What is the effect of the ECB’s Asset Purchase Programme on bank stock returns?

Merijn van der Sluijs 10298312

Track: Economics

Field:

Economics and Finance

Title:

What is the effect of the ECB’s Asset Purchase Programme on bank stock returns?

Name of supervisor: Simas Kucinskas

Statement of Originality

This document is written by Student Merijn van der Sluijs who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are 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.

Introduction

The Asset Purchase Programme (APP), the European Central Bank’s quantitative easing programme, has been ongoing for the past 4 years and is set to continue until at least September this year. It is an unconventional policy tool that has accounted for a total of more than two trillion Euros of

purchases by the European Central Bank and its subsidiary National Central Banks. The programme was launched in a macroeconomic environment where there was a threat of deflation and

conventional monetary tools were exhausted, since the policy rate could not be set lower; the zero lower bound was reached. The goal of the APP was to get the economy of the Euro area away from the threat of deflation and back to the European Central Bank’s policy target of inflation below but close to 2% in the medium term.

The programme started in October 2014 with the Asset Backed Securities Purchase

Programme (ABSPP) and the third Covered Bonds Purchase Programme (CBPP3), programmes that accounted for a total of 10 billion of purchases to be conducted each month. Fairly soon however, in March 2015, the programme was massively expanded because the ABSPP amd CBPP3 combined were judged not to be effective enough by the European Central Bank. The Public Securities Purchase Programme (PSPP) was launched. This programme accounted for another 40 billion of monthly purchases to be made. The total amount of monthly purchases was increased to 60 billion. This Public Securities Purchase Programme entailed the large scale buying of sovereign bonds by the National Central Bank subsidiaries of the European Central Bank. The restrictions on the programme were however fairly tight and this meant it would soon run out of available eligible bonds available. The response to this was to loosen these restrictions, allowing the programme to run for a longer period of time, which is exactly what was done.

The APP was extended to run until at least March 2017 as opposed to its original end date of September 2016, then it was extended until the end of 2017 and then again to the current “end date” of September this year. Parentheses are not out of place in this instance. Not only because the programme already has been extended three times. The European Central Bank has communicated its original end date of September 2016, and every end date after it, with the big caveat that the programme would run until the announced end date “or beyond, if necessary, and in any case until the Governing Council [of the ECB] sees a sustained adjustment in the path of inflation consistent with its inflation aim [of 2%]”.

Since the APP will be going on for the near future and the European Central Bank until now has never named any solid end date, it makes sense for economists to look at what the effects of the ECB’s QE programme are. This is exactly what has been done. For the relatively short time that the APP has run, there has been a lot of discussion and a lot published on the subject.

Demertzis and Wolff (2016) have summarised the discussion as focussing on four aspects: the effectiveness of the programme in achieving the ECB’s inflation target, the duration of the programme, the way it might dispossess savers, and the impact it has on bank profitability. While the results of the other discussion aspects of the discussion will be mentioned for completeness’ sake, the aspect of the discussion that’s relevant for this paper is the APP’s impact on bank profitability.

The APP impacts bank profitability in three ways. Firstly it improves the macroeconomic situation in the Euro area by contributing to the ECB’s inflation target. An improved macroeconomic situation is positive for banks’ stock prices, as it is for each company’s. More directly however, the

APP impacts banks by creating scarcity on the market for sovereign bonds and by lowering and flattening the yield curve.

The scarcity of sovereign bonds, created by the APP, drives op their price, which strengthens banks’ balance sheets. The flattening and lowering of yield curves could have a twofold effect on bank stock prices. The inverse relation between yields and prices means that a lowering of the yield curve drives up stock prices. However banks’ asset returns are more related to long term yields than its liability returns, this means that a flattening of the yield curve reduces bank profitability. This reduction in profitability in drives stock prices down.

This paper uses an event study methodology to estimate the effect of the APP on bank stock prices and seven other sectors, who are used as material for comparison. The sectors used for comparison are “auto and parts”, “banks”, “basic resources”, “financial services”, “food and beverages”, “healthcare”, “industrial” and “ technology”.

There were no results found across industries and the overall response to APP

announcements according to the model in this paper was quite low. However the banking sector did respond the most to APP announcements, with two out of five identified announcements yielding a significant response by the banking sector: the announcement that the amount of monthly

purchases would be increased from 60 to 80 billion, which yielded a significantly positive response, and the announcement that the APP would be extended to September 2018 but also downscaled to 30 billion of purchases per month. This announcement yielded a significantly negative response.

Although the overall effect of the APP remains positive, the significantly negative response to announcement in October last year about the downscaling of the APP, might mean that the macroeconomists interviewed back in 2016 might still be right: the APP should not end for the time being.

Literature review.

The Asset Purchase Programme (APP) of the European Central Bank (ECB) has its roots in the famous “whatever it takes” speech made by ECB president Mario Draghi on the 26th _{of July 2012, in which he} stated that the European Central Bank would use any and all methods available to save the euro “and it will be enough” (European Central Bank, 2012). Within this mandate the European Central Bank announced the Asset Backed Securities Purchase Programe (ABSPP) and the third Covered Bond Purchase Programme (CBPP3) in September 2014, which were officially declared October 19th that year. These programmes accounted for monthly purchases of 10 billion euro’s combined. On March 4th _{2015 the European Central Bank officially confirmed the Public Securities Purchase} Programme (PSPP), which was announced earlier that year on January 22nd_{. The PSPP was a massive} expansion to the APP that would push monthly purchases to 60 billion dollars (European Central Bank, 2015a and European Central Bank, 2015b). The PSPP would expand the horizon of the APP to also include sovereign bonds, securities of supranational organisations and European institutions. However, the restrictions posed on the sovereign and (supra)national agencies’ bonds that would be eligible for purchase under the PSPP would mean the programme would soon reach its limits (Claeys, Leandro & Mandra 2015).

Therefore the PSPP was expanded. There was an expansion of the list of national agencies whose securities were eligible for the PSPP, the issue(r) share limit was expanded and regional and local governments’ assets were made eligible. Principal investments from maturing securities purchased under the APP would be reinvested. Also was the end-date postponed from September 2016 to March 2017 “and beyond if necessary” (Claeys & Leandro, 2016 and European Central Bank, 2015c and European Central Bank, 2015d). This date was then again postponed until end 2017 (European Central Bank, 2016c and European Central Bank, 2017a) and again until September 2018 (European Central Bank, 2017b).

The APP was not just extended however, even with monthly purchases at 60 bn. and

additional bonds made eligible to postpone the ultimate ending date of the programme, this was not judged as enough to reach the inflation target by the European Central Bank. In April 2016 the monthly amount purchased was also increased from 60 to 80 billion (European Central Bank, 2016b), which was announced a month earlier (European Central Bank, 2016a). After a year this was scaled back to 60 billion of purchases per month (European Central Bank, 2016c and European Central Bank, 2017a).

On October 26th _{2017 the ECB officially announced it would decrease monthly purchases to} 30 billion each month at the start of 2018. This situation will continue until September this year “or beyond, if necessary, and in any case until the Governing Council [of the ECB] sees a sustained adjustment in the path of inflation consistent with its inflation aim [of 2%]”. And even now the ECB is ready to return monthly purchases to 60 billion if they judge the path of inflation inconsistent with a return to the inflation aim of an inflation rate close to, but below 2%. The gain from maturing securities will also be reinvested, which will continue for “an extended period of time” after the end of net asset purchases and as long as necessary (European Central Bank, 2017b).

With a market intervention of such durance and magnitude, a large amount of discussion can be expected. The discussion regarding the APP has mostly focused on four aspects. The first one is whether the APP has actually contributed to getting inflation closer to the ECB’s target. About this there is a fair amount of consensus that it did, with several studies concluding it had positive effects on both investment and consumption (Demetzis & Wolff, 2016).

The second point of discussion is when the APP should end. In 2016 the German Council of Economic Advisors argued that the ECB should taper the programme, however 77% of the

macroeconomists interviewed by the CFM thought the APP should continue as it was. Demertzis and Wolff (2016) agreed with this view because the HICP at that time still remained way below the inflation goal of 2%. However since then the HICP has increased from 0.6% to 1.1% in November 2017, so their position might have changed since then.

A third aspect of the discussion regarding the APP is that it would dispossess savers by lowering and flattening yield curves, thereby denying them interest income. However this argument is judged to be mostly political by Demertzis and Wolff (2016), since monetary policy will always have an impact on investors and savers. One could even say that the whole point of any QE programme is to “dispossess savers”, since the goal is to discourage saving and encourage investment. The last aspect of the debate is the effect of the APP on bank profitability. This is of course the most interesting aspect from the view of this thesis.

The APP is meant to improve macroeconomic conditions in the euro area, this should also have a positive effect on the banking sector. By increasing faith in the economy and by extension the banking sector, stock prices should go up. APP should influence banks more directly however. Firstly, demand for sovereign bonds by the ECB under the PSPP creates a “scarcity effect”. The bonds are less available on the common market, which drives up the price. Banks who hold a sufficiently big amount of these sovereign bonds experience noticeable capital gain because of this (Demertzis & Wolff, 2016). This strengthens the banks’ balance sheet and might therefore increase stock prices, since the bank’s stocks have become a more stable investment. However the APP also lowers long term yields in Europe and flattens the yield curve (Cova, Pagano & Pisani, 2015), as it has done in the USA (Montecino & Epstein, 2014). Because returns on assets are more related to long term yields than returns on liabilities, net interest income will decrease (Borio, Gambacorta and Hofmann, 2017). The flattening and lowering of yield curves could have a twofold effect on bank stock prices. The inverse relation between yields and prices means that a lowering of the yield curve drives up stock prices. However banks’ asset returns are more related to long term yields than its liability returns, this means that a flattening of the yield curve reduces bank profitability. This reduction in profitability in drives stock prices down.

Urbschat and Watzka (2017) have done an event study on the APP in the Euro Area to check the effect of the programme on Euro area yield curves. They also found a lowering of the yield curve. While this paper will focus on stock returns, their paper resembles this one most closely. Apart from using a model to estimate parameters that resembles theirs, I will use their methodology for

identifying significant dates regarding the APP.

Because a press release by the ECB might have been preceded by a lot of speculation, or because a journalist might have gotten a scoop beforehand about the contents of the press release, etc., a press release might not contain any real news value. To check for this, Urbschat and Watzka (2017) looked at whether the contents of an ECB press release were featured on page one to three of the Financial Times the next day. This means the contents of the press release were news worthy and thus the date of the press release and the next day were a significant event in their event study on the APP. Any significant news about the APP was assumed to potentially have a significant effect on yield curves. The same logic I apply so stock returns in this paper, so I will follow this

methodology for identifying significant events regarding the APP.

This paper will however not focus on yield curves, as most of the existing literature has done. This paper will look at the effect of the APP on stock prices, defined as daily return. This could

give additional weight to the existing literature since stock prices and yields are well known to be inversely related. One would expect to see a positive effect of the APP on stock prices based on the existing literature since a lowering and flattening of yield curve should translate to an increase in stock price. This paper will also add a clear opportunity to compare the reaction of the stock price index of the banking sector to that of other sectors. This is also something the existing literature has not supplied a clear opportunity for.

Methodology.

As mentioned in the literature review above, the methodology for identifying significant events regarding APP is taken from Urbschat and Watzka (2014). A significant event is defined as a day on which the ECB published a press release or press conference in which new information regarding the APP was published, followed by one or more articles in the Financial Times no further back than page three the following day. I have used the starting point Urbschat and Watzka (2014) have used; the announcement of PSPP. The two significant announcement dates they distinguished after the announcement of PSPP I have logically also used. After the end of the scope of their paper I have identified two more announcements that (should) fit the criteria they listed.

As for the event window I will use the width used by Urbschat and Watzka (2017) for the Euro area, Gagnon, Raskin, Remache and Sack (2010) for the USA and Joyce, Lasaosa, Stevens and Tong (2011) for the UK, in their QE related event studies. This is two days; from closing the day prior to the announcement until closing the day after the announcement. Because I use daily data, this means two data points per event.

They used a two day event window because of the trade-off that every event window in an event study has. On the one side an event window should not be too narrow. Even with the

instantaneousness with which financial markets work in theory, in reality one has to leave time in an event window for the market to react to the event, in this case an APP announcement. On the other hand the event window should not be too wide. Since it is very hard to control for every other variable that enters the creation of stock prices, if an event window is too wide, this lessens the certainty that an effect observed in the event window is just a result of the influencing factor that you’re interested in.

In all three papers the sweet spot in this trade-off was judged to be at an event window width of two days. To fit in best with existing literature and because the judgement of these ten economists seems sound to me, I will also use an event window width of two days.

Event definitions: Table 1: Event definitions

Event Dates

Announcement of PSPP 22-01-2015* – 23-01-2015

Announcement of extension to march 2017 03-12-2015* – 04-12-2015 Announcement of expansion of monthly

purchases from 60 to 80 bn. Starting 04-2016.

10-03-2016* – 11-03-2016

Announcement of extension to december 2017 and downscaling to 60bn. Starting 04-2017.

08-12-2016** – 09-12-2016

Announcement of extension to september 2018 and downscaling to 30bn. Starting 01-2018

26-10-2017*** – 27-10-2017

*identified by Urbschat and Watzka (2017) as significant announcement date. **significant announcement date by Urbschat and Watzka’s (2017) standards.

***Not confirmed to fulfill Urbschat and Watzka’s (2017) criteria, but based on a total of 10

published articles by the Financial Times based about this press release, it is assumed to fullfill their criteria.

Data selection criteria:

A selection of sectors listed in the Euro Stoxx 600 index, retrievable from datastream, for which complete optimised daily data is available from 19-11-2014 until 29-12-20171_{. Sectors were chosen} based on the overall level of the index relating to the Euro Stoxx 600 and the banking sector, assumed correlation with monetary policy of the ECB and the banking sector and and/or volatility. I also aimed to make the sectors cumulatively representative for the economy as a whole. Of all sectors the optimised indices were used because for the banking sector the level of the optimised index was closest to the level of a manual summation of the banks Stoxx 600 index. The eight sectors chosen are “auto and parts”, “banks”, “basic resources”, “financial services”, “food and beverages”, “healthcare”, “industrial” and “ technology”.

Of these sectors financial services is judged to be highly correlated with monetary policy and the banking sector. Healthcare and food and beverages are assumed to not be strongly correlated with monetary policy or the business cycle in general. Basic resources, auto and parts, industrial and technology are judged to not be exceptionally susceptible to specific monetary policy decisions but susceptible to the business cycle. Therefore the set of 8 sectors give a representative image of the economy as a whole. Also the availability of index data from datastream that are subsets of the Euro Stoxx 600 is not a lot larger than this collection of subsectors.

The starting point of 19-11-2014 was used because that is the official starting point of the APP, albeit at a far smaller scale than the expanded APP that started when PSPP started. Most

literature focusses on the PSPP, so using that as my first event makes this paper more in line with existing literature.

Estimation method:

The effect of the earlier defined announcements will be represented by the following model: 𝑅𝑖 = 𝛼 + 𝛽1𝑅𝑚𝑟𝑘𝑡+ 𝛽2𝑃𝑆𝑃𝑃 + 𝛽3𝐸𝑋𝑇𝐸𝑁𝐷 + 𝛽4𝐸𝑋𝑃𝐴𝑁𝐷 + 𝛽5𝐸𝑋𝑇𝐷𝑆𝐶1 + 𝛽6𝐸𝑋𝑇𝐷𝑆𝐶2 + 𝜀𝑖

R = daily stock returns.

i = indicator of sector of interest.

mrkt = market proxy, which is the Euro Stoxx 600 index in this paper.

PSPP = a dummy variable that takes value 1 on 22-01-2015 and 23-01-2015 and zero otherwise. EXTEND = a dummy variable that takes value 1 on 03-12-2015 and 04-12-2015 and zero otherwise. EXPAND = a dummy variable that takes value 1 on 10-03-2016 and 11-03-2016 and zero otherwise EXTDSC1 = a dummy variable that takes value 1 on 08-12-2016 and 09-12-2016 and zero otherwise. EXTDSC2 = a dummy variable that takes value 1 on 26-10-2017 and 27-10-2017 and zero otherwise. The model will be estimated via OLS to make the estimated model:

𝑅̂𝑖 = 𝛼̂ + 𝛽̂1𝑅𝑚𝑟𝑘𝑡+ 𝛽̂2𝑃𝑆𝑃𝑃 + 𝛽̂3𝐸𝑋𝑇𝐸𝑁𝐷 + 𝛽̂4𝐸𝑋𝑃𝐴𝑁𝐷 + 𝛽̂5𝐸𝑋𝑇𝐷𝑆𝐶1 + 𝛽̂6𝐸𝑋𝑇𝐷𝑆𝐶2

The OLS relation to the market return is used as a control variable. If an event dummy parameter is significant despite the already overall estimated relationship of a sector to the market, it means an announcement caused a significant deviation of the sector’s return from the normal return given the market return. Because it is only used as a control variable and is not a parameter of interest, it is not listed among the results in the following sector. The regression itself is performed in stata.

Because each event consists of two datapoints, the estimated effect of an announcement is 2𝛽𝑖 for each dummy variable.

Since each return is relative and therefore not depended on the actual nominal change in stock price, the total estimated effect for a sector equals ∑6𝑖=1𝛽̂𝑖 if all parameters are taken into

account. I will only look at the parameters that test significantly at a 5% level. Hypotheses:

H0: 𝛽̂1= 0; 𝛽̂2= 0; 𝛽̂3= 0; 𝛽̂4= 0; 𝛽̂5= 0; 𝛽̂6 = 0

H1: 𝛽̂1≠ 0 or 𝛽̂2≠ 0 or 𝛽̂3≠ 0 or 𝛽̂4≠ 0 or 𝛽̂5≠ 0 or 𝛽̂6≠ 0

Or, to put it in to words: my null hypothesis is that the estimated coefficients are not significantly different from zero. My alternative hypothesis is that an estimated coefficient is significantly different from zero.

Empirical results:

Using the estimation method described above the following parameters were estimated. Table 2: Empirical results

Sector/Dummy PSPP Extend Expand Extend & Downscale 1

Extend & Downscale 2 Auto & Parts 0.0283259 -0.0143161 -0.0000167 0.0089643 0.0125418 Banks -0.0098889 0.0098671 0.0179096 -0.0057923 -0.0112137 Basic Resources -0.0167209 -0.006042 -0.010038 -0.0075144 -0.0063537 Financial Services 0.0032722 0.0015725 -0.0007282 -0.004588 -0.0023835 Food & Beverages 0.0070201 -0.0040032 -0.0032506 0.0079823 0.0032584 Healthcare -0.0022203 -0.0044861 -0.0039134 0.0099464 -0.0024462 Industrial 0.0006085 0.0012239 -0.003659 -0.0035147 0.0039616 Technology 0.0022231 0.0022965 0.0000361 -0.0019809 -0.0060743 Significant at α= 5% Interpretation of results.

A few things stand out to my when I look at the empirical results. Firstly it’s striking the financial services index does not have any significant response to APP announcements in this model. The first explanation I thought of for this lack of significance is that the financial markets are known to be very volatile and speculative. Therefore any actual announcement by the ECB and any article in the Financial Times might have been preceded by so much speculation about its potential contents that the effect of the actual announcement doesn’t show up anymore. This is however not supported by the data; the standard errors are actually lower for the financial services sector than the banking sector2_{. The explanation the data seems to support therefor is the one that the financial services} market simply has not reacted significantly to APP announcements. Another explanation might be that the financial markets have predicted the contents of each press release correctly and therefore didn’t respond to the announcements.

The inverse of the first striking result is the result for the food and beverages sector. I would have expected no significant results from APP announcements on this sector, since the basic needs industry is supposed to have very low volatility or susceptibility to the business cycle. As can be found in Appendix A, the standard errors for this sector are indeed lower than that of the banking sector. However the only significant parameter outside of the banking sector is the response of the

food and beverages sector to the announcement of the extending and downscaling of APP on December 8th _{2016. The fact that this occurred after the first announcement of extending but} downscaling the APP could be because this sector is not really susceptible to the direct effects of the APP but the first announcement of downscaling the APP gave a clear signal that the economic situation in the Euro area had improved in the eyes of the ECB, this would then be what the stock prices in the food and beverages sector responded to.

As can be seen in table 1, the banking sector is the most responsive sector to APP

announcements of the sectors investigated. Two of the five parameters tested are significant. Both the significant parameters have the sign one would expect. The expansion of the APP from 60 to 80bn. of purchases each month had a positive effect on bank stock prices. The minor downscaling from 80 back to 60bn. of purchases had a negative but not significant effect. The major downscaling from 60 to 30bn. of purchases however did have a significantly negative effect. The total effect of APP announcements on bank stock returns is a gain of 1.34% approximately3_.

It is however notable that only two of five parameters are significant. One might have expected more or even all of the announcements to have a significant effect. However of 40

estimated parameters only three are significant, so the overall responsiveness of the stock indices to APP announcements has not been high among the sectors investigated. Six sectors don’t have any significant parameters at all. In that light the fact that two out of three significant parameters belong to the banking sector means that the banking sector has been by far the most responsive to APP announcements and relatively responsive compared to the overall dataset.

Conclusion.

As measured via the model, data and events in this paper, there is a significant effect of two out of five identified announcements on APP on the banking sector stocks index returns. The

announcement about the expansion from 60 to 80b. of monthly purchases made on March 10th ₂₀₁₆ had a significant positive effect on bank stock returns of approximately 3.6%. This might be an indicator of the macroeconomic channel. The announcement about the extension of the APP to September 2018 but also its downscaling from 60 to 30bn. of monthly purchases had a negative effect on bank stock returns of approximately 2.2%. The other announcements had no significant effect.

The fact that the negative response of the banking sector to the major downscaling of the APP offset most of the positive effect the expansion of it had, suggests that the popular opinion among economists back in 2016 that the APP should not be tapered, might still be true. For investors in the banking sector the signalling by the ECB that the economy is in a better state and doesn’t need the APP as much anymore is clearly still outweighed by the fact that monetary support for the economy is lessening.

The overall effects of the APP announcements were quite small. The only other significant response was by the food and beverages sector to the announcement about the extension of the

APP to December 2017 and its downscaling from 80 to 60 bn. of monthly purchases, made on December 8th _2016.

I expected the effect of ECB’s announcements on the APP to be more widespread across sectors and for a majority of the announcements to be significant for the banking sector. Since there is broad consensus within the existing literature that QE programmes lowered and flattened yield curves in both the USA, UK and Euro area and a widely known economic rule that yields and stock prices are inversely related, I expected a more pronounced response from bank stock returns to APP announcements.

However when one looks at the overall responsiveness to APP announcements, the banking sector did respond strongly relatively speaking. Six out of eight investigated sectors have no

significant effects at all and the banking sector has two. A different kind of regression, more

explanatory or controlling variables or a completely different kind of model (VAR for example) might shed more light on the matter, but that is beyond the scope of this paper

Reference list.

Borio, C., Gambacorta, L., & Hofmann, B. (2017). The influence of monetary policy on bank profitability. International Finance, 20(1), 48-63. doi:10.1111/infi.12104

Claeys, G., & Leandro, Á. (2016). The European central bank’s quantitative easing programme: Limits and risks. Bruegel Policy Contribution, 2016/04. Retrieved from

http://bruegel.org/wp-content/uploads/2016/02/pc_2016_04.pdf

Claeys, G., Lenadro, Á., & Mandra, A. (2015). European central bank quantitative easing: The detailed manual. Bruegel Policy Contribution, 2015/02. Retrieved from

https://www.econstor.eu/handle/10419/126687

Cova, P., Pagano, P., & Pisani, M. (2015). Domestic and international macroeconomic effects of the eurosystem expanded asset purchase programme. Temi di Discussione, 1036. Retrieved from http://www.bancaditalia.it/pubblicazioni/temi-discussione/2015/2015-1036/en_tema_1036.pdf Demertzis, M., & Wolff, G. B. (2016). What impactdoes the ECB’s quantitative easing policy have on bank profitability? Bruegel Policy Contribution, 2016(20). Retrieved from http://bruegel.org/wp-content/uploads/2016/11/pc-20-16-4.pdf

European Central Bank. (2012). Verbatim of the remarks made by Mario Draghi. [Press release]. Retrieved from https://www.ecb.europa.eu/press/key/date/2012/html/sp120726.en.html European Central Bank. (2015a). ECB announces expanded asset purchase programme. [Press

release]. Retrieved from https://www.ecb.europa.eu/press/pr/date/2015/html/pr150122_1.en.html European Central Bank. (2015b). Decision (EU) 2015/774 of the European Central Bank of 4 March

2015 on a secondary markets public sector asset purchase programme (ECB/2015/10). Retrieved

from https://www.ecb.europa.eu/ecb/legal/pdf/oj_jol_2015_121_r_0007_en_txt.pdf

European Central Bank. (2015c). Introductory statement to the press conference (with Q&A). [Press release]. Retrieved from https://www.ecb.europa.eu/press/pressconf/2015/html/is151203.en.html European Central Bank. (2015d). Decision (EU) 2015/2464 of the European Central Bank of 16

December 2015 amending decision (EU) 2015/774 on a secondary markets public sector asset purchase programme (ECB/2015/48). Retrieved from

https://www.ecb.europa.eu/ecb/legal/pdf/oj_jol_2015_344_r_0001_en_txt.pdf

European Central Bank. (2016a). ECB adds corporate sector purchase programme (CSPP) to the asset

purchase programme (APP) and announces changes to APP. [Press release]. Retrieved from

https://www.ecb.europa.eu/press/pr/date/2016/html/pr160310_2.en.html

European Central Bank. (2016b). Decision (EU) 2016/702 of the European Central Bank of 18 April

2016 amending decision (EU) 2015/774 on a secondary markets public sector asset purchase programme (ECB/2016/8). Retrieved from

http://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32016D0008

European Central Bank. (2016c). Monetary policy decisions. [Press release]. Retrieved from https://www.ecb.europa.eu/press/pr/date/2016/html/pr161208.en.html

European Central Bank. (2017a). Decision (EU) 2017/100 of the European Central Bank of 11 January

2017 amending decision (EU) 2015/774 on a secondary markets public sector asset purchase programme (ECB/2017/1). Retrieved from

https://www.ecb.europa.eu/ecb/legal/pdf/celex_32017d0001_en_txt.pdf

European Central Bank. (2017b). Monetary policy decisions. [Press release]. Retrieved from https://www.ecb.europa.eu/press/pr/date/2017/html/ecb.mp171026.en.html

Gagnon, J., Raskin, M., Remache, J., & Sack, B. (2010). Large-scale asset purchases by the Federal Reserve: Did they work? Staff Repor, Federal Reserve Bank of New York, 441. Retrieved from https://www.econstor.eu/bitstream/10419/60927/1/622939319.pdf

Joyce, M.A.S., Lasaosa, A., Stevens, I., & Tong, M. (2011). The financial impact of quantitative easing in the United Kingdom. International Journal of Central Banking, 7(3). 113 – 161. Retrieved from http://www.ijcb.org/journal/ijcb11q3a5.pdf

Montecino, J.A., & Epstein, G. (2014). Have large scale asset purchases increased bank profits?

Political Economy Research Institute, 372. Retrieved from

http://www.peri.umass.edu/media/k2/attachments/WP372.pdf

Urbschat, F., & Watzka, S. (2017). Quantitative easing in the Euro area: An event study approach.

CESifo Working Paper, 6709. Retrieved from