MRIO studies using zip code level data

Comments

5 comments

  • Avatar
    IMPLAN Forum
    Hi Dan, The reason we don't have tradeflow data at the zip-code level is that we don't have zipcode-to-zipcode travel cost data. We only have the cost of travel county-to-county travel cost data (from Oakridge National Laboratory). These costs are used as the county-to-county "distances" in our gravity model (see here for more: http://implan.com/v4/index.php?option=com_docman&task=doc_download&gid=138&Itemid=60). The way you've described it, it doesn't sound like you need MRIO anyway. Did you mean to say that you would like to show the impacts on the county as a whole of the zip-code level impacts? While the level data (e.g., Output, Employment, etc.) will vary by zip-code, most of the relationships (Output per Worker, Earnings per Worker, etc.) will be the same for all the zip-codes as for the county. So, for example, if you only know Employment and are using that to set up your impact, the direct Output, Employee Compensation, etc. will be the same between both of your zip-code level models and your county-level model (see here for more: http://implan.com/v4/index.php?option=com_multicategories&view=article&id=590:590&Itemid=71). Thus, the main driver of different results between the two zip-code level areas will be the RPCs (i.e., the rate of leakage from the area). To approximate MRIO, you could: 1. Run the impacts at the zip-code level. 2. Create a county-level model and edit the study area data for the directly impacted sectors of the combined region to reflect the relationships of those same sectors in the zip-code model (i.e., Employment per Output, Employee Compensation per Output, etc.). Run the same set of impacts on the county model (you can simply import each of the impacts straight from your zip-code model via Activity Options > Import > From Another Model). 3. Subtract the Step 1 results from the step 2 results - this difference is the extra impact from the feedback loops between the zip-code region and the rest of the county. Repeat for the other zip-code region (and with a new county-level model). Note that the county-level impacts will be nearly identical in both cases since the zip-code relationships are based on county-level relationships as described above. It is the leakage/feedback effects that will differ between the analyses due to the different RPCs in each zip-code region. But as far as the county is concerned, it really doesn't matter where the direct impact occurs when using this methodology. With the Econometric RPC method, we know how much of a zip-code-level region's demand for commodity X must be imported, but without zip-code-level trade data, we don't know what proportion of those imports come from within the rest of the county or somewhere outside of the county - and whether this varies depending on which zip-code region we're dealing with.
    0
    Comment actions Permalink
  • Avatar
    Ryan Sharp
    Please provide additional details related to Step 2 in the recommendation to approximate the MRIO analysis. It is not entirely clear what changes should be made based on the description provided. For example, what specific industry relationships should be changed in customizing the study area data for the region/county to match the zip-code level model? Also, when changing the industry relationships for the region/county, should the value added/intermediate expenditure relationship be locked or allowed to change? Thank you for your assistance.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Forum
    Hi Ryan! To create a mock MRIO using ZIP Codes: 1. Create a model and run the impacts on the ZIP Code region of your choice. 2. Create a combined model. This model should include your region from Step 1 and surrounding cities of interest, or you can use the whole county file if you wish to compare to the county. One caveat to this is you will want to ensure that you county Model is built econometric Regional Purchasing Coefficient in order to be comparable. This is modified in File>User Preferences. Once you have opened the User Preferences window, you will want to choose the Analysis tab. On the Analysis tab you will select the Social Accounts tab, and the Econometric Regional Purchasing Coefficient radio button. You will want to change this back to Trade Flows after you build your mock MRIO county or other comparative Models. 3. Edit the study area data for the directly impacted Sector(s) of the combined region (step 2) to reflect the relationships of those same Sectors in the desired ZIP Code region (Step 1) (i.e., Employment per Output, Employee Compensation per Output, etc.), then re-build the model (Options > Construct > Multipliers). Since the impacts are occurring in Step 1 region, you need the combined model to look like this region for the directly-impacted sectors. 4. Run the same set of impacts on the combined model. Fortunately, you can simply import each of the impacts straight from your Step 1 model via Activity Options > Import > From Another Model. 5. Subtract the results from step 1 from the results in step 4 - this difference is the extra impact from the feedback loops between the modeled regions. To navigate to Customize>Study Area Data for Step 3. If your Customize menu is not turned on you will need to go File>User Preferences again. Once you have opened the User Preferences window, you will want to choose the Analysis tab. On the Analysis tab you will select Advanced Modeling, if it is not already checked, and then check the Customize check box. This will activate the Customize menu in both your Standard and Tasks bars. From this menu you will choose Study Area Data. Note all the numbers from the smaller region for this same industry and overwrite these in the larger model.
    0
    Comment actions Permalink
  • Avatar
    Ryan Sharp
    It would be helpful to have a much more detailed description of item 3 per my earlier post. Thank you!
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Forum
    Hi Ryan, Hopefully this article further addresses your question: http://implan.com/index.php?option=com_content&view=article&id=401 Thanks!
    0
    Comment actions Permalink

Please sign in to leave a comment.