Industry Spending Pattern

Comments

16 comments

  • Avatar
    IMPLAN Support
    Hi Jocelyn, When you import an industry spending pattern in IMPLAN and want to put just the cost of goods and services in as the Activity Level (as opposed to total production value/Output)you will first want to normalize the spending pattern. Event Options> Change All> Normalize, as the traditional spending patterns import in with a Sum of Event Values of less than 1.00 equalling their proportion of total Output. Then the LPP field, which should be set to SAM Model Value (less than 100%) for most analyses and Sectors will also act to reflect how much of the regional supply can be used to meet demand based on the Model's RPC value. As regards your results you will want to add back in, as the Direct Effect, your original total Output, as this value is local as well and should be represented. Are you doing an Analysis-by-Parts analysis? If so your final results should include your original 391 Output, Labor Income, Value Added and Employment values as well as the reported spending pattern results and Labor Income Change results. Please let us know if you have any additional questions.
    0
    Comment actions Permalink
  • Avatar
    BPage
    I have a couple of questions loosely related to this thread: 1) When an analysis is done in parts, as described below, and the LPPs are retained as SAM model values is it correct that the model accounts for leakages out of the model geography (e.g., some of the spending from employees at local businesses where the direct expenditures take place would happen out of the model geography) but does not account for "leakages" from surrounding areas into the model geography. . . i.e., some of the direct spending would take place at businesses outside of the model geography, and some of the workers at those businesses may actually live and spend money within the model geography, but impacts associated with that induced spending would not be accounted for in the impact results. 2) I have noticed over the course of past analyses that the characteristics (output per worker, employee comp per worker, value added per worker. . . ) for certain industries are the same for a zip-code defined geography and the county in which those zip codes are located. Is this typical? Thank you!
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hi Britt, 1) That is correct - in a single-region analysis, the model does not capture feedback effects. If you have an estimate of what that returning value is, you can certainly model that in a separate activity. Otherwise, we would suggest Multi-Regional I-O (MRIO). In a typical MRIO analysis, feedback effects from any linked model will be 'recaptured' (as will the indirect and induced effects in the linked region). Unfortunately, however, this is not the case with Industry Spending Patterns (ISPs); while the indirect and induced effects in the linked region will be captured, the feedback effects to the original region will not. A fairly simple work-around is to export the ISP to Excel (Activity Options > Export > To Excel). Change all the commodity codes to industry codes by subtracting 3000 from each of them. Then copy and paste the data into the "Industry Change" tab of the Activity Template Excel file in the Implan User Data\Templates folder of your appliance. The only columns that need to be filled in are Sector, Event Value, Event Year, Retail (which should be all "No" since ISPs are already margined), and Local Direct Purchase. Be sure to also give the Activity a Name, Activity Level, and Activity Year at the top. Then, back in IMPLAN, import the new Activity (Activity Options > Import > from Excel), select "Industry Change", then run the analysis as MRIO. In this case, all feedback effects are captured since you have modeled the expenditures as a series of Industry Change Events as opposed to an ISP. However, you will now get reported 'direct' effects; these are truly the first round of the indirect effects so you would want to add them to the other indirect effects and state your own direct effects as per usual for ABP. 2) Yes, this is to be expected. The only zip-code specific data we have are for population, land area, and establishment count so we use county-level ratios for estimating most of the other data items. This article discusses data creation for ZIP Codes http://implan.com/v4/index.php?option=com_multicategories&view=article&id=590:590&Itemid=71
    0
    Comment actions Permalink
  • Avatar
    BPage
    Hello, Thanks or your reply. I have another question related to ISP impact modeling. How can we tell what proportion of the total value (value entered as "level" in activity screen) is assumed by the model to be spent within the model geography? Thank you! Britt
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hi Britt, If you need to know how much of the Activity Level is kept locally by industry, this can be calculated by Multiplying the Activity Level by the coefficient and the LPP which should be set to SAM Model Value. If you are looking just for the net value lost to imports from the first round of ISP spending that is captured in the LPP Imports field on the left-hand side bar below the Scenario and Dollar Year for View selections.
    0
    Comment actions Permalink
  • Avatar
    BPage
    Hi- That is correct - I am just interested in a total amount lost to imports, not the value by industry. Sorry, but I'm not seeing the LPP Imports field you're referring to. Should I see it when I'm in Setup/Analyze Activities? Or should I be looking somewhere under Explore? Thanks.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Please see this under the Scenario Results screen. It is on the left side of the table results as mentioned in this article: http://implan.com/V4/index.php?option=com_multicategories&view=article&id=529:529&Itemid=71#scentool
    0
    Comment actions Permalink
  • Avatar
    BPage
    I understand which screen you're referring to now - thank you. But the LPP Imports value is $0. I am working with an ISP and a geography comprised of several zip codes. Shouldn't there be imports if I set the LPPs in the activity screen to SAM Model Value? Thanks.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hi Britt, We apologize, older versions of the software still count the first round of Indirect industry expenditures as Direct, as do the Institution Spending Patterns, which also report to this field, so we accidentally misdirected you. However, you can still determine this, but with just a little more calculation by means of the [url=http://implan.com/v4/index.php?option=com_multicategories&view=article&id=565:565&Itemid=14#preview]preview window[/url], which is in the Setup Activities window on the right-hand side of the Activity portion of the screen. This screen provides shows the total value applied to the Multiplier and the total value applied in each Event line. However, it also takes deflators into account, so if you are using an Event Year that is different from base data year, there will be some effect of deflation on the value reported in the preview screen. There are a couple of options to compensate for this. You can export this table to excel and then apply the deflators to each item to adjust each line item's dollar value, or you can calculate an average deflator across all industries and apply this value to the total. If you are using an Event Year that is the same as the data year, not adjustments need to be made. Again we apologize for the confusion, and for the fact that this method is more cumbersome than the cleaner LPP Imports screen.
    0
    Comment actions Permalink
  • Avatar
    BPage
    Not a problem! Thanks for the clarification. Last question - Say you are modeling an impact using ISB and you have total annual expenses plus the percent spent within your model geography but no breakdown of expenses. What is the easiest way to maintain the coefficients (stick with model assumptions about distribution of spending across industries), but change the LPP/import assumption (increase or decrease the amount that would actually be spent at businesses within the model geography). Would you use export the ISP, use a ratio approach to change the LPP for every industry, then re-import the changed ISP into the model? Or is there a simpler approach? Thanks very much.
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hi Britt, We apologize, but we just want to clarify before we answer to make sure we understand, and provide the best answer. As we read this, you have two numbers, a total goods and services expenditure value and a value of sales that are known to be to local business, but no way to know which Sectors the local purchases are attributed too, correct? One additional question on this, sometimes when businesses indicate that they make purchases locally, they are actually referencing a local wholesaler or retailer, rather than a local producer, do you know or have any way to to discern, if the locally spent value is to actual producers of the goods and services purchased as opposed to spending at local retail establishments or through local wholesalers? Thanks!
    0
    Comment actions Permalink
  • Avatar
    BPage
    Good morning, Yes, you are interpreting my question correctly. I would actually be interested to know how to handle both situations you mention - i.e., a situation in which you are certain that the lump sum percent spent locally refers to a dollar amount paid to producers, and a situation where percent spent locally refers to a dollar amount spent at a combination of retailers/wholesalers/producers. Thank you! Britt
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hi Britt, There is no single percentage of local purchases – rather, it varies by commodity and region. Although a business may on average purchase 75% of their goods and services locally, it is likely a mix of purchasing some goods and services 90% locally, some 50% locally, etc., depending on the good or service. IMPLAN’s RPCs make use of local trade data, taking into account the local demand for goods and services and the supply of those goods and services locally and in neighboring counties. Thus, unless you have information as to exactly which goods and services are purchased at the 75% local rate, we recommend leaving LPP as is. That being said, if you want to get a sense of what the average LPP for your impact is, you could calculate a weighted average LPP - weighted by the size of each purchase: 1. Export the activity to Excel 2. Make sure the coefficients sum to 1.00. If they don’t, divide each coefficient by the sum of coefficients – they should now sum to 1.00 3. Calculate a “weighted LPP” for each purchase by multiplying the coefficient by its LPP. 4. Take the average of these weighted LPPs. In the example below, this business buys, on average, 19.28% of its inputs locally: Sector Coefficient LPP Weighted LPP 1 0.2 100 20 2 0.1 0 0 3 0.5 73.69 36.845 4 0.3 67.64 20.292 19.28425 Weighted average LPP Ideally, this value will be in the ballpark of your figure. However, if it is significantly lower than your expectations, this would indicate to me that a lot of the business’ purchase are made from a local wholesaler or retailer. So their figure would represent the % of their purchases that take place locally (in purchaser prices, not accounting for margins), while the calculated value represents the % of local purchases (in producer prices, accounting for margins).
    0
    Comment actions Permalink
  • Avatar
    BPage
    Hello again, We tried the "weighted average LPP" approach you suggested but the resulting figure (less than one half of a percent) doesn't seem to make sense, even if the figure we have for "percent spent locally" includes purchases at retailers and wholesalers as well as producers. Going back to some of the earlier posts in this thread (post #14924 and #14940), if we are simply looking to find out what % of our Activity Level IMPLAN assumes will be spent locally, can we not just use the figure available through the final demand preview window? That is, can we divide the figure from the final demand preview window into the Activity Level amount to get the overall percent that IMPLAN assumes will be spent locally? Thanks very much, Britt
    0
    Comment actions Permalink
  • Avatar
    ljvoves
    Hi Britt, So long as you are not using deflators (i.e., so long as your Event year is the same as the model year) then this would be fine. If your expenditure value is in a different year than the model year (i.e., you are using deflators), then you will need to deflate/inflate that value before you do your percentage calculation. This is because the Preview button applies both LPPs and Deflators to the Event values. Keep in mind that ISPs are already margined (i.e., they are in producer prices) so the % local that you calculate may be lower than the % you might be expecting if your expected % local refers to % local purchases (non-margined, purchaser prices) as opposed to % local production (margined, producer prices).
    0
    Comment actions Permalink
  • Avatar
    IMPLAN Support
    Hi Britt, So long as you are not using deflators (i.e., so long as your Event year is the same as the model year) then this would be fine. If your expenditure value is in a different year than the model year (i.e., you are using deflators), then you will need to deflate/inflate that value before you do your percentage calculation. This is because the Preview button applies both LPPs and Deflators to the Event values. Keep in mind that ISPs are already margined (i.e., they are in producer prices) so the % local that you calculate may be lower than the % you might be expecting if your expected % local refers to % local purchases (non-margined, purchaser prices) as opposed to % local production (margined, producer prices).
    0
    Comment actions Permalink

Please sign in to leave a comment.