Big Data in Action – Let’s Get Started
One of my visionary colleagues at EMC, a leader in digital marketing always ends his presentations with the bold sentence “Let’s Get Started”. So when I was asked to develop the storyboard to demonstrate the specific benefits and applications of Big Data Analytics for clients in capital markets, hedge funds and asset and wealth management, I immediately borrowed this line. We had already spent enough time with theoretical discussions, idea generation and brainstorming workshops. It was time to put the pedal to the metal and demonstrate real results.
We are starting to come out of one the worst credit crises in recent history. Financial firms have learned the lessons of the limitations of incomplete data and piece-meal models for managing portfolios and decision making and become more sensitive to their real time needs. Information driven companies want to harness Big Data Analytics to play pivotal roles to optimize their use of capital and manage their risks. During this year’s Gartner Business Intelligence & Analytics Summit, the key message from Nate Silver, the analytics expert and New York Times columnist was about the need for a practical strategy and specific application for Big Data Analytics. His keynote speech was applauded by a packed audience of more than 1,400 business and IT professionals when he said “Whatever you do, don’t give into the big data hype.”
All complex problems and programs involve a learning curve and tackling Big Data is no exception. Some financial firms are re-thinking existing models and risk management analytics fuelled by readily available, open source Big Data technologies. But they should be cautious of the hidden challenges they pose. Most of these “do-it-ourselves” Big Data Analytics sandboxes (like those on Amazon Cloud) seem to take a “trial and error” approach. To help clients move beyond random explorations and more clearly move into the business realities with Big Data, we started with one solid Big Data Analytics business use case that has common implication and that is a crucial part of the investment process at both the buy side and the sell-side firms.
Leverage Big Data to Analyze Portfolio Exposure
The first question is: what’s the exposure? Exposure is the amount of funds that are invested in a particular type of security and/or market sector or industry and usually expressed as a percentage of total portfolio holdings. Thus, exposure is the amount of funds an investor has at risk that he/she can lose. Investopedia explains ‘Market Exposure‘ as the exposure of a portfolio to particular securities. Markets and sectors must be considered when determining asset allocation. It’s the answers to the critical questions that can help increase returns on capital and, if done properly, also minimizes losses. For example:
- What is your portfolio’s total risk exposure by sector?
- As you read about the financial news in Europe, do you have at your fingertips the information you need to understand its effects on your portfolio and adjust your decisions accordingly?
- What are your long and short portfolio exposures to the Euro currency? How has this changed over the past 2-3 year?
Those questions are as relevant for Portfolio Managers and Traders, Wealth Managers, Financial Advisors as they are for Risk and Compliance Officers, Finance Officers and Internal Auditors who must closely monitor the firm’s investments and adherence to regulatory requirements. But portfolio exposure analysis today is like juggling so many balls – all of different shapes, sizes and weights.
The Challenges – what are the painful points today for Exposure Analytics?
- Difficulty in collecting all available/relevant data from different markets, analyst commentaries, news and events and in real-time
- Cannot predict markets or forecast the future
- Lack the ability to create interactive scenarios modeling for possible events
- Hard to sift through the large numbers of real-time transactions to check for risky positions
- Cannot quickly access large amounts of historical datasets for rapid and complex trade modeling
- Multi-stepped and siloed analysis and reporting performed by different groups using different applications
- Lack of ad-hoc query capabilities and reporting on aggregate exposures across multi-asset portfolios and client accounts
The Opportunities – what are the benefits of Big Data Analytics?
Many cost effective, highly scalable, high performance and low latency Big Data Analytics tools became available in the last few years to assist in the collecting and loading of data from all data sources; from existing data warehouses to internal and/or external feeds as well as 3rd party data files. With the next generation analytics platforms investment management practitioners don’t have to struggle for hours or days to create rich and realistic scenarios to analyze the impact of a certain market, security, or sector exposure on their investments as an event unfolds. They can quickly turn to a single place for instant, accurate information about their portfolio and track multiple dimensions of exposure data for the best course of action. The new exposure analytics solution allows users to:
- Integrate market data, news and events (earnings announcements, quarterly GDP, employment rate, interest rate, inflation, etc.) into exposure analysis
- Analyze multiple levels of exposure (transaction, position, account, counterparty and firm) to various securities, asset classes, sector, as well as market, on-the-fly or near real-time to develop more effective risk mitigating and trading strategies
- Forecast possible event in the future through different scenarios modeled from the news reports and media sentiment
While tracking exposures on individual funds is essential, it’s even more important for portfolio managers to know where the overall portfolio stands at all times and if they’re about to cross a decision threshold. By getting a comprehensive view of geographic, sector, strategy, market cap or stock exposures, investment managers can better decide whether to reallocate or rebalance their portfolios.
Portfolio managers can also compare exposure information with portfolio weights, liquidity, and attribution over time. This way they can spot the important trends from the large volume of historical data points through the dashboards, which provide graphically intuitive data visualization so the exceptions stand out with alerts about exposure anomalies.
Even though this use case requires painstakingly analyzing and interrogating larger data volumes from many external and internal sources to develop accurate, predictive models, every firm has to go through it to make Big Data analytics capabilities more tangible to its users. In my next blog, I will illustrate the key tasks around which different functions within the organization must collaborate to turn the Big Data Analytics into a reality.