Effect of Comprehensive insurance on
aggregate national savings
are implementing comprehensive medical insurance for their people to reduce
financial strain on low income people. This will lead to the reduction of
spending or out of pocket medical expenses for the households. I want to
explore the effect of this introduction of comprehensive health insurance on
aggregate saving rate. I expect that with
this introduction of health insurance, the households will tend to save less
which might be counterproductive for the national economy and might impact
negatively to national growth.
considerable literature that provides evidence of a negative correlation between
comprehensive health insurance and saving. Kotlikoff (1989) in his paper “On
the Contribution of Economics to the Evaluation and Formation of Social
Insurance Policy” showed with a simulation that, when there is comprehensive
insurance available, the household savings is lowest and when agents must get
their own insurance, savings level is highest.
Shawn Kantor and
Price Fishback (1996) test whether the introduction of social insurance has led
to a reduction in private insurance purchases and precautionary saving by
examining the introduction of workers’ compensation. They find that
the presence of workers’ compensation at least partially crowded out private
accident insurance and led to a substantial reduction in precautionary saving.
Buchmueller and Robert G. Valletta (1999) found a strong negative
effect of husband’s health insurance on wives’ work hours, particularly in
families with children. Chung-Ming Kuan and Chien-Liang Chen
(2003) finds that comprehensive insurance has greater impact on the households
with higher income and those with retiring heads, especially on high savers in
(1996) conducted an empirical finding which showed that US households who have
health insurance saved more than those who doesn’t have coverage, which is a
violation with the standard consumption–saving theory.
comprehensive health insurance is introduced, the households face two savings
decisions or this could affect the households saving decision in two ways:
effect or Precautionary motive and
generally risk averse. When a comprehensive health insurance is introduced in
an economy, households face less uncertainty about their future medical
expenses. So, they can reduce their precautionary saving or specifically, the
portion of the precautionary saving that they thought would need to be spent on
future health issues. But, also there is an income effect. When households have
health insurance, they have more to spend compared to the case when they didn’t
have insurance. And, also, they would have to spend less in case of any medical
emergency. So, their income would also increase and they would have more
disposable income to spend. The standard life-cycle model predicts that
increasing personal income will also increase aggregate savings.
In the end, the
result will depend on the overall strength of the two effects. If the
substitution effect is strong, they will increase savings after the
introduction of comprehensive insurance and decrease savings if the income
effect is stronger.
The actual social
insurance is usually followed by the expectation of such. A government will
announce a policy that will result in the introduction of social insurance
policy, starting from a specified date in the future.
At the point of
the announcement there is likely to be a behavioral effect on the part of
individuals who will be affected by this change.
accumulation equation is: At+1
= At + Yt + trt ? Mt ? Ct
The borrowing constraint: At + Yt + trt
? M? Ct ? 0
Relationship between insurance and
know from the definition of total wealth, with access to health insurance,
household will have more total wealth with having to pay less for health care
in the long run in case of sickness. So, we can write this relationship as:
it can be easily inferred that:
with an increase is consumption, savings will fall.
Proposition for the medium run:
The medium-term effect
of the introduction of social health insurance on the aggregate household
saving rate, is negative when the behavioral effect dominates, and ambiguous
when the combined effect of savings and disposable income dominates. When the savings
and disposable income dominates, the effect on the aggregate household saving
rate depends on the relative magnitudes of the increases in Aggregate savings
“AS” and disposable income, “yit”.
So, to observe an
increase in savings, we need to observe a very strong combined effect of
savings and disposable income
In the long run:
long run, it is expected that the behavioral effect will be spread among the
population or among most of the population. So, this effect will become
stronger eventually and the combined effect will most likely be even smaller
compared with the behavioral effect in the long run. So, it is very likely that
in the long run, savings will fall with the introduction of social health
analysis, we can run a OLS regression, a fixed effect regression and a two
stage least square regression to see the effect of comprehensive health
insurance on household savings.
Our model is:
the household asset is the asset level after a specific period after the
introduction of the insurance plan; is a
dummy variable that denotes if an individual has insurance or not. is a vector of time-invariant characteristics and
Xit is a vector of time-varying observed characteristics (like household size);
ui just represents the unobserved household characteristics; and eit
captures the random shock.
there are some limitations with the OLS model. We might face endogeneity problem in case of
OLS. There can be some level of relationship between some unobservable characteristics
and the decision of a household having health insurance. For example, if a
household is more risk lover, they are less likely to hold a health insurance.
So, we may suffer from endogeneity problem in case of OLS regression.
avoid the problems that we might face in case of OLS, we can use a fixed effect
or a 2SLS model to avoid the endogeneity problem.
a 2SLS model, we will need an instrument to use for insurance status. We can
use “insurance offer for a household” as an instrument. Insurance offer is
correlated with the variable insurance status and it have no correlation with
the determinants of household’s asset holdings.
running a 2SLS model we run the following two stages:
first stage regression:
Then in the
second stage, we can use this predicted value of Iit:
this 2SLS model, we can get unbiased estimate of the impact of insurance
eligibility on household asset or savings by the households.