Existing studies document that institutional herding has a stabilizing effect on stock prices, as stock returns are positively correlated with herding over one- to three-quarter horizons. The literature also shows that short-term institutions are better informed than long-term institutions. Motivated by heterogeneity in the level of informed trading between short-term and long-term institutions, this study disentangles the herding effect of short-term and long-term institutions on stock prices. Our results show that herding by short-term institutions promotes price discovery. In contrast, herding by long-term institutions drives stock prices away from fundamentals. Taken together, our findings suggest that the stabilizing effect documented in the existing literature is mainly driven by short-term institutions, and herding by long-term institutions has a destabilizing effect on stock prices.

Magíster en Economía Aplicada / Este trabajo analiza el efecto de la industria de fondos mutuos en la volatilidad del mercado accionario chileno, midiendo el impacto de la propiedad accionaria por parte de los fondos mutuos sobre la volatilidad mensual de los retornos. Utilizando datos de panel con efectos fijos a nivel de firma y tiempo, se obtiene que existe una relación causal negativa entre la inversión de fondos mutuos y la volatilidad accionaria. Este resultado es robusto ante distintas especificaciones, incluyendo variables de control y utilizando un instrumento para controlar por posible endogeneidad. Además se obtiene que el efecto es heterogéneo dependiendo del tamaño de las firmas, siendo más acentuado en acciones con mayor capitalización de mercado. Finalmente se calcula un estadístico para medir el nivel de herding en la industria, encontrando evidencia de que esta es dominante en acciones con mayor capitalización de mercado, por lo que se concluye que este tipo de comportamiento se debe al actuar informado de los fondos mutuos frente a señales correladas del mercado.

Information asymmetry literature has developed models that explain the relation between uninformed traders and informed traders. In general, these models have shown that first, information asymmetry is a driving force for investor buying and selling behavior. Second, the trades of informed investors reveal some of the information they possess suggesting that the trades of informed investors are informative to market makers. Third, when information about a stock enters the market, the characteristics of the firm can change, e.g., a better information environment reduces the cost of capital (Admati, 1985; Easley and O‟Hara, 2004; Wang, 1993). In this study, I apply information asymmetry theory to explore the trading behavior of active equity mutual fund managers and their role as facilitators of information. In the first essay, I study the information environment of firms mutual funds choose to add to their holdings and how it changes after the inclusion. I identify all new additions to the mutual fund holdings universe from 2002 to 2015 and compare them to the available universe of firms not yet owned by mutual funds. I find that active equity mutual fund managers behave as informed investors and prefer to buy stocks with more opaque information environments i.e., firms with larger spreads, lower trading volume, smaller firms with more growth opportunities, and firms that tend to use more accruals. Fund managers also show a preference for firms that have less analyst following, those in which analysts are less likely to agree on their EPS estimates, and firms in which analysts are more likely to err in their predictions. In other words, mutual fund managers prefer firms that are more likely to be mispriced. Once the funds include the firms, I document a strong improvement in their information environment. Firms attract more analyst coverage, reduce its use of accruals, produce more guidance, increase their market cap, and show increased turnover. The second essay focuses on the herding behavior of mutual funds. The study is the first to document the herding of mutual fund managers after creation of toehold positions by portfolio managers. I use a hand-collected dataset consisting of all toehold acquisitions reported to the SEC from 1995 to 2015 to document a strong herding reaction of active equity mutual funds after toehold announcements. This herding reaction is several times stronger than other mutual fund herding events reported by previous literature. I also document that the strength of the herding reaction varies depending on the identity of the filer or the characteristics of the firm acquired. The herding reaction is stronger for toehold announcements of firms with a smaller market capitalization, better growth opportunities, and those that are more illiquid. I also find that the herding reaction is weaker after the filings of hedge fund managers. My results support the informational herding cascade hypothesis. / Includes bibliography. / Dissertation (Ph.D.)--Florida Atlantic University, 2018. / FAU Electronic Theses and Dissertations Collection

This dissertation contains three chapters in financial economics that theoretically and empirically examine how individuals' investment decisions explain aggregate behavior. The first chapter examines how reputational herding between fund managers depends on the fee structure, fund manager evaluation metric, market efficiency, and density of talented fund managers. Results show there are more equilibria involving herding between fund managers when net fund balance growth depends on reputation of talent rather than fund return. These inefficient equilibria are removed when the ratio of the performance fee rate to management fee rate is larger than calculated thresholds that depend on market efficiency and the density of talented fund managers. In the absence of performance fees, lower predictability of investment returns and a higher density of talented fund managers increase the desire for fund managers to deviate from efficient equilibria. The model also shows having fund managers compete against each other induces herding when net fund balance growth depends on fund returns, but removes herding equilibria when net fund balance growth depends on reputation of talent. The second chapter determines what herding networks exist between institutional investors and how herding depends on stock market volatility, degree of portfolio changes, and stock size. Using quarterly holding data from 2000-2010, I find stronger herding networks between similar types of institutions compared to institutions in the same metropolitan area. Furthermore, the herding network between similar types of institutions exists across metropolitan areas. Results show institutions herd more when making major portfolio changes than when making minor portfolio changes. The difference in herding between the two types of portfolio changes is greatest for small cap stocks which exhibit the highest levels of herding under both types of portfolio changes. The relationship between market volatility and herding by institutions is also examined and found not to have a strong correlation using quarterly holdings data. The third chapter answers the question, "Can reasonable wind energy plant cost reductions or efficiency improvements precipitate immediate investment in wind energy in the absence of renewable energy Production Tax Credits?" I analyze a single entity considering an irreversible investment under uncertainty in wind power energy. The investor's decision to invest is dependent on investment cost, energy production efficiency, government policy, current price of electricity, and beliefs on future electricity prices. The results show that even with substantial cost reductions and efficiency improvements, Production Tax Credits are still needed to encourage immediate investment. / text

Em finanças o comportamento de manada (herd behavior) é comumente associado a um importante elemento do comportamento dos investidores nos mercados financeiros, em especial durante períodos de crises financeiras. Muitas pesquisas nesta área procuraram compreender os motivadores teóricos desta anomalia de mercado e um grande número de experimentos procurou identificar e quantificar a presença do herding em mercados desenvolvidos e emergentes. No entanto, por se tratar de uma variável não observável diretamente, a dificuldade na sua mensuração é grande desafio das pesquisas nesta área. Neste estudo pretende-se verificar a presença deste viés comportamental e avaliar a dinâmica desta variável no mercado brasileiro de ações através do modelo proposto por Hwang e Salmon (2001 e 2004), mensurando-se o efeito manada no mercado brasileiro de ações em relação ao índice de mercado através da medida de dispersão transversal dos betas (beta herding) das ações no período entre janeiro de 1995 e maio de 2012. Nesta pesquisa os betas das ações em relação ao índice de mercado foram obtidos utilizando-se as séries de excessos dos retornos diários das ações sobre a taxa DI-Cetip over das ações negociadas na BM&FBovespa, utilizando-se o modelo de mercado de Fama e French (1993) de três fatores, com as séries filtradas e suavizadas por um conjunto de equações dinâmicas de espaço-estado (Filtro de Kalman). O entendimento da dinâmica desta variável ajudaria a explicar melhor o comportamento do investidor em diferentes condições de mercado e o modelo proposto nesta pesquisa parece contribuir neste sentido. Semelhante aos achados de Hwang e Salmon (2001), Amirat e Bouri (2009a) e Hachicha (2010), os resultados encontrados neste trabalho para o mercado brasileiro de ações neste período sugerem que existe um nível base ou estacionário de herding no mercado, independente das condições do mercado. Também se observou a existência de uma componente de feedback herding explicada pela ação dominante anterior ou do nível anterior de herding. Outro achado deste trabalho foi uma diferença entre o senso comum de que o nível de herding aumentaria durante as crises financeiras, verificando-se na amostra analisada fenômeno contrário: uma redução dos níveis de herding durante períodos marcados por crises financeiras. / In finance, herd behavior is a commonly bias associated with an important element of investor\'s behavior in financial markets, particularly during periods of financial crises. Researches in this area try to explain theoretical motivations for this market bias and a large number of experiments try to identify the presence of and measure herding in developed and emerging markets. However, as this is a variable that cannot be directly observed, difficulties faced to measure it comprise a major challenge for researches in this area. The purpose of this study is to verify the presence of this behavioral bias and evaluate this variable dynamics in the Brazilian stock market, using model proposed by Hwang and Salmon (2001 and 2004). Herding is measured in the Brazilian stock market in relation to market index through evaluation of cross-sectional dispersion of equity betas (beta herding) in the period from January 1995 to May 2012. Equity betas, as compared to the market index, were obtained through series of daily stock excess returns over DI-Cetip rate for shares traded at BM&FBovespa, using Fama and French (1993) three-factor model, with series being filtered and smoothed by Kalman filter state-space dynamic equations. Understanding this variable dynamics would help to explain investor\'s behavior under different market situations, and the model proposed in this study seems to contribute. Similar to findings of Hwang and Salmon (2001), Amirat and Bouri (2009a) and Hachicha (2010), Brazilian stock market results found in this study suggest that there is a base or stationary herding in the market, regardless of market conditions. Existence of a feedback herding component, explained by prior dominant action or herding level was also verified. This study also denies common sense understanding that herding level increases during financial crises; in fact, opposite phenomenon was verified: a reduction in herding levels during financial crisis periods.

Em finanças o comportamento de manada (herd behavior) é comumente associado a um importante elemento do comportamento dos investidores nos mercados financeiros, em especial durante períodos de crises financeiras. Muitas pesquisas nesta área procuraram compreender os motivadores teóricos desta anomalia de mercado e um grande número de experimentos procurou identificar e quantificar a presença do herding em mercados desenvolvidos e emergentes. No entanto, por se tratar de uma variável não observável diretamente, a dificuldade na sua mensuração é grande desafio das pesquisas nesta área. Neste estudo pretende-se verificar a presença deste viés comportamental e avaliar a dinâmica desta variável no mercado brasileiro de ações através do modelo proposto por Hwang e Salmon (2001 e 2004), mensurando-se o efeito manada no mercado brasileiro de ações em relação ao índice de mercado através da medida de dispersão transversal dos betas (beta herding) das ações no período entre janeiro de 1995 e maio de 2012. Nesta pesquisa os betas das ações em relação ao índice de mercado foram obtidos utilizando-se as séries de excessos dos retornos diários das ações sobre a taxa DI-Cetip over das ações negociadas na BM&FBovespa, utilizando-se o modelo de mercado de Fama e French (1993) de três fatores, com as séries filtradas e suavizadas por um conjunto de equações dinâmicas de espaço-estado (Filtro de Kalman). O entendimento da dinâmica desta variável ajudaria a explicar melhor o comportamento do investidor em diferentes condições de mercado e o modelo proposto nesta pesquisa parece contribuir neste sentido. Semelhante aos achados de Hwang e Salmon (2001), Amirat e Bouri (2009a) e Hachicha (2010), os resultados encontrados neste trabalho para o mercado brasileiro de ações neste período sugerem que existe um nível base ou estacionário de herding no mercado, independente das condições do mercado. Também se observou a existência de uma componente de feedback herding explicada pela ação dominante anterior ou do nível anterior de herding. Outro achado deste trabalho foi uma diferença entre o senso comum de que o nível de herding aumentaria durante as crises financeiras, verificando-se na amostra analisada fenômeno contrário: uma redução dos níveis de herding durante períodos marcados por crises financeiras. / In finance, herd behavior is a commonly bias associated with an important element of investor\'s behavior in financial markets, particularly during periods of financial crises. Researches in this area try to explain theoretical motivations for this market bias and a large number of experiments try to identify the presence of and measure herding in developed and emerging markets. However, as this is a variable that cannot be directly observed, difficulties faced to measure it comprise a major challenge for researches in this area. The purpose of this study is to verify the presence of this behavioral bias and evaluate this variable dynamics in the Brazilian stock market, using model proposed by Hwang and Salmon (2001 and 2004). Herding is measured in the Brazilian stock market in relation to market index through evaluation of cross-sectional dispersion of equity betas (beta herding) in the period from January 1995 to May 2012. Equity betas, as compared to the market index, were obtained through series of daily stock excess returns over DI-Cetip rate for shares traded at BM&FBovespa, using Fama and French (1993) three-factor model, with series being filtered and smoothed by Kalman filter state-space dynamic equations. Understanding this variable dynamics would help to explain investor\'s behavior under different market situations, and the model proposed in this study seems to contribute. Similar to findings of Hwang and Salmon (2001), Amirat and Bouri (2009a) and Hachicha (2010), Brazilian stock market results found in this study suggest that there is a base or stationary herding in the market, regardless of market conditions. Existence of a feedback herding component, explained by prior dominant action or herding level was also verified. This study also denies common sense understanding that herding level increases during financial crises; in fact, opposite phenomenon was verified: a reduction in herding levels during financial crisis periods.

none

Thesis (M.A.)--Brigham Young University, Dept. of Religious Education, 2005. / Includes bibliographical references (p. 145-156).

My thesis focuses on three related research themes. First, to provide academic rigour to the assertions of the Nyberg Report (2011) that the Irish banks were characterised by both groupthink and herding in the lead up to the recent financial crisis. The Nyberg Report was commissioned by the Irish government to provide explanations for the causes of the Irish financial crisis. In particular my thesis explores whether the board of Anglo Irish Bank (Anglo) was characterised by groupthink tendencies and whether the other Irish banks looked to emulate Anglo's strategy. Second, my thesis will develop a theoretical model which identifies characteristics associated with the increased vulnerability of a board to a poor decision process . In particular the model will focus on the interplay between normative and informative influences on decision process and how these can and do interact with director skill levels. Five research questions are developed and my theoretical model of VPDP is applied to the Irish bank case study. Information will be gathered through a process of semi-structured interviews and an analysis of existing literature, official reports and annual financial statements. Third, a series of recommendations are made which are derived from the model itself and are intended to reduce the likelihood of boards pursuing a poor decision strategy and in increasing the likelihood of a robust boardroom challenge.

No description available.