Patent Issued for Insurance risk scoring based on credit utilization ratio (USPTO 11158005): Cerner Innovation Inc.

2021 NOV 17 (NewsRx) -- By a News Reporter-Staff News Editor at Insurance Daily News -- From Alexandria, Virginia, NewsRx journalists report that a patent by the inventors McNair, Douglas S. (Seattle, WA, US), filed on December 18, 2019, was published online on October 26, 2021.

The patent’s assignee for patent number 11158005 is Cerner Innovation Inc. (Kansas City, Kansas, United States).

News editors obtained the following quote from the background information supplied by the inventors: “Insurance Risk Scores are typically based exclusively on objective, factual credit report information, including consumer accounts such as credit cards, retail store cards, mortgages, and auto loans. Also included in typical insurance risk scores is public record information, including bankruptcies, liens and judgments, and collection accounts. Additionally, Insurance Risk Scores take into consideration consumer-initiated “hard pull” inquiries associated with their requests for new or increased lines of credit. Multiple consumer-generated “hard pull” credit inquiries associated with the shopping for a mortgage or auto loan are de-duplicated on a time horizon of 14 days, to minimize the impact on their score. All of this factual credit information is received by credit rating agencies such as Equifax, TransUnion, Experian, and FICO from tens of thousands of financial institutions, retailers, and court houses on a monthly basis. Typically consumer credit reports and the calculation of a consumer’s Insurance Risk Score do not include medical history and records, consumer buying habits, checking and savings information, income, and prohibited basis characteristics identified by the Comptroller of the Currency, which includes information regarding marital status, race, age, religion, family status, color, receipt of public assistance, disability, gender, or national origin. To date, no basis characteristic related to patterns of credit utilization ratio are known to have been prohibited.

“In their underwriting and pricing process, insurers seek to charge rates that are equitable, adequate and not unfairly discriminatory. These objectives are sometimes difficult to achieve because of regulatory constraints and insurers’ own desires not to discriminate unfairly or act in a manner that is inconsistent with socially acceptable standards. From the company perspective, pricing equity and accurate cost projections are crucial. Credit data can be used to create scores that in fact provide additional predictive information about future losses. However, using credit history is often perceived to be in conflict with what society considers as fair, particularly if the individual’s score is affected by catastrophic events such as divorce, medical problems or loss of a job.

“More than 90 percent of insurers responding (from the top 100 personal lines companies) indicated in an American Academy of Actuaries survey that they currently use credit report data in their auto and property and casualty (non-health) insurance underwriting operations. Ten percent use it for pricing only; 38 percent for underwriting only, and 52 percent for underwriting and pricing. Fourteen percent use credit history on annual renewal; 33 percent use such data during re-underwriting, and 38 percent claim not to use credit report data at all in the renewal process. However, medical insurance underwriting does not typically include credit report data.

“The use of credit data in underwriting and pricing of personal automobile insurance has sparked an intense debate that centers mostly on the following factors relating to statistical correlation between credit data and loss ratio: (1) benefits to consumers, (2) discrimination, and (3) socially acceptable criteria. There are several published studies that show a statistically significant relationship between credit data and loss ratio performance. These studies show that this correlation can change in time-but this correlation, however strong, cannot establish a causal relationship. The use of credit data has allowed insurers to establish that some insured individuals, traditionally classified as “standard,” can qualify as “preferred” when evaluated by these models. Studies have shown that even insured individuals with prior violations or accidents but having good credit behavior can have better loss ratio performance than insured individuals who have no accidents or violations but who have poor credit.

“In the U.S., most state insurance laws prohibit the use of insurance rates that are excessive, inadequate, or unfairly discriminatory. Principle 4 of the Casualty Actuarial Society’s Statement of Principles Regarding Property and Casualty Insurance Ratemaking states that, “A rate is reasonable and not excessive, inadequate, or unfairly discriminatory if it is an actuarially sound estimate of the expected value of all future costs associated with an individual risk transfer.” Thus, the overall average rate level should be set so that the total premium collected from all risks is sufficient to cover the total expected costs. Additionally, the individuals’ rates should be set such that the premium collected from each individual risk, or group of similar risks, accurately reflects the expected costs for that individual risk (or group of similar risks).

“The use of credit data in decision-making, along with having more easily accessible and reliable data, has led to the rapid growth in automated underwriting systems that minimize subjective judgment by relying on more objective, rigorous, data-driven decision processes. Automated systems are more predictive, reliable and can improve the integrity of risk classification systems. The federal Fair Credit Reporting Act (FCRA), 15 U.S.C. 1681 et seq. regulates the consumer reporting industry in the U.S., including firms that furnish data to and use data from consumer reporting agencies. Comprehensive changes to the FCRA were enacted in 2003 when Congress passed the Fair and Accurate Credit Transactions Act of 2003 (FACTA or FACT Act). The FCRA controls the intake and output of consumer reporting data. Many states also have their own credit reporting laws.”

As a supplement to the background information on this patent, NewsRx correspondents also obtained the inventors’ summary information for this patent: “This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of the claimed subject matter.

“Systems, methods, and computer-readable media are provided for the field of insurance underwriting, pricing, and loss ratio estimation. Time series are formed by (a) retrieving an insured individual’s credit utilization ratio (CUR) periodically via ‘soft pull’ inquiries submitted to credit rating agencies, (b) calculating Bayesian power spectra for each time series formed from a plurality of such time-stamped CUR values, © repeatedly randomly sampling the spectra to calculate the median likelihood for each, with Bonferroni or other suitable correction for timeseries length, (d) scaling the median likelihood values so as to be on a scale that is commensurate with the weights calculated by conventional insurance risk scoring, (e) combining each scaled median likelihood with the corresponding conventional actuarial models and basis characteristics, and (g) optionally, rank-ordering the resulting set according to the scores to predict which historical loss patterns most closely resembles the current spectral characteristics of the insured. Insolvency (liquidity, leverage, default risk) represents an instantaneous hazard; as soon as liquidity is restored, default risk abates. But insurance risk effects of financial distress, like health effects are likely to accrue over time, much as occurs with exposure to tobacco or alcohol. Cessation of the exposure does not restore risk to baseline. The system and method of the present technology allow a system to assess the effect of frequent or unexpected changes in an insured individual’s liquidity on physical or psychological stress that may contribute to the insured individual’s health issues, health services utilization, and insurance claims. This is likely to be an effective means of mitigating inaccuracies in estimating the loss ratio.

“An embodiment determines a measure of financial stress, and uses this measure in conjunction with actuarial methods. An embodiment performs credit rating agency “soft pull” inquiries, which may be submitted bi-weekly or monthly, for each insured plan member or policy holder. The impact of frequent or unexpected changes in consumer liquidity on health utilization claims is captured and measured. These frequent or unexpected changes are likely related to stresses experienced by the insured. A credit utilization ratio (CUR), which may be determined as a time-series of outstanding balance of debt as a percentage of credit line available, is used to calculate a Bayesian power spectrum. The CUR enhances estimation accuracy of insurance loss ratio, claims frequency, and probability of excess claims. Further, it augments insurance policy performance characteristics for an individual or for groups of insured individuals.”

The claims supplied by the inventors are:

“1. A non-transitory computer-readable storage media having computer-executable instructions embodied thereon that, when executed by one or more processors, facilitate a method comprising: receiving, via the one or more processors, soft pulled applicant credit utilization data; determining, via the one or more processors, a soft pulled time series for the applicant credit utilization data; generating, via the one or more processors, a structured time series by scaling the soft pulled time series; determining, via the one or more processors, a frequency domain power spectrum based on the structured time series; determining, via the one or more processors, values for a spectrum likelihood measure based at least in part on the frequency domain power spectrum; scaling the values of the spectrum likelihood measure to produce a power spectrum weight (PS_wt) value for an applicant; generating, via the one or more processors, a composite risk score for the applicant by combining the PS_wt values with a plurality of actuarial model basis characteristics; comparing, via the one or more processors, the likelihood measure to a probability threshold; determining, via the one or more processors, a resultant insurance risk category from the result of said comparison; and storing, in an operational data store, the resultant insurance risk category.

“2. The media of claim 1, the method further comprising: providing an incentive to the applicant based at least in part on the resultant insurance risk category wherein the structured time series is scaled based on a standard interval range.

“3. The media of claim 1, the structured time series comprises a predicted future time series.

“4. The media of claim 1, wherein determining values for the spectrum likelihood measure comprises using at least one of a wavelet transform, a discrete cosine transform, a discrete fourier transform, a periodogram method, a Bartlett method, a Welsh method, or an autoregressive moving average estimate.

“5. The media of claim 1, wherein determining values for the spectrum likelihood measure comprises identifying a first set of frequency terms that have a higher frequency than a second set of frequency terms.

“6. The media of claim 5, wherein the first set of frequency terms are discarded.

“7. The media of claim 1, wherein determining values for the spectrum likelihood measure comprises generating repeated random permutations.

“8. The media of claim 7, wherein determining values for the spectrum likelihood measure comprises sampling the resulting permutations utilizing a Bayesian Markov Chain Monte Carlo simulation.

“9. The media of claim 1, the method further comprising determining an entropy of the soft pulled time series or the structured time series.

“10. The media of claim 1, the method further comprising rank-ordering the values of the spectrum likelihood measure.

“11. The media of claim 1, the method further comprising determining a distance between the values of the spectrum likelihood measure and one or more reference spectra; and selecting one or more reference spectra according a classification criteria.

“12. A computer-implemented method comprising: receiving, via one or more processors, soft pulled applicant credit utilization data; determining, via the one or more processors, a soft pulled time series for the applicant credit utilization data; generating, via the one or more processors, a structured time series by scaling the soft pulled time series; determining, via the one or more processors, a frequency domain power spectrum based on the structured time series; determining, via the one or more processors, values for a spectrum likelihood measure based at least in part on the frequency domain power spectrum; scaling, wherein the values of the spectrum likelihood measure are scaled to produce a power spectrum weight (PS_wt) value for an applicant; generating, via the one or more processors, a composite risk score by combining the PS_wt values with a plurality of actuarial model basis characteristics; comparing, via the one or more processors, the likelihood measure to a probability threshold; determining, via the one or more processors, a resultant insurance risk category from the result of said comparison; and storing, in an operational data store, the resultant insurance risk category.

“13. The method of claim 12, the method further comprising: providing an incentive to the applicant based at least in part on the resultant insurance risk category wherein the structured time series comprises a predicted future time series.

“14. The method of claim 12, wherein determining values for the spectrum likelihood measure comprises using at least one of a wavelet transform, a discrete cosine transform, a discrete fourier transform, a periodogram method, a Bartlett method, a Welsh method, or an autoregressive moving average estimate.

“15. The method of claim 14, wherein determining values for the spectrum likelihood measure comprises identifying a first set of frequency terms that have a higher frequency than a second set of frequency terms.

“16. The method of claim 15, wherein the first set of frequency terms are discarded.

“17. The method of claim 12, wherein determining values for the spectrum likelihood measure comprises generating repeated random permutations and sampling the resulting permutations utilizing a Bayesian Markov Chain Monte Carlo simulation.

“18. The method of claim 12, the method further comprising determining an entropy of the soft pulled time series or the structured time series.

“19. A system comprising: one or more processors; and computer-readable storage media having computer-executable instructions embodied thereon that, when executed by the one or more processors, facilitate a method comprising: computer-implemented method comprising: receiving soft pulled applicant credit utilization data; determining a soft pulled time series for the applicant credit utilization data; generating a structured time series by scaling the soft pulled time series; determining a frequency domain power spectrum based on the structured time series; determining values for a spectrum likelihood measure based at least in part on the frequency domain power spectrum; scaling the values of the spectrum likelihood measure are scaled to produce a power spectrum weight (PS_wt) value for an applicant; generating a composite risk score by combining the PS_wt values with a plurality of actuarial model basis characteristics; comparing the likelihood measure to a probability threshold; determining a resultant insurance risk category from the result of said comparison; and storing, in an operational data store, the resultant insurance risk category.

“20. The system of claim 19, wherein the one or more processors is further configured to facilitate the method comprising: providing an incentive to the applicant based at least in part on the resultant insurance risk category.”

For additional information on this patent, see: McNair, Douglas S. Insurance risk scoring based on credit utilization ratio. U.S. Patent Number 11158005, filed December 18, 2019, and published online on October 26, 2021. Patent URL: http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=%2Fnetahtml%2FPTO%2Fsrchnum.htm&r=1&f=G&l=50&s1=11158005.PN.&OS=PN/11158005RS=PN/11158005

(Our reports deliver fact-based news of research and discoveries from around the world.)