Hello thanks for the video. However I did not really understand when you that "this higer implied volatility suggests that the left tail must be heavier than the log-nomal". My question is how can we have an idea about how heavy are the tails of the actual distribution by simply looking at the volatility smile? What's the link (implied volatiltiy-implied distribution)? Thanks :)
"Crashaphobia" may or may not be a contributing factor to option pricing, however, as a theory it fails to explain why there was an implied volatility smile in option pricing prior to 1987. It also fails to explain why the implied volatility smile doesn't adjust as markets become more (or less) risk-on (or risk-off)
There was an implied volatility smile in pricing prior to 1987 because the bid-ask spreads are large relative to the bid, ask, or midway premiums. It does not make sense to buy or sell options that far out of the money when the spreads are so large relative to the premiums (less the implied expiration value if it is in the money). So, the people who do that are a small number and are only willing to sell for a positive profit margin and the people who buy have some reason to do so, or they want a low priced option but are willing to pay double the bid-ask spread in premium despite the implied volatility being quite high (and thus the option a poor value). There should be a volatility smile, and a skewed one, and it should appear even more skewed on a non- loot scale chart, but this isn’t the dominant reason for options with low premiums (say 0 to 4 times the bid-ask spread) having such high implied volatility. Look at that chart. It appears that if you extended it to the left you would see it becoming flatter that makes no sense at a zero strike the implied volatility should be nearly 100 percent. But the smile is actually steeper than that at first. At about 1 sigma, in terms of the minimum implied volatility (let’s consider that a proxy for the normdist sigma) it us already returning an implied volatility of nearly 60 percent. Thus would be easier to visualize if the x axis scale was log, so he equity returns are necessarily log distributed, as they have a limit of zeros .
If the model gives a iv of 25.8 and the market gives you 32.6 iv,would that mean that the option is overpriced (taking into consideration the cost of transaction)
Implied volatility does not imply Black Scholes Merton is flawed or not up to the task. It implies that you are dealing with a non-normal distributed security. That’s all. You also aren’t using a log scale X axis. Nobody really knows how big or skewed the smile should be, because that would imply solving for a complex system, and not strictly speaking stochastic system. Therefore, the implied volatility smile is, in part, based on market participants expectation of the security volatility smile. However, this is based entirely on market price history. However Volatility in the underlying security is higher given shorter periods. And impossibly high given big swings in short periods. 1-day or 1 hour or minute implied volatility approaches 100 percent in reality, when considering the likelihood of gigantic moves in short periods. One could not price an option representing the assumption of a 1- minute 4 percent crash or a 1-day 20 percent crash. The reality of those situations is that the markets were not functioning except for some badly designed algos selling to each other for prices that shouldn’t have gotten filled. In 1987 the traders just accepted the new price as a real price and it took 2 years to make back the 20 percent that was lost. In 2010, other algos came in and bid the price back up very quickly, so it didn’t take long at all to make back the 4 percent losses.
Great video. It is very logical that tail risk and crashphobia cause volatility skew towards lower strike price call options and put options. Is it possible to apply an efficient market hypothesis to this volatility graph? E.g the tangent to every point in the graph will indicate slope of the graph. So can we sell the option where slope is steepest ?
Please use put options on downside and call options on up side. In this way you get correct volatility smile, because put option is not in value on upside and call option is not in value on down side.
I just watched a video with Warren Buffett yesterday and he sort of agreed with you, with the proviso that BSM is a reasonable estimate in a vacuum, if you know nothing about the underlying stock ... but he's never going to to approach a stock from that position of ignorance so for him it is basically worthless.
Another chapter in 9 minutes. Thanks David,
We achieve deep understanding watching your explanations.
May be twice deeper then simply reading Hull.
This is seriously a great video. Thanks for helping me I have my derivatives exam tomorrow.
Great video, this is going to help me with my Derivatives exam this week!
Thank you for watching! We are happy to hear that our video was so helpful! :)
That's nice.It sort of makes you want to smile,doesn't it?
Good explanation,thank you.
Hello thanks for the video. However I did not really understand when you that "this higer implied volatility suggests that the left tail must be heavier than the log-nomal". My question is how can we have an idea about how heavy are the tails of the actual distribution by simply looking at the volatility smile? What's the link (implied volatiltiy-implied distribution)? Thanks :)
"Crashaphobia" may or may not be a contributing factor to option pricing, however, as a theory it fails to explain why there was an implied volatility smile in option pricing prior to 1987. It also fails to explain why the implied volatility smile doesn't adjust as markets become more (or less) risk-on (or risk-off)
There was an implied volatility smile in pricing prior to 1987 because the bid-ask spreads are large relative to the bid, ask, or midway premiums. It does not make sense to buy or sell options that far out of the money when the spreads are so large relative to the premiums (less the implied expiration value if it is in the money). So, the people who do that are a small number and are only willing to sell for a positive profit margin and the people who buy have some reason to do so, or they want a low priced option but are willing to pay double the bid-ask spread in premium despite the implied volatility being quite high (and thus the option a poor value).
There should be a volatility smile, and a skewed one, and it should appear even more skewed on a non- loot scale chart, but this isn’t the dominant reason for options with low premiums (say 0 to 4 times the bid-ask spread) having such high implied volatility.
Look at that chart. It appears that if you extended it to the left you would see it becoming flatter that makes no sense at a zero strike the implied volatility should be nearly 100 percent. But the smile is actually steeper than that at first. At about 1 sigma, in terms of the minimum implied volatility (let’s consider that a proxy for the normdist sigma) it us already returning an implied volatility of nearly 60 percent. Thus would be easier to visualize if the x axis scale was log, so he equity returns are necessarily log distributed, as they have a limit of zeros .
If the model gives a iv of 25.8 and the market gives you 32.6 iv,would that mean that the option is overpriced (taking into consideration the cost of transaction)
Implied volatility does not imply Black Scholes Merton is flawed or not up to the task. It implies that you are dealing with a non-normal distributed security. That’s all. You also aren’t using a log scale X axis. Nobody really knows how big or skewed the smile should be, because that would imply solving for a complex system, and not strictly speaking stochastic system. Therefore, the implied volatility smile is, in part, based on market participants expectation of the security volatility smile. However, this is based entirely on market price history.
However Volatility in the underlying security is higher given shorter periods. And impossibly high given big swings in short periods. 1-day or 1 hour or minute implied volatility approaches 100 percent in reality, when considering the likelihood of gigantic moves in short periods. One could not price an option representing the assumption of a 1- minute 4 percent crash or a 1-day 20 percent crash. The reality of those situations is that the markets were not functioning except for some badly designed algos selling to each other for prices that shouldn’t have gotten filled. In 1987 the traders just accepted the new price as a real price and it took 2 years to make back the 20 percent that was lost. In 2010, other algos came in and bid the price back up very quickly, so it didn’t take long at all to make back the 4 percent losses.
Great video. It is very logical that tail risk and crashphobia cause volatility skew towards lower strike price call options and put options. Is it possible to apply an efficient market hypothesis to this volatility graph? E.g the tangent to every point in the graph will indicate slope of the graph. So can we sell the option where slope is steepest ?
Please use put options on downside and call options on up side. In this way you get correct volatility smile, because put option is not in value on upside and call option is not in value on down side.
What software are using to plot and where can I get access to implied volitality chart for individual stocks. Preferably free. Thanks.
Awesome Video, do you make the excel sheet available for viewers?.
Really good explanation of implied volatility.
Thanks David. Very clearly explained, much appreciated!
understood everything. great video
wow strike at $250. good old days.
@freeactivity90 thank you very much. And punny, too!
So the BSM is worthless!
I just watched a video with Warren Buffett yesterday and he sort of agreed with you, with the proviso that BSM is a reasonable estimate in a vacuum, if you know nothing about the underlying stock ... but he's never going to to approach a stock from that position of ignorance so for him it is basically worthless.
Thanks
Thank you for watching!