results here. They are available upon request.
54
Detailed derivations of the comparative statics are available from the authors upon request.
γ
55
We did not estimate a value for
as that is our key causal variable and in all the illustrations we
either explore its full range or estimate results for
{ 25 5 75 9 1}
γ = . ,. ,. ,. .
0
. We do not consider the
perfectly committed middleman as the case is trivial in this model. In future models where the Boss
wants to achieve a specific level of impact, rather than maximizing his success probability, we will
consider the case where a Boss gets more impact than he wants by hiring a perfectly committed agent,
an agent for whom
γ =
.
56
Our primary interpretation of the Middleman’s discount factor is that it represents how much less the
future is valued than the present. For this interpretation, we can use interest rates to assess how much
an asset needs to return for an agent to forego consuming it today. A more easily identified
interpretation is that the discount factor represents the probability that the game will not end in the next
period due to some exogenous shock; a shock unrelated to the middleman’s potential removal by the
boss due to the failure of an attack. As we will show, the data suggests the same value under either
interpretation.
57
Merari (2004, 11).
58
During the second intifada, Israeli Defense Forces captured Palestinian documents which reported
that one Al Aqsa Martyrs Brigades cell needed approximately 110-150,000 NIS to fund 5-9 operations
per week for what document implies is a 4 week period. Converting this to U.S. dollars at then
exchange rates yields a total cost of approximately $32,000. Dividing this over 28 operations yields a
cost of approximately $1,145 per attack. See "The Involvement of Arafat, PA Senior Officials and
Apparatuses in Terrorism against Israel, Corruption and Crime", Israeli Defense Forces, Ministry of
Parliamentary Affairs, available at: http://mfa.gov.il/mfa/go.asp?MFAH0lom0, accessed 14 March
2005.
59
Lakind and Carmon (1999).
60
According to captured documents amounts from $350 to $800 were paid to specific individuals
involved in organizing and managing attacks.
61
Merari (2004, 11).
62
"I can say today that the central motivation for terrorism in the West Bank is not ideological but
financial. We’re talking about struggles over local control, local gangs, a lot of money that’s injected
from outside the system. There’s no central influence today [meaning, no single authority over the
gangs], not Tanzim, not Hamas and not Jihad [PIJ]. The influence is local and so the infrastructure of
all the organizations is coalesced. No more ’Hamas Nablus’ or ’Jihad Jenin’. Now in Nablus, it’s
Kasbah terrorism, Balata [a refugee camp] terrorism, Kabtya [another part of the city] terrorism. They
work at the neighborhood level, block level, local gang, but they’re still dangerous and can still harm
us." Interview with Ben Caspit, Maariv, 25 May 2004. Translation by Natan Sachs.
63
Of course, it could be that the affect of the Israeli incursions has been to make the operational
environment harder, increasing
α
. However, this interpretation is a reasonable way to ground our
estimates.
64
Al-Zawahiri (2001).
65
Ibid, 6.
66
Ibid, 39.
67
Caroll (2002).
68
The average survival rate for category 4 (finance/logistics) was 81.4% (16.1), for 5
(media/propoganda) was 84.9% 17.3), for 6 (local leaders) it was 82.5% (17.5), for 4&5
(finance/logistics plus media//propaganda) was 82.1% (15.9), and for 4-6 combined it was 82.1%
(16.1).
69
It is only nonzero due to the form of the success probability and the particular parameter estimates in
use.
70
Bueno de Mesquita (2005a).
53
Convention