Quantitative Finance Stack Exchange is a question and answer site for finance professionals and academics. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a time series for a security list with 2 parameters calculated for each time period. For example, for a stock XYZ, I have Param1 and Param2 calculated over various time periods stacked against each other. I calculate the correlation between the two parametric series for each time period.

I see sudden drop in correlation for a certain time period, which I am not able to explain. The correlation drops from 0.5 to 0.02 for the same securities over consecutive time periods.

Can someone help me analyze what is causing the drop in correlation at the parametric level ? What changes/ differences should I look at to explain this change ?

share|improve this question
Could you please say what the securities are, and what is the time period? Otherwise, the best you could hope for is a guess. – Shahar Sep 18 '14 at 19:40
Not sure what goes on: You calibrate two stochastic processes described by two parameters and then you verify the correlation between the two stochastic processes? – pincopallino Sep 19 '14 at 9:07

Depends how you calculating correlation, but probably you have rolling window from what you get high and low for calculation, when you adding samples to window then some samples must exit the window, when sample that exiting are not equal to high and low then it's don't matter, but when high or low is exiting then suddenly everything changes in your calculation.

Can't say more without specifics/details of your algorithm. Long shot.

share|improve this answer
My first thought. I'd suggest the OP to try out his model with non-overlapping windows. – pincopallino Sep 19 '14 at 9:06

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.