 |
The NEMS Residential Demand Module forecasts future residential sector
energy requirements based on projections of the number of households and
the stock, efficiency, and intensity of use of energy-consuming equipment.
The Residential Demand Module projections begin with a base year estimate
of the housing stock, the types and numbers of energy-consuming appliances
servicing the stock, and the “unit energy consumption” by appliance (or
UEC—in million Btu per household per year). The projection process adds
new housing units to the stock, determines the equipment installed in new
units, retires existing housing units, and retires and replaces appliances.
The primary exogenous drivers for the module are housing starts by type
(single-family, multifamily and mobile homes) and Census Division and prices
for each energy source for each of the nine Census Divisions (see Figure
5). The Residential Demand Module also requires projections of available
equipment and their installed costs over the forecast horizon. Over time,
equipment efficiency tends to increase because of general technological
advances and also because of Federal and/or state efficiency standards.
As energy prices and available equipment changes over the forecast horizon,
the module includes projected changes to the type and efficiency of equipment
purchased as well as projected changes in the usage intensity of the equipment
stock.
The end-use services for which equipment stocks are modeled include space
conditioning (heating and cooling), water heating, refrigeration, freezers,
dishwashers, clothes washers, lighting, furnace fans, color televisions,
personal computers, cooking, and clothes drying. In addition to the major
equipment-driven end-uses, the average energy consumption per household
is projected for other electric and nonelectric appliances. The module’s output includes number of households, equipment
stock, average equipment efficiencies, and energy consumed by service,
fuel, and geographic location. The fuels represented are distillate fuel
oil, liquefied petroleum gas, natural gas, kerosene, electricity, wood,
geothermal, coal, and solar energy.
One of the implicit assumptions embodied in the Residential Demand Module
is that, through 2030, there will be no radical changes in technology or
consumer behavior. No new regulations of efficiency beyond those currently
embodied in law or new government programs fostering efficiency improvements
are assumed. Technologies which have not gained widespread acceptance today
will generally not achieve significant penetration by 2030. Currently available
technologies will evolve in both efficiency and cost. In general, at the
same efficiency level, future technologies will be less expensive than
those available today in real dollar terms. When choosing new or replacement
technologies, consumers will behave similarly to the way they now behave.
The intensity of end-uses will change moderately in response to price changes.
Electric end uses will continue to expand, but at a decreasing rate.7
Key Assumptions
Housing Stock Submodule
An important determinant of future energy consumption is the projected
number of households. Base year estimates for 2001 are derived from the
Energy Information Administration’s (EIA) Residential Energy Consumption
Survey (RECS) (Table 7). The forecast for occupied households is done separately
for each Census Division. It is based on the combination of the previous
year’s surviving stock with projected housing starts provided by the NEMS
Macroeconomic Activity Module. The housing stock submodule assumes a constant
survival rate (the percentage of households which are present in the current
forecast year, which were also present in the preceding year) for each
type of housing unit; 99.7 percent for single-family units, 99.8 percent
for multifamily units, and 97.5 percent for mobile home units. Projected
fuel consumption is dependent not only on the projected number of housing
units, but also on the type and geographic distribution of the houses.
The intensity of space heating energy use varies greatly across the various
climate zones in the United States. Also, fuel prevalence varies across
the country—oil (distillate) is more frequently used as a heating fuel
in the New England and Middle Atlantic Census Divisions than in the rest
of the country, while natural gas dominates in the Midwest. An example
of differences by housing type is the more prevalent use of liquefied petroleum
gas in mobile homes relative to other housing types.
Technology Choice Submodule
The key inputs for the Technology Choice Submodule are fuel prices by Census
Division and characteristics of available equipment (installed cost, maintenance
cost, efficiency, and equipment life). Fuel prices are determined by an
equilibrium process which considers energy supplies and demands and are
passed to this submodule from the integrating module of NEMS. Energy price,
combined with equipment UEC (which is a function of efficiency), determines
the operating costs of equipment. Equipment characteristics are exogenous
to the model and are modified to reflect both Federal standards and anticipated
changes in the market place. Table 8 lists capital cost and efficiency
for selected residential appliances for the years 2004 and 2020.
Table 9 provides the cost and performance parameters for representative
distributed generation technologies. The AEO2006 model also incorporates
endogenous “learning” for the residential distributed generation technologies,
allowing for declining technology costs as shipments increase. For fuel
cell and photovoltaic systems, learning parameter assumptions for the AEO2006
reference case result in a 13 percent reduction in capital costs each time
the number of units shipped to the buildings sectors (residential and commercial)
doubles.
The Residential Demand Module projects equipment purchases based on a nested
choice methodology. The first stage of the choice methodology determines
the fuel and technology to be used, the second stage determines the efficiency
of the selected equipment type. The equipment choices for cooling, water
heating, and cooking are linked to the space heating choice for new construction.
Technology and fuel choice for replacement equipment uses a nested methodology
similar to that for new construction, but includes (in addition to the
capital and installation costs of the equipment) explicit costs for technology
switching (e.g., costs for installing gas lines if switching from electricity
or oil to gas, or costs for adding ductwork if switching from electric
resistance heat to central heating types). Also, for replacements, there
is no linking of fuel choice for water heating and cooking as is done for
new construction. Technology switching upon replacement is allowed for
space heating, air conditioning, water heating, cooking and clothes drying.
Once the fuel and technology choice for a particular end use is determined,
the second stage of the choice methodology determines efficiency. In any
given year, there are several available prototypes of varying efficiency
(minimum standard, medium low, medium high and highest efficiency). Efficiency
choice is based on a functional form and coefficients which give greater
or lesser importance to the installed capital cost (first cost) versus
the operating cost. Generally, within a technology class, the higher the
first cost, the lower the operating cost. For new construction, efficiency
choices are made based on the costs of both the heating and cooling equipment
and the building shell characteristics.
The parameters for the second stage efficiency choice are calibrated to
the most recently available shipment data for the major residential appliances.
Shipment efficiency data are obtained from industry associations which
monitor shipments such as the Association of Home Appliance Manufacturers.
Because of this calibration procedure, the model allows the relative importance
of first cost versus operating cost to vary by general technology and fuel
type (e.g., natural gas furnace, electric heat pump, electric central air
conditioner, etc.). Once the model is calibrated, it is possible to calculate
(approximately) the apparent discount rates based on the relative weight
given to the operating cost savings versus the weight given to the higher
cost of more efficient equipment. Hurdle rates in excess of 30 percent
are common in the Residential Demand Module. The prevalence of such high
apparent hurdle rates by consumers has led to the notion of the “efficiency
gap”— that is, there are many investments that could be made that provide
rates of return in excess of residential borrowing rates (15 to 20 percent
for example). There are several studies which document instances of apparent
high discount rates.8 Once equipment efficiencies for a technology and
fuel are determined, the installed efficiency for its entire stock is calculated.
Appliance Stock Submodule
The Appliance Stock Submodule is an accounting framework which tracks the
quantity and average efficiency of equipment by end use, technology, and
fuel. It separately tracks equipment requirements for new construction
and existing housing units. For existing units, this module calculates
equipment which survives from previous years, allows certain end uses to
further penetrate into the existing housing stock and calculates the total
number of units required for replacement and further penetration. Air conditioning
and clothes drying are the two end uses not considered to be “fully penetrated.”
Once a piece of equipment enters into the stock, an accounting of its remaining
life is begun. It is assumed that all appliances survive a minimum number
of years after installation. A fraction of appliances are removed from
the stock once they have survived for the minimum number of years. Between
the minimum and maximum life expectancy, all appliances retire based on
a linear decay function. For example, if an appliance has a minimum life
of 5 years and a maximum life of 15 years, one tenth of the units (1 divided
by 15 minus 5) are retired in each of years 6 through 15. It is further
assumed that, when a house is retired from the stock, all of the equipment
contained in that house retires as well; i.e., there is no secondhand market
for this equipment. The assumptions concerning equipment lives are given
in Table 10.
Fuel Consumption Submodule
Energy consumption is calculated by multiplying the vintage equipment stocks
by their respective UECs. The UECs include adjustments for the average
efficiency of the stock vintages, short term price elasticity of demand
and “rebound” effects on usage (see discussion below), the size of new
construction relative to the existing stock, people per household and shell
efficiency and weather effects (space heating and cooling). The various
levels of aggregated consumption (consumption by fuel, by service, etc.)
are derived from these detailed equipment-specific calculations.
Equipment Efficiency
The average energy consumption of a particular technology is initially
based on estimates derived from RECS 2001. Appliance efficiency is either
derived from a long history of shipment data (e.g., the efficiency of conventional
air-source heat pumps) or assumed based on engineering information concerning
typical installed equipment (e.g., the efficiency of ground-source heat
pumps). When the average efficiency is computed from shipment data, shipments
going back as far as 20 to 30 years are combined with assumptions concerning
equipment lifetimes. This allows for not only an average efficiency to
be calculated, but also for equipment retirements to be vintaged—older
equipment tends to be lower in efficiency and also tends to get retired
before newer, more efficient equipment. Once equipment is retired, the
Appliance Stock and Technology Choice Modules determine the efficiency
of the replacement equipment. It is often the case that the retired equipment
is replaced by substantially more efficient equipment.
As the stock efficiency changes over the simulation interval, energy consumption
decreases in inverse proportion to efficiency. Also, as efficiency increases,
the efficiency rebound effect (discussed below) will offset some of the
reductions in energy consumption by increased demand for the end-use service.
For example, if the stock average for electric heat pumps is now 10 percent
more efficient than in 2001, then all else constant (weather, real energy
prices, shell efficiency, etc.), energy consumption per heat pump would
average about only 9 percent less.
Adjusting for the Size of Housing Units
Information derived from RECS 2001 indicates that new construction (post-1990)
is on average roughly 26 percent larger than the existing stock of housing.
Estimates for the size of each new home built in the projection period
vary by type and region, and are determined by a log-trend forecast based
on historical data from the Bureau of the Census.9 For existing structures,
it is assumed that about 1 percent of households that existed in 2001 add
about 600 square feet to the heated floor space in each year of the projection
period.10 The energy consumption for space heating, air conditioning, and
lighting is assumed to increase with the square footage of the structure.
This results in an increase in the average size of the housing stock from
1,705 to 1,977 square feet from 2001 through 2030.
Adjusting for Weather and Climate
Weather in any given year always includes short-term deviations from the
expected longer-term average (or climate). Recognition of the effect of
weather on space heating and air conditioning is necessary to avoid inadvertently
projecting abnormal weather conditions into the future. In the residential
module, adjustments are made to space heating and air conditioning UECs
by Census Division by their respective heating and cooling degree-days
(HDD and CDD). A 10 percent increase in HDD would increase space heating
consumption by 10 percent over what it would have otherwise been. Over
the projection period, the residential module uses a 30-year average for
heating and cooling degree - days by Census Division, adjusted by projections
in state population shifts.
Short-Term Price Effect and Efficiency Rebound
It is assumed that energy consumption for a given end-use service is affected
by the marginal cost of providing that service. That is, all else equal,
a change in the price of a fuel will have an opposite, but less than proportional,
effect on fuel consumption. The current value for the short-term elasticity
parameter is -0.15.11 This value implies that for a 1 percent increase
in the price of a fuel, there will be a corresponding decrease in energy
consumption of -0.15 percent. Another way of affecting the marginal cost
of providing a service is through altered equipment efficiency. For example,
a 10 percent increase in efficiency will reduce the cost of providing the
end-use service by 10 percent. Based on the short-term efficiency rebound
parameter, the demand for the service will rise by 1.5 percent (-10 percent
multiplied by -0.15). Only space heating and cooling are assumed to be
affected by both elasticities and the efficiency rebound effect.
Shell Efficiency
The shell integrity of the building envelope is an important determinant
of the heating and cooling load for each type of household. In the NEMS
Residential Demand Module, the shell integrity is represented by an index,
which changes over time to reflect improvements in the building shell.
The shell integrity index is dimensioned by vintage of house, type of house,
fuel type, service (heating and cooling), and Census Division. The age,
type, location, and type of heating fuel are important factors in determining
the level of shell integrity. Housing units which heat with electricity
tend to have less air infiltration rates than homes that use other fuels.
The age of homes are classified by new (post-2001) and existing. Existing
homes are characterized by the RECS 2001 survey and are assigned a shell
index value based on the mix of homes that exist in the base year (2001).
The improvement over time in the shell integrity of these homes is a function
of two factors—an assumed annual efficiency improvement and improvements
made when real fuel prices increase (no price-related adjustment is made
when fuel prices fall). For new construction, building shell efficiency
is determined by the relative costs and energy bill savings for several
levels of heating and cooling equipment, in conjunction with the building
shell attributes. The packages represented in NEMS range from homes that
meet the International Energy Conservation Code (IECC)12 to homes that
exceed the IECC by 50 percent. Shell efficiency in new homes would increase
over time if energy prices rise, or the cost of more efficient equipment
falls.
Legislation and Other Federal Programs
Energy Policy Act of 2005 (EPACT05)
The passage of the EPACT05 in August 2005 provides additional minimum efficiency
standards for residential equipment and provides tax credits to producers
and purchasers of energy efficient equipment and builders of energy efficient
homes. The standards contained in EPACT05 include: 190 watt maximum for
torchiere lamps in 2006; Dehumidifier standards for 2007 and 2012; and
ceiling fan light kit standards in 2007. For builders of homes that are
built 30 percent better than the latest code, a $1000 tax credit can be
claimed in 2006 and 2007. Likewise, builders of homes that are 50 percent
better than code can claim a $2000 credit over the same period. The builder
tax credits and production tax credits are assumed to be passed through
to the consumer in the form of lower purchase cost. EPACT05 includes production
tax credits for energy efficient refrigerators, dishwashers, and clothes
washers in 2006 and 2007, with dollar amounts varying by type of appliance
and level of efficiency met, subject to annual caps. Consumers can claim
a 10 percent tax credit in 2006 and 2007 for several types of appliances
specified by EPACT05, including: Energy efficient gas, propane, or oil furnaces or boilers,
energy efficient central air conditioners, air and ground source heat pumps,
hot water heaters, and windows. Lastly, consumers can claim a 30 percent
tax credit in 2006 and 2007 for purchases of solar PV, solar water heaters,
and fuel cells, subject to a cap.
Energy Policy Act of 1992 (EPACT92)
EPACT92 contains several policies which are designed to improve residential
sector energy efficiency. EPACT92 policies represented in the NEMS Residential
Demand Module include the sections relating to window labeling programs,
low-flow showerheads, and building codes. The impact of building codes
is captured in the shell efficiency index for new buildings listed above.
Other EPACT92 provisions, such as home energy efficiency ratings and energy-efficient
mortgages, which allow home buyers to qualify for higher loan amounts if
the home is energy-efficient, are voluntary, and their effects on residential
energy consumption have not been estimated.
The window labeling program is designed to help consumers determine which
windows are most energy efficient. These labels already exist for all major
residential appliances. Based on analysis of RECS data, it is assumed that
the window labeling program will decrease heating loads by 8 percent and
cooling loads by 3 percent. Approximately 35 percent of the existing (pre-2002)
housing stock is affected by this policy by 2030.
The low-flow showerhead program is designed to cut domestic hot water use
for showers. It is assumed that these showerheads cut hot water use by
33 percent for shower use. Since showers account for approximately 30 percent
of domestic hot water use, total hot water use decreases by 10 percent.
It is further assumed that these showerheads are installed exclusively
in new construction.
National Appliance Energy Conservation Act of 1987
The Technology Choice Submodule incorporates equipment standards established
by the National Appliance Energy Conservation Act of 1987 (NAECA). Some
of the NAECA standards implemented in the module include: a Seasonal Energy
Efficiency Rating (SEER) of 10.0 for heat pumps increasing to 13.0 in 2006;
an Annual Fuel Utilization Efficiency (energy output over energy input)
of 0.78 for oil and gas furnaces; an Efficiency Factor of 0.86 for electric
water heaters; increasing to .90 in 2004; and refrigerator standards that
set consumption limits to 976 kilowatt-hours per year in 1990, 690 kilowatt-hours
per year in 1993, and 510 kilowatt-hours per year in 2002.
Residential Technology Cases
In addition to the AEO2006 reference case, three side cases were developed
to examine the effect of equipment and building standards on residential
energy use—a 2005 technology case, a best available technology case, and
a high technology case. These side cases were analyzed in stand-alone (not
integrated with the supply modules) NEMS runs and thus do not include supply-responses
to the altered residential consumption patterns of the two cases. AEO2006
also analyzed integrated 2005 technology and high technology cases. The integrated 2005 technology case combines the 2005 technology cases of the
four end-use demand sectors, the electricity low fossil technology case,
and the assumption of renewable technologies fixed at 2005 levels. The integrated high technology case uses the same approach, but for high technology.
The 2005 technology case assumes that all future equipment purchases are
made based only on equipment available in 2005. This case further assumes
that existing building shell efficiencies will not improve beyond 2005
levels. In the reference case, the 2030 housing stock shell efficiency
is 10 percent higher than in 2003 for heating (6 percent for cooling).
The high technology case assumes earlier availability, lower costs, and/or
higher efficiencies for more advanced equipment than the reference case.
Equipment assumptions were developed by engineering technology experts,
considering the potential impact on technology given increased research
and development into more advanced technologies.13 In the high technology
case, heating shell efficiency in 2030 increases by 22 percent and cooling
shell efficiency by 10 percent, relative to 2003.
The best available technology case assumes that all equipment purchases
from 2006 forward are based on the highest available efficiency in the
high technology case in a particular simulation year, disregarding the
economic costs of such a case. This case is designed to show how much the
choice of the highest-efficiency equipment could affect energy consumption.
In this case, heating shell efficiency in 2030 increases by 26 percent
and cooling shell efficiency by 11 percent, relative to 2003.
Residential Tables
Residential Notes |