Split Incentives in Residential Energy Consumption

Abstract

We explore two split incentive issues between owners and occupants of residential dwellings: heating or cooling incentives are suboptimal when the occupant does not pay for energy use, and insulation incentives are suboptimal when the occupant cannot perfectly observe the owner's insulation choice. We empirically quantify the effect of these two market failures and how they affect behavior in California. We find that those who pay are 16 percent more likely to change the heating setting at night and owner-occupied dwellings are 20 percent more likely to be insulated in the attic or ceiling. However, in contrast to common conception, we find that only small overall energy savings may be possible from policy interventions aimed at correcting the split incentive issues. JEL: Q4,Q5

Pro-environmental Households and Energy Efficiency in Spain

“…This is an indication of household ecofriendly behavior that has been commonly employed by the literature (see e.g. Levinson and Niemann, 2004;Maruejolds et al, 2011;and Gillingham et al, 2012). Moreover, as Figure 1 indicates, space heating is the…”

“…In particular, our survey asks respondents whether their fridge, washing machine, dishwasher and oven have an A or A+ label. We follow Gillingham et al (2012) and restrict the sample to households that have bought their major appliances over the last five years. Therefore, we exclude situations where labels had not been implemented yet, and also reduce possible effects of changing trends on the supply side.…”

“…While income effects were small for temperatures, income was an important determinant of household eco-friendly habits. Gillingham et al (2012) employed the 2003 'California Statewide Residential Appliance Saturation Study' to find the effects of certain variables on morning heating temperatures, changes of heating system or insulation level. Using probit models they identified principal-agent problems both in heating or cooling and insulation.…”

“…Since this may affect the intensity and frequency of residential heating use, we compare the results from the full sample with the results from a reduced sample. This reduced sample, as in Gillingham et al (2012), considers only colder regions 7 .…”

“…Following Gillingham et al (2012) we construct a discrete variable that takes value 1 if the temperature stated by the respondent is less than or equal to 15 º C; value 2 if it falls between the interval 16-18 º C; value 3 for the interval between 19-21 º C; value 4 for the interval between 22-24 º C; value 5 for the interval between 25-27 º C; and value 6 for those temperatures equal or above 28 º C. Then we use an ordered probit model similar to Equation (4) to explain the effects of household socio-economic and environmental characteristics on the winter indoor-heating temperature while controlling for the rest of relevant factors. Table A1.5 (Annex I) contains a description of specific variables used in the regression 6 .…”