comparing infrastructure impacts

Has IMPLAN prepared anything explaining the reasons for differences between various impact models? This question was spurred by this [url=http://www.politifact.com/ohio/statements/2011/oct/26/steve-latourette/us-rep-steve-latourette-says-infrastructure-invest/]article [/url] which suggested that the job impacts of $1B infrastructure investment (in 2007) was 27,800 according to the Council of Economic Advisors. According to my 2015 National model (w/endogenized state/local spending, excl. investment) a generic $1B 'infrastructure' spend in Sector 58 is around 18,000, which is consistent with the findings of UMass researchers (as cited in the article). Additionally, I noticed that the State of CA's Department of Transportation generated impact results about 20% higher for general infrastructure than we see using our CA Statewide model. I appreciate your assistance.
Was this post helpful?
0 out of 0 found this helpful

Comments

1 comment

  • Hello Alex, Variances in studies typically stem from differences in data and assumptions. A data year represents a snapshot of the economy as it was in that year. As economies change over time, a study run with a set variables utilizing one year's data will not compute the same set of results using the same set of variables but utilizing a different data year. As assumptions dictate the variables for a study (such as the assumption mentioned in the article that states will kick in a portion of spending as a result of federal spending), two studies utilizing the exact same data set but with different assumptions will also necessarily compute different results. Which institutions are internalized is also another variable that can have a very large impact on the results of a study (as can study area and trade flow method). Two studies using different data sets and different assumptions will only increase the variance further. The above assumes the same modeling system (such as IMPLAN) is used in every case. Variances can also exist between identical studies conducted using different models, as modeling systems have their own assumptions. Additional thoughts: We often see reports that seem a bit on the high; sometimes this can be intentional (especially when used to lobby for funding) and sometimes this can be an unintended result of not fully understanding the IMPLAN system. It’s very difficult to compare studies if the authors do not include methodology sections with analysis details, few of which do, sadly. If you want to compare the studies, it may be worth your time to reach out to the other study author and request more details about the analysis. While sector 58 seems a likely choice for at least some infrastructure spending, the report seems to refer to highways and bridges and such, which would fall under sector 56.  As is, it's hard to know what to compare with without study details.  Jobs (at least in RIMS II and IMPLAN) are a mix of full-time, part-time, and seasonal, and should not be conveyed as permanent, full-time jobs. Regards, IMPLAN Staff
    0

Please sign in to leave a comment.