The Impacts of Governmental Policies on the Investment
Decision for Renewable Energies in the Swiss Electricity Market
Merla Kubli
System Dynamics Group, University of Bergen
Merla Kubli
Eulerstrasse 15
4051 Basel, Switzerland
+41 77 421 1709
merla@merla.net
Abstract
Switzerland faces two major challenges in the electricity sector. The existing nuclear
power plants will be phased out and at the same time new renewable electricity sources
should increase their share in production. These shifts need to be managed while
ensuring a secure electricity provision. The investment decision for the specific
technologies is a central leverage point in the system. Currently a feed-in remuneration
tariff policy with a fixed tariff is implemented to support new renewable energy
hk 1,
Pp
A System Dynamics simulation model is built to improve the understanding of central
tech ies in their de
developments in the system and the interplay of different electricity technologies in the
electricity production. The model is used to simulate likely developments of the Swiss
electricity power plant park and test the effectiveness of feed-in remuneration policies.
Results are gained on the long-term dynamics of capacity building of electricity
technologies, depending on different public policies. This paper makes a practical
contribution to the management of the energy transition by shedding a more dynamic
light on the capacity expansion in relation to different forms of feed-in tariff policies.
Keywords: Energy, electricity, System Dynamics, Switzerland, feed-in tariff, nuclear
phase out, long-term simulation.
1. Introduction
Switzerland has two self-made challenges in the electricity provision sector to be
solved mutually in the years to come. The Swiss Federal Council and the parliament
decided on the withdrawal from nuclear energy in 2011 (Swiss Federal Council, 2011),
due to the disastrous accident in Fukushima and lacking security of the nuclear
technology in general. The stepwise phase out from nuclear power causes a gap in the
future coverage of the electricity consumption in Switzerland (Prognos, 2007; Prognos,
2012). This gap needs to be filled with locally produced electricity to maintain political
sovereignty (Swiss energy enactment Art. 6; Swiss Federal Council, 2011).
Additionally, a commitment to a more sustainable electricity production was made
(Swiss Federal council, 2011; Swiss energy enactment Art. 3b). Especially the
expansion of hydropower and new renewables energies will be supported.
Nevertheless, the Swiss Federal Council does not consider an electricity provision
based on only renewable energies as feasible.
A System Dynamics model is built to improve the understanding on the dynamic
interplay of central factors in the electricity capacity expansion system and simulate
likely future developments. The focus in this framework lies on the investment decision
taken for the different technologies und how this can be steered by governmental
policies. This simulation model contrasts itself from other energy models currently used
in Switzerland, by the endogenous simulation of the investment decision, which is
driven by the internal dynamics of the system.
Central characteristics of the system as well as policy attack points are tested with the
simulation model. The impact and effectiveness of the currently applied model of the
feed-in remuneration policy is tested and compared with other feed-in tariff models
described in Couture and Gagnon (2010).
The simulation results reveal that a transition towards an electricity system based on
only renewable energies is feasible. Insights are gained on the dependency of the
different technologies on market design and regulations. The widely applied feed-in
tariff policies proof to be a good instrument to push the electricity system in its
transition, but they fail to sustain the system in its new state.
This paper is organized as follows. The theoretical background follows the
introduction. In the third section an overview and detailed description on the simulation
model is given. Results are presented in the forth section. The article closes with a
discussion of the results and further research needed in this area.
2. Background
Energy is a catalyst for every economy. It is the most relevant input for an entire
system, for all kinds of production and consumption. Today we are facing a situation
where the commonly used energies such as oil and gas are getting scarcer but new
renewable energies are not yet completely competitive over the traditional energies
(Jacobsson and Johnson, 2000). Environmental effects of the use of fossil fuels make an
early transition necessary (European Commission, 2011; Dangerman, 2012). The
electricity industry has already undergone multiple transitions, from wood to coal to oil
and gas (Naill, 1992; Jacobsson and Johnson, 2000). Now a transition towards new
renewable energies is necessary. So far the new renewable energies are not yet
competitive over traditional energy sources, which creates the special situation where
the government decides to push the transition. This research focuses on the challenges
of a transition in the area of electricity production within the specific case of
Switzerland.
The coverage of demand for electricity by households and industry in Switzerland is
not guaranteed in the mid-term future. Power plants achieve their maximum lifetime,
import contracts expire, but most important the nuclear power plants will be switched
off, when they don’t fulfil the required security standards anymore (Prognos AG 2007,
Prognos AG 2012). The Swiss Federal council decided on the nuclear power phase out
in 2011 after the happenings in Fukushima (Swiss Federal Council, 2011). No
replacement and any major renovations will be made on the existing five nuclear power
plants. The result is a steadily decreasing electricity production. Figure / visualizes this
problem. In this graph the electricity production based on the currently existing
installed capacity, the expected lifetime of these plants and the planned switch off time
for the nuclear power plants is simulated over 40 years. However, in the essence the
match of the supply with the demand for electricity is much more important. In Figure
1 three demand scenarios are included. The demand scenarios are called “business as
usual”, “new energy politics” and “political measures” and are the same as considered
in the Prognos study (2012). The graph clearly highlights, that no matter which scenario
is chosen, a huge gap in the electricity provision results.
Future electricity production and demand scenarios
90'000
£
=
© 80'000
=
3
5 70°00 === wind
3... Sm biomass
> 60'000
3 © phtovoltaic
= 50'000
5 = hydro
Bags
3 0:00 = nuclear
S 30'000 —demand business as usual
3 i
= 20'000 demand new energy policits
a 4G ———~demand political measures
0
2010 2015 2020 2025 2030 2035 2040 2045
Figure 1: Gap in electricity production without new investments
The obvious question is - how to fill this upcoming gap in electricity provision.
Prognos (2007, 2012) discuss in energy strategy 2035 (Prognos 2007a) and energy
strategy 2050 (Prognos 2050) several constellations of technologies how the upcoming
gap in electricity provision could be filled. These investigations are the major decision-
making basis for the Swiss Federal council. Multiple energy models are combined and
analysed with a scenario method. An extensive bottom up calculation for demand is
made. For supply a static model of the power plant park is used. The investment
decision is considered as exogenous but limited by the physical and economic potential
of the technology. All scenarios designed by Prognos (2012) include gas combined
cycle power plants. An electricity provision with only renewable energies is considered
up front as unfeasible.
Supercomputing Systems Ltd. (SCS) provides a different answer how this gap in
electricity provision could be filled. SCS suggests a power plant park constellation with
only renewable energies (SCS 2013). The electricity model they present is a very
detailed representation of the Swiss electricity production of one year. The simulation
starts with a predefined constellation of the power plant park. Parameters are set for
production costs. Different geographical regions for weather conditions are considered
as determinants of the production of renewables technologies. A priority list is
integrated in the model to ensure that the power plants are operating in the interest of
the overarching system. On the basis of this model several power plant park
constellations are derived that can provide the demand for electricity of 60 TWh per
year with only renewable energies. The major challenge is to compensate for the
volatility of the new renewable energies, determined by their stochastic nature of the
electricity production. With their results SCS are challenging the assumption by the
Swiss Federal Council and Prognos (2007; 2012) that combined heat and power units
and also gas combined cycle power plants are necessary to guarantee a secure
electricity production.
A major capacity expansion would be necessary to achieve a completely renewable
electricity provision, no matter which model is considered. Neither the model by
Prognos (2007; 2012) nor the model by SCS (2013) give an answer how and when
these investments will be realized or whether these investments are an economic choice
by investors or forced by the government. The investment decision for future
investments is a very essential aspect for the future development of the form of the
electricity production. Investments have very long-lasting implications on the
electricity provision system due to the long life times of the power plants. There is a
need for a complementary model, which can simulate the development of the power
plant constellation over time depending on the state of the system. Modelling the
investment decision endogenously is essential to gain knowledge on potential future
developments of the system. A model representing the investment decision into the
various technologies necessarily has to be more aggregated than the SCS model. The
level of detail that the SCS model provides is not desired for a long-term model
focussed on the development of the system. But this depth is very relevant when the
feasibility and reliability of the final state derived by a long-term model should be
tested.
This study provides this long-term model that can simulate the investment decision
endogenously and over the time horizon from 2006 until 2050. It can be seen as the
complement for the SCS model as well as a testing environment for various scenarios
or policies to support renewable energy sources.
The provision of electricity in Switzerland is the task of the electric power industry
(Art. 2, chapter 2, Swiss energy law). Local electricity companies are responsible for
providing their area with electricity. The local electricity companies are working
according economic principles but its shareholders are to a major part the local
governments. In 2011 the public hand held 87.9% of the shares of the electric power
companies in Switzerland (Swiss Federal Office of Energy, 2013). The national
government is responsible to ensure favourable conditions for the energy industry. The
government has the option to introduce incentives, to steer the system into a desired
direction (Art. 2, chapter 2, Swiss energy law).
In the current system a subsidiary support policy for renewable energies, a so-called
feed-in remuneration at cost policy, is established. The general aim of this policy is to
increase the competitiveness of renewable electricity sources over the non-renewables
and reduce the investment risk. The European Commission (2008) observed that feed-
in tariffs are the most effective policy in support renewable energies. Nevertheless the
effect on the different technologies varied. Couture and Gagnon (2010) distinguish
between seven different forms of feed-in remuneration tariffs. Switzerland shifted
applies a fixed price model (Couture and Gagnon 2010, Swiss energy enactment). The
fixed price model is a model independent of the current market price for electricity.
This feed-in tariff (FIT) supports specific energy sources with paying a guaranteed
tariff over a defined period of time per kWh electricity that is fed into the grid (Art.3,
paragraph 2, Swiss energy enactment). The costs of the feed-in tariffs paid to the
producers are transferred to the electricity consumer through a grid charge rate
(Interface et al., 2012). The feed-in remuneration in Switzerland is guaranteed for
specific technologies with individual tariffs. Currently wind, photovoltaic, small-scale
hydropower, geothermal power, biomass power, incinerations and combustion of
sludge are profiting of the support.
Interface et al. (2012) analyse the effectiveness of the applied FIT policy in
Switzerland. They conclude that the FIT policy has the potential to increase
investments into new renewables to reach the goals by the Swiss Federal Council.
Nevertheless, a long waiting list resulted and it is observed that 26% of the receivers of
the FIT policy are free riders, investors who would do their investment anyway also
without the FIT policy. An effect on innovation is not expected. Although the FIT
policy evaluation by interface et al (2012) is fairly extended, an analysis of the long-
term effects of the policy on the electricity market is not made nor is the sustainability
of this policy discussed. SwissCleanTech (2013a) reveal with an economic thinking
experiment, based on some general economic models, that the strong support of the
new renewables will have significant impacts on the electricity market. First of all they
expect that during some times of the day the electricity price will fall to zero or even
become negative. Regulatory electricity technologies will struggle to amortize their
investment. Also new renewables struggle in their profitability due to the gap between
the marginal costs of production and their full costs (including the production unrelated
costs) (SwissCleanTech, 2013b). Furthermore, SwissCleanTech (2013a) fear that after
a stop of the FIT policy there will be no reinvestment into the new renewables.
3. Model
This study aims to increase understanding of the investment decision in the electric
power industry and its dynamic impacts on the electricity provision system. A System
Dynamics simulation model is used to gain insights into the dynamics of the system.
With the simulation of different scenarios knowledge is built how investment decisions
affect the constellation of the power plant park and which structure parts feedback to
the investment decision itself. Furthermore, options are tested how the investment
decision can be steered by public policies. This project sheds an aggregated and long-
term view on the electricity capacity expansion system and focuses on the phenomena
arising during the next 40 years. The simulation timeframe until 2050 is chosen in line
with the planning horizon of the Swiss energy strategy 2050 (Swiss Federal Office of
Energy, 2013b).
System Dynamics is chosen as suitable simulation method to simulate the high
complexity of this system. Major delays in the system, interlinkages between the
physical, economic and natural system require an interdisciplinary and complex method
of analysis. The option to easily conduct sensitivity analysis and scenario testing made
System Dynamics an ideal choice. Furthermore the transparent and visual
representation of the simulation model was considered as a clear benefit.
Insights on likely developments of the power plant park in Switzerland in dependency
of different external conditions are gained. Due to the complex interactions in the
system an investigation based on dynamic simulation is necessary and promises to give
more insightful results than a linear analysis of the problem.
The simulation model used for this study is specifically designed for the purpose of this
analysis. The System Dynamics software iThink 10.0.5 was used for the model
construction and simulation. Simulation results were exported and displayed in
Microsoft Excel.
The System Dynamics model used for this study was constructed in the framework of
the author’s master thesis for the completion of the Erasmus Mundus European Master
in System Dynamics. The project was a collaboration of the University of Bergen
(Norway) and the supercomputing systems Ltd. (Ziirich, Switzerland) under the
supervision of Prof. Erling Moxnes (University of Bergen). The research process was
oriented on the suggestion by Saunders and Lewis (2012). This project setting allowed
that numerous alternatives for model structures were developed, tested, improved or
also rejected. The model version presented here is the version considered as the most
valid, most direct to the point and with the highest explanatory value. A more detailed
description of the model, the underlying assumption and more in depth analysis can be
found in the report on the master thesis: WEBLINK. The web link is currently not
available yet. If you like to receive the full report contact the author under
merla@merla.net.
3.1. Model structure
The model is built on three main sectors. The sector physical system is the core of the
model. It represents the currently installed capacity for the different technologies and
the corresponding capacity supply line for capacity expansion. Also part of the physical
system is the remaining expansion potential for the various electricity sources. The
sector electricity market represents the immediate local electricity production, trade of
electricity and of course the market price for electricity. The section investment
decision is the central determinant for the development of the installed capacity.
The model distinguishes for ten different electricity sources. The array used is called
technology. The elements of this array are: photovoltaic, wind, nuclear power, gas
combined cycle, hydropower - distinguished into run-off-river hydropower, seasonal
storage lakes (called dam in the model) and pumped storage lakes; thermal power from
incineration, biomass and batteries. This separation of technologies is made to allow
understanding the different impacts of the overarching system on the individual
technologies and their development over time. The specific production characteristics
of the different technologies are the most central reason for this distinction. For
instance, while the production of photovoltaic plants is not controllable and totally
dependent on the incidence of solar radiation, biomass plants can produce flexible on
request. In the case of biomass plants the limiting factor are the availability of the input
resources or even more frequent the economic constraints of the production costs.
Treating photovoltaic plants and biomass as the same element in the array would
therefore be strongly misleading. Distinguishing the technologies enables a precise
definition of the seasonal electricity price, which determines production and
investments. Electricity cannot be distinguished by its source, if it is once fed into the
grid. Consequently technologies are heavily interplaying through the electricity price.
Additionally, using this array for technology allows seeing the actual components of the
electricity mix and measure the share of renewable sources. The chosen elements of the
array are consistent with the technologies considered in the SCS model to allow the
exchange of results.
The central dynamics included in the System Dynamics model are represented in a
simplified causal map in Figure 2. In the next section the major feedback loops are
described in more detail.
" im
total supply —
wr
a a
investment A wy ) \
nae | ] :
— A Ba \ market price
investment | ee |
% \ decision = 7 y
. ot accumulated
production
perceived .
es return =
Figure 2: Central dynamics represented in the System Dynamics model
The focus of this model lies on the development of the capacity expansion of the
different technologies and the investment decision steering the development of
capacities. The installed capacities of the technologies determine the production of
electricity at a specific point in time. Here the technology specific production
characteristics influence the amount and time of production. Additionally a feedback
loop for the capacity utilisation is included, ensuring that the flexible producing
technologies only produce at times where it is economic. Trade is represented very
rough. Electricity can be imported or exported to a certain capacity. The actual amount
traded depends of the relation between the local market price in Switzerland and
abroad. The market price is a very quick adjusting stock structure that represents the
Swiss market price in a seasonal manner.
A generalized market oriented investment structure is chosen. The exact number,
specific characteristics and the purchasing power of the investors are not modelled
explicitly. It is assumed that there are multiple investors all making their decisions
based on economic principles. Environmental thinking is not in their nature, as long as
it doesn’t match with profitability criteria. Nevertheless, the investors are not computers
and also don’t behave like homo economicus. Kahneman (2003) highlights that
decision makers (in his work called agents) frequently make intuitive decision based on
what they observe in the system, and not what they are able to calculate. Hampl (2012)
confirms in her three-part dissertation various behavioural and social effects on
decision-making in the energy industry. Investors in this model, although they aim to
make an economic decision, still have biases towards their experience and limited
perceptions. In line with these research the model uses perceived return as the relevant
input for the investment decision. Perceived return is an adjustment process based on
the annual return currently generated with on 1 GW installed capacity. The speed of
The Ebola Crisis and Public Fear Networks
A System Dynamics A pproach
Nasser Sharareh, nshararl @ binghamton.edu
Nasim S. Sabounchi, sabounchi@ binghamton.edu
Hiroki Sayama, sayama@ binghamton.edu
Systems Science and Industrial Engineering Department
State University of New Y ork (SUNY) at Binghamton
Abstract
One of the most important problems during a period of crisis in any country i is how
to respond to the public’s fear. There has been a lot of research in i ic diseast
and the corresponding fear. However, a mere paucity of these have employed a 1 system dynamics
(SD) approach to demonsirate the relationships between pandemics and the public response to
fear, the public perception of epidemic, and related organization's response to the issue. In this
study, an SD model has been developed to study the hidden relations of some irrational
behaviors and the spread of fear that accompanies the spread of disease. More specifically, our
focus is on fear among the population due to the Ebola crisis. This model can also be used for
the other outbreaks, and unfortunately, Ebola is not the last pandemic, and in the near future we
will confront with less and more severe situations again. By learning from the past and using
more systematic approaches, we will be able to stop the spread of disease faster. We conclude
that we should use fear as a leverage point to confront the disease, and try to increase public
attention in order to decrease the infectious rate.
Keyword
Ebola Virus Disease, Fear, Tweet, Irrational behavior, WHO, and Calibration.
Introduction
Changing people’s habits is always difficult, but analyzing fear patterns can make this
process easier for policy makers. It can be a strategy for regulators to implement their plan on
changing population’s habits. These kinds of strategies are called “Just As Well’ strategies
(Rowell, 2014). For instance, we can use people’s fear to teach them to wash their hand, stay
home when they are sick, and start taking the flu shot. These actions help people to take control
over their fear. The author mentions three reasons why these strategies are useful; they give
people the opportunity to overcome their fear by taking their fear seriously, they attract people’s
attention towards actions that can be useful for them, and they give motivation to groups to
defeat risks. Therefore, it’s a useless policy to remove fear which is raised by a fatal and
frightening disease. Rather, the best way to produce positive changes is to show people what
they can do to prevent getting infected, and try to increase public knowledge about the disease.
Overall, there are some factors that lead to the growth of fear from a specific disease.
These include the probability of being infected by some products or contact with people, the
possibility of fatality from the disease, the absence of control over the disease, the speed of
adjustment is determined by the previous experiences by the investor. Wang et al.
(2011) found that investors adjust their perceptions of a stock slower when they have
much experience with the stock, on the other hand the adjustment was much quicker
when they had few experience with the stock. Hamp! (2012) confirms this relation in
the specific field of energy. With the perceived return and the investment costs the
investor calculates the net present value (NPV) of an investor. Investment costs are
altered in relation to the remaining expansion potential of the technology. A scarcity
effect on the investment costs cause the investment costs to raise. Based on the NPV a
distribution of the investments is assumed in an investment function. This function is
multiplied with the existing installed capacity. This relation reflects the investment
power for certain technologies. To prevent a complete lock in effect a minimum
capacity is assumed that new technologies can develop in the model too.
Looking at Figure 2 we realize that the model mainly consists of balancing feedback
loops. This means, that the system has already a strongly self-regulating power. Central
in these dynamics is the market price, which governs the majority of the feedback
loops. Usually in System Dynamics a model focuses more on reinforcing feedback
loops that accelerate the problem under study. In this investigation the relation that
causes problems is the emission of green house gas emissions. This is not explicit part
of this model, but this fact determines the political will to define policies to support
new renewable energies. As this model is designed as a policy testing environment
besides other scenarios, the pressure for change is exogenous and is represented by the
will of the user to apply/test a policy. The same counts for the nuclear phase out and the
desired level of independency.
3.2. Model analysis and validation
The formal validation process was oriented on the suggested procedure by Barlas
(1996). All structure and structure-behaviour tests were conducted and passed.
Statistical behaviour tests were not conducted, since the reference mode is to short to
give reasonable results. However, the simulation results fit the reference data well but
as the reference mode is so short this is not very surprising. As an example, the fit of
the simulated price with the historic data is presented in Figure 3.
10
Market price vs. historic data
160000
sot \AJA
|
20000 ==>)
{
market price per GWh
—arket price data adjusted
Price per GWhin CHF
60000
40000
20000
¢
oP Se
se "
SSS
9 se fF FF FC HF ©
ae
Figure 3: Simulated and historic market price for electricity
S$ SS SS
NB BI I AB AS A A A a iT a
FS MF Pg
FFF PF HF HK HG GF WM HP 8
Sensitivity tests were made based on stable model condition. All policies were
removed. The model does not have a natural equilibrium despite all the balancing
feedback loops. Reason for this is that the model does not contain an automatic
compensation for depreciations. In this model this is deliberately not made. This model
is focussing on the capacity expansion seen from a market perspective. Investment is
purely driven by profitability and the available expansion capacity. Industrial dynamics
by Forrester (1961) as well as the beer game by Sterman (1989) analyse this mode of
behaviour and its determinants in more detail.
Removing the currently established feed-in tariff policy reveals that there would be no
investments into new renewables. The most drastic difference appears at the technology
dam, so the seasonal storage lakes. In the equilibrium model there is no installed
capacity for seasonal storage lakes at all. Today, with the current electricity price and
the investment and marginal costs seasonal storage lakes are simply not profitable.
These facts are supported by the statement of Robert Lombardini, the director of the
board of directors of Axpo the largest electricity producer in Switzerland, in an
interview for Basler Zeitung!.
The exact shape of the investment function investment relative to capacity is very
sensitive in the system. Here changes in the height, shape or base of the curve have a
significant impact in the system. It is observed that the system reacts especially
sensitive to changes in the height of the curve within the area of 0.3 and 0.6.
1 http: ine.ch/wirtschaft/unterne! -und: i ‘Die-Axpo-fragt-sich-Wie-konnte-es-so-
it-k ‘story/19719269 9.6.2014
11
Further more, significant drivers for change are the costs, which are treated as
exogenous in this model. The cost development of new renewable energies will
determine the speed and strength of an upcoming energy transition.
The price abroad and the trade capacity have a very similar and strong effect on the
system. The incentive and ability to import and export electricity lead to major changes
in the local price. A low price abroad, combined with sufficient trade capacity, leads to
a constant underinvestment in the local capacity expansion. A very high price abroad
on the other hand can lead to high investments in the beginning of the simulation
period, which leads to a lower local price in the mid-term. This phase is followed by a
period of high prices in the end of the simulation due to low investment as a
consequence to the previously low price. Trade is in first line working as a buffer for
irregularities, but it also can be seen as a hidden capacity. Altering the transmission
capacity is a politically sensitive policy, but it also has significant impacts on the
investment decision in the electricity provision system.
For the sensitivity analysis four runs with transmission capacities of 0, 1, 2 or 3 GW
were simulated. Here we notice, that trade is in first line working as a buffer for
irregularities. In scenario 1, where there is no transmission capacity, we see that a gap
between demand and supply lead to an enormous shock in price (Figure 4). On the
other hand with a transmission capacity of 4 GW there is only a slight and quite steady
increase in the price. Logically the price is influencing the perceived return of the
technologies and with this it has an impact on the investment decision (Figure 4). In
this light the more balanced price development enabled by the high transmission
capacity gets the negative aspect of blocking new investments. Ochoa (2007) and
Ochoa and van Ackere (2009) analyse this issue in the light of trade liberalization.
Swiss market price (adaptive) Accumulated investment in GW
350000 f\
300000
200000
150000
4
100000
2
50000
I\
250000 | \ 8
| \
| \
0
SSQBSagssesegee “eS eaHRHRRBRBAREES
RRRRRELKRRRRERR ZERERLSSRLERRRE
trade cap 0 CW trade cap 1 GW trade cap 0 GW trade cap 1 GW
"trade cap 2 GW =="trade cap 3 GW ====trade cap 2 GW =="""trade cap 3 GW
Figure 4: Sensitivity test with changes in transmission capacity — market price and
accumulated investment
4. Results
In the previous chapter we got a good overview on the model structure, improved
understanding the sensitive parts of the model and already tested the effect of nuclear
phase out in a deregulated model. In this chapter we are running the model with real
data. We start the simulation in the year 2006 and simulate it until 2050.
As a first step the base run is presented. We look into the major determinants shaping
the base run to understand, where relevant dynamics come from. In the next step we
experiment with policies to support the new renewable energies and analyse their
effectiveness.
4.1. Base run
The simulation run called base run is the basis for our analysis as well as for policy
comparison in the next section. The base run starts in year 2006. Table 2 in the
appendix shows the used initial values. The initial value for the market price is 82520
CHF per GWh, as it was in 2006 (Swiss Federal Office for Energy, 2014).
For the base run the following conditions are included in the model. The fixed price
FIT policy is stopped in the year 2015. For these years the new renewables receive the
FIT tariff according to the historic data. Afterwards the market price at the time of
production time counts for all technologies. The trade capacity is 2 GW at any point in
time. The price abroad is set on 70°000 Swiss Franks per GWh with variations of a
13
sinus curve of an amplitude of 5°000 Swiss Franks per GWh. The political will persists
on the nuclear phase out. The nuclear power plants are shut down according to the dates
currently expected. A hypothetical tax is set on electricity from nuclear power plants
preventing new investments. Production with the currently installed capacity of nuclear
power is allowed and not taxed.
We simulate the model with these conditions. Generally demand is covered in most of
the cases despite the nuclear phase out. Local supply of electricity first increases to
level higher than the initial value and also higher than demand. This rises exports of
electricity, therefore net imports are negative. In course of progressing nuclear phase
out local supply of electricity cannot remain on this high level and drops, after 2035
even under the demand.
Correspondingly to this development is the curve of the electricity price. The market
price first drops slightly in line with the oversupply of electricity. When the last nuclear
power plants are shut down and also the effect of the stopped FIT policy kicks in prices
start to rise again and reach higher levels (Figure 5). Important to notice is that the
fluctuations in the electricity price are increasing with higher share of renewables in the
power plant park and every nuclear power plant that is switched off. The fluctuations
moving along the production characteristics of photovoltaic and wind cause price lows
during their peak production times and price highs when their production is low. With
no nuclear power the share of these fluctuating technologies in the electricity
production increase and cause the price to fluctuate stronger. Interesting to see is that
the annual return for the technologies causing this fluctuations (so photovoltaic and
wind) only increases slightly with the increasing price in the end of the simulation, for
flexible producing technologies such as biomass and pumped hydro power plants the
annual return rises high.
14
Market price from 2006 to 2050 - base run
1180000
4160000
ios |
1120000 |
3
8
—market price per GWh
Price per GWh in CHF
8
8
—adaptive price
Figure 5: Base run — market price
Investments follow for the specific technologies fit the reference mode from 2006 until
2013 in satisfying manner. Afterwards the investments follow a realistic pattern (Figure
6). There is a major expansion of photovoltaic and wind as a consequence of the FIT
policy.
Installed capacity over time - base run
25
Binstalled capacity[gas}
2 20 S installed capacity[biomass]}
ic)
£ Sinstalled capacity[Wind]
ra
3s 6 “installed capacity[Photovoltaic]
a
8 installed capacity[nuclear]
a 10
& S installed capacity[batteries]
%
= s Binstalled capacity[thermal]
Hinstalled capacity[pumped]
installed capacityfriver]
installed capacity[dam]
Figure 6: Base run — installed capacity
After the ending of the FIT policy in 2015 the investments into new renewables fall to
zero. Despite the increase in price, there is no reinvestment into the technologies that
15
were originally supported by the feed-in remuneration policy. In the year 2045 an
increase in installed capacity for gas-fired power plants is observed. In other words, the
FIT policy pushes to system to a real energy transition towards new renewable
energies. But the policy is not sustaining the system in a state with new renewables.
With stopping the policy the transition is removed and the system falls back into
normal patterns (gas replaces nuclear in this moment). This confirms the apprehension
communicated by SwissCleanTech (2013a). The development of the investment into
new renewables is on one hand clearly determined by the Fit policy, as intended, on the
other hand there is also a significant development going on the costs. The data taken
from the Prognos study (Swiss Federal Office of Energy, 2007) are known as rather
conservative. The cost development for photovoltaic is updated with the real data for
2013, since already there the estimation were clearly above the value reached in 2013.
4.2. Policies
The simulation model is used to test different forms of FIT policies to support new
renewable energies and evaluate their effectiveness. We test the currently established
FIT model with a fixed tariff, the spot market price gap model, the premium FIT model
and FIT model granting a percentage of the market price. A set of variables is used to
compare the effectiveness of the policies. The selection of the variables is oriented on
the suggestions by IREA (International Renewable Energy Agency, 2014) but does by
far not reach that level of detail. The set of variables can be seen in Table 1.
Accumulated costs are not discounted.
The fixed tariff FIT policy is the policy applied in the base run. The policy enables a
good start into an energy transition towards new renewable energies. The share of new
renewable energies within the electricity production rises to around 20%, but then drops
down to 11% after the policy is stopped. Investment into new renewables is stopped
completely after the ending of the policy, despite significant cost improvements of the
new renewable energies. In the end of the simulation period there is even investment
into gas-fired power plants.
We analyse the impacts of applying the currently established feed-in remuneration
policy with fixed tariffs for the entire period until 2050. This policy is currently under
revision and will certainly be changed in the future. Nevertheless, we test the impacts
of the feed-in remuneration policy on the system when it is applied in the future with
the current format. For this simulation it is assumed that the feed-in remuneration tariffs
remain on a constant level after 2014. We observe that the effect of the policy goes in
the desired direction — a significantly increasing share of new renewable energies in the
total electricity production results. Initially the development is the same as in the base
run, where the same FIT policy with fixed tariffs is applied but stopped after 2015.
16
With remaining feed-in remuneration tariff the share of new renewables rises to a level
of about 0.25. In the end of the simulation period the percentage dropped a little. This
comes from lacking reinvestment as investments become more expensive with lower
expansion potential. Together with hydropower sustainable energies have a share in the
local production of 87 %. The remaining percentage is covered with imports. Total
investments in general accumulate to a value of 63420 million CHF of which the new
renewables are 607194 million CHF.
The spot market price gap FIT is another market price independent form for a feed-in
remuneration tariff discussed by Couture and Gagnon (2010). The policy ensures a
minimum receiver price for the producers benefiting of that policy with covering the
gap between the market price and the threshold set by the policy. The electricity
producers with new renewable energies receive the market price plus the difference to
the threshold. If the market price is higher than the threshold only the market price will
be paid off. This policy is, from a producer perspective, very similar to the fixed price
model. Theoretically the only difference is that they can receive a higher return when
the spot market price is very high. In practice this policy is usually implemented
without a purchasing guarantee for the produced electricity. So the investors have to
sell the produced electricity themselves on the electricity spot market. This could be a
hurdle for smaller investors such as households (Couture and Gagnon, 2010). This kind
of implications of a policy are not included in this simulation model but have to be kept
in mind when evaluating the policy. Simulation results will therefore be very similar to
the fixed price policy in terms of capacity expansion and price. Nevertheless, it is
interesting to see the difference in the total amount spent for the policy and the total
costs on consumers.
A premium FIT pays a fixed premium for the production of electricity of new
renewable energies. This premium comes in addition to the market price. This is the
system that is most likely to be applied as the new policy instead of the fixed price
policy. For this simulation a constant premium is chosen that leads to a share of new
renewable energies that is comparable to the other policies to allow comparison of
costs. The premium necessary to reach this level is 52000 CHF per GWh. In terms of
implementation this policy is easier to handle and doesn’t create access barriers to small
investors. Nevertheless, the return risk is higher as there is no guaranteed price for the
produced electricity.
An alternative to the previously discussed policies is a FIT that gives a percentage of
the market price to the producers. This policy is artificially accelerating the
fluctuations of the market price in the view of the investors and gives incentives to
produce, when the market price is high. For implementation this policy is rather
complicated, as one would need to know how much every producer was producing at a
specific point in time. Usually the measuring system is not that developed to enable this
17
properly. The percentage was chosen in the manner that again a similar share in new
renewable energies is resulting at the end of the simulation period. 60% is the
percentage reaching this.
In this investigation four alternative policies for the support of new renewable energies
were tested in a dynamic simulation model. The policies are compared in Table 1,
Figure 7. Table | lists the results values for the policy evaluation criteria for the four
tested policies and the base run.
cs a 73878 65°483 64'960 82'243 82364
Sandad 0.23 0.25 0.25 0.19 0.19
deviation price
siete pet 11% 25% 25% 25% 26%
renewables
share renewables ant ox a ae re
plus hydro
accumulated
investment in mio 31'737 63°420 63'433 33/214 34/157
acc investment
into new . , ,
prot] SES 60194 60205 27'053 28°03
poles cess (nmi 9'213 69'025 51653 11'905 11/010
consumer
spendings in mio 205'784 182/398 183013 229'083 229'421
total costs on
consumers in mio 214'997 251'423 234'666 240'987 240'431
Table 1: Policy comparison with evaluation set
The table highlights that all tested policies have a positive impact on the expansion of
new renewables. The share of new renewables increases significantly. The share of
green energies in the total electricity mix reaches levels between 80 and 87 percent. In
all scenarios the coverage of demand also uses imported electricity from abroad. In the
case of the premium FIT and the percentage of market price FIT there is even
investment into gas-fired power plants as can be seen in Figure 7.
18
Policy scenarios - installed capacity in 2050
25 ™ Pumped
® Batteries
0 "Biomass
BThermal
bam
"River
BGas
0
Nuclear
=Wind
5
= °
0
Base run Fit fixed price spot price gap FIT premium FIT —_ percentage FIT
Installed capacity for 2050 in GW
Figure 7: Comparison of policy scenarios — installed capacity
Table | highlights that the costs to conduct the policy are the lowest for the premium
FIT and the percentage of market price FIT. They both cause costs of only around
11000 million CHF. Although only is also here belittling. Those two policies are low
in costs but the market price is on a higher level with these support systems. Therefor
the consumer spendings and the total costs on consumers are high. Oriented along the
costs on consumers the FIT policy based on the gap between the spot market price and
a defined tariff is the most efficient support policy.
Interesting to see is that in this simulation the spot market price gap FIT can reach the
same goal as the fixed price FIT with clearly fewer costs. The money saved is about
20°000 million CHF. This indicates that with a shift from the currently applied fixed
price FIT to the spot market price FIT a lot of money could be saved. However, as
already mentioned earlier, the spot market price gap FIT brings hurdles for small
investors. This could have a significant impact on the expansion of photovoltaic, since
these plants are frequently built on the house roofs of private persons.
However, this investigation will not be able to draw a final conclusion or
recommendation on which policy is best to support the new renewable energies in their
investment. The policies were not tested within their full potential. It was always
assumed that the tariff or the quota remain on the same level. Generally it would be
possible that these tariffs or percentages are adjusted to the current state of the system.
This would allow to steer the system in more precise manner.
However, we are able to draw some general conclusions on the effectiveness of the
tested policies and what might be improved to reach a higher policy effectiveness. All
the FIT policies can significantly increase the expected annual return of an investment
disease spread and a high number of mortalities, the low degree of knowledge and high degree of
uncertainty about symptoms, and the mode of transmission of the disease.
Data analysis and current trends
In Figure 1, you are provided with the total cumulative number of death from Ebola Virus
Disease in three main countries, Guinea, Liberia, and Sierra-Leone. The time period of death toll
study is from March 2014 to March 2015.
20000 Cumulative Total Death
18000 ——=Total Death .
16000 7
14000 — fitted trend rail
12000 line Fear
10000 ne =
8000 Hone
6000
4000
2000
0
Cr SC) o
Bh HH hh yh yh yh
Xr We HM AW WW” eo om a” er A WM Wh Gh
Figure 1 - Cumulative total death of Ebola
As can be seen, the total death had an exponential growth, which can be seen by the dark
dotted line. Our hope is that total deaths from Ebola stops, but our fear is that total deaths
continue to grow.
System Dynamics Model
We have developed a system dynamics model to understand the causal mechanism
underlying the changes in fear and how this impacts the spread of Ebola. We will describe some
of the interesting insights we have captured within the model shown in Figure 3.
In the case of Ebola Virus Disease, the people in United States started to experience a lot
of fear due to the first and second infected nurse (Lyon et al., 2014), who were treating the first
Ebola infected patient in a U.S. hospital. A fter spreading fear through the population, a massive
amount of news was broadcasted and everyone was talking about Ebola, its symptoms, and
strategies to avoid getting infected. However, healthcare organizations like WHO and CDC,
were trying to cover the issue, and as a result, media was releasing news with uncertainty to
make these organizations reveal further information about the disease spread. This whole process
reinforced the increase of fear among population, which is captured by feedback loop (R2 in
Figure 3Error! Reference source not found.).
Meanwhile, healthcare practitioners started to release useful data from their own
experience, and from reliable, credible sources such as CDC website(CDC, 2014) and New
19
and also reduce the investment risk. As the European Commission (2008) correctly
says, the FIT policies have the potential to strongly push the new renewable energies in
their development and kick start an energy transition.
Nevertheless, the feed-in remuneration is in all forms very cost intensive. Simulation
results clearly showed that the policies don’t have a sustainable effect on the system.
Without the policy there is a lack of incentives for reinvestment into renewables.
Therefore when the policy is removed the energy transition is reversed. The necessity
of an external entity to define the tariffs, points towards a lacking dynamic structure of
these policies. Further research is needed to design a policy that can sustain the
electricity provision system in the state after the transition without generating enormous
costs.
Strongly regulated systems and frequent changes in policies bring the risk of confusing
the investors, and therefore increase the perceived risk. It is generally already observed
that investors hesitate to invest in technologies that depend on or are affected by public
policies (Hampl, 2012). There might be very relevant dynamic aspects that are
currently not considered in the simulation model. Incorporating an endogenous
modelling of risk in the model is definitely a considered step for future research.
Policies that are very sophisticated and have the theoretical potential to steer the system
very well might fail in this point and be to complicated for investors and prevent
instead of support their investment. It would also be for example also interesting to see
the effect of the time of communication of the feed-in remuneration tariffs by the
government on the risk perception. A model capturing all these aspects would be
extremely interesting and could lead to very relevant insights.
5. Discussion and conclusion
Switzerland is facing two major challenges in its electricity provision. First, the Federal
council decided on the withdrawal of nuclear power. The stepwise shut down of the
five nuclear power plants of 3.28 GW will cause a major gap in the future electricity
provision. Second, a clear commitment to new renewable energies was made. This
situation brings challenges and chances.
In this investigation a System Dynamics simulation model of the Swiss electricity
production was build. The focus lies on the dynamic interactions of the determinants of
the capacity expansion of the specific technologies, and the investment decision
connected with it. The model captures the development of ten different electricity
production technologies: photovoltaic, wind, nuclear, gas, run-off river, seasonal
storage lakes, thermal power, biomass, batteries and pumped hydro-power. Investments
in this model are made upon a market-oriented investment structure. There is no central
20
planning entity included in this model. Investors are modelled as profit-oriented, but
not perfectly rational. Most important input for the investment decision is the
perception of return, which could be generated with an investment into this technology
per year. This is heavily determined by the market price and the time and shape of its
fluctuations. The production characteristics of a technology define at what time
electricity can be produced and very relevant to which price the technology can be sold.
Analysis of the model reveals that an electricity system, designed as in this simulation
model, always leads to long-term oscillatory behaviour, because there is no central
management compensating for depreciation of installed capacity. In this model gas-
fired power is the technology that is most frequently used to fill this gap, but also
suffers from the oscillations. This is important to know, as the Swiss Federal Council
plans to construct gas-fired power plants to compensate for the phased out nuclear
power plants. Sensitivity tests showed that the capacity for trade of electricity and the
electricity price abroad are very sensitive elements in the system that have the potential
to cause major changes in the model and system behaviour. More investigation is
needed to understand how these elements can be used to support the new renewable
energies. Additionally, the investment function used in the model has very sensitive
areas. More detailed research would be necessary to investigate in the exact shape of
this curve. With increasing shares of renewable technologies the price tends to fluctuate
stronger. In this framework the development of profitable storage options is very
important. Currently the most relevant storage technology, namely the seasonal storage
lakes, are not profitable and no further investments are made. This observation is
supported by the model results.
The model was used to test the effect of the currently established fixed price feed-in
remuneration tariff (FIT) policy and alternative forms of FIT policies. Comparison of
the effectiveness of these policies revealed that FIT policies are good instruments to
boost the initial development of new renewable energies. Market independent FIT
models are very cost intensive, while market price dependent FIT models lead to fewer
governmental costs for the policy. The spot market price gap FIT model caused the
lowest total costs for the consumers. Simulation results indicate that all FIT policy
models cannot bring a sustainable change into an electricity provision system.
Whenever a policy is stopped, the power plant park constellation that just made a
transition towards new renewable energies moves back to an old state. Further research
is necessary, on how these policies can be combined over time to enable an ideal
energy transition. Further more, a dynamic policy should be developed and tested that
can maintain the system in its state after the transition.
This research contributes to the existing knowledge about the Swiss electricity
provision system and its transition to a more sustainable state, with simulating the
21
investment decision for the different technologies endogenously. The simulation
framework was here used to test different models of FIT policies. The developed
System Dynamics model gives options for much broader scenario testing in the wide
field of electricity supply.
Acknowledgement
I would like to thank my supervisor Prof. Erling Moxnes, from the University of
Bergen, Norway, for supervising my thesis and introducing me to the beauty of small
models. This thesis gained from a collaboration with the Supercomputing Systems Ltd.
Company, Ziirich, Switzerland. My thanks go to Alain Brenzikofer and Andreas
Kettner for their feedback, openness to System Dynamics and assistance in the data
collection process. This thesis was written in the framework of the European Erasmus
Mundus Master in System Dynamics (EMSD). My gratitude goes to all people who
made that joint master program possible; and my colleges from the EMSD cohort 2012-
2014 for two very joyful and inspiring years. I would like to thank my parents for
supporting me in all stages of my life and for financially enabling me to participate in
this program.
References
Barlas Y. (1996). “Formal aspects of model validity and validation in system
dynamics.” System Dynamics Review 12: 183-210.
Couture T. and Gagnon Y. (2010). “An analysis of feed-in tariff remuneration models:
Implications for renewable energy investment.” Energy Policy 38: 955-965.
Dangerman J., and Schellnhuber H. J. (2012). “Energy systems transformations.”
Proceedings of the National Academy of Sciences of the United States of
America.
European Commission (2008). “The support of electricity from renewable energy
sources.‘ European Union. Available at:
http://ec.europa.eu/energy/climate_actions/doc/2008_res_working_document_en.
pdf.
European Commission. (2010). “World and European Energy and Environment
Transition Outlook. European Union.” European Union. Available at:
http://ec.europa.eu/research/social-sciences/pdf/publication-weto-t_en.pdf
Forrester, J. W. (1961). Industrial Dynamics. Cambridge: MIT Press.
Hampl N. (2012). “Energy Investment Decision-Making Under Uncertainty: The
Influence of Behavioral and Social Effects — Dissertation.” University of St.
Gallen. Available at:
http://verdi.unisg.ch/www/edis.nsf/SysLkpByldentifier/4083/$FILE/dis4083.pdf
Interface Ltd., Ernst Basler + Partner AG, and University of Geneva. (2012).
“Evaluation der kostendeckenden Einspeisevergiitung (KEV)”. Swiss Federal
Office of Energy. Available at:
http://www.news.admin.ch/NSBSubscriber/message/attachments/28353.pdf
22
International Renewable Energy Agency (2014). “Evaluating Renewable Energy
Policy: A Review of Criteria and Indicators for Assessment.“ International
Renewable Energy Agency. Available at:
http://www.irena.org/DocumentDownloads/Publications/Evaluating RE Policy.p
df
Kahneman D. (2003). “Maps of Bounded Rationality: Psychology for Behavioral
Economics.” The American Economic Review 93(5): 1449-1475.
Moxnes E. (1990). “Interfuel substitution in OECD-European electricity production.”
System Dynamics Review 6: 44-65.
Naill R.F. (1992). “A System Dynamics model of national energy policy planning.”
System Dynamics Review 8: 1-19.
Ochoa P. (2007). “Policy changes in the Swiss electricity market: Analysis of likely
market responses‘. Socio-Economic Planning Sciences 41: 336-349.
Ochoa P., van Ackere A. (2009). “Policy changes and the dynamics of capacity
expansion in the Swiss electricity market”. Energy Policy 37: 1983-1998.
Paul Scherrer Institut (2005). “Neue erneuerbare Energien und neue Nuklearanlagen:
Potenziale und Kosten.“ Paul Scherrer Institut. Available at:
http://www.windland.ch/doku_wind/PSI-Bericht_05-04sc.pdf
Prognos AG (2012). “Die Energieperspektiven fiir die Schweiz bis 2050.” Swiss
Federal Office of Energy.
Saunders M., and Lewis P. (2012). Doing Research in Business & Management.
Pearson.
Sterman, J. (2000). Business Dynamics. McGraw-Hill.
Sterman, J. (1989). “Modelling Managerial behaviour: Misperceptions of Feedback in a
dynamics Decision Making Experiment.” Management Science 43: 301-335.
Supercomputing Systems Ltd. (2013). “SCS Energiemodell — Simulation der
elektrischen Energieversorgung der Schweiz anhand von konfigurierbaren
Szenarien.” Supercomputing Systems Ltd. Available at: www.scs.ch.
SwissCleanTech (2013a, unpublished). “Die Zukunft des Strommarketdesigns.”
SwissCleanTech.
SwissCleanTech (2013b, unpublished). “Investitionskosten-Refinanzierungszuschlag
(IRZ).” SwissCleanTech.
Swiss directive for FIT, photovolatic (Richtlinie kostendeckende Einspeisevergiitung
(KEV) Art. 7a EnG Photovoltaik Anhang 1.2 EnV). Available at:
http://www.bfe.admin.ch/themen/006 12/02073/index.html?lang=de&dossier_id=
02168
Swiss energy enactment (Schweizerische Energieverordnung). Available at:
http://www.admin.ch/opc/de/classified-compilation/1998339 1/index.html
Swiss energy foundation (2006). “SES Positionspapier — Gaskraftwere.” Swiss energy
foundation. Available at:
http://www.energiestiftung.ch/files/downloads/positionenses-strom-
gaskraftwerke/ses_positionspapier_gaskraft.pdf
Swiss energy law (Schweizerisches Energiegesetz). Available at:
http://www.admin.ch/opce/de/classified-compilation/19983485/index.html
Swiss Federal Council (2008). “Press release - Kostendeckende Einspeisevergiitung:
Wichtige Informationen zum Anmeldeverfahren.” Swiss Federal Council.
Available at:
http://www. bfe.admin.ch/energie/00588/00589/00644/index.html?lang=de&msg-
id=18371
Swiss Federal Council (2011), “Press release — Bundesrat beschliesst im Rahmen der
neuen Energiestrategie schrittweise Aussteig aus der Kernenergie.” Swiss Federal
23
Council. Available at: https://www.news.admin.ch/message/?lang=de&msg-
id=39337
Swiss Federal Office of Energy (2007). “Wochenstatistik Elektrizitatsbilanz 2006.”
Swiss Federal Office of Energy. Available at:
http://www.bfe.admin.ch/dossiers/00767/index.html?lang=de
Swiss Federal Office of Energy (2013a). “Schweizerische Elektrizitatsstatistik 2012.”
Swiss Federal Office of Energy. Available at:
http://www.bfe.admin.ch/themen/00526/00541/00542/00630/index.html?dossier_
id=00765
Swiss Federal Office of Energy (2013b). “Energieperspektiven 2050 —
Zusammenfassung,” Swiss Federal Office of Energy. Available at:
http://www.bfe.admin.ch/themen/00526/00527/index.html?dossier_id=05024&la
ng=de
Swiss Federal Office of Energy (2014). “Marktpreis gemass Art. 3 f, Abs. 3 EnV.“
Swiss Federal Office of Energy. Available at:
http://www. bfe.admin.ch/themen/00612/02073/index.html?lang=de&dossier_id=
03136
Wang M., Keller C., and Siegrist M. (2011). “The Less You Know, the More You Are
Afraid of—A Survey on Risk Perceptions of Investment Products.” The Journal
of Behavioral Finance 12: 9-19.
Appendix
I. Complete stock-and-flow diagram
a ve tupuntap C)
oud yo Anonsea
vondeared winies
tu sbuey>
‘umop anys
seaponu jo ays
suonemong sou
‘yep 40 aye
oyu
om ulna sed me
09 Hon 19 9 AdN
sisoD ounsanul
‘500 asanu uo
‘Baye Saaueos
fypedeo or &
‘saneyes weunsanuy
renusied dn pang
g
‘Ausedes 40
Axpqepene yeowuyoar
2e9pon op
‘ued unop anys:
euoseas ‘Ayoedeo je
25
IL. Initial values
Photovoltaic 0.011 0.05 0.029 10.9
Wind 0.002 0 0.012 1.156
Nuclear C) to} 3.278 0.002
Gas 0 oO 0 3.85
River 0.01 0.005 3.652 0.303
Dam 0.001 0.1 7.961 0.298
Thermal 0.01 oO 0.355 0.055
Biomass 0 ° 0.032 0.358
Batteries i} ° i} 2
Pumped 0 0 1383 0.497
Table 2: Initial values used for the stocks in the sector physical system
England Journal of Medicine (NEJM, 2014). This helped people to learn about the facts and
disease, and consequently decrease fear among the population (Lee, 2014). This is demonstrated
by loop B7. Furthermore, as Symplur (2015) has demonstrated through analyzing doctors’ tweet
activity, the more they tweeted and released useful information, the less tweets from the public
were made, and so the less fear there was among people (Figure 2). Tweets about Ebola
represent fear or awareness.
Ebola - Doctor tweets rise while general public’s falls
Twitter, Ju 28 through Sep 19,2014 #Ebola, #bolaDubreak,HbolaVirus,#boleatch
\While general population tweets were
consistently faling, U.S. doctor tweets
Increased by an average of 23%
‘each weak during ths time,
Tweets
oe ee ee eee FE
Figure 2 - Doctor Tweets about Ebola facts
Reference: Symplur
for conftonting disease
sarees amg the counts!
ic mae ote sped i ease
%
Information about at Pintc contact
rk counties tensionangst tbe pevnt
+
Panic and amity
anette Number of closed
News broadcasts = y te tresmting fear Sige, Dance ‘Tratpottion of food and
sical spe ed
Pressure on over
to close the border
+
SX _ Fear hn the +
popula r \
al civuttances
+ 7 ors fomtman:ett “rife peop
ffots of elated organizations Ans) ises to openite
Ps ate wi the public — bodes
about Ebola pannel [Number ofinfected
reported
4 q ri XK __Nabeot Ae)
@ A Disease spread
=
é
& gmptoms vi pictues
Comparison bebween the
feath fom Ebola & other Estimation ofthe
tanec iret contact
(face-to-face)
Public knowledge
about disease
?
Publ awareness ofthe we
possy oh canes
Avoidance of public Space
healthcare practitioners ‘when sick (because of
disease
‘or fu)
Figure 3 - Causal Loop Diagram
Simulation
We have developed simulation models, starting from the basic SIR model. Our objective
is to capture the underlying social and behavioral mechanisms by calibrating our model to
historical number of deaths and cases from WHO.
In each step, we calibrated some variables and did sensitivity analysis on variables to see
which of them had the most significant influence on the output of model.
Conclusion
Overall, the aim of this paper is to establish that public perception and risk management
are two critical points in the case of pandemics. Further, it is vital to consider the whole system
and unintended consequences before implementing our policies because disease spread can
affect all parts of society. Also in the case of the spread of a fatal disease, such as Ebola, it is not
practical to attempt to eliminate fear in the public. On the contrary, we should use fear as a
leverage point to confront the disease, and try to increase public attention in order to decrease the
possibility of getting infected.
We plan to use our refined simulation model calibrated with the WHO data, to test
different policy scenarios in leveraging public fear and awareness to deal with the spread of fatal
diseases such as Ebola.
References
CDC. 2014. CDC. Retrieved March/16. Available from http://www.cdc.gov/vhf/ebola/about.html.
Lee TM. 2014. Ebola on social media shows some revealing insights. Retrieved March/6.
Available from http://www.symplur.com/blog/ebola-social-media-first-look-facts/.
Lyon GM, AK Mehta, J B Varkey, K Brantly, L Plyler, AK McElroy, CS Kraft, ) S Towner et al.
2014. Clinical care of two patients with Ebola virus disease in the United States. New England
J ournal of Medicine 371(25): 2402-2409.
NEJ M. 2014. The New England J ournal of Medicine. Retrieved Macrh/16. Available from
http://www.nejm.org/page/ebola-outbreak.
Rowell A. 2014. Regulating Fear: The Case of Ebola in the United States. Available at SSRN
2513130.
Symplur. 2015. #£ bola healthcare social media hashtag. Retrieved March/16. Available from
http://www.symplur.com/healthcare-hashtags/ebola/.
