independance between monetary attributes

[marianne.lefebvre]

Dear all

I am launching data collection to estimate agricultural producers WTP for an innovative insurance.
The monetary attribute is the insurance premium in percentage of the insured capital: 6 levels from 0,1% to 7%
One of the attribute is the coverage in case of losses: 100% corresponds to an insurance with no deductible, 90% with a 10% deductible (dummy coded, the reference level being 90%).
I have 3 other attributes.

I am wondering whether this is correct to evaluate the WTP for a better coverage by dividing the coefficient for the 100% coverage by the coefficient of the insurance premium.
My concern comes from the fact that both coverage and premium impact income. But the impact of the premium on income is certain and actual, while the impact of the coverage rate on income is uncertain and future. In my literature review on WTP for insurance using DCE, I haven't read any concern about that.

If this is a concern, how to calculate the WTP for the other attibutes as well ?
I hope my question is clear.

Thanks a lot for your help
Marianne

[Michiel Bliemer]

An interesting question. How I see it, the insurance premium is a direct cost of the insurance policy that respondents are asked to choose about. Coverage is an attribute of the insurance policy but does not have a real monetary value because it depends on the damage and perhaps there will never be any damage. The only way to monetise the coverage attribute is by providing the respondent with likelihood of damage and the damage amount in the scenario, and then you could express coverage as a monetary benefit. But without providing this information, the only direct monetary attribute is insurance premium and you should use the coefficient of this attribute for computing WTP.

Note that sometimes there exist multiple monetary attributes, for example in transport there exists fuel cost and toll cost and both can be added as separate attributes in the utility function. But then for computing WTP one has to choose one of the costs to serve as the basis for WTP computations. In your case, this choice is simple, it is the insurance premium.

Note that with dummy coded variables, WTP is typically not simply the division of two coefficients. In your case, the division will work but it will mean WTP expressed in $ for a 10% increase in coverage, it will NOT mean the WTP in $ per percentage of coverage. If you have more than 2 levels, then you will need to determine the difference in utility between any two levels and then determine the WTP from one level to the other.

Michiel

[marianne.lefebvre]

Thank you so much for your fast, precise and reinsuring answer (5 days from launching data collection !).
I will take into account your advice when estimating WTP
Best regards,
Marianne