Energy Policy 39 (2011) 936–949

Contents lists available at ScienceDirect

Energy Policy
journal homepage: www.elsevier.com/locate/enpol

Electric-systems integration in the Andes community:

Opportunities and threats$
Enzo Sauma a,n, Samuel Jerardino b, Carlos Barria b, Rodrigo Marambio b, Alberto Brugman c, José Mejı́a c
a
Industrial and Systems Engineering Department, Pontificia Universidad Católica de Chile, Ave. Vicuña Mackenna 4860, Raúl Déves Hall, 3rd floor, Macul, Santiago, Chile
b
Kas Ingenierı́a S.A., Brown Norte 694, Ñuñoa, Santiago, Chile
c
Estudios Energéticos Limitada, Cr 13a #89-38 Of 424, Distrito Capital, Bogotá, Colombia

a r t i c l e i n f o a b s t r a c t

Article history: This paper identifies some sustainable and technically feasible alternatives for electric exchange through
Received 30 August 2010 interconnections among the electric systems of Bolivia, Chile, Colombia, Ecuador and Peru. In particular,
Accepted 5 November 2010 we assess such interconnections from both technical and economic perspectives, and identify the main
technological, commercial and regulatory barriers for their development. The analysis is carried out at the
Keywords: pre-feasibility level from both private and social point of views, based on the assessment of different
Interconnected power systems investment alternatives in the transmission systems among the aforementioned countries.
Power transmission economics We show that, even when keeping the security and self-sufficiency of the power system of every
Stochastic dual dynamic programming country (i.e., when not altering the generation expansion plans of the countries), the proposed
interconnections have significant economic benefits in the long run. These benefits come from the
supply side, the demand side, the system-cost savings and the environmental side. We also analyze the
commercial and regulatory issues that must be addressed to accelerate the regional energy integration,
and provide some policy recommendations.
& 2010 Elsevier Ltd. All rights reserved.

1. Introduction greenhouse gas emissions, given the better use of the resources of
each country.
The increasing costs of electricity and the difficulties to expand Latin-American countries have diverse and abundant energy
the power generation capacity, as well as the need for increasing resources: oil, natural gas, coal, biomass and other renewable
the energy security levels, have enhanced the potential benefits of sources, as well as a large hydro potential. These resources are not
the electric interconnection among countries and the formation of symmetrically distributed. This asymmetry is precisely what
sub-regional energy markets. The electric interconnection allows underscores the potential for developing important energy
reducing the need for generation capacity reserves and, at the same exchanges in the region.
time, providing a higher security level of supply, given the different There are successful examples of electric interconnection all
production and consumption cycles in the countries, which reduces around the world and an increasing interest in its expansion. In the
energy prices in the long run. Moreover, an important benefit of Andes Community, its members – Bolivia, Colombia, Ecuador, Peru
electric interconnections is the potential contribution to reduce and Venezuela – approved Decision No. 536 in 2002, under which
they agreed the electricity integration of the countries of the Andes
$
Community and established the objectives and rules for the
The work reported in this paper was partially supported by the United Nations
operation of the regional interconnections (CAN, 1969, 2002).
Development Program (UNDP) under a grant associated to the project reported in
UNDP (2010) and by FONDECYT Grant No. 1100434. The work reported in this paper Colombia and Ecuador are electrically interconnected since
is based on the project reported in UNDP (2010), which was required by the energy more than six years, which have helped to partially optimize the
authorities of the countries involved and supported by the UNDP. The main goals of joint energy resources, taking advantage of the market conditions
that project were to identify, characterize and analyze potential electric inter-
that benefit them at every period of the coordinated dispatch. In
connections among the Andes-region countries, to quantify the benefits associated
to taking advantage of the energy complementarities and to face the technical, 2004, Peru and Ecuador built a transmission line (line Zorritos-
economical, organizational and regulatory aspects required to the development of Machala, in 220 kV), which is still inoperative due to normative
the interconnections. The project was made in 2009 and the results were delivered differences between both countries. These examples show there is
in January of 2010. The ideas presented in this paper are only responsibility of the a significant potential for the development of electric interconnec-
authors and they do not represent the position or thoughts of the UNDP nor any of
tion, but also some barriers that limit the use of this potential.
the authorities of the countries involved in this research.
n
Corresponding author. Tel.: + 56 2 354 4272; fax: + 56 2 552 1608. There are some few works in the literature about electric-
E-mail address: esauma@ing.puc.cl (E. Sauma). system integration in Latin America. Hammons et al. (1997)

0301-4215/$ - see front matter & 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enpol.2010.11.019
 E. Sauma et al. / Energy Policy 39 (2011) 936–949 937

analyze the energy market integration in South America by study- Table 1

ing the electric power, the oil and the natural gas industries of Interconnection scenarios and start-up dates of the interconnections.
Argentina, Bolivia, Brazil, Chile, Paraguay and Uruguay from an
Scenario 9 start-up date Co–Ec Ec–Pe Pe–Ch Bo–Ch
institutional-structure perspective. In Hammons et al. (1997), the
authors emphasize some legal, institutional and regulatory issues Base scenario Current
that must be addressed to facilitate the regional energy integration. Scenario 1 2014 2015
They also analyze the potential interactions between gas and Scenario 2 2014 2015 2016
Scenario 3 2014 2015 2016 2017
electricity markets.
In both, Rudnick (1997) and Hammons et al. (1999), the authors
analyze the impacts of the deregulation of Latin-American electric
systems over the power-generation sector and its effect on the transmission systems, as well as the future evolution of such
incentives to accelerate the regional integration of the energy features according to the planning studies performed by the energy
sectors. More recently, Barroso et al. (2008) have studied the authorities of each one of the countries involved.
integration in South America of the liquefied natural gas (LNG) The base scenario corresponds to the representation of the
from the perspectives of the markets, the energy prices and the current state of the electric interconnections among the five
security of energy supply. countries considered here. That is, it comprises only the current
Within this context, this paper identifies some sustainable and interconnection lines between Colombia and Ecuador (since,
technically feasible alternatives for electric exchange through despite the fact that there is an electric interconnection between
interconnections among the electric systems of Bolivia, Chile, Ecuador and Peru, it does not operate due to the lack of commercial
Colombia, Ecuador and Peru, thus allowing optimizing the use of agreements, in addition to the fact that it cannot work in a
the energy resources. In particular, we assess such interconnec- synchronic manner).
tions from both technical and economic perspectives and identify Starting from the base scenario, we build other three scenarios,
the main technological, commercial and regulatory barriers for which integrate countries going from north to south until reaching
their development. The analysis is carried out at the pre-feasibility the interconnections from Bolivia and Peru to Chile. As we will
level from both private and social point of views, based on the describe in Section 3, the design of these scenarios is the result of a
assessment of different investment alternatives in the transmission long-term analysis, made using the Stochastic Dual Dynamic
systems among the aforementioned countries. For doing this Programming (SDDP) model, based on a one-node representation
analysis, we study four interconnection scenarios (including the of the electric system of each country.
base scenario) in the time horizon 2010–2022, in addition to a It is important to underscore that in reference to Chile, we define the
sensitivity analysis for each scenario considering the opportunity non-existence of a future interconnection between the main inter-
prices of natural gas for both exporting and importing countries. connected systems (the Sistema Interconectado Central, SIC and the
From both the private and the social point of view, the analysis Sistema Interconectado del Norte Grande, SING), in agreement with the
focus on quantifying economic costs and benefits through a current Chilean generation and transmission expansion plan. Conse-
methodology that considers different aggregation levels depending quently, the international electric integration with Chile is assumed, at
on the type of agent considered. The modeling of the different all times, being the integration with the SING.2 Given its size,
future economic operation conditions for each one of the con- introducing the interconnection to the SIC–CHILE significantly disturbs
sidered electric systems, and for each one of the assessed scenarios, the obtained outcomes, making no longer credible the self-sufficiency
is a central element of the analysis. of countries’ supply assumption (which is a strong assumption made in
The rest of the paper is organized as follows. Section 2 shows the this article, as explained later).
interconnection scenarios analyzed here. In Section 3, we describe The interconnection scenarios analyzed in this paper (other
the methodology used in the analysis. Section 4 shows the results of than the base scenario) are the following:
simulating a joint and coordinated operation of the main transmis-
sion systems of the different countries, considering some proposed
interconnections. The results show that there are significant social a. Scenario 1: it considers the reinforcement of the currently
benefits due to the proposed interconnections. These benefits come available interconnections between Colombia and Ecuador
from the supply side, the demand side, the system-cost savings and and the construction of a new interconnection between Peru
the environmental side. In Section 5, we analyze the commercial and Ecuador. The rest of the conditions are similar as those in the
and regulatory issues that must be addressed to accelerate the base scenario.
regional energy integration in accordance with local markets. In b. Scenario 2: it adds an electric interconnection between Peru and
this section, we also provide some policy recommendations. Chile (SING) to the conditions stated for Scenario 1.
Section 6 concludes the paper. c. Scenario 3: it adds an electric interconnection between Bolivia
and Chile (SING) to the conditions stated for Scenario 2.

2. Interconnection scenarios analyzed

Table 1 summarizes these interconnection scenarios for the
This work makes a quantitative analysis based on the design of Andes region, indicating the corresponding start-up dates of the
different scenarios of interconnections. These scenarios have a interconnections.
progressive, geographic and sequential order. All these scenarios All the interconnection scenarios considered here use the
incorporate the representation of the main electric systems of the forecasts of the electricity demand shown in Table 2.
countries (based on the official information available for each The forecasted average electric-energy demand growths for the
country).1 Such information includes the current characteristics of 2010–2022 period are as follows: Bolivia: 6.3%; Chile-SING: 5.1%;
the generation supply, the electricity demand and the main Colombia: 3.5%; Ecuador: 5.5%; and Peru: 6.7%.

1 2
The official information of the features of the main electric systems of each Nevertheless, in the appendix of UNDP (2010), it is considered an additional
country was provided by the energy authorities of the respective countries in the scenario that interconnects the SIC and the SING, which allows the quantification of
context of the project reported in UNDP (2010). the effects of integrating the main Chilean electric systems.
938 E. Sauma et al. / Energy Policy 39 (2011) 936–949

Table 2 Table 4
Forecast of the annual electric-energy demand for each country during the 2010– Diesel fuel-oil prices ($/MBtu) considered in each country in the period 2010–2022.
2022 period (GWh).
Year SING-Chile Colombia Ecuador Peru
Year Bolivia Chile-SING Colombia Ecuador Peru Total
2010 16.07 24.44 18.99 21.34
2010 5883 14,320 55,913 18,278 29,956 124,350 2011 21.49 26.85 20.72 21.34
2011 6356 15,034 57,849 21,355 31,637 132,231 2012 22.99 28.86 22.17 21.34
2012 6717 15,784 59,885 22,417 34,095 138,898 2013 23.72 29.84 23.12 21.34
2013 7131 16,573 62,160 23,716 37,665 147,245 2014 24.33 32.07 24.93 21.34
2014 7571 17,401 64,547 25,059 43,477 158,055 2015 25.02 33.01 25.58 26.86
2015 8041 18,302 66,906 26,504 47,661 167,414 2016 25.78 33.38 25.65 26.86
2016 8543 19,249 69,321 28,332 49,478 174,923 2017 26.63 33.45 25.58 26.86
2017 9078 20,246 71,821 30,146 51,427 182,718 2018 27.58 33.82 25.87 26.86
2018 9649 21,294 74,476 31,712 53,811 190,942 2019 28.60 34.02 25.94 26.86
2019 10,253 22,396 77,245 33,181 56,320 199,395 2020 28.60 33.76 25.80 26.86
2020 10,895 23,556 79,734 34,604 58,963 207,752 2021 28.60 33.76 25.80 26.86
2021 11,578 24,778 81,818 36,294 61,748 216,216 2022 28.60 33.76 25.80 26.86
2022 12,303 26,067 84,443 38,096 64,684 225,593

Table 5
Natural gas prices ($/MBtu) considered in each country in the period 2010–2022.
Table 3
Coal prices ($/ton) considered in each country in the period 2010–2022.
Year Bolivia SING-Chile Colombia Ecuador Peru

Year SING-Chile Colombia Peru

2010 1.30 9.21 4.36 3.77 1.74
2011 1.30 10.16 4.45 3.77 1.82
2010 128.03 52.70 150.50
2012 1.30 10.82 4.28 3.77 1.92
2011 133.36 50.64 150.50
2013 1.30 10.82 4.04 3.77 2.05
2012 140.80 47.98 150.50
2014 1.30 10.82 3.93 3.77 2.19
2013 142.63 46.59 150.50
2015 1.30 10.82 3.77 3.77 2.35
2014 143.34 45.95 150.50
2016 1.30 10.82 3.65 3.77 2.51
2015 145.40 45.16 189.40
2017 1.30 10.82 3.56 3.77 2.69
2016 148.71 44.76 189.40
2018 1.30 10.82 3.57 3.77 2.88
2017 152.42 44.80 189.40
2019 1.30 10.82 3.61 3.77 3.08
2018 156.60 45.20 189.40
2020 1.30 10.82 3.64 3.77 3.08
2019 161.17 45.24 189.40
2021 1.30 10.82 3.64 3.77 3.08
2020 161.17 45.20 189.40
2022 1.30 10.82 3.64 3.77 3.08
2021 161.17 45.72 189.40
2022 161.17 45.79 189.40

installed generation capacity considered in the generation expan-

sion plan of each country.
This information and the data bases of the electric systems In all the interconnection scenarios (in the simulations of the
involved were obtained from the official information that was sent operation of the electric systems), we consider 411 thermal and
in the context of the project reported in UNDP (2010), which is 146 hydro power plants. The thermal power plants are distributed
based on the generation and transmission expansion plans that as follows: 66 in Bolivia, 51 in Chile, 104 in Colombia, 119 in
were in force at each country in May of 2009. Ecuador and 71 in Peru. The hydro plants are distributed as: 27 in
For the different scenarios, the simulations of the operation of Bolivia (8 reservoirs), none in Chile, 38 in Colombia (16 reservoirs),
the electric systems were carried out taking into consideration the 35 in Ecuador (2 reservoirs) and 46 in Peru (7 reservoirs).
local prices of the generation fuels, which consider the price setting It is worth to mention that there are countries whose installed
mechanisms existing in every country. The reference fuel prices for capacity up to 2022 is between 2 and 2.5 times the 2010 capacity. On
the thermal power plants in each one of the countries considered the other hand, other countries increase the installed capacity only
here were standardized in similar units in order to be able to between 30% and 50% during the same period. These differences in the
compare the forecasted profiles. The fuel prices used in the increase of the installed capacity with respect to 2010 levels limit
simulations of the operation of the electric systems (in all the aspects such as the use of the interconnections towards the end of the
interconnection scenarios) are detailed in Tables 3–5. Although horizon, the export levels of electric energy towards countries with
this, and seeking an approximation to a situation with international smaller capacity growths, the fluctuations in both operational and
market prices, we also perform a sensitivity analysis for each marginal costs, and the levels of non-served energy.
scenario using the opportunity prices for natural gas, with an With respect to the local transmission systems, in all the
average price of $6/MBTU during the 2010–2022 period.3 interconnection scenarios (in all the simulations of the operation
The generation expansion plans of the countries (provided by of the electric systems), we consider 71 transmission lines in
the respective energy authorities) were not altered in the con- Bolivia, 111 lines in Chile, 151 lines in Colombia, 73 lines in Ecuador
sidered scenarios, excepting for extensions in such plans that do and 97 lines in Peru. For these transmission lines (with a voltage of
not represent the power systems until 2022. Such extensions were either 500 or 220–230 kV), the operation-and-maintenance cost is
proposed by taking into account the official information obtained considered 1.6% of the investment value for the 500 kV lines and
from each country. Table 6 shows a general comparison of the 3.0% of the investment value for the 220–230 kV lines.

3
For the estimation of the international prices of natural gas, we use the
hypothesis that, in both the mid- and the long-term, Peru and Bolivia will be natural- 3. Methodology
gas net exporters to Mexico and Brazil. Also Colombia and Chile will be natural-gas
net importers from Trinidad; Ecuador will be a natural-gas net importer from Peru.
Detailed information on the natural-gas prices used in this paper is available in The pre-feasibility analysis for the development of electric links
UNDP (2010). is, in general, an iterative process that requires considering not only
 E. Sauma et al. / Energy Policy 39 (2011) 936–949 939

Table 6
Installed generation capacity (MW) of each country in the period 2010–2022.

Type Bolivia SING-Chile Colombia Ecuador Peru

Installed generation capacity in 2010 1092 3980 13,440 3198 5952

Hydro capacity in 2010–2022 127 – 3820 3724 5796
Thermal capacity in 2010–2022 1130 1860 1159 1044 3176
Renewable capacity in 2010–2022 97 260 – – –

Total capacity in 2010–2022 1354 2120 4979 4768 8972

the particular situations of each country, but also those technol- equilibrium, transforming the investment decisions into an equili-
ogies that are feasible to be used in implementing the interconnec- brium problem subject to equilibrium constraints (EPEC), which is
tion projects. complex to deal with due to the non-convexities involved. Further
literature about the resulting EPEC when considering locational
3.1. Methodological framework market power is provided in Sauma and Oren (2006, 2007).
In the simulations of the electric-systems operation performed
In general methodological terms, first we perform a long-term with the Ose2000 model, we consider a horizon of 11 years, 3
analysis of the interconnections in order to establish a maximum different blocks of demand, and 20 different hydrologic series.
potential for complementarities among the different countries. In The security effects of the electric systems are reviewed through
particular, we compare the current state of the interconnections static and dynamic simulations in each one of the proposed
(where there are only international links between Colombia and scenarios. In every scenario, we use the PowerFactory electric-
Ecuador) with one in which there are interconnections among all technical analysis model of DigSilent (DigSilent, 2010), which
the considered countries through links with no capacity restric- allows evaluating both the transient and the permanent behaviors
tions and with different start-up dates (according to an expected of the interconnection links under different contingencies. The only
feasibility of implementation). This long-term analysis was carried purpose of this ‘‘security analysis’’ is the verification of the
out using the Stochastic Dual Dynamic Programming (SDDP) technical feasibility of the interconnections in the long run.
model, based on a one-node representation of the electric system Accordingly, the security analysis is only a long-term contingency
of each country. analysis, considering different hydro conditions (in order to
Secondly, the interconnections are designed considering links capture extreme dry condition that last for a long period), and
with sizes and start-up dates that fit the most realistic possibilities we do not perform short-term contingency analysis.
in order to be implemented within the horizon considered here. It is In the security analysis, we consider three types of failures:
important to remark here that the main idea of this article is unexpected disconnection of generators (4 scenarios), transmis-
showing that, even when keeping the security and self-sufficiency sion-line two-phase short circuit to ground without reconnection
of the power system of every country (i.e., when not altering the (3 scenarios), and transmission-line one-phase short circuit to
generation expansion plans of the countries), the interconnections ground with reconnection (2 scenarios). We use a single failure cost
have important economic benefits in the long run. Thus, we seek composed of four steps representing the depth of the failure with
designing links of sizes that allow important exchange of energy, respect to the demand in the corresponding node. In particular, for
but do not induce significant distortions in the self-sufficiency of failures of up to 5% of the demand in a node, the value of lost load
supply of every country. Accordingly, we do not optimize the sizes (cost of failure) is $500/MWh; for failures representing between 5%
of the transmission projects, but only consider economically- and 10% of the demand in a node, the failure cost is $750/MWh; for
feasible transmission lines sizes such that both there are not failures representing between 10% and 20% of the demand in a node,
significant transmission restrictions on the interconnections dur- the failure cost is $1250/MWh; and for failures representing more
ing most of the time and there are not significant distortions in the than 20% of the demand in a node, the failure cost is $1500/MWh.
security and self-sufficiency of supply of every country. In addition, Finally, using the results obtained from the simulations per-
we consider that the development of interconnections is carried formed with the Ose2000 model, we calculate the benefits asso-
out using technologies at efficient voltage levels. ciated to the different interconnection scenarios. Fig. 1 shows the
Then, we perform a mid-term analysis by simulating a joint and general layout used to analyze the different interconnection
coordinated operation of the main transmission systems of the scenarios.
different countries (which incorporate the proposed interconnec- As it can be observed from Fig. 1, the proposed methodology
tions). In doing this, we use the Ose2000 model—developed by Kas allows determining the effects of the interconnection scenarios on
Ingenieria S.A. (2007), which is similar to the SDDP model, break- different aspects: the supply and demand sides of each country, the
ing-down the demand in each system from both the temporary and system costs, the environmental issue, and the system security
the geographic point of view. The joint and coordinated operation concern. In particular, the supply-side effects of the interconnec-
of the electric systems assumes that both electricity demand is tion scenarios are computed through quantifying the operational
inelastic and generation companies bid their true marginal costs4 margin5 of each one of the generators participating in the operation
(i.e., electricity markets are perfectly competitive). This assump- of the systems, in each one of the scenarios analyzed. Similarly, it is
tion could be debatable in the case of the Colombian market, but possible to determine the economic effects of the interconnection
considering the exercise of locational market power implies, scenarios from the point of view of the electric-energy demand,
among other things, that the market equilibrium becomes a Nash through quantifying the value of withdrawing energy at every
demand node of the electric systems.
4
This approach required the exclusion from the analysis of some regulated
charges for generators, that are applied in the calculation of the electricity prices in
5
some countries, as in the case of Colombia where fixed national reliability revenues The operational margin of a generator corresponds to the difference between
assigned to generators are obtained through an energy-reliability charge collected the revenues due to the sale of electricity (valuated at the respective injection nodes)
from the generator’s electricity sales. and the operation costs of each generation power plant.
940 E. Sauma et al. / Energy Policy 39 (2011) 936–949

Fig. 1. General layout of the technical-economic analysis of the interconnection scenarios.

The proposed methodology also allows quantifying the eco- simulations of the system operation must be a multi-nodal and
nomic effects of the interconnection scenarios over the system multi-reservoir simulation, incorporating the hydrologic uncertainty
costs. In doing this, we use a cost function that considers the factor.
operation-and-failure costs in each power system, along with the The availability of limited amounts of hydro energy in the form
international-link transmission costs and the costs of the local of water kept in reservoirs turns the optimal operational problem
transmission-system reinforcements necessary for supporting the into a complex one, requiring the joint evaluation of the decisions
proposed interconnections. in the present stage and the decisions to be adopted in future
In addition, the proposed methodology allows determining the stages. Furthermore, in the absence of perfect hydrological influx
global environmental effect of the interconnection scenarios predictions, the problem is essentially stochastic.
through the valuation of the CO2 emissions avoided on each From the mathematical viewpoint, this problem can be for-
interconnection scenario as compared to the base scenario. mulated as a recursion of a stochastic dynamic program (SDP), as in
It is worth to remark that, in the planning process, any change in Pereira and Pinto (1991)
conditions, assumptions or restrictions that condition the starting
at ðXt Þ ¼ E fMinUtCt ðUt Þ þ bUat þ 1 ðXt þ 1 Þg ð1Þ
point, or in the types of technologies used and their technical- At 9Xt
economic characteristics, will result on obtaining a different
interconnection alternative. Accordingly, the results presented in subject to
this paper are valid within the framework of the assumptions, Xt þ 1 ¼ Ft ðXt ,At ,Ut Þ ð2Þ
considerations and restrictions described here. Thus, a greater level
of demand, a different relation of the fuel prices used by the thermal Rt þ 1 ðXt þ 1 Þ Z0 ð3Þ
power plants, or different investment values for the projects
studied would lead the electric systems to another equilibrium
St ðUt Þ Z0 ð4Þ
point, altering the results (and even the conclusions in the case that
such changes are significant).
where t (t¼ 1, y, T) indexes the stages, Xt is the state vector at
the start of stage t, at(Xt) is the expected value of the operational
3.2. Analytical model cost in state Xt, At9Xt is the probability distribution of inflow vector
At as conditioned by state Xt, Ut is the decision vector for stage t, Ct is
The analyses presented in this article are based on simulations the cost associated to decision Ut, and b is the discount factor.
performed using either the SDDP model or the Ose2000 model. In this problem, Eq. (2) corresponds to the state transition Eq. (3)
Next, we describe the analytical framework of these models. represents the constraints on the state vector and Eq. (4) represents
Facing the transmission planning problem requires evaluating the the decision vector constraints. The state of the system, Xt,
trade-off between the operational costs and the power transmission represents all the information needed, which may affect the future
investment requirements. In hydro-intensive power grids, the operational costs.
 E. Sauma et al. / Energy Policy 39 (2011) 936–949 941

Table 7
Characteristics of the proposed interconnections in the Andes region.

Link Interconnection point Long (km) Main characteristics Start-up date Investment cost
(thousands of dollars)

Colombia– San Marcos-Jamondino 500 kV (Colombia)–Pifo 500 kV (Ecuador) 551 1500 MW–500 KV, AC60Hz April-14 210,942
Ecuador
Ecuador–Peru Yaguachi 500 kV (Ecuador)–Trujillo 500 kV (Peru) 638 1000 MW–500 kV, AC60Hz January-15 174,427
Peru–Chile Montalvo 500 kV (Peru)–Crucero 500 kV (Chile) 645 1500MW–500 kV, HVDC January-16 401,646
Bolivia–Chile Chuquicamata 220 kV (Chile)–Chilcobija-Tarija 230 kV (Bolivia) 489 340 MW–230 kV, AC 50 Hz January-17 163,735

The direct operational costs associated to the decision vector, Ut,

are represented by the thermal-generation costs required for
San Marcos
supplementing the energy supply (Gorenstin et al., 1991)
Jamondino Jamondino
XX
Ct ðUt Þ ¼ Min Cj ðGtj Þ ð5Þ
k A K j A Tk Pifo
subject to Pomasqui
X X X
ri UQti þ Gtj þ Ftkl ¼ Dtk , 8k A K ð6Þ
i A Hk j A Tk l A Ok

Ftkl ¼ gkl ðytk ytl Þ, 8k A K, l A Ok ð7Þ Yaguachi

Gt rG ð8Þ
Trujillo
9Ft 9 rF ð9Þ

where k indexes the system buses (K is the set of buses), j indexes

the thermal plants (Tk is the set of thermal plants in bus k), i indexes the
hydro-plants (Hk is the set of hydro-plants in bus k), Cj is the generation
Montalvo
cost function of plant j, Gtj is the generation of thermal plant j in stage t,
ri is the production coefficient of plant i, Qti is the turbine outflow Montalvo (CC)
Tarija
volume at the hydraulic plant i in stage t, Ftkl is the active power flow in Chilcobija
circuit k–l in stage t, Dtk is the active load in bus k in stage t, Ok is the set
of buses directly connected to bus k, ytk is the voltage angle at bus k in Crucero (CC) Chuquicamata
Crucero
stage t, gkl is the susceptance of circuit k–l, G is the vector of thermal Crucero
generation capabilities, and F is the vector of circuit flow limits.
Solving the system-operation problem by means of the SDP calls CHILE
for the enumeration of all possible combinations of both initial
storage and inflows. Therefore, the computer simulation approach Fig. 2. Proposed interconnections in the Andes region.
grows exponentially in line with the number of reservoirs (Pereira
and Pinto, 1991). Therefore, it becomes necessary to adopt a
methodology that does not require state enumeration, thus alle-
viating the computer simulation requirements of the SDP recur- Table 8
Energy transferences through the interconnections during the 2014–2022
sion. In this context, the focus of the Stochastic Dual Dynamic
period (GWh).
Programming (SDDP) model and the Ose2000 model comes up.
Basically, both the SDDP model and the Ose2000 model are From To Base 1 2 3
modifications of the SDP formulation, where the future cost
function is represented as a piecewise linear function, thus, Colombia Ecuador 14,567 23,435 29,458 29,564
Ecuador Colombia 10,505 15,031 11,211 11,314
avoiding the need for an enumeration of all the states. Peru Ecuador 13,882 4138 4237
Ecuador Peru 12,896 23,321 22,677
Peru Chile-SING 81,663 80,925
4. Results Chile-SING Peru - -
Chile-SING Bolivia -
Bolivia Chile-SING 15,503
4.1. Interconnection projects proposed

As explained in Section 3, the method for determining the

interconnection scenarios starts from a long-term analysis of the
long-term analysis, we design interconnections with a high degree of
interconnections (where we compare the base scenario with a
feasibility in terms of their size, location, and start-up date. In this
scenario in which there are interconnections among all the considered
context, we consider earlier the expansion of the current intercon-
countries through links with no capacity restrictions, modeling each
nections (i.e. the Colombia-Ecuador and Peru-Ecuador links) and later
country as a one-node representation).6 Based on the results of this
other expansion projects whose construction time periods may be
more uncertain due to aspects such as future negotiations among the
6
A detailed description of this long-term analysis is available in UNDP (2010). countries involved and environmental-dealing issues.
942 E. Sauma et al. / Energy Policy 39 (2011) 936–949

Fig. 3. Annual average marginal costs of the systems under base scenario. Fig. 5. Annual average marginal costs of the systems under scenario 2.

Fig. 4. Annual average marginal costs of the systems under scenario 1. Fig. 6. Annual average marginal costs of the systems under scenario 3.

The criteria adopted for selecting the route of the international the interconnection between Peru and Chile-SING. Similarly, the
links are the following: transfers between these two countries decrease in the opposite
direction because the Peruvian energy generation is mainly aimed
 Choosing polygonal sections with the least longitude and the to supply loads in the north of Chile under this scenario. Something
least number of vertices as possible. similar occurs with the generation production of Bolivia in Scenario 3.
 Choosing links close to existing transport routes. It is worth to recall that the design of the interconnections
 Avoiding zones with external geodynamic risk. considered here is based on a capacity sizing that takes into account
 Avoiding archeological zones and national parks. the feasibility of implementation within the time horizon, applying
 Avoiding, when feasible, links that are either parallel to com- conservative criteria regarding the matter. Analyzing ex-post the
munication lines or crossing other power transmission lines. use level of the interconnections obtained in the scenarios studied,
 Avoiding, when feasible, proximity to villages. we verify that the interconnections operate in an expected manner
Accordingly, we consider transmission investments associated at full capacity during most hours in the study period, thus
to feasible interconnections in the Andes region, which produce indicating the existence of a greater energy exchange potential
different benefits for each country in consistency with their than the one established here.
complementarities. The links proposed in the different scenarios
(under an interconnected operation of the electric systems of the
4.3. Effect of the interconnections on the systems marginal costs
five countries) are described in Table 7 and schematically shown in
Fig. 2.
One of the most important effects of the energy transferences
between countries is the effect over the local systems marginal
4.2. Energy exchange levels through the proposed interconnections costs. Next, we describe the impact of the energy exchange on the
‘‘annual average marginal costs of each country’’ according to each
Using the proposed interconnections and the operation condi- interconnection scenario. The ‘‘annual average marginal costs of
tions in each scenario, we perform the mid-term analysis using the each country’’ refers to the weighted average of the generation
Ose2000 model. From the simulation results, it is possible to marginal costs of power plants (weighted by the corresponding
compute the expected use of every link. The results obtained show energy amount annually generated) in each country.
that there are important energy blocks feasible to be transferred
from one electric system to the other. This is due to multiple
reasons, among which it is the possibility of taking advantage of 4.3.1. Base scenario
arbitrage (e.g., price differences), thus generating business oppor- This scenario considers the costs of the fuels provided by the
tunities for the electricity-market agents in each country. energy authorities of each country (UNDP, 2010), which, in the
Table 8 shows a summary of the expected energy transferences particular case of natural gas, refers to prices regulated locally by
through the interconnections during the 2014–2022 period. each country, excepting for Chile that corresponds to the price of the
Going from Scenarios 1 to 2, we observe an important increase in liquefied natural gas placed in the Mejillones re-gasification terminal.
the energy transfers from Ecuador to Peru due to the incorporation of This explains why the profile of the system marginal costs for Chile
 E. Sauma et al. / Energy Policy 39 (2011) 936–949 943

Table 9
Present value of the electric-sector operational margin during the 2014–2022 period
(millions of dollars discounted to April 2014).

Scenario Bolivia Chile Colombia Ecuador Peru Total Difference

with respect
to the base
scenario

Base 372 3139 11,098 5421 6903 26,933 0

1 372 3139 11,546 4903 6789 26,748  185
2 372 1498 12,623 5490 9488 29,470 2537
3 1107 1366 12,548 5463 9367 29,852 2919

Fig. 7. Annual average marginal costs of the systems under natural gas sensitivity Table 10
scenario. Present value of the electric-sector operational margin during the 2014–2022
period, with Natural-gas prices sensitivity (millions of dollars discounted to
April 2014).
turns out to be much greater than those of the other countries (see
Fig. 3). This price condition remains in scenarios 1, 2 and 3. Scenario Bolivia Chile Colombia Ecuador Peru Total Difference
with respect
to the base
4.3.2. Scenario 1 scenario
As we aggregate demand, the most expensive resources are
Base NGs 1884 1854 12,685 6380 9732 32,535 0
replaced by cheaper ones coming from countries with cheaper
1 NGs 1884 1854 14,519 6257 9117 33,631 1096
resources. Thus, some countries, when becoming electricity net 2 NGs 1884 1237 15,360 6766 11,290 36,537 4002
exporters, start using more expensive resources to meet the new 3 NGs 1860 1255 15,446 6853 11,533 36,948 4413
demand. This fact could produce an increment in their system
marginal costs in the short term. In Section 5, we propose a
commercial alternative to alleviate these marginal cost increments Scenario 2. This indicates an important energy transfer from Bolivia to
in electricity net exporter countries. Chile-SING during this period.
In this scenario, due to the expansion of the interconnections The marginal costs in Chile-SING tend to have values slightly
between Colombia and Ecuador (in 2014) and an interconnection lower than $60/MWh, in almost all the rest of the horizon, where
between Peru and Ecuador (in 2015), the marginal costs of these the final increase in marginal cost is lower than that in Scenario 2
three countries tend to be similar in the long term (fluctuating from due to the energy imports from Bolivia.
$30/MWh to $31/MWh). However, the net effect is an increase in
the system marginal costs of Colombia and Peru in reference 4.3.5. Natural gas sensitivity scenario
to the base scenario due to the energy exported towards Ecuador This scenario corresponds to a sensitivity analysis of the
(see Fig. 4). proposed interconnection scenarios, which considers the oppor-
tunity costs for the natural gas prices in the different countries
(from both the export and the import viewpoint).
4.3.3. Scenario 2
The natural-gas prices sensitivity analysis allows isolating, up to a
In this scenario, the interconnection between Peru and Chile is
certain extent, the ‘‘price’’ variable in reference to the natural gas prices,
added. This produces an important increase in the marginal costs
establishing a ‘‘floor’’ for the electricity exchanges. In this case, the
during the first years in Colombia, Ecuador and Peru, in reference to
private decisions about energy exchanges depend only on the com-
the prior scenarios, due to the energy exported towards Chile
plementarities among the electric systems (they do not depend on
(see Fig. 5). On reaching the end of the time horizon, the marginal
arbitrages or price structure fluctuations among countries).
costs in these countries also tend to decrease, approaching
Fig. 7 shows the annual average marginal costs with the afore-
values of around $30/MWh, similar to the stabilization levels
mentioned sensitivity analysis over the natural gas prices, when
provided in Scenario 1. This occurs because, at the end of the
considering the interconnections included in Scenario 3. Under these
time horizon, there is a stronger relation between the installed
conditions, the marginal costs of all countries, particularly in Bolivia,
capacity and demand, including the one coming from the
increase in comparison to the original Scenario 3. A relative matching is
interconnections.
produced in the marginal costs associated to Colombia, Ecuador and
On the contrary, the marginal costs in Chile-SING significantly
Peru in the long-term (at levels between $40/MWh and $45/MWh),
decrease, reaching values close to $60/MWh. This level of the
while marginal costs in Bolivia and Chile-SING tend to be matched (at
marginal cost in Chile-SING remains in almost all the rest of the
levels around $65/MWh). This would suggest the existence of more
horizon, with a small increase at the end of the horizon, which
equivalent energy exchanges between these last two countries within
would indicate that the links are operating at their maximum
the planning horizon considered, with possible bi-directional energy
capacity and it is not possible to inject more low-cost energy to this
flows. This bi-directionality during some periods of the year may open
system.
up the possibility of establishing energy-exchange mechanisms based
on a more balanced equilibrium of exported and imported energy sales
4.3.4. Scenario 3 prices.
In this scenario, the interconnection between the Bolivian electric
system and the SING in Chile is added. As observed in Fig. 6, starting 4.4. Economic benefit analysis
from the entrance into operation of this link, the marginal costs in
Bolivia tend to approach those of Colombia, Ecuador and Peru in the The economic benefit analysis is carried out from a whole-
long-term (reaching values around $34/MWh), while, in the afore- horizon behavior perspective of electricity markets. We quantify
mentioned countries, the marginal costs remains in similar levels as in the economic benefits and/or costs on both the electricity supply
944 E. Sauma et al. / Energy Policy 39 (2011) 936–949

Table 11 Table 13
Present value of the energy withdrawals, valuated at the monthly average marginal Present value of the operations-and-failure costs (millions of dollars discounted to
cost of the system (millions of dollars discounted to April 2014). April 2014).

Scenario Bolivia Chile Colombia Ecuador Peru Total Difference Scenario Bolivia Chile Colombia Ecuador Peru Total Difference
with with respect
respect to to the base
the base scenario
scenario
Base 492 7482 1061 910 2873 12,819 0
Base 909 11,182 12,575 6840 10,257 41,763 0 1 492 7482 1043 670 2957 12,643  175
1 909 11,182 13,032 6076 10,067 41,266  497 2 492 4087 1261 951 4004 10,793  2025
2 909 8803 14,113 6686 12,552 43,063 1300 3 660 3579 1245 929 4006 10,420  2399
3 1541 8483 14,031 6660 12,456 43,171 1408

Table 14
Present value of the operations-and-failure costs, with Natural-gas prices sensitivity
Table 12 (millions of dollars discounted to April 2014).
Present value of the energy withdrawals, valuated at the monthly average marginal
cost of the system, with natural-gas prices sensitivity (millions of dollars discounted Scenario Bolivia Chile Colombia Ecuador Peru Total Difference
to April 2014). with respect
to the base
Scenario Bolivia Chile Colombia Ecuador Peru Total Difference scenario
with respect
to the base Base NGs 1668 7188 1221 1051 4416 15,544 0
scenario 1 NGs 1668 7188 1334 910 4076 15,176  368
2 NGs 1668 4323 1532 1099 5578 14,200  1343
Base NGs 3741 9519 14,313 8082 14,844 50,499 0 3 NGs 1622 4373 1643 1097 5564 14,299  1244
1 NGs 3741 9519 16,098 7683 14,122 51,162 663
2 NGs 3741 8512 16,980 8216 16,183 53,631 3132
3 NGs 3708 8548 17,194 8297 16,451 54,198 3699
Ecuador significantly increase, while the operational margin in
Chile-SING decreases to less than half (due to the energy imports
coming from Colombia, Peru and Ecuador). Something similar
occurs with Bolivia, on incorporating the interconnection between
and the electricity demand sides. We also quantify the total cost
this country and Chile-SING in Scenario 3, keeping the other-
savings achieved in every electric system due to the proposed
countries’ margins in similar magnitudes as those of Scenario 2.
interconnections and the investment costs of both international
It is important to remark that the private benefits obtained by
links and those modifications made to the local transmission
the generation sector of a specific country (due to exporting or
systems (and needed to support the international energy
importing electricity in some scenario) are computed assuming
exchange). In addition, we quantify the environmental benefits
there are no commercial barriers or restrictions limiting the energy
of the proposed interconnections in terms of the avoided emissions
exchanges. In this sense, our results should be understood as
of CO2.
referential values, which could vary in practice according to the
commercial mechanisms being finally implemented for the regio-
nal electric integration.
4.4.1. Supply-side economic benefits
When considering the natural-gas prices sensitivity analysis, we
The private economic benefits for the generation sector are
obtain the electric-sector operational margins shown in Table 10.7
analyzed considering the sector as a single market agent. Such
The results of the natural-gas prices sensitivity analysis show
benefits are calculated based on the ‘‘electric-sector operational
that, even under these highly restrictive conditions, there are
margin’’, which is defined as the difference between the valuation
significant complementarities in the resources use. As previously
(at the monthly marginal cost) of the energy injections coming
mentioned, they also show that the proposed interconnections
from all system’s power plants and the operation costs of all the
have a high capacity use level, with degrees of bi-directionality
thermoelectric power plants. Table 9 shows the present value of the
higher than in the original scenarios.
electric-sector operational margin for every electric system during
the time horizon, in each interconnection scenario. Throughout the
entire article, the economic benefits (and costs) are valuated in 4.4.2. Demand-side economic benefits
mid-2009 dollars and actualized to April of 2014 using a 10% The effect of the interconnections on the demand side is
discount rate. analyzed considering that there is a single buyer in every electric
The main conclusion, from the supply viewpoint, is that the system who consumes electric energy. The demand-side economic
sellers of energy can obtain larger margins as more countries are benefits are calculated, for every scenario, by valuating the energy
electrically integrated. However, these margins are substantially withdrawals of the single buyer, considering he/she pays the
reduced in the case of Chile due to its high cost of supply. monthly average marginal cost of the system (assuming that this
One remarkable result of this work is that there are only private monthly average marginal cost should be the electricity price in the
economic benefits when the interconnections with Chile are taken long run in order to have generators covering their average
into consideration (i.e., scenarios 2 and 3). From Table 9, we observe production costs and their non-covered investment costs). We
that Scenario 1 produce a margin for the whole set of countries that do not consider the effects of the congestion rents on these benefits.
is lower than the one obtained in the base scenario, even though the Table 11 shows the present value of the valuated energy
Colombian generation sector gets higher margin in this scenario withdrawals, considering the single buyer pays the monthly
than in the base scenario due to its exports to Ecuador (which
implies that the Ecuadorian generation sector decreases its mar- 7
The scenarios labeled ‘‘base NGs’’, ‘‘1 NGs’’, ‘‘2 NGs’’ and ‘‘3 NGs’’ correspond to
gin). On incorporating the interconnection between Peru and Chile the equivalent scenarios base, 1, 2 and 3, but incorporating the natural gas prices
in Scenario 2, the operational margins in Colombia, Peru and sensitivity (i.e., using the opportunity costs of natural gas as the natural gas prices).
 E. Sauma et al. / Energy Policy 39 (2011) 936–949 945

Table 15 Table 17
Present value of the total cost function (millions of dollars discounted to April 2014). Per-technology emission factors used to estimate CO2 emissions.

Scenario ITX + O&MCTX O&FC ResidualTX PVTC Difference with Type of power plant Emission factor for CO2 Unit
respect to the base
scenario Coal 2.64 Ton CO2/ton
Natural gas 0.06 Ton CO2/millions of Btu
Base 0 12,819 0 12,819 0 Diesel 3.18 Ton CO2/ton
1 240 12,643 170 12,713  105
2 405 10,793 328 10,870  1948
3 459 10,420 387 10,492  2327
with a high penetration of natural gas power plants) decrease when
considering the opportunity costs of the natural gas.
On the other hand, we calculate the present value of the
Table 16 transmission investments, as well as the costs of the operation and
Present value of the total cost function, with Natural-Gas prices sensitivity (millions maintenance of the international links and the investments needed in
of dollars discounted to April 2014).
the local networks to support the international energy exchanges. All
Scenario ITX + O&MCTX O&FC ResidualTX PVTC Difference with this information can be used for calculating a total cost function,
respect to the base which allows minimizing the total expenditures in the systems in the
scenario same way a centralized transmission investment planner will do it.
Such total cost function is quantified by means of the present value of
Base NGs 0 15,544 0 15,544 0
1 NGs 240 15,176 170 15,245  298 the total costs (PVTC), which corresponds to the sum of the present
2 NGs 405 14,200 328 14,277  1266 values of the operations-and-failure system costs (O&FC) and the
3 NGs 459 14,299 387 14,371  1172 investments, operations and maintenance costs of both the interna-
tional links and the necessary extensions carried out in the local
transmission systems (ITX +O&MCTX), minus the present value of the
average marginal cost of the system (The reader should note that residual value of the transmission investments (ResidualTX).8 We
these are costs (instead of benefits) for the buyers). consider a service life of 30 years for the transmission lines, which
From Table 11, it becomes evident that, from a demand-side explains why we consider a residual value of the transmission
viewpoint, Scenario 1 is the most favorable scenario and Scenario 3 investments when doing the economic evaluation in the horizon
is the most adverse situation to buyers. 2014–2022. As mentioned before, we assume 10% as the discount
Table 12 shows the present value of the valuated energy with- rate. Table 15 shows the present value of the total cost function and its
drawals for every electric system during the time horizon, when components in each scenario.
considering interconnections scenarios that use the opportunity costs The difference between the total cost function in each scenario
of the natural gas. From Table 12, we conclude that considering the and the total cost function in the base scenario corresponds to the
opportunity costs of the natural gas leads, in general, to higher costs system-cost savings due to the proposed interconnections in each
for buyers, which is a direct consequence of the higher production one of such scenarios. These values are shown in the last column of
costs faced by the natural gas power plants and the fact that we are Table 15. From Table 15, we conclude that the most favorable
considering a pricing scheme based on the marginal costs. scenario, from the viewpoint of a centralized transmission plan-
ning, is Scenario 3 (with a net benefit of 2,327 millions of dollars in
reference to the base scenario). On the other hand, Scenario 1 is the
4.4.3. System-cost savings
interconnection scenario that shows the least net benefit from this
The system-cost savings due to the proposed interconnections
point of view.
are quantified by comparing the savings in the system operation
Table 16 shows the present value of the total cost function and
costs due to the interconnections and the investment and main-
its components, when considering the opportunity costs of the
tenance costs associated to such links. The transmission invest-
natural gas.
ment costs involve both the costs associated to the construction of
From Table 16, we observe that the scenarios having opportu-
the international links and the costs associated to the modifications
nity prices for natural gas show a decrease in the net benefit (from
made to the local transmission systems in order to support the
the viewpoint of a centralized transmission planning) in compar-
international energy exchange.
ison to the prior scenarios, with the only exception of Scenario 1.
Tables 13 and 14 show the present value of the operations-and-
This is a result of the impact of the increase in the prices of natural
failure costs, in the analyzed scenarios, for the cases of without
gas over the system costs for those electric systems having an
(Table 13) and with (Table 14) sensitivity analysis for the natural
important participation of such fuel in the generation power plants.
gas prices. The operations-and-failure costs include the variable
production costs (in terms of both fuels and non-fuels costs)
incurred by the different power plants and the failure cost, which
4.4.4. Environmental benefits
is calculated by valuating the expected non-supplied energy. Thus,
The electric integration allows decreasing the overall emissions
the difference between the total operations-and-failure costs in
of both local and global pollutants. Although the environmental
each scenario and such costs in the base scenario corresponds to the
benefits from local-pollutant emission reductions may be very
operations-and-failure cost savings due to the proposed intercon-
significant, in this paper we only quantify the avoided CO2
nections in each one of such scenarios.
emissions given the pre-feasibility nature of this work. Even though
From Tables 13 and 14, it is observed that the more interconnec-
the CO2 emissions were estimated in a very conservative and
tions exist, the lower the operations-and-failure costs. The reason of
approximate manner, we found significant emission reductions
this is the replacement of expensive generation by cheaper energy
due to the interconnections, which grows as the interconnection
available in other country with lower production costs, which is now
scenarios become more integrating. There is no doubt that this is an
available due to the interconnections. In the case of the natural-gas
element that must be incorporated in future negotiations for the
prices sensitivity analysis, the benefits in terms of costs savings
decrease, as it is evident in Table 14. This is because the differences
among the costs of each system (and, particularly, among systems 8
PVTC ¼ ITX þ O&MCTX þ O&FC-ResidualTX .
946 E. Sauma et al. / Energy Policy 39 (2011) 936–949

Table 18
CO2 emissions in each interconnection scenario during the 2014–2022 period (thousands of tons of CO2).

Scenario Bolivia Chile Colombia Ecuador Peru Total Difference with respect
to the base scenario

Base 40,154 190,996 61,910 21,370 75,001 389,431 0

1 40,154 190,996 64,756 16,599 77,356 389,860 429
2 40,154 113,541 71,627 23,255 107,498 356,075  33,356
3 53,572 96,307 70,974 22,773 107,575 351,202  38,229

Table 19
CO2 emissions in each interconnection scenario, with Natural-gas prices sensitivity, during the 2014–2022 period (thousands of tons of CO2).

Scenario Bolivia Chile Colombia Ecuador Peru Total Difference with respect
to the base scenario

Base NGs 41,446 180,269 64,250 22,080 74,901 382,946 0

1 NGs 41,446 180,269 74,458 21,017 69,003 386,193 3247
2 NGs 41,446 117,969 78,575 24,848 97,528 360,365  22,580
3 NGs 40,489 119,711 78,094 24,520 97,330 360,145  22,800

Table 20 Table 22
Present value of the valuation of the avoided CO2 emissions in each interconnection Present value of the total economic benefits in each interconnection scenario during
scenario during the 2014–2022 period (millions of dollars discounted to April 2014). the 2014–2022 period (millions of dollars discounted to April 2014).

Scenario Bolivia Chile Colombia Ecuador Peru Total Item Scenario 1 Scenario 2 Scenario 3

Base 0 0 0 0 0 0 DOM  185 2537 2919

1 0 0  40 61  31  10 DDB 497  1300  1408
2 0 885  128  25  377 355 DCR  137 788 888
3  148 1073  118  18  378 411 DEB  10 355 411
DTC  70  77  72
Total benefit 95 2303 2738

Table 21
Present value of the valuation of the avoided CO2 emissions in each interconnection
scenario, with Natural-gas prices sensitivity, during the 2014–2022 period (millions Table 23
of dollars discounted to April 2014). Present value of the total economic benefits in each interconnection scenario, with
Natural-Gas prices sensitivity, during the 2014–2022 period (millions of dollars
Scenario Bolivia Chile Colombia Ecuador Peru Total discounted to April 2014).

Base 0 0 0 0 0 0 Item Scenario 1 Scenario 2 Scenario 3

1 0 0  132 13 77  43
2 0 699  182  34  250 234 DOM 1096 4002 4413
3 10 682  171  29  246 246 DDB  663  3132  3699
DCR  65 473 530
DEB  43 234 246
DTC  70  77  72
Total benefit 255 1500 1418
exchange of energy (taking into consideration the positive or
negative impacts on reducing the greenhouse gas emissions that
can be produced by a determined electric interconnection). transactions would allow an overall decrease in the total CO2
In order to estimate CO2 emissions in each scenario, we use per- emissions of the Andes region.
technology nominal emission factors for the conventional thermal Table 19 shows the total CO2 emissions in each interconnection
power plants operating in every system. These factors were scenario during the planning horizon, when considering the
obtained from the 2007 Intergovernmental Panel on Climate opportunity prices of the natural gas.
Change (IPCC) report (IPCC, 2007). For those power plants that From Table 19, we observe that, in the case of the sensitivity
operate with a mixture of fossil fuels, we consider the emission analysis for the natural gas prices, Scenario 3 still represents the
factor of the main fuel. Table 17 shows the emission factors used. most favorable situation in reference to avoiding greenhouse gas
Table 18 shows the total CO2 emissions in each interconnection emissions. However, in this case, the CO2 emissions in such
scenario during the planning horizon considered. scenario are very close to those of Scenario 2 (i.e., going from
The most favorable situation (in reference to avoiding green- Scenario 1 to Scenario 2 is where the largest decrease in CO2
house gas emissions) is Scenario 3 due to the fact that, in such emissions is produced).
scenario, an important amount of electric generation coming from For the valuation of the CO2 emissions, we use the price of the
coal and diesel power plants located in the SING-Chile is replaced carbon bonds as a proxy for the value of reducing CO2 emissions.
(with respect to the base scenario) by cleaner generation imported We use the forecast of the carbon bond price made in October of
from the rest of the countries. This is, although the sale of relatively 2009 by the ‘‘European Climate Exchange’’ for December 2012
cheap energy to the SING-Chile market would produce – in (European Market Exchange, 2009), which is equal to 13.32 h/ton
expected terms – an increase in the CO2 emissions in the other (i.e., $19.9/ton approx.).
countries (situation that should be considered at the time of Tables 20 and 21 show the present value of the valuation of the
allocating costs under an interconnection scenario), such energy avoided CO2 emissions due to the corresponding interconnection
 E. Sauma et al. / Energy Policy 39 (2011) 936–949 947

Fig. 8. Present value of the total economic benefit of interconnections in scenarios 1, 2 and 3 with respect to the base scenario.

projects, for the cases of without (Table 20) and with (Table 21) transmission cost in scenario i with respect to the base scenario,
sensitivity analysis for the natural gas prices. Cases showing an and DResidualTX-i is the change in the residual value of the
increase of the CO2 emissions with respect to the base scenario transmission investments in scenario i with respect to the base
have a negative present value of the valuation of the avoided CO2. scenario.
From Tables 20 and 21, we observe that, for both cases (with and Tables 22 and 23 show the total economic benefits for each
without natural gas prices sensitivity analysis), the main benefits interconnection scenario, for the cases of without (Table 22) and
for avoiding CO2 emissions are derived from the emissions reduc- with (Table 23) sensitivity analysis for the natural gas prices.
tions occurring in Chile. The results obtained show that in Scenarios 2 and 3 (that
consider the interconnection with Chile) the environmental ben-
4.4.5. Total economic benefits efits contribute approximately in 15% of the total benefits in case of
The total economic benefits correspond to the sum of all using national gas prices and in 17% in the case of using gas
previously mentioned economic benefits. That is, the supply-side opportunity prices, as it is evident in Fig. 8.
benefits, the demand-side benefits, the benefits derived from
congestion rents, and the environmental benefits, discounting
the investment costs associated to the transmission expansions. 5. Commercial and regulatory framework
The total benefits of the interconnections correspond to the
differences in benefits between the base scenario and the Inter- The power industry requires high-level investments and fre-
connection Scenarios 1, 2 and 3. Mathematically:9 quent maintenance, which entails the need for certainty from
suppliers in covering their costs in the long run. In order to
PVTEBi ¼ DOMi þ DDBi þ DCRi þ DEBi DTCi guarantee this certainty, it is required a set of rules that makes
DOMi DDBi DCRi
zfflfflfflfflfflfflfflfflfflfflffl}|fflfflfflfflfflfflfflfflfflfflffl{ z}|{ zfflfflfflfflfflffl}|fflfflfflfflfflffl{ market agents confident.
¼ DSi DCOPEFi  DBi þ DBi DSi þ DEBi Interconnections must be developed facing power markets with
DTCi different economic characteristics and regulatory frameworks with
zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{
ðDITX-i þ DO&MCTX-i -DResidualTX-i Þ different levels of governmental involvement. Consequently, it is
¼ DEBi DCOPEFi DITX-i DO&MCTX-i þ DResidualTX-i ð10Þ crucial the development of a set of rules that clearly defines: the
electricity price in the frontiers, the dispatch of resources among
countries and among market agents, the distribution of congestion
where PVTEBi is the present value of the total economic benefit in rents (if any), the payments for using the interconnections, and the
scenario i with respect to the base scenario, DOMi is the change in way of solving conflicts, among others. These rules must arise from
the operational margin of power suppliers in scenario i with respect institutions with the technical strength and impartiality needed to
to the base scenario, DDBi is the change in the demand-side benefits build credibility for investors, consumers, and the governments in
in scenario i with respect to the base scenario, DCRi is the change in the long run.
the congestion rents in scenario i with respect to the base scenario, These rules must define the regulatory, commercial, and operative
DEBi is the change in the environmental benefits in scenario i with structures characterizing the international energy exchanges through
respect to the base scenario, DTCi is the change in the transmission a process based on three general principles: (i) graduality of both the
expansion costs in scenario i with respect to the base scenario, DSi is development and expansion of international links and the implemen-
the change in the power sales in scenario i with respect to the base tation of the regulatory and commercial processes associated with the
scenario, DCOPEFi is the change in the operations-and-failure cost energy exchanges, (ii) reciprocity among governments to aim to the
in scenario i with respect to the base scenario, DBi is the change in harmonization of market rules of the countries, and (iii) market
the valuated power purchased by consumers in scenario i with competitivity through objective, transparent, and non-discriminatory
respect to the base scenario, DITX-i is the change in the transmission market rules (EPR, 2000).
investments in scenario i with respect to the base scenario, There are mainly four barriers and asymmetries that must be
DO&MCTX-i is the change in the operations-and-maintenance addressed to go towards the regional energy integration. First, it
is needed a higher institutionalization for: (i) harmonizing the
9
Recall that we consider demand is perfectly inelastic and that generation firms regulatory frameworks, (ii) international operations coordination,
bid their true marginal costs as in a perfectly competitive market. and (iii) coordination in the planning of the national interconnected
948 E. Sauma et al. / Energy Policy 39 (2011) 936–949

systems in agreement with Decision N. 536 and the national plans. mechanisms that modify consumers bills when some interconnec-
Second, the incentives scheme to invest in international links must be tions’ benefits are attained.
revised. Regional electricity integration requires a system of remu- Finally, it is important to remark that the numerical
neration (through transmission tariffs or the transfer of transmission results presented in this paper are valid within the framework of
rights), within a concept of the ‘‘regional grid’’, to incentivize these the assumptions, considerations and restrictions described
investments. Third, it is needed to solve some asymmetries for here. Thus, a greater level of demand, a different relation of the
opportunity transactions such as variations of opportunity power fuel prices used by the thermal power plants, or different invest-
prices among importer and exporter countries and the distribution of ment values for the projects studied would lead the electric
the congestion rents generated by the opportunity transactions. systems to another equilibrium point, altering the numerical
Fourth, the regional integration requires establishing some type of results.
long-term contracts, which allows long-term transactions among
market agents, incentivizes investors of interconnections, and guar-
antees the local supply security. Acknowledgment
With respect to the commercial design of energy exchanges, it is
needed some mechanisms that incentivize market agents to invest The authors gratefully acknowledge the contributions of
in international links in the long run. For this purpose, we suggest Carlos Piña, Raul O’Ryan, Carlos Ferruz, Juan Liu, Andy Garcia, Pia
the simultaneous implementation of short-term and long-term Bravo, and Manuel Maiguashca for their valuable comments.
contracts. Short term, flexible contracts are needed to take advan- The authors also thanks the comments received during the four
tage of the opportunity transactions emerging from variations in meetings organized by the United Nations Development
the short-term economic conditions of the different energy mar- Program (UNDP) in the context of the UNDP project ‘‘Estudio para
kets while long-term, firm contracts are required to give market- Análisis de Prefactibilidad Técnico Económica de Interconexión Eléc-
based incentives for transmission investments in the long run, trica entre Bolivia, Chile, Colombia, Ecuador y Perú’’ (UNDP, 2010).
when congestion rents decrease in importance (as markets become However, the ideas presented in this paper are only responsibility
more integrated). Long-term contracts can be associated with of the authors and they do not represent the position or thoughts
either international financial transmission rights that are tradable of the UNDP nor any of the authorities of the countries involved
among market agents of different countries or an international in this research.
transmission tariff system. Moreover, other mechanisms (such as
the implementation of a public fund for giving financial support to References
invest in international links) can considerably help in order to
accelerate the regional energy integration. Barroso, L., Rudnick, H., Mocarquer, S., Kelman, R., Becerra, B., 2008. LNG in South
America: the markets, the prices and the security of supply. In: Proceedings of
the IEEE Power Engineering Society 2008 General Meeting, Pittsburg, Pennsyl-
vania, USA, July 2008.
Comunidad Andina de Naciones (CAN), 1969. Acuerdo de Integración
6. Conclusions Subregional Andino del 26 de mayo de 1969—Acuerdo de Cartagena (Andes
Sub-regional Integration Agreement of May, 26th of 1969—Cartagena
Agreement). Available on line at: /http://www.comunidadandina.org/norma
We have analyzed the electric integration among five countries tiva.htmS.
in the Andes region: Bolivia, Chile, Colombia, Ecuador, and Peru. We Comunidad Andina de Naciones (CAN), 2002. Decisión del Acuerdo de Cartagena
have determined technically feasible interconnection points that 536: Marco General para la interconexión subregional de sistemas eléctricos e
intercambio intracomunitario de electricidad (Cartagena Agreement Decision
allow establishing the approximate operation of interconnected No. 536: General Framework for the Sub-regional Interconnection of Electric
electric markets. Systems and Intra-communities Electricity Exchange). Year XIX, No. 878, 19 of
Even though we have assumed that the generation expansion December of 2002. Available on line at: /http://www.comunidadandina.org/
normativa/dec/D536.htmS.
plans of the countries were not altered with respect to the plans DigSilent, 2010. PowerFactory V14 Base Package. Available on line at: /http://www.
provided by the authorities of these countries (which corresponds digsilent.de/?p=Software/DIgSILENT_PowerFactory&package=BasicS.
to the worst-case scenario for the energy integration in the Andes Empresa Propietaria de la Red (EPR), 2000. Diseño General del Mercado Eléctrico
Regional (MER) en proyecto SIEPAC (General Design of the Regional Electric
region), we found significant economic benefits of such intercon-
Market (MER) under the SIEPAC project). Available on line at: /http://www.
nections from the supply side, the demand-side, the costs-savings, eprsiepac.com/S.
and the environmental viewpoint. European Market Exchange, 2009. European Climate Exchange (ECX) Emissions
It is possible to increase the economic efficiency among Index. Available on line at: /http://www.ecx.eu/ECX-EUA-IndexesS.
Gorenstin, B., Campodonico, N., Costa, J., Pereira, M., 1991. Stochastic optimization
the considered countries by augmenting the international energy of a hydro-thermal system including network constraints. In: Proceedings of
exchanges. In doing this, it is needed clear market rules and the power industry computer application conference, Brazil, September
the implementation of mechanisms that both adequately manage 1991.
Hammons, T., Corredor, P., Fonseca, A., Melo, A., Rudnick, H., Calmet, M.,
the market risks and are consistent with the development of Guerra, J., 1999. Competitive generation agreements in Latin American
financial contracts. Accordingly, we provide the following policy systems with significant hydro generation. IEEE Power Engineering Review
recommendations: (i) Establishing rights and obligations of gov- 19 (9), 4–21.
Hammons, T., De Franco, N., Sbertoli, L., Khelil, C., Rudnick, H., Clerici, A., Longhi, A.,
ernments, allowing the modifications of local energy operations 1997. Energy market integration in South America. IEEE Power Engineering
and planning policies, (ii) Developing rules that encourage invest- Review 17 (8), 6–14.
ments in both international links and local transmission expan- Intergovernmental Panel on Climate Change (IPCC), 2007. In: Metz, B., Davidson, O.,
Bosch, P., Dave, R., Meyer, L. (Eds.), Mitigation of Climate Change—Contribution
sions needed to support international energy exchanges in a timely of Working Group III to the Fourth Assessment Report of the Intergovernmental
and sustainable manner, (iii) Strengthening appropriate regional Panel on Climate Change. Cambridge University Press, Cambridge, UK and New
initiatives, establishing objective, transparent, and non-discrimi- York, USA.
Kas Ingenieria S.A., 2007. Modelo Multinodal Hidrotermico OSE2000: Manual de
natory market rules, (iv) Creating the correct incentives to encou-
Uso (Hydro-Thermal Multinodal model OSE 2000: User’s Manual). Available
rage the economy, security, and quality of the electricity service in upon request to mailto:samuel.jerardino@kasing.cl.
the region, (v) Establishing a domain for the development of the Pereira, M., Pinto, L., 1991. Multi-Stage Stochastic Optimization Applied to Energy
international energy exchanges through the design of both flexible Planning. Springer, New York, USA.
Rudnick, H., 1997. South American experience in deregulation of the electricity
and firm financial contracts, and (vi) Attempting that the benefits of energy industry. In: Proceedings of the IEEE/IEE Future of the Energy Business
the energy interconnections reach to the consumers through International Conference, Toronto, Canada, November 1997.
 E. Sauma et al. / Energy Policy 39 (2011) 936–949 949

Sauma, E., Oren, S., 2007. Economic criteria for planning transmission investment in United Nations Development Program (UNDP), 2010. Estudio para Análisis de
restructured electricity markets. IEEE Transactions on Power Systems 22 (4), Prefactibilidad Técnico Económica de Interconexión Eléctrica entre Bolivia,
1394–1405. Chile, Colombia, Ecuador y Perú (Study for the Technical-Economic Pre-
Sauma, E., Oren, S., 2006. Proactive planning and valuation of transmission feasibility Analysis of Electric Interconnection among Bolivia, Chile, Colombia,
investments in restructured electricity markets. Journal of Regulatory Econom- Ecuador, and Peru). Available upon request to UNDP.
ics 30 (3), 261–290.