Profile

For George “Tony” Robertson, retirement was not a time to sit back and rest on his laurels. Says the former Wall Street securities analyst, “I did all the things I always said I would, but then I got bored. I didn’t wish to retire mentally.” Motivated to “stay fresh” and intellectually challenged, Robertson enrolled in UMBC’s statistics master’s program two years ago.

Robertson began his career as a mechanical engineer for Sun Oil Company. Due a talent for solving problems of all sorts for his company, Robertson rose through the ranks. Says Robertson, “They’d ask me, ‘can you solve it?’ My automatic answer was ‘yes’. I wanted the chance to try.” Motivated by the challenge, Robinson taught himself whatever was necessary; he earned his MBA, passed the law boards, and taught himself methods of financial analysis. Wall Street was a natural transition for Robertson. He entered the world of investment banking as a trucking research analyst for the Alex Brown in Baltimore and then became a telecommunications financial analyst for Robertson Stephens in San Francisco. “Wall Street was a very stimulating environment,” says Robertson.

A topic right up his Wall Street alley, Robertson’ thesis work pertains to the complex and imperfect process of portfolio selection and optimization. In this day and age, statistical analysis is an invaluable tool for portfolio selection. Robertson’s research is ambitiously aimed at figuring out the best statistical methods for portfolio management—to determine those that best predict returns and that optimize the portfolio selection process.

“There are many techniques that can be used to make financial estimates. Knowing which techniques you should use and why is essential to making good estimates,” he says. Currently, Robertson is investigating the “state-of-the-art” statistical methods that are often used for portfolio selection today. “I want to understand the underlying assumptions of the tools,” he explains. I am investigating the validity of these assumptions and, for each method, compiling a list of limitations for each method. I hope to address these with better statistical techniques.”

In particular, Robertson is working to understand and to ideally improve upon two important statistical methods used for portfolio analysis. The first, mean variance optimization, has been around since the 1950s and has the distinction of being the first mathematical tool developed for optimizing financial analysis. Revolutionary in its day, the method is based on the now widespread concept that the optimal portfolio is one that gives the highest expected return given any level of risk. The expected return and risk is derived by estimating the mean (the average expected return calculated from the historical average) and the variance (how much the expected value can vary from the average) of a stock’s value. “This method has a number of problems, however,” he explains. “For one, it assumes normality when, in reality, the distribution of assets is not an exact bell curve.”

The second method, time series analysis, is a technique that--as its name suggests--makes use of historical data points over time to make future predictions, in this case, stock values. Optimization depends upon time series data; the future projections (of returns) derived from time series are plugged into the optimization analysis and help with the stock selection process. Robertson’s mentor, Statistics Professor Anindya Roy, is an expert in the area of time series statistics. Working alongside Dr. Roy, Robertson explains, “I am exploring different time series techniques in order to discover a way to make better forecasts of asset class returns.”

Says Dr. Roy, “Tony’s different than many students I’ve advised. He already had great insight into the problem he wished to explore. He came to me with the problem and helped me to formulate it, not vice versa. And he’s come up with more ideas for the project than I could imagine.”

Robertson is very satisfied with his experience thus far at UMBC. He’s been highly impressed with the students and faculty he’s encountered. “The caliber of people I’ve run into here is very, very high; I’ve met brilliant students and faculty,” he praises.