Differences between Model Years


1 comment

  • Avatar
    IMPLAN Support
    Hi James, Thank you for your post. Here is some information that may be helpful: The indirect and induced multipliers will definitely vary year to year based on the local availability of the inputs required (i.e., the RPCs), the IE-to-Output ratio, and the LI-to-Output ratio, where IE refers to Intermediate Expenditures and LI refers to Labor Income. Mesa County's IE-to-Output ratio went down very slightly and its LI-to-Output ratio increased quite a bit, both of which are reflected in the lower indirect multiplier and the higher induced multiplier. However, keep in mind that multipliers include all rounds of impact so are also influenced by changes in other sectors' ratios as well! As you noted, the largest change is in the direct Employment Effect, and this involves this sector only so is easier to trace down the cause. The largest factor in this case appears to be the increase in Employment per dollar of Output for sector 11 between 2009 and 2011. Both Employment and Output increased over the time period but Employment increased at a much faster rate than Output (Employment increased by 213%, while Output increased by 41%). Next I will describe how we estimate Employment and Output for this sector: Output: 1. From the USDA’s ERS we get Annual Cash Receipts (our proxy for Output) at the state level. These data only include sales to final demand (not to intermediate demand), so there is no double-counting! 2. These state values are then distributed to the counties by using the ratio of county physical production to state physical production from the latest Ag Census (which comes out every 5 years). For example, if County A has 10% of the state’s corn acres, then it gets 10% of the state’s Annual Cash Receipts value for corn. Since the distribution of the state values to the counties remained the same between 2009 and 2012 (since both years used the 2007 Ag Census), then it must be the state value that increased: 2009 ERS Sales of Cattle and Calves in Colorado: $2,605,779,000 2012 ERS Sales of Cattle and Calves in Colorado: $3,663,529,000 This is a 41% increase so the output value looks good! Employment: The agriculture sectors are particularly difficult to estimate because there are no employment and earnings data collected on a commodity basis, even at the national level. The BEA’s REA data (one of our main source of employment and income data for most sectors) have county-level employment and income data, but these are farm totals that are not broken down by type of agricultural commodity. While BLS' CEW (our other main source of employment and income data for most sectors) publishes some data by agricultural commodity, we do not use them as they only cover about 90% of the REA value for wage and salary farm employment. As a result, IMPLAN developed procedures to estimate employment and income by commodity and county. These estimates of employment and income are then used to distribute the total farm employment value given by the REA data. We use a combination of Ag Census farm counts by commodity (as an indication of proprietors) and employee compensation-to-output relationships from the BEA Benchmark I-O (which have the commodity detail) to get a first estimate of employment by commodity. These are controlled to U.S. REA “Farm Total” numbers for the current data year. The resulting U.S. relationships to output are then applied to state Output values (estimated as described above) to derive state employment and income numbers which are then controlled to each state’s REA “Farm Total”. County numbers are then based on state ratios. When using the Output value and BM ratios, the Employment estimate also grows by 41%, as would be expected from using the same ratios applied to an Output number that grew by 41%! However, bear in mind that these estimates must then a) be controlled to State totals (which themselves must first be controlled to U.S. totals), b) controlled to REA 'Farm Total', then c) re-controlled to state and U.S. Thus, I would not be able to 'prove' the number without looking at all the other farm sectors, all the other counties in the state, and all the other states in the U.S. Since the agriculture data are, to a large extent, derived, analysts with local agriculture data are encouraged to use it when building their IMPLAN models. Thanks!
    Comment actions Permalink

Please sign in to leave a comment.