Tuesday, April 6, 2010

Family Dollar (FDO) - Secret Sauce to Pro Trading: Earnings Vol Analysis/Preview

FDO is trading 37.39, IV30™ is up 6% and earnings are tomorrow BMO (i.e. today is the last trading before earnings) The LIVEVOL™ Pro Summary is below.

I'm oft asked how pros do earnings analysis and trades using options - Often generating 50% returns in a day... Well, here's the secret sauce... sort of...

The company has traded ~ 5,800 contracts today in the first two hours surpassing its total daily average option volume of 5,388. The Stats Tab and Day's biggest trades snapshots are included (click either image to enlarge). For what it's worth, the largest trades have been Apr 34 and Apr 36 put purchases.

A few things we probably know with earnings vol:
(1) Vol will rise into the event
(2) Vol wil drop after the event
(3) The stock will move that day more than average

What we don't know:
Will the underlying move be large enough to make the vol a purchase... Or, put another way, since the vol is most likely going to collapse after the announcement, shouldn't it be an auto sale?

Let's use FDO as our case study, it has earnings tomorrow BMO and presents an interesting case. We'll start simple - the Charts Tab (just the Vol; with IV30™ (red line) vs HV10™ (white line)).

What we're lookin' at:
(1) The first earnings cycle highlighted (Jul '09) we see the white line (HV10™) spiking past the red line (IV30™). In this case, the ATM straddle the day before earnings was a purchase - i.e. it was worth more the next day. Details to come...

(2) The second earnings cycle (Oct '09) showed the opposite. The the white line (HV10™) stayed below the red line (IV30™) - though note that the red line did dip (as expected) and the white line did rise (as expected).

(3) The third earnings cycle (Jan '10) was just like the cycle in Jul '09 - but even more pronounced. Vol was a super purchase as the stock gapped up on earnings more than the implied vol "forecasted."

Now to the details - the Earnings & Dividends Tab snapshot is included (click to enlarge).

What we're lookin' at:
(1) The top ROW is FDO stock price 5 trading days before earnings through 5 trading days after.
(2) The second ROW are the front 2 month ATM straddles for the same period - focus on purple - the front month.
(3) The third ROW is the implied vol for those straddles - focus on the red - the front month. NOTE: The red line always collapses after earnings - this is called the vol crush after earnings.

What do we see?
For the far left earnings cycle (Jul '09) we see specifically vol went from 59 --> 39 after earnings. 20 vol point drop! So clearly vol was a sale right?... Sell high / buy low?... No?... In fact it wasn't a sale...

The straddle went from $2.15 ---> $3.63; a whopping 68% one day move on the stock gap. A perfect example of how dipping implied vol does not guarantee a good a sale. The implied vol dropped b/c the event (news) was over.

Following similar analysis we can see Oct '09 earnings were a sale: Straddle went from $2.35 ---> $1.32 for a 43% one day drop.

The most recent cycle was a purchase: $1.78 --> $3.42 for a 92% one day gain.

So - what's the point?... There are pro-traders on this floor that only trade earnings. They find companies which have significant odds of being either a sale or a purchase - they play that one day game, and make millions a year. As you see, you're looking at 40% - 90% one day changes in FDO. Of course, if you do it badly, you're out of cash and broke really fast.

What do they look for?
Companies with huge pre-dispositions to be sales or purchases... i.e. 10/10 or 9/10 times where one side was a winner. For the record, FDO isn't one of them...

This is trade analysis, not a recommendation.

Legal Stuff:


  1. would be nice to have you be more specific on the trades to take as opposed to being vague and not taking a side.

  2. Another method that chips away gains is to utilize just the first piece of probable knowledge - i.e. vol will usually rise going into earnings - buy volatility (long straddle or strangle) several days before. If you time it well the rise will defeat the time decay (theta) during the same period. You can also profit if the stock breaks out prior to earnings, but you must close before earnings.

  3. Have to agree with the above comment. I also wonder why the "past is prologue" in the case of earnings performance but is subject to serious mean reversion risk in other contexts... There is a self-correcting risk here, isn't there? Just as Aapl became famous for down-mouthing their forward projections (and traders came to discount every word spoken), its terribly 'public' to think "the pros make millions" doing this (betting on the "come line" when a company has a 'hot hand') and implying that retails should simply stay away. If guessing an earnings outcome was a simple as shorting the company that missed 9 or the last 10 expectations, the MMs would be placing the bid ask a mile wide, and volume would be forced to a trickle.... Finally, most floor traders don't carry large net deltas for exactly the reason you state: doing it badly and you're off the floor. If the "winners" are merely the survivors from an initially large pool, chance remains the most likely explanation... again, same as the 'hot hand' at craps table; out of 1000 players, there will be one that has 36 throws without a 7.

  4. In my opinion longing a straddle and closing it before the earnings can hardly make you money, because its too close to expiration at that time and the time decay is very significant. If you look at graphs showing value of the straddle before earnings, its mostly falling a little because theta decay beats the rise in ImpVol. In trading the earnings, only using the very front month will be possible because the farther months do not react a lot. And by using the front month, significant theta decay cannot be avoided.

  5. But for front month, rising vol into earnings delays the theta decay until event has happened. Estimating vol as cheap or expensive into earnings gets quiet trickier, From IV charts we can easily infer if its priced cheap or expensive when compared against how it behaved historically in past earnings announcements. But often people forget to ask themselves if it is priced so expensive , that implies market is expecting a higher risk of strong moves. If it's priced cheap, market's expectations of strong move are low. Don't forget even if a stock moved big few quarters ago after earnings announcement, market will price that risk in present earnings expectations.
    To me that's where your expectations should come in , as a slight miss or beat can make vol priced cheap.