On regional analysis
I am planning to do contribution analysis of forest related industry using county data in Texas. I need to do analysis for some counties (say 43), a region (sum of 43 county), and State as a whole. Obvious approach comes into mind is MRIO analysis. As we know, MRIO is rather complex and requires careful attention to avoid double counting. Therefore, I am planning to do it following the underpinnings of MRIO but rather differently (mathematically). Following approach comes into mind.
A. Obtain overall economic contribution in entire Texas following multi-industry contribution analysis.
B. Obtain overall economic contribution in region (43 county) following multi-industry analysis.
C. Obtain overall economic contribution in other than region B (All county in Texas-43 county) following multi-industry analysis.
D. Obtain overall economic contribution in each 43 county (one at a time) following multi-industry contribution analysis.
Steps for my analysis:
Step 1. While, result summing B and C ( let's say z) should equal A, it will be less than A. It is because B and C don't capture cross region inter-industry trade. Therefore, I want to proportionate this difference (A-z) equally between B and C and get revised value for B and C.
Step 2. I will sum all the value from D. This will also be less than initial value of B for the same region. Again county sum won’t capture inter-industry trade between counties.
Step 3. Now obtain the value of B from Step 1 and get adjusted value for each county depending upon their respective share in the region.
The reason I want to do in this way is because I can utilize final county values to obtain economic impacts for other political bounties (such as state senate, which IMPLAN does not have separate data). After this analysis, my math suggests that if I will sum all the counties, it will equal to the value of region. Similarly, sum of region B (revised) and region C(revised) will equal to total state value A. I am curious to know whether my thought process is making sense. Do you have any suggestions?
Thanks,
Omkar
Was this post helpful?
-
IMPLAN SupportHello Omkar, We apologize about the delay in responding. We believe that what you have described here is a logical approach to your study. We have assisted in other studies that scaled congressional district results to state-level results, and state-level results to national-level results. By doing this we captured inter-regional linkages that were leaked out in the smaller region. Step 1) While splitting z equally to B and C, a better methodological approach would be to split each industry impact in z to B and C weighted by Output in the Study Area data. For example, if "z" for restaurants (sector 413) is $100, and 2012 Sector 413 Output (from Explore Study Area> View By:Industry Detail) is $4 million for region B, and $2 million for region C (total combined is $6 million) then region B would get $66.67 of z (4 mill/6 mill * $100) and region C would get $33.33 of z. This requires a little more work but is more satisfying. Step 2) This step is fine. Step 3) We are not sure what you mean by "depending upon their respective share in the region"..... it sounds like what we suggested as an add on for Step 1). If so, we agree with this approach. However, if we are not understanding what you are describing here, please let us know, so that we can make sure that we give you accurate feedback. Please let us know if you have any additional questions, or if we haven't addressed all your concerns.0 -
You are correct that step 3 is an add for step 1. I can follow all of your steps. Thank you very much! Omkar0
-
I have follow up question to my analysis. Let me explain my case first. My definition of forest sector is sum of many IMPLAN sectors (like 15, 16, 18, 95, 96-100,101-112 etc.).I subgroup these IMPLAN sectors after following steps for multi-industry contribution analysis and obtain the impacts associated with each subgroups. All I am trying to do is to demonstrate the total forestry contribution, but want to see the impact of individual sub-groups of sectors. The values I get by running all sectors in one scenario versus each individually (one at a time) always become same, other than small rounding errors. That helps me to report individual sectors contribution as well as entire forest sector contribution. This is important as reader can see the contribution of logging (sector 16) as well as entire forestry sector to the state. Now, let me come to state, region and county analysis issue. I followed your suggestions to capture inter-regional linkages and obtained the revised regional values. When I added total forest sector values (combination of all sectors) from regions following step 1 (your suggestions), sum of regional forest sector value become equal to the entire forest sector value of the state. That’s what I wanted. However, as mentioned earlier, I want to see the entire forest sector as well as individual sub-group contribution in every region (North, East, South, West). When I added adjusted value of individual IMPLAN sector (sector 16) across all regions, they don't equal to the sum of adjusted value of sector 16 from my statewide model. It won't make sense as sum of sector 16 in region should equal to sector 16 state value. Since regional share of each IMPLAN sector are different (Some region do have one industry others don’t do), the step 1 you suggested in previous post does not seem to work. I hope I clarified my concern. Any suggestions?0
-
IMPLAN SupportHi Omkar. In response to your latest Post, we have developed a Spreadsheet Template, which shows you how to scale your impacts from the county level to the regional level and then from the regional level to the state level. This will capture the trade flow leakages in the multi-regional system and ensure that results from smaller region models add to results reported by larger region models. The spreadsheet Template contains a verbal explanation of the method and can do these calculations automatically. You just have to modify it to match the number of counties and regions under consideration and then input the initial values reported by each model. This process will need to be repeated for each scenario being modeled. With the Spreadsheet, you could scale your total impacts, using one of two approaches. 1) Use the attached Spreadsheet to calculate totals and then only report totals. 2) Use the Spreadsheet and method to calculate new indirect, new induced, and total impacts separately, but then don't calculate totals with the current method. Instead, sum the direct, new indirect and new induced to get new totals. The results may not sum exactly to your original totals since the shares for scaling may be different for each type of impact. However, the differences should be slight but there nonetheless. We hope this resolves your aggregation issue. [attachment=491]Regional_Statewide_Impacts_Balance_Template.xlsx[/attachment]0
-
Thank you very much.0
-
I reviewed the excel file. Do you think this excel file you prepared will solve my situation, which I have attached with the snapshot MS-Word file here.0
-
I have additional comment on the snapshots I attached earlier. Total impacts are obtained from ALL model and is equal to the total impact from forestry, logging, primary solid and secondary solid.Attached is the revised snap shot to avoid confusion.0
-
IMPLAN SupportHi Omkar. We have reviewed your recent Forum Post and attachment. In doing so, we have one important question. First, when entering your data in the events rows, did you select your Event Year first and then enter your event value? This is not a trivial matter. If you are not certain, then you may need to for each event, delete its value, hit enter to make sure the model picks up this change, and then re-enter the value. You will do this for each event in your model. In order for the deflators to be applied correctly and before applying the multipliers, you need to set the Event Year before you enter values into any of the fields. Since your event year is 2012, you should view your results in 2012 dollars. In order for your impact results to reflect the same values as in your Event, you need to set the "Dollar Year for View" in the Scenario Results screen of the to the same year as your "Event Year" (in your case, this is 2012 and not the Default 2014, which is automatically set by the model. Please let us know if we can be of further help to you.0
-
Thanks for the note. I changed dollar year view to 2014 after obtaining results so I don't think it will be problem. Did you had chance to look over matrix I prepared in the attachment? I am interested to know whether following relationship can be maintained via the excel sheet you prepared: forestry(region1)+ forestry(region 2)= Forestry (State) logging(region 1)+ logging(region 2)=Logging (state) .................................................. All industry (region 1)+ All industry(region 2)=All industry(state)0
-
IMPLAN SupportYou are most welcome. Please let us know if this solved your problem.0
-
IMPLAN SupportHi Omkar. We just noticed in your latest post that you changed the Dollar Year for View to 2014. You wan to view the results in 2012 Dollar Year for View and not 2014, unless your data is 2014. Try this and yes the Spreadsheet should sum the regional totals fairly closely to the State totals.0
-
Thanks.0
-
[attachment=503]Regional_Statewide_Impacts_Balance_Template.xlsx[/attachment] Dear Moderator: I conducted regional analysis following the excel balanced sheet (obtained from you). While it provided adjusted regional values in general, there were some problems when industry A is sum of multiple smaller sub-industries.I have attached the excel file with my concern and the background information. I will appreciate if you could suggest how this excel sheet can be revised to solve the issue that I have.0
-
IMPLAN SupportHi Omkar. We have reviewed your recent post and spreadsheet. Although the numbers do differ slightly, they are within the generally accepted range of error (usually 5% or less on either side of the estimate). Also, the impacts will never add exactly to the totals because you are impacting different regions with different economic structures and input demands in your study. The best we can do is to scale the results to narrow the margin of difference between individual regions and the sum total of all regions in the model, and you have done this. In your case, when you calculate the variance between the sum of the individual regions and the overall regions, the percent of variance varies from a -2.98% to a high of 1.23%. We know that any margin of difference is too large, but this is really not too significant considering all of the manipulations that have occurred getting to this point. What we would suggest is that you state in your report or presentation that the differences in estimates are due to rounding of the original data and that after you entered the data into the IMPLAN model, the model further rounded the estimates in performing the analysis. Thus, the results for each region may not necessarily sum to the overall regional total. Beyond this point, if you need additional support, we would suggest that you consider purchasing a “Support Pack” agreement with IMPLAN. The cost is $1,400 and with this agreement you have access to technical assistance and support for a period of one year. This package allows you to have call-in and email access to IMPLAN specialists who can assist you in finishing a project. Alternatively, we also have project consultation at an hourly rate of $350.00, but our recommendation would be as above, to just note the reasons for the variance in report. Please let us know if you can be of further help to you.0
Please sign in to leave a comment.
Comments
14 comments