1. Introduction
As for every species, survival of men is easier in some regions of the world and harder in
others. Figure 1 shows for 137 countries average absolute latitude against the probability for a
child to survive its fifth birthday. In particular tropical regions - defined by an absolute latitude
below 23.5o - provide an unfavorable location for a child to survive whereas survival is almost
certain at latitudes of 40o and higher.1
Figure 1: Absolute Latitude and Child Survival
Child Survival Rate
100%
95%
90%
85%
80%
75%
70%
65% -I-----------------1-----------------1-----------------1-----------------1-----------------1-----------------1
0 102030405060
Latitude
Child Survival Rate=1- Under-5-Mortality Rate, Year 2000, Data from World Bank
(2004) and Masters and McMillan (2001).
Recently, following Acemoglu et al. (2001), a number of studies has found a predominantly
indirect influence of geographic location on income through settler mortality and institution
building. At the same time child survival is to a large extent explained by income per capita
(Pritchett and Summers, 1996), so that the correlation between geography and mortality dis-
played in Figure 1 arises perhaps mainly indirectly. This paper will not deny this hypothesis.
In fact, it will employ an income channel to child survival and argue that additionally geo-
graphic location matters. It will argue that (i) it is ceteris paribus easier for a child to survive
in geographically favorable location (e.g. without malaria prevalence) and (ii) that a unit of
1We focus on child survival rates, which will be the crucial variable in the theoretical model. Similar figures can
be drawn for infant survival and longevity. As Schultz (1999) notes, intercountry differences in life expectancy
are dominated by rates of infant and child survival.