multipliers in county < subset of zips

Comments

15 comments

  • Avatar
    IMPLAN Support
    Please let us know if our response to your prior post does not address your concerns stated here. Thanks!
  • Avatar
    jjohnston
    Hello, As you can see from my second posting, I ran the model again building the County from zip codes and still had the problem. Please take a look at the updated files attached to the second posting, and the text of the second posting. The difference in multipliers seems to be coming from the induced impacts. Thanks, Jenny
  • Avatar
    IMPLAN Support
    Just to verify, you do have your county model built with econometric Regional Purchasing Coefficient for the attached results? Thanks.
  • Avatar
    jjohnston
    In order to avoid having differing geographies from the City model (eg one City zip extending outside the County), I set up the County model the same way as I set up the City model, so by selecting all the zip codes that were listed for Santa Clara County. Therefore, the County model ran the same way as the City, i.e. econometric Regional Purchasing Coefficient. I did not make the adjustments you talked about, since it was run in this way. The results set up in this way are the tables attached to my second posting, NOT the first.
  • Avatar
    IMPLAN Support
    Thank you for confirming Jenny, and we apologize for the confusion. We would suggest comparing the RPCs for some of the most-affected sectors. You certainly can compare saving rates in the SAM for Household's but this wouldn't explain the cases where the Indirect is also higher in the city, which would be more easily explained by the RPC's. You can view these by going into the Explore> Social Accounts> Balance Sheet (Tab), and select View By: Industry Balance Sheet and the Commodity Demand tab. This display can be sorted by Gross Absorption to show you what the highest to lowest demand requirements are, or by RPC for comparisons of highest RPC values or by Regional Absorption. You may also find that some combination of the two are contributing. Please let us know if you are unable to find a reasonable explanation in looking at these methods, or if you have any additional questions.
  • Avatar
    jjohnston
    Hello again. To clarify, when I ran the model the second time, with the County composed of its components zips, I found that some of the sectors I am looking at have a higher total SAM multiplier in the City than in the County, and some not. When I split the components of the multipliers into indirect and induced, EVERY sector had a higher induced multiplier in the City than in the County. ONLY ONE had a higher indirect multiplier in the City, and it is not that important for my purposes, so I am most interested in understanding what is going on with the induced multipliers. This is true for both employee compensation and employment. I am attaching the tables again, with superfluous columns hidden. [attachment=432]IMPLANSJH20Plant2011-CountyEmploymentMultipliers.xls[/attachment] [attachment=433]IMPLANSJH20Plant2011-CountyEmployeeCompensationMultipliers.xls[/attachment]
  • Avatar
    IMPLAN Support
    Hi Jenny, I think at this point it might be best for us to view your models. Can you please zip up your models and email them to implangroup@implan.com. Please include the post number #16561 in the subject line. Thanks
  • Avatar
    jjohnston
    Thanks. I will do that, except now I have a software problem. The models are on my home computer and I am working from the office today. I can run them again very quickly, except that I cannot open IMPLAN on the office computer all of a sudden. Just left a message on the sales and software support line and will be able to send the zips as soon as the software is working. Jenny
  • Avatar
    jjohnston
    Hello again, In working with the models I just ran on the work computer, I am realizing that all the multipliers, and the differences between the City and County, are different than what I ran based on the same sets of zip codes on the other computer. I will send you a zip with those models tomorrow. Now I really can't figure out what's going on! Thanks, Jenny
  • Avatar
    IMPLAN Support
    Hi Jenny, Thanks for providing the models. It certainly does appear to confirm that the city RPC is higher in many cases than the county, thus reflecting that city supplies much of the counties demand. Please see attached. [attachment=434]Book1_2013-09-05.xlsx[/attachment] We can certainly take a look at your other models as well if you would like. If you want to force the city and county to match you can use the Customize>Trade Flows screen, RPC tab to overwrite the county RPC's to match the cities. Please let us know if you have any additional questions.
  • Avatar
    jjohnston
    I don't believe you read the whole of the email I sent last night. I have been trying to figure out a solution to this problem for more than a week now. It is already difficult to communicate with the time difference between California and Minnesota; it would really help me if you took a thorough look at the information I am sending so I don't have to re-iterate in between. 1) As I mentioned in last night's email, when I loaded the same City zip codes to my office and home computers (now you have both these models, the one set up at the office and the one set up at home), the employment and employee compensation Type SAM multipliers for industries 234-258, 329, 340, 371, 372, and 376 were all different. Also, the gross regional demand and employment, as displayed on the opening model information pages, were different. I do not understand why I should get a different result on the two different computers. Possibly it is an error on my part, but I cannot track it down. 2) I do not want to match the City to the County. It seems to me that if an industry is located in a City, and has certain multiplier effects within that City, then when you extend the boundaries of the model, the multiplier effects should increase, not decrease. I believe that IMPLAN does not allow you to pinpoint the geographic location of the industry. If that is the case, what happens, for example, if I have an industry I know is located in Los Angeles, and I want to see its impacts state-wide? Surely the statewide average multiplier effect does not reflect the fact that the industry will have certain synergistic impacts in LA, and then carry over into other areas. Is there a work-around for this situation? 3) We buy the County Plus Package for our analyses company-wide in order to illustrate the impacts within the affected City, and then in the County. If we are not able to create a work-around to represent the fact that we know the industry is located within the City, there is no reason for us, in particular, to use the cheaper County package instead of County Plus. I will be in and out for the rest of the day, but hope we can continue to make progress on these questions. Thanks, Jenny
  • Avatar
    IMPLAN Support
    Hello Jenny, We apologize that we did not see you second email with the second set of models until after you made your previous post. Doing some research, it also appears you have two different devices. Did each of your models come from the different appliances (i.e. did you build them from two different data sets?). We certainly see the variances you are explaining and we are unable to determine how the models that include PLAN in the title (the ones sent in the latter email) were created. When we build the models in house using the same data ZIP Code data, we get the same results as the first files that you sent: Model Information Model Year 2011 GRP $63,909,533,958 Total Personal Income $62,147,880,000 Total Employment 448,978 Number of Industries 310 Land Area (Sq. Miles) 505 Area Count 57 Population 978,413 Total Households 336,694 Average Household Income $184,583 You might try checking to see if you are running two different versions of the software (Help>About). You should have both running either 3.0.17.2 or 3.1.1001.12. If you aren't running either of these versions, we would recommend updating. If you need help with that, please let us know. If you can't access the update files thought the software Help>Check for Updates, you can [url=http://implan.com/V4/index.php?option=com_docman&task=cat_view&gid=160&Itemid=60]download the update files here[/url]. If you are using two different data sets, you may want to try to redownload the one that you created the PLAN models with, and see if that corrects the issue of the variance between the models. In regard to your second concern, the RPC is the answer for that. The RPC is built into the Multiplier calculation, and is based off of the ratio of supply and demand. The reason the city Multiplier's are higher is because the city RPC is higher as demonstrated in our attachment, which also included tabs for the Multipliers and the Household Savings rates to demonstrate to you that that was not the issue. Since the rate of demand is higher at the county level than at the city level, but the rate of supply is either identical or not significantly higher in the county than in the city (which indicates that as we stated earlier the city produces a majority of the counties production (and since it impacts mostly Induced impacts likely the city represents a significant portion of the services and good purchased by Households) this makes the RPC, the amount of local supply available to meet local demand, higher for city, and thus lead to a higher Multiplier for the city. This reflects a real life situation and thus is not an error of the model. You can change this as we mentioned by editing the RPC's in the city level to match. Since the Multipliers build in the RPC's editing the RPCs in the city model should resolve the variance you are seeing in the Induced impact. If the county-plus packages aren't working for you we certainly would encourage you in the future to only purchase county level data, we certainly would never want you to purchase product information that is not useful to you. There are unfortunately many caveats to using ZIP Code data, and sometimes although the Multipliers reflect real life situations, as you are seeing here, they don't necessarily have an equivalent logical understanding (i.e. it is logical that county would have larger Induced impacts than the city, but the way a Multiplier model works and the RPC's function not all economies behave 'as expected'). This was one of the reasons we invented MRIO modeling, and while we love to make it compatible to ZIP Codes, so as to help you and people like you to avoid these issues, there simply isn't sufficient raw data that we are aware of at this time to do so, it is certainly something we are researching for the future. Here are a couple of posts that describe other similar situations: http://implan.com/v4/index.php?option=com_kunena&func=view&catid=80&id=14249&limit=6&limitstart=6&Itemid=35#14261 http://implan.com/v4/index.php?option=com_kunena&func=view&catid=80&id=11987&Itemid=35#11990 http://implan.com/v4/index.php?option=com_kunena&func=view&catid=84&id=13628&Itemid=35#13635 Again we apologize that there isn't a clearer or more direct answer for your concerns.
  • Avatar
    jjohnston
    Thanks for your response. I would have thought the two software versions would be the same, since IMPLAN support helped me re-load IMPLAN on my home computer week of August 25, and on the office computer week of September 1. However, I will check the two software versions. As to the question of the multipliers, I will have to take a look next week. Have a good weekend, Jenny
  • Avatar
    IMPLAN Support
    Hi Jenny, If they are loaded from the same appliance and both updated, they will likely be the same versions, but if they were new loads from different periods, they may be variant. The reason for this is that IMPLAN updates do not edit your original setup files, those remain static to the time of the appliances creation. The files that are updated, are unfortunately wiped out when you uninstall, or are left behind on the previous systems hard-drive. We hope you have a nice weekend as well, and if you do need further help with this please let us know. Thanks, IMPLAN Support Team
  • Avatar
    jjohnston
    Hello, 1) Regarding the different model results from different computers, the versions are in fact the same. I do not understand the explanation about the new loads from different periods, but will call tomorrow from the home computer to see if we can download and update the software to the home computer again, and get everything to match in whatever manner you suggest. The same appliance was used for both software downloads, by the way. 2) Regarding the RPCs and household savings rates, I think I understand the concept you are suggesting, but since the breakdown of the multiplier differences between indirect and induceds, and the RPC results, vary depending on whether I run the model on the home computer or office computer, I propose we first get the model results to match and then continue to dig into that one. Many thanks, Jenny

Please sign in to leave a comment.