6
$\begingroup$

Have seen this being done for years (primarily by J.P. Morgan and a couple other bank research desks) and am attempting to re-create for my own personal research. I’ve read the forums on here but no one has seemed to crack the code yet; here’s what I have thus far —

I calculated the dollar gamma for each SPX call and put option expiring over the next few weeks by taking 100 * open interest * gamma * spot^2 / 100 and aggregated by SPX strike level (in this case, per every $50 strike — 2650, 2600, 2550, etc.). I then subtracted the dollar call gamma from the dollar put gamma for each strike to generate the ‘P-C imbalance.’

So in essence I now have the current net dollar gamma exposure for all weekly/regular expiration options by strike but am unaware as to how to get something even close to the picture attached. What I get is a normal distribution-type graph (i.e. most gamma centered around the ATM strike) which makes sense since the highest gamma is going to be near the ATM strike with generally a large open interest.

Can anyone help me out here? Is there perhaps some weighting scheme I’m failing to incorporate? Do I need access to dealer data to even conduct this analysis?

enter image description here

Improve this question
$\endgroup$
3
  • $\begingroup$ Why do you subtract the call gamma from the put gamma? If you want total gamma should t you be adding them up ? $\endgroup$ – dm63 Jun 4 '19 at 20:30
  • $\begingroup$ I tried calculating this myself. I get the general trend of the chart right but my aggregate gamma values are too small. I filtered for options with 75%-125% moneyness on SPX, calculated the aggregate $-gamma per contract using unitGamma * contractMultiplier * underlyingPrice * openInterest (example: 0.0713 * 100 * 2843.49 * 86212 = 1'747'869'304). Then sum up the strike buckets and take the difference between Puts and Calls. If I use your formula gives me values way too high. Could anyone comment on the methodology? $\endgroup$ – oronimbus Jun 7 '19 at 15:17
  • $\begingroup$ open interest you have is positions of all traders, not just dealers. their research is calculating for only dealers. I know I'm not helping but yes you are missing a few steps of calculations here. $\endgroup$ – user43153 Nov 1 '19 at 16:01
1
$\begingroup$

Consider the problem from a different perspective - you just plotted the existing gammas of the OI, AT THIS MOMENT.

What you modify, is you consider this one iteration, with each iteration being across different spot levels.

What they do is this:

Assume all calls are owned by dealers (positive GEX) and all puts are short by dealers (negative GEX) Using this assumption, calculate the net GEX for the existing options inventory. Iterate through different levels of spot price (use a constant smile assumption for simplicity) and plot your resulting GEX sums across different levels of spot price

Hope this helps

Improve this answer
$\endgroup$
1
  • $\begingroup$ Thanks, this helps a lot and this is exactly what I was thinking. This is inherently a flawed assumption FWIW. $\endgroup$ – Theta Bill Murray Jan 10 '20 at 15:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.