BetaExoticBets: if I were to price it (for real, not as a theoretical exercise), I'd look at the risk neutral binary pricing and at history based expectations as discussed above as a rough. I would take the max of the two and add a huge reserve against all the potential nastiness that very well may come my way. The reserve will probably a few times larger than the TV on any history or risk neutral based model.That being said, I would have to do this because I usually hedge my risks and holding a chunk of unhedgeable risk is a sore thumb in my book (even if I dealt with pure equity structures). You may want to talk to insurance or reinsurance firms about taking the other side of this risk. They specialize in taking tail risks and have a portfolio of risks largely uncorrelated to the one you're looking to offload. This may fit their books pretty well. Knowing them, they would price this risk off of historical probs with some premium on top (I suspect much more sensible than what I would charge). They also won't hedge it, since they usually punt their risks.

I agree with what outrun says, about who to speak to and also about non independence. A lot of those extreme moves all happened around the same date (October 1987), and so perhaps could be considered one event. A rare example of tails being less fat than you'd think!P

Thanks guys.Outrun, I called a few Lloyds insurance brokers today about this. It was funny trying to explain it to them but I got one interested so hopefully he will get back to me tomorrow. I am sure that Lloyds would underwrite the risk if I can demonstrate loss (a legal requirement of insurance as it isn't meant to be a speculative tool) in the event of a market crash. Well, I can certainly do that! We will see; the hunt continues.

As outrun's conclusion it looks does not significant long historical data for realistic modeling of the S&P dynamics. Indeed for estimating what is the probability to fall bellow x% for a day 1-5 days recent data looks more realistic to use then 2-5 years data.Mathematical assumptions regarding a particular distribution as normal lognormal or jumps are good for papers or theses and could be considered for smooth stable period. The macro parameters like indexes in unstable environment as the current moment it might be useful to study the global market of indexes taking into account their statistics as well as volumes. Then falling 3-5% today in Asia or Europe today might implies 2-4% in US tomorrow and vice versa. The source of the global trouble are global events such as wars. Where huge money come from and where they absorb is not clear. The spontaneous period ended about a year ago when $$ declines along with markets. But special rules to pump money to support indexes was accepted by us. This of course leads to the next falls and defaults changes in financial global system. It might look attractive for G- countries FX globalization but on the other end for the developing countries that do not involve in spending money on wars it looks attractive to raise additional restrictions on FX to protect local currencies. Union or over support $$'s will increase risks of fall. Wars define the level of speculations (demand) of oil prices then food prices. Exactly the same events in the same order had happened in Soviet Union at the end of 80-s and at early 90-s. At that time the old economic system were reconstructed and to cover lack of paper money they printed new similar to as one could observe today.Concerning exotic option 1 year to maturity period might be too long today and a month period could be more attractive as well as a forward contact. As far an option is more risky instrument then others it might make sense to change slightly the scheme to convert it in CDS with variable x% as default indicator with upfront or periodic payments? I have a feeling that there no realistic model for 1-year S&P or DJ , FTSE dynamics.

Huh?!

QuoteOriginally posted by: AlanRegarding pricing, I would first try to price the close-to-close version of this option say by studying all the S&P500 optionexpirations with one day to go.For the close-to-close version it should be able to get quotes in the OTC market

Since the historical probability of a >10% drop is about 1/13000, I would assign it approximately that price, e.g., 3K = 300K not'l / 13000 * fudge factor. There is no way to estimate the probability or hedge against an unknown event. You could keep going long a strip of S&P 500 futures as a(n extremely crude) daily hedge but that would be too expensive, so I'd either do nothing or hedge with a front contract if I smelled smoke.Most likely, I would happily pocket the premium. If WWIII broke out, defaulting on a 300K payment would be the least of my worries.