More on Incentives

A lot of my education was just a cover story that looked good a number of years of partying with no job look good on my resume without any real improvement in my long-term skills.  But my time in business school thinking about incentives and later at McKinsey & Co. doing the same for various employee compensation approaches has served me well through my whole life.  It's not that Congress and the media are bad at thinking through inventives -- its that they don't even try.  They accept the motivations and desires of the person proposing a plan as suffiecient gaurantee that the plan will actually reach those results.

I got a lot of mail last week on my post on incentives.  One loyal reader left me this link at Develish Details, a blog analyzing health care reform proposals.  Incentives are a frequent topic on the blog:

While it's true that paying by procedure creates the incentive to perform more procedures, some of which may be unnecessary, an outcomes based payment system has its own drawbacks. It creates the incentive for doctors to choose to treat patients who are less sick over those who are more sick. Very sick patients require a lot of attention and time, but are less likely to have a bonus-worthy outcome.

On the other hand, less sick patients are easier to treat, are likely to have a better outcome, and will offer a better bonus opportunity for the doctor. Doctors' time is scarce, so they must put it to the best use possible to provide for their families "“ and in a pay-for-outcomes system that means choosing easier to treat patients who will generate the highest bonuses. No matter how much we narrow the arbitrary measure of "outcome", the incentive for the doctor in an outcome based system, where "outcome" is defined by a third party, will always be to select the least sick patients at the expense of the sickest patients most in need of care.

Because human beings are complex organisms, defining what constitutes a "good", bonus-worthy outcome is itself a daunting, if not an altogether impossible undertaking for the third party tasked with producing and evaluating those metrics. Medical outcomes depend on many variables, including, but not restricted to the overall health of the patient (not just the condition being treated) and the patient's compliance with the treatment - a factor over which the doctor has no control.


  1. Zach:

    This sounds remarkably like failing government-run schools. Helping at-risk kids in bad schools isn't easily done (i.e. you can't just throw money at it and expect different outcomes) so it's better to just encourage those kids to drop out so you don't look like a failure.

  2. Michael:

    If you take this and the 5% fee on the top 10% of doctors you recently posted about, it's not much of a leap to think the government is trying to get doctors to not provide medical care.

  3. Daublin:

    You should be more precise. I would guess, and you'd probably agree, that politicians themselves are not only good at understanding incentives, but masters of it.

    It is in the plans they offer to the public, and more precisely their explanations for why those plans should help, that we see a failure to consider realistic motives.

  4. Uncle Bill:

    I'm glad somebody finally touched on incentives. No one else seems to think this is a big issue for government-run healthcare. All the discussions seem to be focusing on costs, reductions in benefits so-called death squads, and so on. But I think that reduced motivation could turn out to be a major problem. And the problems it causes will be so hidden that we will never know.

    Let me illustrate what I am talking about with a real-life example. I used to work with a woman who was born and educated in France, a single mom, a PhD engineer. Our company was taken over by a larger company. For almost two years, we didn't know if we would have a job in the new company, so naturally there was a lot of discussion about interviewing, recruiting, and so on. Everyone was putting out feelers, and contacting their friends in other companies.

    One day my friend said with a sigh, "This would be so much easier in France! There, I wouldn't have to worry about health care, so I could take almost any job, and still get by. Here, I have to work and fight and struggle, to make myself attractive to a good company, with good benefits." (Paraphrasing, of course - I don't remember her exact words, but this was the gist.)

    If health care is provided "free" to everyone, whether you can pay for it yourself or not, this will just remove one more incentive to work hard, to get a good education, to be the best employee you can possibly be, so that you are the first hired and last fired. Lots of folks will still do that, just for the personal satisfaction, but many more will take the easy way out. They will work just enough to get by, with the full knowledge that the government will step in to take care of them if they run into trouble.

    But we will never understand what is happening, or see the "what might have beens," the unseens as Bastiat would say. American productivity will drop. Risk takers will become fewer. Inventions will not occur, and companies will not be started. And we will never understand why things have changed so much.

    When will we learn that incentives matter? This ought to be lesson one or lesson two in Econ 101, but apparently it is not. Our politicians certainly do not understand it.

  5. Dr. T:

    Doctors are smarter than the average person. Unfortunately, they aren't more ethical than the average person. So, no matter what compensation scheme is planned (except for slavery), doctors who want to maximize their revenues while minimizing work hours can always find a way to do so that will result in suboptimal outcomes for patients and payors. Therefore, the simplest and most direct payment system works best (less opportunity for "gaming" the system). But, the federal government is involved in health care financing, and its desired outcomes are greater power and more control. Simple solutions don't provide those outcomes, so federal plans always are complex, over-regulated, bureaucratic, and changeable.

  6. Scott:

    Dr. T, no offense, but your opening statement is complete BS.

    Doctors- and Lawyers - are more educated than the average person and in general (in my experience having worked around, with, and for several) think that they are smarter/cleverer than the average person, but that's it. They're just people. And it has been my experience that their insular education and work leaves them less educated and aware of the day to day realities and information that the average person takes for granted.

  7. tomw:

    The tendency to treat patients with a better 'expected positive outcome' is already abundant in hospitals. There are some that will not take the patient that is not the best candidate for survival as it will negatively affect their statistics. They do NOT have the best staff, as a rule.

    Scott, as a general rule, those that can learn enough and retain enough to pass the state boards are above average in intelligence. Slow learners are weeded out. Those that cannot remember are weeded out. For the most part.
    There are some that have a self-opinion that is far over rated, no question. There are some that don't know the cost of a gallon of milk, and couldn't change the oil in their car, but, they are smarter than the average person.



  8. EnglishTeacher:

    The first sentence makes NO SENSE as it is written. What does "A lot of my education was just a cover story that looked good a number of years of partying with no job look good on my resume without any real improvement in my long-term skills" really mean? Is there some punctuation missing? For example, "A lot of my education was just a cover story that looked good -- a number of years of partying with no job that look good on my resume without any real improvement in my long-term skills"...