TCS Daily

The Mythical Health Care Man-Month

By Arnold Kling - April 7, 2005 12:00 AM

"If each part of the task must be separately coordinated with each other part, the effort increases as n(n-1)/2. Three workers require three times as much pairwise intercommunication as two; four require six times as much as two. If, moreover, there need to be conferences among three, four, etc., workers to resolve things jointly, matters get worse yet."
-- Frederick P. Brooks, Jr.,The Mythical Man-Month.

Health insurance generates a lot of paperwork and a lot of overhead. The latest national health care accounts show that administrative costs amount to over $100 billion, or more than 5 percent of total health care spending. I think that some of this cost could be reduced by eliminating some of the unnecessary intercommunication. The legendary software engineer Frederick Brooks points out that as you add interfaces, the number of permutations can quickly get out of hand.


I read Brooks' classic essay on software project management during a particularly stressful period in my career. I was managing a web site that I had launched, and as my partners found more ways to make money from the site, the software became increasingly complicated. I even ended up pushing back against their marketing wizardry, saying that they were creating too many permutations. I confided to friends that I was suffering from bad karma: "After spending so many years berating information technology bureaucrats, I've come back as one."


The point of the "mythical man-month" is that a software project that can be completed in 6 months by 3 people cannot be completed in 3 months by 6 people, as would be the case if the "man-month" were a useful metric. Instead, productivity in software development is inversely related to the number of people who must co-ordinate their actions in order to complete the project.


This insight applies to more than just software. The more entities that must co-ordinate in a business process, the less efficient will be that process. In fact, one of the ways that management consultants earn large fees is by helping businesses to simplify and streamline their operations.


Counting the Interfaces


Taking a management consultant's approach, it is easy to see why health insurance would have a lot of overhead. Consider the list of entities with which a health insurance company must interact.





Other Therapists

Other Specialists (dentists, psychiatrists, etc.)

State Regulators


Other Insurance Providers


A health insurance company cannot simply design a plan for consumers. Because most health care plans are provided by businesses, the insurance company has to tailor its plans to employers.


An auto insurance company has to deal with suppliers of automobile repair services, which is a relatively homogeneous market. For health insurance, the suppliers range widely in size, from independent individual therapists to large hospitals. They range widely in scope, from eye specialists to cardiologists to psychiatrists to occupational therapists.


In any given year, a consumer may not make any sort of claim with their auto insurance or homeowners' insurance. However, the way most health insurance is structured, every time you visit a health care provider you submit a form to the insurance company.


Reducing the Interfaces


A management consultant would recommend eliminating unnecessary interfaces. For example, you could eliminate state regulators and employers from the insurance process. Having 51 different sets of state health regulations means that it is impossible for health insurers to offer the same plans across state lines. This makes it harder to create large risk pools. It also means that various classes of health care providers can lobby their legislators to mandate coverage for their specialties, rather than allowing insurance companies to offer simpler, limited packages of protection.


Another unnecessary interface is with employers. If health insurance were offered directly to consumers, rather than to employers, then any given plan could cover more workers. It would eliminate the need to set up marketing relationships and reimbursement rules that are tied to employers.


People on the Left often argue that having the Federal government be the "single payer" for health care would reduce administrative costs. This might turn out to be true, if for no other reason than that the Federal government would override state regulations and would bypass employers. That would get rid of two interfaces. Since the effect of interfaces is multiplicative, getting rid of two interfaces could reduce costs by a factor of 4.


Of course, if the key to reducing excess overhead is getting rid of the employer and state regulator interfaces, this could be done within the context of a private insurance market. The government could eliminate the tax subsidy for employer-provided health insurance; and the Federal government could allow an insurance company to sell a plan that is approved in one state in another state, regardless of the latter state's regulations.


The Adjuster Model


Recently, one of our cars was damaged while it was parked by a driver who snuck away, and our insurance company did not reimburse the auto body shop. Instead, the company sent out an adjuster, who estimated the cost of the repair, and the insurance company mailed a check directly to us.


Imagine how much simpler it would be if health insurance worked that way. After you break your wrist, you would pay for X-rays and initial consultations. Next, an adjuster looks at the X-rays, as well as the reports of the radiologist and the orthopedist. The adjuster then estimates the cost of the necessary surgery, physical therapy, etc., and the insurance company mails you a check. That would eliminate the whole billing interface between the insurance company and the health care providers.


In addition to reducing overhead, this "adjuster model" would be more consistent with turning health insurance into real health insurance. See my earlier essay comparing event-based health insurance with our current system.


Regulatory Barriers


What is preventing the private health insurance market from reducing its interfaces and adopting something like the adjuster model? One possibility is that there is something wrong with my management consultant thinking, and the current system is actually relatively efficient. However, I suspect that regulatory barriers and other policy distortions are the real culprit. If the Federal government were to eliminate the tax subsidy for employer-provided health insurance and make it easier for health insurance companies to sell plans across state lines, and if at least one state health insurance regulator were to allow event-based health insurance to be marketed, then we could see whether a more efficient health insurance system could emerge from the private sector.


TCS Daily Archives