The organization of health insurance markets

- Loading cost and the behavior of insurance firms

Comsumers can improve their well-being sacrificing (relatively) small, certain premium to insure
againts the probability of a considerably larger loss. It is important how to demonstrate how the
policies will be offered to specific group and why, in fact, some groups will find it difficult to get
insurance at all.

We have referred to the model of a competitive industry, in which the firms will compete to where
profits become zero, or normal. It profits are higher (lower), firms will enter (leave) the market. Only
when profits have become zero, or normal, will entry and exit cease. In this model, the insurance
carries would collect money during the year and pay someof it out. In good years (for the carries),
carries would pay out less than collected; in bad years, they would pay out come. Economic analysis
suggests that the good (and bad) years should be random. It firms have systematically good (bad)
years, it suggests excess profits (losses(, and the probability of entry into (exit from) the industry by
other firms.

We also previously shown how moral hazard can lead firms to offer certain types of coverage and
not others. In particular, firms would shy away from covering conditions that are accompanied by
price-elastic demands for service.

- Impact of loading cost

Firms have additional costs of doing business, often referred to as loading costs. The costs are largely
related to the number and types of costumers and claims that they process. As a result, even in
perfect competition, this costs must be passed on to costumers or else the firms will not be able to
cover all costs and will be forced to leave the market. The incidence of these costs suggests that
firms will shy away from covering events that are almost certain to occur, or those that seldom
occur.

Consider consumers who behave as though they have a utility of wealth function exhibiting
diminishing marginal utility of wealth. Consider figure 11-1, in which the amount the costumer
would be willing topay over the actuarially fair amountis shown by the horizontal distance between
the expected utility line and the (curved) utility fuction, measured in dollars. For example, at point F,
this horizontal distance is FG. Clearly, the horizontal distance between the expected utility line and
the utility fuction is zero if the event never occurs (i.e., if we are at point A). It increase up to some
points as we move in southwest direction, with increased probability of illness, and the decreases, as
the illness becomes more certain, toward point B.
This means that the extra premium people are willing to pay avoid risk is smallest either for rare
events or for almost certain events.

Because insurance is taken against risk, as the probability of the uncertain event approaches either 0
or 1, insurance becomes less desirable. Near point A, the expected loss –that is, the probability ot
the event –multiplied by the loss if the event occurs is not large enough for the costumer to bother
to insure. Near point B –because the event is almost certain –the costumer might as well set the
money aside, and avoid the trouble of dealing with the insurer. Most people do not seek insurance
against losses due to common colds.

Figure 11-1 addresses this issues graphically. The top portion of Figure 11-1 shows the utility
function we have just discussed. Moving southwest from the initial wealth (point A), we see that the
gain from insurance increases initially, then eventually falls back to zero when the probability of
illness becomes 1. Below this graph, we plot the probability of the event on the x-axis and the
expected gain on the y-axis. The lower x-axis is best understood as the probability of not filing a
claim, which decrase from 1 (under point A where the event never occurs) to 0 (under point B where
the event always occurs). Thus, at p=0 (at A where there is no loss) aor at p=1 (at B where a loss is
certain), the expected gain from insurance is zero. Because the marginal benefits are positive in
between, the marginal benefits curve has a hump in he middle.

- Insurance for heart attacks and hangnails

In comparing types of losses, at any value of p, the larger the expected loss, the larger the gain from
the insurance. This is easily seen by cmparing the distances between the expected utility line and
the utility curve for a small loss (line segment EA) and for a large loss (line segment BA). Segment EA
shows asmall distance;segment BA, a large one. Hence, if the costumer has equal probabilities of a
hangnail (small loss) and a heart attack (large loss), the expected gain from the insurance for major
(heart attack) coverage will be alrger than for a minor (hangnail) coverage.

Given the marginal benefits to the costumers, consider now the firms decisions in providing
insurance. Clearly, if the event is almost certain, the costs of administering the policy may exceed
the benefits. In other words, the costs of administering claims are positive. For simplicity, we have
drawn these costs as constant marginal costs –a horizontal line.

From probability 0 to point C, it will not pay to insure claims because the marginal costs axceed the
expected marginal benefits. Between point C and D, expected marginal benefits exceed marginal
costs. To the right of point D, again the marginal costs exceed the expected marginal benefits, and
no insurance will be provided. As the diagram is drawn, no firm could afford to offer hangnail
coverage.

- Loading costs and the uninsured

The analysis provides one avenue for addressing the problem uninsured. Health insurance in the
United States is largely available through participation in the labor market. Those who do not
participate in the labor market, and indeed those employed by small businesses, self-employed, or
sporadically employed, may find it difficult to get insurance.
Although may explanations have been proposed, it is apparent that per-person costs or processing
information and the claims of those individuals who are outside larger organizations (either
companies or unions) may be higher. This would result in an increase in the marginal costs relative
to the costumer’s marginal benefits and would reduce or eliminate the range of services that may
be offered.

The analysis also helps address the impacts of the entry and exit in the insurance market. More
efficient processing and information handling presumably will lower the premiums that must be
paid by the costumers in the market. If we gain address Figure 11-1, we recognize that improved
information handling and processing would not only lead to lower prices, but also would permit
firms to offer services (base on probability of occurrence) taht had not previously been offered.

Consider point C or D, where the expected marginal benefit was previously just equal to (or possibly
just below), the marginal cost. The ability of a firm to ower costs would allow that firm to offer
insurance for types of events that previously were uncovered. Conversely, increased costs, either
due to market forces or often due to mandated types of coverage, would force firms to cut back
coverage on events for which the limited amount od consumer surplus would not allow the firms to
pass the increased costs to the costumers.

EMPLOYER PROVISION OF HEALTH INSURANCE WHO PAYS?

Unike most other types of insurance, for the largest segment of the American population helath
insurance is provided through the workplace. The reason this began was almost an accident. During
world war II, because of the booming economy and coupled with wartime shortages in costumer
goods, not much was available to buy, and wage and price controls were imposed as anti-
inflationary devices. Predictably, compensation needed to be changed in order to attract workers,
because wage controls in a full-employment aconomy would prevent companies from raiding each
other for workers. Fringe benefits, which were not legally consideredas part of the wage package,
were used to provide flexibility in worker compensation, and hence to improved the allocation of
workers among sectors of the economy. One of these fringe benefits was health insurance.

Following Lee (1996), consider amlabor market with a tipically downward-sloping demand for labor,
D, and a tipically upward-sloping supply of labor, S, as noted in Figure 11-2. The demand for labor is
related to the marginal productivity of workers. The supply of workers is related to the wage in this
industry relative to the wage in other industries. Workers will choose to work in this industry as long
as the wage they can earn exceeds the opportunities in other jobs. In figure 11-2, at equilibrium
wage is W1 and the equilibrium quantity of labor demanded and supplied is L1.

Now suppose that workers in the market negotiate a health insurance benefit worth $z per hour at
that margin, and it costs employers exactly $z per hour to provide. What happens? Employers who
were previously willing to pay W1 per hour for workers will now pay W1 less $z. Other points on the
demand curve will shift downward in a similar manner, so the demand curve will shift downward by
excatly $z to D’.

What will happen to the supplu curve? Because the workers were willing to supply various amounts
of labor at various wage rates according to the supply curve before, now that they are receiving a
benefit worth $z they will offer their labor for $z less. Hence, the supply curve will shift downward
by exactly $z to S’. The workers, of course, would like to continue to earn W1 plus the health
benefits, but if they were satisfied with W1 in the absence of health benefits, they would be just as
satisfied to earn W1 less z, if the value the benefits at $z per hour.

What is the result? As with the previous discussion, the net wage remains the same, but the money
wage falls by $z. The equilibrium wage has fallen to W2 at point b’, or by exactly the amount og the
benefit. Workers have been taken their benefit in lower money wages, and the same number of
workers, L1, is employed at the same net wage. For a “real world” example of who pays.

SPOUSAL COVERAGE : WHO PAYS?

An analysis derived from Mark Pauly (1997) helps to examine the issue of spousal insurance
coverage.

HOW THE TAX SYSTEM INFLUENCES HEALTH INSURANCE DEMAND

One of the most important factors in the increased demand for health insurance in the post –World
War II era has been the tax treatment of health insurance.

With increased marginal tax rates, consumers have incentives to increase employer health
expenditures. Employers also benefit from this arrangement because their level of Social Security
taxes will fall. Currently, both employes and employer pay 6.20 percent for the Social Security
portion and 1.45 percent for the Medicare portion. Because insurance is an expense to the
employer rather than a factor payment (on which Social Security and Medicare taxes must be paid),
it is exempt from Social Security and Medicare taxes. In the 1950s, federal marginal rates went as
high as 91 percent, and even today many people pay marginal (federal plus states) tax rates well
over 30 percent.

WHO PAYS –EMPIRICAL TESTS

The trade-off between wages and insurance is important enough to merit serious empirical
invertigation. Morrisey (2001) notes that most empirical studies thus far associate health insurance
with higher, rather than lower, wages. Because compensation (wages plus insurance) is based on
productivity, employers spend considerable effort identifying workers who are better motivated,
more dependable, more highly skilled, and better able to interact with clients and costumers.

Gruber and Krueger (1992) examine workers’ compensation insurance, and Gruber (1994) looks at
mandated maternity benefits coverage.

Studies testing the impact for the individuals have been more tentative. In a model that examines
health insurance, pensions, and vacation time as components of the compensation package, Jensen
and Morrisey (2001) argue that wage and benefits are stimulaneously determined.
OTHER IMPACT OF EMPLOYER PROVISION OF HEALTH INSURANCE

Employer provision of health insurance has other impacts, as well. Because the employer is a large,
single buyer of coverage, the purchase of insurance through the employer provides scale economies
of dealing with insurance providers taht single purchasers could never enjoy. This tends to lower
the effective price of coverage to the employees.

It can be argued that the purchase of insurance by employers minimize adverse selection by
providing a more appropiate pool fir the fixing of insurance rates. This is because most groups
contain a broad mix of risks, by virtue of having been formed for some purpose other than
insurance.

EMPLOYER BASED HEALTH INSURANCE AND LABOR SUPPLY

- Health Insurance and Retirement

While declining health makes retirement more attractive, it also makes amployer-provided
insurance more attractive, especially for those younger than 65 years of age, at which time
Medicare would provide insurance for a large portion of the care. Thus individuals face an
incentive to postpone retirement until they are eligible for Medicare at age 65.
Researchers have generally focused on the impact of retiree health insurance on retirement
behavior. Gruber and Madrian summarize 16 studies, and report that the availibility of retiree
helath insurance raises the odds of retirement by between 30 and 80 percent.

- Health Insurance and Mobility

Health insurance may also affect worker mobility between jobs. Employees may fear losing
coverage for pre-existing conditions, which are generally defined as any medical problem that
has been trated or diagnosed within the past six months to two years.

Madrian uses a simple matrix of the probability of job mobility to consider the impact of job lock.
Because job lock is caused by the potential loss of health insurance coverrage with changing jobs,
one would not expect those with coverage through both their own employment and an outside job
to face job lock.

Employer-Provided Health
Insurance
No Yes
No other health insurance a b
Other health insurance C D

Madrian tests for the magnitude of job lock by examining whether those workers with employer-
provided health insurance and other coverage are more likely to change jobs than those without
alternative coverage, or:

(cell d) – (cell b) > 0

However, if a man is in cell d, it may be due to insurance provided by his wife, who may be providing
income as well; all else being equal, the additional income could lead to increased mobility. Hence,
Madrian derives a secont test: whether having other health insurance increaces mobility more for
those who have employement-based insurance than those who do not, or:

d–b>c–a

which can be written as :

(d – b) – (c – a) > 0

This test is referred to as difference-in-difference, the difference between ( d – b ) and (c – a ).

These result are consistent with anecdotal evidence of job lock as an unintended consequence of the
system of employement-related coverage in the Unated States. It could be addressed through
changes that are broadly consistent with prudent insurance practices. These include elimination of
pre-exsiting condition clauses and the development of health insurance pooling mechanism in local
labor markets that might promote continuity of coverage across employers. Job lock also must be
considered as a consequence of mandated approaches to health insurance benefits, both at state
and national levels.

Employer-Provided Health
Insurance
No Yes
No other health insurance 0.256 0.085
Other health insurance 0.244 0.115
Estimates of job lock
(d – b) 26.1% (13.8%)
(d – b) – (c – a) 31.1% (17.7%)