A passive investment management model was developed via a critical literature review of portfolio methodologies. This model was developed based on the fundamental models originated by both Markowitz and Sharpe. The passive model was automated via the development of a computer programme that can be used to generate the required outputs as suggested by Markowitz and Sharpe. For this computer programme MATLAB is chosen and the model’s logic is designed and validated. The demonstration of the designed programme using securities traded is performed on Johannesburg Securities Exchange. The selected portfolio has been sub-categorised into six components with a total of twenty- seven shares. The shares were grouped into different components due to the investors’ preferences and investment time horizon. The results demonstrate that a test portfolio outperforms a risk- free money market instrument (the government R194 bond), but not the All Share Index for the period under consideration. This design concludes the reason for this is due in part to the use of the error term from Sharpe’s single index model. An investor following the framework proposed by this design may use this to determine the risk- return relationship for selected portfolios, and hopefully, a real return.

Different from the standard treatment discovery framework which is used for finding single treatments for a homogenous group of patients, personalized medicine involves finding therapies that are tailored to each individual in a heterogeneous group. In this paper, we propose a new semiparametric additive single-index model for estimating individualized treatment strategy. The model assumes a flexible and nonparametric link function for the interaction between treatment and predictive covariates. We estimate the rule via monotone B-splines and establish the asymptotic properties of the estimators. Both simulations and an real data application demonstrate that the proposed method has a competitive performance.

[pt] Muitos dos trabalhos em finanças, como os que envolvem modelos financeiros,concentram-se na busca de formas de rejeitar as suposições sobre as quais estes se baseiam. Contudo, uma questão importante, é verificar se um determinado modelo supera ou é superado pelas alternativas existentes.Assim foi feito nesta pesquisa, que tem como objetivo principal mostrar que o gerenciamento ativo de carteiras dos fundos de pensão, com todas as limitações constantes em sua legislação cria valor, se comparado ao gerenciamento passivo. Ou seja, o gerenciamento ativo supera o gerenciamento passivo de carteiras.Basicamente, neste trabalho são apresentadas as limitantes presentes na legislação dos fundos de pensão e metodologias para a construção de carteiras.A carteira passiva foi construída segundo os conceitos presentes no algoritmo de Elton, Gruber e Padberg.A carteira ativa foi construída segundo um processo proposto por Grinold e Kahn de transformar sinais / informações em alphas / previsões.Para a segunda etapa do processo de geração de uma carteira ativa foram utilizadas três técnicas de construção de carteiras: a metodologia das janelas; a metodologia da estratificação; e a metodologia de programação quadrática onde foi utilizado o programa AEGIS 3.0 da consultoria BARRA. Após a construção das carteiras uma comparação, entre ambas, valida o objetivo proposto. / [en] Many of the works in finance, as the ones that involves financial models, are concentrated in fetching the forms to reject the assumptions on which these are based.However, an important question is to verify if one specific model surpasses or is surpassed by the other alternatives. Thus, it was made in this work, which main objective is showing that the active pension funds portfolio management, with all those legislation restrictions, creates value when it was compared to the passive management. In other words, the active portfolio management surpasses the passive management. Basically, in this work, we present the restrictions of the pension funds legislation and the methodology of the portfolio construction.The passive portfolio was built according to the concepts presented in the Elton, Gruber and Padberg algorithm. The active portfolio was built according to the process considered by Grinold and Kahn to transform signs / information into alphas / forecasts. For the second step of the process of the portfolio construction, there are three generic classes of procedures that cover the vast majority of institutional portfolio management, that are used: Screens; Stratification; and Quadratic Programming, in which we used AEGIS 3.0 of BARRA consult. After the portfolio construction we match the results to validate the main objective.

Preview:     SEB Investment Strategy is the function in SEB that supports business units SEB      Private Banking and SEB Retail with investment philosophy and investment            process. The framework of SEB Investment Strategy encompasses to manage a     structured investment philosophy and process to produce a range of investment                    options and portfolios for different target groups. From January 2006 to October        2009 forty “Proposal for fund portfolios” were produced each containing         writing on market condition and expectations plus portfolio recommendations.        Each time four portfolios consisting of six mutual funds was recommended,                    Fund Portfolio 30, 50, 70 and 100. Fund Portfolio 30 (FP30) contained           30% equity fund and 70% fixed-income funds. By same reasoning FP50           contains 50/50 equity- and fixed-income funds, FP70, 70% equity funds and         30% fixed-income funds and FP100 only equity funds.   Purpose:      The aim of this work is to evaluate these SEB Investment Strategy recommended       portfolios for private SEB Retail clients from January 2006 to December 2009.    Evaluation is done by comparing the performance of recommended portfolios       with portfolios produced by applying Vasicek´s Technique and simplified   optimization technique.   Method:     To allow work with Vasicek´s Technique in which we are dependent on a market        portfolio, I have created an Index which includes SEB Mutual Funds and their         share of the Index is determined from each fund´s total assets in relation to the    sum of the total assets under management of all funds inclusive in the Index.   Index consists of 40 mutual funds 2002-2007 and 37 mutual funds 2008         and 2009. The total supply of funds has been reduced to the above numbers by             the following criteria:   Clients must be able to invest in funds through conventional SEB Fund Account. No initiation fees or sales charges. Minimum historical Net Asset Value prices (NAV-prices) from 2nd January 2002. Daily trading and at least 300 million SEK in assets under management. No Fund-in-Fund products. Only SEB or SEB Choice funds.   The closing daily NAV-prices (time series) of these funds have been obtained from seb.se/fonder from 2nd January 2002 to 28th December 2009. With prices daily returns are calculated and used for estimation of historical and average values of variables needed for computing forecasted Alphas and Betas according to Vasicek´s Technique. Mutual funds are then ranked with respect to excess return over forecasted Beta given risk free rate equal to Swedish government 1 month treasury-bill (SSVX1M) at time for optimisation. Top six ranked funds are included in the optimization process. The first optimized portfolio given actual T-bill is then compared to FP100 recommended by SEB Investment Strategy. In order to find optimized solutions to other recommended portfolios premiums are added to actual T-bill rate.

Cette thèse s'intéresse à deux modèles de régression semi-paramétrique permettant de contourner le problème classique du "fléau de la dimension" inhérent aux approches non-paramétriques usuelles. La première partie du travail concerne l'étude d'un modèle de régression dit partiellement linéaire ; le but est d'identifier les régresseurs qui composent la partie non-linéaire de la fonction de régression ainsi que d'estimer tous les paramètres du modèle. Pour ce faire nous définissons des quantités caractéristiques du modèle qui mesurent la linéarité des régresseurs puis nous développons un test du nombre de composantes non-linéaires basé sur cette mesure. La seconde partie porte sur l'étude d'un modèle dit à direction révélatrice unique et consiste à estimer, via des propriétés géométriques, l'axe du modèle et d'en déduire un test convergent et puissant sous une suite d'alternatives locales.

No description available.

The generalized linear model (GLM) is a popular model in many research areas. In the GLM, each outcome of the dependent variable is assumed to be generated from a particular distribution function in the exponential family. The mean of the distribution depends on the independent variables. The link function provides the relationship between the linear predictor and the mean of the distribution function. In this dissertation, two semiparametric extensions of the GLM will be developed. In the first part of this dissertation, we have proposed a new model, called a semiparametric generalized linear model with a log-concave random component (SGLM-L). In this model, the estimate of the distribution of the random component has a nonparametric form while the estimate of the systematic part has a parametric form. In the second part of this dissertation, we have proposed a model, called a generalized semiparametric single-index mixed model (GSSIMM). A nonparametric component with a single index is incorporated into the mean function in the generalized linear mixed model (GLMM) assuming that the random component is following a parametric distribution. In the first part of this dissertation, since most of the literature on the GLM deals with the parametric random component, we relax the parametric distribution assumption for the random component of the GLM and impose a log-concave constraint on the distribution. An iterative numerical algorithm for computing the estimators in the SGLM-L is developed. We construct a log-likelihood ratio test for inference. In the second part of this dissertation, we use a single index model to generalize the GLMM to have a linear combination of covariates enter the model via a nonparametric mean function, because the linear model in the GLMM is not complex enough to capture the underlying relationship between the response and its associated covariates. The marginal likelihood is approximated using the Laplace method. A penalized quasi-likelihood approach is proposed to estimate the nonparametric function and parameters including single-index coe±cients in the GSSIMM. We estimate variance components using marginal quasi-likelihood. Asymptotic properties of the estimators are developed using a similar idea by Yu (2008). A simulation example is carried out to compare the performance of the GSSIMM with that of the GLMM. We demonstrate the advantage of my approach using a study of the association between daily air pollutants and daily mortality adjusted for temperature and wind speed in various counties of North Carolina. / Ph. D.

Bakgrund: Aktieindex används världen över som placeringsalternativ, jämförelsemått inom portföljförvaltning och som underlag för portföljoptimering. Forskare har under senare tid ifrågasatt index viktade efter börsvärdet och alternativa viktningsmetoder för index har framtagits som substitut till det kapitalviktade indexet och prestationsjämförelser har gjorts. Studier har främst gjorts i USA och denna studie ämnar göra en liknande undersökning på den svenska marknaden.  Syfte: Syftet med vår studie är att undersöka alternativa viktningsmetoder på det svenska aktieindexet OMXS30 och dess historiska prestation under åren 1995-2011 i förhållande till det traditionellt kapitalviktade OMXS30. Syftet är vidare att analysera de alternativt viktade indexen som grund för portföljoptimering enligt Single-Index Model.  Genomförande: Fem alternativt viktade index konstrueras i studien där viktningen grundas på fundamentala värden, Sharpekvoter, standardavvikelse, likaviktning och handelsvolym och jämförs prestationsmässigt mot OMXS30. Indexen används sedan vid portföljoptimering enligt SIM där aktiers och portföljers karakteristika analyseras. Indexens prognostisering av betavärden utvärderas i studien för att urskilja om något index är mer träffsäkert gällande aktiens beta för nästkommande period.  Slutsats: Ett flertal av de konstruerade alternativa indexen genererar signifikant högre avkastning till lägre risk i den nedgångsperiod som analyserats varför dessa kan ses som en mer lönsam investering. Tendenser till högre avkastning för den totala perioden finns även om signifikanta skillnader inte föreligger. De alternativa indexen har föranlett skilda allokeringsbeslut vid portföljoptimeringen vilket har gett stora utslag i portföljernas förväntade prestation såväl som faktisk prestation efter optimeringen genomförts. / Background: Stock market indexes are widely used as investment strategies and as a benchmark when portfolios are being constructed and evaluated. Researchers have recently questioned the capital weighted index in favor of other available weighting schemes. By comparing alternative weighted indexes to the traditionally capital weighted index one has been made aware of the significantly lower risk adjusted performance for the capital weighted index.  Aim: Our aim is to investigate in alternative weighting schemes used on the Swedish index OMXS30 and evaluate the historical performance of these alternative indexes in comparison to the traditionally capital weighted index during the period 1995-2011. Furthermore, our objective is to analyze the effects alternative weighting schemes have on portfolio optimization through Single-Index Model.  Completion: To fulfill the purpose of this thesis, five alternative weighting schemes have been applied on the Swedish index OMXS30. The weights have been calculated on fundamental measures, Sharpe ratios, standard deviation, equally weighted and trade volume and they have been compared to the traditionally cap-weighted index. Furthermore, the constructed indexes will be used to optimize portfolios with Single Index Model to compare the portfolios characteristics when different indexes have been used.  Results/Findings: The majority of the alternative weighted indexes generate significantly higher returns in one of our analyzed periods and this was during a market recession. For the total analyzed period no statistical differences among the indexes could be determined even though differences in total return are made clear. The indexes had a big effect on the portfolio optimization in terms of different share allocation.

本研究主要是探討台灣股票市場 79 至 81 年各分類指數的報酬率，以比 較各產業間報酬的多寡與風險的大小。並介紹投資組合的觀念以降低非系 統風險，而系統風險的部分可由單一指數模式估計。再以主成分分析法， 找出影響八個分類指數報酬率的主要成分，得到了第一個主成分做單一指 數模式的迴歸估計，以比較與加權指數為因變數的估計結果。最後探討這 三年來分類指數的報酬率，是否具有隨機性以及分類指數的行為。

In an experiment to understand colon carcinogenesis, all animals were exposed to a carcinogen while half the animals were also exposed to radiation. Spatially, we measured the existence of aberrant crypt foci (ACF), namely morphologically changed colonic crypts that are known to be precursors of colon cancer development. The biological question of interest is whether the locations of these ACFs are spatially correlated: if so, this indicates that damage to the colon due to carcinogens and radiation is localized. Statistically, the data take the form of binary outcomes (corresponding to the existence of an ACF) on a regular grid. We develop score??type methods based upon the Matern and conditionally autoregression (CAR) correlation models to test for the spatial correlation in such data, while allowing for nonstationarity. Because of a technical peculiarity of the score??type test, we also develop robust versions of the method. The methods are compared to a generalization of Moran??s test for continuous outcomes, and are shown via simulation to have the potential for increased power. When applied to our data, the methods indicate the existence of spatial correlation, and hence indicate localization of damage. Assuming that there are correlations in the locations of the ACF, the questions are how great are these correlations, and whether the correlation structures di?er when an animal is exposed to radiation. To understand the extent of the correlation, we cast the problem as a spatial binary regression, where binary responses arise from an underlying Gaussian latent process. We model these marginal probabilities of ACF semiparametrically, using ?xed-knot penalized regression splines and single-index models. We ?t the models using pairwise pseudolikelihood methods. Assuming that the underlying latent process is strongly mixing, known to be the case for many Gaussian processes, we prove asymptotic normality of the methods. The penalized regression splines have penalty parameters that must converge to zero asymptotically: we derive rates for these parameters that do and do not lead to an asymptotic bias, and we derive the optimal rate of convergence for them. Finally, we apply the methods to the data from our experiment.