Clearwater Analytics to Roll Out New Performance, Risk Modules for Flagship Platform

Clearwater Analytics is building out its data management solution to include two new modules for evaluating portfolio performance and risk that will incorporate machine-learning algorithms to produce predictive outputs.

Its flagship platform, which uses a SaaS delivery model, pulls in fundamental and alternative data from investment managers, fund administrators, custodians, brokers, and data vendors. The datasets are then validated, run through accounting processes, and developed into analytical reports.

The performance module will offer predictive analytics and weighted scores on how a security, fund, or sector is expected to perform. The module uses machine learning to analyze and measure a variety of investment return types: time-weighted, money-weighted, book returns, price returns, and income returns. Users will have the ability to conduct historical analysis and drill down into specific data points in the lifecycle of a portfolio.

“When you put together time-series data points—years and years of them in different types of securities—and external data sources, you can really have some pretty powerful knowledge that you can give to your customers to say, ‘This is what’s going to happen tomorrow,’” says Warren Barkley, Clearwater’s CTO. Barkley joined the company earlier this year.

The performance module also enables users to compare portfolios and funds. As part of that capability, middle- and back-office teams can review and compare performance benchmarks and risk-adjusted returns over several years.

Clearwater’s risk management module is also being developed to provide on-demand analytics of a portfolio or fund’s exposure to risk. The technology uses machine learning to map datasets, and automatically build risk profiles of individual securities, companies, and sectors. Users of the risk module can evaluate historical trends, cash flows, different exposure types, and track credit events. The real-time reports will also provide analysis on benchmark comparisons and value at risk (VaR).

“We have a lot of customers in insurance and [other sectors] who want something beyond the traditional statistical mathematics that gives them risk numbers; they want a much more complicated understanding of where their risk lies in a portfolio,” Barkley says.

Both the performance and risk modules are scheduled to be rolled out in the first quarter of 2021.

From an operational perspective, Clearwater Analytics is exploring new ways to speed up the delivery of some of its analytics. Barkley says that the vendor is looking at how machine learning can be used to predict the time of when information—from sources such as custodians—will be sent to its data management solution, which would allow the system to be better prepped from a processing and capacity standpoint to run performance and risk measurements, more quickly. 

The vendor is also using ML to detect anomalies in large datasets, or inaccurate datasets provided by third parties, before that data is funneled downstream for reconciliation, accounting, and reporting.

“As part of the anomaly detection [capability], you also get a quality score, which tells the client that the data from these sources—these types of securities, or this type of data from this bank—is very high quality, and you can both pass that on to the customer … and also use it internally from an operations perspective, because we know that we can trust this data more than we can other data,” Barkley says.