Call Us: US - +1 845 478 5244 | UK - +44 20 7193 7850 | AUS - +61 2 8005 4826


It is obvious that the net effect of such changes on a will be ambiguous. Richer victims are a more attractive target for attack; on the other hand, they invest more in protection. The net impact of victim wealth on the probability of attack can, therefore, go in either direction. Panel B of figure 1 summarizes this situation. However, the effect of an across-the-board change in group incomes is different. To understand this, one must study the technology of protection or defense, because the cost of deploying that technology will 728 journal of political economy This content downloaded from on Tue, 9 Sep 2014 10:18:07 AM All use subject to JSTOR Terms and Conditions vary with group incomes. Think of two components to protection. The first component is human: protection provided by other individuals in the same community. This is ensured, first and foremost, by living in that community, or at least within easy reach of community members.11 Yet that choice cannot but come at a cost. The principal component of that cost lies in the implicit contract of protection. It may well be the case that compatriots would spontaneously defend a potential victim, but such defense is rarely free: by and large, equal contributions will have to be made to the community or obligations incurred, such as the reciprocal protection of others. But the cost of that reciprocity must be commensurate with the opportunity cost of providing protection services, which is related to the average of group incomes to which our victim belongs. We therefore expect that the cost of “human protection” will be proportional to group incomes. The second component of protection largely involves the use of physical capital: the purchase of security through the use of high walls, barricades, and firearms. This sort of protection is generally extremely effective in reducing attack but involves high fixed costs: the purchase of weaponry ðand the hiring of security guards to use themÞ, the erection of high walls around one’s property, and so on. In contrast to human protection, the cost of this component will be less than proportionately related to group incomes and, to the extent that it is fully reliant on physical capital, not related at all. Specifically, we suppose that