At last night’s press conference the President was exactly right when he said:

If we do not reform health care, your premiums and out-of-pocket costs will continue to skyrocket. …

Right now premiums for families that have health insurance have doubled over the last 10 years. They’ve gone up three times faster than wages. So what we know is that if the current trends continue, more and more families are going to lose health care, more and more families are going to be in a position where they keep their health care but it takes a bigger bite out of their budget, employers are going to put more and more of the costs on the employees or they’re just going to stop providing health care altogether. …

One of the things that doesn’t get talked about is the fact that when premiums are going up and the costs to employers are going up, that’s money that could be going into people’s wages and incomes. And over the last decade we basically saw middle-class families, their incomes and wages flatlined. Part of the reason is because health care costs are gobbling that up.

Most of the public and Congressional debate has been about the effect of pending health care reform on the federal budget. While this is incredibly important, it may be less important than the effect of this legislation on private health insurance premiums.

As a reminder, I agree with the President’s core problem definition. Rising per capita health care spending leads to (1) slower wage growth for those with private health insurance, as premiums eat up compensation growth; (2) an increasing number of uninsured who can’t afford the higher premiums; and (3) unsustainable spending trends for state and local governments.

I wrote in mid-May about why health spending continues to grow at an unsustainable rate. Sixty-two to 75 percent of long-term health cost growth is due to the higher costs of improved technology and the increased prevalence of third-party payment. Health care keeps getting more expensive primarily because we use more and better health care each year, and because most of it appears to be paid by someone else. We have to address these sources of cost growth to have any hope of solving the underlying problem.

The President once again correctly identified the core problem, and I compliment him for his emphasis on the need to reduce, or at least slow the growth of, private health insurance premiums. Unfortunately the House “Tri-Committee” bill and the Senate HELP Committee bill move in the opposite direction. They would cause private health insurance premiums to go up, increasing the crunch on wages and the difficulty the uninsured have in affording insurance. The House Tri-Committee bill and the Senate HELP Committee bill contain insurance mandates that would make private health insurance more expensive for most Americans, and would thus exacerbate the problems described by the President.

Two different kinds of mandates
These bills contain two fundamentally different kinds of mandates:

  • “You must” mandates: Individuals and families are required to buy health insurance, and employers (over a certain size) are required to offer health insurance to their employees. Anyone not complying with such a mandate must pay a new tax.
  • “Insurers may not” mandates: Insurers may not sell policies that do not cover pre-existing conditions, or that charge very different premiums to people with different health profiles based on age, gender, or health status. These are usually described as mandates that “insurers must sell policies that do X and Y and Z,” but economically it’s actually a prohibition on selling policies that do not contain X or Y or Z. The government is not requiring companies to sell insurance, but instead prohibiting them from selling insurance unless it meets certain conditions.

Today I want to focus on the latter type of mandates, which have received little attention in the recent legislative debate. Here are four versions of an “insurer may not” mandate:

  1. Mandated benefits – A health insurance plan may not deny reimbursement for mammograms for women meeting certain medical criteria, or for a 48-hour hospital stay after the birth of a baby. Slightly less politically attractive would be a mandate that health insurance plans may not deny reimbursement for chiropractic benefits or substance abuse.
  2. Community rating – A health insurance plan may not charge different prices to different customers.
  3. Guaranteed issue – A health plan may not deny coverage to any individual who applies, regardless of whether they have a pre-existing condition. Guaranteed issue would allow, for instance, a cancer patient to newly enroll in a health insurance plan to get reimbursement for medical treatment for his cancer.
  4. Any willing provider – A health insurance plan may not exclude particular hospitals or doctors from their network.

Benefit mandates are fairly straightforward. The big question is, why should the government be deciding which medical treatments your health insurance must cover? Does it make sense for the government to mandate specific medical care practices? You can tell from my framing of the question that I’m a “no” on benefit mandates. I don’t think that’s an appropriate role for government.

Each benefit mandate raises the price of health insurance a little bit. These increases accumulate.

Community rating and guaranteed issue usually provoke the most active debate because of their enormous distributional effects. They act as cross-subsidies from the healthy to the sick, or more precisely from those more likely to be healthy to those more likely to have higher health costs.

Here is a crash course in community rating and guaranteed issue. I will oversimplify to make it useful.

  • Health care costs tend to be highly concentrated. Most people are fairly healthy and have low health costs. They occasionally get sick or injured, but on average they’re healthy. In a similar way, most houses don’t catch fire each year, but you buy homeowner’s insurance to protect against the small chance that yours will.
  • Most of the health care spending is concentrated in a minority of the population who are frequently or permanently sick or injured. These people have predictably high health costs and therefore cost more to insure in a market without government distortion. Continuing the home/fire example, a house built next to an outdoor flamethrower testing facility will catch fire more often and have predictably higher costs. We would expect an insurer to charge a higher premium for such a home. And a house that is on fire has no risk – the costs/losses are certain. The minority of homes built near the flamethrower testing facility, and those that are already on fire, would account for a large majority of the total costs/losses.
  • If a policy equalizes premiums between the usually healthy and the predictably sick, health insurance will become somewhat more expensive for most (usually healthy) people, and much less expensive for the minority who are predictably high cost. In some cases, a market without distortion won’t even sell insurance to someone who is predictably sick, just as an insurer won’t sell you insurance while your house is on fire. Assuming we want to help the person with cancer, if we do so by requiring companies to sell him insurance and to charge him the same premium as a healthy person, then the cancer victim can buy affordable insurance, cross-subsidized by a large number of relatively healthy people who will pay higher premiums.
  • This is what community rating and guaranteed issue do. They make insurance available and much more affordable for those with predictably high health costs (e.g., someone with incurable cancer, or a sixty-year old man with a family history of heart disease), while raising premiums for most enrollees who are on average relatively healthy.
  • It gets even trickier because some people who are predictably sick and need expensive medical care have inherited or random illnesses that are completely outside of their control, while others use a lot of medical care in part because of their behavior. While both types are high-cost and predictably ill, some people and policymakers come to very different judgments about whether the two cases should be treated the same. If you think someone with a family history of cancer should not be charged a higher premium than someone without, are you comfortable saying that insurance companies should charge smokers and non-smokers the same premiums? Should non-smokers subsidize the premiums of smokers? Does your answer change if the smoker already has lung cancer?

Because the House and Senate bills contain versions of community rating and guaranteed issue mandates, they would benefit those with predictably high costs. At the same time, the President wants to slow the growth of overall (total/average) private health insurance premiums, and we need to understand how much the mandates in these bills would exacerbate that problem.

We have a lot of data on the effects of these types of insurance mandates, because there are today 50 state insurance markets with widely varying requirements. This serves as a natural experiment and allows health economists to tease out whether particular mandates are associated with, and probably cause, higher insurance premiums.

We had three economists from the Council of Economic Advisers study this (internally for the Bush White House) in 2004. They are Mark Showalter, William Congdon, and Amanda Kowalski. They turned their memo into a published paper, which you can access for free here, but only if you’re in an academic institution. For the rest of us, here is the original CEA memo from which this paper was derived.

Their memo analyzed the effect of State insurance mandates on the price of health insurance policies in the non-group (individual) market. Here are three conclusions relevant to the current debate:



  • “Mandated benefits raise the expected price of an individual policy by approximately 0.4 percent per mandate.” For family policies the increase is approximately 0.5 percent per mandate. The typical state has about 20 mandates (with a range from 6 to 48) so a reduction from 20 to 10 mandates would imply a 4 percent decrease in price for individual policies, and a 5 percent decrease for family policies.
  • “Community Rating” laws, which limit insurers’ ability to charge different prices to different customers, raise prices by 20.3 percent for individual policies and 27.3 percent for family policies.”
  • “The difference in price for guaranteed issue laws is $113
    [per month for an individual] (233 – 120), but only a single state in our sample has such a law (New Jersey).”
  • The House and Senate bills both require guaranteed issue, and they both require versions of “modified” community rating, with the specifics to be determined by the States. While neither bill creates new specific federal benefit mandates, they both create a government-appointed board with the ability to create such mandates. Thus the mandated benefits effect is there but indirect.

    Anticipating some of the pushback:

    • Yes, their regressions looked at the effects of State mandates, rather than national mandates, and on the individual market rather than the employer-based group market.
    • It is hard to tell how national mandates would interact with the new exchanges for individual (non-group) purchase. While I will guess that the effects would be similar to those found in the CEA study, that’s just my guess.
    • I am not aware of “any willing provider” mandates in the House or Senate bills, so those are not directly relevant.

    I am not arguing that the numbers from the CEA memo directly translate into the same quantitative effects for either the House “Tri-Committee” bill, nor for the Senate HELP Committee bill. But the numbers are large enough that Congress needs to ask these same questions about the bills they are now considering.

    You may think that community rating and guaranteed issue, which often go together, are fairer than allowing insurers to base premiums on expected risk. You may instead think that health insurance should be like homeowners or auto insurance, in which people with similar risks pool their resources and get charged similar premiums, and those with higher expected risks face higher premiums. This debate is value-driven and often quite intense, and I am not trying to resolve it here.

    I am trying to draw your attention to a more basic analytic point. By including a guaranteed issue mandate, a mandate for modified community rating, and the ability for a new government-appointed body to create new benefit mandates, the House and Senate bills will cause total and average private health insurance premiums to increase.

    How much? I cannot say precisely, because of the differences between CEA studied State mandates and because there are other interactive effects in this bill. But clearly these mandates will increase premiums, and if the numbers are comparable, the neighborhood is quite expensive: +4-5% higher premiums for another 10 benefit mandates, +20-27% for community rating, and New Jersey’s guaranteed issue is associated with 94% higher premiums compared to a similar State without guaranteed issue. Those are potentially astronomical premium increases that would make the problems the President describes far worse than under current law.

    Debates and decisions about the equity effects of guaranteed issue and community rating mandates are why we elect Members of Congress. They will affect the predictably sick and the relatively healthy. These debates need to occur before Members vote on these bills. And Members need to understand the effects these mandates would have on private health insurance premiums. It would come as a harsh surprise if these bills became law and premiums for most Americans suddenly jumped 20%, 27%, or 94%.

    Before voting on these bills Congress needs to ask both CBO and the HHS actuaries two simple questions:

    1. What would be the effects of the guaranteed issue and community rating mandates on average private health insurance premiums, in both the group and non-group markets?
    2. What would be the effects on average private health insurance premiums if the new government-appointed body were to add N more benefit mandates?

    The answers to these questions are at least as important as the effect of these bills on the federal budget.