Scenario Solar Photovoltaic Jobs and Economic Development Impact (JEDI) Model (Text Version)
This is a text version of the video Scenario Solar Photovoltaic Jobs and Economic Development Impact (JEDI) Model.
It is part of the 2013 Solar Technical Assistance Team webinar series. It explains the inputs and outputs of the Scenario JEDI Model and how states and localities can use the model.
>>Courtney Kendall: I would like to introduce our speakers today. Our first speaker is David Keyser. David is a research analyst in the strategic analysis center at the National Renewable Energy Laboratory, or NREL. He manages the jobs and economic development impact suite of models. Prior to joining NREL, David worked as an economist for the state of Colorado, where he was responsible for constructing forecasts, estimating economic indicators, and working with local governments.
Our second speaker today is Barry Friedman with NREL. Barry has over 15 years of experience in the renewable energy and energy efficiency publishing space. Barry conducts research and analysis into costs, markets and policy development, and workforce development in the U.S. and global solar and renewable energy markets. Now I'm going to go ahead and pass the presentation on over to David. David …
>>David Keyser: Thank you, Courtney. Thank you all for having us here today. My name is David Keyser. As Courtney mentioned, we will have time for questions at the end of this. Well, we have about 40 to 45 minutes of material, and then we really encourage you to be thinking about the questions that you have towards the end and then submit questions to us, and we'll do our best to answer them, and if we can't, we'll certainly get back to you about it.
The structure of this presentation today, basically what I want to do is go give you an overview about JEDI. What the model is, what it does, and then talk about some limitations of the model. As with any economic model, there are limitations in ways to interpret results, and give you a little bit of an overview of what results the model produces and what those mean. Then we will get into a lot more information about scenario JEDI and explain how it's different from the other JEDI models and show you some screenshots of the model and how to enter information and use the model.
Why Do Impact Modeling?
Before I get started talking about JEDI or anything about models, I think it is important to ask a question about why we even would do economic impact modeling. It seems really intuitive because we see it all the time. We see it written up in reports and we see it in newspapers and things like that, but really, the way that you're going to be looking at these models and interpreting results and writing them really kind-of changes based on what it is that you're using them for, and there can be a lot of different ways to use a model, maybe some ways that some folks haven't really thought about. One of the most common ways to use these are as a information-providing tool, to inform stakeholders or constituents in the case of government, representing public interest and community development, but another way that these types of models can be used are for local businesses. For example, if a large utility-scale solar installation is going to be installed in rural Colorado, in the San Luis Valley for example, maybe a local business owner might want to know what kind of a labor force is going to be coming into this region in order to install this, and that can be something that's very useful in terms of planning for eventual customers or inventory and things like that, and so there are a lot of different ways that these models can be used for information providing and planning.
Jobs and Economic Development Impacts (JEDI)
The models that I'll be talking about today are the Jobs and Economic Development Impact suite of models, or JEDI suite of models. I'll just head off. First question that a lot of people have, I have no idea where that name came from, and it wasn't from me. So there's that question. JEDI suite of models are freely available tools that NREL offers on its website. Anybody can go on there and download them and use these models. These models are energy-specific, impact estimating tools that contain information that we get from industry experts based on existing projects and developers, and I'll talk more about that in just a minute, and they're models that are really accessible. You can really know a whole lot about a project or you can know a little bit about a project and still get good results. The JEDI suite of models were developed both by NREL and also by Marshall Goldberg of MRG and Associates, and the web URL where you can download the model is at the bottom of your screen. There's also another web-based JEDI PV model. It's just at JEDI.NREL.gov. That's not the scenario model. In order to download the scenario model you have to go through the link that's at the bottom of the screen right now. [Website: www.nrel.gov/analysis/jedi]
JEDI Background
So as I mentioned, JEDI models contain default data that's from industry experts, and the reason why this is basically the way that the model is structured, all JEDI models contain two basic elements. There's the framing of the project, which is done in terms of expenditures, so these expenditures could be things like equipment that you're spending on modules or material that you're purchasing for construction or labor, and then there's the input-output model, and JEDI takes those expenditures and then runs that through the input-output model to calculate economic impact. The expenditure information in JEDI can either be customized if you really know a whole lot about a project, or you can opt to use the defaults, which are based on, as I mentioned, are based on real-world projects or input from industry experts, and we do update those default data periodically, certainly especially with nascent renewable technologies that can change quite a bit over a short period of time, and so we do try to keep current with that. The input-output model that JEDI currently uses is the IMPLAN model. Some of you may be familiar with the IMPLAN model. IMPLAN is one of the most commonly used input-output models in the United States. It's used very commonly by government, academics, private sector, and it's certainly not the only one in the U.S., but it is a very commonly used model.
Input-Output Models
An input-output model is a commonly used way of estimating economic impacts. Basically what an input-output model is input-models have been commonly in use basically since computers were—became widely available. Input-output models provide a snapshot of the relationship between sectors in an economy at a single point in time, and when I say sectors, I'm talking about industries; I'm talking about labor and households. So if you think about it, investments and then returns on investment, and government in terms of collecting taxes and providing services, and then the rest of the world, which shows up as imports and exports. All of these sectors are related in the model based on their inputs and their outputs, as you hear through the name. The inputs that industries use, industries use inputs in order to produce. For example, a sandwich shop might buy agricultural products and then meat. This is what happens when you present over lunchtime. The sandwich shop consumes agricultural products and meat in order to produce sandwiches. They have to have labor in order to put the sandwich together, and then there's capital and business and land and they're paying taxes and also purchasing products from outside of the region of analysis, and then industries are also providing output. The rancher is providing the meat to the sandwich shop, also, and so everything is related within the model.
The reason why these models are popular and the reason why they're widely used is it's a neat way of accounting for everything that goes on in the economy, and also because it can capture ripple effects within a region. When I say ripple effects, think about this. Think about a small town that has maybe 10,000 new households that move into the town. That's a really big number in a small town, and all of a sudden, all of those people are demanding electricity. Let's say that this town is small and basically the only other thing there is a manufacturer that produces turbines in the electric utility, and so in order to supply electricity to those households, the electric utility has to install another turbine, so they go out to the manufacturer and they say, "Hey, can we get another turbine?" Well, the manufacturer has to produce that turbine, and what does the manufacturer need in order to produce that turbine? One of the things it needs locally is extra electricity. If you're looking at the demand for electricity, what actually happens as a result of these households moving in goes beyond the amount that those households actually demanded, and input-output models capture it?
Interpreting Results and Model Limitations
Of course with any economic model, there are caveats; there are limitations in ways to interpret results. One of the most important ones to consider with any input-output model including JEDI is that the results that it produces are gross and not net, and what I mean by that is that it's only looking at impacts that are really kind-of directly associated with a project. It doesn't factor in far-reaching effects that could occur as a result of a project. These could be the things like greenhouse gas emissions; they could be things like changes in property values, public health changes, changes in the economy based on changes in utility rates. Increases in utility rates are going to put downward pressure on economic growth. Gross impacts don't factor that in. Also, input-output models don't estimate supply-side changes, and I don't want to get too much into economic jargon here. When you think about supply side, there's a lot of conversation in politics about supply-side economics, so that's why I say that here. But supply-side changes are basically things that affect producers. So JEDI can't—and input-output models can't factor in things like technological improvements, price changes or changes in taxes and subsidies. So when you look at impacts in an input-output model, if you're looking at those over time, for example, those are assuming today's technology. Those aren't assuming any kind of changes in producer or consumer behavior, and they're assuming the same taxes, same subsidies, and same relative prices for everything. JEDI also doesn't evaluate a project's feasibility or profitability. So if I go and say, if I download the offshore wind JEDI model and say that I want to install a terawatt of offshore wind in Arizona, JEDI will estimate the impacts from a terawatt of offshore wind in Arizona, which brings me to the next point.
NREL is not responsible for how the model is used, applied, or results interpreted. It's freely available tool to the public, we just don't have the time or resources to be able to police how it's used or make comments on it. So it's a tool that's useful for projects, but we really rely on the model users to come up with whether those projects are actually reasonable.
As I mentioned, JEDI assumes constant input-output models and JEDI included, assume constant prices and single technology, no changes in technology. Right now, at the time of this recording, JEDI uses the input-output model from the year 2010. We'll be updating it this winter, to be up to 2012, but regardless, it's a single-year, and this is especially important to consider in the scenario JEDI model. In the scenario model, and we'll be talking about this more in the future, you can look at changes over a time slice, over a period of time, over several years, and so it's important to consider that even though you're looking at that over several years, that's really just a way of categorizing results. All of those results that JEDI is reporting assume the same structure of economy, the same economic structure in the year of the input-output data, which is currently 2010. So it's basically saying, "This is a reasonable representation of how this might look right now," it's not a forecast.
JEDI Results
I talked about the model generally, and I've gone over the caveats to the model and the methodology. Now I want to talk about the metrics that the model reports and how it categorizes those metrics, and then we'll get more into the scenario JEDI model specifically. JEDI reports three metrics right now, jobs, earnings, and gross output. The jobs that JEDI reports are in terms of full-time equivalents or FTEs. One full-time equivalent is the equivalent of one person working 40-hour weeks, 52 weeks a year. Two people working 20-hour weeks all year is the same as one FTE. Earnings are any kind of income from work, so that includes wages and salaries and it also includes employer-provided supplements like retirement and health care. It also includes compensation to sole proprietors. The third metric that JEDI reports is gross output. Gross output is just a general measure of overall economic activity. It's the value of production plus inputs, and I think it's worth emphasizing that gross output is not the same as GDP. GDP is a measure of the value of production. In an input-output framework, GDP is commonly referred to as value added, which I think is a pretty intuitive thing. Most people when you say value added, they understand what that is. It's revenue less the cost of input. Well, gross output is just a measure of everything that's going on, and we do plan on adding value added as a metric in JEDI probably in this next year, but it isn't there yet.
Project Development & Onsite Labor Impacts
JEDI reports impact estimates as they relate to a project scenario. There are three different categories. The one that's closest related to a scenario is project development and on-site labor impacts. These are the ones that are most closely related to a project. These are the people who are getting their hands dirty on site. Things with solar like installers and electricians, technicians, in the case that there need to be cranes brought on board, those are the crane operators, management and sighting. These are really just the direct, direct jobs that are involved with the project.
Local Revenues, Turbine, Module, & Supply Chain Impacts
The second category is local revenue, turbine module and supply chain impacts, for all the turbines in solar. These impacts, these are broadly applicable to a range of technologies as you can see. These impacts are the impacts that really supply those on-site effects, on-site impacts. They can include manufacturers, they can include things like equipment sales, they include business-to-business services, so if an accountant within the region of analysis keeps the books of a solar company, for example, those would show up in there, and …
Induced Impacts
The third category is induced effects. Induced impacts arise as a result of household expenditures. These are expenditures that are made by workers in the on-site effects and expenditures made by the module and supply chain impact effects, and these include things like grocery store, these include effects largely grocery stores, they can be childcare, doctors, dentists, things like that.
Scenario JEDI
Now we've gone over JEDI and talked a lot about basically what the model is, and the methodology of the model and what it reports. Now, we want to get more into the scenario JEDI and how it differs from the other models. In a nutshell, I think the best way to illustrate the difference between scenario and the other JEDI models by asking a question. A normal JEDI model would look at a single project or a single group of projects. An example question that you might answer with a regular JEDI model is what's the impact of 150 kilowatts of small, commercial, rooftop PV installations? Scenario JEDI is much broader than that. It looks at a deployment scenario over time. You might ask what's the impact of 10 megawatts of solar installed over a 5-year period that's split between a number of different types of installations.
Other Innovations New for Scenario JEDI
>>Barry Friedman: And now David is turning it over to me. I'm Barry Friedman, and well done, David. We still have about 10 or 12 more slides, so if you're tempted to go off to a sandwich shop, kind-of hold off for a few minutes. We understand that temptation. A little bit more history here. I was around at NREL 5 years ago and we had our suite of JEDIs, which you can find on the link that David provided earlier. Include wind, land-based wind, offshore wind, biofuels, natural gas, coal, marine, geothermal and petroleum, but in the solar world, we over time, more and more were getting inquiries related to entire solar carve out policies in particular. The demand for something that wasn't designed at a project level but was designed for multi-year policies became strong, became acute and we responded to that demand by developing scenario JEDI. There are a couple of acknowledgments I really want to underscore. David mentioned Marshall Goldberg, our brilliant developer who I worked with developing this tool for over a year, came out late last year. We also received a lot of input on this tool from Vote Solar, so I want to give them a little shout out, as well. There are the basic difference between scenario JEDI and project-level JEDI that David just outlined, there are also some features and characteristics of scenario JEDI that are new for the JEDI suite, and these in particular are in response to, again, a lot of the feedback that we received over the years to make the model more user friendly, to provide more information about methods and sources, items like that, and so we've done that. We've done that in a number of ways that I'm going to get to.
Along those lines, the first thing that we decided we wanted to do since, again, we're talking about a multiyear period for usually a solar carve out, but it can also work if policy makers are looking to say, "What if we had a market that was 10 megawatts over the next 10 years, what would that look like?" so it can be for market scenarios as well as policy scenarios. In some cases, a policymaker in particular would have a certain piece of information that would be a target for the policy, and we realize early on that that is not always a capacity target. They may just know, for example, that they have X amount of dollars to spend on the solar market, and or they may know how many systems they want to put in at the residential, commercial, and utility scale levels. We've generated a system where you can input any of those, megawatts, energy generated, investment, and the model will do the translations for you into the others. That was sort-of a key innovation for scenario JEDI.
The other important piece to take away from the way that we designed the model was that it's meant to be as simple or as complicated as the user is, basically. There is literally a button on the model that allows you to say, "I just want to do this in the simplest way possible," and in that case, you would do the inputs I just mentioned for the three market segments, residential, commercial and utility, and then go right to the outputs and see exactly what the outputs are very simply, and it will prompt you to do that. On the other end of the spectrum, you may have information. You may have specific information on racking costs, inverter costs, electrical wiring, and you may have that information that changes year over year. You may have a sense of how those costs, for example, might be reduced year over year. So on that other end of the spectrum; you can actually input any cost component or any year during your target period. You've inputted a start date of 2012 let's say, or 2013, an end date of 2020 and then there is a spreadsheet that will allow you to input any component costs for any year within that 2013 to 2020 period. That's a very helpful feature because there are people who have that type of information and they can make use of it to improve the accuracy of the outputs. The other thing is there are a lot of help bubbles in the model itself, and I know this may sound like a trivial detail, but we did hear from people that have to flip to FAQs and search through FAQs to figure out things like, how does this work in terms of the methodology here, or where are you getting your default values from, what are your source for your default values, all of that can be found by just looking in that help bubble and see all of that information in there. The other things we've done are a lot of breadcrumbs. There are buttons, they're rather prominent buttons in any of the tabs of the spreadsheet, and there are many tabs in the spreadsheet, that will take you in a logical way to where we think you probably want to go from that tab, and that might be back to where you just were or it might be forward to the next logical place.
Other Innovations New for Scenario JEDI
At the simplest level, you give a percentage of your goal in segments, as I mentioned, and there are some new breakouts that did not appear in the previous JEDI for PV, and those include actually distinguishing large commercial from small commercial, generally in the 10- to 50-kilowatt range. From residential retrofit, which is, over time, becoming a more important distinction as residential new construction grows as a segment unto itself. Then we distinguish fixed mount from single-axis tracking, and silicone from thin film, all of which didn't exist in the previous JEDIs. Then we did something also completely different in this one, which we decided that we wanted to provide some resources for the user, so that these are logical pieces of information that you would otherwise essentially be going online to get and try to find and then going back to the model and creating a really cumbersome process, so the idea here is to provide a more turnkey solution so that as you're running your scenario, there's handy information, rather detailed information, that will inform what you're doing. For example, over time you can see what the status of your market is in your state and generally these are state-level scenarios that you're running. They don't have to be, but they generally are, and you can see how many solar companies are in your state and you can see what the policy is in your state. This would apply to perhaps end-users that really don't have any idea about the requirements in their state for solar. Capacity factors which go into our process of converting capacity in megawatts into energy and megawatt hours and property tax information by state, which is rather detailed and shows how the solar PV project is assessed, where there are exemptions for property tax, that sort of thing.
Scenario JEDI Timeline
And timelines! It's important to realize that the way that the model is structured, and I'll just go right to the example that you see here, if the model takes 2 or 3 years to construct, and during that two or three period you're purchasing components, despite that or notwithstanding that, all of the results are going to show up in the year that the project becomes operational. That's just an important distinction to realize which has more of a relevance to a scenario type model where you are dealing with multiple years, and there are two general categories of jobs depending on how you slice it. David just went over the direct-indirect model, or supply chain and induced, and then here we have sort-of another way to split the jobs into on-site and operating base. Operating is what is sometimes referred to as permanent jobs that are ongoing.
Basic Scenario
In the basic scenario, you put in a start year, you put in a target year, you put in one of the deployment scenario types that I mentioned earlier, you put in the breakout between residential, commercial and utility, and in the simple scenario, there are default cost values for every component for every year that assume some cost reductions over time for the sun shot goals of $1, $1.25, and $1.50 in 2020. The defaults take you in the direction of those goals for 2020, and again, if you want to do this in a more detailed way, you can choose in the entry page an advanced version, and that will take you to places where you can input details for yourself and override the defaults.
The next slide shows you what that basic scenario looks like, and this is your entry page. You can choose your state. There is also an option for choosing a region of the country, and these regions were based on the solar instructor training network, which is divided into eight regional sections, and there was some desire to be able to run scenarios on jobs for sections of the country. In the drop down, you can see those regions and there's detail as to which states those regions are comprised of. It's eight regions and they typically have several states. As I mentioned, here's where you enter your target year, your start year, your type and your segment breakout. All you really need to know to start to see results. I think we took a lot of time to make sure our defaults are as accurate as they can be. Of course, the defaults in terms of costs per year by component are what are going to produce your results in a simple run. With the advanced inputs, I'm going to take this back over to David and so I'll still be here and we've got a few more slides then we'll go to Q&A.
Advanced Inputs
>>David Keyser: Thanks, Barry, and before I get going, I just want to make one point of clarification. Frequently in input-output models, you'll see results categorized as direct, indirect, or induced. Project development and on-site impacts are not the same as direct and supply chain and local revenues, supply chain impacts, are not the same as indirect, and so I just want to make sure that it's clear. Project and development is a little more narrow than direct for those of you who are familiar with it, and then local revenue, turbines, module and supply chain impacts are a little bit more broad than indirect effects in that traditional input-output model. Now what I want to do is go over the advanced inputs in the model and show you examples of how these look in the model and how you can get to these effects and enter them in.
The basic advanced parameters that you can change are deployment over time, so you can change how installations are treated over time by the model, and then also expenditures over time, and I want to issue one word of warning about entering the expenditures. You need to be careful. This has to do with the input-output model representing an economy at a specific point in time. Because of this, there are constant prices, and so if you do specify changes in expenditures, you're specifying changes in expenditures relative to everything else in the economy, so they need to be adjusted for inflation. For example, if you expect laborers to become more expensive over time, and so you're entering in higher and higher labor costs, what you're doing is you're saying that labor is becoming more expensive relative to everything else in the economy. If you expect labor to basically keep up with all the other prices in the economy, then you would want those prices to be flat. The important takeaway is that you're thinking of relative prices here, not absolute prices, so it needs to be adjusted for inflation or anticipated inflation.
This is an expansion of the basic screen that Barry just showed you. You can see down on Row 34 there's the option to select model analysis type, and in this example, advance is selected, and so in this case the model is going to be using specific data that you enter into the model. There are several buttons that are on the bottom of the screen and you can click on those buttons to get to the screens where you can enter it into the data. In this case, I'll click on the review and edit utility systems, do edit utility system data, and this is the screen it takes you to.
You can see that there's the deployment, it's listed up at the top, overtime, and then the project cost data are listed down at the bottom. You can set the deployment by clicking on the goal distribution button or you can set annual costs and other parameters by clicking on the annual costs and other parameters button. If you click on the annual costs and other parameters button, it will take you to this screen.
Costs
And I only had enough room here to put in one year, but these are the inputs that the model allows you to enter for every single year. You can enter in costs of component and then there's also a portion to enter in what's actually purchased locally or manufactured locally, and when it says locally, it's talking about the region of analysis, so that could be the state level or that could be in one of the regions that Barry mentioned.
Deployment
If you click on the deployment or distribution button, it will allow you to edit the deployment for all the types of installations that are in the model. The default is just a constant percentage change over time, so you can change this. I mean, say you think that all the residential installations will happen in 2013 and 2014, you can specify that and then have zeros for the rest of the time or however else you want to split this out. You can change that and customize it.
The results tab is also a little bit different from the other JEDI models because this is multi-year or multi-time estimation. The default when you go to the results tab shows you a sum of impact over all years in the installation, and so this is especially useful for construction period results, which are presented as job years.
Results
In this example, the average annual impacts under the construction phase, the sum over all year is 97, so that's saying 97 job years. If that was a 10-year project for example, that would be an average of 9.7 jobs per year. You can look at that broken out over time by going up to that top button that says, go to all sectors, detailed results, and that will give you annual estimates and I'll show you that in just a minute. The other time slice that it shows is final year for the operating and maintenance jobs. The reason why this shows the final year for operations and maintenance jobs is because these jobs are assumed to be ongoing for the life of the utility, and so the number of jobs that are present in the final year of the scenario are assumed to be the number of jobs that keep on going after the scenario is over.
So if you go up to the top there and if you click on one of those buttons that say, go to all sector, detailed results, I clicked on the go to utilities sector, detailed results, and what it does is it shows you the sum of economic impacts for that sector over the whole time period which is this all years.
This is the first column that you can see here or the first group of columns that you see here in the results, and then it gives you estimates for each year, and that will matter based on how you set up the deployment. With that I would like to thank you all for your attentiveness.
>>Courtney Kendall: Thank you! We would like to thank David Keyser and Barry Friedman for their time today, and we will be posting the presentation slides, audio and transcripts on the SunShot past meetings and workshops page, and I will also be sending out an e-mail with a link to these files as well.
This concludes today's webinar. Thank you for attending and goodbye!
Share