Hacettepe University Graduate School of Social Sciences 

Department of Economics 

 

 

 

 

 

 

 

 

 

ELECTRICITY DEMAND FORECASTING METHODS USED IN 

TURKEY AND THEIR EFFECTS ON INVESTMENTS IN 

ELECTRICITY SECTOR 

 

 

 

Ertan TAŞKIRAN 

 

 

 

Ph.D. Dissertation 

 

Ankara, 2021 

 

 

 



 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



ELECTRICITY DEMAND FORECASTING METHODS USED IN TURKEY AND 

THEIR EFFECTS ON INVESTMENTS IN ELECTRICITY SECTOR 

 

 

 

 

 

 

Ertan TAŞKIRAN 

 

 

 

 

 

Hacettepe University Graduate School of Social Sciences 

Department of Economics 

 

 

 

 

Ph.D Dissertation 

 

 

 

 

 

 

 

Ankara, 2021 

 



ACKNOWLEDGEMENTS 

 

During the preparation of this dissertation, I owe my deepest gratitude to my advisor 

Assoc. Prof. Dr. Nasip Bolatoğlu. In addition, I would like to thank Prof. Dr. Levent 

Aydın with his positive observations and recommendations. Also, I cannot forget the 

sincere and constructive approach of my monitoring committee member, Assoc. Prof. 

Dr Selcen Öztürk. 

 

Again, I would like to express my gratitude to members of my dissertation jury, Assoc. 

Prof. Dr. Ali Murat Berker and Assoc. Prof. Dr.  İzzet Arı, for their great 

recommendations and constructive advices for the completion of the thesis. 

 

Likewise, I would also like to thank the TEİAŞ employees who provided the data used 

in the thesis, and my Colleague Hidayet Yakupoğlu, who made valuable contributions. 

 

Finally, I would like to thank my beloved wife Dilek, who stood by me with her endless 

patience, support and understanding during the this dissertation preparation process.



v 

 

ABSTRACT 

 

TAŞKIRAN, Ertan. Electricity Demand Forecasting Methods Used in Turkey and 

Their Effects on Investments in Electricity Sector. Ph. D. Dissertation, Ankara, 2021. 

 

Electrical energy is one of the most critical elements for accessing applications that 

are related to modern life. Due to the nature of electricity as a commodity that cannot 

be stored on a large scale, demand management should be handled with particular care. 

For this reason, realistic demand forecastings must guide investment decisions. This is 

also an essential element for the balance of payments and investment planning. Thus, 

electricity investments are also essential to determine accurately associated with 

unsaturated electricity demand in Turkey. Furthermore, many tools and materials used 

in the electricity sector are requiring high technology imported goods. Besides, the 

fuels used in electricity generation and the planning of the power plants that use these 

fuels affect all economic activity. As a result, planning is the most important factor for 

power plants. For this purpose, some energy models and forecasting methods have 

been developed for use in many countries worldwide. Some of these models and 

methods have also been used in Turkey, but high rate deviations have been observed 

up to 85% in long-term official forecastings. Even though rapid developments of 

computer and software technology ensure to enhance forecasting accuracy in many 

countries, such a deviation ratio is an unacceptable subject in Turkey. In this thesis, as 

a common and high forecast accuracy rate, the ARIMA Model and Artificial Neural 

Networks (ANN) have been used for future 10-year demand forecasting in Turkey. 

These forecasting results have been compared with the actual amounts and official 

forecasting. Then, the relationship between these results and power plant investments 

in Turkey has been discussed. Accordingly, it is analyzed that existing investments 

and future investments can meet the electricity demand. Finally, the proposals for 

investment in Turkey in the international developments and practical guidance and 

solutions have been introduced. 

 

Keywords: Electricity Market, Electricity Demand Forecasting, ARIMA, Artificial 

Neural Networks. 

 



vi 

 

ÖZET 

 

TAŞKIRAN, Ertan. Türkiye’de Kullanılan Elektrik Talep Tahmin Yöntemleri ve 

Yapılan Tahminlerin Elektrik Sektörü Yatırımlarına Etkileri. Doktora Tezi, Ankara, 

2021.  

Elektrik enerjisi modern insan yaşamına dair bütün uygulamalara ve hizmetlere 

erişebilmek için en önemli araçtır. Doğası gereği büyük ölçekte depolanamayan bir 

mal olduğu için talep yönetimi özel bir özenle yürütülmelidir. Bu nedenle gerçekçi 

elektrik talep tahmin projeksiyonları elektrik sektörü yatırım kararlarına yol gösterici 

olmalıdır. Böylelikle, elektrik sektör yatırımları Türkiye’de henüz doymamış olan 

elektrik talebi ile uyumlu yürütülmelidir. Zira, elektrik sektöründe kullanılan pek çok 

malzeme ve araç-gereç yüksek teknolojili ithal mallardır. Ayrıca, elektrik üretiminde 

kullanılan yakıtlar ve bu yakıtları kullanan santrallerin de planlanması ülkenin bütün 

ekonomik faaliyetini etkilemektedir. Bu amaçla, pek çok ülkede bazı enerji modelleri 

ve tahmin yöntemleri geliştirilmiştir; ancak bu yöntemlerin de kullanıldığı Türkiye’de 

uzun dönemli resmi tahminlerde % 85’e varan oranlarda sapmalar görülmüştür. 

Bilgisayar teknolojileri ve yazılımlarındaki gelişmelerin, Dünya’nın geri kalanında 

yapılan tahminlerde kesinlikler sağladığı görülmekteyken, Türkiye’de yapılan 

tahminlerde bu denli yüksek sapma oranları görülmesi dikkatle ele alınması gereken 

bir husustur. Bu tezde, Türkiye’nin gelecek on yıllık elektrik talep tahmini için, 

yaygınlığı ve tahmin isabeti yüksek olan zaman serisi modellerinden ARIMA Modeli 

ile Yapay Sinir Ağları (YSA) kullanılmıştır. Yapılan bu tahminlerin sonuçları ile resmi 

tahminler karşılaştırılmış, sonrasında tahmin sonuçları ile özellikle elektrik güç 

santralleri yatırımları arasındaki ilişki irdelenmiştir. Böylelikle, mevcut yatırımlar ve 

gelecekte planlanan yatırımlardan elde edileceği öngörülen elektrik enerjisi 

üretiminin, elektrik talebini ne ölçüde karşılayıp karşılayamayacağı da tartışılmıştır. 

Ayrıca Türkiye’de yapılan elektrik sektörü yatırımlarıyla ilgili olmak üzere 

uluslararası gelişmeler ve uygulamalar ışığında bazı öneriler getirilmiştir. 

Anahtar Kelimeler: Elektrik Piyasası, Elektrik Talep Tahmin Yöntemleri, ARIMA 

Modeli, Yapay Sinir Ağları. 



vii 

 

TABLE OF CONTENTS 

ACCEPTANCE AND APPROVAL ...................................................................................... i 

YAYIMLAMA VE FİKRİ MÜLKİYET HAKLARI BEYANI ......................................... ii 

ETHICAL STATEMENT .................................................................................................... iii 

ACKNOWLEDGEMENTS ................................................................................................. iv 

ABSTRACT .............................................................................................................................v 

ÖZET ..................................................................................................................................... vi 

TABLE OF CONTENTS .................................................................................................... vii 

ABBREVIATIONS  ............................................................................................................ viii

LIST OF TABLES ................................................................................................................ ix 

LIST OF FIGURES ................................................................................................................x 

PREFACE  ............................................................................................................................. xi 

INTRODUCTION ...................................................................................................................1 

CHAPTER 1: ENERGY, ELECTRICITY, and MARKET ...............................................6 

1.1 ENERGY .................................................................................................................6 

1.2 CLASSIFICATION OF ENERGY .......................................................................8

   1.2.1 Primary Energy Sources .................................................................................8 

  1.2.2 Secondary Energy Sources .............................................................................9 

1.3 ELECTRICITY ....................................................................................................10 

1.4 FEATURES OF THE ELECTRICITY MARKET ...........................................13  

1.4.1 Electricity Economy .......................................................................................15 

1.5 THE HISTORY OF ELECTRICITY MARKET .................................................17 

1.5.1 World Electricity Market ................................................................................17 

1.5.2 Electricity Market of Turkey ..........................................................................21 

CHAPTER 2: ELECTRICITY DEMAND AND FORECASTING .................................24

           2.1 ELECTRICITY DEMAND .................................................................................24 

           2.2 ELECTRICITY DEMAND FORECASTING ..................................................31 

2.2.1 Electricity Demand Forecasting Models and Methods ...................................36 



viii 

 

2.2.1.1 Electricity (Energy) Demand Forecasting Models .................................37 

2.2.1.1.1 WASP Model .................................................................................37 

2.2.1.1.2 MAED Model .................................................................................37 

2.2.1.1.3 LEAP Model ..................................................................................38 

2.2.1.1.4 NEMS Model .................................................................................38 

2.2.1.2 Electricity Demand Forecasting Methods ..............................................39 

2.2.1.2.1 Simple or Traditional Methods.......................................................40 

2.2.1.2.1.1 Extrapolation or Economic Indicators ....................................40 

2.2.1.2.1.2 Trend Analysis or Time Series ...............................................41 

2.2.1.2.1.3 Direct Surveys ........................................................................42 

2.2.1.2.2 Advanced Methods .........................................................................42 

2.2.1.2.2.1 End-Use Method ....................................................................42 

2.2.1.2.2.2 Advanced Econometric Models .............................................43

       2.2.1.2.2.3 Input-Output Model ................................................................43 

2.2.1.2.2.4 Scenario Method ....................................................................44 

2.2.1.2.2.5 Artificial Neural Networks .....................................................44 

2.2.1.2.2.6 Hybrid Methods .....................................................................46 

2.3 ELECTRICITY DEMAND FORECASTING IN THE WORLD ....................46 

2.4. ELECTRICITY DEMAND FORECASTING IN TURKEY ...........................48 

 2.4.1. Official Forecasting Studies ..........................................................................51 

2.4.1.1 Forecasting of Development Plans .........................................................51 

2.4.1.2 Other Studies ..........................................................................................54 

2.4.1.3 Generation Capacity Projection Reports of TEIAS  ...............................59 

2.4.2. Reasons of Forecasting Deviations ................................................................65 

CHAPTER 3: LITERATURE REVIEW ............................................................................69 

CHAPTER 4: ELECTRICITY INVESTMENTS ..............................................................75 

 4.1 ELECTRICITY INVESTMENTS  .....................................................................75 



ix 

 

 4.2. RESTRICTIONS OF INVESTMENT DECISION  .........................................80 

 4.2.1 Environmental Restrictions ............................................................................80 

 4.2.2 Technical Restrictions ....................................................................................84 

 4.2.3 Financial Restrictions .....................................................................................87 

 4.2.4 Bureaucratic Restrictions ................................................................................92 

 4.3. ELECTRICITY INVESTMENTS IN TURKEY ..............................................92 

 4.3.1. Total Electricity Investment ..........................................................................92 

4.3.2 New Investment Trends ................................................................................101 

4.4 CALCULATIONS OF GENERATION POSSIBILITIES OF TURKEY .....103 

 

CHAPTER 5: FORECASTING MODEL ........................................................................108 

 5.1 FORECASTING METHODOLOGY ...............................................................108 

 5.1.1 ARIMA Model ........................................................................................108 

 5.1.1.1 Autoregresivve Process (AR)  .........................................................108 

 5.1.1.2 Moving Area Process (MA)  ...........................................................109 

  5.1.1.3 Integrated Process (I)  .....................................................................110 

  5.1.2 Stages of ARIMA ....................................................................................112 

       5.1.2.1 Identify ...........................................................................................112 

  5.1.2.2 Estimation.......................................................................................113 

   5.1.2.3 Control-Checking ...........................................................................113 

   5.1.2.4 Forecasting or Diagnostic ..............................................................113 

 5.1.3 Artificial Neural Networks (ANN)  .........................................................114 

5.2 DATA ...................................................................................................................116 

 5.2.1 ARIMA Data ...........................................................................................116 

         5.2.2 ANN Data................................................................................................117 

  5.2.2.1 Economic Growth Rate ...................................................................118 

  5.2.2.2 Population Growth Rate  ....................................................118 



x 

 

   5.2.2.3 Export Growth Rate  ........................................................................119 

      5.2.2.4 Import Growth Rate  ........................................................................119 

 5.3 RESULTS ............................................................................................................119 

5.3.1 ARIMA Model Results .................................................................................119 

5.3.2 ANN Model Results .....................................................................................124 

5.3.3 Comparison of Model Results ......................................................................128 

CHAPTER 6: DISCUSSION AND RECOMMENDATIONS ........................................130 

CHAPTER 7: CONCLUSION ...........................................................................................139 

APPENDIX - A  ...................................................................................................................144 

APPENDIX - B  ...................................................................................................................148 

APPENDIX - C ....................................................................................................................169 

BIBLIOGRAPHY ...............................................................................................................170 

APPENDIX – D ETHICS COMISSION FORM .............................................................182 

APPENDIX – E ORIGINALITY REPORT .....................................................................184 

 

 

 

 

 

 

 

 

 

 

 

 



xi 

 

LIST OF TABLES 

Table 1. Classification of the Energy Sources (1971-2017) .....................................................9 

Table 2. The Share of Total Final Energy Consumption by Energy Sources (1971-2018) ....16 

Table 3. Plant Ownership in Turkey in March, 2021 ..............................................................23 

Table 4. Reasonable Electricity Grid Losses and Leakage Rates  ..........................................26 

Table 5. Electricity Demand Forecasting of 1st and 2nd Development Plans  .......................52 

Table 6. Electricity Demand Forecasting of 3rd Development Plan  .....................................52 

Table 7. Electricity Demand Forecasting of 4th Development Plan  ......................................53 

Table 8. Demand Forecasting of 8th Development Plan Preparation Committee ..................54 

Table 9. 1989 Demand Forecasting Study ..............................................................................55 

Table 10. 1990 Demand Forecasting Study with MAED .......................................................55 

Table 11. 1993 Demand Forecasting Study with MAED  ......................................................56 

Table 12. 1997 Demand Forecasting Study with MAED  ......................................................56 

Table 13. 2003 Demand Forecasting Study of MENR  ..........................................................57 

Table 14. 2004 Demand Forecasting Study of MENR ...........................................................58 

Table 15. Average Deviations Forecasting in Generation Capacity Projection Reports  .......62 

Table 16. Yearly Forecasting Deviations of Generation Capacity Projection Reports  ..........63 

Table 17. MENR- Electricity Demand Forecasting Projections 2019 ....................................63 

Table 18. Forecasting of TEIAS Generation Capacity Projection 2020 .................................64 

Table 19. Electricity Demand Forecastings & Predictions” SCOPUS Search  ......................71 

Table 20. Electricity Demand Forecastings & Predictions” SCOPUS Search  ......................71 

Table 21. Examples of Electricity and Energy Demand Studies for Turkey  .........................72 

Table 22. Life Span (Retirement) of the Power Plants ...........................................................75 

Table 23. Carbon Emissions Produced by Power Plants by Resources  .................................84 

Table 24. Peak-Load Demand Actual and Forecastings .........................................................85 

Table 25. Start Up Time of Power Plants  ..............................................................................85 

Table 26. Power Plants Construction and Operational Costs .................................................88 

Table 27. Number of Electricity Generation Corporations  ....................................................89 



xii 

 

Table 28. Capacity Utilization Mechanism of April 2020 ......................................................91

Table 29. Total Installed Capacity March, 2021 .....................................................................94 

Table 30. Development of Generation Capacity by Source 2009 – 2021/March) ..................95 

Table 31. Plannned Licenced Generation Plants.....................................................................95 

Table 32. Planning Power-Plants Investments and Types ......................................................96 

Table 33. Generation Capacity of Turkey by Years (GWh) ...................................................99 

Table 34. Generation Cost of Renewables 2010-2019 (USD/kWh) .....................................101 

Table 35. Projected Capacity of Power Plants by Type  .......................................................103 

Table 36. Installed Power and Average Capacity Factors of Power Plants in Turkey .........104 

Table 37. Additional Installed Capacity and Generation Capacity Calculations ..................105 

Table 38. Detailed Electricity Generation Capacity Calculation by Source (2020 -2024) ...105 

Table 39. Generation Calculations of Power Plants (GWh) (2020-2024)  ...........................106 

Table 40. Electricity Gross Demand Forecasting of ARIMA 2013 – 2020. .........................122 

Table 41. Electricity Gross Demand Forecasting of ARIMA 2021 - 2030...........................123 

Table 42. ANN Forecasting Results Using MATLAB (2013 -2020) ...................................126 

Table 43. ANN Results for 2021-2030 Using MATLAB .....................................................128

Table 44. Actual Demand and ARIMA-ANN Difference (GWh) ........................................129 

Table 45. Deviations of Official Forecastings Between 2013 - 2020 ...................................130 

Table 46. Comparison of ARIMA, ANN and Official Forecastings (2021-2030) ...............131 

Table 47. Some Forecasting Studies Performance ................................................................132 

Table 48. Available Generation Capacity and Surplus by 2025 ...........................................134 

Table 49. OECD Electricity Generation, Installed Capacity and Consumption  ..................135 

  

 

 

 

 

 

 



xiii 

 

LIST OF FIGURES 

Figure 1. Daily Electrical Load (Demand) Curve ...................................................................12 

Figure 2. Traditional Electricity Sector Stages from Generation to Consumption  ................13 

Figure 3. The Share of Electricity on Total Energy Consumption (1971-2017) .....................17 

Figure 4. Pilfered Demand of Electricity  ...............................................................................27 

Figure 5. Total Electricity Loss-Leakage Rates 1971-2014 (Transmission & Distribution) ..27 

Figure 6. Distribution Loss-Leakage Rate in Turkey (1994-2019) .........................................28 

Figure 7. The Share of Selected Energy Sources in World Total Consumption  ....................31 

Figure 8. Basic Stages of Forecasting  ....................................................................................33

Figure 9. Daily Electricity Load: Actual and Forecasting by Hours  ......................................36 

Figure 10. A Classical ANN Schema  .....................................................................................45 

Figure 11. Deviations of Official Forecasting Studies 1977-2004 .........................................59 

Figure 12. Load Curve of 2020 (8760 Hours) ........................................................................61 

Figure 13. Figure 13. Average Growth Rates of Electricity Demand  ....................................67 

Figure 14. Load and Demand Forecasting Publications by Years  .........................................69 

Figure 15. CO2 Emissions of OECD and World .....................................................................83 

Figure 16. CO2 Emissions Development of Turkey ................................................................83  

Figure 17. Turkey’s Electricity Power-Plant Investments by Source Between 2003-2020 ....94 

Figure 18. Turkey’s Electricity Installed Capacity and Peak Load by Years .........................98 

Figure 19. Installed Capacity to Peak Load Ratio ..................................................................98 

Figure 20. Development of Generation Capacity of Turkey  ................................................100 

Figure 21. ANN Model with Single Layer Four Inputs ........................................................115 

Figure 22. Relationship Electricity Demand and Economic Growth (%) .............................118 

Figure 23. ARIMA Model Output by EViews ......................................................................121 

Figure 24. EViews Output for Actual and Forecasted Demand 2013-2020 Period  .............123 

Figure 25. EViews Output for 2021-2030 Forecasting .........................................................124 

Figure 26.  ANN Schema with Weights and Bias  ................................................................125 

Figure 27. MATLAB Output for Actual and Forecasting Demands (2013-2020)  ...............126 



xiv 

 

Figure 28.  ANN Schema with Weights and Bias  ................................................................127 

Figure 29. MATLAB Output for 2021-2030 Forecasting .....................................................128 

Figure 30. Compares Result of Actual Demand and Forecastings ARIMA and ANN  ........129 

Figure 31. Installed Capacity and Gross Demand Growth Rate  ..........................................133 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



1 

 

INTRODUCTION 

Many services and applications related to modern human life can be accessed with 

electrical energy. To carry out these services continuously and without interruption, 

the interconnected electrical system must work flawlessly. It is very important to 

ensure the supply-demand balance of electricity, which is considered a public good in 

many countries due to this function. Electricity can be generated simultaneously with 

consumption, and for this reason, system balance must be perfectly maintained. If the 

balance cannot be achieved, many services related to modern life are interrupted: 

Metro and train transport, household appliances, communication, health services, 

enlightenment, etc. Thus, it is vital to plan electricity generation to meet the electricity 

demand, and it can be ensured only by accurate demand forecasting. 

There are many studies related to electricity demand forecasting, though many of them 

concerning engineering and electrical industries in Turkey, there have not been any 

academic studies on the effects on investment. In this scope, the relationship and 

interaction between investments and demand forecastings will be examined and aimed 

to contribute to the literature. 

In many developed countries, where electricity demand is saturated, electricity 

investments have concentrated more on renewable energy sources and energy 

efficiency with the effect of both free-market dynamics and the regulatory-

interventionist role of the state.  Although electricity demand has increased an average 

of around 10% in Turkey in the last 50 years, power plant investments have reached 

such a level far above the electricity demand in recent years. The renewable power 

plants have outstandingly increased, but investments in other conventional resources 

have also increased rapidly. For this reason, in the electricity sector, which is a capital-

intensive sector, it cannot be denied that rapid technological developments and highly 

import-dependent sector. Therefore, investments should be carefully planned in this 

sector. For good planning in the efficient investments in the electricity sector, demand 

forecastings are among the most critical indicators. But, demand forecasting studies 



2 

 

carried out in Turkey for many years have not shown a good indicator of their high 

rate of deviation.  

Demand forecasting studies were not elaborated properly during the vertically 

integrated state monopoly period due to the high increase in unsaturated demand and 

cross-subsidies of TEK. The Turkish electricity sector had been operated under the 

state monopoly for many years, and it has turned into private-sector domination with 

the liberalization efforts after 2001. In this period, the idea of supplying electricity to 

every village of the country caused demand forecasts not to be needed or was not taken 

seriously. However, when the share of the private sector in the installed capacity 

increased, its’ effects on distribution and transmission sub-sectors also increased.  

However, the results of demand forecasts made in Turkey have deviated greatly over 

the years. Although there are a slight deviations in the demand forecastings carry-out 

in the World, the deviation rates in Turkey are unfortunately too high to be ignored. 

On the other hand, the reason for the need for investments in the electricity generation 

sector in Turkey is the forecastings that the electricity demand will be very high in the 

future. For this reason, generation investments made by the private sector in Turkey 

continued to increase. However, the increase in the amount of demand remains quite 

below the production capacity. 

Since it is seen that the installed power has doubled in comparison to peak-load 

demand in recent years, forecastings should be made very precisely in terms of 

financing and investment efficiency. Generation planning, in other words, how much 

electricity will be produced, how the generation will be met from which power plants 

and the financing and amortization of the investments to be made are estimated by 

considering the forecastings data. If forecastings are not carried out properly, contracts 

and obligations signed before will cause severe problems in the industry.  

In addition, electricity investments in the world are realized according to the principles 

of profitability and efficiency, taking into account the unique characteristics of 

electricity. Although renewable energy investments increase in Turkey, as in the rest 



3 

 

of the World, with various support mechanisms, other resources are also supported. 

Purchases at guaranteed prices with support mechanisms cause negative effects on 

end-user welfare in terms of price.  

Thus, it is clear that the competitive environment, which is stated to be created by the 

increase in investments, could not be achieved, that market efficiency in the electricity 

market, which was liberalized in 2001, have not realized, and that no benefit could be 

provided for the final consumer. In this process, public power plants were privatized, 

and the share of the private sector in electricity generation reached 80%. This rapid 

transformation experienced in the sector, especially starting from 2010, caused new 

investors to enter the market and the transferred funds from other sectors. These 

investors used foreign credit resources as much as they used their own resources for 

electricity sector investments. Today, the rate of loans used by the sector has reached 

around 50 Billion USD. This situation leads to the realization of idle investments above 

the needs and the exclusion of loan resources from the needs of other sectors. 

In the current situation, the gross demand amount in the electricity generation sector 

is around 304 TWh, while the available generation capacity is calculated as 490 TWh. 

Also, electricity demand in Turkey is being saturated gradually, it is predicted that the 

increase in licensed and unlicensed power plants will increase this idle capacity. 

Although many energy forecasting models and methods with international validity are 

used for estimation electricity demand, the high demand forecast amounts of MENR 

and other public institutions still have been announced, and huge deviation rates have 

been observed too. 

Forecasting studies carried out in many countries, the models and methods, the 

variables and coefficients of these models, calculation tools and justifications are 

disclosed to the public in detail, but in Turkey, only the amounts are disclosed by the 

MENR. Due to this non-transparent process, precise determinations cannot be made 

regarding the reasons for forecasting deviations. Nevertheless, it is understood that it 

is desirable to declare the amount of demand to be high by using overly optimistic 

demand growth amounts. The desire to make a generation investment by making high 



4 

 

electricity demand forecasts cannot be analyzed because the basic principles of public 

administration, transparency, accountability and openness, are not followed and the 

details of the calculations are not known. 

In this dissertation, it is aimed to calculate the state of electricity supply and demand 

balance, based on why official demand forecasts deviate at such high rates, and 

electricity generation investments are realized according to these demand forecasts. 

the high-rate deviations in the gross electricity demand forecasts are examined by 

periods and the reasons for the deviations are examined. In addition, the electricity 

generation plant investments realized from the past to the present and the amount of 

production and available capacity are calculated, and the energy that the plants planned 

to be put into operation could produce in the future according to their capacity ratios 

are  calculated. 

To forecast electricity demand, energy-electricity forecasting models and methods 

used in the world and Turkey are examined briefly and tried very accurate results 

predicted by ARIMA. The increase in investments and the optimum investment level 

are tried to be determined using these forecastings. Then, ARIMA model results are 

compared with the ANN model by using economic and social variables. These results 

are analyzed with actual demand and official forecasts. As a result,  calculations made 

by these two methods show demand quantities far below the official demand forecasts. 

Since the models and variables used for official demand forecasts and the 

considerations are not disclosed to the public, the real reasons for the deviations cannot 

be revealed. However, it has been observed that high annual demand growth rates have 

been determined without considering the slowdown in electricity demand growth rate 

in recent years. It has been observed that the rate of increase in demand has decreased 

since the 2000s in the periods separated into sub-periods since 1923. However, official 

estimates use higher growth rates without considering these declining rates. 

In the light of these explanations, in this dissertation, which consists of seven chapters, 

firstly energy and electricity are dealt with as a commodity, and accordingly, the stages 

of the electricity market in the world and in Turkey are discussed. In the Second 



5 

 

Chapter, electricity demand and forecasting are examined and electricity demand 

forecasting models and methods used in the world and in Turkey are briefly explained. 

Then, demand forecasts in Turkey starting with the Planned period and actual demand 

were compared and electricity demand forecasts for Turkey were mentioned. 

In Chapter 3, electricity demand forecasting literature review is made and explanations 

are made about the established model. In Chapter 4, electricity investments and the 

constraints for making investments are discussed, and current and future electricity 

generation possibilities in Turkey are calculated. 

In Chapter 5, forecasting has been made with the model established with ARIMA and 

ANN and the results were compared with the actual demand amounts. Finally in the 

Chapter 6 with Conclusion, in the light of the forecastings of model, electricity 

generation investments and demand projections in Turkey are discussed and 

suggestions are presented. 

 

 

 

 

 

 

 

 

 

 

 



6 

 

CHAPTER 1 

ENERGY, ELECTRICITY, AND MARKET 

1.1 ENERGY 

Although energy is generally referred to as the ability to do work or the power to 

generate heat, the word energy was first used by the Greek Philosopher Aristotle as 

energeia, the meaning of this word remained uncertain at that time. This uncertainty 

continued for almost two thousand years between experimental philosophy and 

theological debates. Throughout the entire Middle Ages and then even in Galileo, 

Torricelli and Newton, there was no clear definition of energy. Later, in 1717, 

Bernoulli defined energy roughly as “the product of force times the (virtual) way in its 

direction” (Kümmel, 2011). 

The conservation of energy was first described in terms of vis viva. According to 

Leibniz, vis viva was the sum of the mass, and squares of velocity formed the force (p 

= mv2) (McDonough, 2019). This concept was later called kinetic energy and was also 

described as a “living force” for mechanical energy conservation. Then, Albert 

Einstein designated (E = mc2) the relation between energy (E) and mass (m), and 

energy is equivalent to multiplied of mass and square of the speed of light (c2) 

(Kümmel, 2011). 

The energy process can be divided into three parts: Bulk material allocation from the 

natural environment, separation of the raw material, and revealing the energy stored in 

the raw material (Niemes and Schirmer, 2010). The materials needed to produce 

energy are found in proportionally low amounts in the natural environment. This low 

amount of material is separated from the environment where it is located, and then 

energy is released by passing through certain processes.  

While the energy is obtained, all the energy in the material is not revealed. This is 

“anergy.” Anergy is called energy that does not turn into another type of energy or 



7 

 

work under certain thermodynamic conditions. As free disposal costs are inevitable 

for by-products, this is precisely the situation in which we encounter with anergy.  

Also, here we see the diminishing marginal production assumption for each additional 

output, and production is a decreasing function of its inputs. This assumption also 

means an increase in anergy for the by-product because all forms of energy are still 

constant (Niemes and Schirmer, 2010).  

Considering these concepts in the context of energy production, it is necessary to 

explain the concept of thermodynamics. The term thermodynamics was first used by 

Lord Kelvin in his publication in 1849. Thermodynamics is derived from the Latin 

words “therme” (heat) and “dynamis” (power). It can be defined as the branch of 

science that deals with the energy and the deformation of energy. Even today, 

thermodynamics is defined as energy, heat, work, and entropy science (Akdağ, 2009). 

In the mid and late 19th Century, although many scientists adhered to the idea that heat 

is a weightless fluid, empirical evidence supporting the heat-work energy equivalent 

became dominant. Thus the expression of equivalence formulated the First Law of 

Thermodynamics. About the First Law of Thermodynamics, energy (generally heat) 

is conserved in its quantity (including energy equivalent of mass). It can be neither 

created nor destroyed. The total energy in the universe remains constant (Michaelides, 

2012). 

The Second Law of Thermodynamics, which is the main principle for converting heat 

into work, states “no free lunch” (Kümmel, 2011). The meaning of this point for life 

is that plants, people, and animals maintain their lives, and facilities related to modern 

life need additional energy (Hall, 2017). It is implied that if there is not any input, there 

is not production. The second law of thermodynamics states that heat always flows 

from the hot body to the cold body. Not all energy used as input is included in the 

output because some pass to another place in the process (or space). 

As energy turns into another useful form, this process is created through an external 

intervention or a converter. For example, automobile motors are used to provide 



8 

 

motion energy, heaters, and air conditioners are used for heating or cooling, and 

machinery used in production is a kind of converter. The aim here is to explain how 

this transformation works according to the laws of thermodynamics rather than how 

transducers work (Martínez et. al., 2018). Thus, according to the first law, input energy 

is equal to output energy, while according to the second law, input energy is equal to 

output energy plus wasted energy.  

Energy is potentially contained in water before flowing into turbines, coal, natural gas, 

oil before burning, or in radioactive elements before reacting. Kinetically, it manifests 

itself in the form of rivers, streams, and waves in the sea, the wind itself, the radiation 

of the sun. Accordingly, to create a movement other than man's muscle power, creating 

energy by using the substances in nature directly or indirectly is performed under the 

laws of thermodynamics. Thus, primary energy can be obtained as motion energy from 

wind and stream or as heat energy from burning a substance directly. As a result, 

secondary energy can be produced by converting the movement or heat to another 

energy. 

1.2. CLASSIFICATION OF ENERGY  

Although the most common classification of the energy resources is fossil resources, 

nuclear energy, and renewable resources, considering the category of primary and 

secondary energy resources, another classification of energy resources in this 

dissertation. However, it is helpful to explain here one of the standard classifications: 

(i) Fossil sources and nuclear sources called “non-renewable” energy sources. Non-

renewable resources, once passed through the production stages, are resources that 

disappear in nature and cannot be directly used in energy production again. (ii) 

“Renewable” energy sources are resources that can be used as long as they exist in 

natural processes that do not disappear due to production stages. (Aydın, 2018) 

1.2.1. Primary Energy Sources  

The primary energy sources are divided into fossil fuels, nuclear sources, and 

renewable energy sources. While oil, natural gas, and coal are counted in fossil fuels, 



9 

 

hydraulic energy, solar energy, wind energy, geothermal energy, biomass energy, and 

marine energy are among renewable energy sources. 

Apart from human muscle and animal power, they are primary sources such as oil, 

coal, natural gas, water, wind, solar and geothermal energy obtained from nature. As 

some of these can be used directly, such as coal and natural gas, some are offered for 

direct use after various very few physical and chemical processes, such as petroleum. 

1.2.2. Secondary Energy Sources 

Secondary energy sources are energy sources obtained by applying substantially 

chemical and physical processes to primary energy sources. Besides the sources 

obtained by passing through chemical processes such as fuel-oil coke coal and coal 

gas, the most common secondary energy source is electrical energy. Electrical energy 

is generally obtained by producing motion energy by primary energy sources through 

nuclear reaction or physical transformation. In order to obtain products such as 

gasoline, LPG, diesel oil, crude oil must be distilled in refineries (Aydın, 2018). 

Apart from these classifications, a classification can also be made according to whether 

the energy is tradable in the market, whether it is renewable or not, and whether its 

technology is old or new. 

     Table 1. Classification of the Energy 

 

         Source: Aydın (2018) 

 

Renewable Non-Renewable

Animal and Plant Waste  

Windmill & Watermill   

Wood (Sustainable)

Solar 

Small Hydraulic

Wave

Rock Oil                                                

Rock Gas                        

Rock Coal

Energy Classification     

(Tradable or 

Conventional/Renewable)

Renewable or not

C
o

n
v

e
n

ti
o

n
a

l

Commercial      Fossil Fuels

Conventional / Non-

Commercial
  Wood (Non-sustainable)

New- Innovative

Large Hydraulic

Geothermal           

Nuclear



10 

 

1.3. ELECTRICITY 

Electricity is a physical event caused by charged particles that are stationary or 

moving. The basic element in electricity is the accumulation of negatively charged 

electrons, one of the particles of the atom, in one direction or moving in another 

direction. Electricity is the most convenient way of transferring energy. Simply, 

electricity is the flow of electrical charge.   

Most materials are electrically neutral, and these materials have an equal amount of 

positive and negative electrical charges. When an electric field is applied to such 

materials, the positive and negative charges are displaced into the electric field and the 

current direction. As a result, a pair of positively and negatively charged dots with a 

small distance between them occur, called electric dipoles (Matsushita, 2014). 

As is known, particles with the same sign (plus or minus ends) repel each other, while 

particles with different signs attract each other. A positive particle in an atom is a 

proton, and a negative particle is an electron. Neutrons and protons that are uncharged 

are located in the nucleus of the atom, as electrons are free in the orbit of the nucleus. 

The separation of free electrons in many metal materials from their atoms and 

transporting them through a free conductor towards another atom creates electricity. 

To obtain the high amount of electrical energy required in the industry, it is necessary 

to rotate the electric generator connected to the turbine shaft in any power (heat, 

hydraulic, gas, or nuclear reaction) with mechanical energy. The principle of the 

operation of an electric generator is based on the movement of a conductive wire, 

usually copper, made of a material from which electrons are released in a magnetic 

field. The electric generator generates the electric current that is moved from the atom 

to the atom by rotating the wound conductor wires called stator rotating in a magnet 

called the rotor. The electron movement that occurs in this way is transmitted by the 

movement power called voltage. 

The properties of electricity as good can be mentioned as follows: 



11 

 

- Electricity is an unique homogenous commodity. Its frequency measures the 

quality of electricity as good in Hertz, and it’s continuous supply. Although 60 Hz 

frequencies are used in some countries, most countries supply electrical energy at 50 

Hz to end-users. There are also offered to the final consumer of electricity at 220 V 

and 50 Hz in Turkey. Therefore, if it is supplied uninterrupted, everyone can use a 

uniform good of the same quality.  

- Electricity is not consumed directly, and it is used with some appliances and 

equipment. The generated electricity is transmitted at 380 kV or 154 kV voltage levels 

in Turkey, and the substations lower the voltages. After that, a low voltage level (220-

380 V) is used by the industry or households with distribution lines. 

- Although it is similar to other goods, electricity has its characteristics. 

Electricity fluctuates demand by day, hours of the day, seasons, years, and regions.  

- The important feature here is that it cannot be stored economically with current 

technologies. As it cannot be stored economically, it must be used as soon as it is 

generated. 

- One of the most important features of electrical energy is that it can be divided 

into desired amounts. The amount passing through the meters can be easily transmitted 

in kWh units. If technical requirements are met, electricity can be used immediately. 

- It does not directly create environmental pollution.  

- Electricity is transmitted to the end consumer through lines; due to 

thermodynamic laws, technical losses increase as the distance between the place where 

it is generated and consumed is longer. This causes large investment expenditures. 

- One feature that distinguishes electricity from other goods is the sudden 

changes in demand within hours or even seconds. Electricity consumption, also 

generation fluctuate every second of the day. Since the electricity demand is extremely 

volatile, it is a source of energy produced when consumed, requiring perfect planning. 

After the base load is provided, the meteorological events, social events, holidays, or 

non-workdays affect electricity demand. 



12 

 

            

       Figure 1. Daily Electrical Load (Demand) Curve  

 

                 Source: TEIAS, Turkey Daily Electrical Actual Load Curve of 06.01.2021 

It is stated here that the baseload is required for the system to be in balance at minimum 

usage. Baseload is generally provided by thermal power plants such as coal-fired and 

nuclear power plants, which have slow activation and deactivation times. Increasing 

or decreasing consumption, rapidly activated or deactivated power plants use natural 

gas, hydraulic, and wind power plants. This balance is entirely planned by the 

transmission system operator (TSO - TEIAS in Turkey), and the increase and decrease 

in a generation are provided instantaneously with the signals given to the remote 

system-connected plants. 

Since electricity generation plants are diverse, they are activated according to the order 

called merit order. Therefore, the current energy sources are listed according to the 

capacity to be activated in the entire electrical system simultaneously in balance (IEA, 

2001). With the introduction of smart-grid technologies, the effective balance of 

increasing and decreasing generation is analyzed demand changes through real-time 

communication tools and control panels. In particular, when demand reaches its peak 

level in minutes and seconds, instant supply-demand balance can be achieved thanks 

to smart grid technologies (Pratt et. al., 2010). Thus, as stated above, the continuous 

supply of electricity is guaranteed to be readily available if demand increases per 

second.  To ensure this continuity, the following four phases are generally mentioned 

in electricity: Generation, transmission, distribution, and retail. 

0

10.000

20.000

30.000

40.000

50.000

1 2 3 4 5 6 7 8 9 101112131415161718192021222324

C
o

n
su

m
p

ti
o

n
 (

M
W

)

Hours

Daily Electrical Load Curve



13 

 

The electricity generation sub-sector is a highly capital-intensive sector. Economies of 

scale are limited, and coordination economies are seen. In general, many companies 

in this sub-sector show competitive market characteristics. The transmission and 

distribution sub-sectors show natural monopoly properties and have high sunk costs. 

Here there is no competition, or a limited number of actors are empowered. The retail 

sector, where the end-user is located, can generally show competitive features (IEA, 

2001).  

Figure 2. Traditional Electricity Sector Stages from Generation to Consumption 

 

1.4 FEATURES OF THE ELECTRICITY MARKET  

Energy using is an important factor in economic development and is considered one 

of the conditions of development. This relationship which continues with economic 

development, also increases the demand for better energy services. 

In this context, fossil fuels, renewables, alternative energy sources, energy efficiency, 

energy independence and security, and climate changes are also factors to be 

considered. These issues require good planning to ensure continuity as the energy 

supply, and demand balance is related to the energy source where produced or 



14 

 

extracted and where consumed is a different area (Dorsman et. al., 2013). The energy 

sector requires interdisciplinary cooperation as well as regions, countries, and locally 

with special attention because of its technical requirements and capital intensity. 

Besides, it should not be ignored as much as its social dimension also contributes to 

production (Bhattacharyya, 2011). However, there has been more interest in 

environmental and social concerns of development in recent years when markets were 

established, and energy policies were determined. These restrictions impose several 

obligations, particularly under the Kyoto Protocol and the Paris Agreement and civil 

society pressure for decreasing greenhouse emissions.  

Environmental effects are inevitably occurring in all energy production areas. There 

are some critical environmental reactions (about air pollution, climate change, waste 

and water pollution, loss of biodiversity, etc.), especially in power plants using coal 

and oil and nuclear, because the reactions of many people who are interested in 

environmental events other than electricity generation or economics create public 

pressure (Harris, 2006). However, energy is an indispensable element for production, 

such as labor and physical capital, and affects economies heavily when there is a 

supply shortage. Also, energy infrastructures include long-term planning, investment, 

and operation phases. If this process does not harmonize with the economic 

environment and the social environment, it is inevitable to experience supply or 

demand problems in the long run.  Zweifel et. al. (2017) explain briefly; 

- Since energy infrastructures involve long-term planning, investment, and 

operation phases, this process should harmonize with the economic and social 

environment. Therefore, the political effects and regulations of public authorities are 

more dominant than other sectors. 

- As the processes of producing (or extracting), transmitting, and transporting 

energy are the most pollutant factors of water, air, and soil, it is one of the sectors 

where negative externalities are most common. Due to these externalities, it is not 

Pareto Optimum since it does not reflect the actual prices of the energy product. But 

electricity does not directly affect the environment. 



15 

 

- Many energy markets exhibit monopoly or oligopoly characteristics rather 

than full competition. Especially the existence of natural monopolies, lack of 

competition, or being too difficult necessitates a public authority to regulate these 

markets. 

After addressing these issues, energy services considered one of the basic human needs 

such as shelter, food, economic, social, and environmental aspects are all related to 

human development (IEA (c), 2005).  

1.4.1 Electricity Economy 

The electricity economy is mainly divided into two categories: The first is called the 

electricity demand economy, and it concentrates more on the simultaneous balancing 

between production and consumption. The second deals with the optimal allocation of 

resources and electric power source also called the “electricity supply economy” or the 

“electrical energy economy” or “power system economy” (Hu, 2013). 

This dissertation dealt mainly with the demand economy in the first category. Because 

the power systems economy is mostly related to technology, especially production 

technology for electricity generation and the use of resources, it is useful to state that 

the electricity demand occurs simultaneously with the electricity supply. That is, 

electricity is an exceptional commodity that can be produced when it is used 

synchronously. In economic theory, it is quite difficult to find another good that can 

be used in this way, except for free goods. It should be noted here that it is important 

for electricity to increase its prevalence and usage in every field in modern society.  

According to the supply and demand theory, price is the main determinant for the 

demand for a good. Consumers can increase or decrease their demands according to 

the price of the goods within the scope of substitution possibilities. However, 

electricity is a compulsory facility used to access all amenities of modern life such as 

IT, heating, cleaning, cooking utensils. In this respect, substitution for electricity 

consumption is very difficult, sometimes impossible to compare with other goods. 

Therefore, it is seen as a sector that needs to be carefully managed and regulated. 



16 

 

Especially the advantage of electrical energy is its widespread use in many sectors and 

households, and it is difficult to substitute for other energy sources. While it is possible 

to replace the electricity used for some purposes, for example, heating, hot water, and 

cooking, it is still not fully substituted in many areas. Because electricity is the most 

important element in many consumptions and production fields, such as electrical 

appliances, air conditioners, and even cars, this is related to the electrical infrastructure 

and how widely used it. This substitution is directly related to the non-widespread 

electrical infrastructure, the habits of use, and, most importantly, the level of income.  

The low price and income elasticity of electricity demand confirm this situation. Due 

to the low price and income elasticity of electricity, consumers cannot react in a short 

time when the price increases or decreases. Also, using it’s widespread, it is difficult 

to substitute easily in the long term. Accordingly, electricity is a compulsory good 

(inelastic demand) in terms of economic theory.  

Table 2. The Share of Total Final Energy Consumption by Energy Sources (1971-2018) 

Share of Final Energy 

Consumption by Source 

1971 2018 
World OECD Turkey World OECD Turkey 

Coal 14,96 12,64 14,87 10,01 2,68 10,26 

Oil 47,02 55,06 44,88 40,64 46,39 38,03 

Natural Gas 13,71 17,58 0,00 16,21 20,97 24,23 

Electricity 8,89 10,87 4,16 19,31 22,10 21,29 

Other (Biofuels and waste) 13,81 3,14 36,09   3,03 1,62 0,97 

Geothermal+ Solar+Wind 

etc. 
1,61 0,70 0,00 10,80 6,23 5,22 

Total 100,00 100,00 100,00 100,00 100,00 100,00 

Source: IEA, Key World Energy Indicators 2020 

It can be understood from the table above that the share of electricity doubled 

worldwide in 2017 compared to 1971, whereas the share of electricity in Turkey has 

quadrupled in the same period. Based on the increase in Turkey, there is also the impact 

of rural-urban migration and great electrification efforts in the 80s. Mainly, the 

investments made for village electrification corresponded to an average of 16.30% of 

the electricity investments. Likewise, electricity consumption increased 22 times in the 

industry, while there was a 36 times increase in household consumption in this period.  

 



17 

 

      Figure 3. The Share of Electricity on Total Energy Consumption (1971-2018) 

 

                   Source: IEA, Key Indicators of 2019. 

 

1.5 THE HISTORY OF THE ELECTRICITY MARKET  

1.5.1 World Electricity Market 

The electricity industry was organized in a multi-part structure since the first power 

plant was established in 1882 and continued this structure until the 1920s. There was 

intense competition between electricity producers and sellers, which are largely 

privately owned. It is observed that electricity was widely organized as an auto-

producer for the needs of manufacturers necessities in this first period. Besides, 

electricity was mostly used in the lighting of the street and public buildings.  

The electricity market in the World indicated a monopolistic market behavior from 

production to sales mainly invested by the private sector in this first period. Once 

again, regional and national monopolies became active in the sector. In short, the 

prominent feature of this period can be expressed as a private property dominant 

structure, scattered facilities, and the absence of the national network yet. Meanwhile, 

cartels have started to appear in the market. Even electricity was still considered a 

luxury goods by governments.  

0,00

5,00

10,00

15,00

20,00

25,00

1
9
7

1

1
9
7

3

1
9
7

5

1
9
7

7

1
9
7

9

1
9
8

1

1
9
8

3

1
9
8

5

1
9
8

7

1
9
8

9

1
9
9

1

1
9
9

3

1
9
9

5

1
9
9

7

1
9
9

9

2
0
0

1

2
0
0

3

2
0
0

5

2
0
0

7

2
0
0

9

2
0
1

1

2
0
1

3

2
0
1

5

2
0
1

7

S
H

A
R

E
 %

YEARS

World OECD Turkey



18 

 

Later, the public sector intervened dominated the market started from the 1930s and 

continued until the end of World War II, in the second period. The prominent feature 

of this period was the widespread confiscation and nationalization efforts. Starting 

from these years, the idea that electricity was a “public good and service” is a common 

idea until the 1970s. National networks developed rapidly, and large-scale power 

plants for various sources started to be established. 

Meanwhile, many European states were structured in the form of a completely vertical 

integrated natural monopoly (Thomas Edison also proposed such an organization), and 

a large number of small producers were organized nationally as a single monopoly or 

a large regional monopoly. In this period, electricity companies have tended to 

combine structurally as a single company. 

Vertically integrated power companies own transmission and distribution lines as well 

as generation facilities. A company established in this way is a monopoly to generate, 

transmit and distribute the electricity over a specific geographical area with the 

privilege granted to it by law or with its privileges (Kirschen and Strbac, 2004). 

This third period shows that monopolistic structures owned by the public are preferred 

to the monopoly run by the private sector. France established EDF in 1946, and Italy 

constructed ENEL as a monopoly with a vertically integrated structure owned by the 

state in 1962 (Varley, 1999). Turkish Electricity Association (TEK) was established a 

vertically integrated model of EDF in 1971. Also, ECGB (Central Electricity 

Generating Board) in the UK was one of the largest vertically integrated companies 

between 1958 and 1990. Following the functional unbundling in the 1990s, sub-

companies were sold through privatization. 

Only a different model has been applied in the USA, and special monopolies subject 

to state regulation and control have been formed. These monopolies have been 

constantly and tightly controlled by the State, such as tariffs and service quality 

(Atiyas, 2006). One of the most important features of this period is the economic 

completion of the vertical integration between small-scale power plants and production 



19 

 

and transmission due to oil shocks. After the 1970s, especially oil shocks, the period 

of intensive unbundling (vertical separation), competition, regulation, and 

privatization activities began in the electricity sector. The last stage of convergence 

and globalization resulting from technological innovations in electricity is now 

considered other commercial goods (Jentsch, 2001).  

The debate over whether electricity should be treated as a pure public good that 

requires special treatment at the long-term protection, or should be evaluated other 

goods in competitive markets according to supply and demand condition, has 

continued until today (IEA, 1994). 

The following five factors related to private sector dominance are essentially a matter 

of debate (Berrie, 1992): 

1- Especially since the transmission and distribution sectors show strong 

monopoly features, those who use these stages compulsorily do not have a choice. This 

can cause these monopoles to behave arbitrarily. 

2- Thanks to the increased efficiency, where decreasing costs are often seen in 

the electricity sector, this decreasing cost is generally not reflected on the final 

consumer. 

3- The electricity generation sector cannot be held responsible for 

environmental impacts and public impacts of transmission-distribution networks. 

Except for a few simple administrative regulations, an effective responsibility method 

cannot be established in the private sector-dominated sector. 

4- When consumers utilize public services such as direct lighting, they are not 

directly involved in their billing. As this is indirectly contributed, it cannot be easily 

seen how much of a burden is placed on consumers. 

5- People cannot easily understand what the high standard is in the electricity 

industry. While other goods and services can easily be compared, this possibility is 

almost nonexistent in electricity. Still, keeping the voltage constant and uninterrupted 

electricity supply may be the criteria here.  



20 

 

There has been intense debate ever since liberalization work began on whether 

electricity is a public good. Electricity has been defined as a public service in 

continental Europe. For example, about the German Electricity Law, from 1935 to 

1998, electricity was defined as “Daseinsvorsorge” (public service) in Germany, and 

it was stated that consumers could reach it at a reasonable price. (Tehrani et. al., 2013) 

Advocates of the first opinion believe that electricity is an integral part of modern life 

and it must be used to access all other home and industrial applications so that it can 

be provided to the users without any distinction, such as distant and near, poor and 

rich. That is, it should be organized as a specific sector that includes protective 

measures.  

The supporters of the second opinion argued that, although it has unique 

characteristics, electricity should be considered as other energy sources traded in 

competitive markets such as gas, coal, and oil, only because of the complex nature of 

this market, there should be a competitive structure by providing system and supply 

security (IEA, 1994). 

However, these intense debates in the electricity market have recently focused on 

whether electricity grid security is in the public domain or not. Network security in the 

transmission industry is often seen as a public good. The basis of this is that a network 

is a tool that must be used and that the users cannot be excluded from consuming these 

goods or services and cannot be competitive (Kiesling, 2009).  

Nevertheless, unlike this view, there is a transition and institutionalization towards a 

competitive market dominated by the private sector in many countries around the 

world. In this context, starting from Chile in 1982, there has been a transition to 

market-oriented approaches in the electricity industry in Argentina, Norway, England, 

Wales, Australia, Spain, and California-USA. 

Norway is the first European country to create liberalization electricity markets in 

1991. The market operator, which was established as an independent company in 



21 

 

Norway, was named Nord Pool with the partnership of Sweden. Later, Finland and 

Denmark also joined this company. Thus, starting from Scandinavia, with the 

Netherlands following these countries, restructuring occurred in the electricity market 

in almost all of Europe (Dorsman et. al., 2013). 

The European Union also issued many directives for the regulation of the electricity 

market on the creation of a competitive market and the separation of monopolies 

financially and administratively. Expressed here, the electricity market in Turkey in 

2001 with Electricity Market Law No. 4628 has started to join the free market 

transformations without a European Union member because even the member of the 

Union, France, did not have any activities in this regard until 2007. 

1.5.2 Electricity Market of Turkey  

Before the Republican Era, electricity-producing efforts began in 1902 in Tarsus in 

Turkey with a water mill at 2 kW and accelerated after the concessions were granted 

to foreign subsidiaries in 1910. Electricity, which was considered as public goods and 

services in the Ottoman State, was produced and distributed by giving privileges as in 

many other public goods and services. In this period following the II. Constitutional 

(II. Meşrutiyet), privileged foreign companies were established for production and 

sales in some major Anatolian cities.  

After the declaration of the Republic in 1923, the implementation of the privilege of 

foreign companies was continued until the 1930s, and then nationalization and 

confiscation were implemented. The market structure, which operated in this way 

through some state-owned companies, ETIBANK, and municipalities, continued until 

World War II. A new phase that continued until 1960 manifested itself as regional 

privileges given to domestic companies and public companies. 

During the Planned Economic Period, a vertically integrated state-owned monopoly 

Turkish Electricity Authority (TEK) was established in 1971. Except for a few 

domestic privileges given in the 1950s, electricity production transmission and 



22 

 

distribution facilities were gathered in a single company in Turkey. The widespread 

form of organization in the world has been vertically integrated public companies, and 

TEK was organized vertically integrated in the same way by taking the EDF structure 

established in France. 

In the period starting from 1980, it was seen that the laws giving the privilege of 

electricity generation to the organizations outside the TEK were enacted for the 

removal of the TEK monopoly. These applications started as Build-Operate, Build-

Operate, and Transfer and Transfer of Operating Rights, resulted in the enactment of 

the law for the privatization of TEK as a whole in 1993.  However, this law was 

annulled by the Constitutional Court since electricity had strategic importance for 

national security and the possibility of such an arrangement causing cartels. During 

this period, as the increase in electricity consumption could not be met by production, 

TEK had to make regional and hourly mandatory power cuts. After direct privatization 

was failed at this stage, TEK subdivided into two public companies in 1994, and it has 

been separated from generation- transmission (TEAS) and distribution (TEDAS) 

sectors as two parts. Then TEAS (integrated structure of production and transmission) 

was eliminated and three new public companies TEIAS, EUAS, and TETAS, were 

established by the Council of Ministers on 05.02.2001. At this date, also EMRA 

(EPDK) was founded, and the electricity market, and later the natural gas, oil, and LPG 

market, were restructured through regulation.  

Also, the separation in the electricity sector was carried out in two stages. In the first 

stage, the TEK was divided into two, and in the second stage, which started by dividing 

TEAS into three in 2001, the distribution companies affiliated with TEDAS were also 

privatized and separated. Thus, public companies in the electricity sector in Turkey 

were divided into four separate structures. 

The latest market structure is organized as a state monopoly in transmission (TEIAS), 

privileged private oligopoly for distribution, a large number of public (EUAS) and 

private power plants in production, and a large number of retail companies. EPIAS 

operates the wholesale electricity market at any hour of the day in a market organized 



23 

 

as an exchange. All these institutions and companies operating in these sectors are all 

on the market by obtaining a license from EMRA. In this market, tariffs for distribution 

companies and retail companies in Turkey are regulated by the EMRA. These tariffs, 

which are updated every three months, are applied in a uniform standard at the national 

scale by sector. For example, there is an arrangement called a “revenue cap” in the 

transmission sector. Accordingly, after the financing and investment needs required 

by the transmission company are determined, they are carried out with the approval of 

EMRA. As a result, the transmission tariff for generation, distribution companies, and 

companies directly connected to the transmission line are determined. 

However, EPIAS, which is organized as an exchange in the wholesale market, 

determines the system marginal price and market clearing price with future 

transactions for the next day or future. Besides, distribution companies, retail 

companies, or free consumers directly connected to distribution and transmission 

systems can buy and sell electricity at the agreed price with bilateral agreements of 

electricity producer firms. Again, it makes day-ahead agreements with generation 

companies to ensure the system balance for the capacity and ancillary services 

(Balancing Power Market) required by TEIAS.  At present, the share in the installed 

capacity is approximately 68% of private companies, 22% of EÜAŞ, 3% of Build 

Operate and Build Operate Transfer, and the rest is the share of unlicensed power 

plants. Thus the share of the public in the installed capacity has decreased to 22% due 

to privatizations and an increase in private sector investments in the last decade. 

                          Table 3. Plant Ownership in Turkey in March 2021 

Plant Ownership 
Installed 

Power (MW) 

Share 

(%) 
Private Power-Plants 65.675,80 67,66 

EÜAŞ Plants (State) 21.426,60 22,07 

Transferring Operating Rights 2.831,30 2,92 

Build-Operate-Transfer Plants 126,80 0,13 

Unlicenced Plants* 7.007,30 7,22 

TOTAL 97.069,70 100,00 

Source: TEIAS  

* Unlicensed power plants are below 1 MW, do not require any operation 

license from EMRA 

 



24 

 

CHAPTER 2 

ELECTRICITY DEMAND AND FORECASTING 

2.1 ELECTRICITY DEMAND 

As a result of the support of Israel by most of the Western states in respect of the Arab-

Israeli War, the first oil shock had been experienced with the export embargo of oil 

producer Arabic countries. Until then, energy consumption, which took shape 

according to the increasing supply situation, evolved to another dimension with the 

increase in oil prices 3–4 times. Energy supply security has become more critical for 

countries (Laitner et. al., 2003). Energy supply security and increasing reactions and 

restrictions against global warming, and changes in the transactions in competitive 

markets affect the operation of energy markets (Bhattacharyya, 2011). Moreover, the 

increase in energy efficiency affects the demand side of the energy markets and thus 

the supply sector. 

Before 1973, abundant and cheap oil caused no need or little need for local or national 

planning, not only in the oil but also in the electrical, gas, and coal sub-sectors. 

Nevertheless, as there is not a large and cheap oil as in the 1960s, planning has become 

necessary at the national level, whether the market mechanism is working or not. Even 

in the 1990s, in terms of pricing, energy management, and investment, the capacity of 

the country to deal with energy shocks was difficult in the energy and its sub-sectors 

since there was no institutionalization (Berrie, 1992). Like other energy sources, 

electrical energy was also affected by the bottlenecks created by these supply shocks. 

According to this new situation, demand forecasts have to be made with more precise 

and accurate methods to meet the electricity demand. 

In non-OECD countries, electricity using is increasing highly in residential and 

commercial buildings due to income and population growth and spreading access to 

electricity. Also, the use of electricity is growing in the industrial sector as a result of 



25 

 

the expansion in production and the transportation sector with the broadening of 

electric vehicles and electricity-powered subways, and other vehicles (EIA, 2019). 

Demand management changed its’ focus due to various events such as technological 

developments, breakthroughs in communication, improvements in production 

processes, development of better quality with lower costs in the 1990s. The importance 

of demand management has shifted to commercial and industrial demand management 

rather than household consumption. Demand management has also promoted energy 

efficiency for sustainable development due to the industry’s search for lower costs 

(Suganthi and Samuel, 2012). 

As known, the electricity demand is handled much differently than demand for other 

goods, as it cannot be economically stored. Although balance is achieved through 

subsidies or the differences in balance are met by the public in a market dominated by 

state monopolies, there may be some volatility in balance, such as financial problems 

and power outages in a market dominated by the private sector (Cugliari and Poggi, 

2018). A striking example of power outages occurred in California-USA, which started 

in 2000 and continued until 2001. As a result of many market manipulations for 

electricity companies on the stock exchange, deregulations and price quotas, prolonged 

blackouts, and astronomical price increases in the wholesale market were 

observed (Weron, 2006).  

Energy demand management is an important issue for future electrical energy 

planning, selection and prioritization of energy resources, optimization of energy use, 

policy decisions for improvement in energy efficiency, and carbon emission reduction 

(Suganthi and Samuel 2012). However, some difficulties can be seen in developing 

and less developed countries. Bhatia (1986) lists the following problems with forecasts 

in developing countries: (a) using traditional energy is still widespread, (b) many poor 

people not to reach the commercial energy source (c) unsaturated demand pattern and 

(d) height in energy loss-leakage rates.  



26 

 

All of the characteristics above of developing countries largely also possible to see in 

Turkey. Although consumption of traditional primary energy resources is seen in rural 

areas, its share is very low. The slowdown in the demand growth rate compared to 10 

years ago indicates an improvement in saturation. As loss-leakage rates, one of the 

most significant factors affecting electricity demand, tends to decrease, the rate of 

increase in the number of demand decreases.  

Electricity losses-leakages may be due to the technical operation of the system and the 

electrical theft, which is called “electricity pilfered.” This phenomenon is seen 

especially in developing and under-developing countries. Except for technical losses, 

final consumption losses are high due to prevalent reasons such as meter manipulation, 

non-registered connections, and measurement or calculation error, unfortunately, for 

many years in distributing the theft-loss rates at very high levels in Turkey. In this 

losses-leakage rate, also called electrical pilferage, is quite high. Although the loss in 

distribution tends to decrease in recent years, it is still well above the rates of developed 

countries. Regarding losses in electrical energy distribution, it is accepted as 

reasonable loss rates by the American Public Power Association (APPA). APPA takes 

the values in Table-3 as reference.  

                         Table 4. Reasonable Electricity Grid Losses and Leakage Rates 

System 
Losses/Leakages 

(%) 

High/Low Voltage Link 1 

Medium Voltage Link 3,5 

Medium/Low Voltage Link 2,5 

Low Grid and Link 2 

Total 9 

                       Source: American Public Power Association (APPA) 

Pilfered electricity demand is higher ratios in both metered and non-metered 

consumption. Although consumers have meters, they prevent meter control or low 

consumption by interfering with the meters by consumers. Also, usage without meters 

is a common situation, especially in agricultural irrigation.    



27 

 

According to the Figure 4., consumers are accustomed to consuming electricity free of 

charge. In this situation, they consume electricity up to Q0. When these consumers 

start to be charged, their consumption habits will change (lower consumption). They 

will be willing to pay some price (P1), and they will consume lower amounts. 

Electricity providers will save some (Q1-Q0) of the electricity that is already used for 

free, and they will charge a part up to (0-Q1). 

     Figure 4. Pilfered Demand of Electricity 

P 

 

             

 

                                             

    P1                               A                                                                           

                                                            

                                                                                              
        0              Q1                            Q0                            Q                                                                                                                                                                                                     

          Source: Lim and Jenkins (2000)                                              

 

Figure 5. Total Electricity Loss-Leakage Rates 1971-2014) 

 

     Source: Worldbank Electric Power Transmission and Distribution Losses 1 

While the world average of both distribution and transmission losses is around 8%, 

this rate is around 6% on average in OECD countries. As two points in these rates arise 

                                                      
1 Among the OECD countries, lowest total losses-leakage ratio is in Greece, the highest rates, 

respectively, are in Turkey, Mexico and Canada. Losses in Canada mostly occur as a result of the length 

of transmission and distribution lines due to the large area of the country.  

0,00

2,00

4,00

6,00

8,00

10,00

1
9
7

1

1
9
7

3

1
9
7

5

1
9
7

7

1
9
7

9

1
9
8

1

1
9
8

3

1
9
8

5

1
9
8

7

1
9
8

9

1
9
9

1

1
9
9

3

1
9
9

5

1
9
9

7

1
9
9

9

2
0
0

1

2
0
0

3

2
0
0

5

2
0
0

7

2
0
0

9

2
0
1

1

2
0
1

3

Total Electricity Loss-Leakage Rates (%)

OECD World

Willingnes to Pay 

Curtailed Pilfered 

Electricity 

Willgness to Pay 

Retained Pilfered  

Electricity  



28 

 

from the transmission system, the remaining amount is realized at the distribution and 

sales stages. However, this situation reaches incredible dimensions in some countries 

and regions at the distribution stages, and this problem also creates a serious 

inconsistency in terms of correct pricing and real demand. Smith (2004) states that, in 

a developed country such as the USA, it reached 3.5% (approximately $ 10 billion) in 

1999 and it reached 35% in a less developed country like Bangladesh that same year. 

(Smith, 2004) 

The losses-leakages in the transmission system, which has high design standards 

depending on Turkey’s population density and geographical conditions, are at the level 

of 1.75-2% by international performance levels, even if it is better than world averages. 

Also, since the electrical voltage in the transmission system is very high, electrical 

theft or pilfered is almost impossible; losses occur at this stage for technical reasons.  

             Figure 6. Distribution Loss-Leakage Rate in Turkey (1994-2019)2 

 

             Source: Annual Reports of EMRA and TEDAS  

However, it reaches very high rates due to leakages and illegal use during the 

distribution phase. Because this rate in distribution and retail stages has reached of two 

                                                      
2 The Loss-Leakage Ratios were compiled from TEDAS Annual Reports until 2011, the rest were compiled from 

EMRA Annual Reports. However, due to the privatization of TEDAS’s subsidiary companies, there are many 

deficiencies especially in the accuracy of the rates between 2008 and 2012. The loss leakage rates published by 

TEDAS in 2008 and 2009 are around 25%. Moreover, the rates published by EMRA are not exactly correct, since 

calculations made by taking the averages of the percentage rates, The amount of loss and total amounts on the basis 

of the distribution company are not published. Because, while the share of total loss-leakage amount in the 

electricity offered to total consumption should be calculated, EMRA publishes the arithmetic average of the 

percentages of all distribution companies. (https://www.epdk.gov.tr/Detay/Icerik/3-0-24/elektrikyillik-sektor-

raporu) 

15,5

17,3
18,118,7

19,5
20,4

21,621,420,9
19,9

18,6
17,8

15,114,814,4

17,7
18,6

14,615,315,7
14,6

16,4
15,6

14,5
13,112,4

0,0

5,0

10,0

15,0

20,0

25,0

1
9
9

4

1
9
9

5

1
9
9

6

1
9
9

7

1
9
9

8

1
9
9

9

2
0
0

0

2
0
0

1

2
0
0

2

2
0
0

3

2
0
0

4

2
0
0

5

2
0
0

6

2
0
0

7

2
0
0

8

2
0
0

9

2
0
1

0

2
0
1

1

2
0
1

2

2
0
1

3

2
0
1

4

2
0
1

5

2
0
1

6

2
0
1

7

2
0
1

8

2
0
1

9

Distribution Loss-Leakage Rates (%)



29 

 

times the world average and three times the OECD average, it is a big problem for 

Turkey’s electricity demand. Nevertheless, as the loss and leakage rates have been 

decreasing in recent years, it has a decreasing effect on gross demand through meter 

controls and pricing. 

The electricity demand has many aspects, such as consumers’ income and production 

level, the prices of electricity, and the number of electrical appliances. However, it is 

also affected by many temporary factors: changes in weather conditions, power 

outages, and the number of new customer contacts (Lim and Jenkins, 2000). Besides, 

energy demand is not much different from other commodities dealt with in 

microeconomics.  

Nevertheless, the consumer knows and sorts his own preference sets, shows these 

preferences in the utility function, and makes rational choices. That is, the consumer 

depends on his Utility Maximization Problem (UMP). Likewise, producers refer to 

their production function when determining the factor demand required for production: 

Cost Minimization Function (CMP) (Bhattacharyya, 2011). 

However, while the household and industry -services and manufacturing sector, 

consume electricity as a commodity, they also give some reactions. Naturally, both 

households and business sectors pursue economic interests. This interest forms the 

optimum consumption pattern for the household to maximize utility satisfaction in line 

with its income and market prices. Here, the consumer observes her income and 

alternative prices of the goods in her basket. Accordingly, electricity can be replaced 

with natural gas or alternative fuels (for example heating), in areas other than those 

that must be used.  

Additionally, electricity demand is affected by the behavior of consumers and their 

expected income. However, these two factors are very slow to change. Therefore, past 

income and consumer behaviors are important. In other words, consumers react to their 

demands for electricity with a delay (Lim and Jenkins, 2000). 



30 

 

Similarly, companies pursue profit maximization by making choices between the 

production factors and the raw materials that make up the cost item according to their 

production technologies. One condition of profit maximization is provided by cost 

minimization. (Naturally, if productivity growth is kept at least constant). As the share 

of energy costs in total costs is also relatively high, companies prefer energy sources 

according to the prices of the inputs they use and their availability. 

For this reason, energy consumption can vary according to energy demand, input 

prices, income, and output levels in general. It should be noted here that the rate of 

electricity-powered machinery and equipment in companies’ production processes and 

the substitution of electricity. Although the theory of economics emphasizes profit 

maximization according to the tastes of households and production technologies of 

firms, it is limited by psychological and physical factors. Thus, it is not very easy to 

substitute due to the increasing use of modern transportation, home applications, and 

easy access. Therefore, we can say that utility maximization is a bit more important 

than cost minimization.  

While some socioeconomic, demographic factors and climatic conditions affect the 

consumption pattern, and the building architecture, engineering features of the 

building, power appliances and tools existing at the residence, and energy 

infrastructure of the building may not allow direct substitution effects on consumer 

preferences (Gellings, 1996). Because, if there is an electric stove installed in the 

house, and there is no natural gas infrastructure, for example, heating is provided only 

with electric heaters. On the contrary, in a house with both the electric stove and natural 

gas infrastructure, it changes the consumer position according to price fluctuations. 

Typically, while changes in income level and prices of goods affect the consumer’s 

bundle of choice, this is not so easy and speedy with electricity. In this frame, it can 

be even said that electricity is not a flexible commodity. 

As can be seen from the table below, the share of electricity consumption has increased 

in the last 50 years, while the consumption of coal, which is an alternative product, has 

decreased, and natural gas consumption has remained stable at 13-15%. In this 



31 

 

phenomenon, besides the peculiarities of electrical goods, the widespread use of 

electrical appliances and equipment and efficient technologies have a significant role. 

The increasing use of electricity will expand, and it will become a final product for 

which substitution is almost impossible in the long run. 

             Figure 7. The Share of Selected Energy Sources in World Total Consumption 

 

                   Source: IEA, Headline Global Energy Data, 20193 

2.2 ELECTRICITY DEMAND FORECASTING 

In fact, since electricity was widely used in lighting for the first time, there was not 

much need for load and demand estimation. Because the number of lighting lamps in 

the network was known, it was also known how much load would be in the evening 

hours. However, the widespread use of electrical appliances and the use of electricity 

in the industry made this work a little more complicated. Especially in the 1940s, the 

prevalence of air conditioner use caused jumps in daily load curves due to the increase 

in air temperatures and humidity in the USA. Since then, one of the most essential 

tools for electricity demand forecasting has been weather forecasting. (Hong and 

Shahidehpour, 2015) 

Throughout the 1950s and 1960s, interest in demand forecasts was not very common. 

After World War II, which was still growing steadily, the high demand has led many 

                                                      
3 Although the share of coal in the world has decreased by 2000, the increase in the consumption of coal 

after 2000 has a great role in the increase in China’s coal demand. In 2017, 65% of the World’s coal 

consumption is made by China alone. The demand for coal in the cement and steel industry is still high. 

0,00

5,00

10,00

15,00

20,00

1
9
7

1

1
9
7

3

1
9
7

5

1
9
7

7

1
9
7

9

1
9
8

1

1
9
8

3

1
9
8

5

1
9
8

7

1
9
8

9

1
9
9

1

1
9
9

3

1
9
9

5

1
9
9

7

1
9
9

9

2
0
0

1

2
0
0

3

2
0
0

5

2
0
0

7

2
0
0

9

2
0
1

1

2
0
1

3

2
0
1

5

2
0
1

7

Total Final Consumption Share by Source (%)

Electricity NG Coal etc.



32 

 

to think that the extrapolation methods were largely sufficient for planning purposes. 

While the over-estimation for the future was quickly corrected in the next period due 

to rapid demand increases, underestimating was ignored because the new baseload 

power plants were activated due to cheap oil and gas (Électra, 1986). 

Forecasting, which we encounter in all areas of economics and finance, occurs in terms 

of the value that households or firms can take one or more future variables. As known, 

households determine their labor supply according to their expectations of wages and 

savings, while firms decide on their future investment decisions according to expected 

cash flows and interest rates. Similarly, large infrastructure investments in public 

finance are also important to achieve the expected income flow and the targeted goal 

throughout the project's life. (Elliot and Timmermann, 2008). It can even be said that 

forecasting is at the heart of planning for both the power sector and the actors on the 

electricity demand side (Berrie, 1992). 

Many different organizations can make both generation (supply) and consumption 

(demand) forecasting of electrical energy. Industrial companies that consume large 

amounts of electricity, electricity generation, transmission, distribution, or retail 

companies can also make predictions for their own purposes. Forecastings are used for 

various purposes ranging from the real-time operation of power generation plants to 

determining required long-term generation, transmission, and distribution 

development plans. 

Thus, forecasting in the electricity sector is important not only for generation decisions 

and transmission-distribution capital investments but also for financial forecasting, 

fuel purchasing-storing, and using in production, capacity, and reserve planning and 

implementation. Therefore, forecasting is important for electricity generation, 

transmission, distribution, and retail sales companies to make precise and accurate 

predictions (Electra, 1992). 

Electricity generation, transmission, and distribution sub-sectors need high capital 

requirements. Also, since the projects in these sectors are very large, the projects take 



33 

 

a long time. Again, since many years pass from the planning to the construction of 

these projects, it takes a long time to start production too. Therefore, unclear and non-

indicative demand forecasts create a cost regarding making and timing investment 

decisions (CIGRÉ, 2006). Since electricity investments are long-lasting investments 

that depend on technology and need great financing, good planning is required. 

Whether in the public or private sector, good planning depends on making the forecasts 

correctly and meeting the needs. 

Two different types generally make demand forecasting in the electrical industry of 

organizations. The first type of organization is vertically integrated or discrete 

distribution and transmission system operators, generation companies, and electricity 

market actors. Demand forecastings can be range from very short to long term, are 

used for system requirements and network capacity, and operational purposes. The 

second type of organization provides data and information to actors in the first group, 

such as regulators and government agencies (CIGRÉ, 2016). 

The first step in prediction studies is to reach past and present data. Accordingly, 

outputs are obtained by determining the most suitable analysis method. It is necessary 

to establish the right model, which is mandatory to find the lowest deviations in the 

energy or specifically electricity demand forecast process. This process, which does 

not differ much from the stages of the model that we encounter in econometric models, 

increases the reliability of the model. 

 

    Figure 8. Basic Stages of Forecasting 

 
Source: Debnath and Mourshed (2018) 

 

Gellings et. al. (1996) lists in order of seven steps for forecasting as follows;  



34 

 

- The correct definition of the goal: This first step involves defining the 

intended purpose of the forecasting. Accordingly, the purpose and objectives of the 

model can be classified systematically in the economic framework. 

- Identification of the model: In the nature and structure of the economic 

analysis, it is necessary to obtain access data to achieve the purpose of the model 

established and to produce the assumptions desired. 

- Data compilation and review: Careful review of the data to be used in the 

model and order by purpose occurs at this stage. 

- Choice of forecasting method: Since there are many models for energy 

demand forecasts, it is necessary to analyze which model should be chosen well. The 

developing technology and literature also diversify the methods to be used and for the 

intended purpose. 

- Forecast Development-revisions: This phase includes the processes such as 

control of the data, whether alternative data should be used, and reviewing the data. 

Thus, after this, the model can be verified and evaluated. 

- Model Evaluation and Verification: Statistical and parametric tests are used 

at this stage. Forecast errors are calculated accordingly.  

- Forecast Documentation: this stage is important to explain and understand the 

stages we have mentioned above in the future. 

Besides, it will be necessary to consider factors such as electric vehicles, roof panel 

systems that can cause radical changes in electricity consumption because the changes 

in battery technology may affect the consumption amounts as well as the consumption 

patterns. 

CIGRÉ C1-32 Working Group states that the following conditions should be taken 

into account for forecasting electric power and demand (Electra 290, 2017): 

-   Electricity price elasticity for consumers  

-   New production possibilities such as rooftop solar panels 



35 

 

-   Government incentives for energy efficiency 

-  Identifying very small pre-existing power generation facilities for actual 

demand 

-  Government or regulatory agency effects on tariffs: such as ceiling price- price 

limit or instant pricing 

-  Uncertainty in economic variables; changing the demand pattern of exchange 

rates, growth rates, households, or industries that use electricity 

Naturally, those that can be measured among these elements can be included in a 

demand forecast model. At the same time, externally non-measurable variables are 

tried to be included in the model according to their effects, although it is not clear. 

Electricity demand forecasting can be made for a certain time interval (short, medium, 

and long term) as well as for countries, regions, and the whole world in a certain 

geographical area (Debnath and Maurshed, 2018). 

Electricity demand forecastings are divided into three parts in terms of time horizon: 

(i) Short-term, (ii) Medium-term, and (iii) Long-term forecasting (Al-Alawi,1996). 

          i.   Short-term Forecasting: The short-term forecast includes electricity demand 

estimates from the one-hour to one month. Hourly and daily forecasts are mostly 

related to electricity transmission system operation and production planning (such as 

fuel and resource preference). Although some sources also take a “very short time” 

separation from minutes to a few hours, I think it would be useful to have this 

distinction under the short time heading.  

         ii.    Medium-term Forecasting: The medium-term forecast lasts from monthly to 

a few years’ forecasts. It includes some social activities and socio-economic factors 

such as sports events (Olympic games etc.), religious or national holidays, and also 

covers generation planning and plant maintenance. 



36 

 

       iii.    Long-term Forecasting: This period requires a forecasting mechanism, from 

5 to 25 years, to cover all processes from electricity generation to distribution. 

As mentioned in the definitions of short-term and mid-term forecasts, these are 

important for demand fluctuations that may occur mostly due to meteorology and 

random effects such as social events, holidays, and some sports events. Instantaneous 

electrical system balance and working order of the power plants (even maintenance 

and repair activities) are important in these periods. Transmission service 

organizations (TSOs) perform these balancing tasks. That is, technically, balancing 

electricity supply and demand is carried out by TSO.  

Meteorological, economic, cultural, and some special factors affect electricity 

cons