Information Technology Reference
In-Depth Information
The time evolution for ethical agents is reported in Fig. 1c. Here the initial black
market participation is set to
p
bm
= 0 and there is only a very small probability that
one of the agents becomes non-compliant. Since ethical agents avoid black market par-
ticipation the results are also almost independent of the audit probability. Hence, any
positive audit probability would be inefficient in this case.
Finally, Fig. 1d shows the black market share for d-type agents. Since these agents
act by chance their participation in black market activities would be of the order of
∼
50% without any audit. The different enforcement mechanisms then lead to a fur-
ther reduction of this equilibrium value where similar to the a-types the backaudit
mechanism is most effective.
3
Exploring the Shadow Economy
We now turn to the analysis of the shadow economies in France and Germany within
our model which requires the estimation of the corresponding specific agent composi-
tions. In general, tax experiments provide average data and do not provide individual
compliance data of tax payers. The work by Bazart and Pickhardt (2011) is a notable
exception and allows to extract the percentage of fully compliant individuals (i.e. essen-
tially c-type agents in our terminology). Subject to the small group sizes of 5 subjects
Bazart and Pickhardt (2011) obtain a full compliance ratio of 5% and 20% for Germany
and France, respectively
2
.
Unfortunately no data are available on the percentage of d-type agents. However,
Andreoni et al. (1998) report that about seven percent of U.S. households overpaid
their taxes in 1988. If we anticipate that about the same amount of people underpays
their taxes we arrive at a percentage of
15% of d-type agents [cf. also Hokamp and
Pickhardt (2010)]. For simplicity, this percentage is adopted in equal measure for France
and Germany
3
.
In the following we will denote the parameters derived for France (i.e. 20% c-types
and 15% d-types) as parameter set 'F' and the parameters derived for Germany (i.e. 5%
c-types and 15% d-types) as parameter set 'G'. The aim is then to determine the frac-
tion of a- and b-type agents which participate in the shadow economy in these countries.
Calculations are performed for two different enforcement schemes which both are based
on the probability
p
a
for an audit at a given lattice site (agent). For both, France and
Germany we set
p
a
= 0
.
1. Concerning the backaudit enforcement we consider a back-
audit period of
b
p
= 5 time steps which is compatible with the limitation period of 5
years in 2006
/
2007 where the experiments by Bazart and Pickhardt (2011) have been
conducted.
In order to obtain values on the shadow economies we adopt the values from Buehn
and Schneider (2012) which have estimated the shadow economy of 162 countries.
∼
2
Note that the compliance ratio only specifies the percentage of subjects which behaved fully
compliant in each of the rounds of the experiment. The average compliance rate which deter-
mines tax evasion is larger.
3
One may of course expect that due to different tax pressure this number differs between France,
Germany and the U.S.. Future experiments can help to resolve this element of uncertainty for
the percentage of d-type agents.
