Nifty Weekly Analysis using INDIAVIX

5
(2)

After an overwhelming response to the previous article “Analysis of Nifty weekly change“, we got many suggestions to perform analysis using VIX. So, here we are with an interesting read on what we found during our expedition. Do read the previous article to get in sync with this analysis.

The mission of this analysis is to use VIX to find whether we can manage our trades when outlier(Nifty weekly change > +3% and Nifty weekly change < -3%) occurs.

Operation VIX

I started the journey by analyzing the rightly skewed distribution of INDIAVIX and finding its statistical specifications like mean, median and standard deviation.

Mean of VIX17.47
Median of VIX16.34
Standard Deviation of VIX4.66

Here, I considered the data from 2011-2019. Seeing the volatility levels in March 2020, these numbers look very low, but even if I would have used the data of 2020, the mean might not have changed much because it was just a matter of 2-3 months(8 to 12 weekly data points) that VIX cooled off and reverted back to its mean.

Since this is a rightly skewed distribution and not a normal one, we cannot use mean+2*S.D formulae to state that 95% of the data is between this range. So, we will make use of our data to find it.

% weeks when VIX was below 2078%
% weeks when VIX was below 2591%
% weeks when VIX was below 3098%

Please correlate

I thought VIX might be correlated with this week or previous week % change in Nifty. But, I found almost no correlation between the two. So, nothing could be concluded using just VIX and weekly %change.

Correlation b/w present weekly %change and present week VIX-0.13
Correlation b/w present weekly %change and previous week VIX0.07
Correlation b/w previous weekly %change and present week VIX-0.10
Correlation b/w previous weekly %change and previous week VIX-0.13

This was rather disappointing. But I found a myth buster during my stint with the data. As a naive trader, I expected that volatility should act similarly for both the positive(Nifty weekly %change > 3%) and negative(Nifty weekly %change < 3%) outliers in Nifty. I used to think that volatility expands irrespective of the direction of movement. The correlation between weekly %change in Nifty and weekly %change in VIX helped me bust this myth.

Correlation b/w weekly %change in Nifty and weekly %change in VIX-0.48

Although this was a weak correlation between weekly %change in Nifty and weekly %change in VIX, it tells us that volatility shoots up when the market falls and cools off when the market rises. This correlation also proves the fact that “Fear tends to manifest itself much more quickly than greed“.

What outliers told me?

It was also important to have a look at the outliers and find out a particular VIX value below which one can always be safe to trade without worrying about outliers.

%outliers when VIX < 2057%
%outliers when VIX < 2576%
%outliers when VIX < 3095%

It is not at all surprising to see that most of the outliers occurs when VIX is below 30 as the mean and standard deviation of VIX suggests the same. It is worth to note that how we can easily lie using the above table. One could easily conclude that since 95% of outliers occur when VIX < 30, so why not trade strangles when VIX > 30. This happened because 98% of VIX data is below 30. Drawing such conclusions from the above table can be catastrophic for your financial health

Conclusion

It was fun to understand the psychology of traders using just VIX data and prove that fear is indeed stronger than greed. It might be very demotivating to not get any solid conclusions from the above analysis using VIX. But, it is fine to draw no conclusions rather than drawing wrong conclusions from the data.

You can find the analysis report and data* here.

Note: The close price of Nifty in the excel sheet may not exactly match with the adjusted close price of Nifty.This analysis is not any strategy. This analysis is just a base that can help you develop your own ideas or strategies.

That’s it for today.

Stay tuned!

How useful was this post?

Click on a star to rate it!

Average rating 5 / 5. Vote count: 2

No votes so far! Be the first to rate this post.

Spread the love

About the author

Akshay Satpaise

Akshay Satpaise is an Electrical Engineer who loves data crunching. He has interests in personal finance, stock market and data analysis.

View all posts

2 Comments

  • Superb article…. The left skewed distribution (it’s a typo there) implies the vanity of mean+2*SD. This is the Crux

    Bcoz outliers are found in >95% times, using strangles will be catastrophic…. This is contrary to what is sold everywhere. Cud make money, but u need to take extra caution or manage risks

    • Hey Arpit, thanks for the feedback. A right skewed distribution has most outliers on the right side of the distribution. One way to prove this is that the mean is greater than median for right skewed distribution. So, it is not a typo in the article. The VIX distribution is indeed a rightly skewed one.

Leave a Reply to Arpit Patel Cancel reply

Your email address will not be published.