Transcript Willingness to Pay for Low Probability, Low Loss Hazard

Willingness to Pay for Low
Probability, Low Loss Hazard
Insurance
John C. Whitehead
ECO 4810. Senior Seminar
Department of Economics
Appalachian State University
March 19, 2007
1. Motive your research
• From your introduction, provide a slide that
prompts you to tell a short story about the
motivation of your paper
• The goal is to get your readers interested
in what you are doing
Evacuation Cost Insurance
-- Ganderton
et al. 2000, page
272.
“Analysis of
behavior and policy
prescriptions would
not be such a
problem if low
probability natural
disasters had small
consequences.”
2. Describe the previous literature
• Describe what others have found
• Describe how what you are doing is
different
• Be brief
Literature
• McClelland, Schulze
and Coursey (1993)
• Ganderton et al.
(2000)
• Ozdemir (2005)
• Burrus, Dumas and
Graham (2005)
3. Theory
• Present a graph (or equation) that allows
you to explain the theoretical framework
for your analysis
Price
Theory
Supply
Demand (r’)
Demand (r)
Quantity
4. Describe your data source
• Where did you get the data?
• How did you manipulate it?
• What is the sample size?
Data
•March 2001 telephone
survey
•Response rate = 73%
(n = 411)
•Most respondents had
experienced one or
more hurricanes (n =
384)
•50% evacuated for at
least one storm
5. Empirical Model
• Present the equations so that you are
prompted to (1) describe the variables, (2)
introduce the empirical model
Empirical Model: OLS
y   0  1rˆ   2c   3 y
rˆ    j
j
S
j
6. Describe the data/variables
• Devote at least one slide to your
dependent variable: (1) description, (2)
summary statistics
• Devote at least one slide to your primary
independent variable: (1) description, (2)
summary statistics, (3) the rest of your
control variables can be simply listed
Dependent Variable
Many people in coastal areas are
safer if they evacuate their home
before a hurricane strikes. But
evacuating can cost a lot of time,
money and lost income. Some
people would rather stay at home
during a hurricane so that they
don’t spend money and lose
income. To reduce this problem
and increase safety, one plan is to
have a hurricane evacuation
insurance policy to reimburse
people for the costs of evacuation.
Willingness to Pay
• Suppose you could purchase a hurricane
evacuation insurance policy from an insurance
agent before the next hurricane season. The
price of the evacuation insurance is $p each
hurricane season. With insurance, you would be
reimbursed for all of your evacuation expenses
throughout the hurricane season. This would
include reimbursement for travel, lodging, food,
medical costs and lost income. Do you think you
would purchase the hurricane evacuation
insurance policy?
% Willing to Pay
45
40
35
30
25
20
15
10
5
0
$50
$100
$140
$190
$240
Independent Variables
•
•
•
•
Policy price = p (mean = $25)
Evacuation probability = r (mean = .05 )
Evacuation cost = c (mean = $147)
Other factors affecting evacuation
6. Empirical Results
• Describe your primary results. Those most
interested can get the details from your
paper (and remember your goal is to get
members of your audience to want to read
your paper).
• Describe your results qualitatively (see
next slide) or quantitatively (see slide after
that) but not both
Willingness to Pay Models
•
•
•
•
•
•
F-stat is statistically significant
R2 = .11
B(price) < 0 (model 1 only)
B(probability) > 0
B(cost) > 0
B(income) > 0 (model 2 only)
Willingness to Pay Models
Model 1
Model 2
Constant
ln Price
Probability
Coeff.
1.46
-0.99
5.93
t-ratio
0.86
-2.92
1.71
Coeff.
1.66
-1.20
5.29
t-ratio
0.82
-2.84
1.39
Cost
Income
0.003
0.005
2.81
0.93
0.002
0.011
1.78
1.68
F
R2
Cases
22.20
0.11
193
15.94
0.12
193
7. Conclusions
• What are the (policy) implications of your
research?
• Interpret the magnitude of the coefficients
using (1) marginal analysis, (2) elasticity,
(3) consumer/producer surplus
• Don’t restate your results
Summary
• Each 10% increase in the probability of
evacuation increases WTP by $7
• Demand elasticity = 0.64
• WTP = $34 (less than the lowest price of
the insurance product)
8. Future Research
• What arose during the course of the
project that you’d like to fix: (1) better data,
(2) missing variables, (3) a different
empirical model
Future Research
• An unresolved
puzzle: Moral
hazard
• Missing variables:
Subjective vs
objective risk
• Appropriate model:
logitistic regression
vs OLS