Environmental Risks and Disasters
Lecture Notes 2 (1/24/2008)
The risk of dying.
Often when we talk about risk, we are referring to the chance or probability of
dying, or being seriously injured. Risk in this course will often mean just that, but
at other times it will mean the risk of an economic loss. Natural disasters in the
US typically are very costly, but do not cause a great loss of life. This can be
contrasted with disasters in some other countries, where death tolls tend to be
high, but economic losses much smaller. We will return to this difference later,
but we can already see that there is a correlation between higher level of
development and higher monetary losses -- this is quite natural. The reason for
why the loss of human life is so different is not as easily explained in terms of the
wealth of a country and, indeed, some very large death tolls have resulted from,
for example, earthquakes in highly developed countries.
As an example, consider the Kobe earthquake of Japan, which occurred in 1995
and killed around 6,000 people. This earthquake, which was not extremely large,
occurred in a highly developed country that also is very well prepared for
earthquakes. Clearly, we can have high casualties from natural disasters even in
highly developed countries.
As we go on to talk about risks, and in particular risks of dying, it is useful to put
all of this in some kind of perspective. That is, if we are going to talk rationally
about the risk of being killed by a tornado or by drinking Cambridge tap water,
we need to have a method and a context. Presumably we want to be able to
compare risks such as these with all other risks that we face, for example by
crossing Mass. Ave. in Harvard Square, or smoking, or eating in the dining hall.
The risk of dying is, naturally, 100%. That is, we all expect to die, hopefully later
rather than sooner. We all realize that there is some probability that we may die at
any time. Horrible things can happen. However, most people end up living quite
long lives, dying in their 70s, 80s and 90s. One way to see this is simply to look at
the distribution of ages in a population. In stable populations, the number of
people in any age group is essentially the same, with a tapering off above the age
of 70 or so.
It is a very well known observation that if you ask a group of people whether they
believe life is riskier today than some time ago, let's say 20 years, most people
will answer that life is riskier. How does this agree with observations? In order to
test this statement, we need to agree on some kind of measure or metric for risk. If
we agree on interpreting this notion of risk as something related to the probability
of dying, it is not clear exactly what number to choose. Presumably the statement
refers to the population as a whole, and not simply to the individual, so we need a
number that represents the health or "non-dying" of the population.