Calculate covariances for volatility

In MPT, we quantify risk via volatility. The math for calculating portfolio volatility is complex, and it requires daily returns covariances. We'll now loop through each month in the returns_monthly DataFrame, and calculate the covariance of the daily returns.

With pandas datetime indices, we can access the month and year with df.index.month and df.index.year. We'll use this to create a mask for returns_daily that gives us the daily returns for the current month and year in the loop. We then use the mask to subset the DataFrame like this: df[mask]. This gets entries in the returns_daily DataFrame which are in the current month and year in each cycle of the loop. Finally, we'll use pandas' .cov() method to get the covariance of daily returns.

This exercise is part of the course

Machine Learning for Finance in Python

View Course

Exercise instructions

Loop through the index of returns_monthly.
Create a mask for returns_daily which uses the current month and year from returns_monthly, and matches this to the current month and year from i in the loop.
Use the mask on returns_daily and calculate covariances using .cov().

Hands-on interactive exercise

Have a go at this exercise by completing this sample code.

# Daily covariance of stocks (for each monthly period)
covariances = {}
rtd_idx = returns_daily.index
for i in returns_monthly.____:    
    # Mask daily returns for each month and year, and calculate covariance
    mask = (rtd_idx.month == i.month) & (rtd_idx.____ == i.____)
    
    # Use the mask to get daily returns for the current month and year of monthy returns index
    covariances[i] = returns_daily[____].cov()

print(covariances[i])

Edit and Run Code