Panel discussion: Derivatives pricing

Representatives from Calypso, James Caird Asset Management, Numerix and SunGard discuss the latest developments in derivatives pricing.

Derivatives pricing has been one of the most acute challenges facing buy-side firms of all sizes since the start of the financial crisis, and if the slew of recent derivatives research is anything to go by, buy-side firms are not shying away from using derivatives in their portfolios. There have, however, been some significant strides forward in this area, especially in the past two years from pricing data specialists and technology vendors. But arriving at a single, definitive price is only part of the challenge; the other, tricker aspect revolves around the ability to provide accurate, reliable, repeatable and transparent derivatives prices, supported by well-defined methodologies, all of which need to be underpinned by technology. A panel of experts in this field give their views on how this may be achieved.

Victor Anderson: Before we kick off today's discussion, I think we should set the scene in terms of what the current state of play is regarding the pricing of derivatives and the valuation of portfolios containing such instruments. Olivia, do you believe that buy-side firms are being well served at present in this particular area and are they being supported adequately in terms of their business?

Olivia Erby: We are a credit fund predominantly and so a large proportion of our portfolio is in derivatives. On the whole, we value that portfolio daily and produce a daily P&L for our traders, and Citco, our administrator, does the same thing. Do I think we're well supported in terms of being able to do that efficiently and are we producing independent valuations? Yes, on some level, but not all levels.

I think that it depends on the liquidity of the products you're trading, and the sophistication of the system you're using. For the majority of derivatives that we trade, we are able to value those on a daily basis. There is definitely a proportion of our portfolio that is illiquid - it isn't quoted as regularly in the market, and for those instruments, receiving prices from independent sources is just not something that's available to us. I think there are other ways that we could get more independent valuations on those instruments from third-party providers, but we would then have to go through the process of evaluating their models to be able to have any comfort with the valuations that are being produced. There's also a cost involved in that, and a level of time and how quickly third-party valuation services can turn that service around to us.

Paul, is this something that you're hearing from your buy-side clients?

Paul Miller: Absolutely. But if you're a large asset manager and complex derivatives are only a small percentage of what you do, then it is quite a big challenge for the buy-side firms as to where they get their services from and what infrastructure they keep internally to manage that part of their business. What's the cost of that? There's a big overhead in today's market.

Nick Haining: I would second what Olivia said in the sense that we're seeing a differentiation between the traditional processing of vanilla or slightly complex trades where there's a high volume of activity compared with the more complex trades. On the more complex side, which is pretty much where we come from, we're seeing a need for some hand-holding - not just in terms of processing, but actually in terms of understanding.

There are two pressures. One is coming from the regulatory authorities, and the second is being able to understand the transactions you're buying. So that's an internal requirement in order to understand the sensitivities of the trades. The resources that exist on the buy-side are not as great as the sell-side. I think where the challenge lies for the likes of the providers is to do some hand-holding - not just on a processing basis but on a quantitative basis to support the buy-side in understanding the trades they have taken on, the sensitivities, and helping them deal with the regulators and accounting.

Gavin, do you agree with Nick's assessment?

Gavin Lee: Absolutely. We have a valuation service that we provide to our clients and this is one area that we are getting heavily involved in right now. We are seeing the need for consultative services on both sides of the industry.

It's about getting financial engineers in the door to help explain where the valuations are coming from and how they're calculated; understanding the numerical techniques; understanding the modelling assumptions as well - because obviously with a model you have different assumptions you can make; and once you've made those assumptions understanding the limitations of those assumptions within the modelling framework. The transparency has to be there today, which was not the case three, four, or five years ago.

You mentioned transparency, Gavin, which is something that's going to come up later in the discussion - not being totally reliant on a number, but having the ability to drill down and look at the assumptions and the way those numbers were extrapolated. Dave, what is Calypso's take on the state of play on the buy-side regarding the pricing of derivatives instruments?

Dave Wong: I think it depends on who you're talking to. The panellists all alluded to this earlier - it depends on whether you're talking about institutional asset managers or hedge funds. When looking at institutional asset managers, for instance, they are going to be dealing in much lower volumes, but they're looking to basically enhance the yields on their portfolios. They're going to be looking at more of a valuation service-type offering. But when we're looking at hedge funds - the natural consumers of derivatives and the natural innovators - it's not just about the end number, but really the transparency of that number and making the process repeatable. One thing that buy-side firms are asking us for are ways of not just bringing that information into a single platform, but being able to rationalise that information. How do they do comparisons? How do they flag outliers? And how do they do that in a scalable fashion so that as the complexities and volumes grow, they don't necessarily have the overheads growing at the same rate?

What options are available to buy-side firms looking to increase transparency around multiple source pricing, valuations, and the risks within their derivatives portfolios?

Nick Haining: If one looks at the delivery of transparency of modelling and pricing, there are two mechanisms you can offer. One is the valuation service, where essentially what you're doing is providing a price and an explanation of your methodology and calibration. In other words, you're providing the information to the purchaser of the service about how you came up with that price. I don't think it's acceptable any more to just go and dump a price on somebody. You must allow them to understand where it's coming from, within reason, and the reporting requirements that come out must be clear. In other words, the valuation service must provide the clarity and transparency that people need in order to explain to both accounting and risk management how those prices were arrived at.

The second delivery mechanism is for the more sophisticated and larger entities where you're providing software on-site. There you're probably providing models - whether they are integrated models or independent models to the main system. I guess a one-stop shop is nice, but isn't always achievable. If you're delivering the models themselves, you must provide flexibility of understanding of the models. You must provide the papers and you must be clear where the white papers are coming from. You must explain the methodologies that are being used, how the Monte Carlo has perhaps been accelerated or whatever. So you're providing essentially a whole functionality to the risk management group internally, so they can understand what they're using. It comes back to the point about transparency, either as a delivered service explaining what you're doing, or explaining the tools that you're delivering. Both have the same end result, but one is used internally, and the other is an external service.

Gavin Lee: I think that's true. One thing we're seeing is that with many buy-side firms, they're not looking for a single valuation source. We're often asked by firms how they can integrate more than one valuation provider. They need to do a validation. And that comes with an overhead, because once you create a validation process, you then have to have a process to investigate any differences.

You're going to have to set tolerance levels, you have to justify those tolerances, and you set a whole process in action. And that process needs controls around it. It needs people, and it needs systems; it's not something that can just happen. It's important that we understand as we go through this process that we give more transparency, but we also increase overheads as well.

What technology does a buy-side firm need in order to support multiple source pricing, increased transparency and improved risk management of their derivatives portfolios?

Olivia Erby: What we'd like to have is a full front-to-back system where everything from trade entry through to pricing and risk is maintained within one system. We're fortunate that we're a relatively new fund and so we installed Calypso which provides us with full front-to-back functionality. We're multi-assets and so we're able to value the majority of our portfolio using the models that are already within Calypso. But there are assets outside of that where we would look to plug in other pricing models. We'd like to have the functionality where we can take in multiple pricing sources, and have the flexibility where we can input our own logic as to which prices to take.

We'd also like to provide transparency to our administrator as well; we're happy for them to see exactly what we're doing at all times. But all of that comes at a cost, and it's not just the cost of the software, but the people to support that software in-house, and the knowledge base for that as well. Even for us, as a relatively large hedge fund, that's a huge overhead. So whenever new products come out, we need to make quick changes; we have to go back to our software provider and work with them, although we're not their only client.

Nick Haining: I think Olivia summarised the processing side of things: front-to-back, straight-through processing, and so on. That is very much what most people on both the buy-side and sell-side are looking for. It's already been alluded to, though, that there are a lot of specialist requirements out there, specialist firms, specialist tools, and so on, and that's the key, given the cost constraints of the buy-side. How much does it cost to link those specialist items? There are a couple of technological aspects here that are necessary.

In terms of structuring and deal-uploading for the more complex deals, you need to have some technology, maybe something like XML representation, whereby if you have a tool that can price a deal, how do you transfer that trade, that model, that information up into a processing system? One of the ways is - and we use this in our systems - an XML representation so you can pass that up. It's a very IT-centric issue, but it's a key issue that reduces integration cost.

The second thing is that if you have libraries that need to be integrated, for the specialist aspect that is slightly outside the standard STP systems, you need to have libraries that have very good APIs (application programme interfaces) that are easy to integrate. You need to have either Java or C# - Calypso uses Java and NumeriX's libraries are in there because it's easy to integrate into it. In other words, you need to have easy integration that's not going to take hundreds of man-days, but is in the realm of 50 or 40 man-days.

How can multiple source pricing feeds be managed effectively from a technology systems perspective? What about the overall cost of integration and the management of information flow?

Nick Haining: I think it's about having the ease of entry to a core channel. In other words, if you have 17 different systems, then you've got 17 different integrations. You must have some sort of core aspect in there. Obviously James Caird has chosen to go down the Calypso route - everyone will have their system that does that. The key is not going off and buying software that will cause you a year's implementation. So your multiple pricing sources and your cost of implementation will be three times the licence fees. Before going and doing it, think carefully about which systems match with which, think ahead of time - that's in terms of on-desk things.

If you're looking at external systems and external valuation service, it's really a case of information flow and not so much a systems issue. It's about whether the information coming into your organisation is satisfying your needs. Will risk management get the exposures they need and will the trading desk get what they need, and so on? I think it's about thinking ahead and not buying independent software or buying independent valuations that don't satisfy all the needs and then you have to reinvent the wheel and it costs twice as much.

Is that an experience James Caird is familiar with, where you're sitting with disparate systems and there are inoperability and implementation issues?

Olivia Erby: Any software implementation is, from our experience, a long process. We implemented Calypso in six months and I think that was a fairly short time-span. And I think that meant that we came across issues as we went live with using the product as well. We have system inflows and outflows into and out of Calypso. Some of them are easier than others depending on how established that third party is and whether they've got an existing relationship with the software provider.

That's definitely a selling point when we look at outsourced valuations or any type of service that we're thinking of plugging into Calypso. Has it been done before? How long will it take? It definitely puts us off if we can see that it's going to take three months: we really want to see upfront what we're going to get out of that service, which is sometimes difficult for third parties to provide us with.

Gavin Lee: One of the things we offer is the independent valuation service and we have the same issues: we have to take data feeds in from the market data vendors. They're required to feed our modelling so that we can provide the valuations onto the clients. Not one of them has the same API and so you could end up with between five and ten different data feeds coming in, each of them completely separate, each of them completely independent, and each of them requiring their own technology and their own API and their own little idiosyncrasies that go with it. And it becomes very difficult to manage.

Is that going to be addressed any time soon do you think? Are the technology vendors or the data providers actually sitting around a table saying how can we make it easier for buy-side firms?

Gavin Lee: I don't think the data vendors are doing that, but I think the technology vendors are trying to do that.

Dave Wong: I completely agree with that. With Calypso's SaaS offering and with our Fast Track method of deployment, I think one of the things it allows us to do is create a higher touch relationship with our clients. Typically when we were just dealing with software implementations, it was a matter of putting people on-site, observing problems going back, and reproducing problems. Now with the ASP environment, we're able to basically see things as clients see them themselves, which cuts down the time or the feedback loop in terms of getting issues back into our engineering organisation.

In terms of addressing Olivia's earlier comment about another client asking for something from Calypso, Fast Track is our way of standardising all of that market feedback into one place and distributing it to all of our clients.

Paul, a final word from you? I predict that you are going to say something about the issue of standards.

Paul Miller: I've probably said as much as I can on standards, which it seems is always up to the end-users to pick up the mess, regardless of whether you're a distributor or whether you're an end-client. As far as I'm concerned, there's a wider question though. If you're a derivatives specialist, I think there are things you can do in terms of your choice of supporting technology, but if you're not a specialist and you use derivatives as part of an overall asset allocation - part of an overall strategy for a pension fund or a large institution, for example - then all of these problems we're talking about are a sub-set of the wider challenge you face across all the different instruments and asset classes.

It is a wider issue because it depends on where you're looking at it from: from a specialist viewpoint, you've got one set of problems which is you haven't got the scale so therefore you've got to do a lot of this work yourself; but from a larger viewpoint, this is just one small part of the problem, and then you've got that much wider integration across multiple assets and multiple systems.

The panel

Victor Anderson, Panel moderator and editor of Buy-Side Technology magazine

Paul Miller, Buy-side consultant

Olivia Erby James, Caird AM

Nick Haining, NumeriX

Gavin Lee, SunGard

Dave Wong, Calypso Technology.