Induced Spending Discrepancy
I am analyzing the IMPLAN Output based on annual new spending. For the previous year, I ran spending numbers using 2013 State Data and got the following Output:
Direct: $42 M
Indirect: $13 M
Induced: $15 M
Induced/Direct Ratio: 0.35
I ran [b]different but proportional[/b] spending numbers for this year using 2014 State Data and got the following Output:
Direct: $55 M
Indirect: $16 M
Induced: $24 M
Induced/Direct Ratio: 0.44
The Direct Output increased by 30%, but somehow the Induced Output increased by 60%. I expected a more direct relationship between different years of data for the same geographic area. Out of curiosity, I ran [b]the previous year's spending numbers[/b] in a model using 2014 State Data and got the following:
Direct: $44 M
Indirect: $13 M
Induced: $19.5 M
Induced/Direct Ratio: 0.44
It looks like the Induced Output is generally much higher with any inputs using 2014 State Data. What might explain such an increase in the Induced Output from year to year based on equivalent spending values? I would greatly appreciate any explanation or insight on what could cause different Induced Output.
-
Hi! I am more than happy to provide additional details. The two data sets I am using are 2013 State of Arizona data and 2014 State of Arizona data. I decided to just attach the two input files I used. If you would like more information, please let me know and I will do my best to provide it. To be precise, I used SAM Model Values for the following sectors: 402 412 414 442 457 493 496 512 Thank you so much for your assistance! -
Hi Alex, Thank you for the additional information. Would you mind sending to us your 2013 ODF Data File and Trade Flow file to info@implan.com This is so that we can recreate the results that you are seeing. Thanks! -
I emailed the two requested files at info@implan.com. Hope it helps! -
Hi Alex! I apologize, I did not receive your email would you mind resending your email to a different address: implangroup@implan.com I apologize again for the inconvenience. Thanks! IMPLAN Staff! -
Hello, Dropbox will work great! Could you email the link to support@implan.com? Also, implangroup@gmail.com is another address to try. -
Hello, We checked the model set-up and all appears fine. We see now that while the induced/total stayed quite stable, the induced/direct fell for Emp and LI and rose for VA and Output. There can be many reasons for such changes over time, some of which are discussed in these forum discussions: http://www.implan.com/index.php?option=com_kunena&view=topic&catid=80&id=18852&Itemid=1840#19203 http://www.implan.com/index.php?option=com_kunena&view=topic&catid=84&id=17621&Itemid=1840#17623 Since induced effects stem from Labor Income, the first thing to do is look for any substantial changes in the LI/Output of the directly impacted sectors. Indeed, we see that it rose substantially for several of the directly-impacted sectors, most notably sectors 503, 524, and 512. This will generate higher induced expenditures stemming from these sectors, and thus higher induced Output. [ul] [li]In the cases of sector 503 and 512, there are two main causes for the large jump in LI/Output between the two years: 1) in 2014 this sector experienced negative OPI, which reduced its Output relative to the other components and 2) we incorporated an enhanced methodology for estimating Proprietors and Proprietor Income in 2014, which resulted in a marked increase in Proprietor Income for this sector. Regarding the new methodology adopted In 2014, we developed a method to estimate W&S Employment separately from Proprietor Employment (which will allow for better wage per worker interpretations and will allow us to split the State/Local Government tax impacts into further government level detail, which is currently in development). As a part of this process, we incorporated some new data sources (e.g., Census Non-Employer Statistics and CBP Organizations with Employees by Ownership Type) and more involved processes for estimating the proprietor count. These changes are therefore an improvement over past data. While we try to maintain a certain level of consistency over time, we also try to improve our estimates when possible.[/li] [li]Sector 524 is a local government enterprise that is typically subsidized (i.e., it typically has negative TOPI and OPI because ticket sales do not typically cover the cost of public transit operations) and therefore is subject to large shifts year to year.[/li] [/ul] The fact that induced Employment and LI per direct fell indicates to me that some of the sectors from which households make purchases had lower Employment/Output and lower LI/Output, both of which are generally expected trends over time as productivity tends to rise over time in most sectors and most years (exceptions always occur). In sum, we are not alarmed by these differences and we hope this additional information is helpful. Best regards, IMPLAN Group Staff -
Hi, I would like to follow up on this topic and ask once more about this discrepancy. The previous year, I used 2014 data and got the following output: Direct: $55M Indirect: $16M Induced: $24M Total: $95M Induced/Direct Ratio: 0.44 I ordered 2015 IMPLAN data for the same geographic area and analyzed IMPLAN Output based on annual new spending. The direct spending numbers I put in are actually smaller in total than the previous year, but I received the following output: Direct: $50.4M Indirect: $17.9M Induced: $29.2M Total: $97.5M Induced/Direct Ratio: 0.58 Even though these inputs are less than the previous year's inputs by approximately 10%, the Total Output is still greater. This is due to the large difference in the induced/direct spending ratio. Last year, I asked a similar question and learned that induced impacts stem from Labor Income, and the LI/Output of directly affected sectors would likely create such an effect. There were two specific causes for this jump in LI/Output for the largest sectors (negative OPI, and enhanced methodology for Proprietors and Proprietor Income). However, when I look at the directly affected sectors, it seems like the LI/Output ratio generally either remained constant from 2014 to 2015, or even slightly decreased. If someone could help me look into the data and help understand the cause behind a second increase in the Induced/Direct ratio, I would greatly appreciate it. I am more than happy to supply IMPLAN with any additional details of my analysis (Study Region, Sectors, Data Files, Trade Flow files, etc.). -
Hi Alex! Thank you for your post. We are more than happy to review your request. To do so, would you mind sending us your Model. You can attach it to your forum post or email it to: support@implan.com Did you also want us to take a look at the same Sectors as last year: 503, 524, 512 or something different? Thanks!
Please sign in to leave a comment.
Comments
12 comments