Skip Ribbon Commands
Skip to main content

3 Factors Negatively Impacting the Insurance Marketplace

Author: Chris Boggs

Three distinct phenomena negatively impact the commercial and, to some extent, personal insurance marketplace: The Winner's Curse, Submission Bias, and Overconfidence. These phenomena are explained in an international whitepaper titled, "Analyzing the Disconnect Between the Reinsurance Submission and Global Underwriter's Needs."

Awarded the Brian Hey Prize, this whitepaper was developed by an international working party for presentation to the Institute and Faculty of Actuaries (IFoA) and the Casualty Actuarial Society (CAS). Members of this working party hailed from around the world. I'm proud to say I was a member of this group; though I'm not the originator of these market-force phenomena. Professionals much smarter than I am developed these market factors.

The Winners 's Curse

"Winner's Curse" as a market-impacting concept was first introduced and coined by Capen, Clapp & Campbell in their 1971 paper, "Competitive Bidding in High Risk Situations."

The concept postulates that in any scenario where multiple parties are given access to certain, but not necessarily complete, data and asked to estimate a quantity/price from that data, a wide range of estimates typically arise. The average of the parties' estimates is often or can be a good estimator of the unknown quantity. This is the so-called "Wisdom of the Crowds."

This concept is proven over and over in prediction surveys where the mean prediction of the group is consistently better than the predictions of individual respondents. However, there are some circumstances, such as auctions, where the estimate that matters is not the mean estimate but the extreme estimate, i.e. the highest price at auction or the lowest price for a quoted insurance policy (an example of a reverse auction). In such situations, the "winner" is likely to have been "cursed" by either paying too much for the goods at auction or obtaining insufficient premium for the insured risk. – (Winner's Curse GIRO Working Party, 2009)

When an insurer "outbids" its competitors for a risk (by pricing coverage below the competition), but offers coverage at an unprofitably low premium – that's the Winner's Curse. Yes, they "won" the business, but they lost because they underpriced the real risk.

According to research, the existence and impact of the Winner's Curse is directly proportional to the number of competitors. As the number of competitors increases, so does the impact of the Winner's Curse. When there is no competition, theoretically the bidder can gather the necessary detail to adequately price the risk. The only curse of winning in this case is an unexpected event resulting in greater-than-anticipated losses. As competition is introduced, prices drop and the impact of the Winner's Curse is intensified. The "winner" may have underpriced or inadequately priced the risk to win the business.

In addition to competition, the Winner's Curse is exacerbated by poor data gathering methods and/or poor pricing models. For sake of the example, if two insurers are competing for an account, the carrier that gathers the most detailed information is more likely to properly price the risk. Conversely, the carrier that bids based on minimum information is more likely to improperly price the risk, and sometimes to its detriment.

A carrier using inferior data may "win" the account because it failed to gather the information necessary to properly price the risk. The Winner's Curse shall befall this carrier.

Be warned, this is an oversimplified description of the Winner's Curse; but it adequately highlights the ultimate result of competition and inferior data gathering on the winning carrier. Not discussed in this short description of the Winner's Curse is the negative effect of inferior pricing models or a "loose" pricing philosophy.

Submission Bias

Because more and better data allows for a more accurate assessment and pricing of a risk, only the best risks (the best insureds) will provide all the necessary data (and maybe more). Poor risks will be less willing to provide all the necessary data because they know the consequence of too much information is higher pricing or a refusal to offer protection.

Submission bias is also known as Information Asymmetry. If all risks provided the same data, the better risks would pay a lower-than-average premium and the poor risks would pay a higher-than-average premium. This is considered a balanced market. But because there is not a balanced market, assumptions are made about the risks that do not provide detailed information and this increases the possibility that the poor risks are charged less than the technically correct premium.

In effect, poor risks have no incentive to provide detailed information beyond what is required to garner a quote. This is a failure of the underwriter being willing to accept less than is necessary to adequately rate the risk.

Soft markets increase the incidence of submission bias. In a soft market, insurers are more aggressive in writing coverage and may be more willing to offer coverage with less information. Therefore, a bad risk can more easily get away with providing an incomplete or less-detailed submission.

Underwriting managers could mitigate the effect of submission bias simply by standardizing submission requirements. However, carriers that practice underwriting discipline might lose more than they win; but in light of the Winner's Curse, did they really lose? Also, because disciplined carriers should have theoretically better results, they can more competitively price the better risks that give all the necessary information anyway – lessening the effect of the Winner's Curse.


"Good decision making requires more than knowledge of facts, concepts, and relationships. It also requires metaknowledge – an understanding of the limits of our knowledge. Unfortunately, we tend to have a deeply rooted overconfidence in our beliefs and judgments. Because metaknowledge is not recognized or rewarded in practice, nor instilled during formal education, overconfidence has remained a hidden flaw in managerial decision making." (Managing Overconfidence, Russo & Shoemaker, 1992)

Overconfidence is a common trait among decision makers, and underwriters and actuaries are not immune. In one test case, 374 respondents were asked 10 questions related to their knowledge of the global insurance industry; each respondent was to answer with a 90 percent confidence interval (Collins, 2004).

Ideally, respondents would have captured the answer in their ranges nine out of 10 times. However, in this survey, only seven (2 percent) of the respondents were able to supply proper ranges at least nine out of 10 times. Further, 60 percent of the respondents were able to capture the proper range only 3 three or less times (29 of the 374 did not get any of the 10 questions right, even with being offered to give a 90% confidence interval).

Overconfidence results in mispricing of risks. But in what areas are underwriters and actuaries overconfident? Their ability to predict or estimate future contingent events. Yes, an overconfidence on pricing models can improperly affect pricing – up or down.

Market Phenomena that Impact Insurance – And Your Responsibility

These phenomena are relegated to your insurance carrier partners. Your job as an agent is to work with your carriers to help them avoid stepping into the pricing spirals that result from these three phenomena.

Provide them all the data they need to properly price the risks and remind them that pricing models aren't the Gospel. Yes, I know, they probably won't listen to you about their models, but at least you can help them avoid curses and biases.

Information for this article taken from: IFoA / CAS International Research Working Party ("Analyzing the Disconnect Between the Reinsurance Submission and Global Underwriter's Needs.")

Last Updated: October 12, 2018​

​127 South Peyton Street
Alexandria VA 22314
​phone: 800.221.7917
fax: 703.683.7556

Follow Us!

Empowering Trusted Choice®
Independent Insurance Agents.​