While existing underwriting tools that use credit file data are effective when applied to the standard insurance market, these tools are not as effective when applied to a large segment of the nonstandard auto insurance demographic. Specifically, reliance on data from: the major credit bureaus and public files such as LexisNexis produce incomplete, inaccurate and often no results. This can translate into higher premiums, (which can make insurers less competitive in the market), weaker underwriting, and greater exposure to fraud and identity verification challenges. Lack of visibility into this often transient or discreet customer group becomes even more challenging and potentially costly when we take into context the growth of this population of consumers.
There is a high correlation between Non Standard Insurance customers and the underbanked. A recent report by Stone Ridge Advisors states:
...The [nonstandard] sector has evolved to include drivers who purchase insurance policies with the state mandated minimum limits, typically lower income drivers or recent immigrants to the United States. The customer base is also characterized as one that typically pays for an auto insurance policy on a monthly basis, makes purchasing decisions based primarily on the cost of the initial down payment for the policy, and has a high cancellation or non- renewal rate. Higher rates of insurance fraud and staged accidents are also more prevalent among the non-standard auto customer base than among the standard or preferred auto insurance customer base.
This last evolution of the nonstandard definition overlaps well with those within the underbanked segment who have financial challenges that disproportionately weight their insurance purchase decisions. When underwriting policies, many insurance organizations access credit reports' insurance scores, or use their own scoring tools which may included credit information as a variable within their models. Often the underbanked produce little to no information (thin file or no hit) when their credit report is accessed through the three major bureaus.
So what's so bad about the credit bureau data and public file that we use now? When data on the underbanked is returned from credit bureaus, typically in the header field, it is usually stale and subsequently not as relevant. The data can be as much as seven years old. Using trade line data to illustrate this point can be elusive given that the underbanked, as mentioned earlier, often produce thin file or no hit returns. Time at a given address might be better means to demonstrate how applicable traditional bureau data is. According to FactorTrust's Underbanked Index (May 2013) the underbanked average two years at the same address. Given the frequency of changes among this demographic, traditional bureau data which can exceed two years, may no not be as effective at reducing fraud and executing other underwriting practices.
Existing bureau data also fails to use non-traditional data assets to develop a more accurate profile of the underbanked customer. Non-traditional information such as IP addresses and payroll data, can also support fraud prevention, improve employment, and Identity verification along with other stability measures. Payroll data, for example may be good to consider when attempting to improve premium stability. IP address can contribute to fraud preventions efforts.
Finally, public data that is available via products like LexisNexis often does not provide visibility into recent immigrants as they may overlap with the nonstandard auto insurance customer. A recent FDIC Study (2011 National Survey of Unbanked and Underbanked Households) claims that Hispanics — one of the fastest growing consumer segments — make up 28.6% of the underbanked population. Their participation in the nonstandard market makes having the most accurate up to date data critical to success. As the Stone Ridge Advisors report says: "Given the projected growth in the Hispanic population in the U.S. and their growing share as part of the non-standard auto sector, we would expect continued growth in the non-standard auto market."
Growth in the underbanked market will require better data tools to predict behavior, identify consumers and reduce fraud. Here are three ideas:
1. Supplement credit bureau data with non traditional data
Many insurance organizations have integrated third-party data, such as claims records, into their predictive models. Adding non-traditional data is a continuation of this logic. Most nonstandard insurance providers are well versed in handling, interpreting and capitalizing on data, but even the most sophisticated ones are starting to identify and integrate additional third party data in order to improve their models. Actively identify and partner with the kinds of organizations that can furnish your organization with meaningful data assets.
2. Work with a bureau that specializes in the underbanked
Leverage third-party models scoring and data ingredients as you execute your process. Working directly with an FCRA Compliant Credit Bureau with data on the underbanked can improve fraud reduction, identity verification, stability and marketing in a meaningful way. FactorTrust not only provides regular updates on the underbanked, it also develops predictive models and scoring and and raw data elements that insurers can use as ingredients in their home grown modeling and predictive tools.
3. Obtain Real Time Data
Individuals within the underbanked segment tend to operate in time frames of weeks rather than months given their cash constraints. Having current data is important for making underwriting decisions. As you overhaul your systems, work towards real-time updates. Obtaining real-time data should be an objective both for internal systems as well as partner and vendor technology. Having a current update on the underbanked consumer that your servicing, will support your efforts to minimize risk, fraud, and cost. It can also mean the difference between being able to price a product and not having a consumer that meets your underwriting requirements.
About the Author: Grant Brown is senior consultant partner and chief executive officer of Raymore, Mo.-based Profluent ePayment Consulting.