6c Bottom-Up Analysis of Energy Demand PDF

IØ8303: Energy Markets Lecture 6c – Bottom-Up Analysis of Energy Demand Prof. Dr. rer. soc. oec. Reinhard Madlener Lecture Outline 1. Key Issues 2. Basic Concept of ProcessAnalysis 3. Inventory of Appliances, Buildings, Vehicles and Machineries 4. Short-term Determinants of Energy Demand 5. Energy...

IØ8303: Energy Markets Lecture 6c – Bottom-Up Analysis of Energy Demand Prof. Dr. rer. soc. oec. Reinhard Madlener Lecture Outline 1. Key Issues 2. Basic Concept of ProcessAnalysis 3. Inventory of Appliances, Buildings, Vehicles and Machineries 4. Short-term Determinants of Energy Demand 5. Energy Efficiency 6. Innovation and Technical Progress 7. Conclusions 2 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 1. Key issues  What is the basic concept of process analysis?  How are short-term determinants of energy demand influenced and what forecasting methods are there?  How can energy efficiency be defined?  What is the relationship between energy efficiency and energy saving?  What is rebound effect?  Why are there market failures in energy efficiency investments?  How does the concept of contracting work? 3 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 2. Basic Concept of Process Analysis (1/6) Starting point: The demand for each type of energy per unit of time depends on three factors: (1) Energy-using capital stock (appliances, buildings, machinery, vehicles) (2) Intensity of use of this capital stock (e.g. km driven per month) (3) Energy efficiency (e.g. liters of gasoline per 100 km driven) Differentiation of the aggregated energy demand: − by energy source (electricity, heating oil, natural gas, gasoline, diesel, hydrogen, biomass..) − by energy customer (industry, households, small consumers, transport) − but also according to consumption areas / sectors (low / high temperature heat, work, lighting, electrolysis) 4 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 2. Basic Concept of Process Analysis (2/6) Long-term Desired stock Efficiency factors Cap*(t) improvement Age structure Given stock Deviation Adjustment (Capital- Cap(t-1) Cap*(t)-Cap(t-1) ∆Cap(t, t-1) Vintage) Short-term Intensity of use Final energy factors ν(t) demand E(t) Source: Erdmann, Praktiknjo, Zweifel (2017). p. 67 5 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 2. Basic Concept of Process Analysis (3/6) Energy demand E(t) is a function of the stock of energy-using capital Cap(t) and the intensity of use v(t) Net investments is given: ∆𝐶𝐶𝐶𝐶𝐶𝐶 𝑡𝑡 = 𝛼𝛼 𝐶𝐶𝐶𝐶𝐶𝐶∗ 𝑡𝑡 − 𝐶𝐶𝐶𝐶𝐶𝐶 𝑡𝑡 − 1 , 0 < 𝛼𝛼 < 1 (1) Replacements: Belong to net investments Do not change the total stock of appliances, Cap, but energy efficiency and thus energy demand 6 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 2. Basic Concept of Process Analysis (4/6) Cohort ("vintage") model – explicit consideration of the age structure of the equipment stock, i.e.: 𝐶𝐶𝐶𝐶𝐶𝐶𝑖𝑖 𝑡𝑡 = 1 − 𝛿𝛿𝑖𝑖−1 𝐶𝐶𝐶𝐶𝐶𝐶𝑖𝑖−1 𝑡𝑡 − 1 , 0 < 𝛿𝛿𝑖𝑖−1 < 1 𝐶𝐶𝐶𝐶𝐶𝐶1 𝑡𝑡 = ∆𝐶𝐶𝐶𝐶𝐶𝐶 𝑡𝑡 − 1 (𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑜𝑜𝑜𝑜 𝑡𝑡𝑡𝑡𝑡 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦) 𝐶𝐶𝐶𝐶𝐶𝐶 𝑡𝑡 = 𝐶𝐶𝐶𝐶𝐶𝐶𝑖𝑖 𝑡𝑡 𝑖𝑖 δi-1 … rate of depreciation pertaining to a particular vintage i. 7 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 2. Basic Concept of Process Analysis (5/6) Long-term Desired stock Efficiency factors Cap*(t) improvement Age structure Given stock Deviation Adjustment (Capital- Cap(t-1) Cap*(t)-Cap(t-1) ∆Cap(t, t-1) Vintage) Short-term Intensity of use Final energy factors ν(t) demand E(t) Source: Erdmann, Praktiknjo, Zweifel (2017). p. 67 8 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 2. Basic Concept of Process Analysis (6/6) Factors affecting the demand for energy: Long-term factors − Demographic and sociological variables (household size and composition, commuting distances, and lifestyle) − Economic factors (business sales, disposable income and wealth, the rate of interest as a component of capital user cost, and the price of energy relative to other goods and services) Short-term factors − Affect the intensity with which the stock of capital is used (the business cycle, and calendar effects, but also fluctuations in income and energy prices) Improvement of energy efficiency: Technological change, Government policy, Conscious investment / purchase decisions 9 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 3. Inventory of Appliances, Buildings, Vehicles and Machineries (1/7) Indicators of energy demand Source: Erdmann, Praktiknjo, Zweifel (2017). p. 69 10 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 3. Inventory of Appliances, Buildings, Vehicles and Machineries (2/7) Empirical modelling and investigation of the electricity demand of private households (stock of electrical appliances etc.) Equipment inventory as a product of − Number of households (demographic) − and ownership probability 𝑋𝑋𝑛𝑛 (influenced by energy industry and political variables) 𝑋𝑋𝑛𝑛 = 1 (household n owns the type of device) 𝑋𝑋𝑛𝑛 = 0 (household n does not own the type of device) Ownership probability influenced by individually achievable net benefit 11 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 3. Inventory of Appliances, Buildings, Vehicles and Machineries (3/7) Subjective, individual utility depends on several objectively measurable factors, such as Dj : Annual energy demand − Total Cost of Ownership 𝑇𝑇 Annual operating costs 𝑝𝑝𝐸𝐸,𝑡𝑡 𝐸𝐸 + 𝑂𝑂𝑂𝑂 𝐶𝐶 = 𝐼𝐼𝐼𝐼𝐼𝐼 + (2) (1 + 𝑖𝑖)𝑡𝑡 𝑡𝑡=1 − Linear utility V for household n 𝑉𝑉𝑛𝑛 = 𝛽𝛽0 + 𝛽𝛽𝑗𝑗 𝐷𝐷𝑗𝑗,𝑛𝑛 (3) 𝑗𝑗 Subjective, individual utility depends on non-measurable influencing variables: − Random utility model 𝑈𝑈𝑛𝑛 = 𝑉𝑉𝑛𝑛 + 𝜀𝜀𝑛𝑛 = 𝛽𝛽0 + 𝛽𝛽𝑗𝑗 𝐷𝐷𝑗𝑗,𝑛𝑛 + 𝜀𝜀𝑛𝑛 𝑗𝑗 Systematic (deterministic) benefit component Unsystematic (stochastic) benefit component 12 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 3. Inventory of Appliances, Buildings, Vehicles and Machineries (4/7) Logistic function for modeling ownership probability Probability w w=1 w=0 Net utility U(i) Source: Erdmann, Praktiknjo, Zweifel (2017). p. 70 13 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 3. Inventory of Appliances, Buildings, Vehicles and Machineries (5/7) Probability of ownership (logistic function) - as a logit model: 𝑒𝑒 𝑈𝑈𝑛𝑛 1 𝑤𝑤 𝑋𝑋𝑛𝑛 = 1 = 𝑤𝑤 𝑈𝑈𝑛𝑛 > 0 = = (5) 1 + 𝑒𝑒 𝑈𝑈𝑛𝑛 1 + 𝑒𝑒 −𝑈𝑈𝑛𝑛 The logit transformation of the ownership probability w is a linear function of the utility index U: 𝑤𝑤 ln = 𝑈𝑈 = 𝛽𝛽0 + ∑𝑗𝑗 𝛽𝛽𝑗𝑗 𝐷𝐷𝑗𝑗,𝑛𝑛 (6) 1−𝑤𝑤 Solution of the logit model by maximizing the log-likelihood function 𝑙𝑙𝑙𝑙𝑙𝑙 𝛽𝛽0 , 𝛽𝛽1 , … 𝑋𝑋, 𝐷𝐷1 , 𝐷𝐷2 , … = 𝑋𝑋𝑛𝑛 𝑙𝑙𝑙𝑙𝑤𝑤𝑛𝑛 + 1 − 𝑋𝑋𝑛𝑛 ln 1 − 𝑤𝑤𝑛𝑛 (7) 𝑛𝑛 1 with 𝑤𝑤𝑛𝑛 = 1+𝑒𝑒 −𝑉𝑉𝑛𝑛 14 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 3. Inventory of Appliances, Buildings, Vehicles and Machineries (6/7) Marginal impact of independent variables 𝐷𝐷𝑗𝑗 on ownership probability w: 𝜕𝜕𝑤𝑤 1 −𝑉𝑉 1 𝑒𝑒 −𝑉𝑉 =− −𝛽𝛽𝑗𝑗 𝑒𝑒 = 𝛽𝛽𝑗𝑗 = 𝛽𝛽𝑗𝑗 𝑤𝑤 1 − 𝑤𝑤 (8) 𝜕𝜕𝐷𝐷𝑗𝑗 1 + 𝑒𝑒 −𝑉𝑉 2 1 + 𝑒𝑒 −𝑉𝑉 1 + 𝑒𝑒 −𝑉𝑉 at w=0.5 the variable 𝐷𝐷𝑗𝑗 has the largest marginal influence at w=1 or w=0 the marginal influence of variable 𝐷𝐷𝑗𝑗 is zero The changes in probability w in percentage points with a change in the influencing variable 𝐷𝐷𝑗𝑗 by one percent – semi-elasticity: ∗ 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 η = = 𝐷𝐷 = 𝛽𝛽𝑗𝑗 𝐷𝐷𝑗𝑗 𝑤𝑤 1 − 𝑤𝑤 (9) 𝜕𝜕𝐷𝐷𝑗𝑗 /𝐷𝐷𝑗𝑗 𝜕𝜕𝐷𝐷𝑗𝑗 𝑗𝑗 15 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 3. Inventory of Appliances, Buildings, Vehicles and Machineries (7/7) Other models: The multinomial logit model (for K alternatives) - the alternatives available are independent of one another The nested logit model – the alternatives available depend on each other exp (V1 ) w( x = 1) = exp(V0 ) + exp(V1 ) exp (V2 ) w( x = 2) = ⋅ w( x = 1) No car exp(V1 ) + exp(V2 ) One car One car Two cars Source: Erdmann, Praktiknjo, Zweifel (2017). p. 74 16 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 4. Short-term Determinants of Energy Demand (1/5) Short-term factors which affect the intensity of the demand for energy The intensity of use (grid-bound energy source) is significantly influenced by: Time-of-day effects Calendar effects (weekdays, holidays,...) Seasonal effects Temperature and radiation fluctuations, as well as the weather Production fluctuations (strikes, production losses,...) Energy industry control parameters (load management) Special events (major sporting events, festivities,...) Energy price fluctuations Economic trend 17 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 4. Short-term Determinants of Energy Demand (2/5) Example of a weekly electricity delivery schedule Power [MW] Sunday Monday Tuesday Wednesday Thursday Friday Saturday Source: Erdmann, Zweifel (2007). p. 74 18 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 4. Short-term Determinants of Energy Demand (3/5) Open supply contracts (no timetable, call option) Versus TOP (Take-or-Pay) – contracts, sanction in the event of shortfall in procurement Forecasting method: Time series analysis, (artificial) neural networks, cluster analysis (today: numerous commercial tools) 19 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 4. Short-term Determinants of Energy Demand (4/5) Modelling of non-linear relationships between indicator and energy demand Et, e.g. (outdoor) temperature and electricity demand: 𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡 = max 0, 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑡𝑡 − 18 𝐻𝐻𝐻𝐻𝐻𝐻𝑡𝑡 = max(0, 18 − 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑡𝑡 ) 2 4 𝐻𝐻𝐻𝐻𝐻𝐻𝑡𝑡 + 𝐻𝐻𝐻𝐻𝐻𝐻𝑡𝑡−1 𝐶𝐶𝐷𝐷𝐷𝐷𝑡𝑡 + 𝐶𝐶𝐷𝐷𝐷𝐷𝑡𝑡−1 𝐸𝐸𝑡𝑡 = 𝛼𝛼 + 𝛽𝛽 + 𝛾𝛾 +⋯ (10) 2 2 The parameters 𝛼𝛼, 𝛽𝛽 𝑢𝑢𝑛𝑛𝑑𝑑 𝛾𝛾 - can be determined with regression analysis. 20 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 4. Short-term Determinants of Energy Demand (5/5) Non-linear relationship temperature - energy demand Demand growth [%] 10 8 6 4 2 0 -15 -10 -5 0 5 10 15 20 25 30 35 Temperature [°C] Source: Erdmann, Zweifel (2007). p. 76 21 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (1/32) Definitions Economic perspective (optimality – efficiency – productivity) "Optimality is given when the preferences of the demanders are served in the best possible way." Energy industry perspective – productivity energy use Energy efficiency measurement: 𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜  Thermodynamic efficiency η= 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝑦𝑦 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜  Exergetic efficiency η= 𝐸𝐸𝑥𝑥𝑒𝑒𝑒𝑒𝑒𝑒𝑦𝑦 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 Exergy: quantity of energy that can be converted to work (rather thanheat) 22 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (2/32) One-dimensional – increased input of other production factors not considered 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 − 𝑘𝑘𝑘𝑘 𝑆𝑆𝑆𝑆𝑒𝑒𝑒𝑒𝑙𝑙 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑐𝑐𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 , , 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝑦𝑦 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝑦𝑦 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝑦𝑦 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 Monetary efficiency indicators, to express scarcity : 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 , 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝑦𝑦 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝑦𝑦 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 Inverse of energy efficiency indicators energy intensities From a normative perspective, energy efficiency means conversion of energy with the lowest possible losses 23 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (3/32) Energy efficiency vs. Energy saving Energy efficiency can be increased by substitution processes: Substitution of energy by human capital (improvement of energy management) Substitution of energy by capital (use of equipment with improved energy efficiency) Improving energy efficiency differs from: Forced energy saving (often a consequence of crises) Price increases for energy sources (voluntary reduction in demand) Voluntary renunciation of energy services (changed values) 24 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (4/32) Efficiency potential from an economic point of view Cost-optimized provision of energy services (Least Cost Planning) from a business or economic point of view Specified parameters: Additional investment costs – ∆𝐼𝐼 Annual energy savings – ∆𝐸𝐸 Energy purchase costs – 𝑝𝑝𝐸𝐸 PVF and useful life −∆𝐼𝐼 𝐴𝐴𝐴𝐴 = + 𝑝𝑝𝐸𝐸 ∆𝐸𝐸 > 0 (11) 𝑃𝑃𝑃𝑃𝑃𝑃𝑖𝑖,𝑇𝑇 The most economically attractive variant has the lowest annuity (AN) 25 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (5/32) Energy efficiency: economist vs. engineer perspective Example: Thermal insulation of buildings Cost of Conserved Energy ∆𝐼𝐼 1 𝐶𝐶𝐶𝐶𝐶𝐶 = < 𝑝𝑝𝐸𝐸 (12) ∆𝐸𝐸 𝑅𝑅𝑅𝑅𝑅𝑅𝑖𝑖,𝑇𝑇 CCE Cost optimum Thickness of thermal insulation A C Thermodynamic efficiency maximum B Engineer‘s perspective Economist‘s perspective Source: Erdmann, Praktiknjo, Zweifel (2017). p. 80 26 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (6/32) Theoretical and achievable efficiency potentials Cost on conserved energy CCE Market penetration, [EUR/MWh] Rebound Engineering, „Theoretical“ transaction cost potential B (economist Energy price pE [EUR/MWh] perspective) Realistic A potential Cumulated energy demand reduction [MWh/a] Source: Erdmann, Praktiknjo, Zweifel (2017). p. 81 27 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (7/32) Efficiency potentials The following potential terms can be distinguished: Theoretical potential (limited by the laws of physics) Technical potential (limited by current technical feasibility) Economic potential (investment cost) Socio-economic potential (norms, attitudes,...) Reasons for lower actual achievable savings compared to the economic potential: Transaction costs Persistence or persistence effects Rebound effects 28 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (8/32) Energy efficiency and market failure Further reasons why efficiency potentials are not used despite cost advantages? Irrelevance of efficiency-enhancing measures from an individual perspective Divergence of decision-making powers (investor-user dilemma) Short-sightedness of energy customers (high implicit discount rate) Comparison of investment alternatives 29 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (9/32) Energy efficiency and market failure Sample calculation of an investment into energy efficiency Source: Erdmann, Praktiknjo, Zweifel (2017). p. 84 30 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (10/32) Derivation: Cost of Conserved Energy (CCE) − Difference project: NPVA – NPVB < 0; B = more energy-efficient project (−I A − pE ⋅ E A ⋅ RBFi,T )− (−I B − pE ⋅ EB ⋅ RBFi,T )< 0 (−I A + I B ) − (E A − EB )⋅ pE ⋅ RBFi,T 0 purchase costs (€/kWh) RBFi,T ∆E: annual energy savings (kWh/a) ∆I 1 CCE = ≤ pE → Minimal cost of energy saved ∆E RBFi,T 31 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (11/32) Why are energy efficiency potentials, despite their cost-competitiveness, often not exploited? − Irrelevance of the measures from an individual perspective  Consumers (larte and small ones) have scarce knowledge about the possibilities for energy efficiency improvements. Achievable cost advantages do not have much weight for many consumers. − Splitted decision power (Investor-User or Landlord-Tenant Dilemma)  Often the economic gains do not accrue to the investor, causing disinterest in high investment costs (landlord does not want to isolate homes so that the tenant enjoys lower heating costs) − Myopia of energy customers: High subjective discount rates of customers  In the energy business long payback periods are common (investments in long-lived equipment). Customers often only want to invest in projects with short payment periods (several months or years only) 32 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (12/32) What is a subjective discount rate? − The subjective discount rate is a measure for the present preference (or future orientation, resp.) of an individual. The present preference is the stronger the larger the discount rate is. − Subjective discount rates can be computed implicitly by comparing investment alternatives (difference project) and the purchase decision of an individual. Implicit discount rate: − Consumers invest too little in energy-saving measures due to other market barriers, and thus act such that they seem to expect a markedly higher return on their energy saving investment (“implicit interest rate”). − often > 50% ( rarely the result of an explicit profitability analysis) − “Patch” for the aggregate representation of a whole bundle of market barriers regarding the decision behavior of energy consumers against energy efficiency (e.g. lack of knowledge of technical energy options, lack of knowledge regarding economic decision-making tools for valuating investment in private households, lack of interest etc.). 33 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (13/32) Task #1: Comparison of investment alternatives (Refrigerator) − You want to buy a new refrigerator at the beginning of the year − 2 models, A und B, are shortlisted: Refrigerator model A B Difference Purchase price [€] 800 1,000 200 Operating costs [€ p.a.] 600 300 300 − Period of observation: 1 year − Assumption: Operating costs are billed after one year. − At what implicit discount rate or for what present preference you decide in favor of model B? 34 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (14/32) Task #1: Solution ! 600 ! 300  A = B → −800 − = −1000 − (1+ r ) (1+ r ) 800 + 600 ! 300 (300 −600) =1000 + → 800 −1000= (1+ r ) (1+ r )  (1+ r ) 1  = 2 3 → 1+ r = 3 2 (1+ r )  Diskontrate r = 1 2 = Discount rate = 50% 35 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (15/32) Task #1: Interpretation − The additional (energy efficiency) investment yields a return of 50% (return on investment, ROI) − At a return of investment of 50% the client is just indifferentj ( if his/her present preference is < 50 %, then the device B is adopted) − The smaller the required return on investment demanded, the more likely the client invests in the energy efficiency measure. − Discount rates of this magnitude are no exaggeration, but can be observed for purchases of household devices (e.g. discount rates of 243.2% for electrical water boilers have been found in empirical research!) 36 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (16/32) Task #2: Implicit Discount Rate and Cost of Conserved Energy − A private household has to take the purchase decision between a conventional device (A) and an energy-efficient device (B) under the following conditions: Conventional Energy-efficient device A device B Investment costs [€] 20,000 22,830 Annual energy consumption [kWh/a] 13,000 8,500 Electricity price pE [€/kWh] 0.15 0.15 Expected lifetime [a] 10 10 Calculatory interest rate [%] 10 10 a) Assume that the household will decide in favor of device A. What is the implicit discount rate of the household? b) What is the Cost of Conserved Energy (CCE)? Interpret the result! 37 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (17/32) Task #2: Annuity present value factors 38 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (18/32) Task #2: Solution (i) − ∆I ! a) Computation of the implicit discount rate using the annuity: AN = + p E∆E =0 RBFi,T ∆I 2.830 RBFi,T = = = 4,192 = RBFi,10 → see Table → i = 20% pE ⋅ ∆E 0,15⋅ 4.500 ∆I 1 b) Cost of Conserved Energy (CCE): CCE = ≤ pE ∆E RBFi,T 2830 1 CCE = = 0,102 4500 6.1446 CCE = 0.102 < 0.15 = pE → It pays off to invest in the energy saving measure! 39 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (19/32) Energy efficiency projects are often not realized despite of high expected returns on investment. Is such acting necessarily irrational? − Due to credit rationing often investment priorities have to be imposed → Energy efficiency often considered as not being a core issue or problem − Energy does not belong to the core business of many enterprises → Uncertainty − Uncertainty regarding the actual achievement of energy savings → Information asymmetries − If the individual’s period of use < economic period of use (e.g. devices owned by students) → Payback period too long − Solution (among others): Treatment of energy efficiency measure as a real option  Expectation of future technical improvements. By forfeiting an option is created to invest in superior future energy efficiency measures 40 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (20/32) Task #3: Energy efficiency measure as a real option − You are confronted with the decision to replace an old with a more energy-efficient machine. Today, in t0, there is a Model A available. You have information that in the foreseeable future a more energy-efficient model will become available on the market. However, there is uncertainty regarding the timing. It is assumed that the probability of emergence is 50% that it appears in period t2 and t3, respectively. Model A in t0 Model B in t2 Model B in t3 Purchase price [€] 20,000 20,000 20,000 Resale value olf machine 15,000 13,000 12,000 Annual energy consumption [kWh/a] 8,500 6,000 6,000 − The energy consumption of the old machine amounts to 13,000 kWh/a. The market interest rate is 4%, the technical lifetime of both machines 10 a, and the energy price is assumed to be constant over the lifetime at 0.15 €/kWh. − Assume further that Model B for the foreseeable future is the last innovation, that its price remains unchanged, and abstract from considerations after period t14. − Compute the costs and compare them with the value of waiting? Would you, on this basis of the information presented, decide in favor of machine A or B? 41 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (21/32) Task #3: Solution approach (i) − Are the cost of waiting larger / smaller / equal to the value of waiting? (a) Cost of waiting:  Scenario B‘ (Machine B appears in t2 on the market): (p = ½) - Higher energy consumption of the old machine in periods t0 and t1 - Lower resale value (RV) of the old machine after two periods (only € 13,000 instead of € 15,000)  Scenario B‘‘ (Machine B appears in t3 on the market): (p = ½) - Higher energy consumption of the old machine in periods t0 until t2 - Lower RV of the old machine after three periods (only € 12,000 instead of € 15,000) t t0 t1 t2 t3 t4 t5 t6 t7 t8 t9 t10 t11 t12 t13 t14 A (p = ½) B‘ (p = ½) B‘‘ 42 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (22/32) Task #3: Solution approach (ii) − Are the cost of waiting larter / smaller /equal to the value of waiting? (b) Value of waiting:  Scenario B‘: (p = ½) - Lower energy consumption from t2 until including t9 - Later re-investment (twice) → t2 instead of t0 and t12 instead of t10  Scenario B‘‘: (p = ½) - Lower energy consumption from t3 until including t9 - Later reinvestment (twice) → t3 instead of t0 and t13 instead of t10 t t0 t1 t2 t3 t4 t5 t6 t7 t8 t9 t10 t11 t12 t13 t14 A (p = ½) B‘ (p = ½) B‘‘ 43 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (23/32) Task #3: Solution (iii), Cost of waiting – Scenario B‘ (p = ½) − Higher energy consumption of the old machine in periods t0 and t1  Difference between consumption of the old machine and Model A: ∆E ⋅ pE ⋅ RBFi,T = (13,000 − 8,500) ⋅ 0.15⋅ RBF4%,2 = 4,500⋅ 0.15⋅ RBF4%,2 = 675⋅1.8861 = 1,273.12 − Lower and later resale value (RV) of the old machine  € 13,000 in t2 instead of € 15,000 in t0: 13,000 15,000 − = 2,980.77 (1.04 )2 − Cost of waiting = €1,273.12 + €2,980.77 = €4,253.89 44 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (24/32) Task #3: Solution (iv), Cost of waiting – Scenario B‘‘ (p = ½) − Higher energy consumption of the old machine in periods t0 until t2  Difference between consumption of the old machine and machine A: ∆E ⋅ pE ⋅ RBFi,T = 4,500⋅ 0.15⋅ RBF4%,3 = 675⋅ 2.7751 = 1,873.19 − Lower and later resale value (RV) of the old machine  € 12,000 in t3 instead of € 15,000 in t0: 12,000 15,000 − = 4,332.04 (1.04 )3 − Cost of waiting = €1,873.19 + €4,332.04 = €6,205.23 45 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (25/32) Task #3: Solution (v), Value of waiting – Scenario B‘ (p = ½) − Lower energy consumption from t2 until t9.  Difference between consumption of Models A and B: Machine A : E ⋅ p E ⋅ (RBF4%,10 − RBF4%,2 )= 1,275⋅ 6.2248 = 7,936.62 Maschine Machine B : E ⋅ p E ⋅ (RBF4%,10 − RBF4%,2 )= 900 ⋅ 6.2248 = 5,602.32 Maschine A − B = 2,334.3 − Later reinvestment  Reinvestment only in t2 instead of in t0 and in t12 instead of int10: 20,000 20,000 20,000 20,000 − = 1,508.88 & − = 1,019.34 (1.04)2 (1.04) (1.04) 10 12 − Value of waiting = €4,862.52 46 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (26/32) Task #3: Solution (vi), Value of waiting – Scenario B‘‘ (p = ½) − Lower energy consumption from t3 until t9.  Difference between consumption of Models A and B: A → E ⋅ p E ⋅ (RBF4%,10 − RBF4%,3 )= 1,275⋅5.3358 = 6,803.15 B → E ⋅ p E ⋅ (RBF4%,10 − RBF4%,3 )= 900 ⋅5.3358 = 4,802.22 A − B = 2,000.93 − Later reinvestment:  Reinvestment only in t3 instead of in t0 and in t13 instead of int10: 20,000 20,000 20,000 20,000 − = 2,220.07 & − = 1,499.80 (1.04)3 (1.04) (1.04) 10 13 − Value of waiting = €5,720.80 47 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (27/32) Task #3: Solution (vii), comparison − Cost of waiting: 1 ⋅ 4,253.89 + 1 ⋅ 6,205.23 = 5,229.56 2 2 − Value of waiting: 1 ⋅ 4,862,.52 + 1 ⋅5,720.8 = 5,291,.66 2 2 − € 5,229.56 < € 5,291.66 − Cost of waiting < Value of waiting → Waiting pays off: investment in machine B! 48 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (28/32) Energy efficiency and market failure What are the reasons for a reluctance to invest despite high expected returns? Investment priorities Energy not part of core business Information asymmetry Individual life time < economic life time Expectations for the future 49 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (29/32) Contracting Basic principle: − Customer (e.g. owner of a swimming pool) commissions Contractor − Contractor provides new, more efficient energy services (installation, financing, management, maintenance, etc.) − At the end of the contract period, the plants become the property of the customer − The contractor achieves his profit from the energy cost savings (i.e. from the interest arbitrage between the internal rate of return of the Source: Demikhovskiy (2016). Principles of a guaranteed savings energy plant and the standard market contract. p.36 rate of return on capital) 50 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (30/32) Source: Demikhovskiy (2016), Source: Demikhovskiy (2016), Funding a Funding a shared savings project. guaranteed savings project. p.39 p.38 Source: Demikhovskiy (2016), Alternative funding scheme. p.40, based on Kolesnikov (2016) and Nefedov (2014) 51 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (31/32) Contracting Latent problems: Expensive initial evaluation Fluctuating energy demand Rights and obligations of the contracting parties Changes in the legal framework Customer credit risk Legal uncertainties Mistrust 52 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 5. Energy Efficiency (32/32) Practical experiences Ex.: Energy consumption of refrigerators in the US, 1960-2005 Source: Wiel / McMahon (En Pol 2003) NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 7. Conclusions  Main features of the process analysis (+ cohort vintage model)  Determining short- and long-term factors affecting changes in the energy demand  Difference between energy efficiency and energy saving or technical and economic efficiency (energy saving)  Rebound effect and its meaning  Energy efficiency and market failure (and contracting as a remedy)  Innovation and technical progress (invention, innovation, diffusion; economies of scale and learning effects) 54 NTNU IØ 8303, Energy Markets, Part 2, Lecture 6c: Bottom-Up Analysis of Energy Demand | Prof. Dr. Reinhard Madlener | FCN | RWTH Aachen University | Sep 27, 2024 Contact Institute for Future Energy Consumer Needs and Prof. Dr. Reinhard Madlener Behavior (FCN) T +49 241 80 49 820 E.ON Energy Research Center [email protected] Mathieustraße 10 52074 Aachen, Germany http://www.eonerc.rwth-aachen.de/FCN