Credit-Based Insurance Scores (Kucera & McCarty) Flashcards
Kucera: Use of credit-based insurance scores…
helps insurers subdivide risks to determine appropriate rates
Not using insurance score will not lower the overall insurance premium, but will redistribute charges:
a. Risks with lower expected costs will pay more than actuarially fair
b. Risks with greater expected costs will pay less than is actuarially fair
Insurance Score
Numerical score assigned to an insurance risk based on that risk’s underlying characteristics (gender, age, etc.)
Common purpose is to provide useful information in underwriting and pricing insurance
Provides a relative measure of the expected cost of a risk
Credit-Based Insurance Score
Uses items found in a typical individual’s credit report (such as number of inquiries into opening new accounts and account 30 days or more past due)
Models developed by third-party vendors and individual insurance companies (generally the higher the score, the better an individual’s credit rating, therefore lower expected losses)
There is a strong correlation between insurance scores and expected costs associated with the risk (Kucera)
Scores are a statistically reliable tool for segmenting risks with different expected costs
How Insurers Use Credit-Based Insurance Scores
Some insurers use them to determine whether to accept or reject a risk
More commonly used to segment risks into homogeneous groups for rating
a. May be used directly as a rating factor (aka risk classification factor)
b. May be used to assign risk to the appropriate tier
Kucera’s pros of using Credit-Based Insurance Scores
- Relationship of Risk Characteristics and Expected Outcomes (ASOP 12):
a. Actuary should select risk characteristics that are related to expected outcomes
b. Relationship between the risk characteristics and the price is demonstrated if it can be shown that the variation in expected losses correlates to the risk characteristics
—-> Several studies have shown that credit scores reflect differences in expected loss costs - Stopping the use of insurance scores will not lower overall premium collected
-Will just redistribute premium collected
-Risks with lower expected costs will pay more and risks with greater expected costs will pay less - Companies have reported that the use of insurance scores has enable them to write more risks
Discussion of how current economic conditions have affected policyholder premiums related to Credit-Based Insurance scores
-Uncertain how current economic crisis will affect overall insurance costs and prices
-Some regulators concerned that if insurance scores worsen, it will lead to unwarranted premium increases
Counterarguments by Kucera to concerns by regulators in prior flashcard
a. Insurance scores are used to determine rate relationships between classes, not overall premium need
b. If there is a distributional shift (in the economy/insurance scores), actuaries adjust overall rate levels so total premium is adequate.
-This is no difference than pricing homeowners insurance with increasing average home values.
-Any shift in insurance scores due to current adverse economic conditions will not result in any long-term impact on overall premium collected
c. Any distribution shift is likely to have a smaller effect on renewal business
-Some states/companies only allow use of scores for renewals if it reduces the insured’s premium
d. Concern: Some regulators concerned that a dramatic shift in insurance scores could change current relative rates among risks
Counterargument: Competition motivates companies to regularly review rate differences. E.g. insurers adjusted class plans as differential between young males and females changed
e. Concern: Possible insured harm if sudden distribution shift doesn’t show in the data in time to adjust
Counterargument: To date, no quantifiable evidence to show that a dramatic shift has been occurring. Author’s opinion that shift so far has been minor, which could be because majority of business is made up of renewals.
McCarty: Use of credit-based insurance scores…
Disparately impacts certain classes of people such as lower income individuals and protected classes of people.
We must examine the fundamental purpose of insurance and acceptability of rating factors. There are many factors that show mathematical correlation with claims, but that does not necessarily make them fair and valid criteria (health data, DNA testing)
Other Rating Factors Considered to be Inappropriate
Race
-Based purely on actuarial rates, could be used to justify higher rates. This is counter to equal protection for consumers and not sound public policy.
Genetic testing for predispositions of inherited diseases (Outlawed in 1996 through HIPAA)
In 2007, reviewed use of occupation and education as a factor for PPA. Questioned whether it was acting as a proxy for race.
McCarty is sensitive to…
Factors highly correlated with race, ethnicity, religious background, or income level.
Inherent weaknesses in the Credit Reporting System
a. 2000 study by Consumer Reports showed 50% of credit reports contained errors
b. Identity theft can damage the credit of individuals
c. Excessive access to credit as evidenced by the problems in the mortgage industry (when you apply for a mortgage, banks will pull your credit which has an adverse impact on your credit score)
d. Credit reports disproportionately effect recent divorcees, recently naturalized citizens, elderly, disabled, younger individuals who have not established credit histories.
Even if methodology is correct, it is possible that inaccuracies in credit reports may invalidate their use
Downturn in economy could potentially magnify differences in credit scores among vulnerable populations.
Empirical studies show no significant difference in magnitude of claims, only frequency.
-Consumers with lower credit scores file more claims
-It is possible that frequency of insured loss events is the same across populations, but those with higher scores are less likely to file a claims (they may deal with minor claims so claims history is not impacted)
-No studies suggest claims being filed are not legitimate
Methodology used to create scores is opaque to consumers.
-Varies company to company
-Negatively impacted by sound financial decisions (not using credit cards, having too few credit cards, having installment loan)
Disproportionate Impact of Credit-Based Insurance Scores
Race
-There is a strong relationship between credit score and race.
—–> Is this relationship strong enough to prohibit use given American values of equal protection and nondiscrimination
—–> FTC report clearly demonstrates strong correlations
Disparate impact on young people and elderly
Disparate impact on certain religions
—–> May have lower credit scores due to non-use of credit
FTC Report
FTC Report (2007) clearly showed strong correlations between credit score and race
McCarty overall opinion
-McCarty believes negative impact on classes based on race, age, and religions outweighs any suggested enhanced accuracy in pricing and underwriting
-McCarty disappointed by 2007 FTC Report
