This study constructs a quantitative stock selection model across multiple sectors with the application of the Bayesian method. It employees factors from the Taiwan stock market which could explain stock returns. Under this structure, each sector that has different significant factors is allowed to be imported into sub models. The factors are calculated into alpha scores and used to do stock selection. Therefore, the demonstration of both intra and inter-sector alpha scores into sector-specific integration alpha scores is an important concept in this study. Furthermore, an enhanced index fund is built based on the model and related to the benchmark to illustrate the power of this model. Once the contents of a portfolio are decided, this model could provide stock selection criterion based on the predictive power of stock return. Finally, the results demonstrate that this model is practical and flexible for local stock portfolio analysis.

"Enhanced index fund" is an investment strategy, combining active and passive management elements, for index tracking and return enhancing through disciplined market timing, stock selection and leverage activities. Though enhanced index funds have been well developed globally, there is only one enhanced index fund in Taiwan - "Polaris/P-Shares Taiwan Dividend+ ETF". Taiwan's stock market falls between weak form and semi-strong form efficiency. With the growth of Taiwan's mutual fund industry size, the enhanced index funds have very good chance to become the main investment instruments of institutional investors, index investors, and pensions. This study attempts to build enhanced index funds, then analyzes the performance and checks the feasibility of launching such products in Taiwan. In this study, we select "TSEC Taiwan 50 index (TW50)" as the benchmark index. The innovative "Multi-section Adjustment Model" divides the original weights of constituent stocks into two sections. Each section is adjusted through parameters. The "multi-factor model section" is responsible for the delivery of enhanced return, while the "cash dividend yield section" is used to provide excess cash dividend yield. The investment target is set for less than 1.5 percent tracking error, at least 1 percent tracking difference, and higher cash dividend yield than the benchmark. Building methodology can be divided into "fixed parameter model" and "floating parameters model" according to its update frequency. Empirical studies show that: (1) The enhanced index fund built from the "fixed parameter model" not only exhibits risk slightly lower than the benchmark, but also enjoys higher return. (2) In the short-term, the performance of the enhanced index fund built from "floating parameters model" is difficult to predict; in the long-term, however, the risk is lower and the return is higher than TW50. The cumulative return from the "fixed parameter model" is higher than the "floating parameters model" by about 2 percent. (3) The effectiveness of the parameters used to control the optimal weight distribution is decreasing over time, so it is necessary to update parameters regularly. (4) Raising "enhancement multiplier" will cause higher tracking error, but also bring higher tracking difference. This result proves that "multi-factor model section" works nicely and has its contribution. (5) As the "section allotment" and/or "fixed rate" getting lower, there will be more and more weights distributed to the cash dividend yield level, resulting in higher cash dividend yield. It means the "cash dividend yield section" has its merit as well. (6) Regular parameter updates to the "floating parameters model" helps to reduce the tracking error and, at the same time, maintain positive tracking difference. Considering the perpetual life of real world funds, "floating parameters model" should be a better building methodology. "Multi-Section Adjustment Model" has following advantages: (1) Its concept is intuitive and easy to use. (2) Sections can be customized based on investment objectives. (3) It is easy to analyze the impacts and trade-off among the parameters.

Many enhanced index funds are based on a quantitative model to control active risk and to acquire active return. In this thesis we first construct a multiple-factor model (MFM) and then use statistical methods to evaluate the significance and stability of factor explanatory power. Significant and stable factors are utilized to fine tune weights of T50 index fund portfolio by an intuitive weight allocation model to achieve the effect of return enhancement. Empirical studies show that the multiple-factor model can explain the excess stock return effectively; the average R-Square of multiple-factor model reaches 49%. After analyzing the sensitivity of parameter of enhanced index weight allocation, the study finds that the original weight retention rate has linear relationship with active return and active risk of the T50 index fund. Adjusting the retention rate allows us to control the active return and active risk of T50 index fund. Furthermore, adjusting the original weight retention rate according to the Adj-R2 of multiple-risk factor model can effectively improve the stability of active return. The study finds also that the expected rates of return which are calculated by multiple-risk factor model could not differentiate among future performance of the first your guarantee portfolios. Thus, the study adjusts the range of weight allocation to T50 constituent stocks with higher and lower expected return rates. The result shows that this adjustment increased the IR of the enhanced index funds.

The first essay examines cross-sectional differences between money market mutual funds (MMMFs), in the context of the sponsoring fund family. While extant studies have shown that fund family characteristics impact the management of open-end equity mutual funds, results of this study’s analysis find that fund family characteristics also affect the management of MMMF assets, contributing to differences in the maturity of the fund’s holdings, expenses, and realized returns. I find that an MMMF is not simply a transitional account with a short-term low-risk investment objective, but rather, a critical role player within the fund family. Differences in maturity, yield, and expenses in MMMFs can be explained by family-specific characteristics, including diversification and cash management strategies at the family level. The second essay examines implications of substitutability of two similar financial assets: conventional index mutual funds and exchange traded funds (ETFs). I seek to explain the coexistence of these fund types, since both offer a claim on the same underlying index return process, but have different organizational structures. This study compares conventional open-end index funds with matched ETFs on various underlying indexes. Aggregate flows are used to detect substitution and clientele effects. I show that conventional funds and ETFs are substitutes, while ETFs have smaller tracking errors and lower fund expenses. However, I find that these fund types are not perfect substitutes, and their coexistence can be explained by a clientele effect that segregates them into different market niches.

Optimization methods nowadays play a very important role in financial decisions such as portfolio managements, construction of index funds and pension funds.  This Master Thesis is devoted to the problem of an index fund construction. The problem is represented as a linear optimization problem of re-balancing the portfolio at minimum cost and solved using the Primal-Dual interior point method. The index fund is constructed using ten companies from the Dow Jones Industrial Average Index (DJIA). The Primal-Dual interior point method was first implemented in Matlab and later on in Java.

指數基金係提供投資者追隨市場指數成長的投資工具，且投資者僅需考量市場風險即可，其建構方式有完全複製法、分層法、抽樣法、及最佳化法。本論文使用目標規劃模型建構指數基金，此法可歸類為最佳化法。由於模型中每種股票的投資數量設為整數變數，加上控制股票種類數量的0-1變數，因此所建構的目標規劃模型為混合型整數線性規劃問題。此問題在大尺度模型時往往無法求得其最佳解，我們研究此模型的結構提出一組縮小解集合空間的合理不等式，應用切面法加入必需的不等式後再根據本模型的對偶性質發展出有效率的啟發式演算法，最後將此模型及演算法應用在模擬台灣發行量加權股價指數。 / Index fund is an investment tool which tracks a stock-market index and thus is associated with market risk only. Its attraction to investors is low investment risk and low administrative expenses. Four different approaches to index fund construction can be classified as full replication, stratification, sampling, and optimizing respectively. In this thesis, we construct an index fund via the goal programming model with the optimizing approach. The model can be formulated as a mixed integer linear programming. The exact optimal solution can not be obtained when the model becomes large. We then develop a valid inequality and use this valid inequality to develop a cutting plane method. We also propose an efficient heuristic by adopting the dual property. Finally, an empirical study applying to the Taiwan Stock Exchange Capitalization Weighted Stock Index is given to show the efficiency of the algorithm.

通常已建立的指數基金，經過一段時間後其追蹤指數的效能已經無法滿足初期建購時的要求，此時管理者便面臨指數基金投資組合的調整問題。本論文融合建構指數基金的方法及最小化交易成本的概念，提出一個新的混合整數線性規劃模型以調整指數基金投資組合。模型亦考慮實務中交易成本、最小交易單位及批量、固定交易費用比率、以及資產總類數等限制。因此，模型包含整數變數及二元變數，求解也較為困難許多。本論文以啟發式演算法增進求解的效率，並以台灣50指數的相關資料做為實證研究的對象。 / The efficiency of index-tracking in index fund, which has been built, has usually been incapable to meet the needs after a period of time. In this moment, the managers have to face with the problems of the adjusting for index fund portfolio. In this paper, we integrate the methods of constructing index fund and the concepts of minimum transaction cost with it, and propose a new mixed integer linear program model to adjust the index fund portfolio. Moreover, the model also considers some limitations, such as the transaction costs, minimum transaction units and lots, fixed proportional transaction rates, and cardinality constraint in practical operating. For this reason, a set of integer variables and binary variables are introduced. However, they increase the computational complexity in model solution. Due to the difficulty of the MILP problem, a heuristic algorithm has been developed for the solution. The computational results are presented by applying the model to the Taiwan 50 index.

Syfte: Syftet med studien är att undersöka hur aktivt förvaltade fonder presterar jämfört med indexfonder, när avkastningen har justerats för förvaltningsavgiften. Indexfonden representeras av ett jämförelseindex och studien omfattar en tioårsperiod, 2000-2009. Det faktum att en apa vann aktie SM 1993, framför professionella placerare, visar att aktiekurser är slumpmässiga. Varför ska en investerare då lita på att en förvaltare är bättre på att utvärdera marknaden och dess placeringsmöjligheter än andra? Metod: En kvantitativ metod har använts i uppsatsen, där data har erhållits från Morningstar och SIX Telekurs. Det insamlade materialet har bearbetats i Microsoft Excel för att beräkna fondernas avkastning och prestationsmått. Resultatet har redovisats i tabeller och diagram i empirikapitlet, för att sedan analyseras och jämföras med den teoretiska referensramen. Resultat & slutsats: Endast en av tio aktivt förvaltade fonder överträffar index, därmed dras slutsatsen att indexfonder är ett bättre investeringsalternativ än aktivt förvaltade fonder. Resultatet visar därmed att den högre förvaltningsavgiften som fondbolagen kräver från sina kunder inte är berättigat. Förslag till fortsatt forskning: Baserat på de resultat som kommit fram i uppsatsen, voredet intressant att genomföra en kvalitativ studie där fondförvaltarens åsikter är i fokus. Hur motiveras den höga förvaltningsavgiften, när de inte överträffar index? / Aim: The fact that a monkey won the Swedish Championship in stocks in 1993, ahead of professional investors, shows that stock prices are random. Why should an investor trust that a professional manager is better on evaluating the market and its investment opportunities than others? The purpose with this thesis is to investigate how active managed funds perform compared to index funds, after subtraction of the management fee. The index fund is represented by a “comparison index” and the research covers a period of ten years, between 2000 and 2009. Method: A quantitative method has been used in this study, where the information has been received from Morningstar and SIX Telekurs. Microsoft Excel has been used to process the collected data in order to calculate the expected return and the risk measures. The result is presented in diagrams and charts in order to analyse and compare it with the theory. Result & Conclusions: Only one out of ten active managed funds outperform index,therefore draws the conclusion that index funds is a better investment option than active managed funds. The result shows that the higher management fee that stock exchange companies claims is not appropriate. Suggestions for future research: Based on the results in this thesis, it would be interesting to do a qualitative research where the focus is on the fund managers’ opinions. How can they motivate the high management fee, when they don’t outperform index?

投資者面對龐大的股票市場，希望選取少量的股票使如指數基金般達到追蹤市場的效果，傳統的作法是使用指數追蹤的技術，建立一組投資組合使得報酬率與市場報酬率的績效相同。本論文除了最小化指數追蹤的下方追蹤誤差，還加入beta值的限制式，利用不同的beta值建立一組與市場成長趨勢相當或可能超越市場績效的投資組合。論文中使用提出之模型針對不同範圍的beta值進行研究，分析比較標的指數與建立的投資組合之績效表現。最後以台灣股票市場作為實證研究對象，實證結果顯示本論文模型所建立之投資組合在三個月內與標的指數表現相當，並在三個月後超越標的指數。 關鍵字：beta值、指數追蹤、下方追蹤風險、指數基金

建立指數基金時，通常是利用追蹤指數的技巧，選取少量的股票建構指數基金使得報酬率與標的指數(benchmark index)報酬率同步的投資組合。如果能建立包含少量股票的投資組合，就可達到指數追蹤的效果，那麼也能利用少量的股票建立績效可以超越指數基金的投資組合。本論文利用建構指數基金的方法以及大中取小的概念，挑選出一個績效可以超越標的指數的投資組合。本論文提出的模型亦考慮實務上交易所需的各項成本、整數交易單位與資產總類數等限制。因此，模型包含整數變數與二元變數。最後以台灣加權股價指數的相關資料做為實證研究的對象，實證結果顯示本論文提出的模型所建立的投資組合超越標的指數的績效平均年化報酬率25%。 / Setting up an index fund usually uses techniques of index-tracking that choosing few stocks forming a portfolio to obtain the same return rate as the benchmark index. Similarly we can use the same concept to set up a portfolio such that the performance is better than index’s. In this thesis we use index-tracking methods and minimax rule to obtain a portfolio which outperforms the benchmark index. In the proposed mathematical model we will consider the transaction costs, integer trading unit volume, and the total number of assets in the portfolio. Therefore the resulting model is a mixed integer nonlinear programming including integer variables and binary variables. Finally, the empirical study will be performed by using the data from the Taiwan stock market to verify the performance of our model. The empirical study shows that the portfolios created by our models outperform the benchmark index up to 25% in average.