BondIT: Algorithmic Renegade in Portfolio Management

BondIT: Algorithmic Renegade in Portfolio Management

CIO VendorAdrian Gostick, Managing Director, Asia-Pacific
“It’s amazing, but still in 2016, managing fixed income portfolios using spreadsheets and manual efforts remain prevalent,” says Adrian Gostick, Managing Director, Asia-Pacific, BondIT. Gostick’s statement manifests the root cause of problems that range from missed investment opportunities to sub-optimal returns on fixed income assets. Now, compliance with increasingly stringent fiduciary requirements, and investors’ demand for personalized services, necessitates a shift from traditional practices toward deployment of an algorithmic approach to portfolio management. However, the OTC nature of the fixed income markets creates significant difficulty in developing algo-driven portfolio management solutions with zero scope for human intervention. Gostick adds, “Even developing a clean, complete, and accurate database of bonds to apply the algorithms remains a challenge when looking at 100s of thousands of bonds across multiple currencies and jurisdictions.”

The Hertzeliya, Israel based firm’s clientele includes private banks, wealth managers, and asset managers. By using data driven approach, BondIT assists its customers to construct, optimize, and maintain fixed income portfolios. “Our SaaS web-based solution harnesses the machine learning algorithms to construct a portfolio that would achieve the target yield set by an investor, whilst minimizing risk and ensuring compliance with portfolio and bond level constraints,” says Gostick. “Using the platform advisors can construct an optimal portfolio within 90 seconds, tailored to the specific requirements and constraints of the investor,” states Gostick.

Advisors can construct an optimal portfolio within 90 seconds, tailored to the specific requirements and constraints of the investor

The solution’s in-built alerting system enables the user to track and control performance of the portfolio. “Alerts can be set for a number of parameters, such as rating, spread, yield or liquidity changes,” states Gostick. The platform also suggests options for rebalancing or to replace bonds that no longer fit the constraints of a portfolio. Users can choose to allow the system to select bonds to optimize around yield or risk, or from in-house axes.

At all stages of the workflow, a full suite of risk analytics is available, helping customers understand risks of the portfolios and the impact of any changes. As well as average yield, duration, and VaR, the platform highlights where the risk is coming from, e.g. by currency, sector or rating, and where the portfolio sits on the efficient frontier.

“As well as helping clients streamline their processes our platform acts as a sales tool to generate ideas backed by analytics for advisors, and as a compliance tool to help ensure fiduciary responsibilities are met,” says Gostick. The benefits of bringing machine learning algorithms to the world of fixed income is being quickly noticed by the market, with several major financial institutions across Europe and Asia looking to roll out the platform across their businesses.

BondIT is building on its success in Europe and Asia with plans to launch in the U.S. market. “Our aim is to keep innovating for our customers,” adds Gostick. “We’re now starting to integrate into liquidity venues to ensure portfolio and trade ideas take into account the inventory available in the market, as well as to take into account customer behavior to help increase the conversion rate of trade ideas into actual transactions,” says Gostick.