8.7 Containing Crime
159
normal behavior are contagious, too. This can explain the surprising crime cycles that
have often been observed in the past. As a consequence, a crime prevention strategy
based on alleviating socio-economic deprivation might be much more effective than
one based on deterrence.
8.8 Group Selection
Things become trickier if we have several groups with different preferences, due
to different cultural backgrounds or education, for example.
14 Then, a competitive
dynamics between different groups may set in. The subject is often referred to as
“group selection”, which was put forward by Vero Copner Wynne-Edwards (1906–
1997) and others.
15
It is often believed that group selection promotes cooperation. Compare two
groups with high and no cooperation. Then, the cooperative group will get higher
payoffs and grow more quickly than the non-cooperative group. Consequently, cooperation should spread and free-riding should eventually disappear. But what would
happen if there were an exchange of people between the groups? Then, free-riders
could exploit the cooperative group and quickly undermine the cooperation within
it. For such reasons, people often fear that migration would undermine cooperation.
But it doesn’t have to be like this.
8.9 The Surprising Role of Success-Driven Migration
To study the effect of migration on cooperation, back in 2008/09, Wenjian Yu and I
developed a related agent-based computer model.
16 We made the following assumptions: (1) Computer-simulated individuals (called “agents”) move to the most favorable location within a certain radius of their current location (“success-driven migration”). (2) They tend to imitate the behavior of the most successful individuals they
interact with (their “neighbors”). (3) There is some degree of “trial-and-error behavior”, i.e. a certain probability that each individual will either migrate to a free location,
which isn’t occupied already, or change the behavior (from cooperative to noncooperative or vice versa). While rule (1) does not change the number of people who
cooperate, the other two rules tend to undermine significant levels of cooperation.
Surprisingly, however, when all three rules are applied together, a high level of
cooperation is eventually achieved. This proved to be true even in cases where the
simulation began with no cooperative agents at all, a situation akin to the “state of
14 Winter et al. [8].
15 Wynne-Edwards [9].
16 Helbing and Yu [10], Helbing and Yu [11].
159
normal behavior are contagious, too. This can explain the surprising crime cycles that
have often been observed in the past. As a consequence, a crime prevention strategy
based on alleviating socio-economic deprivation might be much more effective than
one based on deterrence.
8.8 Group Selection
Things become trickier if we have several groups with different preferences, due
to different cultural backgrounds or education, for example.
14 Then, a competitive
dynamics between different groups may set in. The subject is often referred to as
“group selection”, which was put forward by Vero Copner Wynne-Edwards (1906–
1997) and others.
15
It is often believed that group selection promotes cooperation. Compare two
groups with high and no cooperation. Then, the cooperative group will get higher
payoffs and grow more quickly than the non-cooperative group. Consequently, cooperation should spread and free-riding should eventually disappear. But what would
happen if there were an exchange of people between the groups? Then, free-riders
could exploit the cooperative group and quickly undermine the cooperation within
it. For such reasons, people often fear that migration would undermine cooperation.
But it doesn’t have to be like this.
8.9 The Surprising Role of Success-Driven Migration
To study the effect of migration on cooperation, back in 2008/09, Wenjian Yu and I
developed a related agent-based computer model.
16 We made the following assumptions: (1) Computer-simulated individuals (called “agents”) move to the most favorable location within a certain radius of their current location (“success-driven migration”). (2) They tend to imitate the behavior of the most successful individuals they
interact with (their “neighbors”). (3) There is some degree of “trial-and-error behavior”, i.e. a certain probability that each individual will either migrate to a free location,
which isn’t occupied already, or change the behavior (from cooperative to noncooperative or vice versa). While rule (1) does not change the number of people who
cooperate, the other two rules tend to undermine significant levels of cooperation.
Surprisingly, however, when all three rules are applied together, a high level of
cooperation is eventually achieved. This proved to be true even in cases where the
simulation began with no cooperative agents at all, a situation akin to the “state of
14 Winter et al. [8].
15 Wynne-Edwards [9].
16 Helbing and Yu [10], Helbing and Yu [11].