29 November 2015

Project-Based Assessments in the Victorian Energy Efficiency Target (VEET) Scheme

Energy Policy and Programs

Department of Economic Development, Jobs, Transport and Resources

GPO Box 4509

Melbourne VIC 3001

Re: Opower’s comments on “Project-based assessments in the VEET scheme

consultation Paper 2015”

To Whom It May Concern:

Opower appreciates this opportunity to submit comments in response to the “Project based

assessments in the VEET scheme October 2015 stakeholder consultation paper.”

We commend the Victorian Government (the “Government”) for continuing to lead on

energy efficiency by working to refine and strengthen the Victorian Energy Efficiency

Target (VEET) scheme, and we welcome the proposed changes.

Opower supports inclusion of project-based assessment (PBA) methods in the VEET

scheme. Further, we strongly support the Government’s proposal to include a method

similar to the Commonwealth ERF Aggregated Small Energy Users (ASEU) methodi as

one of the first PBA methods approved for VEEC crediting in mid-2016. Our responses

to the consultation questions below emphasize the following key points:

Including proven PBA methods in the VEET scheme can dramatically

broaden the scope of energy savings activities taken up in Victoria. PBA

methods for Project Impact Assessment and Aggregated Small Energy Users will

unlock two important sources of cost-effective energy savings at scale:

commercial building efficiency and residential behavioural energy efficiency (BEE).

Enabling forward creation of certificates for rigorous PBA methods will

help ensure a robust market for proven, large-scale EE activities. Forward

creation, subject to ex-post verification of energy-saving performance, would

remove financial barriers to investment in large-scale metered-baseline programs

and ensure a level playing field for all VEET activities.

Adopting an Aggregated Small Energy Users PBA method would unlock

reliable, cost-effective energy savings at scale. Behavioral energy efficiency

programs alone have annual potential to generate 300,000 VEECs and AUD 60

million in customer bill savings on electricity and natural gas.

Experimental program design that requires randomised-controlled trials is

the global “gold standard” for ASEU PBA methods. Such methods are

rigorous and widely accepted in Australia and the United States to measure energy savings resulting from behavioural energy efficiency programmes. The

Government’s proposed ASEU method can draw from well-qualified precedents,

and should not require new investment in infrastructure or capacity development.

About Opower

Opower is an enterprise software company that is transforming the way energy retailers

and distribution companies engage with their customers. Opower’s customer

engagement platform enables these energy providers to reach their customers at

moments that matter through proactive communications that drive energy savings and

demand reduction, increase customer engagement and satisfaction, and lower customer

operation costs. Opower’s software has been deployed to more than 95 utility partners

around the world and reaches more than 55 million households and businesses. Having

run hundreds of large scale field tests, and been subject to over 45 independent

programme evaluations, Opower has amassed the world’s largest body of experience in

delivering behavioural demand-side management programmes.

Opower Response to Consultation

Opower is pleased to provide the following responses to the questions raised by the PBA

consultation. Note that we have taken the opportunity to address consultation questions

3-8 with answers more specifically applicable to an ASEU method than to the M&V

method for project impact assessment (PIA) of commercial building upgrades that is the

primary focus of this consultation. While our direct experience with PIA-type activities is

limited, we engage daily with regulators and energy providers on projects measured

using methods similar to the proposed ASEU method. Though our input may not match

the scope that the consultation has in mind, it is our sincere hope that the information

that follows is helpful to the Government as it proceeds with the larger effort to

incorporate PBA into the VEET scheme.

1. Proposed M&V Method

~ Do you envisage a market for the proposed M&V Method? Why/why not?

Yes. Adoption of the proposed PBA methods can dramatically broaden the scope of

proven, large-scale energy-saving activities taken up in Victoria. However, the extent to

which the Government allows forward creation of certificates from PBA-verified activities

will strongly influence the market for each of the proposed methods.

PBAs unlock substantial savings

In our experience, cost-effective end-use EE at scale results primarily from three types of

activity:

1) Lighting efficiency upgrades;

2) Commercial building efficiency upgrades;

3) Residential behavioural energy efficiency.

The existing VEET scheme has successfully incentivized broad uptake of efficient

lighting. Indeed, more than 90% of VEECs issued in 2015 stem from such activities.ii

Inclusion of PBA methods could unlock substantial energy savings from activity types

that have not enjoyed VEET incentives to date. The proposed ”M&V Method” for project

impact assessment will incentivize commercial building efficiency upgrades, and the

proposed ASEU method will incentivize residential BEE.

Timing of certificate creation matters

Opower strongly agrees with the Government’s observation that “forward creation is a

key feature” for PBA methods, because it removes financing barriers for proven, largescale

energy-saving activities by offsetting upfront costs.iii Commercial building upgrades

and BEE programs alike generate substantially greater energy savings than a typical

VEET activity today, but they also incur higher initial costs, which may limit market

interest if there is an interval between investment and certificate creation.

Moreover, if there were PBA-eligible activities that could benefit from forward creation,

they would be at a disadvantage versus activities that have deemed values (e.g.

replacing incandescent lights with fluorescent ones). This would be despite the fact that

PBA ensures greater environmental integrity than deeming, given that PBA awards

certificates on the basis of verifiable performance and deeming awards certificates on

the basis of strong assumptions about likely performance. Forward creation is already a

key de facto feature for deeming savings from VEET-eligible installed-device activities:

an AP gets credit for 100% of “deemed” lifetime savings for a given installed device

when its installation is certified by the Government.

The Government can ensure a level playing field for all activities by treating the timing of

certificate creation as consistently as possible, despite necessary practical differences in

the means by which savings are deemed or measured.

We therefore support the Government’s proposal to allow forward creation of certificates

subject to the proposed M&V method. We also strongly recommend that the Government

allow forward creation of certificates subject to its proposed ASEU method, to the extent

that it requires rigorous ex-post verification and periodic true-ups.

2. Other proposed project-based assessment methods

~ Do you envisage a market for these proposed methods? Why/why not?

Yes. The Government’s proposal to adopt an “aggregated method for quantifying energy

savings from large groups of small energy users using statistical analysis” corresponds

with global best practice for M&V of behavioral energy efficiency programs. We expect

such BEE programs to deliver substantial energy and bill savings for hundreds of

thousands of Victorians. Market uptake of such programs will depend on the timing of certificate creation for activities subject to the proposed ASEU method. Victoria can build

on precedent from other regulatory jurisdictions to ensure both robust experimental

design in advance of program implementation and rigourous ex-post verification of

measured savings.

Randomized Controlled Trials are Best Practice

Behavioural energy efficiency programs that aggregate energy saving activities

undertaken by many small energy users, such as those administered by Opower, use

RCTs to measure energy savings with over 90% statistical confidence. Under this

approach, the program implementer uses parameters to create an eligible group of

recipients, randomly assigns households to control and treatment groups, tests these

groups to ensure statistical equivalence, implements the intervention (such as sending

reports) only to the treatment group, and measures the difference in energy usage

between the two groups using statistical billing analysis. This approach measures

savings with precision and without bias. Key components of this approach include:

● Statistically equivalent and randomly allocated control and treatment groups;

● Opt-out design;

● Ex-post measurement;

● Panel data methodology billing analysis for comparison of control and treatment

groups.

The following is the basic calculation of impact:

kWh saved (treatment group) = kWh used (control group) – kWh used (test

group) – kWH saved by rebated equipment (product participation) for the

same time period and same customers.

This energy savings calculation can be easily converted into the corresponding amount

of emissions reductions achieved by applying the appropriate grid emission factor to

account for carbon intensity of the local fuel mix.

This methodology matches the prescribed protocols for measuring residential behavioral

energy efficiency programs as detailed by the Australian Ministry of the Environment,iv

the New South Wales Office of Environment and Heritage,v the Essential Services

Commission of South Australia (ESCOSA),vi the US Department of Energy’s National

Renewable Energy Laboratoryvii and numerous state-level jurisdictions in the US.viii

Further, an ASEU approach is consistent with the recommendations of the US

government-led State & Local Energy Efficiency (SEE) Action Network’s definitive report:

“EM&V of Residential Behavior-Based Energy Efficiency Programmes: Issues and

Recommendations.” SEE Action is a consensus group comprised of utilities, consumer

advocates, commission staff, and government officials. This SEE Action report

concludes:

“We recommend using a randomized controlled trial for behaviour-based

efficiency programmes, which will result in robust, unbiased programme

savings impact estimates, and a panel data analysis method, which will

result in precise estimates.”ix

This is a low-risk approach because the results are proven and predictable, but also

because they are measured ex post, so the credit is given for results actually achieved.

This is different from many other efficiency programmes, which have expected values

but no means by which to measure after savings have occurred. It is an approach that

has been recognized as the gold standard across the industry.

RCT enables straightforward avoidance of double-counting

RCT program design allows for robust measurement and exclusion from double-counting

of energy savings that are jointly-attributable to the BEE program, and to another energy

saving activity. This ability does depend on being able to having an accessible and

accurate registry of credited energy savings activities that have been taken by customers

within the BEE program population. Opower’s understanding is that Victoria’s registry

is very good, and will thus facilitate in ensuring that double-counting of savings

does not occur.

For a detailed explanation of how an RCT approach should be used by an ASEUtype

method to identify jointly-attributable savings, please refer to the technical

appendix that covers this very topic in the US Department of Energy’s

“Residential Behavior Protocol.”x

BEE programs deliver energy savings equitably at scale

More than 55 independent evaluations by academics, consultants, and economists have

verified that residential behavioural energy efficiency programs empower customers to

save an average of 1% to 3% on their energy bills. These savings are consistent across

fuel type, geography, income, and age.xi

Opower BEE programs have been proven reliably effective, delivering over 8.2 TWh of

energy savings and AUD 1.3 billion in bill savings for customers at over 95 utilities, while

yielding emissions reductions of over 5.5 million tonnes of CO2e. Opower estimates that

with the inclusion of an ASEU PBA method in the VEET, the Government will unlock the

potential to deliver more than 300,000 VEECs and AUD 60 million in customer bill

savings per year via residential BEE programs.

Rigourous M&V enables prudent forward-creation of certificates

As described in detail above, proven ASEU or aggregated “metered baseline” methods

require ex-post verification of savings. Compared with traditional deemed-savings approaches, this step provides a unique level of certainty that savings have occurred.

Yet in practice, requiring this step before issuing credits (or otherwise recognizing

savings for compliance) penalizes that additional rigor, given that months typically pass

between initial program investment and the first robust measurement of program

savings. This timing issue creates a financial barrier that often limits initial program size

to a fraction of cost-effective potential, and in competitive retail markets such as

Victoria,xii it often discourages investment in BEE altogether.

As touched on in our response to question #1, forward creation of certificates already

takes place as a matter of course in EE markets around the world – whenever an

administrator “deems” energy savings up front for an installed measure such as efficient

lighting. This “standardized baseline” approach to forward-crediting efficiency

improvements implies a tradeoff in favor of financial certainty for the AP or EE program

implementer, at the cost of environmental and regulatory certainty (i.e. that savings will in

fact occur according to assumptions for typical device usage that are embedded in the

deeming formula). This tradeoff is prudent and perfectly reasonable, because in practice,

a requirement to monitor usage from each energy-saving device would create an

insurmountable barrier to market participation.

Opower submits that it is equally prudent to permit forward creation of certificates from

programs designed as randomized-controlled trials and subject to ex post verification

and regular periodic true-ups.

The US state of Michiganxiii applies such a hybrid “deem and verify” approach to the

crediting of energy savings from BEE programs, and to our knowledge, represents its

longest-standing precedent. RCT-based programs in the state that address residential

behavioural energy efficiency are assumed to achieve a 1% - 3% energy efficiency

improvement. This initial savings estimate is deemed based on documented results from

similar programs that have been recently implemented in the same or a similar market,

for the same or a similar duration. At a fixed time each year, actual savings for such a

program are measured ex-post using RCT-based methods to verify that the ex-ante

deemed efficiency improvements were achieved. If verified savings exceed the initial

deemed value, the state credits the program with the additional savings. If verified

savings fall short of the deemed value, the state holds the BEE program administrator

liable for the amount of the shortfall during the following year.