7 / 20
7(20)
We find that individuals assigned to be monitored on average reduce
their later take up rate. We interpret the result as individuals after being
monitored adjust their assessment of the probability of monitoring which
then affects later take-up. The empirical design does not allow us to
estimate the magnitude of the deterrence effect from monitoring. It is
however possible to show that the present level of monitoring save cost
for the government just because it changes the perceptions of the level
of monitoring within the program.
Methodologically the paper by Kleven et al. (2011) is close to our study. In
their study effects on future tax evasion from being assigned to be audited
the year before is estimated. They find evidence of substantial effects on
self reported income the year after. Another highly relevant study is
Engström et al. (2007). Based on a randomised design they sent out
“threat of monitoring letters” to parents eligible for the Swedish temporary
parental benefit.
4
The results show that the parents who received the
“threat of monitoring” letter decreased their use of the benefit by 13 per
cent.
5
This paper, thus, shows that individuals respond on information of
(increased) monitoring. However, as it is not custom for authorities to
inform individuals on future monitoring this may not be an estimate of an
ex ante or deterrence effect and the paper have been criticised in Sweden
for overstating the effect of monitoring. Our results however seems to be
in accordance with theirs.
The rest of the paper is arranged as follows. Section 2 describes the
Swedish temporary parental benefit. Section 3 describes the data. The
analysis is given in section 4 and section 5, finally, concludes.
4
These letters are similar to “threat of audit” letters in the tax evasion literature (cf.,
Coleman (1996), Slemerod et al (2001) and Hasseldine et al. (2007) and Kleven et
al. (2011).
5
Persson (2011) shows that the increased monitoring also had spill-over effects on
parent’s own sickness absence. The parents who were informed about the
increased monitoring were more likely to be absent from work due to own sickness.
The increase in the parent’s own sickness absence corresponded to 43 per cent of
the decrease in the take-up of temporary parental benefit.