Have residential building energy codes and standards reduced household energy consumption- and if so, by how much?

1

• Have residential building energy codes and
  standards reduced household energy consumption-
  and if so, by how much?
• Meredith Fowlie
• discussant
• Conference on
• Green Building, The Economy, and Public
  Policy
• UC Berkeley
• December 3, 2009

2
Jacobsen and Kotchen (2009)

• Estimate the household-level energy consumption
  impacts of a change in Floridas building code
  that took effect in 2002.
• Treated group New homes in Gainesville FL
  built between 2003 and 2005.
• Control group New homes in Gainesville FL
  built between 1999 and 2001.
• To implement their empirical strategy, they use
  (terrific!!) household-level data made available
  under the auspices of the Gainesville Green
  program,
• They find economically and statistically
  significant impacts on average annual electricity
  and natural gas consumption (4 and 6 percent,
  respectively).
• They argue convincingly that the data are
  consistent with efficiency improvements in
  heating and cooling, (versus loads that do not
  vary seasonally).

3
Ex ante expected increases in code stringency by
end use and climate
End Use Florida Climate Zone Florida Climate Zone Florida Climate Zone
End Use North Central South
Cooling 16 13 12
Heating -8 46 45
Hot Water 0 0 0
Overall 4 15 10
Source http//www.energygauge.com/FLARES/new_cod
e.htm
Because these sources of energy demand account
for only half of the energy demand of a typical
Florida residence, the change in code stringency
translates into a 2 percent increase in overall
energy efficiency.
4
Gainesville is increasingly green(independent of
the policy change in question?)

• In October 2002, the City of Gainesville
  introduced the "Gainesville Green Building
  Program," becoming the first city or county
  government in Florida to incentivize green
  building (in the public and private sector).
• Gainesville is one of the most productive markets
  for Energy Star homes nationwide numbers have
  been steadily increasing (Smith and Jones, 2003).
  

Electricity Electricity Natural gas Natural gas
Annual Summer months Annual Winter months
Smith and Jones (2003) 16 23 17 45
Jacobsen and Kotchen (2009) 4 6-8 6 15-25
5
Aroonruengsawat, Auffhammer, and Sanstad (2009)

• Authors consider the average impacts of
  residential building codes introduced nationwide,
  1970-2006.
• Compare annual, state-level residential energy
  consumption across states and time periods
  different energy building code regimes.
• Recognizing that a states choice of building
  code regime could be endogenous, they employ an
  instrumental variables strategy.
• Estimate average reductions in residential
  electricity use of 3-5...Why not look at
  residential natural gas consumption?

6
Estimating the effects of building codes

• SHARE coefficient is hard to interpret because it
  picks up the combined effect of new construction
  activity and new building codes.
• Turnover/vintage of housing stock is an important
  determinant of residential electricity use
  (independent of building codes).

7
Addressing endogeneity concerns

• Fixed effects control for some omitted
  variables.. but not potentially significant,
  omitted factors whose effects vary with time.
• Instruments are plausibly correlated with
  omitted, time-variant factors (such as private
  investment in efficiency measures, other
  state-level efficiency programs, etc).
• Why not also instrument for intensity?

8
Building codes are neither randomly assigned
nor implemented in isolation
Policy category Weight
Utility demand side management/public benefit programs 20
State government efficiency initiatives 7
Building energy codes 7
Combined heat and power 5
State appliance and equipment efficiency standards 3
Source The 2009 State Energy Efficiency
Scorecard. October 2009. www.aceee.org

• State performance in different efficiency policy
  categories is evaluated annually by ACEEE.
• Categories are weighted based on approximate
  savings impacts.
• State scores tend to be highly correlated across
  categories.
• Some attempt to control for DSM programs seems
  warranted?