Hello !
I am working on a research project to understand the obstacles to repair household appliances using a DCE.
There are two monetary attributes that need to be considered;
- the cost of repair
- the repair bonus (a financial bonus applied by the government to encourage repair)
Is it possible to have two monetary attributes in a DCE ?
Would this mean that I can calculate two WTP? One relative to cost of repair and the other one relative to the repair bonus ?
Thank you in advance
Having two monetary variables in a discrete choice experimen
Moderators: Andrew Collins, Michiel Bliemer, johnr
-
Michiel Bliemer
- Posts: 2055
- Joined: Tue Mar 31, 2009 4:13 pm
Re: Having two monetary variables in a discrete choice exper
Yes you can have multiple monetary attributes, in transport it is common for example to have fuel cost and toll cost attributes. You can estimate different parameters for each attribute, e.g. b1 * fuel + b2 * toll, or you can assume the same parameter, e.g. b1 * fuel + b1 * toll. Economists argue that a dollar is a dollar, but often we find that people have more disutility for tolls than fuel.
If you use b1 * fuel + b2 * toll then indeed you can compute different WTPs. Alternatively, you can use a weighted average of the two cost coefficients to obtain a single WTP.
In your case, you could have cost = (repair - bonus), so you could estimate b1 * (repair - bonus) or you could estimate b1 * repair - b2 * bonus if you believe that people have different sensitivities to money for repair or bonus. It would be easier to assume b1 * (repair - bonus) as this would result in a single WTP and it is defendable since people would only pay (repair - bonus) out of pocket. In that case, in Ngene you could use something like: b1[-0.01] * repair[10,20,30] + b1 * bonus[-3,-6,-9].
Michiel
If you use b1 * fuel + b2 * toll then indeed you can compute different WTPs. Alternatively, you can use a weighted average of the two cost coefficients to obtain a single WTP.
In your case, you could have cost = (repair - bonus), so you could estimate b1 * (repair - bonus) or you could estimate b1 * repair - b2 * bonus if you believe that people have different sensitivities to money for repair or bonus. It would be easier to assume b1 * (repair - bonus) as this would result in a single WTP and it is defendable since people would only pay (repair - bonus) out of pocket. In that case, in Ngene you could use something like: b1[-0.01] * repair[10,20,30] + b1 * bonus[-3,-6,-9].
Michiel