Aggregating contribution analysis

Comments

21 comments

  • Avatar
    ljvoves
    Hi Christine. Based on your question, it looks like you want to do a separate MRIO analysis of a particular industry in a set of target counties at the county, regional, and state level in your state. But caution is advised not to not cross link the models when doing MRIO analysis, because cross-linked models will create issues in the trade flows and that will cause some double counting or increases in the results. The following lays out an example of how we think you want to do this using 5 generic counties: County#1, County#2, County#3, County#4, and County#5. You should create folders in your models folder for each county that has a Direct impact (Event) occurring in it. So your 5 folders would be County#1, County#2, County#3, County#4, and County#5. Inside EACH folder you should build a model for your Event region and models for the linked region: [u][b]County MRIO Models[/b][/u] [b]County#1:[/b] County#2, County#3, County#4, and County#5; and Rest of the State less County#1, County#2, County#3, County#4, County#5[br] [b]County#2:[/b] County#1, County#3, County#4, and County#5; and Rest of the State less County#1, County#2, County#3, County#4, County#5[br] [b]County#3:[/b] County#1, County#2, County#4, and County#5; and Rest of the State less County#1, County#2, County#3, County#4, County#5[br] [b]County#4:[/b] County#1, County#2, County#3, and County#5; and Rest of the State less County#1, County#2, County#3, County#4, County#5[br] [b]County#5:[/b] County#1, County#2, County#3, and County#4; and Rest of the State less County#1, County#2, County#3, County#4, County#5 [u][b]Regional MRIO Model[/b][/u] County#1, County#2, County#3, County#4, and County#5. [u][b]State Level MRIO model[/b][/u] All Counties in the State less County#1, County#2, County#3, County#4, County#5 Please let us know if we can be of further assistance.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hi Christine. Based on your question, it looks like you want to do a separate MRIO analysis of a particular industry in a set of target counties at the county, regional, and state level in your state. But caution is advised not to not cross link the models when doing MRIO analysis, because cross-linked models will create issues in the trade flows and that will cause some double counting or increases in the results. The following lays out an example of how we think you want to do this using 5 generic counties: County#1, County#2, County#3, County#4, and County#5. You should create folders in your models folder for each county that has a Direct impact (Event) occurring in it. So your 5 folders would be County#1, County#2, County#3, County#4, and County#5. Inside EACH folder you should build a model for your Event region and models for the linked region: [u][b]County MRIO Models[/b][/u] [b]County#1:[/b] County#2, County#3, County#4, and County#5; and Rest of the State less County#1, County#2, County#3, County#4, County#5[br] [b]County#2:[/b] County#1, County#3, County#4, and County#5; and Rest of the State less County#1, County#2, County#3, County#4, County#5[br] [b]County#3:[/b] County#1, County#2, County#4, and County#5; and Rest of the State less County#1, County#2, County#3, County#4, County#5[br] [b]County#4:[/b] County#1, County#2, County#3, and County#5; and Rest of the State less County#1, County#2, County#3, County#4, County#5[br] [b]County#5:[/b] County#1, County#2, County#3, and County#4; and Rest of the State less County#1, County#2, County#3, County#4, County#5 [u][b]Regional MRIO Model[/b][/u] County#1, County#2, County#3, County#4, and County#5. [u][b]State Level MRIO model[/b][/u] All Counties in the State less County#1, County#2, County#3, County#4, County#5 Please let us know if we can be of further assistance.
    0
    Comment actions Permalink
  • Avatar
    ccooper
    Thank you for your response. This is seeming rather complex. For clarification, would each of the MRIO models be modified so that they are contribution models? Is this what you mean by not cross-linking them? I am hoping that this method would produce consistent results such that the counties' contributions will add up to a region's contribution and all regions' contributions will add up to the state's overall contribution (as if it had been run by itself). Does this seem feasible?
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hello Christine. Just to be clear on what we think you are asking. Yes this could be a time-consuming process as you will need to link and run many models if you want to do a county-by-county and county-to-rest of state (ROS) comparisons. Just to illustrate in terms of the number of models that would have to be run. If there are five counties in your state and you wanted to do a county-by-county and county-to-state analysis, you would have to run 5 x 5 or 25 total county models plus 5 x 5 or 25 total county-to-rest of state (ROS) models for a total of 50 models. Each county model would be modified to remove the backward linkages in the target industry that you want to show its economic contribution to the county’s total economy and linked to another county so as to show the target county’s economic contribution in the link county as well. To remove the backward linkages in the target industry in the county, you would take the inverse of the Detail Type SAM Multiplier for the target industry and divide it into the target industry’s total output in the county. You can get to these multipliers by navigating to Explore>Multipliers>Detail Multipliers and selecting the industry you are working with. Next, you would create an industry change activity and model the modified total output value for your target industry as if you were modeling an economic impact of this industry. The results will look exactly like the results you would get if you had modeled the change as an increase in total output. The difference is that with a contribution analysis, the results indicate the target industry’s contribution to the county’s total economy and not an expansion of the total economy. A key sign in determining whether you have performed a contribution analysis versus an economic impact analysis is that the target industry’s total output value should equal the original total output before you divide it by the inverse of the Detail Type SAM Multiplier for the target industry. If it is greater than the original/unadjusted total output value, then you did something wrong. Here is a link to a paper on the IMPLAN Website that goes into great detail and shows you how to do a single industry and multiple industry contribution analysis. https://implan.com/v4/index.php?option=com_multicategories&view=article&id=660:660&Itemid=14 We suggest that you read this paper before starting your analysis. An alternative to the county-to-county and county-to-ROS approaches is to group the counties into a maximum of 9 regional models and run MRIO with those models. The maximum number of linkages the software can handle is a total of 10 individual models (the original/target model + 9 regional models). The regional models will contain all of the counties in the state, except the original/target county. Again, you would want to make sure not to cross link the models when doing MRIO analysis, because cross-linked models will create issues in the trade flows and that will cause some double counting or increases in the results. Under this approach, you will get some sense of the regional differences in contribution resulting from your target industry, and makes some sense given that some counties may be small and rural in nature.
    0
    Comment actions Permalink
  • Avatar
    ccooper
    To simplify things, I think we'll only do county level contributions and the state level. I assume then we'd want to do each county as an MRIO with itself and the rest of the state. Hopefully these will add up to the state level contribution.
    0
    Comment actions Permalink
  • Avatar
    ccooper
    Thank you, and yes I could see how many models would be needed for the analysis we were considering. I like the idea of forming regional models, especially since the state is so diverse. I will still need the county-level contributions though, but that seems doable if they only need two models each.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Christine, Forming the regions would certainly reduce the number of models that you would have to create. How many counties will be involved in your MRIO Contribution Analysis? Depending on the number, the regional approach seems more doable. We have this helps in addressing your problem.
    0
    Comment actions Permalink
  • Avatar
    ccooper
    There are 58 counties. The "industry" is really a combination of industries, so an MRIO of each county with the state less that county will be time-consuming as we remove the cross linkages and rebuild each model twice... We could do a simple contribution analysis but I expect that then the individual county results would not add up to the state contribution, which is also important. Any advice on making this more efficient would be greatly appreciated.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hello Christine, Probably the best advice we could give you is to use our batch template process for building the models. While this doesn't reduce the time necessary to edit the county models where you Direct Effect is occurring, you can drastically reduce the amount of time you spend building the models themselves. The Batch Template process includes a list of all the counties by state, so you can use this list as a copy/paste (removing the county where the Direct impact is taking place from the list for each Rest of State (ROS) model). Information on the Batch Template can be found at this[url=http://implan.com/index.php?option=com_multicategories&view=article&id=685:batch-model-processing&Itemid=71] link[/url].
    0
    Comment actions Permalink
  • Avatar
    ccooper
    We've been working our way through the counties in our multi-industry contribution analysis and have encountered some strange results in Sacramento County (2012 data). After running the scenario, the direct value added in sector 337 is negative, and in the tax impact, the direct corporate profits tax and dividends are also negative. The opti for this sector is negative and LARGE. I'm pretty sure this has to do with it being the state capital, but can you provide an explanation, and should we modify this model (and if so how)? Thanks very much for your help.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hi Christine. We are sorry to hear about the problem you have encountered. Would you mind telling us which counties you linked to Sacramento County? This would allow us to reproduce your model and perhaps determine exactly where the problem is. Thanks!
    0
    Comment actions Permalink
  • Avatar
    ccooper
    Actually, we are not running MRIO, just simple contribution analysis. I can send the model if you like.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Christine, That would be great. Please compress the model by opening the model and navigating to Options>Compact Model File and save when finished. You can send the model to: implangroup@implan.com Thanks!
    0
    Comment actions Permalink
  • Avatar
    ccooper
    I did send the model, but it was still pretty big. Just want to make sure it was received.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hi Christine, Thanks for checking. It unfortunately doesn't look like it did come through. The issue maybe that it looks like we accidentally mistyped our response to you. The file needs to be zipped, rather than compressed. Would you mind trying again sending a zipped file? We apologize for the confusion.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hi Christine, In addition, this information may help, and may be able to circumvent you needing to send in the Model. The Taxes on Production & Imports for the Sector are not negative, only the Other Property Type Income is negative. This leads to negative corporate income taxes being paid. The negative Other Property Type Income for pipeline would not/could not be a long term situation so zeroing it out in Sacramento and re-constructing the multipliers would be a reasonable solution. If you know this isn't a long-term trend (since we believe you have several years of data available to you) then we would definitely recommend this stratagem. Please let us know if this doesn't answer your question, and if you are still sending in a Model. Thanks!
    0
    Comment actions Permalink
  • Avatar
    ccooper
    Thank you. I did try to zero it out, but then value added is greater than output. I'm not sure what type of adjustment is needed, so I have zipped the file and sent it to you. Let me know if you have received it and if you have any advice. Thanks so much.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hi Christine, We apologize, but we still have not received it. Would it be possible to setup a time with you to do a logmein to transfer the model to us? Alternatively, we could try a different email address (implangroup@gmail.com). Please feel free to give us a call to setup the transfer if that is how you would like to proceed, and if all our agents happen to be on the lines, please leave us a message and we will get right back to you. We apologize that we are having trouble receiving your model.
    0
    Comment actions Permalink
  • Avatar
    ccooper
    Let me try the other email. Otherwise, we can try a logmein next week since this week is pretty busy for me.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hi Christine, It worked! Thank you for your patience and continuing trying. We will get a chance to look at your model here shortly and get back to you with what we find. Thanks!
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hi Christine, Thank you for providing your Model. The data shows over several years a gradually increasing loss of profits for Sector 337, but the 2012 is definitely more marked. There are a couple ways that you can choose to handle this. Since Sector 337 is showing a loss, you could certainly zero out their corporate tax and recalculate the tax impacts, Value Added and Output for this adjustment. Alternatively, you could re-run the impacts with the OPI zeroed out, but accounted for in the Output value in the customized screen, as a way to balance the table, without running into the situation that you are seeing where Value Added > Output. To do this we would recommend recording all the values in the Study Area screen before you do any editing so that you can verify that nothing other than OPI and Output have changed. You will first want to adjust the 'totals' and increase Output by the value of the OPI. You can then zero out OPI and this should allow you to model without running into this issue. Please let us know if you need any additional assistance with this, or if you have any additional questions.
    0
    Comment actions Permalink

Please sign in to leave a comment.