none

[pt] Este estudo objetiva investigar se existe uma relação entre posições agregadas de fundos de investimento brasileiros e o desempenho passado, presente e futuro das ações. Utilizando dados mensais das carteiras de 30.416 fundos entre 2006 e 2018 e dados de mercado de 84 ações, diversas regressões com dados em painel foram feitas para testar a correlação entre retornos totais das ações e o nível e a mudança de posições compradas e vendidas. Além disso, os dados foram utilizados para testar uma estratégia long-short de investimento segundo a qual as ações que ocupam posições altas nos fundos são compradas e as ações com baixas posições são vendidas. / [en] This study aims to investigate whether there is a relationship between aggregate Brazilian mutual fund positions and past, concurrent, and future stock performance. Using monthly fund portfolio data for 30.416 funds between 2006 and 2018 and market data for 84 stocks, several panel data regressions were ran to test the correlation between stock total returns and both the level and change in mutual fund aggregate long and short positions. In addition, the data was used in testing a long-short investing strategy in which stocks with high fund ownership indicators are purchased and stocks with low fund ownership indicators are sold.

<p>  </p> <p>In the first chapter, we developed a dynamic equilibrium model of multiple stocks with extrapolators under the framework of Barberis, Greenwood, Jin, and Shleifer (2015a). Our model builds on the fact that extrapolative investors assign different relative weights of recent versus distant past return when forming their beliefs. We find that stock price increases in its own past performance measure, and is additionally associated with the past performance measure of the other stock if their dividends are correlated. The stock with higher relative weight have higher stock price, higher stock volatility, and lower risk premium. Both the own stock’s and the other stock’s past performance measure negatively predicts future stock price changes. </p> <p><br></p> <p>The second chapter includes Adem Atmaz, Stefano Cassella, and Huseyin Gulen as co-authors. In this chapter, we document considerable cross-sectional variation in survey expectations about aggregate stock market returns. While most investors are extrapolators who expect higher returns after a good market performance, some are contrarians who expect lower returns after a good performance. More notably, compared to extrapolators, contrarians have less persistent expectations that are corrected more quickly. Accordingly, we develop a dynamic equilibrium model accounting for these differences in expectations and find that the equilibrium stock price exhibits short-term momentum and long-term reversal as in the data. Furthermore, we test the key predictions of the model linking the shortterm momentum to observable differences between extrapolators and contrarians and find supportive evidence for our mechanism. </p> <p><br></p> <p>The third chapter includes Adem Atmaz and Suleyman Basak as co-authors. In this chapter, we develop a dynamic model of costly stock short-selling and lending market and obtain implications simultaneously supporting many empirical regularities. In our model, investors’ belief disagreement leads to lenders and short-sellers, who pay shorting fees to borrow stocks from lenders. Our main novel results are as follows. Short interest predicts future stock returns negatively and has a stronger predictive power than the corresponding dividend-price ratio. Higher short-selling risk can be associated with lower stock returns and less short-selling activity. Stock volatility is increased under costly short-selling. An application to the GameStop episode yields implications consistent with observed patterns. </p>

The hefty price increases in heavily shorted stocks in the beginning of 2021 indicates that short interest might be an underrated yet important key figure for investors when deciding on whether to take on an investment strategy or not. Most stock exchanges release information regarding the short interest only once a month leaving investors having to make decisions on outdated information. No previous research was found on whether there exists a linear relationship between some variables to estimate the short interest. Through the collection of mostly financial data and some regression analysis, a large­sample linear regression model was constructed. Although the problem complexity may seem of higher degree than linear, this study suggest that there exists a linear relationship between the variables studied when estimating short interest. These results suggest that one can use the mathematical model presented in this study for estimation and also get a better understanding of which underlying variables that impact short interest the most. The final model consisted of 6 variables, number of outstanding shares, average trading volume, stock price, volatility, equity andaverage weighted cost of capital. / De kraftiga prisökningarna på kraftigt blankade aktier i början av 2021 indikerar att mängden blankade aktier kan vara ett underskattat men ändå viktigt nyckeltal för investerare när de beslutar om den investeringsstrategi som skall användas. De flesta börser släpper information om mängden blankningar en gång i månaden, vilket gör att investerare måste fatta beslut på föråldrad information. Ingen tidigare forskning hittades på om det finns ett linjärt samband mellan vissa variabler för att uppskatta antalet blankningar. Genom insamlingen av mestadels finansiell data och med hjälp av regressionsanalys konstruerades en regressionsmodell baserat på ett stort urval. Även om problemkomplexiteten kan verka av högre grad än linjär tyder denna studie på att det finns ett linjärt samband mellan de studerade variablerna. Dessa resultat tyder på att man kan använda den matematiska modellen som presenteras i denna studie för uppskattning och också få en bättre förståelse för vilka underliggande variabler som påverkar mest. Den slutliga modellen bestod av 6 variabler, antal utestående aktier, genomsnittlig handelsvolym, aktiekurs, volatilitet, eget kapital och genomsnittlig viktad kapitalkostnad.