How to evaluate water for municipal and industrial use?

Thanks for the reply, Michael!

I think I follow the approach you laid out. That said, I am still wondering if/how this would address the fact that many water services are provided publicly by local government (sector 534) while only a small portion is provided by private entities (sector 49). In general, Reclamation provides water diversion, storage, and transmission services and then sells the water to local water systems for treatment and distribution, and those systems are often public (sector 534) and sometimes private (sector 49). It seems one could perhaps use the approach you laid out and look at both sectors 49 and 534, but I imagine this would require knowing the breakdown between the two, correct? Related, I have looked at using only sector 534 since it seems that in most areas this sector provides more of commodity 3049 than sector 49, but in this case the 534 sector seems too broad and doesn't seem to properly focus on water services. This is what led me to focus on 3049 and using a Commodity Output event, as this seems to incorporate both sectors 49 and 534 while also focusing on water services. Thoughts?

Thanks!

Matt

Matt,

You are certainly correct that the majority of commodity 3049 is going to be produced by sector 534, however commodity 3049 only accounts for about 1/4 of the total commodities produced by sector 534 (US totals). So running any amount of spending through that Industry would effect various other commodities that you are not interested in. 

We have had some back and forth on my team and are kind of at a stalemate. So taking a step back, what exactly are you trying to quantify here? "I am trying to evaluate the contribution of water services provided by the Bureau of Reclamation for municipal and industrial use." is what you stated here in your first post, however we are having a hard time drilling that down to something more specific. Is it a dollar value? Or just more generally the services they provide? Any clarity you can add here would assist us in determining the best practices for setting up your project.

Thanks!

Michael Nealy 

The goal is to quantify the economic contribution of Reclamation activities which support municipal and industrial (M&I) water use. As mentioned, Reclamation generally provides water diversion, storage, and transmission services and then "sells" the water to local water systems for treatment and distribution. The intended value to plug into IMPLAN would reflect the transaction price with the local entities (generally the contract price, or a proxy for the value of the raw water prior to treatment, which generally reflects the cost of Reclamation capital and OM&R expenditures associated with that service) multiplied by the volume of water. That said, this piece is also still an open question as I have come across various studies that aim to look at the "contribution of M&I water" in IMPLAN which take a very different approach than this (e.g. focusing on water as a final G/S as well as industries that use water as an input). However, these approaches seem more akin to quantifying the contribution of "water" itself as they make strong assumptions like assuming all downstream production that uses water is proportionally reduced if water availability is reduced and that there is no substitutability for water etc. As you might imagine, these approaches generate some very LARGE estimates in an "extraction ICA framework" and do not seem appropriate for our purpose. For Reclamation's purpose, we are interested in only the contribution of services that our agency provides.

For a bit more context, I am actually working on a guidance document for the ideal methodology to use to quantify the contribution of Reclamation activities that support M&I water use, so any insights or suggestions you might have are much appreciated (note we also look at agriculture, hydropower, and recreation, but those are not part of this effort and we already have some established methods in these areas). On a related note, it seems that IMPLAN now has environmental data, part of which includes water. It looks like this option is grayed out when I try to explore it, but I am curious if some of that info might be useful to incorporate into the framework? I have also recently started to explore extraction ICA vs base ICA to see if base ICA might be more appropriate when looking at water. As mentioned, any insights you folks might have would be much appreciated. Thanks so much for your time assisting on this!

- Matt

Another way to potentially frame this, is to model the BOR under Institution 11001 - Federal Government NonDefense, as I can see that they have a small market share for commodity 3049. This is where the BoR operations would fall in the data. You could set this up as an Institutional Spending Pattern Event and edit the Spending Pattern to more accurately reflect the commodity purchases supported by your reclamation activities. Your Event Value here would be that same quoted transaction price * volume sold, and your Spending Pattern would likely be made up of a significant amount of Commodity 3049.

Also, on the environmental note: that is a recent addition to the IMPLAN application, and would run about $4500 to add to your package.

Best,

Michael Nealy

Thanks for the idea. I will explore this option and let you know if I have any further questions.

Matt

Hello Again,

Some follow up questions. I compared Commodity Event for 3049 to Institutional Spending Pattern for 11001 with 100% in 3049 and the results are nearly identical, with the same top 5 sectors, but then some minor sector differences beyond that (note I purposely excluded employment and payroll under 11001 since that is generally handled separately). The similarity between these 2 approaches is not too surprising, but I am curious if one is preferable for what I am trying to do? Also, its seems that with both of these approaches there is no restriction on buy backs so I am still looking at impacts, not contributions, which is the goal. Is there a way to use either of these event types to look at contributions? As you know, I cannot use ICA event type with 3049 or 11001 so I am at a bit of a loss...

Thanks,

Matt

Good afternoon Matt!

Institutional Spending Pattern Events are going to be treated as if you ran a series of Commodity Output Events for the Commodities that are included in your modeled Spending Pattern, which is because Institutions are Final Demanders in IMPLAN. So as you have noted, the Results from a Commodity Output Event for 3049 would likely be extremely similar to an Institutional Spending Pattern Event if Commodity 3049 make's up a sizable portion of the Spending Pattern. 

I would lean towards using a Institutional Spending Pattern Event in your case, simply because that would be the more straight forward approach between the two. Completely rebuilding that Spending Pattern with a single Commodity Output Event for each Commodity demanded by your institution could become tedious/complicated in a hurry. With the Spending Pattern, you can simply add or delete any Commodity or edit it's percentage of total spending all within the same window. 

Unfortunately, neither of these two methods would restrict buybacks to the Industries that produce the Commodities so you would indeed be looking at an Impact and not a Contribution. To fully restrict buybacks to the Industries that produce Commodity 3049, you would have to define an Industry you wish to model your production level through and include an ICA Event for that same Industry in the same group. At this time, there is no established method for conducting a Contribution Analysis for Institutional Spending or Commodity Output in IMPLAN, only for Industries.

Best,

Michael Nealy  

Thanks for the explanation and suggestion. In regards to restricting buy backs, it seems that sectors 49 and 534 are the key sectors that produce commodity 3049, in which case I am wondering if I could run an Institutional Spending Pattern event as suggested, but then also include two ICA events, one for sector 49 and another for sector 534, with $1 as the input value. Would this effectively limit buy backs for sectors 49 and 534 to provide an estimate of contributions instead of impacts?

Thanks!

Matt

That is an approach that I had previously considered for the sake of your analysis. This would effectively restrict buybacks yes, but unfortunately would not work as we intended. Because you are specifying $1 as the Direct Effect in that Industry, that will be exactly what you see in the results, $1 of Direct Effects to that Industry ($0 in Indirect/Induced). Being that those Industries are the primary producers of Commodity 3049, if we include those ICA Events with another Event to model spending on Commodity 3049 we would produce results that are not intuitive. 

For a point of reference - I ran an Institutional Spending Pattern of $1M (100% of spending on Commodity 3049) coupled with an ICA Event in Industries 49 and 534, and got a Direct Output of around $50k.

Best,

Michael Nealy

I see, thanks for testing this and the explanation. Sounds like my best option is an Institutional Spending Pattern event with a caveat that the estimates include buy backs. I will be in touch if any other questions come up related to this, thanks again for your assistance.

-Matt