1. IMPLAN - Support
  2. Community
  3. Analysis Questions

Trade Flow vs. Econometric RPC

Avatar
nkemper@uark.edu
I've just finished reading the conference paper by Thorvaldson, et al(2011) "Updating and Enhancing IMPLAN's Econometric Regional Purchase Coefficients" from the MCRSA conference. Great discussion in the paper. The paper also prompted me to reexamine my previous methods in doing contribution analysis and I've also come up a bit confused and need some advice on what to do. Here's an example of my problem: Say I want to estimate the contribution of the soybean production and processing industry in Arkansas and I define that industry as: 01-Oilseed farming and 45-Soybean and other oilseed processing. Also, it is assumed that in Arkansas those sectors are exclusively soybean and we assume 100% of the value in those sectors to be related to soybean production and processing. [b]Here is my previous procedure (summarized)[/b] 1. Build IMPLAN AR state model and Economic RPCs 2. update 01-Oilseed farming output value with latest NASS/ERS value of production for Arkansas 3. rebuild model to incorporated study area data change 4. zero out RPC for sectors 01 and 45 and "save" then rebuild model 5. enter impact events and analyze scenarios So I have two topics I need help with based on the above: 1) Tradeflow vs. Econometric RPCs and 2) modifying commodity production. [b]Tradeflow vs. Econometric RPCs[/b] - Thorvaldson, et al (2011) states: "...tradeflow RPCs not responsive to edits to the underlying study area data. In these instances, econometric RPCs are needed." In my case, I have to update the value of production data in the agricultural sectors to reflect state level estimates for Arkansas. Does this mean that I should not use the tradeflow RPCs if all I am doing is making a slight adjustment to the baseline value of output? I have assumed that I should use the Econometric RPCs. Because tradeflow RPCs are preferred to econometric I would like to be able to use them; also I need them for multi-regional models as well. Which should I use in this case? [b]Modifying Commodity Production[/b] - assuming I should use Econometric RPCs, applying what you describe in the FAQs regarding multi-industry contribution analysis, that I should edit the 01-soybean industry commodity production so that the primary commodity equals 1; I also do this for sector 45-oilseed processing. However, in the past I have not done this as part of my procedures (as noted above in my example). Is this recommendation on modifying commodity production specific to using Trade Flow RPCs? Or does this advice hold if I am using Econometric RPCs? In the case of the soybean industry, modifying commodity production in these two sectors results in an overall higher impact (just slightly); I guess the point is that they are different and I'm just wondering what you would recommend in this case? As always, your advice is greatly appreciated! Nathan
Was this post helpful?
0 out of 0 found this helpful

Comments

3 comments

Sort by Date Votes
  • Avatar
    jenny
    Hi Nathan, One solution would be to build a copy of your model but use econometric RPCs, make the output changes, and rebuild with econometric RPCs. Then you can edit the RPCs for the affected sectors in the tradeflows model, based on the RPCs given by the econometric model. That way, you can continue using the tradeflows model but with edited RPCs for the sectors for which the data was edited. However, please note that the econometric RPC method currently available in IMPLAN is somehwat outdated and can sometimes produce unexpected results. So if it gives an RPC that is significantly different from the tradeflow RPC, you may want to keep the tradeflow RPC unedited or use some average of the two. In the second set of edits, you are not actually changing the value of Output, so using econometrics (or adjusting tradeflow RPCs) is not necessary.
    0
  • Avatar
    nkemper@uark.edu
    Jenny, Very interesting proposition on editing the trade flow model with econometric RPCs for only the sectors for which data is adjusted. I'm going to have to tinker around with that. Thank you. One additional point of clarification on the second situation: in a contribution analysis (using either trade flow RPCs or econometric RPCs) would you recommend modifying the commodity production of the sector of interest so that the primary commodity equals 1 and thereby setting industry production from a mix of intermediate demand and final demand to only final demand (exports)? Just want to make sure this would be the recommendation in the case of both trade flow RPCs and econometric RPCs. Thanks so much, I appreciate your thoughtful suggestions. Nathan
    0
  • Avatar
    jenny
    That is correct. Additionally, since you are assuming that the food processing plant only processes soybeans, you will want to edit the food processing sector so that the only agricultural crop it purchases is Oilseeds. You can make this change by going to Customize > Industry Production, then scroll down to sector 45. Add the Coefficients for grains, cotton, and all other crop farming products to the Coefficients for oilseeds, then set the Coefficients for grains cotton and all other crop farming products to 0. Put a checkmark in the "Fixed" boxes next to all these changed Coefficients, then click "Balance" and "Save". You will need to rebuild your model after these changes.
    0

Please sign in to leave a comment.