The Analogy

In the United States, cultural and government mandates have tended to place healthcare in its own category, distinct and siloed off from other industries. It’s true that both medical practice and the financing thereof are, in many ways, unlike any other industry, but that’s not to say they need to be treated differently or that Healthcare can’t learn from other industries.

Healthcare delivery can learn from the supply chain management of chain restaurants and their ability to deliver consistent quality across locations, time, and people by using standardization (Gawande, Big Med, 2012). Healthcare can learn from the airline industry and power of the checklist in quality enforcement (Gawande, The Checklist Manifesto, 2009). Even though the healthcare silo even delineates between a Master’s in Business Administration (MBA) and a Master’s in Health Administration (MHA), the fundamental principles of business, growth, and management are still more or less the same (Hekman, 2010), even if they’ve tended to be kept separate.

Further, I believe that we—patients, citizens, consumers, policy makers—can learn a great deal about healthcare financing by looking at other industries. Health insurance is unlike other types of insurance, but I would argue that this represents a misunderstanding the application of insurance to healthcare, rather than a difference in the fundamental nature and category of healthcare and health insurance. In this case, I believe we could learn a great deal about a better way to implement health insurance by looking at the auto insurance industry.

I’ve often made this analogy, but I want to flesh it out, and see how far it really goes.

Parallels between Healthcare and Car Repair

Because many people balk at the comparison initially, I want to start by highlighting areas where these two, seemingly disparate industries are, in fact, quite similar.

Specialized knowledge of complex system

Although the human body is more complex and less well understood than the internal combustion engine, they are both sufficiently complicated as to require specialized knowledge to diagnose and fix/treat. In both cases, therefore, the learning curve of acquiring that specialized knowledge and the costs—money, time, etc.—creates a natural barrier to entry to supply these particular services. The fact that such a barrier exists means that specialization and division of labor is not just advantageous, as it usually is (Roberts, 2010), but truly necessary: self-sufficiency is not an option for anything beyond the most rudimentary oil change or wound care. The fact that the human body is more complicated or that physicians receive commensurately more special training and certifications does not change this fact. It only means that physicians require more reimbursement than mechanics (both in pecuniary form and in prestige), which, in the US, they receive.

Asymmetric information

As a result of the specialized training and knowledge of mechanics and physicians, we have a situation that economists call “asymmetric information,” meaning that one party in a transaction knows something that the other party doesn’t (Arrow, 1963). This includes things on the consumer side where the consumer doesn’t disclose germane information to the provider (mechanic or medical provider), whether due to embarrassment or shame—for example, sexual practices or driving habits (Caine & Tierney, 2015)—or due to not believing that the information is relevant. This informational asymmetry also means that the suppliers have a much better understanding than the consumers of how reliable a diagnostic test is, whether or not a given service is truly necessary.

Fiduciary responsibility and history of bad actors

Because consumers don’t always know better, they tend to trust the experts—in this case, mechanics or physicians—about what services they do or don’t need, and an unscrupulous provider could use that to his or her advantage by recommending services that are not in the best interest of the consumer. Although the trope of the dishonest mechanic who recommends unnecessary parts and services is an old one, the same behavior in physicians—over-ordering unnecessary or low value care due to greed or defensive medicine—has recently come until the public discussion as well (Gawande, 2009; Gawande, 2015).

Insurance to spread out risk

Although we expect to require minor maintenance services for both our cars—oil changes, tire rotations, etc.—and our persons—routine physicals, medications, etc.—we also expect to have larger maintenance needs, but only occasionally. I might need an annual physical an annual tire rotation, but I also expect that there’s a possibility that I will get in a motor vehicle accident and require both medical care and car repairs, both of which will be quite expensive. It’s possible I’ll go through my whole life without a serious accident, but just in case, I’ve bought insurance to cover the costs of that emergency.

Furthermore, as of 2014, both health insurance and car insurance are required for almost everyone. New Hampshire and Virginia do not require car insurance and there are some exceptions to the individual mandate portion of the Affordable Care Act (ACA; aka ‘Obamacare’), but in general, insurance is required for everyone with health and/or a car.

Service is a small part of a critically important outcome

Both good health and functioning cars are critically important to the survival and well-being of most people. Obviously, a working car lacks the obvious life-and-death-hanging-in-the-balance mystique of trauma surgery and isn’t critical in densely urban areas, but reliable transportation is a major determinant of health. Our cars get us to and from work, to and from grocery stores, and generally play a role in most aspects of daily living in suburban and rural areas (RHIhub, 2017).

However, a mechanic isn’t needed to drive the car, just to fix the car when it breaks; just like a physician isn’t needed to help us perform activities of daily living: health is. But so much of the both our physical health and the “health” of our cars has to do with our environments and our behavior. Salty roads take a toll on the body of a car; and where we live (i.e. how rural) affects how many miles we put on our car. By the same token, where we live affects access to healthy foods and clean air. Our behavior also affects our health and the health of our cars. By some estimates, medical care only accounts for 20% (or less) of health outcomes, and social determinants of health (McGovern, Miller, & Hughes-Cromwick, 2014). I don’t know if anyone has done something similar for automotive outcomes, but it seems plausible to me that this is also the case for cars: regular access to high-quality mechanics account for only 10-20% of vehicle life.

Market structure

Finally, there’s market structure. The majority of mechanics and the majority of physicians are in small, private practices, with a handful of assistants to handle scheduling, billing, etc. Additionally, dealerships and hospitals hire a number of mechanics and physicians, respectively. The prices at dealerships and hospital-owned physician practices tend to be higher, but more reliable. The quality of independent practices tends to vary a lot more, such that you can probably get a better deal on a physical or repair by going to an independent provider, but the transaction costs to identify the high-quality providers can be considerable.

Where the Similarities End

Despite their similarities, auto repair is obviously not health care. Most notably, cars are not sentient beings; they are man-made machines for which we possess a complete blueprint. Thus, the ethical considerations associated with auto-repair are really just commercial/economic ethics, whereas medicine has separate set of ethical considerations apart from economic activity.

Several additional differences between auto repair and healthcare exist in the market dynamics of both auto repair and insurance purchasing. With health insurance, the majority of individuals get coverage through their employer or the US government. Auto insurance is virtually entirely private, with individuals purchasing their own insurance. Additionally, price transparency is virtually non-existent in healthcare. While prices still function on an estimate-only basis in auto body repair, the prices are still widely published online and all mechanic shops will provide an estimate in advance at no cost, which cannot be said of medical providers.

Nevertheless, the non-ethical differences between healthcare and auto repair are really just differences of degree, rather than fundamental differences in nature of each service. The human body is not fully understood the way cars are, but from the point of view of both the consumer and the mechanic, the result is more or less the same: cars a still a complex system that the individuals don’t fully understand, though the mechanics understand the system better than the customer; the same is true of healthcare, just with a greater degree of complexity in the subject.

What We Can Learn

The point of all of the above was to show that healthcare and auto repair are somewhat analogous, so what can we learn from this analogy? What lessons—successes or failures—from the auto repair industry can we apply to healthcare?

Lemon Law and Malpractice

When a physician orders, say, a total shoulder arthroplasty and the patient survives and is able to move her arm, then the surgery is considered a success. From the perspective of the patient though, the surgery should only be considered a success when it fixes the underlying problem—in the case of a shoulder arthroplasty, the pain or limited motion is resolved. Even if it wasn’t a success and the patient died on the table, the physician would get paid, which is understandable in that surgery is risky and we shouldn’t incentivize physicians to avoid caring for risky patients, but it’s no small surprise that physicians tend to over-order invasive procedures—procedures that other countries avoid due to cost and limited success (Reid, 2010)—physicians see all the benefits with little-to-no risk of a downside.

In the rest of the economy, particularly for automobiles, we have “Lemon Laws” that protect consumers—or at least give them a viable recourse—from being taken advantage of by the proverbial used car salesman. Specifically, the Magnuson-Moss Warranty Act of 1975 (MMWA) mandates federal standards for warrantees on consumer products and vehicles, and most states have passed similar, additional laws. MMWA and the subsequent state laws are far from perfect. Warranties are full of legalese and there’s a reasonable argument to be made that such consumer protection laws make us less safe because we all have warning fatigue. Nevertheless, a legal apparatus exists such that warranties need to meet certain standards and are enforceable, and there is a competitive advantage to be gained by having a warrantee or guarantee (Tommy Boy notwithstanding).

In medicine, we have malpractice lawsuits, but those are for cases of gross negligence or obvious misdiagnosis; a “successful” procedure that didn’t deliver the advertised results shouldn’t result in a malpractice suit because the physician did everything right in the execution of a given procedure; just not in the advertising and prognosis. Still, physicians should be held accountable for their recommendations, given that they are acting in a fiduciary capacity in an environment of asymmetric information. If they recommend a procedure that empirically has a low rate of solving the problem from the patient’s perspective, they should not receive financial benefits from it, at least not equal to the financial reimbursement they receive when they recommend and perform a procedure that does solve the patient’s problem.

Would it be possible to implement a medical warrantee? Politically, the answer is almost certainly no, given America’s history with health reform (Steinmo & Watts, 1995). But such laws have mitigated the negative effects of bad actors acting in a fiduciary role. In the auto industry, false advertising isn’t in the long run financial interest of car salesmen. If we were to implement similar laws in healthcare, we could prevent over-ordering of invasive procedures from being in the long term financial interest of physicians. Such laws wouldn’t penalize physicians with malpractice suits, but they could prevent physicians from being paid as much for doing a procedure that didn’t benefit the patient. For example, rather than being paid the same rate by Medicare whether the patient survives, recovers, or improves, Medicare could pay 100% of current rates for survival (the current standard), 110% for improvement, and 65% for recovery without improvement. Note that while this would seem to incentivize a physician to kill the patient on table (and make it look like an accident) rather than risk a recovery without improvement, I think it’s fair to say that (1) the vast majority of doctors would never intentionally kill a patient and (2) that would clearly fall under malpractice and, if proven, would result in a revoked medical license and possibly criminal charges.

Insurance Markets

Insurance, by definition, disperses risk. For high-cost, low probability events, we disperse the risk such that even in that low probability event, we aren’t wiped out by the cost. For example, home insurance costs less than a dollar a day, but in the unlikely event that your house burns down, you aren’t wiped out financially and left homeless. How insurance is implemented though seems to vary between healthcare and auto insurance. Here are some of the key differences in auto insurance, relative to health insurance, that I believe could be implemented in healthcare to the benefit of patients.

Not covering routine services

No car insurance covers oil changes, tire rotations, etc. These things are considered givens, and not part of the risk that insurance covers. After all, the probability of needing an oil change is 1, so the actuarial value of an oil change is simply the price an oil change. Similarly, a routine annual physical is an expected expense. There’s no “risk” of having an annual physical—it’s a given—so insurance, by definition, doesn’t make sense. In the auto industry, we follow this definition of insurance, but for some reason, we treat healthcare differently.

With very few exceptions, almost everyone can afford a $70-120 annual physical (adjusting for local cost of living), especially with the benefit of health savings accounts to save away $10 a month. By removing these and other routine medical expenses from the monies insurance are expected to pay, premiums will fall by the expected cost of an annual physical (price * probability an individual in the insurance pool gets a physical) times the overhead markup (about 20% in the US). Furthermore, if individuals are responsible for purchasing their own medical services, demand for price transparency will increase, and competition between providers will drive down prices.

These claims make two assumptions, neither of which are entirely true in the current US healthcare market: competition between insurers, and price transparency of medical services. However, I would argue that both of these unique elements of healthcare are a result of the unique setup of employer-sponsored healthcare, rather than anything fundamentally different about medicine. In both medicine and auto repair, the specifics of the fix will vary wildly by the patient (human or auto), and therefore be more or less expensive, but the fact that most physicians will not, or are unable to, give an estimated cost of services ex ante is more the result of a lack of demand, than an impossibility of doing so. Removing expected services from insurance coverage will, and has already begun to, increase the demand for price transparency in healthcare.

Price discrimination

Under the ACA, limits were placed on the premium increase insurers can charge on sicker (read: more expensive) subscribers, relative to healthy subscribers. Specifically, the ACA allows at most a 3:1 ratio between the lowest risk level and the highest (the American Health Care Act, the potential Republican replacement to the ACA, would allow a 5:1 ratio). While this limits the increases unhealthy patients—elderly, smokers, obese patients, sedentary patients, etc.—see in their premiums, less risk stratification also means that healthy patients will pay more for insurance, subsidizing their more expensive counterparts in the insurance pool.

In the market for auto insurance, we can see some evidence of the opposite effect: auto insurers give discounts for things like age (esp. over 25), good grades, driving record, and other negative correlates to getting in a car accident. Granted, these big data algorithms are not without negative social consequences: things like zip code may be highly predictive of an individual’s health and driving record, but also tend to discriminate against the already disadvantaged (O’Neil, 2016). Still, the allowance of such price discrimination, as long as it is done responsibly, can mitigate the negative impacts of an “individual mandate” as seen in auto insurance by minimizing some of subsidy low utilizers give to high utilizers and therefore preventing an insurance “death spiral.” I’m not advocating that this necessarily be a part of health insurance policy, but is worth noting that this has been a successful policy in an analogous market and therefore incentives like this could be a tool to mitigate rising premiums for healthy, young adults.

Who purchases insurance?

As previously mentioned, the majority of working-age Americans, get insurance through their employer. This means that health insurance salesmen need to cater to HR managers, rather than directly to patients. Auto insurance, on the other hand, is primarily sold directly to drivers. Health insurance is complicated, but so is auto insurance—many of the terms of insurance like “deductible” and “copay” that allegedly cause so much frustration and confusion in healthcare are still used auto insurance. The difference is that auto insurers like esurance have a strong incentive to make their product more understandable, and, in my experience, they have succeeded in doing so.

The human body is more complicated than a car, so it’s understandable that insurance for health will be more complicated than insurance for a car, but it doesn’t follow that insurance for health should be impossible for the average purchaser to understand, as long as insurers have the incentives to make their policies understandable. Removing employer-sponsored health insurance would increase employee pay (by transferring the portion of insurance that employers pay directly to the workers) and incentivize insurers to demystify some of the nuances of health insurance for subscribers.

Individual Mandates

The individual mandate portion of the ACA was highly contentious with voters. Even many (~46%) liberal voters have an unfavorable view of the individual mandate (KFF, 2016). This begs the question: why? We have an individual mandate on auto insurance, why is that not as contentious? My view is that this is primarily the result of a well-executed political theater on the part of conservative politicians and pundits, but given the importance in an individual mandate in preventing a death spiral, liberal politicians would do well to make this comparison. Auto insurance requirements are not contentious, so it’s unclear why similar requirements in healthcare should be so contentious.

Closing Remarks

Political implications

As noted throughout the “what we can learn” section of this post, I believe that treating health insurance more like auto insurance would be beneficial to consumers. I also acknowledge that this is a tall order politically, given the tendency of individuals—politicians, pundits, and voters alike—to view healthcare as exceptional. That’s really the main point of this post: healthcare may be unique by degree, but it is not unique in its fundamentals. Framing the discussion differently and putting health insurance in terms more people understand demystifies the topic and, I believe, increases the likelihood of finding common ground and a consensus.

Notes about this post

One of my goals for this year is to write more, and this post is part of that goal. That said, I’m not entirely happy with this post, especially the “what we can learn” section. However, I’ve spent enough time writing this, and believe that done is better than perfect. I would like to come back at a later date to revisit some of these lessons with mathematical models and data where it’s available. But for now, I ask you, my readers, to not focus on the specific lessons learned, but rather on the main point: healthcare may be unique by degree, but it is not unique in its fundamentals. The lessons are more of examples than specific recommendations.


Arrow, K. J. (1963). Uncertainty and the Welfare Economics of Medical Economics. The American Economic Review, 53(5), 941-973.

Caine, K., & Tierney, W. M. (2015). Point and Counterpoint: Patient Control of Access to Data in Their Electronic Health Records. Journal of General Internal Medicine, 30(S1), 38-41. doi:10.1007/s11606-014-3061-0

Gawande, A. (2009). The Checklist Manifesto. Henry Holt and Company.

Gawande, A. (2009, June 1). The Cost Conundrum. The New Yorker. Retrieved from

Gawande, A. (2012, August 13). Big Med. The New Yorker. Retrieved February 13, 2017, from

Gawande, A. (2015, May 11). Overkill. The New Yorker. Retrieved from

Hekman, K. (2010). Curiosity Keeps the Cat Alive. Holland, MI: Trillium Arts Press. Retrieved from

KFF. (2016, December 1). Kaiser Health Tracking Poll: November 2016. Retrieved from

McGovern, L., Miller, G., & Hughes-Cromwick, P. (2014, August 21). Health Policy Brief: The Relative Contribution of Multiple Determinants to Health Outcomes. Health Affairs, 1-9. doi:10.1377/hpb2014.17

O’Neil, C. (2016, October 3). Cathy O’Neil on Weapons of Math Destruction. (R. Roberts, Interviewer) EconTalk. Retrieved from

Reid, T. R. (2010). The Healing of America. Penguin Group LLC.

RHIhub. (2017). Social Determinants of Health for Rural People. Retrieved March 5, 2017, from Rural Health Information Hub:

Roberts, R. (2010, Feb 8). Roberts on Smith, Ricardo, and Trade. Retrieved from EconTalk:

Steinmo, S., & Watts, J. (1995). It’s the Institutions, Stupid! Why Comprehensive National Health Insurance Always Fails in America. Journal of Politics, Policy, and Law, 20(2), 329-372.