Low Carbon Economy, 2010, 1, 8-17
doi:10.4236/lce.2010.11002 Published Online September 2010 (http://www.SciRP.org/journal/lce)
7
Behavioral and Technological Changes Regarding
Lighting Consumptions: A MARKAL Case Study
Emmanuel Fragnière1,2 , Roman Kanala3, Denis Lavigne4, Francesco Moresino2, Gustave Nguene2
1University of Bath School of Management, Bath, UK; 2Haute Ecole de Gestion (HEG), Carouge, Switzerland; 3University of Ge-
neva, Institute of Environmental Sciences, Carouge, Switzerland; 4Collège Militaire Royal de Saint-Jean, Succursale Forces, King-
ston, Ontario, Canada.
Received August 15th, 2010; revised September 10th, 2010; accepted September 15th, 2010.
ABSTRACT
The present study aims at assessing the joint impact of awareness campaigns and technology choice, on end-use energy
consumption behaviour. Actions to achieve energy savings through the use of more energy efficient end-use technology
are included. A new MARKAL framework, the Socio-MARKAL, was recently proposed by the authors. As opposed to the
traditional MARKAL framework based on technical and economic considerations, the Socio-MARKAL concept inte-
grates technological, economic and behavioural contributions to the environment. This study takes into consideration,
technological improvements on the demand side by consumers as well as behavioural changes minimizing carbon di-
oxide emissions and encouraging rational use of energy. The study presented in this paper, “Lighting Consumptions
Habits in Geneva Households”, was conducted from September to December 2009. Based on the statistical analysis of
this survey, we have determined coefficients to feed the database of the Socio-MARKAL model (an IEA/ANSWER data-
base is available for testing purposes).
Keywords: Demand Side Management, Energy and Environmental Planning, MARKAL, Sociological Surveys,
Sustainable Development
1. Introduction
The original MARKAL model is a multi-period linear
programming formulation of a reference energy system
(RES). The constraints of the model describe all energy
flows, production of electricity and centralized heat, in-
dustrial processes, consumption by end-use technologies
and lastly energy services. The objective function in the
linear programming model is the discounted sum, over
the time horizon considered (usually between 30 and 45
years), of investment, operating and maintenance costs of
all technologies, plus the cost of energy imports. The
model also accounts for emissions of atmospheric pol-
lutants (NOx, SO2 and CO2) and is currently used in many
countries and regions around the world for the assess-
ment of energy pollution abatement policies. As the ex-
isting literature shows [1-4], the MARKAL family of
models and similar energy optimization models to a cer-
tain extent are appropriate to answer questions such as:
how do technologies and policies affect the environ-
mental impacts of energy use (i.e., GHG and emissions
of other pollutants)? How do actions on the demand-side
affect the supply-side and vice versa? How to model the
dynamics of technology (e.g., the switch from one tech-
nology to another)?
However, among the existing MARKAL models, the
social/sociological aspect of energy use on the de-
mand-side is not taken into account. In particular, none
of the above mentioned models takes into account the
contribution of end use consumers’ behavioural change
as a reliable resource for energy efficiency, energy sav-
ings, and emissions reductions. Saving energy requires a
change in the behavior of consumers, either to improve
the energy efficiency of their technology (e.g., technol-
ogy switch), or to improve the way they use energy with
non-efficient technologies (e.g., better use of a less effi-
cient technology). As such, information can be an im-
portant driver, as it can positively influence people’s be-
havior towards energy and the environment [5]. Do we
know enough about factors that can affect the behavior of
energy consumers? Is it reasonable to integrate behav-
ioral and techno-economic parameters? How behavioral
parameters and data can be obtained?

Behavioral and Technological Changes Regarding Lighting Consumptions: A MARKAL Case Study9
Such considerations are accounted for in the Socio-
MARKAL concept developed at HEG [6]. Behavioural
contributions are modeled through virtual technologies
built from sociological surveys, in order to capture the
perception of the population in terms of attitudes and
behaviors regarding energy consumption. These socio-
logical and intangible technologies are therefore com-
bined with traditional and tangible technologies. As a
result of this combination, it will be possible to model the
actual behavior of consumers as well as economically
rational technology choices. These days, environmental/
behavioral campaigns are becoming increasingly sophis-
ticated, going far beyond standard information-only pro-
grams. Consequently, it is essential to define a clear and
systematic protocol for socio-technological evaluations
based on the Socio-MARKAL concept.
2. Protocol for the Socio-MARKAL
The Socio-MARKAL concept is based upon the intro-
duction of a virtual technology built from sociological
surveys. The purpose of such a concept is to capture con-
sumers’ perceptions towards their energy consumption
trends, with an emphasis on their attitudes and behav-
iours. To this end, the virtual, i.e., “sociological” tech-
nologies are associated with tangible technologies, al-
lowing planners or analysts to model, analyse and assess
the actual behaviour of consumers as well as technology
choices which are economically rational. A detailed pre-
sentation of the Socio-MARKAL concept can be found
in Fragnière et al. [6].
Below, we present a method for collecting social data
in the context of the Socio-MARKAL project.
1) Hypothesis generation: qualitative research to iden-
tify potentials of behavioral change regarding energy
consumption, handled through empirical methods (semi
structured interviews, observations, social experiments).
The energy-saving benefits (without a reduction in per-
formance) as well as the essential character of behavioral
change must be clearly explained. If the interviewees or
respondents express interest in the campaigns, the aware-
ness program must be designed so as to remove all the
barriers – e.g. lack of information and motivation, cost of
changing the technology, as well as its installation.
2) Hypothesis testing: survey research to test and mea-
sure hypotheses generated during the first phase (ques-
tionnaire, rank and sample statistical analyses).
3) Behavioral change scenario process: construction of
long-term scenarios including behavioral change, in par-
ticular expert-built scenarios from the collected data.
4) Design: transformation of the Socio-MARKAL (ab-
breviated SOMARKAL in Figure 1) data and scenarios
to feed the MARKAL data base.
The survey questionnaire is elaborated with the ex-
press purpose of assessing the potential contribution of
behavioral change in end-use energy consumption pat-
tern, and thereby in climate change mitigation. The me-
thodology is structured as presented in Figure 1.
3. The Survey Research
The questionnaire can be constructed based on the con-
sideration that climate change is due to greater energy
use by humans. In line with this, two approaches appar-
ently inclusive can be used [7]: 1) improving the effi-
ciency of end-uses of energy and 2) not using or con-
serving energy. For Rudin, the proponents of the first
approach seem to denigrate the overall notion of suffi-
cient and limited energy use.
The main focus of this paper is on lighting technolo-
gies. Our survey is based upon the analysis of a sample
(probabilistic) made up of 393 valid questionnaires ad-
dressed to the populations living along Lake Geneva.
This research has been conducted by the laboratory of
market research (LEM, Laboratoire d’Études de Marché)
of Geneva Haute École de Gestion, whose objectives are
to develop locally-based survey research in Economics
and Business Administration and to train students to
marketing survey techniques. The study presented in this
paper, “Lighting Consumptions Habits in Geneva House-
holds”, was conducted from September to December
2009. In the questionnaire, we also included questions
based on Contingent Valuation Methods (hypothetical
scenarios) to assess the individuals’ possible behaviors in
specific contexts. Relationships between classes as well
as relationships between variables have been investigated
and analyzed in depth. Then, research hypotheses were
verified on the basis of non-parametric statistical tests
before being introduced in the Socio-MARKAL database.
This survey questionnaire has been elaborated with the
express purpose of assessing the potential contribution of
behavioral change in end-use energy consumption pat-
tern, and thereby in climate change mitigation.
The survey questionnaire was distributed to a random
sample of people, equally distributed between men (52%)
and women (48%). These individuals are aged between
Figure 1. Outline of the methodology.

Behavioral and Technological Changes Regarding Lighting Consumptions: A MARKAL Case Study
10
15 and 75 years old, with a mean of about 38 years old.
They have the following occupations: employees (39%),
students (31%), high-level executives (12%), independ-
ents and retirees, as well as housewives and unemployed
(4% respectively). As for the ownership, 31% of the re-
spondents own their house while 66% are tenants. Addi-
tionally, 68% live in apartments, while 32% live in a
house. In this short paper, we just focus on some descrip-
tive statistics. In particular, we present the results related
to the question “What proportion of low consumption
bulbs do you have at home”.
Table 1 below indicates that, firstly, the great majority
of the respondents (58.84%) have a proportion of low
consumption light bulbs ranging between 25% and 75%.
Secondly, the lowest response rates are on the extremes,
namely those who have only low consumption bulbs
(8.04%), and those who have none (14.83%).
The first observation shows that 58.84% of the popu-
lation can potentially switch technology or change their
consumption behavior through better use of incandescent
bulbs. The corresponding maximum of incandescent light
ranges from 25% to 75% of the bulbs.
However, 20% of the respondents responded “I don’t
know”, which seems to indicate that they have no infor-
mation about the technology (incandescent or not) they
have been using for lighting.
Among other results provided, it is interesting to men-
tion the following elements: 82.3% of the respondents
indicates that they have been taught by their parents to
switch off the lights when leaving a room; the respon-
dents had also to give the most important peculiarities (2
choices) associated to a light bulb and it appears that
“lighting intensity” is the most important one (26.8%),
followed by “consumption” (22.0%), “lifetime” (20.2%),
“purchase price” (10.6%). Summing up, economic pa-
rameters (lifetime, price and consumption) represent
52.8%, while parameters related to comfort (light inten-
sity, color and ambiance) represent 40.3%. Ecological
parameters (origin, manufacturing, and disposal) are of
least concerns to the respondents (6.9%). The other ques-
tions have been designed in order to provide social data
to these so-called virtual technologies in MARKAL.
The questionnaire includes questions on Contingent
Valuation Methods (hypothetical scenarios) to assess
individuals’ attitudes and behaviors towards more effi-
cient end-use technologies (i.e., incandescent light bulbs
vs. low consumption bulbs) (see for example [8-11]).
Based on this recent survey, we have determined an
environmental and energy planning scenario that simul-
taneously takes into account technological as well as so-
ciological aspects.
4. Modeling the Bulb Demand Devices in
Socio-MARKAL
The representative parameters of the Socio-MARKAL
have been designed so as to keep the traditional MAR-
KAL formalism. This will ease the use of MARKAL plat-
forms such as ANSWER. ANSWER [12] is the data base
management system of the MARKAL-TIMES models
generators developed by IEA-ETSAP, the International
Energy Agency Implementing Agreement for a Program
of Energy Technology Systems Analysis.
As mentioned on the previous sections, energy con-
servation may require the introduction/adoption of meas-
ures aimed at promoting rational use of energy. These
measures include: 1) a better use and management of ex-
isting equipments or technologies and/or 2) technology
switch. In this study, we assume that people/consumers
who are willing to adopt one or more of these measures
are driven by the desire to change their energy consump-
tion behaviour. This willingness could be explained by
many factors, such as their sensitivity to marketing/
awareness campaigns, training, their education, the qual-
ity of information they have been receiving, as opposed
to the assumption of perfect economic rationality gener-
ally used in the traditional MARKAL family of models.
Behavioural change in Socio-MARKAL requires in-
troducing virtual technologies, whose purpose is to trigger
Table 1. Summary statistics showing the proportion of low consumption vs. incandescent light bulbs.
Proportion of low consump-
tion bulbs [%]
N, number of respon-
dents Percentage [%] Percentage without “does not know” [%]
Cumulative percentage
[%]
0 43 14.83 17.13 17.13
25 76 24.44 30.28 47.41
50 55 17.68 21.91 69.32
75 52 16.72 20.72 90.04
100 25 08.04 9.96 100.00
Does not know 60 19.29 - -
Total 311 100.00 - -
Note:
Number of questionnaires missing or not filled by the respondents. N = 82; Percentage: 20.87% in a total of 393 questionnaires

Behavioral and Technological Changes Regarding Lighting Consumptions: A MARKAL Case Study 11
the survey research are transformed to feed the bulb de-
mand devices section of the Socio-MARKAL model. In
this short paper, we will concentrate on a single case.
Note that due to the qualitative nature of the data, we will
be using a narrative scheme to make the case.
behavioural change among energy consumers. For the
bulb demand devices section of the Socio-MARKAL
model, we ended up with the following representation,
using structures defined by MARKAL, as outlined in
Figure 2. We have two sets of data for the residual capacities for
incandescent bulbs RLD1 and low consumption bulbs
RLD4. One data set comes from MARKAL data points
that are obtained from statistics of observed actual con-
texts. The second data set about the number of bulbs and
their split comes from the survey and it is not exact be-
cause it corresponds to elements of perceptions. The
qualitative scale is defined over: zero, 1/4, 1/2, 3/4, all
bulbs, and “do not know”.
We’ve got four demand devices. Parameters RLD1 and
RLD4 represent the real and tangible lighting technolo-
gies, receiving electric power as input, and generating
residential lighting.
Parameters RLD2 and RLD3 represent the virtual tech-
nologies.
As opposed to the real technologies, virtual technolo-
gies receive inputs that are intangible, leading to energy
savings or technology switch. These devices are presented
below: Residual capacity of behaviour change in favour of
electricity savings RLD2 and for technology switch
RLD3 is zero at the beginning of the optimisation. Then,
we have the evolution of the environment that corre-
sponds to Socio-MARKAL scenarios. For instance, there
are people who will never change their behaviour and
will only use incandescent bulbs RLD1 as long as they
can buy them. Then, another case concerns people who
are not aware of the advantages of low consumption
bulbs and consequently might change their behaviour
spontaneously as they get or receive more information.
- RLD1, “existing incandescent bulbs”
- RLD4, “existing low consumption bulbs
- RLD2, “moderate use of incandescent bulbs”, and
- RLD3, “switch to low consumption bulbs”
- MRKRP2 and MRKRP3 are marketing/awareness
campaigns which have the effect of changing the be-
haviour of energy consumers. MRKRP2 and MRKRP3
are respectively supposed to trigger the “Moderate use of
incandescent light bulbs”, and the behaviour towards low
consumption bulbs, i.e. “Technology Switch towards low
consumption bulbs”. Finally, there are people who would not change their
behaviour without an information campaign; however
with an exposure to information, they will start saving
energy and/or switch technology. In order to measure this
part of behavioural change, we have asked the following
questions in the survey:
5. Transforming Sociological Data into
MARKAL Parameters: An Example
In this section, we show how the statistics obtained from
Figure 2. Structure of the reference energy system (RES).

Behavioral and Technological Changes Regarding Lighting Consumptions: A MARKAL Case Study 12
- Question 10: “Did you know that low consumption
bulbs may consume up to 5 times less energy than the
incandescent ones?” (possible answers: Yes, No).
- Question 11: “Did you know that low consumption
bulbs have a lifetime up to 10 times superior to the
incandescent bulbs?” (possible answers: Yes, No).
- Question 12: “If you were better informed about the
economic advantages of the low consumption bulbs,
would you be ready to abandon completely the
incandescent bulbs?” (possible answers: Yes, No).
Thanks to these questions, we can identify the part of
consumers who were not informed about the energy
consumption and lifetime of the technology. But once
they got this complement of information, they claim to
be ready to undertake technology switch. People who
know about the advantages of low consumption bulbs
and despite that, are not willing to change, influence the
decrease of RLD1 to a steady incompressible level. For
people who did not know about the advantages of low
consumption bulbs and are then informed, there will be
some of them switching to low consumption bulbs. We
have then added one more question formulated as hypo-
thetical scenarios:
- Question 13: “Did you know that if a household
changes half of of incandescent bulbs for low con-
sumption ones, about 200 CHF per year can be saved?”
(possible answers Yes, No).
- Question 14: “Based on this information, would you
change at least half of your bulbs?” (possible answers:
Yes, No, I did it already).
This question enables us to identify the part of people
who are well informed but will not change, those who
were not informed but did the change and finally those
who would do the technology switch thanks to the in-
formation.
In order to assess the drivers of energy savings, we
have asked about the reasons why people turn the lights
on when they enter into a room. If this is due to their
poor eyesight or irrational fear of obscurity, they will
probably not change their behaviour. However people
who say it is just a habit, or if they do so for comfort,
aesthetics, or do not know, they could possibly make the
effort to change their behaviour. Likewise, people who
say they leave light on when watching TV, could switch
it off. But those who do have already done this or who
leave only a small spot to reduce contrast and eyestrain,
cannot make further savings. This is typically the kind of
hypotheses we need to set up in order to get a proxy of
the parameters that will be entered into the model.
For instance, in Table 1, related to Question 7 “What
proportion of low consumption bulbs do you have at
home”, we had roughly 20% who wouldn’t know about
the number of low consumption bulbs they originally had.
Based on cross table analyses involving different ques-
tions of the questionnaire, and referring to the above hy-
potheses, we propose a new presentation for the propor-
tion of low consumption bulbs in Table 2 above. This
table is now usable to determine the RESID of RLD 1.
6. Preparing the Data for Socio-MARKAL
This section aims at determining the parameters for all
the demand technologies (i.e., RLD1, RLD2, RLD3,
RLD4) presented in Figure 2. More descriptive details
about these technologies can be found in the appendix, in
Table 5. We start with RLD3, which corresponds to
“Switch to low consumption bulbs”.
6.1. The Case of RLD3
Our evaluation is based on the cross-analysis of ques-
tions 13 and 14. The respondents (3.9% of the sample
and 15.3% of people who answered “Yes” to question 13)
were already informed about the economic advantage of
choosing low consumption bulbs and are not willing to
change. These respondents represent an incompressible
ratio that enters into the efficiency coefficient of the
MARKAL model. To this figure, it is necessary to add
14.4% of the sample (i.e., 19.2% of people who an-
swered “No” to question 14). These latter persons did not
have the information before and were informed during
the survey. Still, they indicate that they won’t change
their behavior.
On the other hand, 9% of the sample and 35.7% of the
people answering “Yes” to question 13 were already in-
formed about the economic advantages of switching to
low consumption bulbs. Moreover, they indicate that
they are willing to make a technology switch soon, but
did not do it yet. We believe that for these respondents,
this new exposition to information represents a reminder.
Likewise, 44.2% of the sample (i.e., 59.1% of all people
for whom the information is new) claim that they are
willing to undertake the technology switch in order to
replace half of their incandescent light bulbs by low-
consumption bulbs.
Finally, people who have announced that they already
made the change before the survey, are deducted from
Table 2. Descriptive statistics showing the proportion of low
consumption light bulbs.
Proportion of low
consumption bulbs [%]
N, number of
respondents Percentage [%]
0 53 17.04
25 96 30.87
50 65 20.90
75 72 23.15
100 25 08.04
Total 311 100.00

Behavioral and Technological Changes Regarding Lighting Consumptions: A MARKAL Case Study
13
the final ratio that is inputted in MARKAL (53.2% of the
people who answered “Yes” to question 14).
6.2. The Case of RLD2
Here we explore and analyze the answers in question 15,
“Your electricity consumption would likely change for
the following reasons” (respondents can make up to 2
choices among six options). The results (normalised) are
presented on the table below.
In order to determine the parameters for RLD2, we
have considered three different aspects. Firstly, we have
decided to evaluate the direct impact of an advertising
campaign (i.e., “An information campaign on the media
or advertising”) as our main explanatory parameter for
RLD2. Secondly, the indirect effect can be represented
by both “The opinion of a relative” and “A request from
our children”. Thirdly, economic criteria include the in-
come modification (i.e., “A change in your income”),
and “Electricity price increases”. The former and the
latter are excluded because these parameters (effects) are
in fact characterized by their economic rationality, which
is compatible with the standard MARKAL formulation.
Our evaluation shows that 37.5% of all respondents
(24.7% of the normalized total, as shown in Table 3 above)
declare that an awareness campaign can influence their
electricity consumption. Based on this information, we
can set the efficiency of the awareness campaign to
24.7%. Setting the efficiency as composed of both the
direct and indirect effects, we get 27.5+37.5+11.1 = 76%
of all replies (i.e . , 50.2% of the normalized total). How-
ever this figure represents the declared likeliness of the
respondents influenced directly by an information cam-
paign or indirectly through a third-party person (i.e.,
relatives and parents).
6.3. Determining Investment Costs for MRKP2
and MRKP3
The following table corresponds to the results provided
for question 16, “What kind of information means would
likely change your behaviour”. These results, presented
in Table 4 below, are associated with percentages that
have been normalised. In this table, we have included a
new column titled “Cost per individual”. These estimates
represent calculations based on the results of question 16
and hypotheses related to marketing costs (obtained
through a discussion with a marketing expert operating in
the Geneva region). In the case of articles from newspa-
pers, we consider the cost to be zero because it is an in-
dependent editorial initiative.
6.4. Inputting the DM Tables of MARKAL
We have implemented a MARKAL model using AN-
SWER, the IEA platform. Consequently this kind of
simulation will be directly available to ANSWER users
and we hope that it will enable them to develop their own
Socio MARKAL scenarios (the IEA/ANSWER mdb file
of the case study presented in this paper is available by
contacting the authors).
The following snapshot, exhibited in Figure 3 below,
presents an example of this kind of Socio-MARKAL
Table 3. Distribution of responses to question 15, “Your electricity consumption would likely change for the following rea-
sons” (respondents can choose among up to 2 options).
Proportion of respondents (N = 311)
Option 1. The opinion of a relative, a friend or a neighboor 18.20
Option 2. A request from your children 7.27
Option 3. An information campaign on the media or advertising 24.74
Option 4. A change in your income 12.53
Option 5. Significant electricity price increase 30.71
Option 6. Nothing would change my behaviour 6.62
Total 100.00
Aggregates
Information (options 1 to 3): 50.13%
Economic conditions (options 4 & 5): 43.25%
Table 4. Distribution of responses to question 16, “What kind of information would likely help changing your behavior”.
Means Percentage [%]
Cost [CHF] per
individual
A Web page filled with useful information 10.50 5.00
A Web page with an energy savings calculator 17.60 5.83
Articles from newspapers/TV reports 19.40 0.00
Leaflets from utilities or green organisations 14.50 0.03
Doorstep awareness campaigning 2.40 20.00
Selling points with information (e.g. posters) 12.30 20.00
Advertisement on TV and radio 14.20 1.00
Advertisement campaigns on public transports 9.10 0.25
Total 100

Behavioral and Technological Changes Regarding Lighting Consumptions: A MARKAL Case Study
14
Figure 3. Example of Socio-MARKAL scenarisation developed in IEA’s ANSWER.
scenarisation in IEA’s ANSWER. We are aware that this
kind of scenarisation is associated with uncertainties and
that this kind of development is still in its infancy. How-
ever, it offers to policy makers a tool to test environmen-
tal policies that involve social change.
Recently, new regulations imposing low consumption
bulbs have been enforced in the EU as well as in Switzer-
land. Consequently, this kind of scenario can be useful to
understand alternate evolutions if this kind of regulations
were not in place.
We have seen that it is particularly difficult to devise
hypotheses regarding investment costs. We would like al-
so to remind that it is possible to produce shadow prices
from MARKAL, because it is based on convex optimiza-
tion. So, it is always possible to determine what should
be “the rational” cost of awareness campaigns.

Behavioral and Technological Changes Regarding Lighting Consumptions: A MARKAL Case Study15
7. Illustration
In this section, we will assess the Socio-MARKAL mo-
del with an illustration based on three scenarios. In this
illustration, we consider an evaluation over a span of 20
years (i.e., from 2005 to 2025) spread over 4 periods of 5
years.
7.1. Assumptions
A number of assumptions have been introduced, specifi-
cally regarding both the demand investments for residen-
tial lighting technologies.
The overall demand for light bulbs is expected to grow
by about 50% over the evaluation period, i.e., from 1442
hundreds units (144'200 bulbs) in 2005 to 2500 hundreds
units (250'000 bulbs) in 2025. Residual capacity is split
between RLD1 and RLD4, respectively for 80% and
20%.
Investment costs for RLD1 are expected to rise by 40%
over the evaluation period, i.e., from 1 CHF/bulb in 2005
to 1.40 CHF/bulb in 2025. However, for low consump-
tion bulbs (RLD4) whose costs are set to 10 CHF/bulb in
2005, we assume decreasing costs over the time periods
of respectively, 7 CHF/bulb in 2010, 6 CHF/bulb in 2015,
5 CHF/bulb in 2020, and 4 CHF/bulb in 2025.
The lifetime of new light bulbs is set to 1 year for in-
candescent bulbs (RLD1) and 2 years for low consump-
tion bulbs (RLD4).
The energy carrier input, (i.e. MA (ENT) in AN-
SWER), remains constant over the evaluation period. For
RLD1 and RLD4, it is set to 0.0328 TJ/unit and 0.0065
TJ/unit respectively.
7.2. Scenarios and Results
We consider three scenarios.
Scenario 1: the first scenario (see Figure 4) is charac-
terized by investment costs in marketing technologies
which are so high that virtual technologies do not appear
in the optimal solution. Here, we have a case of classical
MARKAL competition between incandescent bulbs
(RLD1) and low consumption bulbs (RLD4). Both tech-
nologies have bounds. In the case of incandescent bulbs,
there is a constraint related to the proportion of people
who claim they are not willing to undertake the technol-
ogy switch, i.e., from incandescent to low consumption
bulbs. On the other hand, for low consumption bulbs, we
have put a lower bound equal to the installed capacity on
the first period, assuming that in the following periods,
their penetration should never go below that value.
Scenario 2: the second scenario presented in Figure 5,
is an example of a case when the cost of information cam-
paigns in favour of technology switch is getting lower so
0
500
1000
1500
2000
2500
2005 2010 2015 2020 2025
RLD4
RLD3
RLD2
RLD1
Figure 4. Outline of scenario 1, “Marketing technologies too
costly”.
0
500
1000
1500
2000
2500
2005 2010 2015 20202025
RLD4
RLD3
RLD2
RLD1
Figure 5. Scenario 2, “Marketing for technology switch
becomes competitive”.
that the campaigns are worth to be financed. In this case,
and for two periods, the number of incandescent bulbs
decreases, replaced by low consumption bulbs, which
appear as the result of the campaign in favour of low con-
sumption bulbs.
Scenario 3: the third scenario (Figure 6) shows that if
the cost of marketing technologies is low, then the en-
ergy savings and technology switch may dominate the
spontaneous purchase of real bulbs.
8. Conclusions
This study implements the Socio-MARKAL framework
that would take into account consumers’ technological
0
500
1000
1500
2000
2500
2005 2010 2015 2020 2025
RLD4
RLD3
RLD2
RLD1
Figure 6. Scenario 3, “Both marketing technologies appear-
ing in the optimal solution”.

Behavioral and Technological Changes Regarding Lighting Consumptions: A MARKAL Case Study
16
improvements and behavioral changes minimizing car-
bon dioxide emissions and encouraging rational use of
energy. In this paper, we firstly present and discuss the
results of a survey research related to attitudes and be-
haviors towards lighting consumption.
We show how to transform the sociological data into
parameters for the MARKAL model. Secondly, we prove
with this paper that, sociological data can be integrated in
a model of technological choices such as MARKAL. The
IEA’s platform, ANSWER, has been used for that. Thirdly,
our study shows that it is possible to develop environ-
mental and energy planning scenarios that simultane-
ously take into account technological as well as socio-
logical aspects. This study also shows that we have been
able to move from an idea (i.e., integration of behavioral
aspects of energy consumption into a model of techno-
logical choices) to the concept proven (i.e., the Socio-
MARKAL).
We are currently working to extend the concept to the
transportation sector. A new sociological survey about
attitudes and behaviors regarding passengers is currently
conducted in order to feed the Socio-MARKAL model
with additional relevant data.
The conclusions drawn from one of our previous stud-
ies [10] show that the sociological/behavioural approach,
i.e., data collection through surveys and sociological ex-
periments are powerful tools that can help people under-
stand their (personal) energy use and for motivating their
actions to reduce carbon emissions. This means that
awareness campaigns can stimulate behaviour change to
conserve energy. Consequently, both technological and
behavioural contributions can be integrated into a single
strategy. In turn, this is enough to justify an extension of
the current MARKAL family of models, with the inte-
gration of data collected through surveys and/or aware-
ness campaigns.
9. Acknowledgements
We would like to thank Dr. Christian Decurnex, Director
of the Municipal Utility of Nyon (Switzerland), for his
strong commitment to the application of the Socio-
MARKAL model in a real context. Finally, we would
like to express our gratitude to Professor Jean Tuberosa,
Director of the Market Studies Laboratory at the Geneva
School of Business Administration.
REFERENCES
[1] L. D. Hamilton, G. A. Goldstein, J. Lee, A. S. Manne, W.
Marcuse and S. C. Morris, “MARKAL-MACRO: An Over-
view,” BNL-48377, Brookhaven National Laboratory, Up-
ton, 1992.
[2] E. Fragniere and A. Haurie, “A Stochastic Programming
Model for Energy/Environment Choices under Uncer-
tainty,” International Journal of Environment and Pollu-
tion, Vol. 6, No. 4-6, 1996, pp. 587-603.
[3] E. Fragnière, A. Haurie and R. Kanala, “A GIS-Based Re-
gional Energy-Environment Policy Mode,” International
Journal of Global Energy, Vol. 12, No. 1-6, 1999, pp. 159-
167.
[4] H. Turton and L. Barreto, “Long-Term Security of Energy
Supply and Climate Change,” Energy Policy, Vol. 34, No.
15, 2006, pp. 2232-2250.
[5] H. Hondo and K. Baba, “Socio-Psychological Impacts of
the Introduction of Energy Technologies: Change in En-
vironmental Behaviour of Households with Photovoltaic
Systems,” Applied Energy, Vol. 87, No. 1, 2010, pp. 229-
235.
[6] E. Fragniere, R. Kanala, D. Lavigne, F. Moresino, A. De
Sousa, C. Cubizolle, C. Decurnex and G. Nguene, “Socio-
Markal (Somarkal): First Modeling Attempts in the Nyon
Residential and Commercial Sectors Taking into Account
Behavioral Uncertainties,” 2009. http://ssrn.com/abstract=
1522143
[7] A. Rudin, “Why We Should Change Our Message and
Goa l from “Use Energy Efficiently” to “Use Less En-
ergy,” Proceedings of the ACEEE 2000 Summer Study on
Energy Efficiency in Buildings, Washington D.C., 2000,
pp. 392-340.
[8] R. Hoevenagel, “An Assessment of the Contingent Valua-
tion Methods,” In: R. Pethig, Ed., Valuing the Environ-
ment: Methodological and Measurement Issues, Kluwer
Academic Publishers, Dordrecht, 1994, pp. 195-227.
[9] G. Catenazzo and E. Fragniere, “La gestion des services,”
Economica, Paris, 2008.
[10] S. Weber, A. Baranzini and E. Fragnière, “Consumers’
Choices among Alternative Electricity Programs in Ge-
neva,” International Journal of Global Energy, Vol. 31,
No. 3-4, 2009, pp. 295-309.
[11] G. Catenazzo, E. Fragniere, B. Ribordy and J. Tuberosa,
“Is the 2008 Financial Turmoil Increasing the Risk of a
Bank Run?” Journal of Modern Accounting and Auditing,
Vol. 6, No. 1, 2010, pp. 29-45.
[12] R. Loulou, G. Goldstein and K. Noble, “MARKAL Users’
Guide: Documentation for the MARKAL Family of Mod-
els,” Technical Report, 2004. http://www.etsap.org

Behavioral and Technological Changes Regarding Lighting Consumptions: A MARKAL Case Study17
Appendix
Table 5. Technologies and their description.
Parameters Value Units Type
ELC Energy imports TJ Resource
CELC Energy carrier TJ Resource
MRKP2 Awareness campaign “Moderate use of incandescent light bulbs” CHF/capacity
unit
Process tech-
nology
MRKP3 Awareness campaign “Technology switch towards low consumption bulbs”CHF/capacity
unit
Process tech-
nology
MRKPRC2 Energy carrier MRKP2 TJ Process
MRKPRC3 Energy carrier MRKP3 TJ Process
RLD1 Existing Incandescent Bulbs Hundreds of light
bulbs
Demand
device
RLD2 Moderate Use of Existing Incandescent Bulbs Hundreds of light
bulbs
Virtual de-
mand device
RLD3 Technology Switch toward Low Consumption Bulbs Hundreds of light
bulbs
Virtual de-
mand device
RLD4 Existing Low Consumption Bulbs (a mix of new technologies) Hundreds of light
bulbs
Demand
device
RLD End use residential lighting Hundreds of light
bulbs Demand