MassMutual Shares Life Insurance Underwriting Algorithm

by on Mar 15, 2018

MassMutual subsidiary LifeScore Labs now sells its LifeScore360 underwriting algorithm, an unusual development in an industry that tends to keep these calculations private. The release anticipates a shift toward shared underwriting models and improves speed and transparency to attract new customers. The model draws from 15 years of data about the firm’s life insurance applicants and generates a mortality risk score. LifeScore Labs, a division of MassMutual, now sells the use of the algorithm and corresponding data visualization to other firms via third-party vendor iPipeline. According to data scientist Sears Merritt, the firm expects the industry to turn toward a standard risk score and sees the LifeScore360 launch as a way to lead that transition.

LifeScore360 uses data collected over MassMutual’s long history, a significant benefit in a data-based field. This may level the playing field for life insurance startups, which often lack the databases of established firms and thus may partner with them instead of underwriting independently. For this model, LifeScore Labs published a white paper describing the algorithm’s foundation, promoting it as more effective than traditional statistical models based on population-level trends. The firm used 15 years of applicant history—about 908,000 applications—to train the model. An overview of training data lists inputs based on information collected during underwriting, such as blood protein levels, indicator tests and histories of heart conditions or cancer. Along with LifeScore360, MassMutual envisions LifeScore Labs as a way to share future projects, and it is hiring a data engineer to work with the algorithm and conceptualize new models.

LifeScore360 White Paper – Graph

In addition to drawing from data, MassMutual promotes LifeScore360 as a transparent resource for applicants, adding clarity and trust to an opaque process. A sales brochure from iPipeline describes the resulting report as easily interpretable for underwriters, advisors and applicants. While the firm has yet to release specifics or a data visualization sample, a preview on the LifeScore Labs website depicts a graph of how factors like blood protein and kidney function affect the mortality risk score. There is no indication that the LifeScore product will include information about how firms collect data, potentially leaving customers ignorant of how details from third-party sources may impact their underwriting results.

LifeScore Labs Score Preview

Promotional materials note that the score will improve business through expedited underwriting. The website, brochures and press releases promise instant scoring, and the product page on the iPipeline website warns that slow underwriting will drive applicants to abandon sales and turn to competitors. The materials position speed as an asset in attracting Millennials, recommending LifeScore360 for the quick underwriting decisions expected by younger applicants.

For more innovative ways firms are working to improve underwriting and attract younger customers, check out recent blog posts about expedited underwriting, Haven Life and health-focused underwriting.