Capital asset pricing models predict the tax penalty capitalized into expected returns reflects the average tax rate of all investors in a market (average rate theory). To test this theory, I argue stock markets in developed European countries and the U.S. form an integrated market, where the tax capitalized reflects the average rate of investors across these countries. If this is the case, when the U.S. dividend tax rate was cut by 60 percent in 2003, the average rate theory predicts a decrease in the tax capitalized in European stocks. In contrast, firms in less integrated European countries should react significantly less to the U.S. tax cut. Finally, I test a prediction from Desai and Dharmapala (2011) that because of market integration the magnitude of the reaction to the tax cut should be the same for firms in the U.S. and developed European countries. The results in this paper support these predictions.

Esta tese investiga a relação entre disclosure e custo de capital próprio no Brasil. Foi utilizada uma amostra composta pelas 50 empresas mais líquidas da BOVESPA com dados para os anos de 1998, 2000, 2002, 2004 e 2005. Como proxy para o nível de disclosure foi desenvolvido e utilizado o Índice de Disclosure Brasileiro (IDB) que mensura características essenciais de evidenciação ao longo de seis dimensões: (i) informações gerais sobre a empresa, (ii) relações com empregados e administradores, (iii) informações não-financeiras sobre mercado, vendas e produtos, (iv) projeções, (v) discussões e análises de dados financeiros e (vi) outras informações. Para calcular o custo de capital foi utilizado o modelo de Ohlson - Juettner-Nauroth. Os resultados mostram que o nível geral de disclosure aumentou e que a dimensão relativa às projeções dos administradores apresenta os resultados mais modestos. Foi verificada relação negativa e estatisticamente significante entre custo de capital e disclosure para todas as empresas. No entanto, essa relação é mais pronunciada para empresas que recebem menor cobertura por parte dos analistas de mercado e possuem estrutura de propriedade mais dispersa. Este trabalho contribui para a literatura nacional e internacional na área ao demonstrar os benefícios econômicos diretos das ações das empresas no que tange ao nível de evidenciação. / This thesis investigates the association between disclosure and cost of equity capital in Brazil. Results are based on a sample of the fifty most liquid shares traded at BOVESPA using data from 1998, 2000, 2002, 2004 and 2005. To proxy for disclosure level a Brazilian Disclosure Index (IDB) was built and used. IDB measures essential features of disclosure across six components: (i) general information about the company, (ii) relationships with employees and management, (iii) non-financial data about market, sales and products, (iv) projections, (v) analysis and discussion of financial data, (vi) other information. To measure cost of equity capital the Ohlson- Juettner-Nauroth model was used. Results show that the general disclosure level improved along the sample period and that the projections\' component of IDB presented the poorest scores over the years. There is a negative and statistically significant relation between disclosure and cost of equity capital for all firms in the sample. This association is exacerbated for firms with receive less attention from the analysts and have a more dispersed ownership structure. This thesis extends the Brazilian and international literatures by showing the actual economic benefits of superior disclosure levels.

Este trabalho propõe um modelo para estimativa do custo de capital próprio ex-ante baseado no índice lucro/preço, E/P, na taxa de crescimento esperada para o lucro na perpetuidade, g, e no coeficiente ?, o qual mede a taxa de conversão de crescimento em rentabilidade e, portanto, em valor. Tal modelo é testado com base em uma amostra brasileira, utilizando metodologia econométrica de regressões multivariadas em primeira diferença de séries temporais. São utilizadas como variáveis explicativas medidas alternativas para a taxa livre de risco e para a taxa esperada de crescimento do lucro na perpetuidade. Diversas das medidas utilizadas foram relevantes para explicar os índices E/P no Brasil, corroborando as hipóteses formuladas. Os resultados mostram, também, que as medidas mais individualmente adequadas para a taxa livre de risco a serem utilizadas para precificação de ações no Brasil são aquelas baseadas em yields de títulos soberanos emitidos pelo governo dos EUA adicionados do prêmio pelo risco Brasil, medido com base no índice EMBI+ Brazil. Identifica-se que as taxas de juros locais, apesar de apresentarem poder explicativo individualmente inferior, são fatores com poder explicativo conjunto significativo, resultando em um modelo com uma taxa livre de risco média ponderada. Quanto à taxa esperada para o crescimento na perpetuidade, a variável mais significativa foi a projeção de consenso de mercado de curto prazo para o crescimento do PIB. Por fim, estima-se o prêmio pelo risco de mercado ex-ante no Brasil, com base no modelo proposto, utilizando coeficientes e variáveis identificadas nos testes econométricos. Tal estimativa mostrou-se substancialmente mais precisa quando comparada àquelas baseadas em retornos ex-post, inclusive em mercados internacionais, bem como quando comparada a outras metodologias ex-ante no Brasil. / This work proposes a model for the estimation of the implied cost of equity. Such model is based on earnings/price ratios, E/P, on the expected perpetual earnings growth rate, g, and on the ? coefficient, which measures the rate of conversion of growth to return, and, therefore, to value. The proposed model is tested on a Brazilian sample, using multivariate first-difference time series regressions. The explanatory variables include several alternative measures for the risk-free rate and the expected perpetual earnings growth rate. The results show that most of the selected measures were relevant in explaining E/P ratios in Brazil, confirming the proposed hypothesis. The results also show that US sovereign bonds, combined with a measure for the Brazilian risk premium, the EMBI+ Brazil index, are the most relevant measures for the risk-free rate to be used in equity valuation in Brazil. Additionally, we conclude that local interest rates, albeit having individually lower explanatory power, remain relevant in conjunction with international bond yields, resulting in a weighted average risk-free rate. The most significant measure for the expected perpetuity growth rate was the short term consensus forecast for the GDP. Finally, we estimate the ex-ante market risk premium in Brazil, using the proposed model with coefficients and variables selected based on the econometric results. Such estimate is substantially more accurate when compared to ex-post estimates, including those for international markets, as well as other ex-ante estimates for the Brazilian market.

O trabalho seminal de Amihud e Mendelson (1986) abriu caminho para uma grande quantidade de pesquisas no âmbito internacional sugerindo que a liquidez poderia ser um fator relevante na explicação dos retornos das ações. A premissa central é que ativos menos líquidos devem apresentar taxas de retornos superiores a dos ativos mais líquidos, por representarem mais riscos a seus detentores. Assim, o objetivo principal da presente tese consistiu em verificar se há prêmios pela liquidez no mercado acionário brasileiro com o uso de uma vasta quantidade de medidas de liquidez, formas de cálculo e períodos de retenção das carteiras, bem como se o modelo de precificação de ativos de 2-fatores de Liu (2006) - formado pelo beta de mercado e pelo fator liquidez - é válido para o mercado acionário brasileiro e, em caso positivo, se é superior ao CAPM, ao modelo dos 3-fatores de Fama e French (1993) e ao modelo dos 4-fatores de Carhart (1997), na explicação das variações dos retornos cross-section das carteiras de ações. Para isso, foram usadas todas as ações listadas na BM&FBOVESPA, de 1995 a 2011. Os procedimentos metodológicos para obtenção das variáveis e testes para verificação da existência de prêmios pela liquidez seguiram, essencialmente, o estudo de Liu (2006). Os procedimentos para validação e comparação dos modelos de precificação de ativos foram efetuados seguindo o modelo de testes preditivo de Fama e MacBeth (1973). Foram testadas doze medidas de liquidez, dentre as mais recorrentes adotadas na literatura internacional. Os resultados encontrados evidenciaram fortes coeficientes de correlação entre muitas medidas, o que levou à manutenção dos testes com as medidas menos correlacionadas entre si: Índice de Negociabilidade da BM&FBOVESPA, Turnover, Return-to-Volume e Coeficiente de Variação do Volume Financeiro. Dentre estas, observou-se a existência de prêmio pela liquidez estatisticamente significante no mercado acionário brasileiro, na maioria das estratégias testadas, com o uso do Índice de Negociabilidade da BM&FBOVESPA e com o Coeficiente de Variação do Volume Financeiro. Estes resultados indicaram a existência de relação consistente e negativa entre o retorno das carteiras e a liquidez das ações e relação positiva entre a volatilidade da liquidez e o retorno das carteiras. Os prêmios encontrados com o uso do Índice de Negociabilidade da BM&FBOVESPA se mostraram robustos aos testes de subamostras, subperíodos e efeitos sazonais. Em relação aos testes empíricos dos modelos de precificação de ativos, o modelo dos 2-fatores se mostrou válido para explicação das variações dos retornos cross-section das ações no mercado brasileiro. O fator liquidez se mostrou complementar ao fator beta de mercado, aumentando o poder de explicação do modelo quando comparado ao CAPM, especialmente nas carteiras compostas por ações de baixa liquidez. O modelo de 2-fatores se mostrou também superior aos modelos 3-fatores e 4-fatores. Os resultados se mostraram robustos aos testes efetuados quanto a possíveis vieses de seleção do período amostral. Assim, mesmo que nenhum dos modelos tenha se mostrado suficiente na explicação das variações cross-section dos retornos no mercado acionário brasileiro, os resultados parecem indicar que a liquidez é uma direção especialmente promissora para a continuidade das pesquisas sobre o tema. / The seminal study of Amihud and Mendelson (1986) opened the way to a large quantity of researches in international environment suggesting that liquidity may be a important risk factor to explain stock returns. The central premise is that less liquid assets should present rates of return higher than the most liquid assets, because they represent more risk to their owners. Thus, the main objective of this thesis is to check if there is liquidity premium in the Brazilian stock market through the use of a vast amount of liquidity measures, forms of calculation and portfolios retention periods, as well as if the 2-factor pricing model developed by Liu (2006) - formed by the market beta and the liquidity factor - is valid for the Brazilian stock market and, if so, whether it is superior to the CAPM, the 3-factor model of Fama and French (1993) and the 4-factor model of Carhart (1997), in explaining the cross-section variations of assets portfolio returns. For this, this study used all shares listed on the BM&FBOVESPA from 1995 to 2011. The methodological procedures for the variables construction and tests to verify the existence of liquidity premiums followed the study of Liu (2006). The procedures for validation and comparison of asset pricing models were made following the model of predictive tests of Fama and MacBeth (1973). This thesis tested 12 liquidity measures, among the most recurrent adopted in the international literature. The results showed strong correlations between many measures, which led to the maintenance of the tests with measures less correlated: BM&FBOVESPA Negotiability Index, Turnover, Return-to-Volume and Coefficient of Variation of Financial Volume. Among these, were observed the existence of a statistically significant premium for liquidity in the Brazilian stock market, in most of the strategies tested, using the BM&FBOVESPA Negotiability Index and the Coefficient of Variation of Financial Volume. These results indicated that there is consistent and negative relationship between portfolio returns and shares liquidity and positive relationship between liquidity and volatility of portfolio returns. Prizes found using the BM&FBOVESPA Negotiability Index were robust to tests of subsamples, subperiods and seasonal effects. Regarding the asset pricing models empirical testing, the 2-factors model proved valid explanation for the cross-section variations of returns of the shares in the Brazilian market. The liquidity factor proved to complement the market beta, increasing the explanatory power of the model when compared to the CAPM, especially in portfolios composed of stocks with low liquidity. The 2-factor model was even superior to 3-factor and 4-factor models. The results have not changed even after the robustness tests regarding possible sample period selection biases. So even though none of the models has been shown enough in explaining the cross-section variations of stock returns in the Brazilian market, the results seem to indicate that liquidity is a particularly promising direction for continued research on the topic.

Esta dissertação procurou comparar os modelos CAPM, três fatores de Fama e French (1993) e quatro fatores de Carhart (1997) com o objetivo de verificar qual possui o maior poder de explicação das variações dos retornos dos ativos no mercado brasileiro. O estudo considerou 512 ações listadas na Bolsa de Valores de São Paulo no período de 1995-2011. Utilizou-se metodologia preditiva, que aplica regressão em dois estágios - série temporal e corte transversal - com erro padrão calculado de acordo com técnica desenvolvida por Fama e Macbeth (1973). Foram calculados os retornos mensais das ações, que foram agrupadas em carteiras e utilizadas como variável dependente nas regressões. Já foram utilizadas como variáveis independentes os fatores carteira de mercado, tamanho, índice book-to-market e momento. Os resultados observados indicaram que, apesar de o modelo de três fatores de Fama e French (1993) ter apresentado maior poder preditivo em relação ao de quatro fatores de Carhart (1997) e ao CAPM, nenhum dos modelos foi suficiente para explicar as variações dos retornos das carteiras formadas. Os fatores tamanho e momento não foram estatisticamente significantes, o que indicou que não foi possível observar no mercado brasileiro os mesmos efeitos documentados por Fama e French (1993) e Carhart (1997). Já o fator mercado foi significante somente nos modelos com intercepto, e o fator índice B/M foi significante somente nos modelos sem intercepto. O intercepto foi fortemente significante nos modelos de três e quatro fatores, o que sinaliza, juntamente com o baixo poder de explicação dos modelos, que outros fatores não incluídos nos modelos, seriam capazes de explicar as variações dos retornos dos ativos. / This dissertation seeks to compare the CAPM model, three factor model of Fama and French (1993) and four factor model of Carhart (1997). Its goal is to verify the one that has the best capability to explain the stock return variations in the Brazilian market. This study considered 512 stocks listed in São Paulo Stock Exchange (BOVESPA) along the period 1995-2011. Predictive methodology, which applies a two stage regression (time series and cross-sectional), was used. The standard error was calculated in accordance to the technique developed by Fama and Macbeth (1973). Stocks monthly returns were calculated and grouped in portfolios that were employed afterwards as dependent variable in the regressions. The market portfolio factor, the size factor, the book-to-market index factor and the momentum factor were used as independent variables. The observed results indicated that, despite the tree factor model of Fama and French (1993) had showed a better predictability over the four factor model of Carhart (1997) and CAPM, none of these models were enough to explain the return variation of the formed portfolios. The size and momentum effects weren\'t statistically significant, which indicates that it was not observed, in the Brazilian market, the same effects documented by Fama and French (1993) and Carhart (1997). The market factor was significant only in the models with the intercept, and the B/M index factor was significant only in the models without the intercept. The intercept itself was strongly significant in the tree and four factor models, which, combined with the poor power of explaining the models, signalizes that other factors not included in the models would be able to explain the stock return variations.

Este trabalho se baseia na análise de 125 Laudos de Avaliação emitidos no contexto de Ofertas Públicas de Aquisição (OPAs) no Brasil realizadas no período entre 2006 e 2017. As OPAs movimentaram dezenas de bilhões de Reais, envolvendo companhias avaliadas, no total, em mais de R$ 300 bilhões. Os objetivos principais são três: (i) mapear as metodologias de avaliação utilizadas, (ii) identificar erros cometidos, comparando as metodologias utilizadas com o referencial teórico e (iii) medir vieses de posição na preparação das avaliações. Os resultados indicam (i) diversidade de metodologias aplicadas, o que pode gerar inconsistência e viés nos resultados, (ii) existência de erros, inclusive relacionados a conceitos básicos de avaliação de empresas e (iii) viés de posição por parte dos avaliadores, especialmente empresas independentes de consultoria, que tendem a emitir resultados consistentes com os interesses dos contratantes, potencialmente causando expropriação dos minoritários. Tais resultados mostram a necessidade de aumentar tanto a qualidade técnica dos avaliadores como o nível de controle sobre possíveis conflitos de agência. Possíveis mitigadores incluem a aplicação de legislação e regulação mais rigorosas, com maior controle do processo de avaliação pelos minoritários, e exigências relacionadas à qualificação técnica dos avaliadores, às metodologias aplicadas e à responsabilização do avaliador. / This work is based on the analysis of 125 valuation reports (Laudos de Avaliação) issued in the context of tender offers (OPAs) in Brazil during the 2006-2017 period. Such offers had a total value of tens billions of Reais, with firms valued at over R$ 300 billion. The main objectives are (i) describe the valuation methodologies, (ii) identify errors, comparing the adopted methodologies with the theoretical framework and (iii) measure valuation biases. The results indicate (i) diversity of methodologies used by practitioners, possibly causing inconsistencies and biases in the results, (ii) existence of errors, some of them related to basic valuation concepts and (iii) valuation biases, especially in reports prepared by independent consulting firms, which tend to issue results consistent with the controlling shareholders\' interests, potentially causing expropriation of the minority shareholders. Such results indicate the need to improve the practitioners\' technical quality as well as the controls against agency conflicts. Possible mitigators include stricter legislation and regulation, with greater control by minority shareholders of the valuation process, and minimum requirements regarding the practioners\' technical qualification, acceptable methodologies and accountability of the report\'s issuer.

Using an APT model where global temperature shocks are a systematically priced factor, the risk premium is significant and positive. Evidence is provided that positive exposure to temperature shocks is related to increasing CO2 emissions by industry. The global impact on the cost of equity could be as high as 2.8% per year, implying a global GDP loss of $2.2 Trillion per year due to global temperature shocks.

No description available.

No description available.

This study examines how diversification changes firms' key characteristics, which consequently alter firms' value. The reason why I focus on this topic is because of the mixed findings in literature about the valuation effect of diversification. This study offers deeper insights to the influence of diversification on important valuation factors that are already identified in finance literature. Specifically, it examines if diversification affects firms' information asymmetry problem, firms' cost of capital and cash flow, and firms' production efficiency. The study looks at both the financial industry and non-financial industry and the chapters are arranged in the following order. Firstly, empirical studies show that investors do not value BHCs' pursuit of non-interest income generating activities and yet these activities have demonstrated a dramatic pace of growth in the recent decades. An interesting question is what factors drive the discontent of the investors with the diversification endeavors of the BHCs in non-interest income activities. The first chapter examines the subject from the view point of information opaqueness, which is unique in the banking industry in terms of its intensity. We propose that increased diversification into non-interest income activities deepens information asymmetry, making BHCs more opaque and curtailing their value, as a result. Two important results are obtained in support of this proposition. First, analysts' forecasts are less accurate and more dispersed for the BHCs with greater diversity of non-interest income activities, indicating that information asymmetry problem is more severe for these BHCs. Second, stock market reactions to earning announcements by these BHCs signaling new information to the market are larger, indicating that more information is revealed to the market by each announcement. These findings indicate that increased diversity of non-interest income activities is associated with more severe information asymmetry between insiders and outsiders and, hence, a lower valuation by shareholder. Secondly, since Lang and Stulz (1994) and Berger and Ofek (1995), corporate literature has taken the position that industrial diversification is associated with a firm value discount. However, the validity and the sources of the diversification discount are still highly debated. In particular, extant studies limit themselves to cash flow effects, totally overlooking the cost of capital as a factor determining firm value. Inspired by Lamont and Polk (2001), the second chapter examines how industrial and international diversification change the conglomerates' cost of capital (equity and debt), and thereby the firm value. Our empirical results, based on a sample of Russell 3000 firms over the 1998-2004 period, show that industrial (international) diversification is associated with a lower (higher) firm cost of capital. These findings also hold for firms fully financed with equity. In addition, international diversification is found to be associated with a lower operating cash flow while industrial diversification doesn't alter it. These results indicate that industrial (international) diversification is associated with firm value enhancement (destruction). Given the fact that the majority of the firms involved in industrial diversification also diversify internationally, failing to separate these two dimensions of diversification may result in mistakenly attributing the diversification discount to industrial diversification. Thirdly, financial conglomerates have been increasingly diversifying their business into banking, securities, and insurance activities, especially after the Gramm-Leach-Bliley Act (GLBA, 1999). The third chapter examines whether bank holding company (BHC) diversification is associated with improvement in production efficiency. By applying the data envelopment analysis (DEA), the Malmquist Index of productivity, and total factor productivity change as a decomposed factor of the index, are calculated for a sample of BHCs over the period 1997-2007. The following results are obtained. First, technical efficiency is negatively associated with activity diversification and the effect is primarily driven by BHCs that did not diversify through Section 20 subsidiaries before GLBA. Second, the degree of change in diversification over time does not affect the total factor productivity change but is negatively associated with technical efficiency change over time. This latter effect is also primarily shown on BHCs that did not have Section 20 subsidiaries before GLBA. Therefore, it can be concluded that diversification is on average associated with lower production efficiency of BHCs, especially those BHCs without first-mover advantage obtained through Section 20 subsidiaries. These chapters explores the possible channels through which diversification could alter firms' valuation. They contribute to the literature by offering further knowledge about the effect of diversification. / Business Administration