While front and back offices in commodity trading companies have seen significant impact from digitization, risk functions have lagged behind. That may be about to change…
Digital technologies are transforming the commodity trading value chain. At the front-end, machine learning and algorithmic trading have had a significant impact on pre-deal analytics and revenue generation, while in the back office, process automation has resulted in efficiency gains. This contrasts with functions within the middle office such as risk, product control, and compliance, where digital adoption has been much slower.
There are multiple reasons for the slower progress. Unlike the front office where there exists the carrot of increasing revenues, risk and middle office functions have tended to see the benefits of digitization more in terms of efficiency gain. Given the unpalatable potential consequences of getting it wrong, risk and middle office functions have been less willing to act. Additionally, increasing and constantly changing regulatory compliance reporting pressures have seen risk officers running to keep up. Consequently, there has been little appetite for implementing a digitization agenda.
The situation is changing, however, with chief risk officers recognizing that there are many benefits to be realized in their market, credit and operational risk functions, through a judicious and targeted application of digitization. Many benefits will come from increased automation and resulting efficiency gains, but there will be true “value-add” benefits too.
In my view, there are a number of clear areas in which the digital agenda will have an impact.
Credit risk management continues to be reliant on largely manual processes and individual judgment. The ability to integrate data sources, both internal and external, and overlay the resulting large amount of disparate data with machine learning and advanced analytics, will allow the credit manager to extract better information and make better decisions. The use of analytical capabilities for pattern recognition of company failures, for example, will aid credit risk managers in making better informed and faster decisions around credit approval and credit limit setting.
Increased automation of credit risk management processes will also drive gains in analytical capability: exposure-explained to decompose daily credit exposures, CVaR-explained to better analyze complex investigations, and analysis of expected loss (EL) figures to determine changes in probabilities of default.
Operational efficiencies will also be realized. Automation of the largely manual and cumbersome know your customer (KYC) process, for example, will drive standardization and decrease the time taken to execute the end-to-end process.
Market risk in commodity trading companies is inevitably reliant on a variety of data sources and systems, particularly where multiple commodities are traded, with numerous price data sources and a variety of transaction entry platforms. Automation will have a clear impact on the often manual manipulations required to bring disparate data into Monte Carlo and VaR models, for example, reducing operational errors inherent in the process. Further, automation will be used in regular and manually intensive processes such as the end-of-day process, again driving efficiency and reducing errors.
Advanced analytics and cognitive computing will be used to analyze complex scenarios, perform path-dependent risk analysis, and predict risk events, giving the market risk manager sophisticated tools to supplement existing history-based probabilistic risk calculations.
More immediately, analytics can be used to track the proximity of exposures to limits and to assess if positions are moving faster than market liquidity, allowing market risk managers to better assess the true risk in a trading portfolio and act appropriately.
Operational risk and compliance
In the compliance arena, there has already been some progress in the use of digitization. The detection of wash trades and the application of analytical capabilities to detect abnormal behavior and potential market manipulation are two examples. With regulatory regimes such as MiFID II and MAR requiring not only detection of fraudulent trade execution, but additionally the monitoring of market abuse intent, companies must now monitor and analyze an enormously expanded amount of data. Critically, much of this additional data and information is unstructured, making the use of advanced analytics and machine learning an imperative.
Advanced analytics may also be used to sew together information held in silos in different departments of a trading business, spotting informational differences. In numerous past cases of fraud, different departments have held information which, if viewed and analyzed holistically, could have provided the alert for fraudulent activity. Unfortunately, these information anomalies have only become apparent during a post-fraud event audit. Advanced analytics will be able to take data from different departments, map the information and spot potential anomalies, thus providing an alert to further investigate a trade or position.
Although few commodities trading companies are undertaking large-scale risk digitization transformation programs, the benefits are clear, coming from both efficiency gains leading to reduced operating costs, and amplified effectiveness of the risk function itself.
The best path forward for companies may be to test the waters with small-scale and targeted digitization to demonstrate value, before embarking on a broader initiative.
Cyrus is a Partner with Infosys Consulting and manages our energy trading and risk management portfolio with top clients. He is an energy trading and risk management professional with over 25 years of experience in a variety of management consulting and trading / risk positions with leading global organizations.