The Gateway to Algorithmic and Automated Trading

VolX: Realising Volatility

Published in Automated Trader Magazine Issue 20 Q1 2011

The Volatility Exchange (VolX®) bears little resemblance to a conventional exchange and is probably best thought of as being a virtual exchange. Its realised volatility contracts can be based upon products already listed on any existing exchange or trading platform to provide a new market complex of uncorrelated instruments. Automated Trader’s founder Andy Webb recently caught up with Robert Krause, the exchange’s CEO, and Chief Business Development Officer Charles Barwis, to find out more.

Robert Krause

Robert Krause

Andy Webb: What were the origins of the realised volatility contracts the exchange provides?

Robert Krause: In 1993 the CBOE contracted with Robert Whaley to design some kind of volatility product. This resulted in the VIX - essentially an index based on implied volatility.

At the time, I didn't believe that a contract on implied vol was necessarily the best direction to take, and so I started thinking about the possibility of an instrument based on realised volatility instead. Volatility and variance swaps (which settled to realised volatility) were already being traded over the counter, so it seemed to make sense to try to standardise a product that would mimic a vol swap, rather than try to develop a completely new approach from scratch.

However, I didn't actually put anything down on paper until 2000 when I attended a conference and sat next to a patent attorney. I asked him if it was possible to patent such a concept, and he said yes, so we got the application underway. That wasn't a quick process; it took until 2008 for the patent to be granted.

Andy Webb: You've taken an unusual approach in that you aren't building a physical exchange from scratch with all the infrastructure overheads that implies. Was it always obvious from the outset that this was a concept that would work best as a partnership with existing markets?

Robert Krause: It was really a case of research. Even by the time the patent was awarded, the most suitable business model for exploiting it wasn't obvious. It was only after we started talking with exchanges and others in the industry that the idea of conducting business in this manner emerged.

Charles Barwis: Part of the key to the success in launching a new product is distribution. How easy is it for market participants to get access to your product? We aren't really directly competing with anybody with this product, so we don't have to set up our own infrastructure, matching system, and clearing facilities, etc. Whenever you do have to go down that route you end up in the classic chicken and egg situation. Some clients will want to wait until the liquidity develops before they commit to the development work to write to the API or pay for a new ISV connection. And that behaviour in turn reduces the available liquidity needed to get the ball rolling. We don't have to do that as we have a symbiotic product that is complementary to whatever the underlying asset happens to be. Therefore, we can work with entities that already have distribution, be they exchanges, over-the-counter firms, or whomever.

Robert Krause: We also came to the conclusion that if we simply arranged straightforward licensing deals with individual exchanges or other parties then we would lose control of the message and the final product. It would essentially no longer be our product, although it might still carry the Volatility Exchange name. Licensees would be able to tweak the formula and so on, immediately giving rise to fungibility issues and making inter-market arbitrage less attractive. We therefore decided that the best way to exploit the product would be to brand it, define it, and standardise it. (We have made one exception to this rule for a variety of business reasons, but going forward we do not intend to use a licensing model.) As a result, it doesn't really matter whether the underlying market is a currency pair or poultry - the formula always remains the same.

Andy Webb: Is it just the method of calculation you wanted to keep standardised?

Robert Krause: No; we also wanted to make sure that the product always expired on the same day as any related options, so as to provide the best possible hedging tool. From the volatility contract's perspective, it doesn't really matter whether the contract expires at one time or another, but if it expires at the same time as the options, then that is far more useful to anyone trading options and having a volatility exposure that he or she needs to hedge.

Andy Webb: What was the initial response when you started talking to exchanges about the new contracts?

Charles Barwis: We received consistently favourable feedback, whether we were talking to an exchange, market-maker, hedge fund, retail firm, or even CFD firm. The first two groups we approached were market-makers and exchanges. Market-makers have been extremely willing since the outset to make markets in different groups of realised volatility contracts. The exchanges have also been very positive, because they can immediately see the potential.

Andy Webb: You mentioned hedging earlier. How do you see the application of realised volatility contracts for this differing from existing hedging methods?

Robert Krause: Market-makers have all kinds of risks in their option book - delta, gamma, vega, theta, etc. At present they hedge their delta risk with futures or the underlying. In so doing, they are effectively trying to marry a curve with a line, which isn't so easy.

You can more or less draw the tangent to the curve, which is what you get when you hedge with futures, but there is a constant adjustment required - and that is delta-neutral hedging.

When you delta-neutral hedge an options book you are still left with the vega and gamma risk. The vega risk is the risk of gain or loss resulting from changes in implied volatility, while the gamma risk is risk of gain or loss resulting from (sometimes abrupt) changes in delta. At present, vega and gamma risks can only be managed with options. For example, if you have large positive vega, you need to sell some options to get flat. (If you are a VIX trader you can obtain a measure of mitigation by using the VIX for the vega, but to date, this is viable on only one asset, the S&P 500.)

The realised volatility contract effectively has both vega- and gamma-hedging qualities embedded in it, which opens up further hedging opportunities/efficiencies. We have conducted research that shows that straightforward delta hedging typically reduces the risk of an options book by between 50% and 75%. However, if you add a volatility contract hedge on top of that, you can reduce the residual risk by a further 50-75%. Therefore early adopters of realised volatility contracts will enjoy reduced hedging costs and more accurate hedging, thereby allowing them to make slightly sharper markets than their peers.

It's also worth noting that option market-makers are in a unique position regarding realised volatility contracts. Not only are they likely to be the first adopters to make markets in the product, they are probably also the primary natural hedging users of the product.

Andy Webb: What about directional or speculative trades?

Charles Barwis

Charles Barwis

Robert Krause: There is an interesting day-trading possibility. If the market moves 2%, it doesn't matter whether the move is up or down because it doesn't affect the calculation; it is still just a 2% move. But what if you are up 2% at midday? Let's assume the same scenario that the market could go up or down 2% in the afternoon. If it goes back down, then we will be unchanged on the day, but if it goes up, then that represents a 4% move, which is considerable. Therefore there is a very strong directional trade available intra-day. That trade isn't available inter-day because at the end of each day you mark the closing price and start again, but intra-day there is a very strong directional component during the Realised-Volatility Period. However, I would stress that this is only available occasionally during the Realised-Volatility Period (see box "VolContracts" for definition and mechanics of the contracts).

Andy Webb: And arbitrage or spread trades?

Charles Barwis: There should also be a near arbitrage (it isn't perfect) between the one- and three-month contracts; and if we proceed as planned with the roll-out of a yearly contract as well, then that will obviously provide further spreading opportunities. We are also talking with some strategic partners about possibly rolling out a suite of stock indices, so there could be some interesting spread trades between the realised volatility contracts traded on each of those indices.

Another interesting characteristic of these contracts is the range of behaviour they display. For example, the one-month contract is very volatile compared to the three- or twelve-month contracts. The one-month VolContract should be highly volatile, in the order of 80%, 100%, and sometimes a lot more, depending on the asset. The three-month vol of vol is about one-third of the one-month version. And, the twelve-month product is about one-third of the three-month.

Robert Krause: This is basically down to the averaging period. There are days when markets go crazy and make a big move, but if you are averaging

over more days than that, the move is dampened. As a result, this creates three totally different trading animals. The one-month is very volatile, while the three-month is more akin to the volatility of many commodities and stocks, and the one-year is closer to the risk of a government bond in terms of vol.

VolContracts

VolContracts (Realised Volatility Contracts) are futures-like financial instruments that capture the inter-day realised volatility of an underlying asset, index, or instrument. The one-, three-, and twelve-month contracts allow participants to hedge, spread/arb, or speculate on realised volatility in the short-, medium-, or long term.

Contract expiry dates typically coincide with associated options expirations and the contacts are applicable to any asset classes with reasonable liquidity, including equity indices, currencies, rates, and commodities.

The formula used to settle any VolContract on its expiration day is as follows:

Where:

Vol = Realised Volatility

n = number of trading days in the period

Rt = continuously compounded daily return as calculated by the formula:

Where:

Ln = natural logarithm

Pt = Underlying Reference Price at time t

Pt-1 = Underlying Reference Price at the time period immediately preceding time t

Availability

The first listed instruments based on realised volatility are due to launch Feb. 7, 2011. VolX and CME Group announced that CME will begin offering FX VolContracts™ - a set of FX realised volatility futures based upon the major currency pairs that it lists. These contracts will be listed with, and subject to, the rules and regulations of CME.

For further details and contract specifications, see http://www.cmegroup.com/trading/fx/fx-realized-volatility-futures.html. For more on The Volatility Exchange, realised volatility and VolContracts™, see www.volx.us.

Andy Webb: Is there an implication for the trading demographic here?

I believe day traders and retail traders are very interested in the directional aspect of volatility, but anyone who is not an options market-maker struggles to access the volatility market at present. If you are an options trader, it is labour intensive to create a pure volatility exposure with options; and once created, positions require constant monitoring - especially if you are short volatility.

Charles Barwis: Right now, if you want to trade volatility in gold, for example, or any other asset other than S&P or EuroStoxx, you are either an options trader or you must have access to the OTC volatility/variance swaps market. All of the other asset classes are therefore, in a sense, closed to professional trading firms and retail traders that don't trade options, or don't have access to volatility/variance swaps.

The fund community is in a similar situation and they have a strong inclination to trade volatility. In general, they see volatility as an interesting asset class in its own right because it is uncorrelated with other products they trade, which is always an attraction.

There is also a subset of the fund community that has emerged in the past few years that exclusively trades volatility. Some of that community has access to volatility swaps, which are similar to realised volatility contracts. They tell us that while they will continue to trade volatility swaps, our products are attractive because they are exchange traded and reduce counterparty risk while increasing transparency. In addition, they are looking for more products to trade in the volatility space.

However, the majority of funds trading volatility don't have access to the volatility swap market, perhaps because they lack the capital to access the market, or their mandate does not permit it, or some other reason. For these funds, the more products they can trade in the exchange-listed volatility space the better.

Andy Webb: So how would you describe the essential difference between a realised vol contract and an implied one such as the VIX?

Robert Krause: VIX is based on implied volatility, which is essentially the market's forecast of future volatility. Theoretically the market is supposed to be assessing what the risk is in the future and putting a price on the option based on that. If you trade a futures contract that settles to implied you need to bear in mind that futures are also a forward forecasting mechanism, because everyone in the futures market is trying to figure out where the contract will be at expiry. Therefore, a futures contract on VIX is trying to forecast the VIX index, which itself is a forecast of future volatility. So what are you trading when you buy a futures contract on the VIX? You are effectively trying to forecast the future forecast of volatility! This is an intangible result that would leave most market participants wanting.

A vital point to grasp is that if I think 2011 is going to be really volatile and I trade the VIX, it doesn't matter whether 2011 actually is really volatile, but instead it matters whether everybody thinks it's going to be volatile. That's a key difference.

Another interesting distinction is that a VolContract™ actually incorporates implied volatility into its pricing as well. As far as risk goes, VIX tries to hedge your implied volatility risk (or vega risk, in option terms), while the VolContract™ hedges both vega and gamma risk.

Andy Webb: So where do you see the opportunities with realised vol contracts for automated traders?

Charles Barwis: I believe that in view of its relatively high volatility, the one-month contract could be a logical point of interest for high-frequency traders.

The arbitrage opportunities among one-, three-, and twelve-month contracts could also present opportunities for some specialised trading firms. Then there is the opportunity to spread-trade across realised volatility contracts based on various underlying instruments, or with those underlying instruments themselves. Finally, there is the chance to take directional positions in instruments in any time frame, while simultaneously hedging out the risks of exogenous vol shocks in those instruments. That minimises the need for human/manual judgement on every trade where questions arise such as: "Has this trend now really broken down so I need to follow the stop loss, or is this just a one-period vol spike?"

Robert Krause: There is also a very strong embedded directional component within the trading day that is not available over longer periods because of the very nature of the contract design (just using daily settlement prices in its calculation). This effect would allow day traders the ability to tap into the liquidity of the futures market for spreads, offsets, and additional trading opportunities with VolContracts. Automated traders would be best positioned to take advantage of this embedded directionality.

Robert KrauseCharles Barwis