Clarus Financial Technology

Variance Swaps and Other Equity Derivatives on SDR

Tod Skarecky

You’re likely aware that we’ve been the de-facto “janitors” of the SDR data.  However, to date, our focus has been on swaps traded in Interest Rates, Credit, and FX.  This week I turned to the world of equity derivatives to see what’s lurking in there.

To begin with, I had to grab some raw data.  Out of the 4 SDRs that we monitor (DTCC, CME, ICE, and BSDR), the only place I could find any equity derivatives data is on DTCC.  So I gathered up all of trades reported in March 2016, a mere 847,648 trades!

First thing I did was remove the 7,860 trades dated prior to March 1, 2016 (late reports, including 20 trades reported with execution timestamps in the year 1970!).  This brings us down to just 839,788 trades.

Being the good janitors that we are, I next went on to observe the various data types:

For today’s analysis, I don’t really care about terminations and other post-trade data, I only want to see “New Trade” activity.  This means I get to delete 68% of the records – and back down to a more manageable 266,045 trade records.

Next up, let’s look at the Cancellations and clean up the cancel & replace garbage.  Of the 266,045 Equity Derivatives that we’re playing with:

We need to be careful here.  First, I removed all “NEW” and “CORRECT” trades that had been cancelled by a later report.  Then I remove the CANCEL messages.  This leaves us with just those “NEW” and “CORRECT” trades that are currently unique.  Great, now we are down to just 203,285 trades, and this should be the universe of new equity derivative swaps dealt in March.

HEADLINE STATISTICS

Let’s look at some headline Stats after doing this first round of culling:

So what currency are all of these equity derivatives denominated in?

List of Currency Denominations for Equity Derivatives (March 2016)

So just slightly more than ½ the trades were in USD.  The next largest amount was in Korean Won, which is an active equity derivatives market on their own shores.  There were some equity derivatives denominated in currencies that I did not immediately recognize; of note, there were 6 trades in the Pakistani Rupee, and 4 in the Uganda Shilling!

Let’s also not worry about the 1,494 trades that do not have a currency – whoops!

PRODUCT TYPES

So the more interesting aspect here is what kind of trades are reported.  Here is a breakdown of the Taxonomies (product classifications) within the Equity data:

List of Taxonomies for Equity Derivatives (March 2016)

If those taxonomies don’t make much sense to you, you’re not alone.  At least I’m with you.  Luckily, ISDA has a decent description of some of the taxonomies here.

Generally speaking, the translation seems to be:

CAN WE MAKE SENSE OUT OF ANY OF IT

Looking at the data, I was drawn to the “Single Index” data.  Primary reason being the Single Name data gets massive, with thousands of underlying names; whereas for indices, I can make out some common standards such as the S&P 500 index.

So if I look at the top three Single Index activity, I need to choose between:

It turns out, the cleanest and most manageable data set is the “Swap:ParameterReturnVariance:SingleIndex” – aka Variance Swaps.  So let’s look at those, and walk before we run.

WHAT IS A VARIANCE SWAP

A variance swap is a pure play on volatility.  Two parties agree to exchange a future payment based upon the actual (realized) volatility over a period of time.

For example, if implied 1M volatility in the S&P 500 is 15%, and I think the actual realized volatility will be higher, I can enter into a variance swap struck at 15.  If the actual observed volatility over the next month ends up being 20%, I make money on the difference (20 – 15).

More sophisticated readers will see that I have mixed the terms variance and volatility there for convenience.  The payout of a variance swap is actually based upon – wait for it – the variance.  So to be clear, we need to remind ourselves that Volatility (σ) is a measure of the standard deviation, while Variance (σ2) is the square of volatility.

Some quick highlights of variance swaps:

Of course, both parties need to determine and agree the observed variance using official daily closing prices in order to determine the final observed variance.

If you want to get more under the hood, here is the JP Morgan document I referenced for a refresher.

LET’S SEE SOME REAL TRADES

So let’s go back to our 726 variance swap trades in march 2016, and see if we can make sense of them.  For starters, here are the indices we see trades against:

List of Indices for Equity Variance Swaps (March 2016)

726 trades in the month of March works out to be 33 trades per day.  In the S&P index, the 397 trades arrives at a figure of 18 trades per day.  Not bad.

As is typical with SDR trade reporting, the data is all over the place:

Alas, after all of the scrubbing and data enrichment, we get to a pretty clean set of data.  Here is a sample ticker of the cleansed data:

Cleansed Trade Ticker for Variance Swaps

Some help on the fields:

FRUITS OF OUR LABOR

So now we get to do some nice aggregation.  For starters, how about asking how many trades are dealt each day, and average trade size:

Trade Counts and Vega Traded per day (March 2016)

And how about which tenors trade:

Distribution of Variance Swap Tenors (March 2016)

So:

Also:

As a quick sanity check, we can also throw together average strikes (volatility) over the course of the month:

Average Strike Prices (Vol) for S&P Variance Swaps (March 2016)

Granted this term structure is computed from prices observed over an entire month, but I thought it might at least demonstrate we’re on the right track.  The structure is inverted from what I had expected (would have expected short term vol to be higher) but then again I might be guilty of thinking in terms of implied vol rather than realized vol.

And lastly, let’s take our most active tenor (6M) and look at its price history over the course of March.  This is 84 trades:

Vol for 6M S&P Variance Swaps (March 2016)

Hmm, this seems to say that S&P 500 vol traded down over the course of the month.  OK I’d believe that.  But I am a bit suspicious of the large variance (excuse the confusion) in the prices.  Some of this is likely due to the large “broad” brush I have applied to grouping trades into tenors (For example I’ve combined both 170 and 200 day trades into “6M” in my sample).

It’s a decent first step.

SUMMARY

We’ll have to leave it there for today.  Few things I’ve learned:

Bonus Question:  I know we looked mostly at Indices, but why are there Single Name Equity swaps on the (CFTC) SDR in the first place?  Don’t they belong on the upcoming (SEC) SBSDR?  I don’t know the answer – so please let us know.

And as always, get in touch with any of your comments or suggestions.

Stay informed with our FREE newsletter, subscribe here.

Exit mobile version