ESTIM IMATIN ING M G MEASUR URE L LIF IFE F FROM BEHAVIO - PowerPoint PPT Presentation

ESTIM IMATIN ING M G MEASUR URE L LIF IFE F FROM BEHAVIO IORAL P PROGR GRAMS Hannah Arnold BECC 2013 November 20, 2013 Opinion Dynamics

…What we know, What we need to know, How it challenges our thinking Opinion Dynamics

Wha hat i is t the he i issue a and nd w why d y does i it ma matter? Estimating Measure Life from Behavioral Programs 3

Why a y are w we t talki lking ng a about t thi his i issue? § Program administrators (PAs) have been offering behavioral programs for a relatively short time and their fate as an effective program intervention depends on their associated costs and benefits. § As part of cost effectiveness calculations, PAs look to Effective Useful Life (EUL) an estimate of the median number of years that a measure installed under a program is still in place and operable. § However, EUL is less clear when considering actions taken by participants due to behavioral programs particularly when those actions are not equipment purchases or constitute multiple equipment installations. Opinion Dynamics Estimating Measure Life from Behavioral Programs 4

Purcha hase a and nd no non-p n-purcha hase a actions ns a are b both a h aspects o of beha havior Waste Profile End Use Profile Percentage of Annual Percentage of Total Annual Lighting kWh kWh Waste : ¡How ¡much ¡of ¡the ¡ remaining ¡usage ¡is ¡due ¡ Lighting to ¡inefficient ¡equipment ¡ vs. ¡wasteful ¡behavior? ¡ Efficient ¡Usage : ¡How ¡li=le ¡ Current ¡Usage : ¡How ¡ ligh>ng ¡energy ¡could ¡be ¡used ¡if ¡ much ¡electricity ¡ all ¡customers ¡installed ¡efficient ¡ actually ¡goes ¡to ¡each ¡ lamps ¡and ¡turned ¡lights ¡off ¡when ¡ ¡ end ¡use? ¡ not ¡needed? ¡ Opinion Dynamics 5

De Determi mini ning ng a an E n EUL UL f for b beha havioral p l programs ms i is c current ntly ly conf nfound nded b by t y two i issues Uncertainty Application § Uncertainty as to what occurs with non-purchase energy efficiency actions without continued program intervention § The practical application of an EUL given how these programs are implemented Opinion Dynamics Estimating Measure Life from Behavioral Programs 6

Unc ncertaint nty A y Around nd t the he S Source o of E Ene nergy S y Saving ngs § It is unclear what specific actions are driving savings, and therefore how long those savings might persist § Once a behavior is internalized, it is known to persist without continued prompting from outside sources. However, behaviors can decay and not are habituated indefinitely § If non-purchase behaviors persist, we have no empirical evidence about the length of time they persist Opinion Dynamics Estimating Measure Life from Behavioral Programs 7

Practical A l Appli lication o n of a an E n EUL UL i in C n Cla laime med S Saving ngs § Unlike standard programs with a specific “installation date,” behavioral programs treat customers over periods of time, usually multiple program years, in which customers can take a wide range of actions. § This complicates how a program team might claim savings that extend beyond one year. § One example is claiming an EUL for each year of continued treatment Opinion Dynamics Estimating Measure Life from Behavioral Programs 8

Cha halle lleng nges o of P Potent ntial E l EUL UL a appli lication n § Under this scenario, savings would be effectively double counted year-over-year. Years Savings are Claimed PY1 PY2 PY3 PY4 PY5 Actual PY1 * Treatment PY2 * Year PY3 * Note: orange shading represents double counted savings for customers who are present the first year, as well as subsequent years. Opinion Dynamics Estimating Measure Life from Behavioral Programs 9

Wha hat w we kno know Estimating Measure Life from Behavioral Programs 10

The he C Current nt S State o of E Evidenc nce f from A m Ana nalys lyses § Persistence with treatment § Persistence without treatment Opinion Dynamics Estimating Measure Life from Behavioral Programs 11

Persistenc nce w with T h Treatme ment nt § Our review of behavior-based programs demonstrates that savings continue to persist with treatment year-over-year when the program continues to provide information to participants. SMUD UD Persistenc nce Nationa nal Gr l Grid Puget S Sound nd E Ene nergy (Int Integral A l Ana nalyt lytics, , with T h Treatme ment nt (Opini nion Dyna n Dynami mics, 2 , 2012) (KEMA, 2 , 2010) 20 2012) 12) 2009 2010 2009 Coho hort Wave 1 Wave 2 Electric Gas Electric Electric Gas Year 1 1 1.80% 1.60% 1.61% 1.25% 0.81% 1.71% 1.17% Year 2 2 2.40% NA 2.06% 1.63% 1.25% 2.00% 1.46% Year 3 3 2.40% NA 2.21% 2.12% 1.43% - - Year 4 4 2.10% NA NA NA NA 2.60% 1.30% Opinion Dynamics Estimating Measure Life from Behavioral Programs 12

Persistenc nce w with T h Treatme ment nt ( (Cont nt.) .) Lift i in U n Uptake Measure L Life ( (Years) (Treatme ment nt % % - C - Cont ntrol % l %) Average a and nd R Rang nge Building Building Envelope 7.0% Envelope Consumer Consumer s asures 6.8% Electronics Electronics Low-Cost Low-Cost Measures 6.2% Measu Measures Appliances 5.0% Appliances Me Light Light Fixtures 2.2% Fixtures Heating / Heating / Cooling 1.8% Cooling Consumer Hot water 4.2% electronics usage Consumer Hot water usage 4.2% haviors electronics Space Lighting 0.8% heating Beha Space Heating / 0.8% Lighting Cooling Refrigerato Refrigerator Maint. -0.7% r HVAC HVAC Maintenance -4.7% maintenan -10% -5% 0% 5% 10% 15% 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Chart o order ered ed f from h highes est t to l lowes est l lift i in Range b e bands r rep epres esen ent m minimum a and m maxi ximum uptake b e by t trea eatmen ent g group mea easure l e life o e of m mea easures es w within ea each g group Opinion Dynamics Estimating Measure Life from Behavioral Programs 13

Example le o of P Potent ntial B l Beha havioral S l Saving ngs o over T Time me d due t to the he E EUL UL o of M Measures PY1 Poten>al ¡persistence ¡of ¡HER ¡program ¡savings ¡ 1.7% savings 100% electronics ¡& ¡ligh.ng ¡ measures, ¡4-­‑6 ¡yrs ¡ 80% ngs l as Saving level a recycled ¡refrigerator, ¡5yrs ¡ PY1 S 60% ngs le l Saving of P ntage o nnual S 40% Percent Annu showerheads ¡& ¡ aerators, ¡10 ¡yrs ¡ ¡ 20% 0% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Years f from P m Program S m Start Opinion Dynamics Estimating Measure Life from Behavioral Programs 14

Persistenc nce w witho hout T Treatme ment nt § Program administrators and evaluators are studying the persistence of savings from behavioral programs once the treatment (i.e., the messages from the program) has been stopped several jurisdictions, but it’s still early § There are current analyses within the industry that indicate that savings persist for longer than a year, but these analyses do not help answer how long they last. Opinion Dynamics Estimating Measure Life from Behavioral Programs 15

Wha hat w we ne need t to kno know Estimating Measure Life from Behavioral Programs 16

At p present nt, t , the here i is i ins nsufficient nt e evidenc nce f for a a s specific E EUL UL estima mate § Current studies do not provide conclusive evidence of standard, predictable actions taken as result of the programs. § Program design and implementation matters: § Empirical research has demonstrated that savings magnitude and persistence with treatment varies based on target population and program model (opt-in vs. opt-out). § Further, the frequency and duration of behavior interventions has a big impact on the persistence of the behavior being promoted by the intervention. Opinion Dynamics Estimating Measure Life from Behavioral Programs 17

Ind Industry r y research o h on E n EUL UL i is ne needed § Cohorts must be dropped from treatment to assess persistence and some experimentation is happening § There are two primary ways to determine the EUL of these programs: § Conduct a longitudinal persistence test: Remove treatment of reports and observe how savings change over time § Conduct annual Poten>al ¡persistence ¡of ¡HER ¡program ¡savings ¡ PY 100% survey research: electronics ¡& ¡ 80% ligh.ng ¡ Determine measure measures, ¡4-­‑6 ¡ 60% yrs ¡ recycled ¡ refrigerator, ¡ installations due to 5yrs ¡ 40% the program showerheads ¡ 20% & ¡ aerators, ¡10 ¡ 0% yrs ¡ ¡ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Opinion Dynamics Estimating Measure Life from Behavioral Programs 18