I have been looking at some equity IV surfaces and encountered this one, that I don't yet understand. This is a representation of the PUT IV taken on 8/14/2015 after market close for SPY. There is a discontinuity between the 4 longer term contracts VS the 4 shorter term contracts. The CALL Volatility Surface graph for this date does not reflect this discontinuity. I observe this on other dates as well, but not when SPY is in backwardation. I am curious if anyone else has observed this, and if so, have an explanation. Consider only Moneyness values <0 for this oddity. 2 pictures are provided to draw focus to the interesting area. A penny for your "constructive" thoughts.
If you are a trader or strategy developer (as opposed to an academic researcher) then in your position I would think very practically and a step further: Have I maybe made an important discovery that I can turn into profit? etc. It should be easy to answer that question if you have the market data around these events...
First, double and triple check your data for accuracy. This surface seems very odd for a liquid contract like SPY. But, if accurate, then various debit strategies would seem to be in order.
panzerman: I agree. Most likely I have an error, or the data source is faulty. Attempting to resolve, but do not have reliable data source to compare.
I believe this is likely due to the fact that you're showing the surface using puts only. Often times the listed closing mark on a given day for ITM options (in this case the upside puts) is distorted. Via put-call parity, options of the same strike should trade at the same IV (at least for European). I've always understood most market participants use the OTM contract as the benchmark IV. Hence, the upside surface would be based off of call IV and the downside off of puts. I think this is why you don't get the same distortion on the upside when you run the surface with calls. The deep ITM options are very illiquid, so the posted midpoint mark in your data set may be well off of where the option would actually trade in the market (at the same vol as the OTM call of the same strike).
I can back that up with antectodatal experience trading SPX options. The deep ITM puts/calls often don't trade near the mid, while OTM and less ITM reliably hit 5 cents off the mid almost every time. If I take the deep OTM vol and imply a price for the deep ITM opposite option, I can usually trade at that price on the deep ITM option.
Longthewings: Thnx for your feedback. Unclear if your comments resolve the anomaly (which may be data or my coding error). I just re-plotted using the prices 30Minuites prior to Closing. I am posting the Call graph as well, which does not have this anomaly. The plots contain all available prices for all put and call strikes and the indicated expirations (8 series). Look only at negative moneyness for OTM. The IV values plotted are all derived individually by extracting the required IV to match the "MARK-price" of each option. Since there were comments regarding liquidity that seem to be off mark, I
Since there were some comments regarding liquidity that seem to be misplaced, I am posting the picture again (this is 30 min prior to close), with the interesting region circled.
Sorry...from the original post I thought you were discussing a different area of the graph. Looks to me like a jump condition priced in for an event (i.e. dividend).
Duh! You hit the nail on the head! Thnx. Sometimes it is difficult to see the forest for all the trees! ;-) -- Dividend caused the step! -- RE-evaluating my formulas to determine if my handling of dividends is proper. Thnx!