Like any swap, you agree that at some point in the future you going to get paid the difference between some current level and the value that actually realizes. In this case, it would be the value of realized variance against the pre-agreed strike K: Code: return[i] = log( SPX[i]/SPX[i-1] ) V = sum{i=1 to N} return[i]^2/N pnl = Notional * (V - K^2)/(2*K) If you take a simple example, let's say on Feb 21 you agreed to sell a March var swap (meaning you are receiving fixed, paying floating) for a vega notional of $100k and the initial strike of 17.8. On March expiration, you see that SPX realized 64.0. Your PnL will be 100k * (17.8^2 - 64*2)/(2*17.8)
Sle. Thanks for helping, this very fascinating. If you don’t mind, I will keep the questions coming: Why does a dealer only trade on the close? Why not midday and then flatten? (If advantageous). Would a retailer try to analyse the magnitude of the swap flows indirectly from the cash index, the futures, or the replicating options strip? Or can someone attempt to measure the flows from VIX products? Does the intraday volatility impact the size of the offsetting flows on close? Or is it only the close-to-close distance? Perhaps intra-day movements are altogether irrelevant to the dealer (only close-to-close?)
I apologize to everyone for the decidedly "off" post, but sometimes 'not responding' to a troll sends the wrong message to all concerned. 'What it's about' is a person with a practiced lack of self-awareness of deep personal issues, on flagrant display through the history of their ET posts. It ends up to be a close call sometimes, whether non-response or responding in-kind is a better path. But when the history of someone's posts include a large measure of "queefs" "twats" "vags" or "why don't you just die and save us the trouble of ___whatever____" -- that sez a lot about a person, issues or not.
@tommcginnis relax, I have been called names much worse than that in the years I am here. It doesn't bother me a bit. Why should I? They told me I was dumb, and I admit. They said I was a loser and I admit. They said I was a moron, no big deal.... It is an anonymous forum we are all anonymous, I only care about folks who are kind, generous and helpful and there are too many to mentioned here. Take care, be careful and make sure you practice physical distancing and psychological distancing.
On the other hand, in his defense, @PoopyDeek is a really great trader, outspoken, yes. He did us a great service exposing many frauds and fakes here to prevent them from taking advantage of newbies like me. So I am also his fan.
I am not Sle and Sle is not me. Very important point, should SEC ask Assuming the dealer has a perfect hedge on (i.e. the strip vs the variance swap), he does not have any gamma intraday, he only going to have to hedge the daily reset. I.e. he does not actually have any gamma because his convexity is perfectly matched. Granted, this is not always true (e.g. large gaps are going to disconnect your hedge from the swap) but it's true enough that most dealers are not gonna bother unless the move is huge. That's the tricky question. You can use the clearing house history to see the flows and attribute them, pretty sure that data is publicly available from the clearing house. Alternatively, you can look at the strength of the previous impacts and try to back out the size of the outstanding gamma. Well, the dealer only cares about the variance swap payoff that's calculated from the close to close returns. So from his perspective, even if he's marking the close, he's golden (if anything, he likes that impact if he's not fully hedged). But from the markets perspective, the market makers will pull back when they see the target-close imbalance and will buy it back into the settlement (thus making money to the retail guy that decided to play this game )
For the purpose of this discussion, dealers are the guys trading variance swaps, usually these are index volatility desks at banks. Market makers are market makers in index futures, who are usually high frequency shops like Jump et al. Very different entities.