Risk Population and the Environment

1
Risk Population and the Environment

• ARE 298
• David Zilberman

2
On Population and the Environment

• The environment consists of many interacting and
  interdependent population of species.
• It is evolving, yet in the short run it seems to
  be at equilibrium.
• In traditional systems, none of the species was
  dominant enough to suppress the others.
• We have seen that when one of the species
  (humans) has the capacity to dominate others
  (whale, kelp) it does it at its own peril.

3
The Role of Policy

• Natural selection and laissez faire lead to
  environmental sustainability when none of the
  species dominate.
• Government intervention and control of behavior
  are needed to control humans with potentially
  destructive technology.
• Policy intervention can be in terms of
• Incentives
• Direct control
• Education
• Institutions

4
Policy and Risk

• The impacts of policies are uncertain, and the
  environment is subject to stochastic forces.
• Methodologies to both model risk and analyze
  choices under risk are crucial for effective
  policymaking.
• There are alternative approaches to risk.
  Economic and decision theoretic models measure
  risk as deviations from the norm or average.
  They emphasize assessing the impact of such
  deviations on behavior and their cost.
• Public health develops risk assessment techniques
  that define risk explicitly as the probability of
  data outcome.

5
Properties of Risk Assessment Models

• Risk probability that a member of a population
  will die or get sick during a period of time.
• Risk-generating functions relationship between
  risk and processes that cause it.
• The knowledge needed for risk-generation
  functions is interdisciplinary. It provides the
  base for both estimation and policymaking.
• Risk assessment models can be used to assess
• Human health risk
• Environmental health risk (risk to fish)
• Food security

6
Chemical application risk
Pollution control policies
Barriers/filters
Protective clothing
Medical treatment
contamination
exposure
Risk
Transfer fate

Dose/ Response



Risk of chemical residues of can be reduced
by Reducing application levels through taxes,
direct control,etc Blocking movement of residue
to and in bodies of water (can be induced by
incentives) Reduce human exposure by
filters,clothing treatment incase of poisoning
and injury
7
Farm worker risks of chemical

• Let r represent individual health risk where
• r f1(x,B1) f2(B2) f3(B3)
• initial exposure exposure dose/response
• X pollution on site (i.e., the level of
  pesticide use)
• B1 damage control activity at the site (i.e.,
  protective clothing re-entry rules)
• B2 averting behavior of individuals (i.e.,
  washing fruits and vegetables)
• B3 the medical control of pollution dosage.

8
Modeling environmental risk

• The modeling principles used to model human
  health risk from pesticides also apply to
  modeling risk to say birds.
• There are processes of contamination transfer
  and fate exposure and dose response (transfer and
  fate and contamination are most important in this
  context)
• These processes are controlled through policies

9
Policy optimization under risk

• A Precautionary policy making approach-
• The objective is to maximize economic welfare
  subject to the constraint
• Probability ( Risk lt R ) gt ?
• R target level of risk
• ?. safety level (measures the degree of
  social risk aversion)
• ? might represent the degree of confidence we
  have in our risk estimate.
• For example policy makers may aim to maximize
  economic surplus given that risk from pesticides
  can not exceed 1/million with 95 probability.
  

10
Uncertainty and assessment

• Use of higher degree of statistical reliability,
  ????? leads to with risk targets.For example, if
• R risk of death from chemical ABC where
• Acontamination .5 with probability .5
• 1 with probability .5
• Bexposure .1 with probability .5
• .2 with probability .5
• ADose/response .1 with probability .5
• .2 with probability .5

11
Distribution of risk

• Risk is equal to
• .04 with probability .125 (A1,B.2,C.2)
• .02 with probability .375(A1,B.2,C.1, OR
  A1,B.1,C.2 or A.5,B.2,C.2)
• .01 with probability . 375(A1,B.1,C.1, OR
  A.5,B.1,C.2 or A.5,B.2,C.1)
• .005 with probability .125 (A.5,B.1,C.1)
• Risk lt. .03with probability .875
• Risk lt .02 with probability .5
• Risk lt .1with probability .125
• If A.5 with probability 1 risk is
• .02 with probability .25
• .01 with probability .5
• 005 with probability .25

12
Policy and change of risk

• W.P. with probability
• New policy costlier
• If constraint is
• Risklt .2 W.P gt.6 choose II
• Risklt .2 W.P gt.45 choose I
• Risk .1 W.P gt.5 choose I

Risk Probability I original Prob II A.5
.04 .125 0
.02 .375 .25
.01 ,375 .5
.005 .125 .25
13
Sources of variability

• Coefficient of risk generation functions vary
• We may not have a one reliable number
  representing Coefficients of specific processes.
• The risk function may be r AlfabetagammaX
  and the coefficients may be stochastic
• The causes of variability
• Heterogeneity-can be handled by more specific
  analysis
• Randomness
• Uncertainty(lack of knowledge)-can be reduced by
  learning

14
Risk Assessment Example

• Probability of malaria is
• A Measure of area of swamps.
• B Pest intensity
• C Exposure
• Each is a random variable.

15
(No Transcript)
16
Risk Distribution
17
(No Transcript)
18
(No Transcript)
19
There is a significant difference between
average risk and a risk level that is restricted
and occurs 5 of the time (.0005). The mean
risk is .012, but there is less than a 2
probability that the risk is .06, and about a 10
probability that the risk is .03. Suppose that
the pest intensity can be reduced by 50 by
spending a 1 million on chemicals, and the
flooded area may be reduced by two-thirds by
spending 3 million on drainage. If the
objective function is to minimize cost, subject
to the constraint that average risk is below.01,
then it is sufficient to invest 1 million on
chemicals, and average risk will become .006
.012/2.
20
If the objective is to minimize cost subject
to the constraint there is less than .05
probability that risk is below .01, then it is
optimal to spend 4 million on pest control and
drainage. In this case risk levels will be
divided by 1/6 and maximum risk is .01 and will
occur with probability of 1/64. Suppose
1,000,000 people are exposed to these risks. The
risk distribution can be translated to the
distribution of the malaria cases. With
probability 1/64, without treatment, we will have
60,000 malaria cases annually, and the average
number of cases is 12,000.

21
If policymakers aim to minimize treatment
costs expected health costs, where expected
health costs mean risk X cost of one sick
person-V IF the area is sprayed the expected
health cost 10000006000 V The optimal policy
is
Do nothing if V ( cost of sick)
.
Apply pesticides if V is
.
Drain and apply pesticides if V is
.
22
Value of statistical life