6 / 20
6(20)
An additional question is the consequences on take-up rates when the
individuals do not have full information of the detection probability. If
individuals update their probability of being detected misusing the
insurance when being monitored this would reduce the take-up probability
later on given that there is a deterrence effect.
2
This would result in a
form of equilibrium effect from monitoring. In contrast, if the detection
probability was known the individual’s future take-up of the insurance
would not be affected by being monitored today even if a deterrence effect
exists. Furthermore if there is no deterrence effect then the updating of
probabilities of detection will not affect future take up either.
There is some support of this type of equilibrium effect through the
updating of probabilities of being monitored in the SI literature. Johansson
and Lindahl (2013) find that screening affects future take-up rates of
sickness benefits. de Jong et al., (2011) finds that that the increased
screening in the SI program reduce the applications to the disability
insurance program. This effect is denoted self screening in de Jong et al.
(2011).
This paper studies the existence of an equilibrium effect of monitoring or
self screening by using data from the Swedish temporary parental benefit
program. The program is well suited to test for an equilibrium effect on
insurance programs as the frequency of using the insurance is high. In
addition we have the advantage that the SSIA picks out parents randomly
to be monitored which enables us to estimate the effects of being randomly
assigned to be monitored on future (up to one year after the assignment)
temporary parental leave payments. If individuals do not know the
probability of detection but make inferences about the probability of the
detection by using heuristic rules, e.g. based on their own experience then
they are likely to "overestimate" the probability of detection some time
period after being monitored even when the level of monitoring is constant,
which is the case for the program in the studied period.
The temporary parental benefit compensates for the loss of earnings
when staying at home from work to take care of a sick child (below the
age of 12). In the SI program the employer pays the sickness benefits
the first fourteen days and there is one waiting day before receiving the
compensation, while there is none in the temporary parental benefit
program. Both the employer period and the waiting day gives monetary
incentives of using the insurance for the parents own illness.
3
In order to
claim benefits from the insurance parents simply report that the child is ill
to the Swedish Social Insurance Agency (SSIA). This implies that parents
could claim for benefits for own illness but also as a means of increasing
incomes, that is, they can work and let the child attend day care or school
during the claimed benefit period. The sanctions when misusing the system
are small or not existing and the degree of monitoring or screening is low
(around 5 percent of all claims). It is hence very likely that any finding of
an economic significant equilibrium effect could be carried over to other
programs with less lenient sanctions and to programs with more
monitoring.
2
This prediction builds on the idea of bounded rationality, that is, we humans make
decisions based on probabilities that are estimated using simple heuristic rules (cf.
Kahneman (2003)).
3
Since the cap in the temporary insurance is lower than the cap in the sickness
insurance the incentives differ across individuals. The incentives may be the
opposite for high income individuals who expect to be long term absent.