Evolution of an Energy-Efficiency Forecast


Building a model that works across states and programs.

Fortnightly Magazine - January 2013

For more than two decades, the six New England states have made significant investments to advance energy efficiency in the region. State spending on energy efficiency (EE) in New England reaches hundreds of millions of dollars annually, with almost $1.2 billion spent on state-sponsored EE from 2008 through 2011. And it’s estimated that from 2015 through 2021, nearly $5.7 billion will be available to fund energy-efficiency programs.

Backed by the public policies that have spurred this financial commitment, energy efficiency comprises an increasingly important part of the New England electric energy resource mix. For 2015 to 2016, more than 1,500 MW of EE is committed to provide capacity, representing 4.5 percent of the region’s total power system requirement of 34,000 MW.

It’s clear that New England’s state-sponsored EE programs are having an effect on electricity usage in the region, but until recently the magnitude of the long-term impact wasn’t fully quantified. Rapid increases in EE investments have raised important questions about how to measure and incorporate the demand-reducing effect of these EE programs into the 10-year power system planning process. Since 2009, ISO New England, state regulators, and other regional stakeholders—including market participants with EE programs—have been working together intensively to find the answers.

For the short term, ISO New England knows how much EE is committed through the annual Forward Capacity Market (FCM) auctions, which procure generation and demand-side capacity, including energy efficiency, to meet the region’s resource needs three years into the future.

Historically, the ISO’s traditional long-term load forecast incorporated EE by virtue of the fact that the baseline load level was lower than it would have been absent these energy-saving programs. The ISO’s load forecast also included the future impacts of federal appliance-efficiency standards. However, the amount of EE procured in the FCM was held constant past the third year in the 10-year load forecast, so incremental growth in EE wasn’t estimated for years four through 10 of the load forecast or transmission needs assessments.

Because the New England states and other regional stakeholders believed that the growing effect of EE on demand—beyond the three-year FCM timeframe—should be accounted for, the question became, how can energy-efficiency savings be predicted so far into the future? And if a forecast methodology could be developed, would it estimate energy and peak demand savings with enough accuracy to be used in planning for a reliable power system? The challenge in this process is magnified by the diversity of the dozens of unique EE programs in the six states, differing state priorities, and a variety of methods of reporting performance in New England.

The answers to these questions can be found with New England’s first forecast of long-term energy-efficiency savings—believed to be the first multistate EE forecast in the nation. Developed after three years of research and data collection, the forecast uses a model that factors in budgeted and actual spending on EE programs; historical energy savings by EE program type; and a production cost, by state, of each megawatt-hour of energy saved. The model also accounts for uncertainties such as future revenue streams, inflation, technology transformation, and program diversity and penetration. This model allows the ISO to project how many gigawatt-hours of electricity consumption and how many megawatts of peak demand will be avoided in years four through 10 of the long-term forecast, beyond the FCM horizon.

Finalized earlier this year, the energy-efficiency forecast reveals a dampening effect on the growth of both total electricity consumption and peak demand through 2021, both regionally and in each state. The findings from the first EE forecast are being applied in transmission planning analyses. The EE savings, along with an updated load forecast and some recently completed equipment upgrades, have resulted in a revised transmission needs assessment for Vermont and New Hampshire. While many proposed reliability upgrades are still needed, the ISO determined that 10 proposed transmission upgrades, totaling an estimated $260 million, can be deferred. In this case, the EE forecast had a significant influence on transmission planning decisions, representing an important development for the region.

With a robust collaborative effort, ISO New England and the region’s stakeholders were able to develop an energy-efficiency forecast that provides the level of confidence needed for transmission planning decisions.

Energy Efficiency in New England
Figure 1 - Production Costs (Of Energy Efficiency Savings) Over Forecast Period ($/MWh)

Each of the six New England states—Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island, and Vermont—has made energy-efficiency measures a significant policy focus. In general, state public utility commissions (PUC) in New England have been responsible for approving EE program scope, costs, and implementation. Utilities often serve as program administrators (PA) and manage the state-sponsored EE programs, although some states have created separate entities to provide EE services.

In 2010, more than 125 individual EE programs, including education and loan programs, were in place across the six states. The state-sponsored programs deliver services to residential, commercial, industrial, and low-income electric energy customers. These EE programs provide financial incentives to promote the replacement of inefficient electrical devices or processes and also fund the cost difference between standard-efficiency and higher-efficiency devices in new construction.

Primary funding sources for these EE programs are the “system benefits charge” on ratepayers’ bills, revenues from Regional Greenhouse Gas Initiative (RGGI) auctions, revenues from the region’s FCM, and amounts included in state budgets to support policy directives. Each state has a different funding approach, including funding sources, budget periods, and program duration—typically one to three years.

In 2009, New England’s stakeholders wanted to determine whether the states’ investments in EE could be reflected in the ISO’s long-term forecast. The ISO launched the Regional Energy Efficiency Initiative (REEI) to begin a focused discussion of the effect of EE in New England. The REEI gathered high-level information about state EE programs and established a forum for ongoing discussion about EE in New England. The formal process to establish an EE forecast began soon after.

Figure 2 - Total Estimated Budget Dollars ($Thousands)

Initially, the ISO worked with the New England states, the EE program administrators, and other key stakeholders to quantify EE projections through 2021 and to identify any EE resources not already accounted for in the FCM. Results indicated that nearly all the EE resources in these programs participated in the FCM.

ISO New England also surveyed other ISOs and RTOs and found that PJM Interconnection, which has a capacity market similar to New England’s FCM, accounts for EE much as the ISO had done—carrying forward only the amount offered into the market without projecting future investment. The other ISOs and RTOs, with the exception of the New York ISO (NYISO), don’t forecast energy-efficiency savings.

NYISO collects data from six investor-owned utilities, the state’s two power authorities, and the New York State Energy Research and Development Authority. NYISO’s methodology is based on budget projections for energy-efficiency programs, historically based on megawatt-hours saved per dollar spent, program participation rates, and net-to-gross savings ratios. The approach developed by NYISO is designed to provide guidance to NYISO stakeholders as it tracks New York program administrators’ efforts to meet state goals for energy and demand reductions. NYISO’s projections of energy-efficiency impacts become part of the forecasts used for transmission planning analysis. ISO New England determined that New England’s forecast could build on the NYISO approach by including multistate information, but it wouldn’t be subject to any specific state goal, program, or funding cycle. ISO New England then began to develop a forecast methodology in 2011 that included a proof-of-concept forecast using preliminary estimates of production costs and budgets. 

ISO New England also evaluated a proposed methodology to forecast future EE beyond the three-year FCM horizon using a running average of EE procured in the annual FCM auction. However, the ISO concluded that this methodology would be subject to significant risks by assuming future performance would be consistent with past performance. ISO New England determined that a simple averaging and forward projection of EE captured by the FCM wouldn’t be robust enough to support long-term forecasting, and that a more detailed analysis of individual state spending and performance of EE programs would serve as a more stable foundation upon which to build a forecast.

Figure 3 - Energy Savings (GWh)

With the results of the proof-of-concept forecast in hand, the ISO and the New England stakeholders concluded that EE could be forecasted using a production cost model. Such a model would provide the confidence needed to incorporate the EE forecast into long-range system planning processes.

Forecast Model

In early 2012, ISO New England established the Energy-Efficiency Forecast Working Group (EEFWG), which is made up of utilities, regulators, and other stakeholders with expertise in energy-efficiency programs, to assist the ISO in developing the annual forecast, including data collection, assumptions, and model parameters.

The proof-of-concept process determined the types of data elements and level of granularity required to implement the forecast model. The next step was to develop data specifications and data reporting tools for use by program administrators and state regulatory agencies. This resulted in the collection of multiyear, annualized data for each program administered by the region’s utilities. Specific metrics included achieved and committed annual and lifetime net energy savings in kilowatt-hours; demand savings in kilowatts; planned and expended budget dollars; program type, such as building retrofit or new construction; customer class (i.e., residential, commercial and industrial, and low income); EE measure type, such as lighting; and project status. In addition, state energy regulatory agencies provided projections for budgets ranging from one to 20 years into the future.

The forecast model is a budget-based estimate of peak demand savings derived from historical annualized net energy savings, the ratios of peak demand savings to annual energy use savings (MW/MWh), production costs, and future sources of revenue (See sidebar “EE Forecast Model Concept”). The model includes factors designed to capture the uncertainty associated with future revenues, budgets, production costs, and changing technology.

Figure 4 - Forecasted Effects of EE on Energy in New England (GWh). * RRSP12 = 2012 Regional System Plan (RSP) forecast; RSP12-EE in FCM = RSP12 minus energy efficiency (EE) cleared in Forward Capacity Market (FCM); RSP12-EE in FCM-EEF = RSP12 minus EE in FCM minus savings captured in EE forecast.

To address the major challenge of predicting future budgets, the ISO developed a budget uncertainty factor, with input from the EEFWG, based on expectations that changes in state priorities and legislative budgets could occur. The uncertainty factor is modified annually with actual budget figures.

Production costs also were escalated annually by 5 percent in most states to account for increasing EE saturation rates and standards or policies that drive up costs for new technology. Program administrators’ experience shows that, over the long run, the cost of saving a megawatt-hour of electricity is likely to increase as less expensive measures are widely implemented. A 2.5-percent inflation rate was factored in as well.

Additionally, for states where production costs were significantly lower than the regional average, a second cost-escalation factor was applied to bring these programs’ costs closer to the mean. In states such as Massachusetts and Rhode Island, where program budgets are expanding rapidly, a larger discount to future budgets was applied to reflect the potential challenges of meeting these spending targets.

Finally, the forecast estimates the cost of producing EE savings in each state. For instance, the forecast estimates the cost in Maine will rise from $300 per megawatt-hour ($/MWh) in 2014 to $555/MWh in 2021. At the high end of the production cost scale, the forecast estimates a Massachusetts production cost of $530/MWh in 2014 rising to $880/MWh in 2021 (See Figure 1).

2012 Energy-Efficiency Forecast 
Figure 5- Peak Demand Savings (MW). * RRSP12 = 2012 Regional System Plan (RSP) forecast; RSP12-EE in FCM = RSP12 minus energy efficiency (EE) cleared in Forward Capacity Market (FCM); RSP12-EE in FCM-EEF = RSP12 minus EE in FCM minus savings captured in EE forecast.

ISO New England developed its EE forecast on the same timeline as its traditional load forecast, with the results of both forecasts incorporated in the ISO’s annual Regional System Plan and the annual Capacity, Energy, Loads, and Transmission report. The first EE forecast, which ISO NE posted in April 2012, shows estimated savings in energy and peak demand for the region and in each state from 2015 to 2021. As noted, the first three years (2012 to 2014) of the long-term forecast incorporate the savings from EE procured through the FCM.

New England’s first EE forecast projects that the six states will spend nearly $5.7 billion on energy-efficiency measures between 2015 and 2021, producing substantial energy savings (See Figure 2). The forecast shows an average annual energy saving of 1,343 gigawatt-hours (GWh), and total savings of 9,399 GWh from 2015 to 2021 (See Figure 3). Currently, the region consumes about 130,000 GWh annually. The states’ average annual energy savings ranged from a low of 56 GWh in New Hampshire to a high of 786 GWh in Massachusetts.

Overall demand for electricity is projected to remain flat in the region, while in Vermont and Rhode Island, energy demand is projected to decline (See Figure 4).

The forecast also estimated average annual peak demand savings for the region at 206 MW, while projected cumulative peak savings totaled 1,444 MW from 2015 to 2021 (See Figures 5 and 6). The states’ average annual peak savings ranged from a low of 8 MW in Maine to a high of 122 MW in Massachusetts.

Figure 6 - Forecasted Effects of EE on Peak Demand in New England (MW). * RRSP12 = 2012 Regional System Plan (RSP) forecast; RSP12-EE in FCM = RSP12 minus energy efficiency (EE) cleared in Forward Capacity Market (FCM); RSP12-EE in FCM-EEF = RSP12 minus EE in FCM minus savings captured in EE forecast.

Now that it’s in place, ISO New England’s long-term EE forecast will be applied to load levels used in studies looking beyond the three-year FCM timeframe, including long-term transmission needs assessments and solutions studies, and other studies such as generation interconnection or transfer limit analyses, if the study timeframes permit.

Refining the Model

While the FCM provides ISO New England with an accurate and comprehensive understanding of energy savings to be expected from EE resources over the short term, the ISO has worked with its stakeholders to collect and analyze information gathered from the region’s EE programs, and the analysis of this information has advanced the ISO’s understanding of EE from the anecdotal to the empirical.

The result is the nation’s first multistate, long-term forecast of energy-efficiency savings. System planners and stakeholders now have a tool that includes the most complete EE data set possible to serve as a resource for future analysis in a changing energy landscape. When incorporated into the ISO’s planning activities, this forecast should ensure that the region’s large investments in EE are fully accounted for and have an appropriate effect on regional transmission planning decisions.

Monitoring the actual delivery of EE savings to the system and refining the inputs to the forecast model will continue to be an important part of implementing the EE forecast in planning. The ISO’s Energy-Efficiency Forecast Working Group will continue to provide input to the forecast assumptions and feedback on model results as the region gains experience with the forecast. With a 10-year planning horizon, timely corrections to the forecast will be possible as system planners seek solutions to the needs identified in long-term transmission studies.

As the trend of energy-efficiency spending continues to increase across the country, New England’s multistate EE forecasting model is helping the region measure the growth, success, and effects that state-sponsored EE programs are having on the power system.