Communication of the European Central Bank and contagion on financial markets
|Název práce v češtině:||Komunikace Evropské centrální banky a nákaza na finančních trzích|
|Název v anglickém jazyce:||Communication of the European Central Bank and contagion on financial markets|
|Klíčová slova:||Finanční krize, nákaza, vzájemná závislost, coexceedance, komunikace centralní banky|
|Klíčová slova anglicky:||Financial crisis, contagion, interdependence, coexceedance, central bank communication|
|Akademický rok vypsání:||2014/2015|
|Typ práce:||diplomová práce|
|Ústav:||Institut ekonomických studií (23-IES)|
|Vedoucí / školitel:||prof. Roman Horváth, Ph.D.|
|Řešitel:||skrytý - zadáno vedoucím/školitelem|
|Datum a čas obhajoby:||23.06.2016 09:30|
|Místo konání obhajoby:||IES|
|Datum odevzdání elektronické podoby:||12.05.2016|
|Datum proběhlé obhajoby:||23.06.2016|
|Oponenti:||PhDr. Michal Hlaváček, Ph.D.|
|Seznam odborné literatury|
|Bae, K.-H., Karolyi, G.A., Stulz, R.M., 2003. A new approach to measuring financial contagion. Review of Financial Studies 16 (3), 717– 763.
Baur, Dirk & Schulze, Niels, 2005. Coexceedances in financial markets--a quantile regression analysis of contagion, Emerging Markets Review, Elsevier, vol. 6(1), 21-43.
Beck, Melanie-Kristin & Bernd Hayo & Matthias Neuenkirch, 2013. Central Bank Communication and Correlation between Financial Markets: Canada and the United States, Journal of International Economics and Economic Policy, Springer-Verlag, vol. 10(2), 277-296.
Beirne, John & Jana Gieck, 2014. Interdependence and Contagion in Global Asset Markets, Review of International Economics, Wiley Blackwell, vol. 22(4), 639-659.
Born, Benjamin & Michael Ehrmann & Marcel Fratzscher, 2014. Central Bank Communication on Financial Stability, Economic Journal, Royal Economic Society, vol. 124(577), 701-734.
Chevapatrakul, Thanaset & Tee, Kai-Hong, 2014. The effects of news events on market contagion: Evidence from the 2007–2009 financial crisis, Research in International Business and Finance, Elsevier, vol. 32(C), 83-105.
Christiansen, Charlotte & Ranaldo, Angelo, 2009. Extreme coexceedances in new EU member states' stock markets, Journal of Banking & Finance, Elsevier, vol. 33(6), 1048-1057.
Forbes, Kristin J. & Roberto Rigobon, 2002. No Contagion, Only Interdependence: Measuring Stock Market Comovements, Journal of Finance, American Finance Association, vol. 57(5), 2223-2261.
Knutter, R., B. Mohr, & H. Wagner, 2011. The effects of central bank communication on financial stability: A systematization of the empirical evidence.” Fern Universitat in Hagen, Discussion Paper No.463.
Poon, S.-H., Rockinger, M., Tawn, J., 2004. Extreme value dependence in financial markets: diagnostics, models,
and financial implications. Review of Financial Studies 17 (2), 581–610.
|Předběžná náplň práce v anglickém jazyce|
Financial markets’ linkages are of great importance, especially in periods of market turmoil as recent crisis confirmed. Transmission of crisis to other countries is present mainly in financial system characterized by high degree of interdependence and is usually connected with contagion. Transmission strengthens the linkages and this, therefore, decreases investors’ benefits resulting from diversification and at the same time, it makes policymakers more attentive towards financial stability. However, it is crucial to differentiate between contagion and interdependence. Forbes et al. (2002) define interdependence as a situation with ‘’high level of comovement during all states of world (during crisis as well as more stable periods)’’, while contagion is according to latest papers (e.g. Beirne et al., 2014) defined as structural break in international propagation mechanism during a crisis period. Therefore, linkages must be assessed before, during and also after the crisis in order to make conclusion about the presence of contagion.
Since researchers do not agree on one particular definition of contagion, measures of contagion differ as well. One of the measures is coexceedance proposed by Bae at al. (2003). They define coexceedance as number of joint occurrences of extreme returns. Baur et al. (2005) apply quantile regression model to analyze the degree and occurrences of coexceedances. Their approach tries to identify linear as well as non-linear linkages between contagion and its determinants. Results of the research confirm predictability of contagion within and across regions. Additionally, regional (world) market return and its volatility affect contagion and the effect is stronger for extreme negative returns. Study by Christiansen et al. (2009) focuses on new EU member states’ stock markets and occurrence of extreme coexceedances. They conclude that coexceedance of old EU states are more influenced by US stock market and price movements of other asset classes.
Most of the studies implementing coexceedance as a measure of contagion focuses on identifying its determinants such as conditional volatility, exchange rates and interest rates. However, central bank communication might also play a role to substantial extent. Research proves that various means of central bank communication influence financial stability. Knutter et al. (2011) provide evidence that speeches and press conferences can be considered as one of the most effective channels of central bank communication in maintaining financial stability. In similar manner, Born et al. (2014) show that speeches and interviews do not have a significant impact on stock market returns during tranquil times, but on the other hand, they have substantial effect during the global financial crisis. Beck et al. (2013) analyze comovements of stock and bond markets between the United States and Canada as well as within Canada. Applying diagonal-BEKK models it is found out that communication of the central banks significantly affects correlation of these financial markets. This suggests that contagion could be affected by central bank communication in form of speeches and interviews.
1. Hypothesis #1: ECB communication affects contagion immediately.
2. Hypothesis #2: ECB communication has long lasting effect on contagion.
3. Hypothesis #3: Nature of the statement (hawkish or dovish) matters in the manner the contagion spreads.
As suggested in paper by Bae et al. (2003), coexceedance will be used in order to determine periods of contagion despite the possibility of usage of other measures such as correlation coefficient and cointegration analysis, since correlation coefficient is a linear measure sensitive to heteroscedasticity. This thesis will follow the approach by Bae et al. (2003) and apply ordered logit method. This particular method is chosen due to discrete and ordered nature of dependent variable. Moreover, it allows to assess the probability of further spreading contagion to other countries.
To test the first hypothesis, we extend the model by Baur et al. (2005), which incorporates main determinants of coexceedance, e.g. dummy crisis, market returns, conditional volatility and include additional variable communication whose coefficient will represent the effect of the European Central Bank’s communication on contagion. Since the model includes also conditional volatility, the estimation will be performed in two steps similarly to Chevapatrakul et al. (2014): firstly, conditional volatility will be calculated using general autoregressive conditional heteroscedasticity (GARCH) model, secondly, ordered logit regression will be estimated. Nextly, lagged variable communication will be added to the original model. Therefore, we will be able to assess whether market participants keep in mind content of speeches published during previous days and hence, contagion is affected by current as well as historical communication. This enables to address the second hypothesis. The third hypothesis will be tested by adding another variable statement to original model which is equal to 1 if the statement is identified as hawkish or containing tightening surprise with respect to prevailing trends or it is equal to 0, if it is identified as dovish.
Data for variable communication and statement will be created by coding daily statements of individual representatives of the ECB released on Reuters. If the statement suggests good economic outlook, communication will be equal to 1. On the other hand, speeches indicating neutral outlook, no cut or no rise, will be given value 0. Finally, statements regarding bad outlook will be assigned -1. Moreover, all these statements will be sorted based on their dovish/hawkish nature. Information about the rest of the variables needed for the models is expected to be obtained from Reuters Wealth Manager.
Effect of central bank communication on financial stability has been studied to large extent. However, any research has not been performed on the relationship between communication of the ECB and contagion on financial markets (to the best knowledge of the author of this proposal). Since it is so far unexplored field, this thesis is expected to contribute substantially to the current academic discussion. It might prove that there is additional significant effect of the central bank’s communication which could serve as a higher incentive for effective communication and thus, ability to regulate even financial markets’ behaviour in a purpose to decrease the contagion.
1. Introduction – Firstly, transmission mechanism and importance of contagion will be described. Nextly, motivation for the examined hypothesis as well as description of thesis structure will be included.
2. Literature review – In this section, brief summary of the studies will be presented in order to examine what has been done so far in this field. Attention will be paid to different definitions of contagion and its channels as well as determinants of coexceedance.
3. Data description – Data sources, statistical description of the main variables will be presented. Additionally, detailed description of creation of the ECB communication dataset will be included.
4. Empirical methodology – Various methods of measuring contagion, their advantages and drawbacks are expected to be discussed. Furthermore, concept of coexceedance will be described in details. Choice of the particular estimation framework will be supported by arguments following from econometric theory and econometric model will be presented.
5. Results – Interpretation of the results will be part of this section. Moreover, discussion will follow and possible policy applications will be suggested.
6. Conclusion – Main findings of the thesis will be summarized and recommendation for further research will be made.