Aggregating a model

Hi Janet, If you have the ability to set up the analysis with the individual Sectors, IMPLAN 126, 214, 215, 216, 222, 223, 250, 270, and 271, this is the preferred method, as it avoids [url=http://implan.com/V4/index.php?option=com_glossary&id=227&Itemid=57]aggregation bias[/url]. Whether you aggregate results or aggregate the models you will start in the Model Aggregation screen, which is found in the Options>Aggregate Window. From this window you can create a custom aggregated Sector and give it your name, Clean Technology Manufacturing. Once this is created you will choose: [ol] [li]To Aggregate the Model: By choosing the "Aggregate Model" button at the bottom of the screen. This will require you to [url=https://www.google.com/url?q=http://implan.com/v4/index.php%3Foption%3Dcom_multicategories%26view%3Darticle%26id%3D555:555%26Itemid%3D14&sa=U&ei=Vp04UtHBL8PF2AXU7oDwDQ&ved=0CAsQFjAC&client=internal-uds-cse&usg=AFQjCNHdHKtuCS2UAt0indx63n2jRN3S4w]rebuild the Model [/url]as all new Multiplier relationships and a new Sectoring scheme, will need to be developed.[/li] [li]To Aggregate the Results: By choosing the "Close" button. You will not need to change the Model with this technique, but you will be able to proceed with your analysis. Once the analysis is complete you will be able to go into the Detail results and choose the Options>Aggegate>User Scheme function to see the Detail Results with the Aggregate Industry representation. This method aggregates by summing the results so the Multiplier and Sectoral identities of all the included industries are maintained.[/li] Another advantage of the second method (aggregating results) is that you can remove (saving if you want to) the previous aggregation scheme. Create a new one and repeat within the same Model. If you aggregate the Model, you will need to build a new model from the .odf files for each aggregation scheme. Here is [url=https://www.google.com/url?q=https://implan.com/v4/index.php%3Foption%3Dcom_multicategories%26view%3Darticle%26id%3D556:556%26Itemid%3D14&sa=U&ei=fpw4UpvVBcbs2QWEo4GoBg&ved=0CAcQFjAA&client=internal-uds-cse&usg=AFQjCNFmeA29OW_4rUaDwRX_572GieTPBg]a link [/url]that includes additional information on aggregating Models. Please let us know if you have any additional questions. [/ol]

Thanks for the instructions, but I need a little clarification (I'm a novice, but hope to get better!). After I select the industries to aggregate, then close the aggregation screen to set up my analysis (100 new jobs in the created sector), the new sector is not among my choices ... it remains 126 Basic chemicals and not Clean Tech Mfg. Should the sector list reflect my newly created sector ... or am I missing a step in the way I've done it: I started by creating a New Model, selecting the counties that will be part of the study area (in this case, the entire state). I next Aggregated my sectors and closed the window (without aggregating the model). Then, I selected Setup Activities to begin the analysis, but my new sector is not there ... do I select Sector 126, anyway? Thanks for your help!!

Hi Janet, If you only have 100 jobs in the aggregate Sector with no way to break them between the individual Sectors that you are aggregating you have no choice but to actually aggregate the model for each aggregation scheme ou need to create. The second method that allows you to aggregate your results and avoids the bias and different model issues is only an option if you can split the jobs amongst the respective Sectors. To complete, you will just need to open up the Options> Aggregation screen again, choose the "Aggregate Model" option and then [url=https://www.google.com/url?q=http://implan.com/v4/index.php%3Foption%3Dcom_multicategories%26view%3Darticle%26id%3D555:555%26Itemid%3D14&sa=U&ei=Vp04UtHBL8PF2AXU7oDwDQ&ved=0CAsQFjAC&client=internal-uds-cse&usg=AFQjCNHdHKtuCS2UAt0indx63n2jRN3S4w]rebuild the model [/url]. You will then just need to start you next aggregation analysis in a new model and repeat the process. Sorry we didn't make that clearer in our earlier post. Please let us know if you have any additional questions.

Thanks ... aggregating the model is the way I ran it the first time, but didn't like the numbers I got. (My indirect jobs were greater than the initial new 100 jobs. If that's even possible, maybe that's an example of the aggregation bias.) I'll rerun and compare numbers ... maybe I did something wrong the first time. Thanks for confirming that I aggregate the model since I don't have a separate new jobs breakdown for each industry.

Glady, If you run into the same problem again, please let us know. If you can also give us an idea what the region is we can see if we can duplicate what your seeing in house and give you an idea of what might be going on. IMPLAN Support Team

I tried again - aggregating the model - and got the same results. My data - all the counties in West Virginia. I first aggregate my sectors - IMPLAN 126, 214, 215, 216, 222, 223, 250, 270,271, naming them Clean Technology Manufacturing. I aggregate the model; close when complete; reconstruct the model. Setup new activities and events - New activity: 100 New Clean Technology Manufacturing Jobs. Event - new sector 126 Clean Technology Manufacturing and add 100 new jobs. Next, and run the analysis. Direct jobs: 100 Indirect jobs: 225.2 Induced jobs: 126.3 Total jobs: 451.6 OMG ... what am I doing wrong? (So hoping I can come to your NC class in December!) Thanks again for ALL your help! Janet

Hi Janet, Your employment multiplier looks fine – it is only output multipliers that tend to be around 2 or less. When looking at the multipliers of the individual sectors prior to aggregation, you’ll notice that sector 126 is the largest of the 9 combined sectors and thus will be given the most weight when the sectors are aggregated, and that its employment multiplier is 3.92. Indeed, when we created a 2011 WV and calculated a weighted average employment multiplier for these 9 sectors, for comparison, we got 3.89, which is very close to the employment multiplier of 3.97 from our aggregated model and very close to what you are seeing (see attached spreadsheet). If you are working with a model other than 2011, that would explain the difference between your figures and ours. If not, we’ll want to investigate why your multiplier is even higher. We look forward to meeting you if you can make the December class. Please let us know if this addresses your concern or if you have any additional questions.

Thanks for all your help. Based on your response, I guess my numbers aren't out of whack (I'm using a 2009 data set). One last item and I'll put this little exercise to bed. The latest reponse referenced an attached spreadsheet, but I didn't see access to it. (I hope it's not right in front of me and something that I overlooked!!) Will you mind resending? Thanks!!! Janet PS - Have a great weekend :)

Nope, not you. We forgot to attach. Sorry about that. Hope you have a great weekend too. [attachment=440]MultiplierComparison.xlsx[/attachment]

Thanks much for the spreadsheet!! So ... my earlier calculations using the 2009 data are likely correct - that creating 100 new jobs in the clean manufacturing technology "sector" translates to: Indirect jobs: 225.2 Induced jobs: 126.3 Total jobs: 451.6 Pretty good sector, if you can get it :) Hope all your teams win this weekend!