i am interested in measuring the volatility of a variable in the financial statements. previous studies measured it using the rolling standard deviation of the item that i am interested in measuring it divided by average total assets over the last four quarters. i don't understand how to calculate this measure. Also, in some articles they use quarters while in others they use years, my question is how can i choose the right period(quarters, years). Another question: what is meant by the standard deviation of a time series. could anyone please help? Thanks
1 Answer
$\begingroup$
$\endgroup$
The calculation part depends on the medium you use, and the exact syntax differs from the program or programming language you are using. Randomness in a phenomenon manifests variability or dispersion in the observations. There are many statistical objects you can use to model this variability, which come with their own advantages and disadvantages, standard deviation being a common one of them. Standard deviation measures how spread out your observations are from the mean. Under the independent, identically distributed assumptions (all your data points are independent and come from the same data generating process) standard deviation of the whole dataset can be quite explanatory. However, the statistical dispersion of observations in the case of a time series changes over time, depending on the underlying process. You can compute the standard deviation of the total series -what I think you are implying in your question- but most likely it will not be as explanatory as moving standard deviations.
Therefore, it ends up being a modeling choice. If you are interested in monthly variabilities, you can compute the standard deviation in each month; there is no right or wrong way of doing this, it depends on the question in your mind.
