Monetary policy announcements and bank stock return.
Jan Willem van Arkel
Bachelor thesis Finance & Organization 10835024
Table of contents
1. Abstract 2. Introduction
3. Past literature and testable hypotheses 4. Empirical Analysis 4.1 Method 4.2 Sample 4.3 Model 4.4 Data 5. Results 5.1 Event study 5.2 All sample period
5.3 Sovereign debt crisis period 5.4 Post-crisis period
6. Conclusions and research limitations 7. References
Statement of Originality.
This document is written by Jan Willem van Arkel, 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. Abstract
Banks are interest rate sensitive and are therefore affected by central bank policies. Since the financial crisis several studies have investigated the effects of monetary policy decisions on the stock prices of banks. Especially the effects of non-standard policy measures are interesting in this context since they were created specifically for countering the effects of the financial crisis. This paper investigates the effect of a variety of
monetary policy measures on the stock price of large European banks for the period of the sovereign debt crisis and the post-crisis period. This is done by conducting an event study to measure cumulated abnormal returns around policy announcements over the period January 2010 – January 2016. Results show that banks are more sensitive to monetary policy shocks than what would be expected with the normal market model. Banks are also more sensitive to non-standard policy measures than conventional policy measures and are affected the most by contractionary measures. Results show heterogeneity in reaction for different periods in time. In addition, cumulated abnormal returns show a large variance during both the sovereign debt crisis and the post-crisis period which implies heterogeneity in individual banks responses to monetary policy announcements.
2. Introduction
Since in 2007 the financial crisis hit markets and economies globally, much attention has been paid to the role of banks, governments and regulatory institutions in creating and solving the existing problems. According to Fiordelisi et al., (2014), monetary policy interventions of central banks around the world played a very important role in stabilizing markets, especially financial markets. Among those central banks was the European Central Bank, which set interest rates to never before seen low levels and conducted new types of policy interventions; nonconventional interventions like monetary easing by asset purchasing programmes (Ricci, 2015). These interventions were partially designed to secure liquidity and stimulate banks to lend to private sectors (Galloppo et al., 2014).
There is no clear-cut answer to the question how specific policy measures affect bank stock returns. This is because the effect is dependent on economic circumstances, economic mechanisms and individual characteristics and risk profile of banks (Altunbas, 2010).
Several studies have been published looking at the effect of how monetary policy affects bank stock prices. Literature from before the crisis suggest heterogeneous response from banks to monetary policy in the United States and the European union. The most important factor concerning the effect of expansionary or contractionary measures on
individual banks is the liquidity of these banks. The size of the bank does, unlike in the U.S., not matter when concerning reaction to policy shocks (Ehrmann et al., 2002).
The effects of monetary policy on bank’s stock returns during the financial crisis is researched by Fiordelisi et al. (2014). However, this was analysed for the U.S. market. The research of Ricci (2015) is comparable with that of Fiordelisi, but focuses the empirical research of bank’s stock reaction to policy shocks on the European Union. What is also interesting about her research is the fact that it includes the effect of both standard and non-standard policy measures from the European Central Bank. This was not analysed before.
Her research builds hypotheses on findings of Yin and Yang (2013), who found that banks have a very heterogeneous response to monetary policy. When this was further
analysed in her research she found that banks that have weak balance sheets and have higher risk are more sensitive to policy interventions. She also found that higher proportions of deposit funding in the bank’s capital structure leads to less sensitivity to monetary policy shocks. Furthermore, in countries where banking is seen as more risky it is found that banks profit more from expansionary measures. Systematic risk of countries may be seen as an important factor of heterogeneity among bank stock prices reaction to policy shocks.
In addition to similar conclusions from Fiordelisi et al (2014), she found that bank stock returns are more sensitive to non-standard policy interventions. Interest rate decisions were not as effective. The strongest (negative) effect was found in the occurrence of the stopping of liquidity provisions and asset purchasing programmes. However, these findings are heterogeneous in different stages of the financial crisis. Since these findings concern the financial crisis period it is unclear if the same effects hold in the longer run.
This paper will study the effect of monetary policy on the stock prices of large
European banks in the period 2010-2016. It will add to previous research by investigating the effect of policy changes on bank stocks in the period after the financial crisis, as well as comparing these results with effects found on the period in the sovereign debt crisis period (2010-2013), as is partly investigated by Ricci (2015).
How banks react to monetary policy decisions may be useful for portfolio
Considering the previous literature, difference in sensitivity between contractionary and expansionary measures will be researched.
How do sovereign debt crisis and post-crisis ECB monetary policy announcements affect European bank stocks?
Are banks in post-crisis period more sensitive to contractionary policy shocks than to expansionary policy shocks?
To answer the questions considering how policy announcements affect bank stock prices and if contractionary policy measures have more effect than expansionary measures, an event study will be run to estimate cumulated abnormal returns around announcement dates.
3. Past literature and testable hypotheses
Bank equity value is dependent on common and bank specific factors (Molyneux, 2010). One very important factor is monetary policy, because bank assets and liabilities are interest rate sensitive (Yin and Yang, 2013). Interest rates affect loan demand, availability of guarantees, valuing securities portfolios and the expected rate of return. Literature concerning asset pricing and interest rates mostly focuses on non-financial equities, while bank equity is more interest rate sensitive (Kim et al., 2013). A policy intervention by the Central bank will therefore directly affect a bank’s assets and liabilities.
Monetary policy seems to have more effect on highly leveraged firms (Angoli et al., 2003). Banks are interest rate sensitive and often highly leveraged, therefore it is expected that policy announcements affects them significantly (Haitsma et al., 2016). Chuliá et al (2010) found that of a variety of industries, financial institutions’ stocks are affected the most when the Federal Reserve unexpectedly announces policy.
Yin and yang (2013) found that banks that are more reliant on non-deposit funding are more sensitive to monetary shocks, but that banking activity mix does not affect the bank’s sensitivity to monetary shocks. Ricci (2015) also found that banks that are more dependent on non-deposit funding are more sensitive to policy shocks.
Most of the previous research done on asset pricing and monetary policy considers U.S. firms and the effect of monetary policy on the New York Stock Exchange. There are a
few studies that focus on monetary policy changes and bank equity pricing as is discussed in the introduction. Literature from before the crisis concerning the effect of policy measures on bank stock prices suggests that this effect is dependent on the liquidity of individual banks. It also suggests that effects on European and U.S. based banks differ from each other (Ehrmann et al., 2002).
Research done on the period of the crisis from Ricci (2015) concerned European banks and European Central Bank policy, while most of the literature focuses on the U.S. markets, for example Fiordelisi et al., (2014).
Ricci (2015), continued expanding existing literature considering this subject on especially two fronts. For the first time empirical analysis was done on the effects of non-conventional monetary policy interventions on the stock price of European banks. Second, she analysed the effect of a variety of bank characteristics on sensitivity to monetary policy
changes. This is important because bank equity prices are not solely relevant for investors, but the effect of monetary policy changes on these equity prices may be seen as a test of the effects of monetary policy decisions (Yin et al., 2010).
Research from Fiordelisi et al, (2014) included nonconventional measures but didn’t analyse the effects considering individual stock returns. However, he concluded that the strongest positive reaction on stock prices changes comes from nonconventional policies like monetary easing programs, enhancing funding for banks. This was confirmed by Ricci (2015), who concluded that non-standard policy measures affect stock returns more than standard measures. The strongest (negative) effect was found in the reaction of bank stock prices to contractionary measures. For example the stopping of liquidity provisions or asset purchasing programmes.
As mentioned earlier, this paper will distinguish two periods in the sample. These periods will be the sovereign debt crisis and the period after the crises. The 2010-2013 period is seen as the sovereign debt crisis. This period was after the global financial crisis but in the Eurozone there were huge shocks to the cost of borrowing for countries, threatening the stability of the European Union financial system (Beltratti & Stulz, 2017).
Research from Berger et al., (2017) analysed a special action taken by the Federal Reserve during the financial crisis. The Federal Reserve decided to create the Term Auction Facility. This facility gave banks the option for short term lending anonymously to prevent negative impact of lending because markets interpret short term lending by banks as weak. The usage of this lending facility showed to be enormous. Compared to the short term borrowing by banks prior to the crisis, the average amount borrowed per day multiplied by
thirteen-hundred. Research showed that this borrowing resulted in growth in lending to the private sector. It also showed the scale of the liquidity problems banks faced during the crisis. Similar liquidity provisions were offered by the European Central Bank to European banks during the crisis. It is expected that liquidity provisions have had a positive effect on the stock price of banks during the crisis. The liquidity provisions allowed the banks not only to stay liquid but also expand their commercial and retail banking activities. Although the European Central Bank decreased most of liquidity provision programmes in the period after the sovereign debt crisis, some new initiatives were announced and some programmes prolonged (ECB, 2014). It is expected that these liquidity provisions have had a positive effect on bank stock returns for the period after the crisis.
Hypotheses
Stock prices of banks are more sensitive to monetary policy announcements by the European Central Bank than the stock market.
Bank stocks are more affected by contractionary measures than expansionary measures.
Liquidity provisions have a positive effect on bank stock prices during and after the sovereign debt crisis.
4. Empirical Analysis
4.1 Method
Event study
To be able to calculate the reaction of bank stock prices to policy announcement there is need to find abnormal returns. For every bank these will be estimated using the method proposed by (MacKinlay, 1997). More specifically, the standard market model will be used to estimate expected normal returns and calculate the difference in stock price reactions. The standard market model will include daily returns of each bank and the daily return of a broad market index. Following Ricci (2015) and Ait-Sahalia et al, (2010) the policy interventions will be divided in two categories: Expansionary and Restrictive/inaction. Policy interventions
come in three different types. Liquidity decisions, interest rate decisions and monetary easing decisions. Decisions on interest rates are to decrease (IR_CUT), to increase and unchanged. The latter two will be put in a single category (IR_INC_UNC). Monetary easing policy includes the purchase of government bonds or corporate bonds in primary or secondary markets (MONEASE). When the central bank announces that quantitative or credit easing will stop, it is treated as a contractionary measure, (CONT). Liquidity policy announcements contain decisions considering the provision of domestic currency, extension on accepted collaterals or provision of foreign currency swaps (LIQ). When interest rates don’t change or increase, monetary easing programmes are stopped and of liquidity restrictions are put in place, this is considered restrictive/inactive monetary policy (INA_RESTR). When monetary easing programmes are announced, interest rates are decreased and liquidity provisions in foreign and domestic currency are offered, it will be labelled an expansionary measure (EXP_MS). In some cases the central bank announces several policy measures on the same date. Following Fiordelisi et al., (2014) this will be dealt with by determining the importance of the decisions and treat them as the single, most important, event. When interest rates are cut this is always the most important event.
4.2 Sample
Policy interventions of the European Central Bank and 25 European based banks will be analysed. It considers the period January 2010 to January 2016. The sample of the paper consists out of 25 large European banks that met requirements for the research.
First, it has to be based in the European Monetary Union. Second, it must be publicly traded on a European stock market. Third, it must have been subjected to the 2012 credit test or the 2011 and 2016 stress-test of the European banking Authority. The European banking authority conducts stress tests and credit tests to assess the soundness of European banks in hypothetical scenario’s. It covers around 70% of the Banking assets in Europe and therefore the combined banks subjected to this test are a good representation of the European banking sector. The testing also attempts to ensure that banks behave accordingly in terms of risk taking, especially considering all the systematic important banks are included (EBA, 2016).
Table 2
Large European banks are included in the sample. These are selected following three criteria. First, they have to be based in the European monetary union. Second, they have to be traded publicly. Third, they have been subjected to the 2011 and the 2016 stress test conducted by the European banking authority, or the 2012 credit test. Data retrieved from official European banking authority website, (2011, 2012, 2016).
Number Name Country
1 Allied Irish Banks plc Ireland
2 Alpha Bank Greece
3 BNP Paribas France
4 Banca Monte dei Paschi di Siena S.p.A Italy
5 Banco BPI Portugal
6 Banco Bilbao Vizcaya Argentaria S.A. Spain 7 Banco Comercial Portugues SA Portugal 8 Banco Popular Español S.A. Spain 9 Banco Santander S.A. Spain 10 Banco de Sabadell S.A. Spain 11 Banco espirito santo Portugal 12 Bank of Ireland Ireland
13 Commerzbank AG Germany
14 Crédit agricole France 15 Deutsche Bank AG Germany 16 Erste Group Bank AG Austria 17 Eurobank Ergasias Greece 18 ING Groep N.V Netherlands 19 Intesa Sanpaolo S.p.A Italy
20 KBC Group NV Belgium
21 Pireaus Bank Greece
22 Raiffeisen Bank INTL Germany 23 Société Générale S.A. France 24 UniCredit S.p.A. Italy 25 Unione Di Banche Italiane Italy
4.3 Model
The market model is as follows for every security i.
Rit = α + βRmt + ε
With E(εit=0) and var(εit) = σεit ARit = Rit – E(Rit|X)
ARit is the abnormal return, Rit the actual return and E(Rit|X) the expected return for
the security calculated with CAPM following the market model. The model relates the return of the security to the return of the market portfolio.
ARit =Rit – α –βRmt
After the abnormal returns for every event are calculated, they are aggregated to get cumulated abnormal returns. CAR(t1,t2) = ∑AR
This aggregates the abnormal returns over time, the aggregated abnormal returns for all the securities is calculated as follows. AR = 1/N*∑ARit
To test the null hypothesis that the abnormal returns are zero the proposed standard method from MacKinlay (1997) is used. It is appropriate to use the sample variance measure from the market model regression in the estimation window. H0 can be tested using:
Z = CAAR(t1t2) / S(CAAR(t1t2)) – N(0,1)
Although MacKinlay advises not to include the event window in the estimation window, this will not be followed. This will has two reasons. First, the period that is used for the estimation is considerably larger than the 250 days proposed, six years of daily returns. This will numb outliers and suggests a more robust estimation. Second, the years before the event windows include the years of the global financial crisis. In these years never foreseen liquidity problems occurred but also never before used expansionary measures have been introduced (Ricci, 2015). Therefore to avoid risking a biased estimation through nonstandard data these years will not be included. Furthermore, when a 250-day estimation window will be used, estimation windows will include other, comparable measures which would also bias the estimation.
Considering all the banks in the sample are European monetary union based and the fact that we consider European Central Bank announcements, the MSCI European Monetary Union price index will be used as a market proxy. Partially following Ait-Sahalia et al, (2010) short windows will be used to estimate the abnormal returns, specifically five day (-1,+3), three-day (-1,+1), two day (0,1) and one day (0,0).
For the risk-free rate an average of the monthly data s of 10-year German treasury bonds will be used for the period January 2010- January 2016. The reason for this is the fact that the German 10-year yield is seen as the benchmark for the Euro Area (Abad et al., 2010).
4.4 Data
For the collection of the policy announcement dates the official press releases on the European Central Bank website will be used. The final sample, after correcting for
overlapping events, contains 85 policy announcements from the period January 2010 till January 2016. After correcting for overlapping events there are a few cases in which the events are close to each other, so close that in the 5-day window of the event window they overlap. These are not excluded from the sample to avoid selection bias. The stock prices of the banks are retrieved from datastream.
Table 1
All monetary policy interventions by the European Central Bank are listed in this table. Information was collected from the ECB official website over the period 01/01/2010 – 01/01/2016. (IR_CUT) means a decrease in the interest rate. (LIQ) stands for the provision of liquidity and (MONEASE) for the announcing of monetary easing programmes. These fall in the category (EXP_MS); Expensionary measures. (IR_UNC/INCR) indicates and increase in the interest rate or unchanged. (CONT) means the ending of liquidity provisions or monetary easing programmes. These two form the category (INA_RESTR); no action and restrictive measures.
IR_CUT LIQ MONEASE EXP_MS IR_UNC/INCR CONT INA_RESTR TOTAL
ECB 08 29 07 44 34 07 41 85
5. Results
5.1 Event study
The results of the analysis of the cumulated average abnormal return are presented in table 3. It is clear that after testing for abnormal returns to be equal to zero not many results are statistically significant. Furthermore the 1% significance level has not held on any policy measure. However the fact remains that the table shows both positive and negative
significant cumulated average abnormal returns. This states that under certain policy measures the expected bank stock returns are higher or lower than the actual bank stock returns. From this there may well be assumed that banks are more sensitive to monetary policy shocks than the whole stock market, or MSCI EMU index in this sample. Therefore showing some
evidence for the first hypothesis. The cumulated abnormal returns showed a large variance over the whole period, the sovereign debt crisis period and the post-crisis period. This implies a strong heterogeneous response to policy measures by individual banks and requires more research on the determinants of these responses. What is also shown in the table is a heterogeneous response to monetary policy announcements in different periods of time.
5.2 All sample period
When looking at the results of the whole sample some other conclusions can be made. First the contractionary measures. The effect of contractionary measures is the largest and most significant effect in the whole period. This was also found by Ricci (2015) and hypothesized in this paper. A reason for this was also hypothesized by her, reasoning that contractionary measures are as effective as they are because of the common belief that banking systems are very dependent on monetary policy interventions and central bank funding. The other restrictive/inactive measure shows also a negative effect on bank stock return over the whole period, although only significant for the three- day window at the 5% significance level.
Considering expansionary measures first will be looked at the monetary easing programmes. As expected and in accordance with Fiordelisi et al., (2014) the results show a positive effect in bank stock return over the period. However, what does not align with the research from Fiordelisi et al., (2014) is the fact that it is barely significant. Only the three day window at 10%. Liquidity provision programmes also show results that do not align with the expectations derived from previous literature. Liquidity provision programmes show a negative effect on bank stock returns, significant for the five, three and one day window. These findings directly oppose the suggested third hypothesis concerning the effect of liquidity provisions on bank stock returns. Interest rate cuts show a positive effect, however only slightly significant for the two day period.
5.3 Sovereign debt crisis period
When looking at the sovereign debt crisis period some different results are shown. First however, contractionary measures have a significant negative effect on stock returns as the whole period shows too. Second Interest rate increases and unchanged also have a negative impact, but show no significance over the sovereign debt crisis period.
Concerning expansionary measures, monetary easing programmes show a significant positive effect on bank stock return. These results almost compare with the effect of contractionary
measures, but with a different sign. It is clear that non-standard policy measures have the greatest effect on bank stock returns, in accordance with Fiordelisi et al., (2014). The fact that monetary easing has such a positive influence in this period may be found in investor beliefs. As central banks introduced nonstandard monetary policy measures during the financial crisis, investors may believe monetary easing programmes are specifically designed to support banks and their funding. Therefore provoking positive effects on bank stock returns when monetary or credit easing programmes are announced.
The effect of liquidity provisions has a positive effect in this period, significant for the five and the one day event window. This was expected from earlier research from Berger et al., (2017) and hypothesized in this paper. Interest rate cuts however do not show any
significant effects during the sovereign debt crisis. A possible explanation for this may be that investors interpret interest rate cuts in this period as an indication for negative economic expectations from the central bank and severity of the situation.
5.4 Post-crisis period
The post-crisis period shows some very different results from the earlier period. First and most surprising, the contractionary measures seem to have a significant positive effect on bank stock returns in the period after the sovereign debt crisis. Concerning inactive and restrictive measures in this period it shows a clear significant effect of interest rate increases or no changes. This is a negative effect on bank stock return.
Expansionary measures show some surprising results in this period too. Starting with monetary easing programmes. It seems that monetary easing programmes have a negative effect on bank stock returns in the period after the crisis. Although, this effect is barely significant for only the five and two day window. Liquidity provisions show a negative effect too, only significant for the five day window and 10% significance level. As opposed to the sovereign debt crisis, interest rates cuts show a positive effect on bank stock return.
Table 3
This table shows European banks’ stock price reaction to central bank policy announcements. Contains the statistics of cumulated abnormal returns estimated over event windows for 85 monetary policy announcements. The period investigated is January 2010-January 2016. The banks investigated are found in table two. To estimate daily abnormal returns, event study methodology of MacKinlay(1997) is used. Specifically, the market model. The market proxy that is used is the MSCI European monetary union index. The significance of the statistics of the cumulated average abnormal returns is calculated using the basic approach for the event study methodology from MacKinlay (1997). Sovereign debt crisis: 01/01/2010 – 01/06/2013. Post crisis period: 02/06/13 – 01/01/16. (IR_CUT) means a decrease in the interest rate. (LIQ) stands for the provision of liquidity and (MONEASE) for the announcing of monetary easing programmes. These fall in the category (EXP_MS); Expensionary measures. (IR_UNC/INCR) indicates and increase in the interest rate or unchanged. (CONT) means the ending of liquidity provisions or monetary easing programmes. These two form the category (INA_RESTR); no action and restrictive measures. ***,**,* indicate significance at 1, 5 and 10%. Data from European Central Bank official website and Datastream.
ALL Sample period IR_CUT CAAR Z-stat 212 obs LIQ+ CAAR Z-stat 678 obs MONEASE CAAR Z-Stat 139 obs IR_UNC/INC R CAAR Z-stat 802 CONTR CAAR Z-stat 162 obs (-1,3) -0.0184 -0.401 -0.0751* -3.206 0.0056 1.326 -0.0024 -1.112 - 0.009** -2.183 (-1,1) 0.0024 0.443 -0.006** -2.032 0.0572* 1.768 -0.0071** -2.843 -0.0260* -1.801 (0,1) 0.0072* 1.856 -0.0505 -1.520 0.0577 0.431 -0.0058 -1.443 -0.0124* -1.857 (0,0) 0.0065 1.428 -0.0079* -1.771 0.0173 0.922 -0.0169 -0.618 -0.0193** -2.348 Sovereign Debt crisis
IR_CUT Z-stat LIQ+ Z-stat MONEASE CAAR Z-Stat IR_UNC/INC R CAAR Z-Stat CONTR CAAR Z-stat (-1,3) -0.0069 -1.038 0.0170** 2.604 0.0286** 2.014 0.0112 0.337 -0.0528* -1.978 (-1,1) -0.0254 -1.797 0.0065 1.283 0.0321** 2.202 -0.0375 0.495 -0.0423** -2.653 (0,1) -0.0078 -0.801 0.0264 1.322 0.0106 1.348 -0.0173 -0.331 -0.0125 -1.319 (0,0) -0.0021 -0,13 0.0109* 1.91 0.0151 1.161 -0.0043 -0.59 - 0.023* -2.115 Post-Crisis IR_CUT CAAR Z-stat LIQ+ CAAR Z-stat MONEASE CAAR Z-stat IR_UNC/INC R CAAR Z-stat CONTR CAAR Z-stat (-1,3) 0.0275 0.396 -0.0069* -1.994 - 0.0135* -1.876 -0.0061** -2.288 0.0108* 1.819 (-1,1) 0.0043 1.013 -0.0041 -1.081 -0.0079 -1. 021 -0.0023* -2.459 0.0319* 2.123 (1,0) 0.0026** 2.139 -0.0349 -0.743 -0.0079* -1.884 -0.0087* -2.153 0.0123** 2.127 (0,0) 0.0028** 2.322 -0.0185 -0.237 0.0292 1.524 -0.0052 -1.219 0.0142* 1.812
6. Conclusions and research limitations
To answer the research questions an event study was done to find and analyse
cumulated abnormal returns. These abnormal returns were measured around monetary policy announcement dates and give an indication of bank stock price reaction to monetary policy. It was found that there is an heterogeneous response to monetary policy in different periods of time. Respectively, the period of the sovereign debt crisis and the period after the crisis.
Significant positive and negative reactions to monetary policy announcement suggests confirmation of the first hypothesis. Using a normal market model expected returns were calculated which did significantly differ from the observed returns around certain policy announcement dates. Although these reactions are heterogeneous in different periods in time it can be concluded that for most policy announcements, bank stock prices are more sensitive than the stock market. However though, these results were not overwhelmingly significant.
What can be concluded is that the strongest reaction was found in the bank stock return response to contractionary monetary policy during the sovereign debt crisis. This is in accordance with earlier literature from Ricci (2015). Considering the second hypothesis concerning the difference in sensitivity to contractionary and expansionary policy, it can be confirmed that in in the overall sample and in the period of the sovereign debt crisis there was a stronger reaction to contractionary policy measures. However, surprisingly to this finding was that for the period after the crisis there seems to be a positive response to contractionary measures.
The third hypothesis, concerning expected positive reactions to the announcement of liquidity provisions, can’t be confirmed. Built on earlier research from Berger et al., (2017) it was expected that liquidity provision announcements would have a positive impact for both the sovereign debt crisis period and the period after the crisis. However while liquidity
provisions do have a slightly significant positive effect during the sovereign debt crisis, this is not the case for the period after. Which shows a barely significant negative effect.
There are a few possible explanations for the fact that few results were significant. First of all, the event windows were included in the estimation windows. MacKinlay (1997) proposes that it is typical that the estimation window and the event window don’t overlap. He proposes this because otherwise the parameters of the normal return model are influenced by the returns around the event. However, for the proposed 250-day estimation window he proposed there would also be constant overlap with comparable events in the sample. Therefore, a six year daily return estimation window was applied.
The second reason for non-significant results may be the amount of observations, since there are only a fixed and limited amount of central bank policy announcements available. This may also be the reason for the fact that contractionary monetary policy announcements seem to have an positive effect on bank stock return in the period after the crisis. There have been only four contractionary measures in that period. These contractionary measures were the stopping of liquidity provisions.
In the post-crisis period banks were forcibly higher capitalized because of Basel III regulation, and less dependent on central bank liquidity provisions. Therefore it may be that returns were high enough in that period for contractionary measures not to make a negative impact. Similar conclusions are outlined by Gambacorta & Song Shin (2016). They find that when banks are higher capitalized, contractionary monetary policy affects them less.
Uninsured financing is easier available. It may therefore well be that because of the Basel III regulation, banks were better capitalized post-crisis and therefore did not react as negatively as would be expected with the normal market model.
Another possible reason may be found in positive investor expectations. During the sovereign debt crisis, monetary easing measures had a positive effect on abnormal returns and contractionary measures had a negative effect. Investors expected that expansionary measures would make banks more profitable and with contractionary measures less profitable. In the period after the crisis investors may have been unexpectedly optimistic about financial
markets, so much that contractionary measures would not have a significant impact on bank’s profitability.
A shortcoming to the research can be found in the fact that for the calculation of abnormal returns there has not been controlled for the effects of monetary policy set by other central banks. For the sake of accuracy it might be needed in further research to control for the spillover effects of different central banks. Ricci (2015) found that although the Swiss, English and Japanese central banks do not cause spillover effects for European banks, the Federal reserve does.
An important finding is the fact that the cumulated abnormal returns show a large variance over the whole period, the sovereign debt crisis period and the post-crisis period. This large variance implies high heterogeneity in bank responses to monetary policy announcements. It may be useful in further research to find the determinants of this
heterogeneous response. Especially for portfolio management purposes it can be valuable to be able to allocate specific bank characteristics to certain responses to monetary policy announcements.
7. References
Abad, P., Chulia, H., Gomez-Puig, M., (2010). EMU and European government bond m market integration. Journal of Banking & Finance, 34, 2851-2860.
Aït-Sahalia, Y., Andritzky, J., Jobst, A., Nowak, S., and Tamirisa, N. (2012) Market response to policy initiatives during the global financial crisis. Journal of International
Economics 87, pp. 162–177.
Altunbas, Y., Gambacorta, L., Marquez-Ibanez, D., (2010). Does Monetary policy Affect Bank Risk-Taking? ECB working paper no. 1166
Beltratti, A., & Stulz, R.M., How important was contagion through banks during the
European sovereign crisis ? Fisher College of Business Working Paper Series, 06/17.
Berger, A.N., & Bouwman, C.H.S, (2013). How does capital affect bank performance during financial crises? Journal of Financial Economics, 109, 146-176.
Berger, A.N., Black, L.K., Bouwman, C.H.S., Diugosz, (2017). Bank loan supply responses To Federal Reserve emergency liquidity facilities. Journal of Financial
intermediation, 03/17.
Bernanke, B.S., & Kuttner, K.N., (2005). What explains the stock market’s reaction to Federal Reserve Policy? Journal of Finance, 3, 1221-1257.
Gambacorta, L., & Song Shin, H., (2016). Why bank capital matters for monetary policy.
Journal of Financial intermediation, 10, 1-13.
Haitsma, R., Unalmis, D. & de Haan, J., (2016). The impact of ECB’s conventional and unconventional monetary policies on stock markets. Journal of Macroeconomics, 48,
101-116.
Mackinlay, A.C., (1997) Event Studies in Economics and Finance. Journal of Economic
Paligorova, T., Santos, J.A.C., (2017). Monetary policy and bank risk-taking: Evidence from the corporate loan market. Journal of Financial Intermediation. 30, 35-49.
Ricci, O. (2015). The impact of monetary policy announcements on the stock price of large European banks during the financial crisis. Journal of Banking & Finance, 52, 245-255.
Rosa, C., (2011). Words that shake traders: The stock market’s reaction to central bank communication in real time. Journal of Empirical Finance, 18, 915-934.
Yin, H., & Yang, J., (2013). Bank characteristics and stock reactions to federal funds rate target change. Applied Financial Economics, 23, 1755-1764.
