Events

The world of asset management is a notoriously complex and competitive one. Since the rise of computing the finance world has seen a fundamental shift towards algorithms and mathematics. Our group has worked on a set of software tools that provide solutions to both the business and analytic aspects of finance management.

We have designed a web application for the Oregon State Investment Group (OSIG). OSIG manages around $3 million for the university spread across a variety of different assets. Our work is broken into two different tasks; tools that help expedite the research process, provide a central platform for the group’s information, and automate common tasks. Our other task is designing a deep learning model that helps the groups portfolio managers decide the best allocations of assets underlying the portfolio. 

Our deep learning work is based on combining the insights of classical economic theory with the computational power of deep learning. From the economic side of things, we are measuring portfolio performance using the Sharpe Ratio, which describes the expected return of a portfolio while taking into account the risk of the investments. For the deep learning model, we are using a LSTM architecture that analyzes a variety of features of the underlying assets (daily returns, financials, etc.) and finds a set of asset allocations that maximize the Sharpe Ratio.  

Because there is real money on the line, we have taken care to make our models as rigorous as possible. We have created a test suite that allows us to get a grasp on the expected performance using real financial data. Thus far our model is doing excellent. We have applied it to the groups large cap portfolio, which comprises over 20 assets, and have seen substantial returns with the model out competing both the S&P and the OSIGS current allocation strategy. 

OSIG members will be able to visit the web application through osig.oregonstate.edu. The home page for this site will display a calendar for the groups key events. The calendar will include dates for upcoming and past stock pitches. It will also include the results for those pitches on whether or not the group decided to buy, hold, or sell the stock that was pitched.

Also made available to members of OSIG is a stock research page for insights into a company's financials. Members will be able to use a stock's ticker symbol to be able to pull data from the financial API, IEX Cloud, like income statements, cash flows, and balance sheets for the past 4 years. Users will also be able to pull key statistics and metrics related to a company like beta, Enterprise Value, EBITDA, PEG ratio, profit margin, P/E, P/S, etc.

Excel workbooks are a pivotal element to OSIG's stock evaluations and pitches. Another aim of the stock research page is to automate this workbook building process. Members will be able to enter a stock's ticker and 5 other competitors’ tickers to generate a zip folder containing the original stocks 10-Ks and 10-Qs pulled from the SEC.