output Gaps and Structural Budget Balances.” OECD Economic Studies 24: pp. 167-209
Gorter, J. and Jacobs, J. and De Haan, J. (2008). “Taylor Rules for the ECB using Expectations Data.” Scandinavian Journal of Economics, 110(3), pp. 473-488
Hayo, B. and Méon, P-G. (2013). “Behind closed doors: Revealing the ECB’s decision rule.”
Journal of International Money and Finance, 37
Heinemann, H. and Huefner, F.P. (2004). “Is the View form the Eurotower Purely European? – National Divergence and ECB Interest Rate Policy.” Scottish Journal of Political Economy, 51(4), pp. 544-558
Kahn, G.A. and Benolkin, S. (2007). “The Role of Money in Monetary Policy: Why Do the ECB and the Fed See it so Differently?” Economic Review 2007(3): pp. 5-36
Kool, C.J.M. (2005). “What Drives ECB Monetary Policy?” Tjalling C. Koopmans Research
Institute Discussion Paper series, 05(03)
Meade, E.E. and Sheets, D.N. (2002). “Regional Influences on U.S. Monetary Policy: Some Implications for Europe.” International Finance Discussion Papers, 721
Meade, E.E. and Sheets, D.N. (2005). “Regional Influences on FOMC Voting Patterns.”
Journal of Money, Credit and Banking, 37(4), pp. 661-677
Peersman, G. and Smets, F. (1999). “The Taylor Rule: A Useful Monetary Policy Benchmark for the Euro Area?” International Finance, 2(1), pp. 85-116
Sturm, J-E. and Wollmershäuser, T. (2008). “The Stress of Having a Single Monetary Policy in Europe.” KOF Working Papers, 190
Taylor, J.B. (1993) “Discretion versus Policy Rules in Practice.” Carnegie Rochester
Conference Series on Public Policy 39: pp. 195–214
Von Hagen, J. and Brückner, M. (2002). “Monetary and Fiscal Policy in the European Monetary Union.” Monetary and Economic Studies, 20, pp. 123-165
(
VII.! APPENDIX Ramsey-RESET test:
Durbin-Watson test for autocorrelation:
( (
White’s Test for homoscedasticity:
( ( Chow Test: 3 = 450 46+ 47 /9 46+ 47 /(:6 + :7029)! Prob > F = 0.0000 F(3, 56) = 15.52 Ho: model has no omitted variables
Ramsey RESET test using powers of the fitted values of EONIA
Durbin-Watson d-statistic( 5, 64) = .1916415 Total 34.54 15 0.0029 Kurtosis 1.15 1 0.2831 Skewness 11.34 4 0.0230 Heteroskedasticity 22.05 10 0.0148 Source chi2 df p Cameron & Trivedi's decomposition of IM-test
Prob > chi2 = 0.0148
chi2(10) = 22.05
against Ha: unrestricted heteroskedasticity White's test for Ho: homoskedasticity
Table 5: Hypotheses Chow Test Hypothesis Taylor rule
H0 <6, <7, <>, <?@ABCDE.&.GEAHBCI = <6, <7, <>, <?JEK".LM."NE.OPALQLCE
H1 <6, <7, <>, <?@ABCDE.&.GEAHBCI ≠ <6, <7, <>, <?JEK".LM."NE.OPALQLCE
Table 6: Chow test F-statistic
Components Score 4S 0.011140523 41 0,002377317 42 0,004943998 9 4 :1 64 :2 64 :1 + :2029 120 T 15.6497
This table shows the Chow test statistic. The critical F value 2 = .01 is 3 > 4.9774
( ( ( ( (
