Data processing power has helped to transform the financial industry, thanks to the rise of GPU-powered applications; GPUs are Graphics Processing Units, commonly used by video game developers. Today, not only hedge fund managers but all asset managers, are looking for ways to deploy next generation technologies to improve the way they absorb and process vast data sets. The aim being to improve data management, liquidity management, cost of execution, margin management. The list goes on.
Ultimately, technology advances have pushed fund management into a new data-rich realm.
And in some ways, this is unsurprising given that we experience a data-reach world in our private lives, where every leisure activity now appears to have an associated application giving us a stream of information we never before knew.
“At this stage, I think all market participants expect more; greater returns, lower fees, greater transparency,” says John Hack, Senior Vice President and Head of Technical Relationship Management and Product Consulting at Arcesium, a New York-based middle- and back-office technology solutions and operations provider.
“It’s analogous to increased expectations we have in our personal lives. You can get all manner of data today through your mobile phone – so shouldn’t that extend to gaining insight into your hedge fund investments?”
Trusting data is vital for data management
Hack believes that the next iteration of technology will be more self-service in nature, where all key players in the hedge fund industry, including the front-office team, will have greater access to data.
“I don’t expect all portfolio managers to become Python programmers but I also don’t think it will be acceptable to have a two-week turnaround to wait for new data fields in a report they are used to receiving. Technologists are going to spend more time building tools to push that data upstream to users, and give them actual access to the data. That way, someone looking at the data will know what alerts were generated, what issues were checked and cleared; all that contextualised data will go directly to the end users where they can manage it themselves, knowing it is the same high quality data as that shared with their prime brokers, fund administrators, etc,” comments Hack.
This comes down to trusting the data the PM, CRO or CCO is looking at any point in the day. As more data fields and sources are utilised, the more critical it becomes that hedge funds trust what they are looking at. This is especially important when a hedge fund seeks to widen out its investment strategy, perhaps to include a new trading instrument or market segment.
Suddenly, the portfolio manager might start trading credit or swap instruments which the operations team have never had to consider in relation to compliance processes and reporting.
Arcesium is at an advantage here, to support managers as they look to scale their business, as it has a strong technology backbone, given that it spun out of one of the world’s largest and most complex multi-strategy firms (DE Shaw). With combined client assets north of USD125 billion, Arcesium is well equipped to help hedge funds flex in one direction or another, as they evolve their investment strategies.
Gianluca Lobefalo is Head of Quant Strategies at Algebris Investments, a USD12 billion global asset management group founded in 2006. Most of its investment activities are fully systematic and as such technology plays a key role.
Lobefalo says he sees two trends unfolding. The first is that there are more and more non-systematic strategies utilising technology. This is becoming a feature not just of the quant or systematic trading space but of all investment strategies.
“Secondly, solutions that would have required huge capital investment are becoming much more accessible to managers thanks to technological advances,” says Lobefalo.
These two trends are increasing demand for technology he says.
Limits of machine learning
“Market structure has also changed,” continues Lobefalo. “Now there is a strong dominance of trade volume being concentrated in passive strategies, but only some are systematic passive strategies. This is driving a concentration of liquidity at a specific time of the day, typically at the close of market, when there is a rebalancing among passive strategy portfolios.
“Active strategies need to keep up, in terms of processing new information and speed of trading, with these passive strategies. Even traditional long-only strategies are being forced to incorporate technology advances into their normal investment strategies because of the speed required in the closing option.
“Having a technological component to investment and execution is becoming a more relevant part, no matter what strategy you follow.”
When asked how technology is being used by the quant team at Algebris, in terms of portfolio construction, Lobefalo says that the markets are so complex and continually changing that it requires an equally continual evaluation of data sources.
“It would be extremely challenging to run a systematic strategy without re-setting data and analysing new sources of data,” he says, adding that machine learning is one of the most used and least well understood terms in the investment industry.
“There is little evidence yet of the successful application of machine learning in investing. It’s an interesting field to explore and we’ve been using machine learning techniques since 2009. Humans always look for solutions to fix all our problems. Unfortunately, life is not so easy. Machine learning tools can be used successfully in some fields, such as data analysis, and to a less extent investing but for marketing purposes, to say that you use machine learning has become very fashionable.
“We take a prudent approach. Any technological innovation cannot replace common sense. Human beings will always be in the driving seat of investing. Machines can help you be more disciplined in what you do. They can help to optimise processes but they cannot replace the core of the investment process; the human mind and judgment.”
Machine learning is good at finding patterns in data and matching up disparate data sets. That can be directly applicable in data reconciliation from an in-house system to the Street, and it also can be applicable to the challenge of relating alternative data sets to tradable assets and determining whether there is a genuine trading signal to act upon.
“Machine learning is playing more of a supporting role in both cases, and I think that will be the case for some time,” says Hack.
Bill Neuman is Senior Managing Director, Product & Engineering at SS&C Eze. He argues that the biggest advances in machine learning, across all industries not just finance, have been where large amounts of data have been available.
One element that he thinks will help to take machine learning to the next level in finance, will be to have large amounts of data, which by implication, will mean greater collaboration between fund managers.
“I am suspicious of any offering that says to an asset manager, ‘We will use just your data to help optimise your execution quality’,” comments Neuman. “If it were easy or cheap to look at an aggregated view of all trade activity in an anonymous fashion, and use that as inputs to improve trading algorithms, everybody would be doing it.
“As it is, this remains an interesting data problem. The question is, to what degree would hedge fund managers be willing to participate in such data sharing exercises? Even in full knowledge they would be anonymous, to help them, but also their competitors, do a better job? That’s a big question to ponder today.”
To help support its hedge fund clients in respect to data management and analysis, Siepe, a leading provider of data management and IT solutions, provides a range of tools that allow managers to analysis alternative data sets overlaying their historical portfolio data. This is leading portfolio managers and analysts to uncover new insights and to understand what impact that alternative data might have had on portfolio positions.
“It’s as much looking back as it is looking forward,” comments Jilbert El-Zmetr, Head of MSP Operations at Siepe. “I think there’s a lot of power in this for existing managers who have a lot of their own historical position data, to fuse it with alternative data sets to see how different the portfolio would have looked if it had been positioned differently, based on insights from that alternative data set layer.
“We’re looking at partnerships with some alternative data set providers to integrate those two things together. I think it could really change the way people invest, not only from a research and idea generation perspective but also an execution perspective.
“There’s so much data that can be used for idea generation for future trading.”
Arcesium’s Hack observes that because technology providers are making hedge funds more efficient, it is enabling managers to allocate some of their middle- and back-office staff to front-office supporting roles “to give their traders more power and improved trading capabilities”.
“That’s what you want to see as an investor: the manager really focusing on the trading strategy, not on reconciliations and accounting. That’s what Arcesium can do.
“There is true evidence of technology directly improving hedge fund managers on the efficiency side; if they can save money in one area they can deploy it elsewhere and expand the investment team. They can also benefit from faster time to market with trading strategies. People are turning around launching hedge funds more quickly than ever before,” opines Hack.
Flexible technology… and agility
We live in an era where cloud-based server deployment, the creation of AI protocols and the creation of smart workflows are now realistic tools for managers to use. It’s a new era because of the computing power that now exists but the challenge people face is that we are still in the early stages of this new data era and as such the standards have not yet been set.
It’s important to therefore maintain a flexible structure; not just managers, but their service providers such as the fund administrator, who holds all of the fund data.
This is the opinion held by Mike Canni, COO, Opus Fund Services, who says: “If you build your product solely based on current capabilities, by the time you’ve finished building it the technology would have changed. It can take years to build something, by which time the chances are the technology has become obsolete and so rigid that it would take another decade to change it; it becomes a Sisyphus-like challenge.
“Any technology solution created today needs to be built on flexible foundations and modular in nature.
“Suppose next year there’s a new cloud computing solution that is ten times faster but requires some parameters for you to plug into it? If you’re not able to take a piece of your software and attach it, you won’t be able to take advantage of that new higher speed development.
“The way you collect and store data and output the data – those connections need to be as flexible as possible. It is easier to achieve this by operating under agile development methods.”
To illustrate the point, Canni refers to one USD1.5 billion hedge fund that has 2,000 investors and who was recently audited by the SEC. The SEC said, ‘these investors reside in various states across the US and we’re no longer going to allow your fund to send any material to any of them by email’.
“The manager had to switch overnight from using email and our team was able to roll out an online portal for them based on their exact requirements,” relates Canni. “When a regulatory body comes in and says you can no longer send sensitive information related to the fund over email, if you don’t have a good flexible technology backbone to make a change, you’re going to get stuck.”
Canni provides a second illustrative example of how flexible technology benefited this same manager: “Once a year, the manager would pay out an income distribution received from some of its fixed income products to investors. A request was made by some investors to increase the frequency of those payments to quarterly or monthly. We were able to shift our technology to fully automate the calculation of interest per investor and automate the creation of the bank wire, as well as automate notifications to the custodian. With a click of a button, the manager is now able to pay out interest distributions to all 2,000 investors, and report it to them via the online investor portal.
“Fund administrators using off-the-shelf software products could in no way adjust and make those changes.”
One area of technology development that Siepe is looking at, and where it is spending a lot of time with managers, is utilising Office365, which traditionally people think of as just email, word and excel document management but as El-Zmetr points out, “the tools from Microsoft have really evolved”.
“We are starting to provide clients with more functionality through some of the workflow tools and collaboration tools that are available in Office365 that allow us to integrate electronic communication with instant chat messaging, for example, where people can discuss ideas.
“Integrating such tools into their daily workflow, and ensuring they remain compliant from a retention perspective, is one aspect, in addition to giving them dashboards and tools to use for reporting and analytics; not just on the portfolio management and data warehousing side, but also in relation to research ideas, using the Microsoft platform to generate alerts when specific conditions are met.
“Funds have a process they need to follow to get approval to trade off of a restricted list. If someone wants to trade from a restricted list, they can use the Microsoft platform to generate request forms, or acceptance and rejection-type alerts for their compliance teams, without having to build bespoke systems.”
Cloud technology has unequivocally changed the way that hedge funds run their operations today. It is not an exaggeration to say that the public cloud, led by the growth of Amazon Web Services and Microsoft Azure, has been a game changer and arguably given small hedge funds the opportunity to compete on a level playing field with the biggest, most technologically sophisticated hedge funds.
A few years ago, the debate over public versus private cloud raged on, but many of the early concerns surrounding privacy and data security have largely been banished.
“The industry has completely accepted it as the way forward – we run our managed services in the public cloud using AWS and Microsoft,” affirms El-Zmetr. “We are now at a point where the biggest funds, the prime brokers, and the most successful service providers are all using the public cloud. Discussions around security and business continuity are no different in the public cloud than they are in the private cloud; that’s why the adoption of the public cloud has really accelerated as managers look for the most modern and innovative solution to grow their business.”
Managers face shrinking margins and increased performance expectations, so they want to spend as little as possible on the operational side of the business; the public cloud helps them do that and systems that use the cloud: investment management systems, PMS/OMS/EMS systems, are sitting in the cloud. This is giving managers a more reasonable, lower cost solution as they look to get their fund up and running.
“Gone are the days when everybody had to have a dedicated OMS, a dedicated accounting system, spending hundreds of thousands of dollars a year on these systems. The time it takes to launch a fund has really shrunk because technology has evolved and allowed that to happen,” adds El-Zmetr.
At SS&C Eze, while some of the components of the Eze Investment Suite were originally built as on premises apps, they are now commonly cloud deployed with positive results. Its Eze Eclipse platform, however, was natively built in the cloud from the start. The rapidly growing platform was aimed initially at emerging hedge fund managers, but is now growing to other market segments.
“It’s been interesting to see what’s happened over the last year,” comments Neuman, discussing cloud technology. “What I’m seeing is that the asset management industry, at all levels, has turned a corner relative to their perspectives on the cloud.
“A couple of years ago, some of our clients were cautiously optimistic around cloud technology, with larger managers, in particular, saying ‘We are happy to see you in investing in this area but it’s not for us’.
“In the last 12 months, however, speaking with those same clients about our product roadmap and the fact that we have Eclipse, some are now saying ‘The cloud is imperative for us today’. So I do think the market has shifted its views on the cloud.”
Security remains at the forefront of managers’ minds, says Neuman. He says that SS&C Eze processes hundreds of technical due diligence questionnaires from clients every year, with cloud security front and centre.
What asset managers need to ask themselves is, if they deployed a system within their own office building, would that be more or less secure than a cloud service provider with global clients and who spends millions of dollars on its data centres and physical and network security measures?
“We are one of relatively few firms who have earned the ISO 27001 security approval as well as the specific cloud security certification, ISO 27017, and cloud privacy, ISO 27018,” says Neuman. “Those are important seals of approval for a cloud vendor to demonstrate its commitment to security.
“We designed Eclipse with the principles of ‘zero trust’ and “least permissions” in mind. No single service must trust other services in the cloud, and both users and services operate with only the permissions required to serve their stated functions. These principles, combined with strong firewalls and employees access restrictions, help to prevent attacks from within or outside of the system.”
“We have mainly used the Private Cloud to date,” states Phillip Chapple, COO at Monterone Partners, a European equity long/short hedge fund. “This has given us a flexible but resilient infrastructure which enables us to develop the business while providing the security and controls expected.”
When asked what technology tools/systems had recently been prioritised by Monterone, based on available IT budget, Chapple confirms: “We implemented a new OMS/PMS last year, this was due to a combination of increased regulatory requirements, the need for shadow accounting and enhanced risk tools.
“In an environment where there seems a stream of new regulatory requirements, it is critical to try and future-proof as much as possible by dealing with vendors who can be pro-active and provide solutions as the new requirements occur. Otherwise the need to then layer multiple systems can create additional workflows and risks to ensure all data is reconciled and interpreted correctly.”
What lies ahead?
Asking industry professionals – both managers and technologists – what the next technology trend might be invites a wide range of opinions and responses.
From a front-office perspective, SS&C Eze’s Neuman says “the one thing we can’t overlook is regulation and the impact that could potentially have on the front office’s ability to execute”.
He says without automated systems to help keep portfolio managers out of trouble, in terms of how and when they want to shift their investment strategies as markets move, hedge funds will struggle to remain agile.
“Automation requires a strong system to help them avoid any guardrails being put in place; whether it’s beneficial ownership reporting, margining rules, transparency reporting under MiFID II, etc.
“Our compliance tools are one of the ways asset managers can solve the regulatory problem.
“We see a lot of customisation of strategies, as investors opt for separately managed accounts and this is one way our manager clients are looking to differentiate themselves. This requires good technology system capabilities to trade and monitor a well-defined investment strategy as per the investor’s tailored needs. I think in the front-office, that is becoming an important trend,” explains Neuman.
Over at Algebris Investments, Lobefalo believes the next leap forward in technology will happen when, rather than if, quantum computing becomes available.
“It will offer a completely different scale of calculation but I think we are still a few years away,” he says.
Quantum computing will allow managers to analyse many more sources of data, “and in terms of optimisation, whereas I referred to parallel computing reducing the time to a few hours, I think quantum computing would reduce the time to a few seconds. It will allow us to analyse exponentially many more data sets, and many more markets,” says Lobefalo.
On the back-office side of things, the one area that excites Opus Fund Services is how technology has the potential to change how investors subscribe to hedge funds.
Currently, Opus services approximately 700 hedge funds, all of which have their own unique subscription document with different fields that investors submit via email. Overall, it is a very inefficient process and one that affects the whole industry.
“We see a big opportunity here,” says Canni. “We are in a discovery phase at present with a number of law firms to see what can be done to standardise subscription documents, and move them into an electronic format. Why have 700 different formats, which are all asking for the same information from investors, just in different ways?
“We’ve built an online tool that allows the fund manager to type in the name of a prospective investor’s email and allows the investor to log in to the portal to fill out all the necessary information. We’ve built it with the intention to show it to law firms and ask them, ‘If we roll this out further, would you advise your clients to use it?’ The feedback has been very good so far.”
The industry is ready to move towards something like this, especially because cybersecurity is becoming a much more serious topic. All these subscription documents moving between the manager, the investor, the fund administrator, the custodian…they are being emailed everywhere and in different formats and it allows for malicious activity by sophisticated hackers to look at the patterns and act as if they are the investor and make fraudulent transaction requests.
“It’s a real risk. Greater standardisation of the subscription process, however, and moving to an electronic format, could be a way to reduce that risk and improve the security of how sensitive fund information is shared,” concludes Canni.
Mon 15/07/2019 - 11:18
Thu 11/07/2019 - 10:40
Thu 11/07/2019 - 10:24
Tue 09/07/2019 - 13:40
Fri 05/07/2019 - 10:06
Mon 15/07/2019 - 10:03
Thu 11/07/2019 - 09:51
Wed 10/07/2019 - 14:03
Wed 10/07/2019 - 10:35
Tue 09/07/2019 - 13:51