PJM, Energy Efficiency & Utility Rebates

[Background: We are publishing the contents of a letter that we recently sent to PJM about how energy efficiency activity driven by the use of utility rebates affects PJM's load forecast. We are publishing this a service to community to educate about the various data sources and limitations available to the general public. We've excluded the general business letter formatting and the introduction to Encentiv Energy and myself] 

9/26/2024

I’m bringing to your attention that deficiencies in how the US Energy Information Administration(EIA) accounts for state and utility energy efficiency programs in the Annual Energy Outlook(AEO) limits and underrepresents the impact of energy efficiency activity in your load forecast. To be clear, I don’t believe there is any issue with your forecast methodology. It seems sound and reasonable. And most of the 3rd party information feeding your forecast model is excellent. However, the AEO has some issues regarding energy efficiency incentives. The AEO is an excellent product and the staff at the EIA should be commended for the hard work and insightful model they have put together with NEMS, it is truly an invaluable resource. But, utility energy efficiency incentives are an arcane and difficult to collect and standardize data set that are independently created by hundreds of unique utility entities under the jurisdiction of 51 different governing authorities and legislatures. 

My research indicates that the AEO does not adequately account for these state and utility energy efficiency incentives to a substantial degree and since the AEO is such a fundamental part of your analysis, it has a concerning impact on the accuracy of your forecast. An accurate load forecast is obviously very important to you and to your constituents so I want to share my findings and see if there is a way I can help. I’ve reached out to several PJM stakeholders and energy efficiency organizations to review my research and they all suggested to reach out directly to PJM to share the results. 

PJM’s recent re-evaluation of energy efficiency resources in the capacity market sparked my curiosity on how utility energy efficiency incentive activity was captured in the load forecast and caused me to initiate research into that topic.

Research & Analysis

I reviewed the following related documentation, primarily:

• PJM Model Review Final Report From Itron - https://www.pjm.com/-/media/planning/res-adeq/load-forecast/pjm-model-review-final-report-from-itron.ashx
• PJM 2024 Load Forecast Supplement: https://www.pjm.com/-/media/planning/res-adeq/load-forecast/load-forecast-supplement.ashx
• PJM Interconnection, L.L.C. submits response to the 06/20/2024 Complaint of Joint Consumer Advocates under EL24-118. https://www.pjm.com/-/media/documents/ferc/filings/2024/20240710-el24-118-000.ashx
• PJM Interconnection, L.L.C. supplements its 07/10/2024 response to the 06/20/2024 complaint of Joint Consumer Advocates under EL24-118. https://www.pjm.com/-/media/documents/ferc/filings/2024/20240729-el24-118-000.ashx
• The National Energy Modeling System: An Overview - https://www.eia.gov/outlooks/aeo/nems/overview/pdf/0581(2023).pdf
• Commercial Demand Module - NEMS Documentation https://www.eia.gov/analysis/pdfpages/m066index.php
• Residential Demand Module - NEMS Documentation https://www.eia.gov/analysis/pdfpages/m067index.php
• Assumptions to the Annual Energy Outlook 2023: Commercial Demand Module https://www.eia.gov/outlooks/aeo/assumptions/pdf/CDM_Assumptions.pdf
• Assumptions to the Annual Energy Outlook 2023: Residential Demand Module: https://www.eia.gov/outlooks/aeo/assumptions/pdf/RDM_Assumptions.pdf

The key activity here is that in the residential and commercial demand modules the EIA conducts a technology decision analysis for when a building decides to do an upgrade. They use prices and efficiencies of various product types to determine the economic factors that decision makers would rely on to determine whether to purchase average efficient products or higher efficient products, typically using a lifetime cost or payback analysis. This cost/benefit analysis is where utility energy efficiency incentives are utilized. The EIA computes what percentage an average utility rebate makes up of a higher efficiency product’s cost which in turn is used in the Residential and Commercial Demand Module’s technology choice algorithm. So the higher the rebate, the more likely a buyer is going to choose a higher efficiency product. The EIA uses several data sources to compute an average utility rebate per various product categories in each of their census zones. An example of those tables are included in the “Assumptions to the…” publications.

The EIA’s published average rebate as a % of product cost in the EIA documentation did not seem right to me based on our product experience with rebates and product manufacturers, so I began a conversation with the economists at the EIA. Over several weeks and various conversation threads my conversations with the EIA economists pointed me to their data sources and revealed several policy misunderstandings about how utility rebates affect the market.

I’ll start with the data sources. The EIA primarily depends on the following sources for nationwide incentive information:

• CEE Program Summaries - https://cee1.org/program-resources/
• ENERGY STAR Program Summaries(DIME) - https://www.energystar.gov/DIME

Both the CEE and ENERGY STAR are providing a great service to the community.  But both initiatives are reliant on voluntary contributions of incentive information from busy utility program personnel. There isn’t a national standard for how to design energy efficiency incentives, so each utility is submitting as they see fit. 

I examined both source closely and compared them to our database of incentives. I found the following deficiencies:

• Age- CEE in particular is always 2-3 years behind the publication of the AEO forecast.  Here in September 2024, as the EIA works on the 2025 AEO, the CEE data is for utility incentives as of 2022.
  • Does this matter? Yes. Utility rebates have gotten more volatile especially for common measures like Lighting and HVAC.
  • My company, Encentiv Energy, produces an annual webinar and report for changes in commercial energy efficiency programs. It indicates common changes in programs rates of 50% or more. A link to the most recent one:https://blog.encentivenergy.com/news/analysis-2024-utility-business-commercial-industrial-rebate-trends
• Quality/Consistency– Both CEE and ENERGYSTAR sources are updated voluntarily by utilities. While admirable, it leads to errors and inconsistencies.  Examples:
  • ComEd - Only providing midstream incentives for lighting
  • PECO HVAC - Incentive levels are too low by 80%
  • BGE HVAC - Incentive levels are too low by 40%
• Omissions – There is a lot of missing information. Some measures, utilities and some whole states are not represented in the data. Examples:
  • Commercial Outdoor lighting - Missing
  • Chillers - Missing
  • Virginia utilities(Dominion and AEP) - Missing
  • First Energy PA - Missing
  • PPL - No data
  • Potomac Edison - Missing
  • AEP Indiana Michigan Power - Missing

As to the policy misunderstandings, I learned the following:

• Midstream/Upstream programs are not accounted for. In the EIA’s view, their Residential and Commercial choice models do not allow for the introduction of other actors in the rebate process.  It assumes a direct interaction between the buyer and the utility in terms of whether a rebate can be introduced into the upgrade economics or not.  The EIA considered the fact that midstream or upstream programs introduce a 3rd party, a retailer or distributor, meant that those incentives have to be ignored in their analysis. 
  • Does this matter? Yes, a lot.  The midstream and upstream program designs have been a staple of residential programs and increasingly commercial programs. Additionally, many utilities have been and continue to move to using these models as their sole or primary means for dispersing incentive dollars. This policy represent an existing and increasingly important issue for energy efficiency data accuracy
• An EIA assumption that midstream incentives are included in the public pricing that distributors display on their websites.  The EIA checks its product cost assumption by comparing them to distributor public sites. It is NOT the standard practice for distributors to include net rebate pricing in their public site since midstream programs have a customer eligibility requirement to receive the incentive and distributors don’t want to reduce product prices for ineligible transactions that they can’t be reimbursed for.
  • Does this matter? Similar to the previous point, yes, a lot. To the outsider it would be logical to assume that a distributor would publish the lowest price possible, until you learn about the deeper workings of these programs and the risk of loss that distributors would have to take on to publish net rebate prices
• PJM’s EER program. The EIA was unaware that PJM and other grid operators are operating energy efficiency programs, so those savings are also not accounted for in their model.

In the subsequent discussions with the EIA, they are examining how they can update the documentation to more explicitly describe their treatment of midstream and upstream programs.

Impacts:

My research and investigation leads me to find the following impacts:

• PJM load forecasts are missing load reductions for higher efficiency equipment that are sold with midstream/upstream rebates. This would cause there to be an underrepresentation of the amount of higher efficient equipment installed and lead to the EIA to publish higher energy use intensities than is correct across all geographies.
• PJM load forecasts are distorted by misrepresentations of product incentives by downstream rebates. This would lead to mostly underrepresentation of higher efficient product sales in specific utility territories for specific product types. 
• PJM load forecasts are deficient with regards specific categories of products, like  commercial outdoor lighting and chillers which are very significant sources of electrical demand

 I’d like to offer a number of potential solutions, some of which are already in progress:

• Short-term work with EIA. The EIA has received permission from Encentiv to start examining our UtilityGenius site to help address some shortcomings in their data set. 
• Longer term with EIA. The EIA published an RFI (https://sam.gov/opp/60ca5c649bb943cab33a66871575b458/view) and will move to an RFP eventually to hire a contractor to improve the energy efficiency incentive collection process
• Allowing qualified energy efficiency measures to participate in capacity market. Both options for the EIA will take some time and may be years before fully addressed. It is possible to identify energy efficiency measures by type and geography that explicitly couldn’t have been included in the EIA use intensities. For example:
  •  The measures above that have been omitted from analysis like commercial outdoor lighting
  • Where a utility’s rebate as a percentage of product cost is higher than the EIA’s census average
  • Projects where midstream rebates would be qualified
  • Others
• I’ve been trying to think of a way you could adjust your load forecast to account for these deficiencies, but the way the EIA aggregates and averages these values into census regions and then runs their technology choice algorithms would make that mathematically impossible. 

Conclusion

I hope this research and analysis is helpful to you. Utility energy efficiency incentives are difficult to work with in aggregate because each utility program can design them independently, so we have run into issues like this in other business areas. I am more than happy to answer any questions that you have.  We are also happy to work with PJM to help determine solutions that can resolve these deficiencies.

Mike Cham

CTO, Encentiv Energy