University of Amsterdam
Quantitative Easing at the European
Central Bank
The effect of the European Central Bank’s QE
announcements on the state of the economy
Bas Loman Bachelor Thesis Faculty of Economics & Business Economics & Finance Supervisor Dhr. dr. D.F. Damsma Date June 29, 2016 Abstract As of October 2008 the European Central Bank (ECB) started to use non-standard monetary policy measures to repair and stimulate economic activity. Applying an event study methodology this research analyzes the effect of the non-standard quantitative easing (QE) related announcements by the ECB on the European stock market and the expected inflation from October 2008 to March 2016. Both show in general positive, and at specific events significant impacts of the announcements. The study provides evidence that suggests that QE had a positive immediate impact on the state of the economy. Keywords ∙ Quantitative easing ∙ Monetary policy ∙ European Central Bank ∙
Statement of Originality
This document is written by Bas Loman 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.
1
Introduction
Before the financial crisis erupted in 2008 central banking was a rather boring profession. At the European Central Bank were used fairly normal tools. Main Refinancing Operations (MRO’s) serve to steer the short-term interest rate, to manage the liquidity situation and to signal the monetary policy stance in the euro area. These measures were used to help and steer the economy where needed in order to maintain price stability, the primary objective of the European Central Bank (ECB, 2011). When in 2008 the financial crisis started central banks were made aware and found out that the regular tools were not sufficient to reach their target. A set of new tools were introduced and used by multiple major Central Banks including the ECB. This set of tools included policy measures that could be qualified as quantitative easing (Rodriguez & Carrasco, 2014). However economists still debate on whether these new tools are in fact quantitative easing (QE). In the next chapter this will be explained in a deeper sense.
The quantitative easing project of the ECB has been and still is debated much more than the similar projects at the Federal Reserve and the Bank of England. The main difference is that the sovereign debt crisis followed up the financial crisis in 2010 (Fawley & Neely, 2013). Many Dutch economists, including the current president of the Dutch Central Bank Klaas Knot and former director of the Dutch Central Bank professor dr. Lex Hoogduin, and also foreign ones like professor Mark Blyth have reacted with aversion to the policy of mister Draghi, the president of the European Central Bank.
There is a significant conflict between two thoughts. On the one hand the more aggressive and long-term use of unconventional monetary tools and the belief of mister Draghi and economists that it will help the economy eventually restore (Martin & Milas, 2012). On the other hand, the more skeptical thought that keeping the interest rates low and not directing the money directly enough to the people will feed a damaging cycle of booms and busts (Blyth & Lonergan, 2014).
Both proponents and opponents however, show empirical background, and while there has been done research regarding the impact of QE on the economy, it is difficult to say, with the real effects hard to measure, to what extent QE impacted the economy. It could be interesting to look at more measurable variables with one more focused on the financial sector and the other one on the real economy. This leads to the
following research question: “What has been the effect of QE announcements by the
European Central Bank on the financial sector and the real economy?”
This research makes use of an event-study with the QE related monetary policy announcements of the ECB as the events, the S&P-EURO stock index as indicator of the financial sector and the expected inflation in the euro area as indicator of the real economy. The event timeline runs from the first announcement of fixed-rate full-allotment for main refinancing operations and the lowering of the key interest rates in October 2008 to the latest extension of the asset purchase program (APP) with the corporate sector purchase program (CSPP) in March 2016. In the next chapter an outline of the current theory and the implementations of the ECB regarding QE will be addressed. Then a clear explanation will be given in the methodology about the way of testing and on what grounds it is based. Furthermore the results will be analyzed in the analysis and a summary and concluding remarks will be given in the conclusion.
2
Current theory regarding QE in the Eurozone.
In the literature review the following points will be addressed. First there will be an outline of the current literature on what quantitative easing is and in what sense the European Central Bank uses this monetary tool to influence the rather poor state of the economy the last seven years. With this clarity in mind the proponents and opponents of the quantitative easing project and the influence it had on the economy will be described. Finally we will be looking at what the current literature says about the impact of quantitative easing on the one hand on the financial market and on the other hand on the real economy. The empirical data that most of these articles have used to support their findings will be addressed as well. 2.1 An explanation of the origin of QE at the ECB. First of all to understand what quantitative easing is and in what sense the ECB uses this program, it is necessary to address the objectives of the ECB and the monetary policy instruments it is using in order to meet this objectives. As described in the introduction
the main goal of the ECB is to maintain price stability. In 1993 the signing of the Maastricht treaty by all the member states formed the European Monetary Union (EMU). At that time maintaining price stability was thought to be the best contribution monetary policy can make to economic growth, job creation and social cohesion (ECB, 2009). Therefore price stability became the primary objective of the ECB. Because the treaty does not give a clear description of price stability, the governing council of the ECB announced the following definition in 1998: “Price stability is defined as a year-on-year increase in the Harmonized Index of Consumer Prices (HICP) for the euro area of below 2%. Price stability is to maintained over the medium term.” In 2003 the Governing council of the ECB has undertaken a thorough evaluation regarding its policy. They agreed that in the pursuit of price stability and to provide a safety margin against the risks of deflation it will aim to maintain inflation close to 2% over the medium term. The ECB finds this to be low enough for the economy to fully reap the benefits from price stability (ECB, 2011). The ECB uses a set of monetary policy instruments, which provide the operational framework for the implementation of decisions in practice. They determine which money market interest rate is needed to maintain price stability and use the instruments and procedures to achieve this. In normal times without any crises and major macroeconomic shocks the ECB uses Open Market Operations (OMO) and standing facilities. The OMO firstly consists of the Main Refinancing Operations (MROs), which are weekly and have a maturity of a week as well, and secondly it consists of Longer Term Refinancing Operations (LTROs), which are used monthly and have a maturity of 3 months. By setting the rates on the standing facilities, which allows credit institutions to obtain overnight liquidity or to make overnight deposits from and with the central bank, the ECB determines the corridor in which the overnight rate can fluctuate. The overnight rate is measured by the euro overnight index average (EONIA) and is generally close to the rate on the MROs. The MROs are the most important open market operations and play the key role in setting the short-term interest rate (ECB, 2011). The transmission of the policy goes through transmission mechanisms in the Eurozone and has several stages. The money market is the first market, which is affected by changes in monetary policy. A good working money market is essential because it ensures an even delivery of central bank liquidity and an equal level of short-term interest rates through out the Eurozone (ECB, 2009). Furthermore it is typically assumed that short-term interest rates are changed by monetary institutions to have
influence on asset prices. These prices will have an effect on the state of banks to lend, firms to invest, or people to consume which combined potentially affects the level of output and employment (Fawley & Neely, 2013).
‘There was a time, not too long ago, when central banking was considered to be a rather boring and unexciting occupation [...] I can confidently say that this time has passed’. Thus said Mr. Draghi on April 15th 2013. Since the onset and the outburst of the
financial crisis in 2007 central banking’s boring and normal times are passed. The central banks have to deal with and solve new problems and face different challenges these days. A much more active monetary policy is used by the European Central Bank and other major central banks all over the world (Rodriguez & Carrasco, 2014). The financial crisis was followed by years of serious recession and economists refer that to as the ‘Great Recession’. This period in time consisted of a decline in economic activity across most major economies. This recession has not only been severe but also enduring and the economy is still trying to recover (Bowdler & Radia, 2012). The central banks that were facing the prospects of a deep economic downturn had to act, but with the short-term interest rates near the zero lower bound the use of regular monetary tools ran out. Therefor a new set of tools and a range of measures that may be described as unconventional have been used ever since to still meet the objectives of the central banks (Bowdler & Radia, 2012).
At the Federal Reserve and the Bank of England these new balance-magnifying measures were directly stated as quantitative easing. On the measures taken by the ECB however there is still an ongoing debate on whether it is QE or not. The ECB itself never used the term QE until the end of 2014, but used non-standard instead (Fawley & Neely, 2013). The two most important measures that the ECB took before that time were reducing the MRO rate to 1 percent and the enormous amount of 1 trillion of LTROs with increased maturities of 6 and 12 months. With these LTROs the ECB has managed a massive expansion of it balance sheet and it has been called the Eurozone equivalent of quantitative easing, as done by the FED and the BoE (Pisani-Ferry & Wolff, 2012). The big argument here is that the ECB did not purchase any government bonds and therefor state that it did not use QE. Whether or not the ECB used QE is debatable. There are several different definitions and descriptions of QE and a few will be described and laid along one another.
The traditional clarification of quantitative easing gives three conditions: 1). An explicit target for bank reserves providing liquidity to realize a current account targeted excessively of the required reserves. 2). Conditional commitment to maintain high reserve levels into the future. 3). Increased purchases of long-term government bonds to facilitate the attainment of the target bank reserves (Ugai, 2007). ECB measures do not particular fit these requirements, but on the other hand Mark Spiegel (2001) characterized quantitative easing as aimed to reduce long-term interest rates when policy rates are close to zero. This description of quantitative easing fits the ECB much better compared to Ugai (2007). The former president of the Federal Reserve Ben Bernanke split QE into credit easing and pure QE. Where credit easing is used to reduce specific interest rates and restore market function, pure QE characterizes policies that unusually increase the magnitude of central bank liabilities, currency and bank reserves, mostly at the zero lower bound (Fawley & Neely, 2013). The ECB has initiated lending programs that in this way could be considered as pure QE, because of the targeted reserves and the wide range of assets as collateral. To enumerate monetary policy has been in an accelerating change since the start of the financial crisis. Bearing this in mind makes it more difficult to match the new measures with old categories. One could say that a new standard of quantitative easing has been set, regardless of calling it non-standard or unconventional (Rodriguez & Carrasco, 2014). All in all the conditions and clarifications above lead to an explanation were QE is defined as non-standard mostly balance magnifying monetary policy measures, which try to provoke or stimulate low economic activity and try to reinstate economic variables which have non-standard values when regular mechanisms fail to do so. 2.2 Outline of the implementation of QE in the Eurozone. The aim of the non-standard measures, as the ECB names them, before 2014 was to remedy the impairments in various stages of the transmission mechanism. (ECB, 2015). Rodriguez and Carrasco (2014) did a chronological analysis of the policy responses between 2007 and 2014. In the first phase between September 2008 and April 2010 the ECB decided to install a fixed rate system with full allotment tenders. Due to this policy banks with liquidity shortages could refinance at a fixed rate and a
predetermined period. This has made ECB money and bank reserves largely endogenous-determined since October 2008. The MRO (Main Refinancing Operations) interest rate was reduced from 4.25 to 1.00 and the corridor was narrowed from 200 basis points to 100 and back to 150 at the end of 2009. Narrowing the corridor width reduces the volatility margin of the interbank rate and helps to steer it more precisely. Furthermore the Long-Term Refinancing Operations (LTROs) were introduced with maturity of three and six months in 2008 and 12 months LTROs in 2009. These LTRO’s were part of regular toolkit of the ECB, but became non-standard due the large quantity and the prolong maturity. Together they had a value of 742 billion euros. Finally the ECB came up with the Covered Bond Purchase Program 1 (CBPP1) to stimulate the ongoing decline in money market term rates. The aim of this program has been to support funding conditions for banks and enterprises, to encourage credit institutions to preserve and increase there lending to clients. Next to that it had the goal of enhancing market liquidity in important segments of the private debt securities market. As a result For all that, and despite the expectations this program only reached the small amount of 61 billion euros. If all liquidity that was injected is summed up it increased the balance sheet of the ECB with 30 percent in a year whereas in regular times the annual growth rate was about 4 percent (Rodriguez & Carrasco, 2014). The ECB still made the presumption that the crisis would be short-lived and that all non-standard policies were temporary in nature (ECB, 2011). In Phase II from May 2010 to August 2011, the first part of the sovereign debt crisis prevailed. At that time the ECB was still reluctant in acting as a lender of last resort. The sole measure worth observing is the Securities Markets Program (SMP). It was launched to address several damaged securities markets, which appeared to threaten the mechanism which is used to transmit the policy rates to the market interest rates of longer maturity and in other market segments, as well to the real economy and prices. Not more than a year later in March 2011 Greece, Ireland and Portugal were bailed out by the intervention of the ECB in the secondary bond markets (Ghysels, Idier, Manganelli, & Vergote, 2014). As of August 2011 both financial and the sovereign debt crisis intensified and the ECB did the important announcement of becoming a true lender of last resort and declaring that non-standard measures would be there as long as necessary. In 2011 and 2012 the ECB purchased 220 billion of stressed sovereign bonds and a total of 1018 billion euros of Very Long-Term Refinancing Operations with 36-month maturity. Because negative bond markets
developments were noticed in Italy and Spain the MRO interest rate was cut to 0,75 percent (Rodriguez & Carrasco, 2014).
In June 2014 the European Central Bank started introducing a new set of tools. The purpose of these tools and measures was to enhance the transmission mechanisms but also to reinstate the accommodative monetary policy with the weak inflation outlook in mind (ECB, 2015). In other words the standard operating framework proved insufficient and needed to be stimulated for two reasons. First, the mechanism from which the price of the central banks reserves are transmitted to the broader financing conditions that affect investment and consumption of firms and households, was severely damaged. Second, the duration of the global financial crisis led to the provision of monetary boost to the economy, because short-term rates were at the zero lower bound (ECB, 2015). At this time the ECB used quantitative easing for the first time to name its program. This probably was the case because the inflation was approaching the deflation area and because the ECB new what the following Purchase programs in September would look like. The Key ECB interest rate was cut for the first time below the zero bound. A negative rate on the deposit facility of -0.20 percent was a fact. In September 2014 two asset purchase programs were installed as part of the Asset Purchase Program (APP). The asset-backed securities purchase program (ABSPP) and the third covered bond purchase program (CBPP3). The ABSPP is seen as providing the asset class with a credibility boost and CBPP3, which exits of private assets purchases, has had a downward impact on covered bonds spreads. Furthermore, they intended to enhance and facilitate credit provision to the euro area economy and to generate positive spillovers to other markets. As a result it should ease the monetary policy attitude of the ECB and contribute to a return of inflation to a stable level near 2 percent (ECB, 2015).
The last big change in policy was the announcement of the expanded APP in January 2015. The policy rates at this point were still at the lower bound. This program was used to ease the monetary policy stance as the ECB described it itself (ECB, 2015). The APP program consists of the CBPP3, the ABSPP and was added with the Public Sector Purchase Program (PSPP). The purchases under this program were 60 billion per month and grew to 80 billion euros per month as of March 2016. At the same time the program was extended with the Corporate Sector Purchase Program (CSPP), which
means that there will be outright purchases of bonds issued by non-bank corporations. The program is intended to have duration until September 2016 (ECB, 2015).
To put all the events discussed above regarding QE in a nutshell, one could make a distinction between the events that do and the events that do not or nearly fit the profile of being Quantitative Easing stated at the end of paragraph 2.1. While all measures described are non-standard, an enlargement of the balance sheet is considered as the most pure form of QE (Bernanke, 2008). This contains the extended LTRO’s, CBPP’s and the APP. The SMP program may be defined as QE as well, despite the fact that due to sterilization the balance sheet did not increased. Although the fact that the ECB only started using the term QE after the APP program was announced one could make up that most non-standard monetary policy measures since 2008 have been or nearly are quantitative easing. 2.3 Positive implications regarding QE in the EMU. Cour-Thimann & Winkler (2012) argue that the non-standard measures taken after the start of the financial crisis in 2008 as described above have been competent in stabilizing the financial system, the economy and furthermore it ensured price stability, the main goal of the European Central Bank. Looking from a empirical perspective, a set of model-based exercises with counterfactual scenarios was used by Lenza et al. (2010) to finger point the question of what would have happened if the ECB had not implemented the non-standard policy measures following the beginning of the financial crisis. A structural vector autoregressive (VAR) analysis, in which the unconventional measures are being segregated from the standard policy interest decisions as discussed at the start of this chapter, indicates that the ECB can effectively stimulate the economy by increasing the size of the balance sheet or the monetary base.
When in 2010 the sovereign debt crisis started with acute market expectations about a possible default of Greece impacting on Ireland, Portugal, Spain and Italy, the ECB established its Security Market Program (SMP), which helped to avoid an uncontrolled increase in sovereign bond yields. This helped keeping down the financing costs for the economy and thereby avoiding adverse implications for price stability (Cour-Thimann & Winkler, 2012). Furthermore the SMP helped to avoid contagion across
other countries and protected the monetary transmission in most of the euro area (Cour-Thimann & Winkler, 2012).
In the ECB Research Bulletin of the summer 2015 Lenza, M published an event study on the effect of the announcements of the ECB about the Outright Monetary Transactions in Italy, Spain, France and Germany. All in all the study suggest that the OMT announcements were associated with an increase in economic activity, consumer prices and bank loans in Italy and Spain. In Germany and France the effect was moderate but still of measurable size and importance. These results suggest that the non-standard policy tools acting through the changes in asset prices, due to for example large scale asset purchases, may have non standard effects on the main macroeconomic variables in the area (Altavilla, Giannone, & Lenza, 2015). In addition after a presidential ECB meeting in December the president of the ECB mister Draghi came with a statement that included several important matters about the APP program, which the ECB states as QE, and its outcomes (Draghi, 2015). He stated that the 10-year government bond yields fell by around 120 basis points between June 2014 and December 2015. The same decline holds for bonds issued by firms and banks. Due to this the cost of credit for the whole euro area fell by approximately 80 basis points. The ECB produced a figure which stated that in the absence of the measures of the ECB the inflation would be at least half a percentage point lower in 2016, and around a third of a percentage point lower in 2017 (Draghi, 2015). Draghi (2015) also stated that the impact on GDP was very reasonable. The policy measures are committing to raise GDP by almost 1 percent in the years 2015 to 2017. All in all these measures have had and will be having an effect on the economy (Draghi, 2015). 2.4 Contravening thoughts regarding QE.
Several economists and studies are skeptical regarding QE. A number of these will be addressed here. To start with the LTROs measures of the ECB, in 2010 and 2011 these measures helped banks and sovereigns to deal with short-term liquidity issues, but the intended short-term fix did actually increase the long-term yields slightly (Tempelman, 2012). In addition Tempelman (2012) argues that it is inappropriate for the ECB to carry out outright large purchases of government debt. The ECB is not going to buy debt issued by top-rated countries like Germany and the Netherlands, but rather the countries that are in the so-called periphery of the Eurozone. For example Italy and
Spain. The debt issued by these countries is not considered to be credit-risk free, and that means that by purchasing these the ECB transfers credit risk from capital market participants to European taxpayers. Martin & Milas (2012) did a study on QE in general in 2014 and found that QE programs did succeed in lowering bond yields at the longer end of the yield curve. Nonetheless econometric studies propose that these effects may have been only temporary. There is little evidence that successive QE programs have much effect (Martin & Milas, 2012). There was also evidence that QE had been a rather weak policy instrument. Initial large asset purchases had an effect comparable to lowering the policy rate with 200-300 basis points. Subsequent programs had little effect (Martin & Milas, 2012). For these subsequent rounds of QE Putnam (2013) argues that the reviews are clearly diverse and densely dependent on the assumptions embedded in the economic assumptions used by researchers. When a closed economy is assumed the long-term rates are affected much more then when it is part of a global integrated economy (Putnam, 2013). Econometric estimates of the responses of the QE programs on the real economy, measured by output and inflation are broadly similar. QE increased GDP growth with 1-3 percent with a comparable effect on inflation. This suggests a positive effect on reducing the downturn in economic activity. On the other hand does it suggest that QE is not strong enough to light up an economic recovery (Martin & Milas, 2012). This argument is tentative, because there is relatively little evidence on the effects of QE and most of these studies use the same methodology (Martin & Milas, 2012). Neely (2014) argues that economists should respect more the limited predictability of asset prices in estimating dynamic relations. These asset prices are used in the SVAR analyses. Forecasts of asset prices should be met with caution. There is a good reason to believe that asset prices should display limited predictability and empirical background literature has confirmed this multiple times (Neely, 2014).
Just to point it out, Blyth & Lonergan (2014) came with a more radical and transformative thought. They recalled that al the measures that the ECB has taken brought the Eurozone to a zero and in some countries even negative inflation level. They argue that in a poor condition the current monetary policy could lead to further instability and stagnation and that the system requires a new monetary policy framework (Blyth & Lonergan, 2014).
2.5 All in all the literature summed up.
To sum up, the current literature of what is feasible is mostly positive about the implementation of the quantitative easing or non-standard policy measures taken by the ECB. The result of Lenza et all (2010) showed that an increasing balance sheet positively affected the economy and furthermore in 2015 Alta villa, C et all (2015) found that OMT announcements had a positive effect on the economic activity, consumer prices and bank loans. On the contrary critics found that the ECB was short-term oriented (Tempelman, 2012), and that the long-term rates were not affected the right way. Further the credit risk is transferred from capital market participants to the taxpayers. Putnam (2013) stated that economic assumptions being made influenced the outcome on the long-term yields too positively. In addition Neely (2014) states that economists should bear in mind much more the limited predictability of asset prices in estimating dynamic relations. At that point some economists go even further with questioning the role and the institutional framework of the European central bank (Blyth & Lonergan, 2014).
As one can see, most studies regarding QE and the effect it has on the economy find positive relations, although small and not with certainty. Existing and current literature measured most of the effects through interest rates, GDP growth and bond asset prices. Given all theories and considering the transmission mechanism of the ECB, a good measures of the state of the economy needs to be addressed. Regarding the transmission mechanisms the following is observed. The interest rate is used by monetary institutions to have influence on asset prices, which effects the state of banks to lend, firms to invest, and people to consume which combined potentially affects the level of output and employment. Because the lag in output and unemployment rates is to big and to unsecure they will not be good variables for measuring immediate effects of QE. The state of the economy will be measured by the stock market and the expected inflation. These are both very liquid, which makes them appropriate to use for immediate effects. Again Lenza et all (2010) and Altavilla et all (2015) already performed research regarding the effect of ECB announcements on economic activity, but did not use the correct variables to describe this subject. With the Stock market and the expected inflation two corner stone economic figures are going to be researched (Cour-Thimann & Winkler, 2012) (Moessner, 2015). From the theory above and the research that is going to be done I derive the following two hypotheses.
Hypothesis 1 “The QE related announcements of the ECB result in abnormal positive returns for the European stock market.” Hypothesis 2 “The expected inflation experience higher growth rates due the QE announcements in the Eurozone.”
Both these hypotheses will be tested using a event study. In the next chapter the research method of testing both hypotheses will be explained and the way the data was collected.
3
Methodology
In this section, the methodology that is used to test the data is described and discussed. First of all, there will be an outline of the data that is used in the research, at what resources the specific data was found and selection criteria were adopted. In the following paragraph I provide a summary of descriptive statistics concerning the data. From there on I provide knowledge about the testing methodology. First I will conduct what impact the ECB announcements, which are part of the non-standard measures, had on the stock market in the Eurozone. Furthermore, the relation between these announcements and long-term expected inflation is tested. There will be described what tests are used and in what sense they could contribute to the findings in this research.3.1 Data
Because an event study is going to be used to test the data, the events need to be collected and specified. Twelve ECB announcements ranging between October 2008 and March 2016 were collected by hand from ECB Press releases and financial ECB timelines in multiple articles (Altavilla, 2014; Ricci, 2015). The events were selected under the criteria stated in chapter 2.1 and 2.2. These criteria defined QE as non-standard mostly balance magnifying monetary policy measures, which try to provoke or stimulate low
economic activity and try to reinstate economic variables which have non-standard values when regular mechanisms fail to do so. Table 1 provides a list of the twelve events with a short description of what measures were taken and an explanation of what they were meant to do.
Table.1. QE announcements ECB
Event Date Description of measures
#1 08-10-08 Introduction of the fixed-rate full-allotment (FRFA) for MROs. And to lower its 3 key interest rates by 50bp
#2 07-05-09 ECB announced 3 one-year maturity LTROs, and launched the first covered bond purchase program.
#3 10-05-10 Introduction of SMP, and reintroduction of the FRFA for LTROs
#4 06-10-11 CBPP2 announced, and two one-year LTROs
#5 03-11-11 Details on CBPP2, and Lowered key interest rate by 25 bp to 1.25%
#6 08-12-11 Two 3-year LTROs and Looser rules for collateral on ABS. The interest rates were lowered by 25bp.
#7 02-08-12 The ECB hinted to the fact that, within its mandate to maintain price stability over the medium term and in observance of its independence in determining monetary policy, may undertake outright open market operations of a size adequate to reach its objective.
#8 06-09-12 ECB announces new program of buying sovereign debt. Details on the implementation of the OMTs and measures to preserve the collateral availability for refinancing operations.
#9 05-06-14 Actual interest rate lowered by 10bp to 0.15%. Deposit rate first time below zero to -0.10%. Launch of the Targeted LTROs.
#10 04-09-14 Launce of ABSPP and CBPP3, and lowered the 3 key interest rates. Interest rate reduced to 0.05% and the deposit rate to -0.20%.
#11 22-01-15 Announcement of Expanded APP adding Public sector Purchase
Program in addition of ABSPP and CBPP3.
#12 10-03-16 ECB announces to add Corporate Sector Purchase Program to the APP. And increased combined purchases from 60 to 80 billion per month.
Regarding the European stock market data, there is made use of the S&P EURO index. This is a sub-index of the S&P Europe 350 and it includes all Eurozone domiciled stocks from parent index. The index is designed to be reflective of the Eurozone market, yet efficient to replicate. The S&P EURO contains constituents from within the Eurozone only. The index constituents are leading companies from all the ten sectors within the Global Industry Classification Standard (GICS). The data was collected from S&P indices, which provides indices for markets all over the world. Furthermore for the expected inflation there was made use of the Euro Inflation Swap Forward 5y5y. This series is a measure of expected inflation (on average) over the five-year period that begins five years from today. The rate is a common measure, which is used by central banks and dealers, to look at the market’s future inflation expectations. The rate is calculated using the following formula: 𝐹𝑊𝐼𝑆𝐸𝑈55 = 2 ∗ 𝐸𝑈𝑆𝑊𝐼10 − 𝐸𝑈𝑆𝑊𝐼5 Where 𝐸𝑈𝑆𝑊𝐼5 and 𝐸𝑈𝑆𝑊𝐼10 are the five and ten year expected inflations swaps of the euro area. These swaps are traded daily and therefor have a good functional form to perform an event study. Historical prices were collected from Bloomberg. Next to five-year, five-year forward inflation swap the EUSW15 is used to estimate what the effect of the non-standard measures of the ECB were. This five-year Euro Inflation Swap is the expected inflation at a given date five years from that specific date. These series were also collected from Bloomberg. 3.2 Descriptive statistics A short summary of the descriptive statistics is given in table 2 below. The lowest return on the S&P EURO was on October 6th 2008, the day after the financial crisis in Iceland emerged. The highest return on the S&P EURO was on October 13th 2008. This was after the weekend where the European leaders met to announce the plans of recapitalizing major European banks. Where the 5y5y forward inflation had its highest return similar as the S&P EURO on the 13th of October 2008, the EUSWI5 is much more volatile and had
is peak return in December 2014 and the biggest downfall on January 6th 2015 with the
oil price dropping below $50 for the first time.
Table. 2. Descriptive Statistics
Variable N Mean Std. Dev. Min Max
S&P EURO 2133 1209.38 205.53 724.80 1752.78 Return on S&P EURO 2133 0.0008% 1.5187% -7.8291% 10.6559% FWISEU55 2166 2.21875 0.32225 1.361 2.7985 Return on forward Inflation Swaps 2166 -0.01515% 1.441% -7.82502% 9.58515% EUSWI5 2166 1.61223 0.49108 0.4335 2.9405 Return on EU 5y inflation swaps 2166 -0.0210% 2.4225% -18.4384% 18.2599% 3.3 S&P EURO return on ECB announcements
To measure and estimate to what extent the ECB announcement were affecting the economy two event studies will be pursued. The first one is the effect of the non-standard ECB measures on the European stock market. As table 1 in shows I conducted 12 events running from the first reduction of the key interest rates and installing the fixed rate full-allotment procedure in October 2008, to the latest expansion of the APP with Corporate Sector Purchase Program in March 2016. With an event study one looks for a significant abnormal return within the event window. To calculate the abnormal returns one must estimate the normal returns, which would have occurred without the event. For the calculation of the abnormal and normal returns the theory of Mackinlay (1997) is followed. The normal returns for each individual event are predicted in the estimation window ten days prior to the announcement date. Comparing the estimated
normal return with the actual returns results in the abnormal returns for the S&P EURO in the event window. To estimate the normal returns the constant mean return model is used. Also commonly used is the standard market model, but because we make use of market data this model is not applicable. The constant mean return model, which is used to estimate the expected normal returns, looks like this: 𝑅!,! = 𝜇!+ 𝜖!,! With 𝐸 𝜖!,! = 0 And 𝑉𝐴𝑅 𝜖!,! = 𝜎!"!
Where R it is the rate of return of the S&P EURO at date t. The error term 𝜖!,! is assumed to have a zero mean to be independent of 𝜇!. The parameter µ is estimated by the arithmetic mean of the estimation window returns: 𝜇! = 1 𝑀! 𝑅!,! !! !!!!!! Where 𝑀! is the number of non-missing returns over the estimation window. The time frame of an event study as stated by MacKinlay (1997) has the following image: As said the estimation window L1 is used to estimate the normal return. This window has been set at 90 days. It is typical for the estimation window and the event window not to overlap. Including the event in the estimation window could lead to the event returns having large impact on the normal estimated returns. L2 is the event window
and commonly L2 is set larger than the actual estimated event window due to the danger of duplication. L2 is set at 20 days (-10,+10). The abnormal returns are calculated as the difference between the real return and the estimated return 𝜇!. The abnormal return for the S&P EURO at event date t is: 𝐴𝑅!&!",! = 𝑅!&!",!− 𝜇!
Because it could take some time before the market has adopted the news of the ECB, although you want to avoid the risk of other news events affecting the estimate, a two day event window is used (0,+1). Because investors try to have rational expectations the day prior to the event could be impacting as well. We therefor conduct a second event window of five days (-1,+3). An exception is made for event #1 and #3. These events were announced on Monday and Sunday and the Friday return is therefore not relevant. To determine whether the abnormal returns are significantly different from zero, they are all tested. Every abnormal return is provided with a t-statistic 𝑡𝐴𝑅𝑖𝑡. 𝑡𝐴𝑅𝑖𝑡 =𝐴𝑅𝑖𝑡 𝑆𝐴𝑅𝑖 Where 𝑆𝐴𝑅𝑖2 =𝑀1 𝑖− 2 (𝐴𝑅𝑖𝑡) 2 𝑇1 𝑡=𝑇0
The 𝑆𝐴𝑅𝑖2 is calculated from the estimation window, and therefore divided by 𝑀𝑖 with two degrees of freedom.
To estimate if the total abnormal return in the event window is significantly different from zero more or less the same method is used as the one above. First 𝐶𝐴𝑅𝑖(𝑡1, 𝑡2), which is defined as the Cumulative Abnormal Return from 𝑡1 𝑡𝑜 𝑡2 (the event window), is calculated. It is just the sum of the abnormal return in the event window. 𝐶𝐴𝑅𝑖 𝑡1, 𝑡2 = 𝐴𝑅𝑖𝑡 𝑡2 𝑡=𝑡1 The t statistic to see if the CAR is significantly different from zero is the following: 𝑡𝐶𝐴𝑅=𝐶𝐴𝑅𝑖 𝑆𝐶𝐴𝑅
Where 𝑆𝐶𝐴𝑅2 = 𝐿2 𝑆𝐴𝑅𝑖2 L2 is the length of the event window in days. So for the first estimation (0,+1) L2 is 2 and for second estimation (-1,+3) L2 is 5, with exceptions for event #1 and #3 where it is 4 days. The ECB announcements could either have a positive, negative or zero effect on the European stock market. Although one would expect to find similar results as were found in the current literature it could be the world upside down, because investors believe that the announced monetary policy is not capable of resulting in the desired outcome or that the announcement reveals a state of the economy that the market was not aware of.
3.4 Expected inflation on ECB announcements
For this part mostly the same methodology is used. The main difference is the event window which is reduced to one day and 3 days (-1,+1). This is because these inflation swaps are highly liquid and react and adjust more quickly than the asset prices in the S&P EURO index (Moessner, 2015). Because the event windows have become smaller, L2 is 1 and 3 for the one and three day event window. The remainders of the calculations are the same as with the S&P EURO. So for the one day window the t-value is calculated as follows: 𝑡𝐴𝑅𝑖𝑡 = 𝐴𝑅𝑖𝑡 𝑆𝐴𝑅𝑖
For the three day event window the cumulative abnormal return is calculated and divided by the standard deviation of it: 𝑡𝐶𝐴𝑅=𝐶𝐴𝑅𝑆 𝑖 𝐶𝐴𝑅 Where 𝑆𝐶𝐴𝑅= 𝐿2 𝑆𝐴𝑅𝑖2
3.5 Cumulative average abnormal return over all events
Given the results of the abnormal returns on both the S&P EURO and the expected inflation on the ECB QE announcements I examine the average abnormal results of all events across time. To estimate what the average effect was when an ECB announcement took place, I will use the aggregation over time of the events a described by MacKinlay (1997). The abnormal returns can be aggregated using the estimated 𝐴𝑅𝑖𝑡 for each event period. Regarding that you have N events the average abnormal aggregated abnormal returns per period are: 𝐴𝑅𝑡=𝑁1 𝐴𝑅𝑖𝑡 𝑡2 𝑡=𝑡1 With its variance 𝑉𝐴𝑅 𝐴𝑅! = 1 𝑁! 𝜎!"! ! !!! The average abnormal returns can be accumulated over the event window in the same way as with the individual events. The Cumulative Average Abnormal Return (CAAR) is calculated as follows: 𝐶𝐴𝑅𝑖 𝑡1, 𝑡2 = 𝐴𝑅𝑖𝑡 𝑡2 𝑡=𝑡1 Where the VAR is 𝑉𝐴𝑅 𝐶𝐴𝑅! 𝑡!, 𝑡! = 𝑉𝐴𝑅 𝐴𝑅! ! !!! To test the null hypotheses that the abnormal returns are zero per day and across the event window I use the t-value for individual days: 𝑡𝐴𝑅𝑖𝑡 =𝐴𝑅𝑆 𝑖𝑡 𝐴𝑅𝑡 And the following test for the 𝐶𝐴𝐴𝑅 over the event window: 𝑡𝐶𝐴𝑅𝑖𝑡 =𝐶𝐴𝑅𝑖𝑡 𝑆𝐶𝐴𝑅𝑡
The CAAR is both used to test the S&P EURO and the two data sets of expected inflation. The same event windows were used as with the previous tests above.
To recap the methodology that is going to be used, the event study is the base of this research. Abnormal returns will be observed and examined for the S&P EURO, FWISEU55 and the EUSWI5, around the events dates. Finally the CAAR is used to look at the average abnormal returns over all events.
4 Empirical Results
This chapter will describe and analyze the main results obtained through this research. To outline the sequence, first the results of the event study on the effect of the ECB announcements on the S&P EURO stock index is discussed regarding the significance of the cumulative abnormal returns. Second, a closer look is taken on the event study regarding the ECB announcements on the 5-year, 5-year forward inflation. Furthermore the results of the ECB announcements on the 5-year expected inflation are discussed. At last, the overall immediate economic impact of the ECB announcements will be addressed.
4.1 Results of announcement effect on the stock market
Using a constant mean return model approach I investigated the reaction of the stock index on non-standard QE related monetary policy announcements of the ECB from October 2008 through March 2016. The findings and results are given in table 3, which incorporate the raw and abnormal return at the event date from the S&P EURO index. The abnormal return is appended with the t-value to view the significance. Furthermore, the cumulative abnormal returns (CAR) of the two and five day window are added. These are equipped likewise with the t-values. Looking just at the actual returns it should be noticed that seven out of twelve events show positive returns on the event date and the others show negative returns. This displays a diverse short-term (one day) effect on the stock market. Table 3. S&P EURO Stock index returns on ECB monetary policy announcements dates Event Date Raw return at
event date % Abnormal return at event date % Cumulative abnormal return (0,+1) %
Cumulative abnormal return (-1,+3)
October 8, 2008 #1 -6.0515 -5.9283 (-3.06)*** -7.95718 (-2,9049)*** -15,60509% (-4,0283)*** May 7, 2009 #2 -1.0376 -1.1379 (0.496) 0.83178 (0,25638) -0,08007% (-0,0156) May 10, 2010 #3 9.2274 9.1618 (8.330)*** 8.27579 (5,32064)*** 9,63655% (4,3809)*** October 6, 2011 #4 3.1358 3.5158 (1.640)* 4.61718 (1,5188) 11,78112% (2,4510)** November 3, 2011 #5 2.5958 2.8054 (1.141) 0.94145 (0,27076) 3,35465% (0,6102) December 8, 2011 #6 -2.3150 -2.2320 (-0.923) -0,10220 (-0,02987) -3,49967% (-0,6469) August 2, 2012 #7 -2.7187 -2.6983 (1.972)** 1,60243 (0,82798) 4,40110% (1,4382) September 6, 2012 #8 3.1914 2.9816 (2.016)** 3,40266 (1,62716) 3,48556% (1,0542) June 5, 2014 #9 0.8606 0.8332 (0.901) 1,66815 (1,27455) 2,17586% (1,0514) September 4, 2014 #10 1.5870 1.6352 (2.069)** 1,69625 (1,51771) 2,13970% (1,2108) January 22, 2015 #11 1.5975 1.5757 (1.123) 3,25874 (1,64171)* 3,82355% (1,2183) March 10, 2016 #12 -1.4662 -1.1859 (-0.685) 2,43384 (0,99426) 3,64310% (0,9413) Notes: *, **, *** are significant at 10, 5 and 1 % level. As with Lenza et all (2011) looking at the CAR with the five day window in nine out of twelve events the market reacted positively on the QE announcement, but only two of which were significant. The events that resulted in significant cumulative abnormal returns were the one’s on May 5th 2010 and October 6th 2011. On the 5th of
May the ECB introduced the SMP and in October 2011 the second CBPP was introduced next to two one-year LTRO’s. If we look at CAR of January 22nd 2015 we see a positive
effect with significance of 10 percent at the two-day event window. At this day the expended Asset Purchase Program was introduced, which added the Public sector purchase program and apparently the Eurozone reacted positively. The significant negative CAR, which is occurring on October 8th 2008, is most certainly that big because
the statement of introducing Fixed-Rate Full-Allotment the negative effect was one to expect.
If we look back at the single day abnormal returns six were found to have significant abnormal returns, four of which were positive. Next to May 10th 2010 and
October 6th 2011, both September 6th 2012 and September 4th 2014 show significant
positive abnormal returns on the event dates. At in September 6th 2012 the ECB announced a new program of purchasing sovereign debt, and the details about the outright monetary transaction (OMT) were communicated. The significant single day abnormal return event, next to October 8th 2008, was on August 2nd 2012. At this point
mister Draghi hinted that that within the mandate of the ECB to maintain price stability over the medium term it may undertake OMT’s of size adequate to reach his objective. This news could have scared investors that the state of the economy was even worse than they thought, but the following day on August 3rd 2012 the S&P EURO showed a
daily return 4,3%. This adds to the idea that there was no one-way thought about the monetary policy of the ECB. Even within the euro area the reactions were scattered in different parts of the zone as Tempelman (2012) described. 4.2 The effect of the ECB announcements on the expected inflation For the effect on expected inflation the event window was adjusted to one (0) and three (-1,+1) days. Looking at the numbers in table 4 of the effect on the expected 5-year, 5-year forward euro inflation swap the most noticeable is that ten out of twelve real returns at the event date are positive and show positive abnormal returns as well. Four of these event returns were estimated with significance. Event number 2, 3, 10 and 11 were the events when the first CBPP was introduced, the SMP was announced and the launch of the APP in 2014 and the expended version in 2015. Table 4. FWISEU55 returns on ECB monetary policy announcements dates Event Date Raw return at event date % Abnormal return at event date % Cumulative abnormal return (-1,+1) % October 8, 2008 #1 -0,0631 -0,2069 (-0,0935) -0,1345 (-0,03502)
May 7, 2009 #2 2,9994 2,9218 (2,0585)** 1,6679 (0,6784) May 10, 2010 #3 3,3225 3,3744 (2,8456)*** 6,7087 (3,2663)*** October 6, 2011 #4 1,7544 1,8986 (1,0261) 0,77069 (0,2405) November 3, 2011 #5 2,6815 2,9033 (1,5937) 4,6881 (1,4857) December 8, 2011 #6 -0,2179 -0,1135 (-0.0655) -2,4083 (0,8029) August 2, 2012 #7 0,4308 0,5578 (0,3992) 2,2599 (0,9338) September 6, 2012 #8 0,0000 -0,1214 (-0,0932) -0,3642 (-0,1615) June 5, 2014 #9 0,1661 0.2379 (0.3285) 1,9840 (1,5817) September 4, 2014 #10 1,7893 1,8547 (2.5423)** 0,0736 (0,0582) January 22, 2015 #11 4,1979 4,4645 (2,8512)*** 1,4571 (0,5373) March 10, 2016 #12 1,5038 1,8137 (1,5431) 1,9624 (0,9639) Notes: *, **, *** are significant at 10, 5 and 1 % level. If we look at the CAR with the three day event window (-1,+1), there were three events that had a negative CAR, namely event #6 where two 3-year LTRO’s were announced and the interest rate was lowered by 25 basis points. There was only one significant with a three-day window. Event #3 where the SMP was introduced was positively significant at a 1% level. Furthermore, when the details of CBPP2 were announced and the launch of the targeted LTRO’s was stated in event #5 and #9 a reasonable just not significant positive CAR is observed. On the other hand it is also interesting to see that, while the 5-y, 5-y Swap is commonly used as a measure of monetary policy, most announcements did not had an immediate significant effect on the 5 year, 5-year forward expected inflation.
If instead we look at the EUSWI5 that presents the five year expected inflation we observe greater volatility then with the FWISEU55. That is what you would expect, because with actions today you tend to have more influence in the near future in comparison with the time period ten years from now. As shown in table 5, five out of twelve events show a significant effect on the event date, and of these events only the
first event in October 2008 is negative. Similar as with 5-y, 5-y SWAP the events in May 2010 and January 2015 were significant with a 10% level. Considering the CAR with the three days event window (-1, +1) events #11 and #12 both show a significant effect of the ECB announcements. These were the announcement of the CBPP3 in combination with the launch of the APP in September 2014, and the expansion of the APP with the Public Sector PP in January 2015. Table 5. EUSWI5 returns on ECB monetary policy announcements dates Event Date Raw return at event date % Abnormal return at event date % Cumulative abnormal return (-1,+1) % October 8, 2008 #1 -3,7661 -3,6686 (-1,79)* 0,0606 (0,02) May 7, 2009 #2 2,8077 2,9219 (1,08) 2,7553 (0,59) May 10, 2010 #3 7,1429 7,2353 (6,29)*** 2,1223 (1,07) October 6, 2011 #4 0,3546 0,5751 (0,30) 0,9645 (0,0,29) November 3, 2011 #5 3,4890 3,6753 (1,86)* 3,1181 (0,91) December 8, 2011 #6 -1,5112 -1,4108 (-0.75) -0,9258 (-0,28) August 2, 2012 #7 1,6159 1,9691 (0,85) 2,7933 (0,69) September 6, 2012 #8 -0,3261 -0,4459 (-0,19) 0,0317 (0,01) June 5, 2014 #9 1,7964 1,7792 (1,58) 2,6057 (1,34) September 4, 2014 #10 2,3423 2,6843 (1,88)* 4,4995 (1,82)* January 22, 2015 #11 13,0268 14,1500 (2,71)*** 27,8050 (3,08)*** March 10, 2016 #12 0,2656 0,9198 (0,28) 1,3153 (0,23) Notes: *, **, *** are significant at 10, 5 and 1 % level. It is clear that the introduction of the SMP and the APP both had a considerable big impact on the expected inflation.
4.3 Cumulative Average Abnormal Returns over all Events In this section the CAAR will be investigated. The CAAR is used to examine the average effect of a QE related announcement of the ECB. The returns are accumulated from five days before the event date to five days after (-5, +5) over all of the twelve events, and then the average was taken. In table 6 below is shown that regarding the S&P EURO index four days around the event date show significant average abnormal returns. Only the event day it self shows just over the significance level of 10%. The event window of two days regarding all events has an average abnormal return of 1,72% and is significant with a 10% level. Also shown in the table is that both FWISEU55 and EUSWI5 have a significant average abnormal return at a 1% level. Furthermore is the CAAR for EUSWI5 with event window (-1, +1) also significant. Regarding the graphs in figure 1 which are built up from the CAAR of the S&P EURO, the FWISEU55, the EUSWI5 and the average of all three of them, there is a clear positive effect starting two days before the event date and ending one day after. It is also worth noticing that after event day +1 the average abnormal return is close to zero regarding the flattening of the graphs. Table 6. 𝐶𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑣𝑒 𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝐴𝑏𝑛𝑜𝑟𝑚𝑎𝑙 𝑅𝑒𝑡𝑢𝑟𝑛 Notes: *, **, *** are significant at 10, 5 and 1 % level. The empty booths are there because different event windows were used for the variables.
Variables CAAR (0) CAAR (0,
+1) CAAR (-1, +1) S&P EURO 0.77718% (1.53) 1.72241% (1.69)* …. FWISEU55 1.63208% (3.79)*** …. 1.71739% (1.33) EUSWI5 2.53206% (3.48)*** …. 3.92878% (1.80)*
To sum up the above findings, the majority of the abnormal returns are positive with three events standing out. The SMP, APP and APP extended were al all significantly positive for all three variables. Furthermore has the CAAR showed a positive relation between the window (-2, +1) and the average abnormal returns as shown in the graph.
5. Discussion
In the previous section the results were described and provided with some economic theory. To get a better view we should carefully interpret the results obtained above. This section points out the limitations and discusses matters that could be examined more thoroughly.
The first and most important issue to address is that an events study only measures the immediate effect of new information on a specific asset price. With a small event window it is not possible to detect what the effect on a longer-term would look like. A way to cover this issue is to increase the event window. This on the other hand brings the risk along that more economic news and shocks blur the results. It is possible to control for, but controlling will get more difficult the bigger the event window becomes. Another point to address is that some of the events occurred close to each other. If an event takes place in the estimation window of following event it might lead to a bias in the constant mean return model. Regarding the results above it is striking that the SMP, APP and APP extended showed significant results on all three variables. It would be a logical step to investigate these three programs closer. At last, there is a
-2% 0% 2% 4% 6% 8% 10% -5 -4 -3 -2 -1 0 1 2 3 4 5 Da il y CAAR Days, where 0 is the event date.
CAAR
EUSWI5 FWISEU55 S&P EURO Average total CAARserious average downward effect two days before the event dates. More then 80 percent of the abnormal returns two days before the events were found to be negative. I would be interesting to investigate this issue and to take a closer look at how the announcements are being processed by the ECB.
6. Conclusion
In the turbulent economic time the last eight years central banks all over the world do everything in their power to stimulate economic activity. In this research it was examined what the effect of implementing non-standard QE related monetary policy is on the state of the economy. To address this subject the abnormal returns of the Eurozone stock market and the expected inflation were analyzed around the dates where the ECB announced these non-standard QE related monetary policy measures. Twelve events were chosen between October 2008 and March 2016. First the events were investigated separately whether the abnormal returns on the event dates and pre specified event windows were significant. Based on the literature a positive reaction on the S&P EURO and higher growth rates on the two inflation variables were hypothesized. Furthermore there was examined if there was a total average effect of a QE related ECB announcement and if it was significant.
The results show that the investors on the stock market reacted differently to the various events. Five events showed negative abnormal returns of which two were significant. This could suggest that these announcements made investors more aware of the bad situation the economy found itself in. On the other hand, the positive abnormal returns are observed around the events that meet the profile of being QE the best. Looking at the expected inflation, mostly positive abnormal returns on the 5-year expected inflation and more diversified abnormal returns for the 5-year forward, 5 – year expected inflation were found. Three events however showed an overwhelming effect on inflation, namely the announcement of the SMP and the launch and expansion of the APP. That the APP had a positive effect was intended by the ECB because the Eurozone was facing very low rates of inflation at that time. An important matter to address is that all returns became positive as of June 2014. This might be due to the low inflation, which started to have impact at that time. The abnormal returns for the expected inflation are much higher and more significant in comparison with the Stock
market returns. One could say that the ECB managed to increase the expected inflation when it was needed.
The Overall effect, which was displayed in figure 1 showed that especially around the event date the abnormal returns were significantly positive. Therefore you could conclude that the QE related announcements had a positive effect on average on the state of the economy. But is it also clearly visible that the returns flattened out when a few days past. As mentioned in the discussion above is the size of the event window an important problem with regard to the long-term view. Subsequent studies should address this manner more specific.
Bibliografie
Altavilla, C., Giannone, D., & Lenza, M. (2015). The financial and macroeconomic effects of the OMT announcements. ECB Research Bulletin , 22. Baumeister, C., & Benati, L. (2012). Unconventional monetary policy and the great recession: Estimating the macroeconomic effects of a spread compressionat the zero lower bound. Bank of Canada Working Paper , 21. Blyth, M., & Lonergan, E. (2014). Print Less but transfer More: Why Central Banks Should Give Money Directly to the people. Foreign Aff (93), 98-109. Bowdler, C., & Radia, A. (2012). Unconventional monetary policy: the assessment. Oxford Review of Economic Policy , 28 (4), 603-621. Briciu, L., & Lisi, G. (2015). An event-study analysis of ECB balance sheet policies since October 2008. European Economy Economic Brief , 1. Cour-Thimann, P., & Winkler, B. (2012). The ECB’s non-standard monetary policy measures: the role of institutional factors and financial structure. Oxford Review of Economic Policy , 28 (4), 765-803. Draghi, M. (2015, december). Introductory Statement to the press conference. ECB. (2011, July). The ECB's Non-Standard Measures-Impact and Phasing-Out. ECB Monthly Bulletin , 55-69. ECB. (2009). The Eurosystem and the European Central Bank. ECB Bulletin . ECB. (2011). The Monetary Policy of the ECB. European Central Bank, Eurosystem .ECB. (2015, 7). The tranmission of the ECB's rencent non-standard monetary policy measures. ECB Economic Bulletin . Fawley, B. W., & Neely, C. J. (2013). Four Stories of quantitative easing. Federal Reserve Bank of St. Louis Review , 95 (1), 51-88. Ghysels, E., Idier, J., Manganelli, S., & Vergote, O. (2014). A high frquency assessment of the ECB securities markets programme. Working paper series , 1642. Joyce, M., Lasaosa, A., Stevens, I., & Tong, M. (2011). The financial market inpact of quantitative easing in the United Kingdom. International Journal of Central Banking , 7 (3), 113-161. Kapetanios, G., Mumtaz, H., Stevens, I., & Theodoridis, K. (2012). Assessing the Economy-wide Effects of Quantitative Easing. The Economic Journal , 122, F316-F347. Lenza, M., Pill, H., & Reichlin, L. (2010). Monetary Policy in exceptional times. Economic Policy , 25, 295-339. MacKinlay, A. C. (1997). Event Studies in Economics and Finance. Journal of Economic Literature , 35, 13-39. Martin, C., & Milas, C. (2012). Quantitative Easing: a sceptical survey. Oxford Review of Economic Policy , 28 (4), 750-764. Moessner, R. (2015). Effects of ECB balance sheet policy announcements on inflation expectations . Applied Economics Letter , 22 (6), 483-487. Neely, C. (2014). How Persistent Are Unconventional Monetary Policy Effects? . Research Division Federal Reserve Bank of St Louis . Pisani-Ferry, J., & Wolff, G. (2012). Is LTRO QE in disguise? CEPR's Policy Portal . Putnam, B. (2013). Essential concepts necessary to consider when evaluating the efficacy of quantitative easing . Review of Financial Economics , 22, 1-7. Rodriguez, C., & Carrasco, C. A. (2014). ECB Policy Responses between 2007 and 2014: a chronological analysis and a money quantity assessment of their effects. No. workingpaper65 . Tempelman, J. (2012, July/August). Against Quantitative Easing by the European Central Bank. Financial Analysts Journal , 4-6. Ugai, H. (2007). Effects of the Quantitative Easing Policy: A survey of empirical analyses. Monetary and Economic Studies, Bank of Japan , 25 (1), 1-48.
