Treynor Black Model Report: Apple Regression Analysis

It’s a group assignment, I did Apple. Just talk about my part will be fine.

BMGT 343 “Investments” Group Project Treynor-Black Model Written Report Guidelines

Who are the group members?

Which market index did you use?

Which securities/funds did you choose? Why? What was your data source?

How did you compute the monthly total returns? Discuss your observations about the regression results:

Parameter results; model “fit”; t-stats, etc.

How did you estimate alpha for each security or fund

What method did you choose? Why?

Discuss data collected, analysis performed, and value you chose for alpha

-Be sure to input monthly data into the Treynor-Black spreadsheet model Be sure the input scales are consistent

-Report the final total portfolio recommended by the Treynor-Black model Discuss these results

-Sharpe ratio of the market index portfolio vs the total portfolio Portfolio holding weights and input parameters (alpha and residuals) Anything else you wish to acknowledge

BMGT343

Treynor Black Model Report

We chose to use Netflix (NFLX), Disney (AAPL), and Disney (DIS) as our individual securities for this project, since they are all well-known corporations who are often making headlines in the news. Another interesting tie between all of them is that they all provide streaming platforms. While Netflix has done this for a while, Disney and Apple have just recently introduced theirs. For our monthly returns, we sourced our data from Yahoo! Finance to find monthly adjusted-close stock prices for our individual securities between 7/31/2014 and 9/30/2019. We then calculated our monthly returns by determining the rate of change that occured in adjusted-close prices from month to month. We utilized the S&P 500 as our market index for this project. We did not need to calculate the S&P returns over this time period, as they were previously given to us.

After running the regression of NFLX’s monthly returns vs. the market’s returns, we found that NFLX had a significant beta value of 1.22. We concluded that this alpha value was significant based off of its high t-stat value and p-value below .05. Our alpha value proved to be insignificant, as it’s t-stat value was 1.26. We didn’t worry about that however, since that alpha value is not indicative of NFLX’s true alpha, since the value we received was backwards looking. We applied that same thought process to the rest of our analyses, and we went on to determine the forward looking alpha values later. This regression gave us an R-square value of 10.92%, showing that fluctuations in market returns predict 10.92% of the variation in NFLX’S returns. Ideally this would be higher, but it is not absolutely terrible.

Running the regression of DIS’s monthly returns vs. the monthly returns of the market resulted in us determining that DIS’s beta value is .962. We are very confident that this is accurate, since the t-stat value is over 5, and the p-value is about 0. Even better, our R-squared value came in at 34.29%, which is good for an individual security. We disregarded the alpha yet again, since it is backwards-looking and extremely insignificant with a t-stat value of just .049. The scatter plot clearly displays this strong trend, as there is an obvious positive relationship that can be seen between market returns and those of Disney. 

Estimation for each securities alpha was calculated using the CAPM method and utilizing a market index. We used data from Yahoo! Finance, NASDAQ and CNN Business. Netflix does not pay out a dividend, while both Disney and Apple had dividend payouts of $1.76, and $3.08 respectively. We utilized an ERP of 5.09%, based off of the November 2019 rate determined by Damodaran. We used the current 1- year treasury bill rate of 1.57% which we found on the US Treasury Department’s website. We determined the alpha values by determining the expected excess returns of each individual stock (based off of expected price and dividend payments), then subtracted out beta*ERP to find the abnormal rates of return in excess/deficiency of CAPM. After doing this for each data source, we averaged them out to determine our alpha values.  Overall, our utilization of CAPM showed us that NFLX’s forward-looking alpha is about 2.22%, while DIS’s is about -0.07% and AAPL’s is about -0.61%. In other words, we determined that NFLX is expected to outperform the market, while DIS and AAPL are expected to underperform when compared to the market, according to CAPM.

Since we had two negative alpha values, we ended up having two negative weights within our active portfolio. These negative weights ultimately lead to the portfolio alpha value being negative as well, which occurs as a result of our negative portfolio alpha value of -11.66. This makes the numerator of our total active portfolio weight equation negative.  We also ended up with a large residual variance of 3122.03. Since this value is so high, it drives down the absolute value of the weight invested in our active portfolio because of a large potential for firm-specific risk. The overall composition of our active portfolio makes sense, as NFLX, with the highest alpha value, commands the most weight, and AAPL, the lowest alpha, commands the smallest weight. We can expect to be short DIS, AAPL, and as a result, the majority of our active portfolio. We believe that this model is telling us to sell DIS and AAPL in the short term and invest the proceeds in to the market, since we can expect to get a greater return there, with the intention of buying back the securities later. The Sharpe ratio of our market index came in at .1232, while the Sharpe ratio of our total active and index portfolio was .2423. This proves that our combined portfolio is a superior ratio of portfolio risk premium to standard deviation than the index. When examining our results, we looked for anything that could have the potential to change our end results. When looking at our alpha values, it was clear to us that NFLX has a strong positive value, while AAPL has a clear negative value. Disney however, had several different 1-year price expectations, where some sources provided a positive alpha and others provided a negative. The end result was an alpha very close to 0, but still positive. We think that it is positive that Disney’s true forward-looking alpha value is positive, which would then result in it have a positive weight in our portfolio. Further, our NFLX regression only provided us with an R-square value of  roughly 11%. It is possible that our estimated beta value is not very accurate, in which case we could look towards utilizing an adjusted beta value, or running the regression for a longer window of time.