In this latest update on structured warrants, a simple multi regression has been performed on structured call / put warrants (with FBMKLCI as underlying) in order to understand the impacts of (1) remaining days to expiry (2) strike price ; and (3) conversion ratio on warrant premium / discount.
The current list of structured call / put warrants (FBMKLCI) is summarised as follows:
* The above warrant prices are based on bid prices
Regression on structured call warrants
As expected, there is significant statistical relationship between premium / discount of warrants and days to expiry / strike price. However, the Exercise Ratio factor does not appear to have significant statistical property in this case. Based on the above regression results, we may make a broad prediction on the premium of the call warrants and thereby, allowing us to estimate a revised theoretical warrant price (based on the predicted premium):
It is important to note that warrants that are nearing to their expiry dates, are subject to significant volatility. Hence, the above regression analysis should technically exclude these expiring warrants.
Regression on structured put warrants
In relation to put analysis, both days to expiry and strike price represent statistical significance. Similar to that of structured call warrants, we are able to perform simple forecast on the premium / discount of the warrants:
A quick comparison between call and put warrants shows that a higher number of put warrants appear to be trading at higher relative premium (i.e predicted premium < current premium) due to prevailing negative sentiments associated with emerging markets (post the recent US election).
It is important to note that the above analysis is a simple desktop analysis. Further study is required. It is also important to consider the confidence interval in order to account for potential errors in the statistical analysis.
DISCLAIMER: THIS IS A PERSONAL BLOG AND SHALL NOT BE RELIED IN WHATSOEVER MANNER BY ANYONE. ALL ARTICLES CONTAINED IN THIS SITE ARE FOR INFORMATION AND ILLUSTRATIVE PURPOSES ONLY AND DOES NOT PURPORT TO SHOW ACTUAL RESULTS. IT IS NOT, AND SHOULD NOT BE REGARDED AS INVESTMENT ADVICE OR AS A RECOMMENDATION REGARDING ANY PARTICULAR SECURITY OR COURSE OF ACTION. SOURCES USED IN THIS SITE HAVE NOT BEEN INDEPENDENTLY VERIFIED FOR ACCURACY, COMPLETENESS AND TIMELINESS. YOU SHOULD SEEK INDEPENDENT AND PROFESSIONAL INVESTMENT ADVICE IN REGARD TO YOUR INVESTMENT DECISIONS. THE AUTHOR MAY HOLD POSITIONS IN THE SECURITIES OR RELATED INSTRUMENTS MENTIONED ABOVE.
Ken,
In your Regression on Structured Call Warrants, if you find that the Exercise Ratio is insignificant, should you not take it out and recalculate a new set of coefficients before calculating your Predicted Premium?
Instead of:
Predicted Premium = 0.000145729*DaysToExpiry – 0.0000376915*ExerciseRatio + 0.000567001*Strike – 0.920459
Should it not be:
Predicted Premium = a*DaysToExpiry + b*Strike + c
whereby a, b, c are the coefficients of a new regression without the ExerciseRatio variable.
LikeLiked by 1 person
That is the recommended regression approach (i.e removal of non-statistical significant variable and re-run of the regression). My intention was to look at the options valuation from a broad framework & regression analysis instead of using BS or other appropriate methods. Further refinements are definitely required. Cheers.
LikeLike