case studies on top-down methods: case of solar … studies on top-down methods: case of solar water...
TRANSCRIPT
Case studies on top-down methods: case of solarwater heaters and new cars
Didier Bosseboeuf, ADEME
Bruno Lapillonne, Enerdata
Nathalie Desbrosses, Enerdata
European Expert meeting,
La Colle-sur-Loup
June 5 2007
List of selected case studies (1/2)
Residential sector
(i) Building shell and heating systems (energy consumption indicator)
(ii) Household electricity use excluding thermal uses (ie electric appliances as
a whole including lighting) (energy consumption indicator)
(iii) Specific white goods (e.g. cold appliances, dryers) (market diffusion indicator)
(iv) Solar thermal collectors (market diffusion indicator)
Transport sector
(i) New cars (energy consumption indicator)
(ii) Improvement of the car, bus and truck stock (energy consumption indicator)
(iii) Modal shift in passenger transport (energy consumption indicator/ modal
split indicator)
(iv)Modal shift in goods transport (energy consumption indicator/ modal split
indicator)
List of selected case studies(2/2)
Industry sector(i) Industrial thermal energy use (excluding electricity) (energy consumption
indicator)
(ii) Industrial electricity consumption (energy consumption indicator)
(iii)Industrial CHP (market diffusion indicator)
Tertiary sector(i) Building shell and heating systems (energy consumption indicator)
(ii) Electricity end- uses excluding thermal uses
General policy instruments(i) Energy taxation
(ii) Focused information campaigns with high impact
Pilot case studies to test the methodology
Pilot case studies selected so as to have one based on a marketdiffusion indicator and one based on an energy consumptionindicator
Select end use with a good data coverage :Solar thermal collectors (market diffusion indicator)
New cars (energy consumption indicator)
Comparison of approaches:
With market diffusion indicators, need to express them in terms of
energy savings whereas with energy consumption indicators energy
savings obtained directly
Market diffusion indicators more rapidly updated (e.g. 2006 available for
solar compared to 2004 for new cars)
Estimation of energy savings linked to policy
measures: why a top down evaluation
Energy savings linked to the development of solar water heaters can be
both assessed with bottom-up and top-down methods. Why consider it with
top down methods?
With bottom-up approach, assessment from a given policy measure (e.g. if
grant available, installed capacity derived from the amount of grants given)
However in most countries several measures are often available at the
same time (e.g. In France tax credit on the cost of the equipment +
subsidies from regional/local organisations + soft loans+ reduced VAT on
the cost of installation
risk of double counting if looking at policy measures individually
ex post evaluation of measures not always available (e.g. tax credit)
it may be easier to look at the overall market development (e.g. sales of water
heaters or installed capacity)
Solar water heaters
Estimation of energy savings linked to policy
measures: methodology for solar heaters (1/2)
Development of solar water heaters will be measured from the installed
stock in m2
The diffusion of solar water heaters and hence the related energy
savings will be explained by the following factors/ variables:
Autonomous trend
Energy price
Energy policy measures (subsidies, tax credit) (After / before 1995)
residual hidden structure effects and direct rebound effects neglected
Total energy savings are calculated by multiplying the number of m2 by
an amount of energy saving per m2 depending on the country
ESD energy savings calculated by difference: totals savings minus
trend and price related savings
Estimation of energy savings linked to policy
measures: methodology for solar heaters (2/2)
Identification by country of a first period over which policy
measures either are negligible or have a limited impact (graphic, plus
analysis of policy measures) by country over that period diffusion
mainly linked to autonomous trend and possibly energy prices
Modelling over that period of the diffusion of solar water heaters
through regression analysis with two variables:
Time to capture an autonomous trend
Average price of energies used for water heating to measure the
impact of prices
Ln (IC) = T X ln (t) + A X ln (P) + K
T: trend
A: price elasticity (>0 as price increase should increase penetration of solar
water heaters)
P: energy price
Solar heaters installed stock (m2): Odyssee from
Observ’er, plus:
• European Solar Thermal Industry Federation
• IEA Solar Heating and Cooling Programme
Energy price: Odyssee+ Enerdata database (based on IEA, Eurostat, nationaldata
Average price of energies used for water heating
or average household price
or price of the dominant source for water heating
Energy policy measures
MURE database, plus:
• IEA global renewable energy policies and measures database
• National sources as transversal sources are not always exhaustive (inparticular for new member countries)
Source of data and information on energy
policy measures for solar heaters
Classification of countries
Three groups of countries:
“Policy pushed markets”: countries for which a policy was implemented after
the mid –nineties period with autonomous trend clearly visible (net and rapid
take off from one year):
Case of France, Belgium, Italy, Finland, Ireland, Netherlands, Spain, UK, Hungary,
Sweden, Portugal
“New markets”: recent policies (same as above but diffusion starting from a
very low level close to zero no autonomous trend ):
Case of most new EU member countries
“Mature markets” with already a high penetration of solar water heaters
resulting from policies implemented before 1995 most difficult case:
what is eligible with respect to ESD?
what is part of the trend and what is linked to policies
Case of Germany, Austria, Cyprus, Greece, Slovenia and Denmark
Policy pushed markets (1/2):
Acceleration of the diffusion when a policy measure isimplemented
Belgium
0
2
4
6
8
10
12
1990 1992 1994 1996 1998 2000 2002 2004 2006Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Finland
0
1
1
2
2
3
3
4
1990 1992 1994 1996 1998 2000 2002 2004 2006Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
France
0
2
4
6
8
10
12
14
16
18
20
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Ireland
0
1
1
2
2
3
3
4
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Italy
0
2
4
6
8
10
12
1990 1992 1994 1996 1998 2000 2002 2004 2006Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Netherlands
0
5
10
15
20
25
30
35
40
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
For sake of comparison all graphs are shown in terms of installed area in m2 per 1000 capita
Policy pushed markets (2/2):
Acceleration of the diffusion when a policy measure is implemented
Spain
0
5
10
15
20
25
1990 1992 1994 1996 1998 2000 2002 2004 2006Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
United Kingdom
0
1
1
2
2
3
3
4
4
5
1990 1992 1994 1996 1998 2000 2002 2004 2006Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Sweden
0
5
10
15
20
25
30
35
1990 1992 1994 1996 1998 2000 2002 2004 2006Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Portugal
0
2
4
6
8
10
12
14
16
18
1990 1992 1994 1996 1998 2000 2002 2004 2006Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Hungary
2
3
3
4
4
5
5
1990 1992 1994 1996 1998 2000 2002 2004 2006Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
For sake of comparison all graphs are shown in terms of installed area in m2 per capita
Policy pushed markets: case of FranceHistorical development (1/4)
0
2
4
6
8
10
12
14
16
18
20
1990 1992 1994 1996 1998 2000 2002 2004 2006 In
sta
lle
d s
toc
k p
er
ca
pit
a (
m2
/10
00
in
ha
b)
Solar water heater plan in
DOM
Solar planReinforcement
of Solar plan
Policy pushed markets: case of FranceModelling (2/4)
Regression with autonomous trend and average price of energies used for
water heating between 1990 and 2000
Ln (IC) = 0.27 X ln (t) – 0.07 X ln (P) + 1.96 + et
t-stat (6.6) (0.2)
R = 0.96 => Good correlation (R near 1)
F-stat = 104 => regression is globally significant (F-statistic is > 4.5 )
T-stat > 1.9 for time but <1.9 for prices and negative elasticity=> price effect not
significant
Regression with autonomous trend only between 1990 and 2000
Ln(IC) = 0.28 X ln(t) + 1.54 + et
t-stat (15.2) , R = 0.96, F-stat = 232
Good correlation, regression globally significant
Policy pushed markets: case of France (3/4)
Modelling of installed capacities in m2
0
2
4
6
8
10
12
14
16
18
20
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
sto
ck
pe
r c
ap
ita
(m
2/1
00
0 i
nh
ab
)
Real data Trend + prices Trend only
Solar water heater plan in
DOM
Solar planReinforcement
of Solar plan
Policy pushed markets: case of FranceEnergy savings (4/4)
0
5
10
15
20
25
30
35
40
kto
e
2000 2001 2002 2003 2004 2005 2006
Total energy savings Autonomous progress effect
Energy savings due to policies
Annual energy savings calculated from installed area of solar collectors and a coefficient in toe/m2
(useful energy provided by the solar energy); could be also calculated in terms of final energy
replaced (e.g. electricity or gas)
Annual energy savings from solar water heaters
Mature markets: Old policies and/or continuous policies
Austria
0
50
100
150
200
250
300
350
400
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Greece
0
50
100
150
200
250
300
350
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Germany
0
20
40
60
80
100
120
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Denmark
0
10
20
30
40
50
60
70
80
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
What is the autonomous trend? When does the regression end? What is eligiblepolicy according to ESD?
Cyprus
0
100
200
300
400
500
600
700
800
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Slovenia
30
35
40
45
50
55
60
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
For sake of comparison all graphs are shown in terms of installed area in m2 per capita
Mature markets: case of Germany Historical development (1/4)
0
20
40
60
80
100
120
1990 1992 1994 1996 1998 2000 2002 2004 2006Ins
tall
ed
sto
ck
pe
r c
ap
ita
(m
2/1
00
0 i
nh
ab
)
EPR-Environment and Energy Saving Programme
Market stimulation Programme
Solarthermie 2000
Solarthermie 2000Plus
Several policies implemented for solar water heaters ( 1993, 1995, 1999), plus
ecological tax in 1999. Over which period do we do the regression? The diffusion
mainly significant since 1999.
Old
policies
Mature markets: case of Germany(2/4): installed capacities
0
20
40
60
80
100
120
1990 1992 1994 1996 1998 2000 2002 2004 2006
In
sta
lle
d s
toc
k p
er
ca
pit
a (
m2
/10
00
in
ha
b)
Real data Price + trendTrend only Price w/o Ecotax + trend
EPR-Environment and Energy Saving Programme
Market stimulation Programme
Solarthermie 2000
Solarthermie 2000Plus
Regression until 1999: Ln (IC) = 1.08 X ln (t) + 1.03 X ln (P) - 4.96 + t
Old
policies
Impact of
measures
(incl ecotax)
Impact of
market
prices
Mature markets: case of Germany Energy savings (4/4)
0
20
40
60
80
100
120
140
160
180
200k
toe
2000 2001 2002 2003 2004 2005 2006
Total energy savings Autonomous progress effect
Prices effect Ecotax effect
Policy and measures effect
Annual energy savings from solar water heaters
Energy savings linked to policies include ecotax (around 20% of total savings) ;
limited impact of ecotax (equal around 5% of average energy price since 1999)
New markets: Recent policies
Is the beginning of the diffusion due to a policy or is there a data problem?
Luxembourg
0
5
10
15
20
25
30
35
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Czech Republic
0
1
2
3
4
5
6
7
8
9
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Estonia
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Latvia
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Lithuania
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Malta
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Poland
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
Slovak Republic
0
2
4
6
8
10
12
14
1990 1992 1994 1996 1998 2000 2002 2004 2006
Ins
tall
ed
ca
pa
cit
y p
er
ca
pit
a (
m2
/ca
p)
New markets :Malta Case Study
0
10
20
30
40
50
60
70
1990 1992 1994 1996 1998 2000 2002 2004 2006Ins
tall
ed
sto
ck
pe
r c
ap
ita
(m
2/1
00
0 i
nh
ab
)
Top down assessment of energy savingsfor solar heaters: first conclusions (1/2)
No data limitations: all countries are covered even if for most EU NewMember Countries data start in the late nineties
Definition of trends or baseline difficult for countries with maturemarkets
Role of energy price variable: negligible for France, very high inGermany:
quality of the estimate of price elasticity questionable as shortperiod ;
use of same price elasticity for all countries? (default valuecalibrated on a longer period and all countries)
Impact of price elasticity (case of Germany)
0
20
40
60
80
100
120
1990 1992 1994 1996 1998 2000 2002 2004 2006
In
sta
lled
sto
ck p
er
cap
ita (
m2/1
000 in
hab
)
Real data Trend onlyPrice + trend (elasticity = 1) Price + Trend (elasticity = 0,3)
Impact of
measures
(elasticity 0.3)
Impact of
measures
(elasticity 1)
Price elasticity of 0.3 about EU average
Top down assessment of energy savingsfor solar heaters: first conclusions (2/2)
Need of country specific coefficient of energy saving to account fordifference in solar flows; however decision has to be taken on the way tomeasure the savings; in terms of energy displaced? In terms of solarinputs?
No need to account for the lifetime of energy savings: removal/replacement of solar heaters at the end of their lifetime implicitly takeninto account in the measurement of the installed solar area
Similar approach seems applicable for other case studies with marketdiffusion indicators
New cars
Estimation of energy savings linked to policy
measures: methodology for new cars (1/3)
Change in new car efficiency measured from the test specific consumption
of new cars sold every year in litres/100km
Trends in the specific consumption of new cars and hence the related
energy savings can be explained by the following factors:
Change in the average size of vehicles (in terms of weight, or horse power or
engine size in cm3) (“hidden structure effect”) ( towards larger or more powerful cars,
energy savings are underestimated
Autonomous trend (in technical efficiency)
Motor fuel price
EU policy (ACEA/JAMA/KAMA agreement) and national energy policy
measures (tax on motor fuels, subsidies/ tax on vehicles): after 1995/ before
1995
Effect of change in the size of vehicles interesting to consider but limited in
practice due to data availability ; only tested for one or two countries
Direct rebound effect not taken into account at this stage (may be introduced later
through the annual distance travelled by new car)
Estimation of energy savings linked to policy
measures: methodology for solar heaters (2/3)
Identification of the period over which policy measures either are
negligible or have a limited impact so as to get the autonomous trend
4
5
6
7
8
9
10
11
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
l/1
00
km
Germany UK Italy Spain France Netherlands EUR15
Very different trends
before and after 1995,
mainly as result of the
ACEA/JAMA/KAMA
agreement
What is the
autonomous trend:
before or after 1995?
Specific consumption of new cars (test value)
Estimation of energy savings linked to policy
measures: methodology for solar heaters (3/3)
Modelling over that period of the trend in the specific consumption of cars (SC)(by country) (in litre/100km)* :
Time to capture an autonomous trend
Average price of gasoline and diesel
To clean the impact of fuels substitution between gasoline and diesel, totalenergy savings will be calculated separately for gasoline and diesel vehicles aswell for alternative fuels vehicles and then added together.
Ln (SC) = T X ln (t) + A X ln (P) + KT: trend
A: price elasticity (<0)
P: motor fuel price
The energy savings associated to price changes will then be split into twocomponents: energy savings linked to tax increase (policy related) and savings(>0 or <0 savings depending on the variation) linked to change in crude oil price(market related)
* or litre/100km/kW if car size taken into account
Data on the specific consumption of new cars (test value):
assessed in the annual monitoring of the ACEA/JAMA/KAMA agreement for
all EU-15 countries since 1995 (available in ODYSSEE)
not available for new EU members.
for a few countries, available as long time series (e.g. since 1980 for
France, Denmark, Austria, and since 1990 for Italy, UK and Germany) with a
break due to a change in the way to measure the test consumption
Data by car size available for a few countries from national sources (e.g.
France, Ireland); from international sources, only available for ACEA, which
represent a decreasing market share of cars sales in the EU
Gasoline and diesel price:
from Enerdata database*
Number of new cars registered
from ODYSSEE
Source of data and information on new cars(1/2)
*Source IEA, Eurostat
Annual distance driven by new cars:
not available
only available for the average for all cars underestimation ofenergy savings as new cars travel more than the stock average
Technical coefficient accounting for the difference between the testvalue and the actual value for the specific consumption:
only estimate will be calibrated from actual data available (e.g. estimatedin a range of 15-20% from Secodip survey in France) and/ or by comparisonbetween the simulated gasoline or diesel consumption of cars and the actualgasoline or diesel consumption of cars.
Energy policy measures:MURE database
IEA energy efficiency database
WEC data base on energy efficiency policies
Source of data and information on new cars(2/2)
Classification of countries
Two group of countries:
“Countries with national measures: e.g. Germany, UK, Finland, Sweden,
Austria, Denmark, France
Tax on motor fuels (increased or new) : Germany, UK, Finland, Sweden
Tax on car purchase linked to energy efficiency and/or CO2 emissions*:
Austria, Denmark, UK
Annual registration tax linked to energy efficiency and/or CO2 emissions*:
Germany, Denmark, UK, Sweden
“Countries without national measures trends mainly influenced by
ACEA/JAMA/KAMA agreement, market price and autonomous trend
* France: exist measures since 2006 on commercial cars only
Countries without national measures: case ofFrance: diesel cars
5,0
5,5
6,0
6,5
7,0
7,5
8,0
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
l/1
00
km
Real data After 1995 trend Before 1995 trend
EU Average trend Average slower trend
What autonomous trends to be considered? Trend before 1995 (ie before the ACEA/JAMA/KAMA agreement) (“before 1995 trend”)
Trend since 1995 (“after 1995 trend”) reference used in the following case studies
EU average trend
Trend of the average of the 3 countries with the lowest autonomous trend (“average slower trend”)
Specific consumption of new diesel cars
Countries without national measures: case ofFrance: gasoline cars
6
6,5
7
7,5
8
1995 1996 1997 1998 1999 2000 2001 2002 2003
l/1
00
km
Real data Trend + prices Trend only
Prices effect not significant
Trend: national since 1995 (regression 1995-2003: Ln (IC) = -0.010 X ln (t) + 2.05 + t
Specific consumption of new gasoline cars
Trend: national trend since 1995
Countries with national measures: case ofGermany Diesel cars
5,0
5,5
6,0
6,5
7,0
7,5
8,0
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
l/1
00
km
Real data Trend + prices Trend only
Price effect not significant
Trend: national since 1995; regression 1995-2004: Ln (IC) = -0.012 X ln (t) + 1.99 + t
Specific consumption of new diesel cars
Countries with national measures: case ofGermany: gasoline cars
6
6,5
7
7,5
8
8,5
9
1995 1996 1997 1998 1999 2000 2001 2002 2003
l/1
00
km
Real data Trend + prices Trend only
Prices effect not significant
Trend: national since 1995; rRegression 1995-2003: Ln (IC) = -0.008 X ln (t) + 2.13 + t
Specific consumption of new gasoline cars
Countries with data on engine size: case ofdiesel cars in France
4
4.5
5
5.5
6
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
l/1
00
k
m/1
00
0c
m3
Real data Trend only
Specific consumption of new diesel cars per cm3
The impact of change in average car size can be done if data are available engine
size in cm3 seems to be the most relevant indicator Trend of –3,4% (close to technical trend) instead of –2.3% (technical and non technical trend)
Better estimate of energy savings ( underestimated if no correction of size)
Engine size data available ONLY for ACEA members
Energy savings of new cars: preliminaryconclusions
Data limitations: no data for most EU New Member Countries(data not covered yet by the EU monitoring)
Definition of trends or baseline to be decided
Role of energy price negligible so far, which is not surprising
Need of define coefficient to account for difference betweentest values and actual values
Similar approach seems applicable for other case study withenergy consumption indicators, although this case study may besimpler as other end-uses