1) Testing the Size Premium In this exercise we investigate the idea that small stocks make more than large stocks

1) Testing the Size Premium In this exercise we investigate the idea that small stocks make more than large stocks. 
(a) Start by going to Ken French's data website. After clicking on the "Data Library" link on the left, download data related to "Portfolios formed on Size." 
(b) Extract the ZIP file and open up the underlying CSV, in Excel. Within the spread-sheet you will find monthly value-weighted returns that correspond to different sorted portfolios. Each portfolio looks at the average return of stocks based on their size. Some portfolios will be tilted towards small cap stocks while others will be tilted towards large cap stocks. 
Find the portfolios (the columns) that relate to the 10 deciles (that is from "Lo 10" to "Hi 10"). The "Lo 10" portfolio will relate to stocks that fall into the 10% smallest of size. The next portfolio ("2-Dec") will look at 10 — 20%; the third will look at 20 — 30% and so on. 
Taking the date column and the 10 columns mentioned above, save them in a brand new CSV file. 
(c) Load the freshly saved CSV file into your Python environment. How many rows and columns do you have? 
(d) What is the average return of all 10 portfolios? Do small stocks make more than large stocks? 
(e) Using Matplotlib, lets plot a graph of the average returns per portfolio. Let the bottom axis indicate the size portfolio of interest (1, 2, ... , 10) and the y-axis be the average return for that portfolio. Is there a consistent relationship throughout? (f) Some fund managers believe that the size premium has more recently become less prominent. Let's test that. 
Conditioning on data after 2000, examine the average returns as we did in part (d). Do we see that there is still a size premium? Has the shape of the pattern changed? 
2. Investment Grade vs. High Yield Corporate Bond Spreads (a) Go to the St. Louis FRED and download 3 series: (i) the ICE Bank of America US Corporate Index OAS, (ii) the ICE Bank of America US High Yield OAS, and (iii) the monthly industrial production index. Place all of these in one CSV 
2 
file along with the date. Only keep rows where data are available for all 4 columns. 
The ICE BofA Corporate Index represents credit spreads (average borrowing risk-iness) for investment grade firms, while the high yield index represents credit spreads for the riskiest firms. 
(b) Load the data file into Python. You should have 4 columns. Compute the monthly growth rate of industrial production. 
(c) Using techniques we used in a previous live session, convert the dates to be a datetime object at the end of the month. For example if the original date is 1/1/2001, make it 2001-01-31 as a datetime object. 
Hint: look back to the "freddata" example in the 2nd code file. 
(d) What are the means, standard deviations, and correlations across the data series? 
(e) Using Matplotlib, plot the series of high yield spreads against the investment grade spreads. Show that the high yield spreads are strictly greater throughout the plot. Qualitatively, why would be expect this to be the case? 
(f) We will focus our attention on the sensitivity of the two credit spread series to adverse events — the global financial crisis (GFC) and the recent pandemic. 
Compute and report the economic growth and change in each spread series for two periods: • End of December 2007 to the end of December 2008 (GFC period) • End of December 2019 to the end of March 2020 (US onset of pandemic) Which credit spread series seems to react more to adverse events? Does this make sense?