This master´s thesis deals with testing the efficiency of capital markets. The subject to verification of the stock markets of Austria, Hungary, Germany, Great Britain, Czech Republic and Poland during the 2006-2016Q1. The aim of this work is to test the weak formo f efficiency in these markets. The hypothesis was tested using the ACF test ADF and KPSS tests, variance ratio test, run test and test of January effect.

This thesis is focused on testing the weak effectiveness of the US, Japanese, German and Czech market in the period 1995 - 2015. The first part contains a theoretical basis for the theory of efficient markets, the conditions, characteristics and models. Further test methods of weak market efficiency are presented and semistrong and strong effectiveness mentioned. The practical part deals with the introduction of tested stock exchange indices and by testing the weak effectiveness of these markets through tests of randomness, variance ratio test and serial correlation.

Made available in DSpace on 2016-04-25T16:44:37Z (GMT). No. of bitstreams: 1 Andre Nardy.pdf: 1089149 bytes, checksum: bf93de2a1a852c7d9ef44cfa8f114323 (MD5) Previous issue date: 2014-02-12 / The dynamics of abnormal returns , volume and betas is analyzed for Bovespa s stocks included or excluded from the Ibovespa index between 2003 and 2012, in a phenomenon known in the financial literature as the index effect, one of the oldest reported anomalies. Event studies are used with different settings of estimation window to measure abnormal returns and assess its effect on the calculation of return for the market model , since the calculation of the theoretical portfolio of Bovespa is known beforehand and is based on marketability and liquidity. No abnormal return is veryfied for shares on the date of their effective entry on the index, only abnormally high volumes. On the date of the first preview of inclusions positive abnormal returns and volumes are observed, and so on for excluded stocks. However, when we exclude from the sample companies with IPOs up to 3 years of its entry into the Bovespa Index and those assets included during the crisis of the financial markets, it appears tha abnormal returns do occur on the effective date, consistent with previous literature on the theme. The betas of the stocks included tend to covariate with greater force after inclusion in the index . With the results achieved market efficiency in the semi-strong form cannot be challenged for the Brazilian stock market, but there is a possible change in the occurrence of the index effect for the period studied, compared with previous studies / Analisa-se a ocorrência para o Ibovespa de dinâmica anormal de retornos, volume e dos betas para as ações incluídas ou excluídas do índice, entre 2003 e 2012, em fenômeno conhecido dentro da literatura de finanças como Efeito índice, uma das anomalias mais antigas relatadas. Utilizam-se estudos de eventos em diferentes configurações de janela de estimação para medir os retornos anormais e avaliar o efeito da mesma na apuração de retorno pelo modelo de mercado, dado o cálculo da carteira teórica do Ibovespa ser conhecido de antemão e baseado em negociabilidade e liquidez. Não se encontram ocorrências de retorno anormal para a data de efetiva entrada das ações, apenas volumes anormalmente altos. Na data de primeira prévia das inclusões ocorrem retornos e volumes anormais positivos, o mesmo ocorrendo para exclusões. Entretanto, ao se excluir da amostra de inclusões as empresas com IPOs realizados até 3 anos de seu ingresso no Ibovespa e aqueles ativos incluídos durante a crise dos mercados financeiros, verifica-se retornos anormais na data de efetivação da nova carteira teórica, coerente com a literatura precedente. Os betas das ações incluídas tendem a covariar com maior força após a inclusão no índice. Com os resultados não é possível questionar a eficiência na forma semiforte para o mercado acionário brasileiro, porém verifica-se uma possível mudança na ocorrência do efeito índice para o período estudado, em comparação com estudos anteriores

Studying whether new trading rules provide higher returns than the buy-and-hold strategy is relevant for both finance theory and the asset management field. In this thesis, we examine the profitability of the newly proposed trading strategy based on the concept of price overreaction on eight developed stock indices. In comparison to other studies, we extend a definition of price overreaction with an inclusion of a minimum volatility threshold. Based on the Ordinary Least Squares model, we find that a volatility condition significantly improves the predictability of return reversals after positive price overreaction. For comparison with the buy-and-hold, we use Hansen's Superior Predictive Ability test that corrects the data snooping bias. Despite better annualised returns during in-sample and out-of-sample periods, the results show that the proposed strategy is not superior to the buy-and-hold at any stock index due to heavy reliance on the predictions of the largest declines. Nevertheless, we confirm the effect of decoupling (flight to quality) that can positively affect our strategy, but only when we do not take into account transaction costs. In the end, we summarize behavioural concepts that lie behind our strategy as the overreaction and decoupling are mostly justified with cognitive biases.