Time Analytics - Is Correlation and Jiseki connected?


Jiseki is a performance ranking idea for an asset in a group for different holding periods. Conventional tools don’t look at sectors as a proxy of a group but not components perse. Our idea of extreme reversion is designed to understand performance cyclicality.

This week we look at the Indian Energy sector correlation from Oct 2009 to 2011 and juxtapose it with change in quarterly performance rankings during that period. The correlation of the energy sector components was made with Reliance.

What we could observe? The higher the correlation of the stocks with the Reliance the worst the stocks did in performance and the lower the correlation of the stocks with Reliance the better they did in performance.

Petronet for example moved up 50% in Jiseki performance rankings during the period along with Castrol. While the high correlation and positive correlation stocks with reliance like Seamec, Suzlon saw not only a drop in rankings by 50% but also absolute price loss.

The connection of correlation with Jiseki change in rankings might seem strange. But it’s not because if Reliance is at the top of the group, it will underperform and because it’s a sector leader, anything correlated positively correlated with it will also underperform.

This is simple logic, which the data expresses visually. Does the data tell us something else? The histogram also tells us that the only stock which should have fallen and did not fall as much was Reliance. This is what we may call a sector leader’s premium, which keeps us away from Reliance as a buy opportunity confirming what we wrote recently in the article, “should I buy Reliance?”

Avinash Barnwal 
is a final year master student of Statistics and Informatics at IIT Kharagpur. He has worked on human response time at Department of Psychology, University of Amsterdam.  He has worked on marketing analytics for the Customer Intelligence Unit, HDFC Bank . Avinash is passionate about developing statistical models and believes that statistics could address temporal challenges.

2 Responses to “Time Analytics - Is Correlation and Jiseki connected?”

  1. Debi says:

    Hi Avinash,
    Orpheus had another story saying Correlation is cyclical. Whether that will also hold true with respect to this theory. Can you please check from 2000 onwards?


Post a comment


Spam protection by WP Captcha-Free