The quest for futuristic solutions and models that bring positive changes in Data Management has always been to transform sourcing, distribution and managing the golden copy of data. Unlike sourcing, chaffing and maintaining data at every stage of the trade lifecycle, the data in itself is transforming into a platform with centralized data management principles. This article explains the application and benefits of using semantic technology combined with ontological structures allowing semantic requests to source business-driven data.
Challenges of Financial Data Management
Financial Data Management has become more important and complex over the years due to the following reasons:
Increased regulatory compliance: The 2008 financial crisis has led to tighter regulatory norms and more stringent compliance and reporting requirements. This meant better financial performance and financial data management.
Increased Cost: When firms are operating in a low returns model, managing data is proving to be more cost intensive and inefficient.
Timeliness and Efficiency: From the perspective of consumers, it is of utmost importance to receive data from the aggregators and deliver it internally in time to enable optimized batch processes. Given the dependencies on batch processing timelines, it has become even more important to deliver data with high quality and minimal human interventions. Refer to Figure 1 for what consumers are looking for.
There is much more focus in the market on improving data quality overall due to regulatory, risk management, and business pressures. Timeliness and accuracy of data are both of primary importance in a market where financial and reputational damage can be caused by inaccurate reporting to regulators and clients. Getting the right framework in place to measure data quality and to ensure the taxonomies are consistent across an organization is therefore very important.
-Virginie O'Shea Senior Analyst, Securities and Investments Aite Group
Figure 1: Results of survey asking respondents the criteria used to measure success of a data management implementation
Redundant data: Financial Data Management is done in independent silos by organizations, which results in higher costs, further aggravated by redundant data subscriptions and more technology platforms to cleanse and manage data.
Duplication of efforts: Same data is maintained separately by the front-office, middle-office and back-office teams, because their usage varies. This leads to added complexity and duplication of efforts.
Lack of single golden source: Firms have to source data from multiple vendors for various asset classes. There is no universal standard, leading to many data vendors and firms using proprietary standards to identify and to enrich their data.
Ever-changing technology and skills: Technology is changing at a rapid pace and so is the demand for specific skillsets. Firms pay a premium in this catch-up game and end up focusing on non-core businesses to remain competitive.
Semantics and Ontology
Semantic technologies are meaning-centered that include tools for auto-recognition of topics, concepts, information, extraction, and categorization. Semantic technology provides an abstraction layer above existing layers to enable bridging and interconnection of data and processes. At the same time, they provide level of depth that is far more intelligent, capable, relevant, and responsive interaction than with information technologies alone.
Semantic technology when combined with the Ontological representation of data enables machines as well as people to understand, share, and reason data and its file content during execution time. Figure 2 explains the ontological structures of Basic Financial Instruments concepts like stock exchange market, trading day, and analysis.
Figure 2: Ontological structures of Basic Financial Instruments concepts like stock exchange market, trading day, and analysis Source: Employing Seman
Application of Semantics and Ontology to Financial Data Management
Application of Semantics technology has been limited to academic purposes barring a couple of exceptions. With respect to financial data, the use of Semantics and Ontology will bring considerable changes to the industry:
The role of semantic integration defines a wrapper around data sources and establishes dynamic links between Ontology entities (e.g. classes, properties,, etc.) and the data sources. This is best explained with an example where an asset manager is interested in sourcing data for “Bonds with yield greater than 7% in APAC”. There are two aspects involved here:
Figure 3: Ontological representation for a subset of data for getting yield rates for a specific category of fixed income security
The Future Operating and Pricing Models
Financial Institutions are forever reshaping their operating models and processes to bring in more data credibility and reduce the total cost of data-centric operations and ownership. The application of Semantic technology can enable operating models and cost of sourcing to significantly change.
Traditionally, the data was sourced from multiple vendors/aggregators based on the quality and the prioritizing needs of the data. However, the future operating model as depicted in Figure 4 fundamentally relies on the following principles:
a) Aggregators provide data based on semantic language
b) Data credibility is maintained and managed by the data aggregators
c) Target model fundamentally shifts the golden source of data residing in consumer’s firm to the data aggregator’s firm
d) This model aggregates the data from various sources that meet the needs of the data consumers
Figure 4: Target Operating Model
Invariable pressure on the data aggregators to provide better credibility and lower costs to consumers shift the focus from providing huge volume of data to the more relevant and meaningful data that satisfies the firm's business. Given better options, Financial Institutions can pay for the data that they actually consume in their business rather than the entire offering.
Figure 5 below depicts data consumption usage against various asset classes and data characteristics.
Figure 5: Data pricing using Traditional Model
The proposed Semantics and Ontology-based model would price data based on various parameters such as type of instrument, instrument attributes, geography, complexity of sourcing data, quality of data, and time delay factor. This model helps Financial Institutions to move from a uniform pricing to a pay-as-you-go model.
Figure 6 represents the various dimensions such as asset classes, type of instrument and other factors (quality, complexity, etc.) for pricing and the semantic usage of data sourcing.
Figure 6: Data pricing using the Proposed Model
Acquiring data from multiple sources, integrating with the existing systems and processes, and ascertaining data quality has always been a challenge. Although in a nascent stage, the advent of Semantics and Ontology has been transforming with the development of cognitive technology and tools. The application of these technologies in the financial services industry is certainly a revolution that would only succeed with the changes to the underlying data organization and management.
With the increasing thrust on operations improvement and reducing cost of ownership, this technology would be adapted slowly but with great caution and care. To reap the benefits of this shift in data management, it is imperative that this initiative and implementation should come from the data aggregators. Without the initiative from the data aggregators, this technology would remain forever un-incubated. This change would eventually happen as the consumers see a huge potential in terms of outsourcing the management and maintenance of "Golden Source of Data" outside the firm’s environment to benefit from the economies of scale.
Giridhar Vugrala is a Management Consultant with the Capital Markets business at Wipro Technologies. His area of expertise spans business architecture and development of solutions in areas of data management, exchanges, front office and product control. Giridhar has been involved in executing strategic multi-million dollar signature programs of global reputation and visibility across North America, Europe, Middle East and Africa as well as Asia Pacific.
Giridhar has hands-on experience in functional, technology and delivery of systems across Fixed Income, Derivatives, Wealth Management and Trading, Quant and Investment Systems.
Giridhar is a graduate in Computer science from Osmania University and masters in Finance and Systems from JNTU. He is based out of Bangalore and can be reached at: email@example.com.
Gaurav is a Business Analyst with the Securities and Capital Markets vertical at Wipro Technologies. He has pursued his Masters of Business Administration in Finance from IIM Shillong. He is based out of Bangalore and can be reached at: firstname.lastname@example.org.
Sonali is a Business Analyst with the Securities and Capital Markets Industry Advisory Group at Wipro Technologies. She has pursued her Masters of Business Administration, specializing in Finance from Symbiosis Center for Management & HRD, Pune. She is based out of Bangalore and can be reached at: