This Project is an extension from Project 1 and Project 2 attached.

Directions:

1. It is an individual project ? however, students could help each other.
2. Do it yourself and submit your copy by the due date.
3. No Excuses and HARD COPY PLEASE ? if you?re absent on the due date, drop your copy in my mailbox any time before the due date.
4. Write a professional report. Include all your results/GRETL output.

Teaching Goals:

1. To estimate the probability of recessions.
2. Based on the above estimates, determine a hedging strategy.

Steps you may need to follow:

1. 1. Assume that you have invested $500 million in SPY.
2. 2. You want to use S&P500 futures to hedge based on the estimates of the probability of recessions.
3. 3. To accomplish this, you need to estimate the probability of recessions. You need the following data.Steps to calculate the estimate of the probability of recessions:
4. a. Download 1985-2016 10-year treasury monthly data fromhttps://fred.stlouisfed.org/categories/115. Name this variable 10-year.
5. b.Download1985-2016monthly3-MonthTreasuryBill:SecondaryMarketRatedatafromhttps://fred.stlouisfed.org/categories/116.Namethisvariable3-month.
6. c. Download1985-2016NBERbasedRecessionIndicatorsfortheUnitedStatesfromthePeriodfollowingthePeakthroughtheTrough(collectmonthlydata)fromhttps://fred.stlouisfed.org/search?+data.NamethisvariablesRecession.
7. d. In Excel, gather all of the above and make sure they are aligned by date. Note that you have 4 variables: SPY prices, yields on 10-year and 3-months Treasuries, and the recession binary variable.
8. e. Import the data set to GRETL. And create a new variable TERM, which is the difference in yields on 10-year and 3-month Treasuries. Create another variable: SPY returns.
9. f. You need to run a probit model. Select, Model Limited Dependent Variable Probit Binary. The reason for those steps is that you want to find the probability of recessions. Since recession is a ?binary? variable (see the data), and you want to find its probability, you want a probit model.
10. g. Select ?Recession? as dependent. TERM as a regressor.
11. h. Select robust standard error (left bottom of the window).
12. i. At left bottom of the window, select lags. Here you select different lags of the regressors. Select specific lag selection. Every lag represents one-month.
13. j. Now find the estimates of recession probabilities at different forecast horizons.
14. k. If you select TERM(-1) you have predicted/forecasted recessions one month before recessions.
15. l. Run probit models for 1 to 15 forecast horizons.
16. 4. At what forecast horizon do you get the best estimates of recessions? The ?best? is determined by McFadden R-squared values (higher is better). Make a simple table to show McFadden R-squared values for different forecast horizons.
17. 5. Plot the estimates of recession probabilities for the best model. Select GRAPH fitted actual plot against time.
18. 6. Based on the plot, when do you think you should have hedged your $500 million SPY portfolio just before the last recession?
19. 7. For how long you would have hedged? How many futures did you need to hedge?
20. 8. What was your net profit/loss during the hedge period?Hint: part 7 and 8 are similar to project #2.
21. 9. Now assume that you have invested $500 million in the portfolio you created in Project #1 or Project#2.
22. 10. Repeat steps 6 through 8 for the portfolio you?ve created. Hint: Since your data for this project is from 1985-2016 and your portfolio data is from 2006, you need to create an excel file that has all information you need and use this excel file for the analysis.