MONDAY’S STUDY: Numbers Show Energy Efficiency Is Key To Cutting NatGas
Cost of saving natural gas through efficiency programs funded by utility customers: 2012–2017
Steven R. Schiller, Ian Hoffman, Sean Murphy, Greg Leventis, Lisa C. Schwartz, May 2020 (Lawrence Berkeley National Laboratory)
Energy efficiency programs for customers of natural gas utilities provide multiple benefits, including improving energy affordability and resilience and easing gas pipeline constraints. Policymakers, state public utility commissions, utilities, and other program administrators rely on cost performance metrics, such as the cost of saved energy (CSE), to assess energy savings potential and design and implement programs in a cost-effective manner. In resource planning and implementation processes, accurate assessments of efficiency cost performance help ensure reliability at the most affordable cost.
Berkeley Lab earlier conducted an analysis of the cost of saving natural gas for the 2009–2011 period (Billingsley et al. 2014) through energy efficiency programs funded by customers of investor-owned utilities (IOUs). In this new study, Berkeley Lab collected publicly available cost and savings data for 2012-2017 reported by IOUs and other program administrators (PAs) in a dozen representative states (AR, CA, CT, IA, MA, MI, MN, NJ, NY, OK, RI and UT) that provide geographic representation in all four U.S. Census regions. Our analyses focus on estimating the PA levelized cost of saved energy (PA CSE) for three core sectors for natural gas: residential, low income, and commercial and industrial (C&I). We also aggregate these sectors to provide regional and national values.
We report savings-weighted averages and unweighted medians and interquartile ranges of the levelized PA CSE for natural gas in constant $2017/therm (Figure ES 1). The savings-weighted average PA CSE for gas efficiency programs in these 12 states over the 2012-2017 period was $0.40/therm. This is similar to our earlier estimate for the period 2009 to 2011—$0.38/therm.
For context, this figure represents the avoided costs for the natural gas commodity, plus transportation, delivery, and any storage costs. Some jurisdictions also consider additional avoided costs, including environmental factors and reduction in price risk.
Other findings include the following, subject to additional research:
• C&I programs provided the lowest savings-weighted average cost of gas savings ($0.18/therm), yet represented a minority of overall spending (about 20%). The cost of savings for residential and low-income programs was $0.43/therm and $1.47/therm, respectively. Residential and low-income programs accounted for about three-quarters of national gas program spending in our dataset.1
• The PA CSE for gas programs in our dataset varied by geographic region, with the largest differences between the Midwest ($0.25/therm) and the West ($0.59/therm). Some of this difference is probably driven by the large amount of spending on low-income programs in our Western regional dataset and perhaps the difference in savings opportunities between cold versus temperate regions.
• Within our dataset, the average PA CSE for natural gas trended downward from 2012 to 2017. Part of the driver for this change appears to be a shift toward longer-lasting measures. Verifying this shift would require additional data collection and analyses.
Within each state, the number of PAs studied varies from one to a dozen. The number of programs and PAs for which data were collected also varies each year due to changes over time in state policy, PA reporting or data availability. Depending on the year, the Berkeley Lab dataset covers 32 to 37 PAs that account for about 50% to 70% of annual national spending on natural gas efficiency programs.2
We made significant efforts to ensure accurate representation of PA-reported costs and savings in our database. However, we experienced several data quality and data screening challenges, similar to those indicated in prior Berkeley Lab reports.3 Thus, the CSE values provided in this report should be considered estimates for the PAs sampled.
With respect to data collection and analyses, we were able to generate defensible estimates for the PA cost of saving natural gas covering a large share of national program spending, despite some data challenges. Reporting of gas program data has improved in many states, but significant and meaningful opportunities remain for greater transparency, rigor and comprehensiveness, including estimating and reporting savings and measure lifetimes.
This work could be strengthened and supplemented with the following additional research:
• Expand data collection to other states for fuller geographic representation, larger sample size and greater confidence in the results
• Provide technical guidance and support to states and PAs for more consistent reporting and improvements in evaluation, measurement, and verification (EM&V) and estimation of savings from natural gas efficiency programs
• Conduct analyses to identify the drivers of a downward trend in the average PA CSE from 2012 to 2017, as well as factors such as climate that drive differences in the cost of saving gas across states
• Analyze PA CSE for gas by efficiency program type (e.g., residential new construction, industrial process, commercial heating, ventilating and air-conditioning)
• Develop a cost curve for gas efficiency programs
• Assess the potential impact of PA size on cost
• Estimate the Total CSE for natural gas programs, including participant costs
Program, Sector and Portfolio Years
Program years represent spending and savings data for a single program for a single year. For example, data covering four years of spending and savings for a particular program represent four program years.
We rolled up all program year data to the market sector level for each program administrator to create sector years—the administrator’s total spending and savings data for all programs targeting a market sector for a specific year. The market sectors include Residential, Commercial and Industrial, Low Income and Cross Cutting (for programs and spending that span two or more sectors). We then aggregated the sector years by region and for the full dataset.
At the sector level, costs and savings are weighted averages of individual program costs and savings. Sector-level CSE values therefore reflect the influence of higher-volume programs in that sector and can be pulled up or down by those programs.
Sample sizes (n) in charts and tables are in portfolio years for results for the full dataset or all data for a region. Sample sizes are in sector years for results by market sector.
Observations and potential next steps
This targeted collection and analysis of natural gas efficiency program data in 12 states demonstrates the potential for estimating, evaluating and reporting the cost of saving gas at national and regional scales and for multiple market sectors. This information could serve as an important new resource for utilities, state and national decision-makers, and the energy industry as a whole. We are not aware of other, geographically representative estimates for the cost of gas savings.25 Furthermore, while this effort underscores continuing challenges with completeness and rigor in savings estimation and program reporting, it also highlights opportunities for improvements as well as examples of exceptional estimation and reporting practices.
We observed wide ranges in the cost of gas savings by region and among market sectors. Yet within the residential and C&I sectors in most regions, we see fairly tight ranges. Identifying the nature of regional variability was beyond the scope of this study, but may be the result of just a few drivers, such as climate and the scope of low-income efficiency programs. We also noted relatively small differences between savings-weighted averages and medians at the market sector level, which suggests that the relationship of the reported costs and savings is roughly similar among PAs, regardless of size.
Further examination of the relationship of gas prices and efficiency program cost-effectiveness could provide meaningful insight into future gas savings from voluntary programs as an energy resource and source of savings for utility customers. In particular, this work could be expanded and enriched in several ways:
1. Expand data collection and analysis – Collecting additional states’ data would provide fuller geographic representation, larger sample size, more diversity and greater confidence in results.
2. Provide technical guidance and support to states or utilities for improved reporting – More comprehensive, consistent and rigorous reporting pays dividends for utilities, public utility commissions, program implementers, trade allies and other stakeholders (including researchers). Technical support could address improved reporting as well as underlying issues in EM&V, estimation of assumed savings and measure lives, and allocation of dual-fuel program costs by fuel and portfolio-level costs to programs. Berkeley Lab created guidance documents with recommendations on data collection and reporting for efficiency programs (Rybka et al. 2015).
3. Analyze drivers of cost trends – Analysis is needed to explain the drivers of the downward trend in the average PA CSE we observe from 2012 to 2017. In addition, analysis of heating degree days for states or PA territories could help isolate how much of the variation in the cost of gas savings is attributable to climate versus other factors.
4. Estimate PA CSE for gas by efficiency program type – Utility filings collected for this project contain data at the program level. Much of the data in major states (e.g., CA, NY, MA) were collected and standardized prior to sorting by sector and analysis. Thus, the data could be analyzed at the program level. A national program-level analysis would require additional time for data collection, not resolution of new challenges, and yield insight into the magnitude of savings and cost performance of various submarkets, measure classes and implementation strategies. We perform and publish such programlevel analysis for electricity efficiency programs.
5. Develop a cost curve for gas efficiency programs – In 2018, Berkeley Lab introduced a cost curve for electricity efficiency programs that simultaneously depicts the magnitude of savings for individual program types and the cost of acquiring those savings (Hoffman, et al. 2018). A similar waterfall chart for gas efficiency programs would illustrate where PAs rely most for savings, how cost performance of those programs stacks up against gas prices, and how magnitude of savings might change in the future.
6. Analyze costs by PA size – Annual energy savings tend to be correlated with the size of the utility—its retail electricity load. Similar to our most recent comprehensive study of electricity efficiency programs (e.g., Hoffman et al. 2018), we could segment PAs into groups by annual energy savings to assess the potential impact of PA size on the cost of saving a therm.
7. Estimate the Total CSE for natural gas programs – Participating customers typically pay for a portion of efficiency project costs. In some cases, participants pay a significant share. Similar to our studies for electricity efficiency, a future study could collect participant costs for natural gas programs to develop estimates of the Total CSE—PA CSE (administrative costs and incentives) plus participant costs.