Unbiased Analysis of Today's Healthcare Issues

Empirical Model of Insurance Markets

Written By: Jason Shafrin - Nov• 05•09

Most economists begin to learn about health insurance market with the classic models developed by Rothschild and Stiglitz. In these models, individual risk and characteristics are measured along a single dimension. Insurance contracts pay a lump sum amount in the case of a loss. In the real world of health insurance, however, things are more complicated.

Einav, Findelstein and Levin (2009) review empirical studies to study insurance market in a more sophisticated manner.  For instance, adverse selection is a common problem in insurance markets.  Adverse selection occurs when individuals have private information about their own health that health insurers do not observe.  When insurance companies offer a menu of coverages, high-risk individuals will gravitate towards more generous insurance.

However, adverse selection may be more or less important in reality.  High-risk individuals may be risk loving and thus may prefer less generous insurance contracts; low-risk individuals may also be risk averse and prefer more generous insurance contracts.  Finkelstein and McGarry (2006) find that this occurs in the market for long-term care insurance.  On the other hand, Cohen and Einav (2007) find a positive correlation between risk and risk aversion.

Other real-world insurance complications include:

  • Adverse selection was found to be strong when individuals receive a default level of coverage and incremental coverage is competitively priced.  In general, however, the welfare affects of adverse selection tend to be small in practice.
  • The level of “observable” patient characteristics is endogenous.  Insurers have the ability to conduct more or less scrutiny in underwriting.
  • Moral hazard has different effects in different markets.  For instance, the moral hazard problem for annuities is small (i.e., few people want to kill themselves) while it is likely larger for elective medical care.
  • Most studies examine a set of contracts offered by insurance companies.  It is also important to look at what types of contracts are not offered. In some markets, certain types of coverage are not offered due to concerns about extremely adverse take-up.  Government coverage mandates for mental health and in vitro fertilization coverage may be considered responses to this problem.
  • Insurance mandates eliminate the problem of adverse selection, but create another problem: the limitation of individual choice.  When individual insurance preferences differ, insurance mandates may be welfare destroying.
  • Search frictions impair the efficiency of health insurance markets.  We see wide price variation across insurers even when the products offered are nearly identical.
  • Health insurance markets are not examples of perfect competition.  In most markets, large health insurers act in a oligopoly.

In order to truly understand how health insurance markets operate in reality, it is important to take into account these nuances.


You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.


  1. Steven Hales says:

    Question: In current HCR bills in the House and Senate won’t insurance companies save a certain percentage of their SGA expenses when it is no longer possible under law to take into account the possibility of adverse selection?

    Follow up: If true, then given that health insurers have MBRs in the high 70’s to low 80’s and some of the bills mandate MBRs of at least 85, will the savings be enough to prevent the average MBR from rising beyond profitability?

  2. Guy Thomas says:

    the welfare affects of adverse selection tend to be small in practice.

    Why do economists always think of adverse selection as a cost? It’s a mis-measure of the efficacy of insurance. Correctly conceived, (a modest degree of) adverse selection incrreases loss coverage, which is a benefit. See “Loss ocverage as a public policy objective for risk classification schemes” http://tinyurl.com/cgal3g
    “Demand elasticity, risk classification and loss coverage: when can community rating work?” http://tinyurl.com/da5z7y

Leave a Reply

Your email address will not be published. Required fields are marked *