Regulation and more detailed operational due diligence questionnaires used by institutional investors are pushing hedge funds to the limit of data management. Transparency is as great as it has ever been, but with that comes the need to process enormous volumes of data. Right now, however, large numbers of hedge funds do not have the required operational infrastructure to run a ‘Big Data’ strategy.
According to the results of a survey published by Thomson Reuters on 12 August 2014, 41 per cent of respondents (including broker-dealers, asset managers and hedge funds) currently lack a big data solution. The survey, ‘Big Data in Capital Markets: At the Start of the Journey’, was produced in conjunction with Aite Group and shows that capital market participants have been slow to adopt big data strategies.
Outside of the Financial Services industry there are examples of how big data has helped firms improve their operational efficiency, marketing strategies, business decision making and better understand their customers. However, according to research performed by SEI, a leading global investment operations and fund administrator, in conjunction with global financial institutions consultancy, Alpha FMC, the asset management industry has been slower to adopt so-called “big data” initiatives.
Asset Managers understand that data is important and critical to running their business, but they are less clear as to where to start or what they should focus on from a budget, governance and initiative perspective with respect to managing their data more effectively.
Separating the wheat from the chaff is a major theme for many fund firms in terms of their use of data. Fund providers sit on a treasure trove of data, but sorting through it can be extremely challenging if not daunting – this is the Big Data challenge. The growing number of data sources, coupled with advances in technology, allows firms to capture multiple pieces of information about clients and prospects. Making the best use of data is both challenging and a priority for fund firms.
Those that are using big data are doing so on a piecemeal basis and applying it to specific areas of the business; in particular trading and quantitative research. For hedge funds, utilising big data within the front office makes sense as it is the engine that drives alpha generation and returns for investors. But going forward, hedge funds should look to implement a big data strategy firm-wide so as to not only support trading but assist in sales & marketing initiatives, improve compliance and drive down operational costs.
Indeed, the Thomson Reuters survey found that half of respondents either currently employ or plan to hire a data scientist within the next 24 months.
Yet big data solutions will not be effective in organisations where there is not an effective data foundation strategy alongside robust and thoughtful governance.
At SEI, the firm has been developing effective data aggregation capabilities to support its hedge fund clients as they face increased complexity managing data. In an informal poll, they found that 83 per cent of participants agreed or strongly agreed that they would like to make more use of big data to help them make business decisions.
Moreover, 94 per cent of managers interviewed said that in the next one to three years they would be improving their data management infrastructure.
There is, then, a growing realisation that big data is becoming an integral part of a hedge fund’s operational strategy. However, getting all the pieces of the jigsaw in place to handle large volumes of data across the firm – front through back – is a tall order and requires a huge investment in time, resources and technology spend. This is frankly not an option for many hedge fund managers who are looking to limit the size of their operations teams in a bid to protect the bottom line.
This is why service providers like SEI are seizing on the opportunity to develop automated data management solutions – by aggregating internal, manager and third party data and delivering it through web-based applications like its Manager Dashboard, Investor Dashboard and global regulatory reporting solution all of which are components of the overall SEI operational platform.
Joe Henkel, part of the Global Solutions team at SEI’s Investment Manager Services division in Dublin explains: “We had a meeting in London recently with a large institution who uses a number of different administrators. Because they trade all types of asset classes and run multiple fund structure products including UCITS, hedge funds, and managed accounts, they are struggling to find one system that could provide them with OMS and PMS capabilities.
“We showed them our Manager Dashboard and whilst we typically don’t sell it in that capacity as an OMS or PMS, this tool can provide the types of reporting that managers typically gain from those types of systems.
“Our data model and processing engines are agnostic as far as product structure, domicile, or asset classes utilised, so for this client it was a unique tool that could aggregate and normalise data from the investment manager, the prime brokers and the custodian, as well as external market and industry sources, and then publish it in a usable, relevant and insightful manner.”
Risk and portfolio attribution are due to be integrated into the dashboard to bring added functionality, according to Henkel. What this is doing is to help push clients more towards the adoption of outsourced operations to handle the data management challenge. As Henkel points out:
“We have a number of large institutional clients who had multiple in-house systems and not only did they decide to outsource administration and investment processing to us but also the hosting of their middle office systems. We think people who view big data as being just about aggregating, housing and securing data are missing a critical point – it is also about governing the data.
“SEI brings a significant amount of experience in this area which can allow managers to jump ahead of the curve, harness relevant data and analyse it to find answers that enable them to optimise their offerings, become more efficient, make smarter business decisions and focus on their core competencies.”
The Data Management Highway
One of the critical elements to data warehousing is that a firm like SEI, with its extensive team of industry experts, not only ensures that the aggregated data is cleansed and normalised, but they also ensure that trade breaks and exceptions are identified and fixed. It is essentially a quality control exercise, based upon exception management and streamlined workflow, that requires deep knowledge and skill.
What this does is create a sort of golden repository that the manager can access and mine for all manner of operational requirements: portfolio management, risk analytics, shadow accounting, regulatory reporting, middle office reporting and so on.
Ross Ellis is Vice President and Managing Director of the Knowledge Partnership in the Investment Manager Services division at SEI. In his view, a so-called golden record may be a great vision to aim for but reality dictates that there may need to be more than one golden repository as the data being accessed by the front office is more time sensitive than the middle and back office.
“The front office wants data in real time 24/7 to make trading decisions whereas middle and back office teams have to do reconciliations, check on corporate actions etc. The timing for getting that right is not as fast or time sensitive as it is for the front office.
“The accounting book may not be the same as the trading book so do you have two golden repositories? From an academic standpoint a golden repository makes great sense but in practical terms maybe that’s not what the manager needs. Maybe they need different repositories depending on the strategies and instrument types the front office trades,” says Ellis.
This would seem logical. By having accurate portfolio and risk data in real-time, traders can react much quicker and more confidently to market events. They operate in the outside lane as it were, where speed is essential. Often, the needs of different groups are in reality different views of the same sets of data or different point in time views of that data.
The transparency within a hedge fund, therefore, differs according to the activity.
“In terms of a manager’s operational infrastructure, it has to be such that it isn’t a one size fits all. It has to be built in a flexible, customisable fashion. The data warehouse is where the data is cleansed and normalised but what you decide you’re going to send, to whom and in what state of cleanliness, will vary,” says Ellis, who continues:
“Regulators want data reported to them at a single point in time, e.g. at the end of the quarter. It’s not like they need data throughout the day. Nor in fact do most investors want data intraday. They are happy to receive it when it’s been reconciled at the end of the day or period. When managers or service providers build their infrastructures, they have to think ahead. You need to anticipate how fast interested parties are going to require data and in what form. What types of data might be useful that you don’t currently house?”
To continue the highway analogy, it’s not just thinking about what lane you are in. The weather must also be taken into account – what we mean here is unstructured data that comes from market rumours, social media, events and so on. Market noise basically.
How does one incorporate unstructured data into a data management model? After all, market news is vital when it comes to making investment decisions. Until now, however, most data models have not supported unstructured data because there was no clear way to categorise it.
Unstructured data might, for example, include macroeconomic data. Think back to when Long Term Capital Management imploded after Russia defaulted on its bond payments in 1998. The fund wasn’t equipped to handle such an unlikely stochastic event. Today’s hedge fund managers, whatever strategy they are running, are increasingly relying on their service providers to handle unstructured data and improve their ability to model stochastic elements that directly impact the price of the portfolio.
SEI has begun to design methods to marry this macroeconomic data, from a big data perspective, to investment and investor data processed by SEI for their clients. Henkel explains:
“We don’t make any precursor decisions based on who the manager is; e.g. they’re a long/short equity manager so they don’t need bond market data. Well actually, indirectly they do because if bond prices are rising, equities will typically be falling. We keep the data warehouse architecture open.
“When we talk to clients, part of our implementation strategy is understanding what it is they are interested in and then providing them with as much data as possible, in an organised fashion.”
The Three Vs
Henkel uses a key word in the above quote: organised. It’s all well and good having a big data strategy in place but the trick is to make sense of it. To mine specific chunks of data that can reveal market trends and patterns perhaps within specific sectors or asset classes.
As Ellis says: “Different data sets can lead to different reactions and from a manager’s perspective, it could give them insight into what securities to buy and sell, insights on investor interest and so on. It’s not only getting data but figuring it out and taking the appropriate action within the necessary timeframe.”
Without the necessary infrastructure and tools in place to handle big data, managers are confronted with so much information that they find it hard to separate what’s relevant from what is not relevant. They become paralysed into making no decisions at all.
“The three Vs apply when talking about big data,” adds Ellis. “These are volume, velocity and variety. Data is faster, there’s more of it, and it’s coming from everywhere. Managers have to address these three factors to make sense of it.”
Data is valuable but if you don’t do anything with it, it’s wasted. It’s not just figuring out smart ways of getting data, therefore, it’s about how to monetise that data; it could help with product development, following investor leads for capital raising, making smarter portfolio decisions. Incremental gains (or efficiencies) have the potential to put managers in a stronger position to run sustainable, successful businesses.
Don’t forget the investors
It’s easy for the front office to be sidetracked into only focussing on big data for the benefit of the trading book. However, data aggregation is important not just for the investment side of the equation but also that of the investor.
“In some cases, this is as important, if not more important. Investors are looking for managers today that can provide something different to the firm down the street. They clearly look at the AuM and the performance track record but increasingly institutional investors are also looking to see how they answer due diligence questions, how do they report information to investors, what level of transparency do they provide?” says Henkel.
SEI’s Investor Dashboard and Manager Dashboard put more control in the hands of managers in running their business as well as for investors in gaining comfort about their investments. Investors can log in and interrogate portfolio data taking the day-to-day pressure off managers’ IR personnel. The idea for doing this goes back 15 years, according to Ellis. Back then, managers wanted a purely internal dashboard until one particular client said, ‘We really like the functionality. How about you take all that data and create an investor-driven dashboard?’
“The manager wanted to respond to investors one day and three days after month-end with preliminary results and to send investors data that they could slice and dice to see what impact that investment had on the rest of their portfolio.
“To do that they needed access to aggregated data that they could interrogate the way they wanted. That was the start for us to make sure that significant cleansed data was available to investors as well as managers,” says Ellis, who confirms that investors are increasingly looking for exceptions processing.
An investor may see a large drawdown in the fund and want to know why – is it because the manager has lost their edge? Is it down to the trader making one bad decision? Everyone makes mistakes but as long as investors understand why it happened by digging deeper, that extra information can help them when talking to the manager.
This is an important advantage of a big data strategy. Combined with its application for trading in the front office and populating myriad regulatory reports as part of ongoing compliance, managers with the right infrastructure in place can achieve a high degree of operational efficiency.
As Ellis concludes: “When talking about an operational platform, it’s not just about technology. Anybody can buy the best technology. What makes it a platform rather than just a collection of systems is how you integrate the various systems together so they work as a team and not disjointly. It’s the expertise that you have within your organisation that understands what these systems can do, how they can talk to each other.
“A good platform is based on the strategies and securities traded not having the arbitrary product structure dictating rules. Central to this is one data model that can power the entire platform.”