3. The Residential Sector

Introduction

More than 90 million single-family, multifamily, and mobile home households encompass the residential sector. Households use energy to cool and heat their homes, to heat water, and to operate many appliances such as refrigerators, stoves, televisions, and hot tubs.

The energy sources utilized by the residential sector include electricity, natural gas, fuel oil, kerosene, liquefied petroleum gas (propane), coal, wood, and other renewable sources such as solar energy.⁵

Early in the 1980's total energy consumption in the residential sector started to decline, although personal consumption expenditures continued to increase (Figure 3.1). Since this slight dip in the early 1980's, energy use has remained near its 1980 level to the present time, while personal consumption expenditures have continued to rise. During this time, when personal consumption expenditures were increasing, the percent of energy in total personal consumption expenditures was falling. Additionally, even though households were purchasing more energy for appliances such as televisions and dishwashers, many of the appliances were using less energy than older models.

The National Appliance Energy Conservation Act (NAECA) of 1987 mandated minimum energy efficiency standards for several types of household appliances and equipment such as refrigerators, freezers, room air conditioners, television sets, furnaces, water heaters, and heat pumps. This followed the earlier voluntary appliance efficiency targets of the Energy Policy and Conservation Act (EPCA) of 1975 and various State appliance-efficiency standards. In response to these various standards, manufacturers have improved the energy efficiency of household appliances and equipment over the past 20 years.

As, Figure 3.1 shows, an increasing gap has developed between primary and site energy consumed in the residential sector. Site electricity consumption as a percent of total site energy has climbed from 27 percent in 1982 to 33 percent in 1993. ⁶ Greater usage of electricity, due mainly to more widespread use of electric heat pumps, central air conditioning, and appliances, is largely responsible for this widening gap. The losses in the generation, transmission, and distribution of electricity are more than twice the amount of electricity delivered to the household. These losses are incorporated into the primary energy estimates.

Chapter Organization

In this chapter, the major data source, EIA's Residential Energy Consumption Survey, is described first. This is followed by a discussion of energy consumption in the residential sector and the necessity of adjusting consumption for changes in weather before any analysis is undertaken. Next, the demand indicators--households, buildings, household members, and floorspace--are described along with the trends in these demand indicators. Four energy-intensity indicators are presented, followed by a discussion of the strengths and limitations of these energy-intensity indicators.

Major Data Sources

Energy Information Administration

Residential Energy Consumption Survey (RECS)

The RECS is conducted every 3 years by the EIA. It is a national sample survey of more than 5,000 residential housing units and their energy suppliers. The RECS is the only comprehensive source of national-level data on energy-related information such as energy consumption for the residential sector.

The RECS is a national multistage probability sample survey. Housing unit and household characteristics data are collected via personal interview with the householder. Householders are asked to sign authorization forms allowing their suppliers of energy to release billing information about their households. A mail survey is used to collect this information on energy consumption from the energy suppliers. Using the RECS data, end-use estimates for space heat, air conditioning, water heating, and appliances are obtained by disaggregations of actual bills using a nonlinear technique. ⁷

Energy Consumption in the Residential Sector

Background

Residential energy provides numerous services associated with household living, including space heating and cooling, water heating, cooking, refrigeration, lighting, and the powering of a wide variety of other appliances. Space heating is the dominant energy service. In 1990, 52 percent of the total site energy was used to provide space heating, followed by 25 percent used for appliances (Figure 3.2).

Energy demand is shaped by a variety of factors, including type of housing unit, location, and climate. The dominant housing type in the residential sector by far is the single-family detached housing unit. In 1990, 62 percent of the 94 million housing units were detached single-family units (Figure 3.3). Single-family detached housing units consumed 72 percent of total site energy (Figure 3.4). In fact, single-family detached homes are the most energy-intensive housing type, consuming 113 million Btu per household compared with an average of 98 million Btu per household for all housing types in 1990.

Location is an important factor in determining energy consumption. Different regions of the country face different weather patterns, building codes, energy-related behavior, fuel mixes, and equipment needs and suitability. In 1990, the Midwest used 31 percent of the total site energy, whereas the West used only 16 percent, half the amount that the Midwest used with only 16 percent fewer households (Figure 3.4). Weather played a large part in this difference. Eighty-five percent of the housing units in the Midwest were located where heating-degree-days (HDD) ⁸ for 1990 exceeded 5,500. On the other hand, only 24 percent of housing units in the West were located where the HDD's for 1990 were more than 5,500.

Energy consumption in the residential sector is examined for two time intervals, a growth/ growth interval (1984-1987) and a growth/ recession interval (1987-1990). ⁹

Energy Trends

Among all households, the growth of total site energy consumption was nearly stable, with a recorded growth of approximately 1 percent per time interval (Figure 3.5). However, this pattern of growth was not the same for each of the individual housing types. Total site energy consumption grew during the growth/growth interval for each of the housing types with the exception of multifamily units (Figure 3.6). This trend continued partially during the growth/recession interval and reductions in total site energy consumption were even more pronounced for all households in large multifamily buildings. From 1987 to 1990, the number of occupied units in large buildings decreased by about half a million.¹⁰ Attached single-family housing units, which had experienced growth in total consumption during the growth/growth interval, actually experienced reductions during the growth/recession interval.

The weather was mild in the years studied (1984, 1989, and 1990) compared to 30-year averages. In 1990, the weather was 14 percent warmer than the average. In 1987, the weather was 8 percent warmer than the average while in 1984, the weather was 2 percent colder than the average. When weather effects like these are not considered, reductions in total energy consumption due to milder weather, could be regarded as increases in energy efficiency when indeed, they are not.

Total residential site energy increased by 2 percent from 1984 to 1990, but adjusting for weather suggests that if the weather had been average, total energy consumption would have grown by 9 percent during 1984-1990. Comparison of weather-adjusted trends against the unadjusted total energy consumption trends suggests that the weather-induced "conservation" should be removed before any attempt is made to develop energy-intensity indicators.

In this chapter, total site energy consumption is weather-adjusted before any attempt is made to present energy-intensity indicators. Weather-adjusted total site energy consumption is calculated for each Census region. For 1984, 1987 and 1990, space-heating energy consumption is adjusted by a heating-degree day (HDD) factor and the air- conditioning demand is adjusted by a cooling-degree day (CDD) factor, to which was added unadjusted demand for all other end uses. The methodology used to develop the HDD and CDD factors and adjusted total site energy consumption estimates is presented along with examples in the residential section of Appendix A under "Degree-Day-Adjusted Estimates."

As seen in Figure 3.7 covering intervals of growth/growth and growth/recession, if total site energy consumption is not adjusted for the mild weather, growth in total site energy consumption is underestimated.

The main demand indicators or drivers for energy services in the residential sector are the number of households, population, buildings, and the amount of floorspace (Figure 3.8).

Trends in Demand Indicators

Households

The housing stock has changed significantly since 1984. The number of households grew by 4.9 percent during the growth/growth interval and 3.8 percent during the growth/recession interval.

Much of this growth was in the South. Figure 3.9 shows that the greatest percentage of older households (at least 10 years old) is largest in the Northeast and smallest in the South. The growth in the construction of new housing units in the South is also one reason for the increase in households with air-conditioning. New housing units are more apt to have central air-conditioning. Between 1984 and 1990, the number of air-conditioned households throughout the country increased by almost 24 percent (Table 3.1). Throughout the period, the fuel mix also changed in United States households. The increased use of the electric heat pump, especially in the South, drove the 48 percent increase in the use of electricity as the main heating fuel.

Population

The number of households is growing faster than the U.S. population. The population grew by only 6 percent during 1984-1990, whereas the number of households grew by 9 percent. This led to a decline in household size from 2.73 members per household in 1984 to 2.65 members per household in 1990.

Furthermore, in 1990, 25 percent of the households had only one member--in 1984 this percent was 23.6. ¹¹

Floorspace

Total residential floorspace increased by 17 percent between 1984 and 1990, much faster than the growth in the number of households. The size of housing units has grown over time from 1,673 square feet per household in 1984 to 1,800 square feet per household in 1990.¹²

Buildings

The number of residential buildings between 1984 and 1990 grew faster than the number of households (9.8 percent growth for buildings and 8.9 percent for households), indicating greater construction of single-family homes. This was the case during the growth/recession interval but less strong during the growth/growth interval between 1984 and 1987 (Table 3.1).

Demand indicators in the residential sector are influenced by behavioral and structural effects that can mask the effects of energy-efficiency changes. The major behavioral influences on residential energy consumption include changes in family size, income, average length of daily occupancy, age of household members, thermostat setting, and the number of employed members of household.¹³  Structural changes include household size (floorspace), building age, location (region), and fuel mix used.

These behavioral and structural effects must be taken into account in comparisons of energy-intensity indicators over time. For example, increases in floorspace per household member (612 square feet per household member in 1984 to 680 square feet in 1990) increase the demand for space heating and air conditioning, lighting, and convenience appliances such as dishwashers.

A behavioral change affecting energy consumption is the growing number of single-member households. Declines in household size also result in more floorspace per person, higher heating demand, and lower water heating demand per person. Households with fewer members will often acquire the same number and size of major appliances, and consume more energy per household member than a larger household. At the same time, though, single persons may spend less time at home than families, resulting in fewer hours demanding energy.

Over the whole population, disposable income per capita has increased. During the growth/growth interval, disposable income in constant dollars increased 16 percent whereas during the growth/recession interval, this growth was limited to 4 percent.¹⁴ While single-member and two-person households may spend less time at home, thereby using less energy, they and larger households may also increase the demand for energy services if disposable incomes have increased. Purchase of such energy-using equipment as computers, hot tubs, home theater systems, and swimming pools increase residential energy consumption.

Illustrated are only a few of the behavioral and structural influences that affect the demand for energy services in the residential sector. The next section presents choices of energy-intensity indicators for the residential sector. The behavioral and structural influences described in this section need to be considered when the comparisons of the energy-intensity indicators are presented over time.

Energy-Intensity Indicators for the Residential Sector

Energy consumption and the drivers of the demand for energy in the residential sector, the demand indicators, have been presented in detail. The next step is to construct energy-intensity indicators for the residential sector. Box 3.2 lists the indicator choices presented in this section.

Each of the indicators have their own strengths and limitations, which will be discussed in the next section. Additionally, although the indicators presented in this chapter are based only on site energy consumption, the indicators may also be presented based on end-use site energy consumption such as space heating. However, not all of the demand indicators are suitable to use in the development of end-use intensity indicators. For example, the energy used for space heating may be presented as space-heating energy per square foot, but it would not make sense to use the square foot demand indicator to develop a water-heating intensity indicator. A more suitable energy-intensity indicator would be water-heating energy per household member.

Energy-Intensity Indicator Trends

Table 3.2 shows estimates for the various energy-intensity indicators for 3 of the RECS years. Two sets of estimates are presented: unadjusted and weather-adjusted. Given the variety of indicators available, caution is warranted in the interpretation of residential energy-intensity indicators since the magnitude and direction of the energy-intensity indicator changes are dependent on the choice of the demand indicator. Figure 3.10a shows the changes in these indicators over the growth/growth interval. In the absence of adjusting consumption for weather deviations from the normal average, energy-intensity appears to fall during the growth/growth interval, and energy efficiency may have increased, no matter which indicator is used.

When weather adjustments are made, the decreases in energy-intensity are less pronounced. The story is slightly different when comparing changes in these energy-intensity indicators over the growth/recession interval (Figure 3.10b). All of the energy-intensity indicators registered a decrease in intensity, suggesting an increase in energy efficiency. Weather-adjusted energy-intensity indicators, with the exception of million Btu per square foot, actually registered increases in energy- intensity suggesting decreases in energy efficiency over the growth/recession interval.

Individually, the demand indicator, square feet, registered the largest percent increase over the growth/growth interval, partially causing the energy-intensity indicator, million Btu per square foot, to register the largest percentage decrease in energy-intensity. Underlying this was the growth in the average size of the housing units, a structural change. Conversely, the smallest decrease in an energy-intensity indicator, million Btu per person, reflected the fact that the growth in population registered the slowest percent change of all of the demand indicators. When adjustments were made for weather effects, this intensity indicator actually increased.

Some of the decreases in the intensity indicators during both the growth/growth and growth/recession intervals may be a reflection of energy-efficiency increases. However, the above discussion gives the reader a clearer understanding of the difficulties in trying to assess energy-intensity changes as a reflection of changes in energy efficiency. Efficiency gains or reductions are taking place along with structural and behavioral changes. To unbundle the intensity indicators and obtain "true" or "pure" energy efficiency is impractical if not impossible. Adjustments such as the degree-day adjustments and awareness of the behavioral and structural influences are about the most that may be done, especially in an environment of limited data, time, and resources. Standardizations can be done to take care of the structural changes such as: changing distribution of the household by household type.

Another aspect in the development of energy-intensity indicators is the ability to make comparisons within a particular indicator, such as million Btu per household, but over characteristics such as type of housing unit. These comparisons may be more fruitful. The million Btu per household is not a particularly robust indicator in that energy use per household incorporates changing household member compositions, size, and housing type distributions. The energy-intensity indicators may become more robust if the indicator is decomposed by characteristics, such as housing type.

Table 3.3 does just that for two characteristics, type of housing unit and Census region. One can see the differences in the energy-intensity indicator across a particular characteristic, whether the energy-intensity indicator was unadjusted or weather adjusted. Taking the characteristic, type of housing unit as an example, will demonstrate the advantages of comparing within an energy-intensity indicator and across the characteristic. The more detailed Amicro-residential energy-intensity indicator is more robust.

Table 3.3. Unadjusted and Weather-Adjusted Residential Energy-Intensity Indicator, 1984, 1987, and 1990
(Million Btu per Household)

The largest decreases in the intensity indicator, million Btu per household, was in the single-family attached and large multifamily housing units (Figure 3.11). Since 1984, the number of occupied large multifamily housing units declined by almost a half-million units in 1990 as well as total consumption for all large multifamily units.

After weather adjustments, intensity actually increased during the growth/recession interval in the single-family detached and smaller multifamily housing units.

Strengths and Limitations of the Energy-Intensity Indicators

Four energy-intensity indicators were presented in this chapter that may be used as the basis for the measurement of energy efficiency. All four indicators are imperfect. One imperfection can easily be addressed: the influence of changes in the weather. Adjusting the intensities for weather, especially since the recent years have been mild, can explain a considerable portion of the reductions in these energy-intensity indicators. Structural and behavioral influences affect all four indicators, some more than others. Energy use per person account for population growth but does not address other issues. Energy per household does not accounts for the expansion in household floorspace whereas energy per square foot does. Energy per square foot may be appropriate for some end uses, e.g., space-heating energy, but not for others such as water heating energy.

No single energy-intensity indicator for the residential sector stands out as clearly superior to the others. The choice of indicator depends on the questions asked and on data and resource availability.

End Notes

⁵In this chapter, total site energy consumption includes only: natural gas, electricity, liquefied petroleum gas, fuel oil, and kerosene.

⁶Based on preliminary data from the 1993 Residential Energy Consumption Survey (RECS) since final 1993 RECS data were not available at the time of this analysis.

⁷See Household Energy Consumption and Expenditures 1990, Appendix D, "End-Use Estimation Methodology," for details on the procedures used to calculate the end-use estimates. This EIA publication also describes the RECS sample design, data collection procedures, and limitations of the RECS data.

⁸See Heating-Degree-Days and Cooling-Degree-Days in the general terminology section of the Glossary.

⁹The word "recession" is used to describe a period of slow or not economic growth. A recession is officially defined as two consecutive quarters of falling Gross National Product.

¹⁰RECS does not collect data on vacant or seasonal housing units.

¹¹Energy Information Administration, Housing Characteristics 1990, DOE/EIA-0314(90), Table 11; Housing Characteristics 1984, DOE/EIA-0314(84), Table 17.

¹²Energy Information Administration, Housing Characteristics 1990, DOE/EIA-0314(90), Table 15; Housing Characteristics 1984, DOE/EIA-0314(84), Table 20.

¹³U.S. Congress, Office of Technology Assessment, Building Energy Efficiency, OTA-E-518, May 1992, p. 20.

¹⁴U.S. Department of Commerce, Bureau of Economic Analysis, Survey of Current Business Patterns, August issues.

Table of Contents

Commercial Building Sector

Contacts

Specific questions on this topic may be directed to:

Stephanie Battles

Stephanie.Battles@eia.doe.gov

(Phone: (202) 586-7237)

FAX: (202) 586-0018

Contact Us

URL: http://www.eia.doe.gov/emeu/efficiency/ee_ch3.htm

File Last Modified: October 17, 1999