I am estimating a DEA for an industry with a multiproduct technology (two inputs and two ouptuts). When I run the "complete" model I find ambiguous results regarding scale economies. When I run the model separately for both outputs, in one case, I find scale diseconomies while in the other scale economies.
I suspect that this is not the correct way of estimating "product-specific" scale economies so, I would be grateful of receiving suggestions and references.
It needs to be for both inputs and outputs together. There can be a couple of reasons for ambiguous results: First check that are you using increasing or decreasing returns to scale. Sometimes wrong model selection may result in unexpected results. Second, is the data transformation. Make sure all data is measured on the same scale, it is mostly recommended to take the log of whole data. Third can be the software codes. This is a common issue that some codes work with one data and may not work well with other data. May try some adjustments in your codes.
Thank you Salah. I am using the Variable Returns to Scale (VRS) model. I will try introducing your suggestions. Thank you.
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India