Analyze Rice Milling as a Separate Industry

That is true of analysis by parts. You have to add the direct effects (employment, income and output) of the rice mill to the IMPLAN results. Note that there is a database (dummy model) that contains the 509 sets of industry production functions which can be imported and are usable in the current models which can supplement the data you have.

Doug, Thanks. I found the dummy model - didn't know that existed. That's helpful. However, I have tried exporting industries 48 and 49 from the 509 activities to the activity library, but when I try to import into the new model, there are no activities in the activity library. I have the same issue when trying to export custom production functions. Nothing is available in the production function library after I export. Ideas what I am doing wrong? My software is up to date. Kat

Hi Kat, The necessary steps are as follows: 1. In the Setup Activities screen, click Activity Options > Import > From Another Model 2. Navigate to the Impland3 Appliance > Implan User Data > Utilities folder 3. Double-click on the file "SpendingPatterns_for_Operations_by_509Industries_MIGActLib440.impdb" file. 4. Select the industry of interest then click Import. 5. Set the Activity Level to the entire industry sales value (or total budget including payroll) via the Edit Activity button. Note that the Sum of Event Values for the spending pattern will be less than one because it does not include value-added, so you will need to run a separate Labor Income Change Activity to get the induced effects.

Jenny, Thanks. That works for the activity import but there is no option to import production functions from another model - only import from the production function library. How do I get a production function into the library?

Hi Kat, I'm sorry, I misunderstood what you were trying to do. When you export the industry spending pattern to the Activity Library, this is different than and separate from exporting a production function to the Library of production fuctions. Since the software treats the industry spending pattern as an Activity rather than a production function, it goes to a separate library when exported and it cannot be imported as a production function. You could create the production function manually, but this is unneccessary for impact analysis. Instead, you could use the analysis-by-parts approach for which there is a tutorial (http://implan.com/V4/index.php?option=com_docman&task=doc_download&gid=108&Itemid=7) and numerous forum discussions. In short: a. Run the industry spending pattern for the indirect effects. b. Run a Labor Income Change Activity for the induced effects. c. The rice sector's sales and associated employment and labor income are the direct effects.

Thanks Jenny. I think I am clear on how to use the industry spending pattern and I will definitely look at the ABP methods for analyzing rice milling, but I guess the second part of my question was more of a technical, rather than analytical, question. If I customize a production function (under "customize > industry production") for sector 43 in one model and then do "options > export > save," I thought this custom production function should be available to use in other models. However, when I open a new model and try to import the customized sector 43, there are no production functions in the library. I realize this is a separate issue from importing the activities from another model, but I just noticed that neither importing activities nor production functions works when I try to use the activity or production function libraries.

Hi Kat, This appears to be a bug with the software - we have been able to export and import production fuctions but usually only after 2 or 3 attempts. I would suggest closing your models and then reopening them and trying again. We have alerted our technical team to this issue. Also note that the models need to be of the same model year and they cannot be aggregated.

Doug and Jenny, First, I think this subject should be in the analysis section, but I don't know how to move it. Sorry. Thanks for your suggestion to use the ABP method with the imported industry spending patterns from the 509 dummy model. I am using this method. Did you know that the tutorial you link to for analysis-by-parts is using IMPLAN V2, not V3? Is there any tutorial that uses V3 software? Anyway, I have been working on the analysis and have had some questions arise. Just to be clear, the final goal of this analysis is to measure the contribution of rice milling separately from flour milling within the larger contribution of agriculture (defined by us as sectors 1-19, 41-112, 295-297, 300-301). Here are my methods for the contribution/ABP econometric RPC model with questions sprinkled in: 1. I customized the employment for industry 43 to remove the number of direct jobs in rice milling. I also customized output and value added for sectors 1-14 according to NASS data. For sector 13, I also use an updated employment number. To customize data, first uncheck box under "Lock," change "edit options" to "Edit totals then update per worker values," replace data, then click "Update." Then model must be reconstructed through multipliers. Questions: Why does the industry spending pattern for 49-Flour Milling from the 509 dummy model have no event for 3002- Grains? What is the relationship between coefficients in events in an industry spending pattern and the gross and regional absorption coefficients? According to the V3 manual, the absorption coefficients are "Coefficients are equivalent to the fraction of each dollar of outlay that is spent on purchasing the listed Commodity." Does outlay include industry sales and payroll? As I understand it, the coefficients in the event would be the gross absorption coefficients and 1 - sum of event coefficients is the value added component. If correct, could I use the coefficients from the Flour Milling industry spending pattern as gross absorption coefficients in the production function since that is newer data (2007 vs. 2004)? 2. Imported industry spending pattern for 49- Rice Milling (old 509 scheme number) for analysis by parts. 3. Set activity level to industry sales value (output). 4. Set all events to study year (2010). 5. Set LPP for ag sectors in events to 100% and left other LPP's at default values. This step assumes all ag products are purchased locally, which is a constraint of our study, so this step is not mandatory: leaving all sectors at default values is also an option for other ABP. Questions: What are the default LPP values? I don't think they are the SAM values. Also, in Jenny's post #12273, you advised someone to normalize their events so the coefficients summed to 1, but did not suggest I do that. Jenny said, "At this point, the coefficients sum to 1.00 and represent the proportion of the non-payroll budget that is spent on each input." When should I normalize and when shouldn't I? It seems that normalizing would remove the value added component. 6. Create rice milling labor income change activity and set the event's labor income value to that of employee compensation plus proprietor income. 7. Create Industry Change activities for my ag sectors for contribution analysis and set all these events to year 2010 and LPP=100%. 8. Set commodity byproducts except for primary commodity (i.e., sector 70 only produces commodity 3070) for all my ag sectors (1-19,41-112, 295-297, 300-301) to zero, save, and reconstruct through multipliers. 8. Set all ag sector RPC's to zero, save, and reconstruct through multipliers. Steps 10-11 per http://implan.com/V4/index.php?option=com_multicategories&view=article&id=618:618&Itemid=14 good info on this page! 10. Run rice milling industry spending pattern activity and labor income change as a scenario1 (level=1) to get indirect and induced effects of rice milling. 11. Create and run scenario2 for all other ag sectors (level=1) and run to get all other direct, indirect, and induced effects of agriculture. 12. Manually add direct, indirect, and induced effects for rice milling that I have collected to agriculture direct effects totals. Question: Am I allowed to run the two scenarios (rice milling and ag contribution) as one scenario so I would only have to manually add in direct effects of rice milling? It seems that would be fine to me. Thanks so much for reviewing these methods for me. I will edit these steps according to your suggestions so it is accurate for other users to follow.

For some reason, when you import a 509 industry spending pattern, the Events are not automatically sorted by Sector. Try clicking on the “Sector” heading and you should see commodity 3002 (Grains) at the top. After removing the rice employment from sector 43, did you click “Update” so that the associated Output, Employee Compensation, Proprietor Income, Other Property Income, Indirect Business Taxes, and Intermediate Expenditures were also removed? You must do this to keep the industry relationships (output per worker, income per worker, etc.). As an alternative, you can skip the step of even removing the employment in the first place. It isn’t the levels that are important for impact analysis but the ratios. Step 5 assumes that all ag product purchases by the rice milling sector are made locally – just wanted to double-check that you feel this is accurate. Normalizing needs to be done if you’ve edited any of the coefficients but you want the same proportion of IE to VA. For example, if you increase one of the coefficients but do not normalize, then you are saying that not only does the sector spend a greater proportion of its IE on that commodity, but also that it spends more on IE relative to VA. If you normalize you must then set the Activity Level to just the non-payroll portion of the budget – otherwise you would spending the entire budget on IE. If you are not customizing any of the coefficients you do not need to normalize and you want to set Activity Level to the entire budget since only the proportion of the budget represented by the Sum of Event Values will be spent on those items. In either case, a separate Labor Income Change Activity is needed. Just to be clear on step 7, you don’t want to zero out all by-products, just those other than the primary commodity. You can certainly run these together as one Scenario; however, you would not be able to see the contribution of the rice milling sector separately in this case. Just curious – any particular reason you are using the econometric RPC method?

Jenny, Thanks! You've been so helpful. I feel much better about this analysis now. I found the grains in the list of events, and am laughing at myself for that one! After editing data for sector 43, yes, I updated the other values. I just forgot to put in the methods. Can you explain a little further why you say that is not necessary. I thought not editing the sector after moving jobs out of 43 and into the industry spending pattern for rice milling would end up double counting the rice milling jobs in final contributions, which would in turn also inflate LI, VA, etc. Step 5 is an assumption we make in the study across the board for all industries, so yes, making only local purchases is what I was intending. Can you address what the default LPPs are in the imported industry spending pattern? (See previous question under step 5) I will clarify step 7. I understand what you mean. We use the econometric RPC method because I inherited most of these methods from previous years. Each year we do this study, and using the same model makes time series data more comparable. Do you suggest we use trade flows methods? If so, will you explain why trade flow would be better? Thanks for your time.

Hi Kat, It seems counterintuitive, but a sector’s multiplier does not depend on the sector’s overall size; just on the relationships (e.g., output per worker) of that industry. For example, suppose the output and employment for the grain milling sector were $1 million and 100, respectively. The output per worker is then $10,000. Then, when you run a $1 million impact on the sector, you will get 100 workers. If, on the other hand, you reduced the grain milling sector’s output to $500,000 and rebalanced, the employment would be reduced to 50 and the output per worker remains the same at $10,000. So if you now run a $1 million impact on the sector, you will still get 100 workers. However, since you are doing a contribution analysis, I assume you are specifying the entire industry output amount as the impact and in that case, yes – you want to specify the reduced amount. So you are probably better off reducing the study area data for the industry just so that it matches your direct impacts and makes it clearer to understand. It was really just an unnecessary parenthetical comment – I apologize for adding confusion. The default LPPs are the RPCs for each of those commodities, which represent the percentage of local demand for the commodity that is met by local suppliers. It is an average for the whole region so in effect you are saying "We don't know how much of this commodity is purchased locally by the millers, so our best guess is the regional average for this commodity." Without any additional information, this is definitely the recommended approach. We suspected that that might have been the reason for using econometric RPC and that is fine. If a user is not trying to replicate a previous study, we typically recommend the trade flow method because it is newer and allows for multi-regional I-O (MRIO) analysis. However, for your purposes, the econometric method is fine.

Jenny, I have been unable to collect intermediate expenditures and proprietor income for the rice milling industry, so I can not calculate direct value added. What I do have is employment, employee compensation, and industry sales. I am wondering if it would be acceptable to get the IMPLAN program to estimate the missing direct effects for me before doing the ABP. This was my idea: edit "Industry Production" for sector 43 to reflect the rice milling industry spending pattern from the 509 sectors dummy model, input the values I know in "Customize > Study Area Data", and then "update" the sector and use those estimates. For my final results, I need to be able to present direct effects for rice milling for employment, employee compensation, labor income, and total value added. Any other ideas to estimate the missing values would be appreciated as well. Thank you.

Hi Kat, That is exactly how I would have recommended doing it.

I think you just made my day when you said "That is exactly how I would have recommended doing it." But it's not working as I expected. Just to be clear: the event values in an industry spending pattern can be used as regional absorption coefficients in the "Customize > Industry Production," correct? When I use this new industry production, the total absorption value and value added coefficients for the sector are changed also (I used Doug's methods from post #12643), but the value added for that sector in the Study Area Data is not changed. I thought that editing the production function data would change the default study area data since the relationship between absorption and value added coefficients was changed. I really want a method that would force IMPLAN to change the values of the components of value added instead of keep the default ones (the ones I have not been able to collect: proprietor income, other property type income, and indirect business tax). Also, should I "Lock" the relationship between TVA and IE or not? I think not. Thanks for any suggestions you can provide.

Yes - the Event Values in an Industry Spending Pattern are the same as the Coefficients in a production function (so long as you have not normalized the events in the Industry Spending Pattern). If you are editing a production function via Customize > Industry Production, the software makes you click the Balance button before you can hit the Save button. By re-balancing, the Total Absorption Value stays the same. If you want to change the Total Absorption Value, you will need to change Intermediate Expenditures via Customize > Study Area Data. You will not be able to directly specify IE on this screen but IE will change if you change Output or if you uncheck the Lock box and change any of the value-added values. If, instead, you are making the changes via Access, I believe you will need to change the Intermediate/Value-Added ratio via Customzize > Study Area Data. Once you uncheck the Lock box and you can adjust Intermediate Expenditures.

Thanks for the tip on editing industry production but not balancing in the software. I found it easier to edit those in Access due to how many I needed to change. I understand I can change the ratio of IE to VA in the software if I uncheck the lock box. When I do this, IMPLAN changes the value of Other Property Type Income to compensate (keep TVA the same), but leaves proprietor income and IBT unchanged. Why does it make the changes to OPTI only? To clarify, here are the data I have (made up numbers to avoid data disclosure) and what I would like IMPLAN to estimate since I have no data (using the new production function): employment= 1000 employee compensation= $80,000,000 output= 1,800,000,000 OPTI= no data PI= no data IBT= no data IE= have production function, total abs.= 0.88, VA coeff= 0.12 (can I read this as for every dollar of output, there are 88 cents of IE and 12 cents of VA? if so, I think this might solve my problems, as I can calculate IE and VA) TVA= no data I think I am mostly unclear on why changing the production function doesn't change the default values for the components of value added since I have changed the composition of IE and the relationship between IE and VA (correct?). I'm sorry, this is really confusing to me. I know there is a Regional Value Added Coefficients table another poster mentioned (in the same thread I got Doug's instructions on editing the production function). Should I be changing these to get different default data? If so, is there somewhere I can get more information on the Regional Value Added Coefficients? Thanks for your patience.

The reason the software uses OPTI as the balancer is because OPTI is treated as a leakage - i.e., it generates no additional indirect or induced impacts - so its value is not important in terms of running analyses. If you uncheck the Lock box, Intermediate Expenditures will adjust when you change any of the Value-Added components. Intermediate Expenditures will also adjust (and Value-Added will not) when you change Output. Yes - you can simply multiply your output value by 0.88 to get direct Intermediate Expenditures (IE) and by 0.12 to get direct VA. If you want values for direct PI, OPTI, and IBT, I would just use the existing ratios for sector 43 - for example, if PI makes up 30% of sector 43's VA, then multiply your direct VA figure by 0.3 to get direct PI. The regional absorption coefficients table can be found by going to Explore > Social Accounts > Balance Sheets tab > Commodity Demand tab and then selecting the sector of interest from the drop-down menu. In that table, you will see Gross Absorption - these coefficients (listed here in percentage format) match those that are in the Customize > Industry Production screen as well as those in the Industry Spending Pattern. The Regional Absorption is found by multiplying the Gross Absorption by the RPC. Gross Absorption is what they buy per $1 of output, while Regional Absorption is what they buy [u]locally[/u] per $1 of output. When you run an Industry Spending Pattern or an Industry Change Activity, the software takes care of that multiplication and then applies the results to the multipliers. There is no place to change these coefficients other than via Customize > Industry Production or the method Doug proposed using Access. When you do this, does your Total Absorption value match what you specified in the Access table?

[quote="Jenny Thorvaldson" post=12893] Yes - you can simply multiply your output value by 0.88 to get direct Intermediate Expenditures (IE) and by 0.12 to get direct VA. If you want values for direct PI, OPTI, and IBT, I would just use the existing ratios for sector 43 - for example, if PI makes up 30% of sector 43's VA, then multiply your direct VA figure by 0.3 to get direct PI. [/quote] Thanks for this suggestion. I will use this method, but one question: since I was able to collect a reported number for employee compensation, I would like to use it. So, should I use the existing ratio for LI to VA and make PI equal the remainder of LI not covered by EC? If not, my new numbers would not match the original ratios if using the EC I collected. Which is more accurate? [quote="Jenny Thorvaldson" post=12893] The regional absorption coefficients table can be found by going to Explore > Social Accounts > Balance Sheets tab > Commodity Demand tab and then selecting the sector of interest from the drop-down menu. In that table, you will see Gross Absorption - these coefficients (listed here in percentage format) match those that are in the Customize > Industry Production screen as well as those in the Industry Spending Pattern. The Regional Absorption is found by multiplying the Gross Absorption by the RPC. Gross Absorption is what they buy per $1 of output, while Regional Absorption is what they buy [u]locally[/u] per $1 of output. When you run an Industry Spending Pattern or an Industry Change Activity, the software takes care of that multiplication and then applies the results to the multipliers. There is no place to change these coefficients other than via Customize > Industry Production or the method Doug proposed using Access. When you do this, does your Total Absorption value match what you specified in the Access table?[/quote] OK, the Gross Absorption listed in the commodity demand table is the same as the number in the Industry Production window and the same as the number I entered in the Regional Absorption table in Access. Is that correct? I did notice that in the commodity demand tab, the Regional Absorption listed is NOT the Regional Absorption I entered in Access. It seems that the table in Access should have been named "Gross Absorption"? Remember that also, for our contribution analysis, we set up the industry change activities so that all ag inputs are purchased locally. Does that information matter for this discussion of the default industry components of value added? Thanks for your continued support for this analysis.

Yes, I would use your EC figure and give any leftover LI to PI. Local knowledge always trumps estimated values. While the table is named "RegionalAbsorption", you will see that the column name is "GrossAbsorptionCoefficient" so the Regional Absorption listed in Commodity Demand table will not match the numbers you enter into Access - rather, they will match the numbers you entered into Access multiplied by the RPC for that commodity (which is also listed in the "RegionalAbsorption" table. Did the values you entered in Access match the Gross Absorption values in the Commodity Demand table? If you are not changing Output for any sectors, then after making the changes in Access, you can simply choose Options > Construct > Social Accounts > Social Accounts (SAM) and then Options > Construct > Multipliers (do not run Options > Construct > Social Accounts > Regional Absorption or Options > Construct > Social Accounts > All Steps, as this would wipe out any changes you made). If, on the other hand, you are changing Output, you must follow the procedure shown in post #12643 (especially step 4): http://implan.com/v4/index.php?option=com_kunena&func=view&catid=84&id=441&limit=6&limitstart=6&Itemid=35#12643 When you change IE by changing the GrossAbsorptionCoefficients in the Access table, the software will use those coefficient values but will not adjust total Value-Added to balance with the new sum of the GrossAbsorptionCoefficients - you will need to change Value-Added manually. The software will never automatically change VA to adjust to changes in IE. No, the LPP edits should not affect the value-added.

Jenny, Earlier you suggested that we could use the sector 43 TVA ratios from 2010 to calculate estimates for rice milling's OPTI, PI, and IBT. We have data for 2006, the last year of the 509 scheme, so we also have the 2006 ratios for rice milling's value added components. Which, in your opinion, is the better option between using the new 2010 ratios for sector 43 or using the rice milling ratios from 2006? If we use 2006, we could also use the industry production to create an industry spending pattern to be consistent as to what year we base our estimates. Using sector 43 ratios seems attractive since the estimates are newer, but less accurate since it also includes flour milling and malt manufacturing components. Thanks for your input.

Just checked the Census Annual Survey of Manufactures for 2010, and the data does not separate out Rice Milling. My preference would be to use the more current data and live with the mix with flour and malt. If you are working with the industry, they might be willing to correct the implied values you derived from the data.

Can you think of a rational explanation of why my direct effects are smaller but indirect and induced effects are larger than another similar study? I understand why the discrepancy in direct effects: different source data. However, my intuition says that a smaller direct effect could not result in increased indirect and induced effects. Also, can you run duplicate analyses to confirm my methods? Please let me know how to proceed if so. Thanks so much.

You don't mention if you are referring to output or employment impacts which can make a difference. Intermediate impacts (output - value added) drive the indirect effects. Labor Income (employee compensation plus proprietor income) drive the induced effects. Indirect business taxes and other property income are treated as leakages and cause no further effects. So if the output is less but other property income is significantly less, then you may have a lower direct output impact yet higher indirect and induced impacts. If you are seeing these differences in the employment impact, then a higher output per worker would explain the lower direct and higher indirect and induced (more goods and services purchased per worker). This is all assuming you are using the same region, the same intermediate production function and same trade flow RPCs in both cases. Any difference in any one of the three could also cause your differences.

In the ABP methods you have taught me so far, we use an industry spending pattern and a labor income change activity to calculate the indirect and induced effects. I have attached some data given to me that includes a custom industry spending pattern (not one of those specified in the 2007 dummy model) and household income information (see specifically tab "Mill Data" cells A64:F78). Is there a way to use this information to do the ABP? It seems that I would use the industry spending pattern specified in the spreadsheet (except that they include payroll as part of the ISP, which I have not been doing) and then somehow use the household income change activity to estimate labor income. Can you explain how to use an ISP that includes a portion devoted to payroll and how to use a the household income data to estimate labor income in an analysis by parts? Thank you for being so helpful and a vast wealth of knowledge on the subject!

I'm unclear as to which mill(s) the data in the spreadsheet are from and which mill(s) you are trying to model. Is this U.S. average data and you are trying to apply it to a specific mill? Any details you can provide about your study would help us better help you. Hopefully the following information will help; don't hesitate to follow up with additional questions. It was my understanding that you already had a known Employee Compensation value - this is what you would use for your Labor Income Change Activity. The data in the spreadsheet are much less detailed than IMPLAN's ISPs so I would just use IMPLAN's ISP if I were you, though you could use the information in the spreadsheet to customize IMPLAN's ISP. It is also helpful for determining the total direct Intermediate Expenditures (i.e., the Activity Level to set to a normalized IMPLAN ISP). The proportions in the spreadsheet are analogous to those in an ISP imported from IMPLAN (i.e., they represent proportions of total output). However, if you are using a normalized ISP, then you would need to calculate new proportions from the spreadsheet by dividing each expenditure by just the total intermediate expenditures (i.e., cells B65 through B75). Note that contract labor is not payroll but goes to the ag support sector (sector 18 in the 509 scheme). Suppose you were working solely with the data in the spreadsheet rather than using IMPLAN's ISP. You would: 1. Run $150.88 as a Labor Income Change Activity 2. Create a new Industry Spending Pattern Activity, enter the proportions from cells C65 through C75, then set the Activty Level to $2968.39. (Alternatively, you could create an Industry Change Activity and enter each of the values in cells B65 through B75 as the Industry Sales values for separate Events). 3. Report cell B77 as direct output, cell B64 as direct EC, cell B75 as direct OPI. I don't see a line item for direct IBT (i.e., taxes on production).

Jenny, This data is not for my study, but for another study whose methods we are trying to replicate (with our data we have collected). I do have an employee compensation value, but this study said that "Employee Compensation was reported as HH Income < $50,000" so there is some link between the labor income change activity I use for ABP and the household pattern in cells E64:F76. They also stated that "Profits were reported as HH Income > $50,000" Can you explain those relationships? The specific mill would be US mills. The results from the study are broken out by state, and we just want to replicate the study modeling only the portion attributable to Arkansas. Thanks for your time!

When they say that "Employee Compensation was reported as HH Income < $50,000", it looks to me like they ran the EC value as a Household Income Chnage Activity for that household income category. If the EC figure is loaded payroll, it needs to be run as a Labor Income Change Activity - in which case you do not need to choose a household income category as it will spread the EC across all household income categories proportional to the number of households in each category in your study area (as reflected in the SAM). See post #12967 in this forum thread for more on a Labor Income Change vs. a Household Income Change: http://implan.com/v4/index.php?option=com_kunena&func=view&catid=83&id=12965&Itemid=35#12967 Similarly, when they say that "Profits were reported as HH Income > $50,000", it looks to me like they ran the profits as a second Household Income Chnage Activity, this time for a higher household income bracket. We don't usually recommend doing this because it assumes that the business owners live within the study area and spend their profits in the same way they spend their labor income. In reality, we usually don't know where the business owners live, let alone where or how they spend/invest their profits. As for cells E64-E75, I really have no idea what they are doing. They are taking the coefficients (the expenditures per dollar of revenue) and multiplying them by 1000000 - I don't know the economic meaning behind that or how it would be used in IMPLAN. The typical approach is to: 1. Either enter the values themselves (cells B65-B75) and leave the Level at 1 or enter the coefficients (cells C65-C75) and set the Level to total output. 2. Run a Labor Income Change Activity with Event value = cell B64 for the induced. 3. Add back in the direct.

In step 2 (of post #13173) you said "Create a new Industry Spending Pattern Activity, enter the proportions from cells C65 through C75, then set the Activty Level to $2968.39." The proportions calculated in the spreadsheet to come up with the value in C77 include the proportion attributable to Total Payroll (C64), and similarly, the "SUM" (which I gather is the absorption coefficient for the industry) reported in B77 includes the value of Total Payroll (B64). Then the Total Value Added Coefficient is the value reported in C76. I am a little confused because since EC is included as a portion of TVA, it can't also count as a portion of the Intermediate Expenditures. So, my issue is with using the C65:C75 proportions in an ISP since those seem to be incorrect by including EC. Or does this have something to do with normalizing the coefficients that I am missing? I do understand what you mean when you explained how to normalize the coefficients if I wanted to, but that would not include payroll (or profits line 76) either, correct? If I try to model EC as HH Income change as they did, with the <$50,000, what do I use in the event values? Is there a good thread you can suggest on HH Income changes as I am not very familiar with this aspect of IMPLAN. Thanks again.

If you are going to set the Activity Level equal to total Output (cell B77), then the proportions in the spending pattern need to be expressed as expenditures per dollar of output, which is exactly as they are in column C. Note that I said to include only those coefficients in celss C65 through C75 - these lines do not include any value added and only sum to 0.85. Thus, the softwared will only spend 85% of the Activity Level (i.e., 85% of the total expenditures from cell B77) on these items. 5% will go to payroll (which you will take care of via the Labor Income Change). The remaining 10% goes to profits (a leakage so no need to model it). If you normalized the ISP (forced the coefficients to sum to 1.00), then in that case you would only set the Activity Level to $2535 (the sum of cells B65 through B75) because the software will "spend" 100% of the Activity Level since the Sum of Event Values is now 1.00. If you are going to use a Household Income Change Activity, then you need to first remove payroll taxes and any net in-commuters. The easiest way to do this is to: 1. Go to Explore > Social Accounts > IxC Social Accounting Matrix tab and sum up the Employee Compensation tab's payments to the Household rows (10001 through 10009). 2. Divide this sum by the Employee Compensation column total. This is the proportion of EC that households get to take home with them. 3. Multiply your known EC value by the proportion from step 2. This is your Household Income figure.

Why is cell B75 included in the ISP but is also reported as OPI, part of value added? To clarify step 3 in the HH income change figure: would I take the figure I calculated and distribute it across the income levels and use activity level = 1? Otherwise, I only have an activity level and no events?