Chapter 7on others (rather than on the individual concerned), that is,
through channel B (rather than channel C).
To put numerical values on this influence is not easy, and
at this stage we can do no more than give an illustrative
example— hopefully, this will become a major area of re-
search. We take our example from an important study of
the effect of the local crime rate on local life- satisfaction.^9
This was based on the British Household Panel Survey and
included a measure of the quarterly crime rate in the in-
dividual’s local Police Force Area. It found that individual
mental health (0– 10) fell by 0.14 points for each unit in-
crease in the log of the local crime rate. This in turn implies
that each crime reduces the life- satisfaction (0– 10) of the
population by roughly 1 point- year, when the effects are cu-
mulated over the whole local population.^10
Comparing the Two Effects
It would be interesting to compare this external effect of 1
point- year (channel B) with the own effect on the individual
concerned (channel C). On channel C, our data show that
each arrest between 16 and 34 reduces the criminal’s life-
satisfaction (0– 10) at 34 by 0.05 points.^11 Thus, supposing
the effects of an arrest last 20 years, each arrest reduces the
criminal’s cumulated life- satisfaction by roughly 1 point-
year. Since crimes exceed arrests in the ratio 3.6:1, each crime
reduces the criminal’s life- satisfaction by 0.3 point- years.^12
These 0.3 point- years are considerably less than the effect
each crime has on the rest of the population, which was
approximately 1 point- year. This serves to illustrate why the
measured impacts of a policy need to include not only the
well- being of those directly affected but also that of others