Monday, July 9, 2007

Implied Volatility and Earnings Spikers

A reader asked for some more details about how I use implied volatility (IV) and/or historical volatility to help identify companies with a high potential to spike following an earnings announcement. Specifically, she was looking at the 17% gains in Schnitzer Steel (SCHN) today and wondering if IV could have tipped her off to the probability of a big move.

Regular readers will probably recall that I beat the earnings spiker horse rather severely during the ill-fated (more on their part than mine) CNBC Million Dollar Portfolio Challenge, but for anyone who is interested in some details, I laid out the bulk of my thinking in “How to Find the Spiker Before the Earnings Announcement.”

Since part of the query touched on why I thought IV was better than beta for determining volatility (past and future), here are three reasons why I think IV is superior to beta:
  • Yahoo Finance, Google Finance, and other data providers sometimes list betas of 1.0 for issues they apparently have not calculated a beta for, particularly newer issues and foreign stocks
  • highly volatile stocks that go in the opposite direction of the market for awhile can sometimes have low betas -- think small oil/gas exploration companies, gold miners, etc., but also consider that some tech stocks may countertrend for a long period and thus acquire a smaller beta than their volatility would suggest (historical volatility would be a better number to watch here, because it focuses entirely on the magnitude of the moves and does not care whether these moves are correlated to the broader markets)
  • implied volatility is forward looking, so it automatically adjusts to account for scheduled earnings announcements, a pending FDA drug decision, a legal issue that is due to be resolved, buyout rumors, terrorism, violence in the Middle East, a hurricane that is bearing down on the US, etc.

Getting to the meat of the question, since I am looking forward in time to earnings, I pretty much ignore historical volatility and focus entirely on IV. I usually try to target the top 10% or so most volatile companies that are due to report in a 24 hour period, so that if it is mid-June and there are very few reports, I might look at IV as low as 35 (I generally consider 40 to be a minimum IV), but by the end of July to early August earnings peak, there will be so many small and extremely volatile companies reporting that it might be possible to screen out all companies with an IV below 50. Also, once you get over about 60 or so, I am not sure that a higher IV really translates to incremental future volatility in the short-term (unless we stray from earnings and talk about FDA decisions, etc.)

Regarding specific numbers, I use the current IV Index call number (the 44.20% from the SCHN iVolatility link), but the current IV Index put and current IV Index mean are usually so closely correlated that it doesn't matter which one you pick -- as long as it is a current number.

I also recommend that readers consider looking not just at the raw numbers but also at the 12 month volatility charts (such as this iVolatility chart for SCHN) where it is often much easier to visualize the size of the current pre-earnings volatility run-up – and also compare it to similar up-trending volatility patterns that preceded earnings in the past few quarters. In the case of SCHN, you can see a big jump in volatility during the past week and a sustained move since about mid-April. Implied volatility may not have been screaming “Buy!” in an unambiguous manner, but one can reasonably argue that at an 11 month high just prior to this morning’s earnings announcement, it was warning of the increased possibility of a big move in one direction or the other.

Finally, I would be would be remiss in not reiterating that directional earnings plays are highly speculative and usually carry a formidable risk/reward profile, so I recommend that anyone who plays the earnings lottery considers limiting their exposure with an options play or by making a bet on volatility instead of direction, as volatility typically – and much more predictably – decreases 15-20% the day after earnings are announced.

6 comments:

  1. There's an article in this week's Barron's that might interest you. The gist of it was that VIX calls are the best thing ever invented. It seemed to have a lot of mistakes, which is weird considering Larry McMillan wrote it. I thought he was an options expert. I stand corrected.

    ReplyDelete
  2. Bill, one thing you said in your article was thought provoking, and that was the i.v. of the calls and the puts being the same. They have to be, otherwise something is up with the stock; usually the stock has such a huge short interest that there is no stock to borrow. I wonder if that characteristic--different i.v. for mean calls and puts--adds to the potential for a spiker?

    One other thing, i.l. said that Larry's article had "a lot of mistakes". I am curious what i.l. finds mistaken. Not trying to be a wise guy. Just curious.

    ReplyDelete
  3. FWIW, the only mistake that jumps out at me in the Barron's article is in the first sentence, where McMillan (or some junior editor) cites 1,600+ trading days since the beginning of the March '03 bull market when actually it is 1,600+ calendar days.

    I think I will probably post about McMillan's thinking here tomorrow.

    Anon, regarding the pricing of puts and calls, there is usually a small difference, but rarely a significant one. Today, I see that DNDN's put IV is about 1.7% higer than the call IV, per iVolatility's numbers.

    ReplyDelete
  4. Ok, one mistake at each line was a bit of an exageration.
    However, this was shocking:
    "When they expire (VIX options), they settle to the price of the index."
    Don't they settle at the VIX futures settlement price and NOT the VIX index?

    ReplyDelete
  5. I believe that McMillian is correct here. According to the CBOE's VIX Options FAQ, VIX options prices essentially revert to the VIX index price at expiration.

    From the FAQ linked above:

    "5. Will VIX options always reflect current, real-time, VIX values?

    Probably not, at least not until you get close to expiration. The underlying for VIX options is the expected, or forward, value of VIX at expiration, rather than the current, or "spot" VIX value. This forward value is estimated using the price quotations of SPX options that will be used to calculate the exercise settlement value for VIX on the expiration date, and not the options used to calculate spot VIX. For example, VIX options expiring in May 2006 will be based on SPX options expiring 30 days later - i.e.; June 2006 SPX series. In fact, June SPX options do not even enter into the spot VIX calculation until April 17, 2006.

    Some VIX options investors look at the prices of the VIX futures to gain a better general idea of how the market is estimating the forward value of VIX.

    VIX option prices should reflect the forward value of VIX, which is typically not as volatile as spot VIX. For instance, if spot VIX experienced a big up move, call option prices might not increase as much as one would expect. Depending on the value of forward VIX, call prices might not rise at all, or could even fall! As time passes, the options used to calculate spot VIX gradually converge with the options used to estimate forward VIX.
    Finally, at VIX options expiration, the SPX options used to calculate VIX are the same as the SPX options used to calculate the exercise settlement value for VIX options."

    ReplyDelete
  6. Bill,

    I realize that the at-the-money options almost always trade at the same i.v. The only time the puts and calls deviate is when the stock has a sky high short interest--so high that you can't find stock to borrow so that you can't arbitrage the options and the stock. That means put and call pricing can deviate to the point that implied volatility is wildly divergent, even at the same strike.

    That's my point. When i.v. gets to the point that you see it as a potential spiker, AND you see the implied volatility of the puts and calls deviate, you know you've got a situation where the short interest is so high that you can't short the stock because you can't find stock to borrow. That means if you get a positive earnings surprise, you've got the potential for major fireworks.

    That's what I mean. Don't just look for a high implied volatility like you suggest, but also look for a discrepancy in the put and call implied volatility for an even bigger spike.

    Just a thought.

    As far as VIX options are concerned:

    Settlement of Option Exercise: The exercise-settlement value for VIX options (Ticker: VRO) shall be a Special Opening Quotation (SOQ) of VIX calculated from the sequence of opening prices of the options used to calculate the index on the settlement date.

    Settlement of VIX futures: The final settlement value for VIX futures shall be a Special Opening Quotation (SOQ) of VIX calculated from the sequence of opening prices of the options used to calculate the index on the settlement date.

    That last section means that both the VIX options and VIX futures settlement is based on SPX options.

    ReplyDelete