Based on the paper: "The impact of market design and clean energy incentives on strategic generation investments and resource adequacy in low-carbon electricity markets" by Kwon et al. (2023).
Abstract Summary: Our electricity grid needs to be reliable and increasingly clean. This paper explores how different rules for electricity markets (like how generators get paid) and incentives for clean energy influence companies' decisions to build new power plants (like wind, solar, batteries, gas) and how reliable the overall system ends up being.
Our society relies heavily on electricity. With more things going electric (like cars and heating) and the urgent need to tackle climate change, our power grid is undergoing a massive transformation. We're seeing more Variable Renewable Energy (VRE) sources like wind and solar, plus new technologies like large batteries (Energy Storage - ES). At the same time, extreme weather events linked to climate change put more stress on the system.
Planning the future grid is complex. We need enough power generation to meet demand reliably, even on the hottest or coldest days (this is called Resource Adequacy). We also want the power to be clean and affordable.
In many parts of the world, including large areas of the U.S., the grid isn't planned centrally by a single entity. Instead, independent power companies, called Generating Companies (GenCos), decide whether to build new power plants or retire old ones based on whether they expect to make a profit. These decisions happen within complex Wholesale Electricity Markets run by Independent System Operators (ISOs) or Regional Transmission Organizations (RTOs), who also manage grid reliability.
The core question this paper tackles is: How do the specific rules of these markets, and policies like clean energy incentives, influence GenCos' investment decisions, and ultimately, the reliability and cleanliness of our future power supply?
Traditionally, grid planners used models to find the cheapest way to build enough power plants to meet future demand reliably. This is the Least-Cost Generation Expansion Planning (LC-GEP) approach. It assumes a central planner optimizes the whole system for minimum cost.
Imagine a planner needs to add 1000 MW of new capacity. They have options: Solar, Wind, and Gas, each with different costs. The goal is to meet the 1000 MW target at the lowest total cost.
Total Cost: $800,000
This is highly simplified. Real LC-GEP considers operating costs, fuel costs, reliability constraints (like Planning Reserve Margin), transmission, etc.However, in deregulated markets, there's no single central planner making all investment decisions. Instead, multiple GenCos act strategically, aiming to maximize their own profits. They look at expected revenues from selling energy, capacity (if available), and potentially clean energy credits, and compare that to the cost of building and operating a plant.
This paper uses a more complex model called the Strategic Capacity Investment Model (SCIM). It simulates this competitive environment using game theory, specifically an Equilibrium Problem with Equilibrium Constraints (EPEC). The key idea is that each GenCo makes its best decision assuming what other GenCos will do, and the model finds a stable outcome (an equilibrium) where no GenCo wants to unilaterally change its strategy.
Imagine you're a GenCo deciding whether to build a new 100 MW Solar plant. You estimate its lifetime cost and the potential revenue from selling energy in the market.
Estimated Profit: $10 Million
Decision: Invest
This is extremely simplified. Real GenCos forecast complex market prices over decades, consider financing, risk, competitor actions, different revenue streams (energy, capacity, etc.), and operating costs.Crucially, the collective outcome of these individual profit-driven decisions might be different from the system-wide least-cost solution. The SCIM allows researchers to explore these differences.
The paper examines how different market structures influence GenCo decisions and system outcomes. It focuses on three main types:
In the simplest design, GenCos only get paid for the actual energy ($MWh$) they produce and sell into the wholesale market. Prices can vary significantly, sometimes spiking very high during periods of scarcity (when demand is high and supply is tight). The theory is that these occasional high prices should provide enough revenue, even for plants that don't run often (like gas "peaker" plants needed for reliability), to cover their costs and incentivize investment.
Can infrequent price spikes cover a plant's fixed costs? Let's simulate a year of hourly prices for a hypothetical peaking plant that costs $50,000/year to maintain (fixed cost) and only runs when prices are high.
Click 'Run Simulation' to see price spikes.
Simulated Annual Energy Revenue: $0
Plant Fixed Costs: $50,000
Net Result: ---
This simulation uses random price generation for illustration. Real prices depend on complex supply, demand, and grid conditions. The paper's "ORDC" refers to Operating Reserve Demand Curves which help shape scarcity prices.Finding from Paper: In SCIM simulations, Energy-Only markets often led to lower investment levels and lower Planning Reserve Margins (PRM) - the buffer of extra capacity above peak demand - compared to what a least-cost planner would choose or what happens in markets with capacity payments. This suggests potential reliability concerns if solely relying on energy prices.
To address potential under-investment in EO markets, many regions use Capacity Markets. Here, GenCos receive payments not just for energy produced, but for being available to produce power when needed, measured in capacity ($MW$). The ISO/RTO determines how much capacity the system needs to be reliable (often based on a target PRM) and runs an auction to procure it. The price is typically set where the supply of available capacity meets an administratively determined Capacity Demand Curve.
Generators offer their available capacity (derated by factors like outage probability or fuel limits, called "capacity credits"). The ISO has a demand curve representing the value of reliability. The intersection sets the clearing price and quantity.
Cleared Capacity: 0 MW
Clearing Price: $0 /MW-day (Illustrative)
This uses a simplified linear supply curve and variable demand curve. Real auctions are complex. The paper explores different demand curve shapes (CM1, CM2, CM3) and capacity credit rules (Peak Av. Output, ELCC).Finding from Paper: Capacity markets, simulated in SCIM, generally resulted in PRMs closer to the target levels compared to EO markets. However, the specific design of the capacity market (e.g., shape of the demand curve) significantly influenced which types of power plants were built (e.g., favoring gas vs. solar depending on the curve).
This is a newer concept, explored in the paper, designed to directly incentivize clean energy generation. Similar to a capacity market, there's a target amount of clean energy attributes (like Renewable Energy Credits - RECs) needed. Generators eligible (like wind, solar, nuclear, hydro in the paper's setup) offer their expected clean energy production, and an auction clears based on a demand curve reflecting the value of clean energy (e.g., related to the social cost of carbon or policy goals).
Eligible clean generators offer their expected annual clean MWh. The market operator has a demand curve for these clean attributes.
Cleared Clean Energy: 0 %
Clearing Price: $0 /MWh (Illustrative)
This is a conceptual illustration. The paper models a specific downward-sloping demand curve based on Brattle Group proposals.Finding from Paper: Adding a CEM promoted more investment in wind, solar, and storage compared to not having one, especially when combined with a capacity market. However, at the reference price used ($1.30/MWh), its impact in the energy-only market setting was minimal, suggesting the price signal might need to be stronger or combined with other mechanisms.
Beyond market structure, specific policies also influence investment:
A GenCo compares building a new Gas plant vs. a Wind plant. Adjust the policy levels to see how the relative profitability changes.
Baseline Costs/Revenues (Illustrative):
Gas Plant: Cost $80M, Revenue $90M (Profit +$10M)
Wind Plant: Cost $120M, Revenue $125M (Profit +$5M)
Adjusted Profitability:
Gas Plant Profit: $10M
Wind Plant Profit: $5M
More Profitable Option: Gas Plant
Assumes $15/MWh PTC adds ~$5M revenue, $40/ton CP reduces Gas revenue by ~$8M in this example.Finding from Paper: Both tax credits and carbon pricing significantly shifted investments towards cleaner resources in the SCIM simulations. Carbon pricing, by changing the relative operating costs, had a broad impact on which plants run and when, leading to substantial increases in renewables and storage, and higher PRMs in the EO case. Tax credits also boosted renewables but had less impact on retiring existing fossil plants.
The paper's core message is that the rules of the electricity market profoundly impact investment decisions and grid outcomes. Here are some key takeaways visualized:
Let's compare the resulting generation mix and reliability (PRM) under different scenarios from the profit-driven SCIM model versus the idealized Least-Cost (LC-GEP) model. Select scenarios to compare.
Planning Reserve Margin (PRM): 0%
Planning Reserve Margin (PRM): 0%
This research underscores the critical importance of well-designed electricity markets and policies for navigating the transition to a reliable, clean, and affordable power system. Key takeaways include:
The paper uses sophisticated modeling (EPEC solved with diagonalization and progressive hedging) to capture these complex interactions. While this interactive exploration simplifies many details, it aims to convey the fundamental insight: the rules we set for our electricity markets will play a major role in shaping our energy future.