Mexico I-O Tables, Regionalizing & Missing Data
This question is not regarding IMPLAN per-se, but could be relevant to modifying local models. I'm in search of someone with experience working with building regional I/O models from national I/O matrices. I'm working with Mexico's national I/O table at the 4-digit NAICS code level, regionalizing it to the state level using the CILQ method. The industries I'm interested in are agriculture and fisheries, and the national matrix doesn't provide ag industries by 4-digit NAICS code. Fisheries and ag need to be separate because I'll be comparing impacts in the two sectors. I have some rough estimates of value of production for ag, but I would be using this to fill in the gaps for state GDP data on ag (which is not available) in order to build my location quotients. Are there any resources available on the best methodologies for filling in data gaps for regionalizing I/O matrices? Thanks for any suggestions you can provide
-
Hi, Thank you for your post. I apologize for the delay of our response. There is no best-practice method that will get you there every time. There are many ways to do this type of work and each of them has strengths and weaknesses. The only way to determine which methods best serve your particular project is to evaluate the project details thoroughly. This includes evaluating existing data, alternative data, and supplemental data, as well as the project objectives and requirements to determine which of these methods provides both theoretical and structural validity as well as functionality. Without this information it would be difficult to provide any direct suggestions. For regionalizing a national I/O, you will need to make certain assumptions about the study region (state) relative to the nation. With location quotient techniques like CILQ, that assumption is generally that the state’s share of national employment (by sector) is equal to the state’s share of national output (or something conceptually similar to this). These assumptions provide a bridge that allows national I/O data to be regionalized, but they also come with some drawbacks (in this example, output per worker in each industry would be equal to the national output per worker so, in your example, comparison of the output/worker ratios of the traditional Ag. Sector and Fisheries would be the same as studying these ratios nationally). As for disaggregating the Agriculture industry. In this case you need two things: (a) the assumption that is going to provide the bridge from the aggregated industry to the disaggregated industry, and (2) the data required to make that assumption actionable. That is, in order to disaggregate the agricultural industry, you’ll need a data source that has the desired aggregation scheme (this could be Employment, Value-Added, previous year’s data, data from a proxy region, or any number of things); then you’ll assume that your aggregated industry disaggregates according to this new data source. Again, please be careful to understand and document any assumptions you make in the process. As for combining the two processes (the regionalization of the national I/O and the disaggregation of the Ag industry), this should be dealt with separately. However, you have the freedom to determine which process you do first. You could regionalize the national I/O and then disaggregate Ag based on a similar state-like region or you could disaggregate Ag based on a similar nation-like region and then regionalize the resulting national I/O (which would now have detailed Ag industries). IMPLAN does offer services to construct such models for you. These models operate within IMPLAN for fully functional impact analysis including full Multiplier reports. We have extensive experience with the issues and have constructed Models using data for Mexico in the past. If you are interested in using an existing Mexico IMPLAN Model or having us construct an IMPLAN Model for you using your data, please let us know! Thanks!
Please sign in to leave a comment.
Comments
2 comments