The Takeaways:
- Insurance is a future oriented business: Insurers (insureds) charge (pay) premiums to prepare for risk and uncertainties in the future, not in the past.
- The business model of the insurance industry that everyone has been talking about is to make money from collecting and investing premiums. But we should add in risk sharing, which leverages the uneven and asymmetrical chances of filing claims by different policyholders to protect underwriting profit.
- Risk sharing differs from risk pooling. The former adds an additional step of selecting for good risk to the latter.
- Risk sharing has two forms: people sharing and money sharing. Cost sharing, a term familiar to most in the health insurance market that takes the forms of deductibles, co-payment and coinsurance, is a part of money sharing, together with others like different premium, lapsed policies and universal premium rate changes in a state.
- Risk sharing works for both predictable and unpredictable risks. In health insurance the law prohibits discrimination against preexisting medical conditions, both in enrollment and in premium charged. While this makes risk prediction irrelevant, risk sharing still works through enrolling as many people as possible when they are entitled to and eligible for Medicare.
- Risk sharing may have a reputation of being unfair to different policyholders, but that’s wrong because life is fair in terms of all humans facing the same future related uncertainty and risk. While those with insurance claims are being financially rewarded by insurers, those without filing insurance claims are gaining peace of mind. Nobody’s premium is wasted.
Surprisingly, as far as I know nobody has talked comprehensively about how insurance companies (or “insurers” for short) make money as a whole. In my view, if you don’t know the business model of an industry or a profession, you don’t know the industry at all no matter how long you have been doing it.
Unfortunately most businesspeople focus on how to make money for themselves, not the fundamental business model. As a result, this topic has been rarely touched upon even at the most basic conceptual level. We end up seeing “trees” but not “forest” even though no tree exists alone and the ecology of forest shapes each and every tree.
A quick sidenote: A business model is just a strategic plan of how a company, a business entity, or in our case, the entire industry, will make money. Here making money means earning a profit by receiving more money than spending it.
Insurance Is a Future Oriented Business
But don’t take the topic for granted or presume it’s meaningful. Instead, it’s always healthy to ask why we should be bothered with a question or a topic. Compare insurance with other businesses helps.
Consider retail businesses (e.g., supermarkets, department stores, or online retail distributors like Amazon), which make money by selling goods or services to customers. Insurers receive money from policyholders’ premium rather than from, and insurers pay money through fulfilling insurance claims (to designated beneficiaries, policyholders, healthcare service providers, the courts, employees, third-party claimants, to name just a few).
There, we see that insurance is a future oriented business, meaning insurers receive and spend money for events in the future, while retailers do business by selling previously made goods and/or existing services. They typically won’t collect consumers’ money ahead of time without offering goods or services in exchange.
The Challenge of Charging a Price for the Future
Insurance is not the only future oriented business, many other (e.g., renewable energy, space technologies and even healthcare, in which a crucial part is to develop new drugs and new treatment regimens) do that.
Going deeper, almost all businesses have to deal with elements of the future. It’s not that some businesses are exclusively working for the past while others exclusively for the future. Therefore, I won’t say insurance is unique because it focuses on the future.
Of course, a future focus does have its ramifications, chiefly among them is future related uncertainties or risks. For example, it is relatively simple for retailers to decide how much to charge customers: They can always use the production cost, plus transportation or distribution cost to get started. The nice part is that both production and distribution costs are known before sales, retailers only need to add the desired profit margin to set the final price.
Insurance has a different story. We must determine the price (i.e., premium) of our product (i.e., insurance policy) ahead of time or before an accident or an undesirable event happens to policyholders. We don’t have the luxury of seeing the event first, and then figure out the cost and consequences to decide how much the premium should be.
Insurers Have a Unique Business Model
Determining the right premium is a risky and uncertain business. If we charge too much, we lose business to the competitors; if we charge too little, we fail to cover the bottom lines to grow.
But risk is everywhere, they just come in different shapes and forms. Just like all businesses have to deal with the future, they also face risk and uncertainties. Retailers take the risk of buying the wrong amount of goods from wholesalers, for example. If they buy too much, more than consumers would buy, they suffer a loss by having to cut down sales price; if they buy too little, consumers will move to competitors with more stable supply.
Maintaining “just right” inventories is just one example. More risk arises from predicting consumers’ future taste and preferences. We have seen stories where retailers missed the trend of demand and drove themselves out of business altogether.
Risk Sharing Makes Insurance Unique
Once again, my point is that the uniqueness of insurance is not in dealing with risk, even with future oriented risk, but in its unique business model. This is where the idea of “risk sharing” comes into play.
There have been posts, articles and websites on how insurers make money. A good example is the article by Policygenius.com published April 2023. It summarizes four ways life insurance companies make money: charging premiums, investing premiums, cash value investments, and lapsed policies.
I won’t get into details but will point out that the above four ways really apply to all insurers, not just life insurers. All insurers share the same business model that has been functioning for centuries.
The problem is that the presentations so far have all overlooked one fundamental part of risk sharing.
That’s right, risk sharing plays a fundamental role in all lines of insurance, be it personal or commercial, life or health, properties or liabilities. It does not negate or eliminate the role played by premium investment, but does cut down its relative importance.
Simply put, before collecting and investing premiums, the two ways discussed by pretty much everyone, we need something else in place, something that has unfortunately been severely underappreciated.
The “Law of Large Number” Differs from Risk Sharing
Before defining what risk sharing is, let’s find out what it is not. We do hear the “law of large numbers” that has been taught as one of the fundamental axioms for insurance business. It states that the larger the number of exposure units independently exposed to loss, the greater the accuracy of the prediction of loss.
As a statistical axiom, this law is applicable everywhere, not just insurance. In population surveys or polling, one can be easily convinced that the result from a sample of 1,000 is more reliable than that from a sample of 100 people, for example.
But the law of large numbers is not the same as risk sharing, even though both are relevant and related to each other. The former helps insurance companies estimate the possibility of making insurance claims to be paid by insurer. Once the risk of claim is estimated right, insurers then charge the corresponding premium to cover the estimated claim losses, which is the key step in insurance underwriting.
Risk Pooling vs Risk Sharing
The law of large numbers is sometimes interpreted as “pooling of risks,” which is inaccurate. In its original and official sense, the law simply tells us that the bigger the sample size, the more accurate the sample estimates will be — other things equal, no more and no less.
If the law really means pooling risks, it must gather diverse risks of different types and shapes, as gathering similar risks makes little difference. To use an extreme example, talking to 1,000 people of identical age, gender and/or health record is the same as talking to one person. Pooling risk only makes sense when it works with risk diversity.
But here is what “risk pooling” differs from “risk sharing:” While the former works with diverse risks, risk sharing favored by insurers works with selected risks. Whenever possible, insurers always seek policyholders who possess low or “good risk,” which ultimately comes down to a low probability of making a huge insurance claim in the future.
Insurance firms do not hide their preference. When one applies for an insurance policy, insurers will typically ask questions regarding one’s claim history. A history of frequent claims, especially those involving a large amount of money, will make one’s application more likely to be rejected, or being charged a higher premium.
Predictable vs Unpredictable Risks
Selecting good risk matters because it is the main way for insurers to boost underwriting profit or reduce underwriting loss (i.e., the net profit or loss an insurer makes from issuing insurance policies without counting in investment gains).
Let me use a hypothetical example to illustrate underwriting profit or loss: Say an insurer has 1,000 policyholders, and collects $1 million premium each month but on average pays out $1.5 million on claims. The insurer will have a net underwriting loss of $0.5 million, even though it has an average gain of $50,000 each month from investing the premium in financial markets.
But no matter how carefully insurers select for good risk, bad risk will enter the population of policyholders, especially with simplified underwriting that skips medical exam and collects less information from applicants but relies on third-party sources to gather information about the applicant, such as their prescription drug history and driving record.
Bad risk or the number of claims also rise dramatically during natural disasters (for property insurance like hurricanes in Florida) and pandemics (for health insurance like Covid-19).
But “good” and “bad” risks are all relative and can be turned into each other. This is because insurance handles more or less unpredictable events. An auto insurance policy covers only injury and damage from a traffic accident but not intentional damages, wars and losses from regular wear and tear, all because these losses are highly predictable, unlike an accident.
Similarly, a homeowner policy protects your home from rare or accidental perils like fires, hail, theft, windstorm, smoke, lightning, explosion, riot or civil commotion, tree falling and volcanic eruption — but not from wars, earthquakes, landslides, floods, or even a large scale power failures, which all share the feature of incurring more predictable and heavy losses.
But here is the thing: A rare and accidental event can turn a “good” risk into “bad” and does the opposite for an originally “bad” one. Consider a hypothetical example: An auto policyholder who has never even had a single traffic ticket for 30 years is certainly considered a “good risk” in the eyes of insurer. However, this poor guy has recently lost his wife of 25 years and decided to drink a little. The next thing we know is that he hit five cars in a row and incurred injuries and damages at more than $1 million.
My point is that insurers can only do so much in selecting good risk, and no matter how hard they try, they will have to deal with some bad risks in the real world — the question is how. It’s never enough for insurers to predict risk — they must share it among policyholders.
Defining Risk Sharing
A formal definition of risk sharing is the constantly functioning mechanism that allows insurers to leverage the uneven and asymmetrical chances of filing insurance claims by different policyholders to protect underwriting profit and/or reduce underwriting loss.
The reason risk sharing works all the time is due to the silent fact that some policyholders will contribute premium without making any claim, or making fewer and smaller claims, while others may file large and /or frequent claims. People in the former group effectively pay premiums that end up covering claims by policyholders in the latter group.
It is risk sharing that makes the insurance business model unique, more so than the other two parts of collecting and investing premiums.
Two Forms of Risk Sharing
Risk sharing has two forms: people sharing and money sharing. Cost sharing, a term familiar to most in the health insurance market that takes the forms of deductibles, co-payment and coinsurance, is a part of money sharing.
Money sharing takes other forms, such as charging different premiums for different policyholders with different risk profiles. Universal money sharing is also possible in a state where rate increase applies for all existing and future policyholders, like we see in Florida, California and other states recently.
Finally, money sharing also comes when some policyholders had trouble making premium payments, and their lapsed policies will leave money to the insurer, indirectly funding insurance claims.
People sharing is the least known, but works quietly and forms the foundation for money sharing. Without people staying, either among policyholders with the same insurer or among policyholders in the same state, money sharing is very difficult or even impossible. After all, it can only happen among the people in the same state or with the same insurer.
It is not much different from a bank having a diverse clientele base like Bank of America or JP Morgan Chase versus a bank of highly homogenous clientele like Silicon Valley Bank. In the former different clients involuntarily assist each other due to different needs and preferences, making the bank financially more stable, while in the latter the missing diversity makes the bank financially more vulnerable.
Another simpler analogue is a housing co-op, where residents share a meal plan with all residents paying the same or similar amount of money, but those eating less will quietly help those eating more every meal without making it a big deal.
Risk Sharing with Predictable Risk
Risk sharing works not only for unpredictable risk (e.g., in an auto policy) but predictable ones (e.g., in health insurance, where insurers have no choice but to enroll everyone eligible, even knowing the extra cost associated with people of preexisting medical conditions.
Yes, in health insurance the risk is somewhat predictable — to the extent that people with preexisting medical conditions tend to have higher chances of making medical claims while enrolled in Medicare, Medicare Advantage, Original Medicare (Parts A & B) and Medicare Part D.
It is the laws and regulations that make risk prediction largely irrelevant. Since 2014 the Affordable Care Act (aka Obamacare) has changed the previously more or less accepted practice of discrimination against pre-existing medical conditions. Now no eligible enrollees can be denied enrollment nor be charged higher premiums. This means insurers won’t spend time figuring out who is more likely to file a claim.
The good news is that even when risk prediction becomes irrelevant, risk sharing still works for programs like Medicare. Here is how.
While the law prohibits discrimination against preexisting medical conditions, it also places penalties to enroll a sufficiently large number of people when they first become entitled to Medicare Part A (in patient services) and eligible to enroll in Part B (out of patient services). It’s a number’s game and with all enrollees on board, there will be enrollees who may never need intensive healthcare or will need it much later in life, by the time they would have paid enough premium to cover themselves and perhaps even provide surplus money for others.
Risk Sharing Is Still Win-Win
Why is risk sharing not openly discussed? Fairness and justice is perhaps a key concern. If some policyholders are essentially “money doners” while other recipients, it does not sound fair. On the country, the law of large numbers sounds “scientific” and entirely fair and safe to talk about.
But the seemingly strong argument against risk sharing has a problem: It ignores the fact that future insurance risk can never be completely and accurately predicted. Consider health insurance: Yes, people with preexisting medical conditions are risky in the future, but so are people with a very clean medical record. We have all heard stories of sudden death of presumably very healthy individuals. The truth is no one can guarantee you a “claimless” life. A fancier way to say this is that your probability of getting sick is never 0 nor 1. It falls somewhere in between. Life is fair at least in this fact involving probabilities.
Some policyholders or enrollees will be rewarded financially through claims, while others gain peace of mind without claim. We can even make a point on the “fairness” ground: Having insurance coverage for everyone provides a fair and safe environment for everyone in society. Consider the FDIC insurance for banks, when all depositors are protected, everyone gains directly or indirectly from a sound banking system, even though most banks may never need the FDIC protection. It seems fair to say that nobody’s premium will be wasted, with or without making a claim.