Introduce the members and our PFUG: computational finance We are here today to tell you a little about our study of the stock market and its complexities This title may not have much meaning to you right now but don’t worry, it will make sense later on
You all know there is an inherent risk in investing in the stock market, that is why “risk management” plays a huge role in investing. Traditionally, risk management has focused on the individual firm and the effect of its own actions on itself, so if company A has some management problems, we tend to focus solely on company A and maybe we’ll sell the stocks that we own for company A but we tend neglect the effect of company A on the rest of the market. If any of you follow the Wall Street Journal, a few weeks ago you may have noticed a quote by Timothy F. Geithner, President and CEO of the Federal Reserve Bank. In this quote, Geithner points out that the firms in the stock market are much more correlated than traditionally believed. This calls for further study and development of models that can account for these “systematic issues”.
Geithner’s quote is closely related to our objective, which is “To develop a mixture model that captures the complex correlations (“systematic issues”) present in a given portfolio” Again, this statement might be a little unclear, but this objective will become more clear and more meaningful later on
In stocks and stock prices section: The closing price of a stock is the price of the last stock traded. We looked at the closing prices
Oil Growth = S&P 500 top performers in the oil sector in the past year
Note: We have two data sets. While the first uses MC as one of the criteria for data selection, the second data set is selected based strictly on market capitalization
Pseudo world = a microcosm representative of the entire stock market, defines the system in which to find the “systematic issues”
Weights are important because they decide how much influence each of the companies will have on the whole portfolio.
This slide is a glimpse of what we’ll be covering in the next section: exploratory data analysis, which is what we have been focusing on in the past few weeks. After the portfolio is complete, we downloaded data from Wharton’s CRSP Database (CRSP=Center for Research in Security Prices) One of the first things we did was to look at how each of the sectors performed over the past year. With 10,000 dollars invested with equal weights in each sector, we compounded the daily returns to get the growth. You can see some general trends in how different sectors performed relative to others. Ex. The whole energy sector grew while airlines and finance sectors declined etc. This is very basic but provides a general stock behavior. In the next section, we’ll show some of the statistical analysis that we used to better understand our data and decide if the data have potential of revealing the complex correlations that we are looking for
Good. You might change “normal” to “Gaussian” in the slide. Then when speaking, say “also known as normal or ‘the bell curve’ ”
Good, although you might just go with mean and variance. Then you won’t need the definition of variance on slide 10. forget about st dev.
Assume variance constant in time
The first four models have been developed
Weights evolving in time
1. This shows what we’ve been focusing on in the past few weeks 2. Look more for factors that influence the response of the model, building the model
In addition we would also like to thank Ricardo, Sarah, Dr. Ensor and Dr. Driskill for their help with this project.