Factors Affecting Renters' Electricity Use: More Than Split Incentives.

AuthorBest, Rohan

    How much extra electricity do renters use relative to comparable non-renters, and what explains any difference? Multiple factors may be relevant. In cases where landlords pay the energy bill, renters do not face a price-based incentive to reduce electricity use. This is referred to as a "consumption split incentive". (1) Landlords may also underinvest in energy efficiency due to information asymmetries, as is often likely to be the case for insulation given the difficulty of observing its quality and quantity. This is called an "efficiency split incentive".

    Other explanations, such as possible behavioral channels and the use of substitute fuel types, have been less analyzed. It is possible, for example, that renters display different average behaviors in terms of the number of hours spent watching television and using lighting. Renters may also be more reliant on some certain electrical appliances, such as space heaters, especially if there are barriers to the use of natural gas in the home. (2)

    Understanding potential renter effects on electricity consumption is important for a number of reasons. Renters make up over 30% of U.S. households (Pew Research Center 2017) and tend to be more disadvantaged from a socioeconomic perspective. The 2015 Residential Energy Consumption Survey (RECS) indicates that over 30% of renting households had annual gross household incomes of less than $20,000, compared to 10% of non-renters (U.S. Energy Information Administration 2018a). Any factors that would place upward pressure on their electricity bills would contribute to financial stress. Inefficiently high electricity usage quantities would also contribute to increased emissions of carbon dioxide and other pollutants from the electricity sector. Steps that could help to improve energy efficiency would help to alleviate these issues, while also allowing for greater enjoyment of energy services.

    Previous findings on the effects of renting on energy use and expenditure have been mixed, perhaps in part because the sets of control variables that have been used in the analyses are different. Melvin (2018) found that a split-incentive rental market failure has led to 3% higher total energy consumption in the U.S., based on the 2009 Residential Energy Consumption Survey. Rehdanz (2007) found lower expenditures on space heating and hot water supply for owner-occupiers in Germany. Some studies have found an effect in the opposite direction for other developed countries, with Wood et al. (2012) for example finding that energy expenditure by private renters in Australia is lower than that for homeowners. For the United Kingdom, Meier and Rehdanz (2010) found a positive effect of home ownership in explaining heating expenditure per room.

    A number of prior studies focus on specific split-incentive effects that could raise energy consumption. For example, Levinson and Niemann (2004) investigated consumption split incentives and found that the amount of extra rent for households who do not pay their energy bill is less than the cost of the energy used. Making renters responsible for energy bills is thought to potentially reduce energy consumption by around 25% (Elinder et al. 2017; Brewer 2018). Principal-agent problems affecting residential energy efficiency have also been extensively considered (Gillingham et al. 2012; Myers 2020; Houde and Spurlock 2016; Joskow 2016).

    There are many economic and social factors that are relevant for residential energy consumption. Jones and Lomas (2015) investigated socioeconomic determinants of high electricity consumption in the United Kingdom, finding that income is a key factor but education is not. Age of the survey respondent is another potentially important determinant, with younger respondents tending to report higher household electricity consumption in China, other things equal (Chen et al. 2013). It is also important to control for affluence given its negative correlation with renting. This can be done either directly or in an indirect way by controlling for variables that reflect resource differences across households, such as income, education, age, property size, and household contents.

    This paper has two objectives. First, it aims to quantify the magnitude of the effect of renting on electricity consumption after conditioning for a larger set of covariates than prior studies. The sequential addition of control variables will provide information on how conclusions about the renter effect rely on the controls that are included. Second, the paper seeks to examine the possible channels through which renter effects may occur.

    A key finding is that a negative unconditional effect of renting on electricity use turns into a positive conditional effect when suitable controls are added. Renters use 6% more electricity than non-renters on average after controlling for basic socioeconomic factors. This increases to around 9% when controlling for quantities of appliances that are less popular in rented households on average. The paper then considers the various channels for this type of effect, including split incentives and behavioral differences. Finally, the paper quantifes the effect of renting on uptake of appliances. The findings suggest that renters are more likely to have electric space and water heaters and that there is evidence of some relevant behavioral differences. For example, renters are more likely to have the main television on for at least four hours on a weekday, all else equal.

  2. DATA

    The data are from the 2015 Residential Energy Consumption Survey (RECS) of the U.S. Energy Information Administration (2018a). This is a nationally representative survey of 5,686 U.S. households. Data collection in the 2015 RECS was carried out through a combination of computer-assisted personal interviews and self-administered reporting. The survey asked many questions about energy use over the prior year. In the current paper, households for whom electricity use is imputed will be excluded from the electricity use regressions. (3) The survey covers single-family homes, units, and mobile homes. Group quarters such as prisons, military barracks, and nursing homes are excluded. Probability weights are available, so each surveyed household can potentially be weighted according to how many households in the U.S. population it represents. These weights are used in weighted least squares regressions (see robustness tests available through the online code).

    The survey data provide evidence of substantial differences between renting and non-renting households. Electricity consumption is 35% lower for renter households on average, as evident in Table A.1. However, it is important to consider key factors that could explain this difference, as is done in the econometric analysis. Renter households on average have lower income, live in smaller homes, and are more likely to have recently occupied their residence (i.e. moved in within the last five years).

    Figure 1 reveals the link between income and electricity consumption. For non-renters, higher income tends to be associated with higher electricity consumption. This relationship is less clear for renters. 86% of renters are in the bottom four income bands shown in the Figure, meaning that only a small share of renting households had annual gross household incomes above $80,000 as of 2015. Only 62% of non-renters are in the bottom four income bands.

    Renters also differ in terms of the quantity of appliances in their household. Renters generally have less electrical equipment: only 83% of renters have air conditioning of some type (compared to 89% of non-renters), while only 48% of renters have electrical clothes dryers (compared to 75% of non-renters). (4) Exceptions include that renters are more likely to use electricity for some key purposes such as space heating, water heating, and cooking--as evident in Table A.1. As many as 47% of renters have electric space heaters in the home, compared to only 30% of non-renters.

    Natural gas, the main alternative residential energy source, is available in the neighborhood for a larger percentage of renters than non-renters. However, renters are nevertheless less likely to use natural gas than non-renters (U.S. Energy Information Administration 2018a). On some other dimensions, renters and non-renters are actually quite similar. For example, there are 2.6 people per household in non-renting households on average, compared to 2.5 for renting households (see Table A.1).

    As can be seen in Figure 2, non-renters are more likely to have adequate insulation. Non-renters also have higher uptake rates of every type of energy-star appliance shown in the Figure. For example, 87% of non-renting households have adequate insulation in the home, compared to 72% of renters. In relative terms, more than twice as many non-renting households have access to energy-star-rated equipment such as windows, dishwashers, and freezers.


    3.1 Electricity consumption model

    An initial regression model explaining household electricity consumption is given in equation (1).

    [lnE.sub.h] =[theta]+[alpha][R.sub.h] + [L.sub.'h][beta]+ [S.sub.'h][gamma]+ [F.sub.'h][delta]+[rho][lnP.sub.h] +[[epsilon].sub.h] (1)

    The dependent variable is the log of electricity consumption by household h, ln [E.sub.h]. The key explanatory variable is R, a binary variable equal to one for renting households and zero for non-renters.

    L is a vector of location variables, including binary variables for metropolitan status, neighborhoods that have natural gas, and climate zones. There are 11 climate zones in the United States based on the International Energy Conservation Code (IECC) classification (Baechler et al. 2015). S refers to socioeconomic and related characteristics. This includes series of binary variables for income and education categories, variables related to socioeconomic characteristics such as the physical size of houses in square...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT