Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

MFIN704 Final Exam

Due: 2020-04-22

Time: 17:30

Instructions:

You are being asked to provide consulting services to a company to develop a new product offering. Currently the company has several investment vehicles which are all actively managed but are looking into developing a series of rule based passively managed funds. The rise of passively managed investment vehicles which offer lower management fees are drawing customers away. To compete in     this market segment, managers are looking at developing a series of funds that will have the investment decisions controlled algorithmically based on a series of rules set out by the portfolio manager. These rules can be anything from specific factor weighting to cash reserve requirements. During your initial meeting with the funds’ management you learn details about their operations, plus the managers have  raised some additional questions. You are to prepare a report to answer their questions and provide recommendations with rationale.

Currently, each of the portfolio managers store data locally in Excel within their respective teams. This is creating a large amount of duplication, requires complex links and daily manual data entry. Managers would like to know of a more efficient method to store and access data.

Portfolio managers mention their investors are highly concerned with value-at-risk, which is no different for these new products. Currently the company uses a parametric distribution to measure the 5% value  at risk. They like this method because it is simple to implement and easy to explain to investors. You have some concerns with this approach and suggest that this could be done using bootstrapping or Monte Carlo simulation versus the parametric method. The company wants you to outline the advantages and disadvantages of the three methods, plus provide the rational behind your recommendation. Does your recommendation change if the metric was the average value at risk at the  5% level, instead of just the 5% value-at-risk?

One of the managers is interested in some additional details about the Monte Carlo method, specifically how the modelling is done, if it is modelling the individual securities or the portfolio as a whole? Is it possible to model different economic states, and if so, how? Also, how confident are the results and is there anything that can be done to increase confidence in the results.

The first passively managed fund follows a simple mean-variance framework following Markowitz’s portfolio theory. The constraints include no short selling, limits on investments in a firm, limits on investments in an industry, and a requirement of cash reserves. Mangers would like details on how the system will calculate the optimum weights for this fund.

The second passively managed fund will follow a factor-investment strategy. The managers did not provide specifics but mentioned that they are using a non-linear function with factors as inputs and want to maximize the expected value with respect to these factors. Managers are concerned about verifying the result and have asked how they can be confident that it is an optimal solution. If managers were to introduce constraints to the investment function, how would this change the approach of finding the optimum?