I have a question about the estimation of RPCs in IMPLANS latest version. A previous study had found that IMPLAN was interpreting distribution centers as production facilities and thus systematically overestimating RPCs. Has this problem been resolved in the latest release? The study, performed in 2002, is titled: IMPLANS Weakest Link: Production Functions or RPCs. Its authors are William F. Lazarus, Diego E. Platas and George W. Morse.
Was this post helpful?
0 out of 0 found this helpful


1 comment

  • Hello George, We suspect the article might be referring to three possible things here: 1) Foreign export data - we are not using Census "Origin of Movement" export data because they do not distinguish between brokers and producers. As a result, exports by certain counties of commodities will vastly exceed production. This is why we use an across the board foreign export ratio - eg, if county x has 10% of the US production of marbles, it gets 10% of the US export of marbles. 2) For our trade flow model, our target miles moved by commodity are based on Census Commodity Flow Survey which did not used to distinguish point of manufacture v. brokerage for production, nor point of consumption v. broker for destination. We had to modify these targets accordingly - but this has always been done. If you are interested in learning more about the Trade Flow Model that has been incorporated in our data starting with the 2007 data year release, you can read more about it [url=https://www.google.com/url?q=https://implan.com/v4/index.php%3Foption%3Dcom_docman%26task%3Ddoc_download%26gid%3D138%26Itemid%3D60&sa=U&ei=06eCUqyJFdO34APbxoCACg&ved=0CAcQFjAA&client=internal-uds-cse&usg=AFQjCNEEeAzHy33UikiBtzvdRkanSLv5yw]here.[/url] 3. This article is old enough that it might be referring to the original IMPLAN produced by the USFS which used Supply-Demand pooling ratios which maximized trade flows. However, that was only true for the mainframe version of the software. When the forest service produced the first stable PC version of the software in 1989, they had moved onto econometric RPCs. While this is not the likely reason, we wanted to at least let you know about this possibility. Our RPCs can be higher than other studies for a given commodity, and/or for a given region, but not for the reason given. Our RPCs are, in general, lower than those derived by location quotients (eg RIMS II).

Please sign in to leave a comment.