📈 economics

Trading Frictions in Dynamic Cap-and-Trade Markets

This paper develops and quantifies a dynamic stochastic model of cap-and-trade markets to demonstrate how the interaction of slow participation, limited intermediation, and heterogeneous information creates a unique equilibrium premium that non-additively amplifies price responses, using 2.7 million EU ETS transactions to reveal that roughly 40% of operators do not trade annually and that purchases concentrate in April when returns are systematically high.

Original authors: Nicola Borri, Yukun Liu, Aleh Tsyvinski, Xi Wu

Published 2026-06-03

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Nicola Borri, Yukun Liu, Aleh Tsyvinski, Xi Wu

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a giant, global game of "Hot Potato" where the potato is a permit to pollute the air. In this game, the government sets a strict limit on how many permits exist (the "cap"). Companies that pollute must hold enough permits to cover their emissions by a specific deadline every year. If they don't have enough, they have to buy them from others. This system is called Cap-and-Trade.

Theoretically, this should be a perfectly smooth market: companies with extra permits sell them to companies that need them, and the price settles at a fair level. But in reality, the market is messy. This paper, by Borri, Liu, Tsyvinski, and Wu, investigates why the market is messy and how three specific "frictions" (obstacles) make it inefficient.

Here is the story of their findings, told through simple analogies.

The Three Obstacles (Frictions)

The authors identify three main reasons why the market doesn't work like a perfect video game:

Slow Participation (The "Lazy Shopper"):
Imagine you need to buy groceries for a party next week. In a perfect world, you'd go to the store as soon as you realize you're missing an ingredient. But in this market, many companies are like shoppers who wait until the very last minute. They don't want to bother with the hassle of logging in, checking prices, or getting internal approvals. So, they sit on their hands until the deadline hits.
- The Result: About 40% of companies don't trade at all in a given year. They just hope their initial allocation is enough.
Limited Intermediation (The "Bottlenecked Bridge"):
When companies do decide to trade, they often need help. They use intermediaries (like brokers or banks) to buy and sell. But these intermediaries have a limit on how much traffic they can handle at once. Think of them as a narrow bridge. If everyone tries to cross the bridge at the same time, it gets jammed.
- The Result: When many companies rush to buy permits at the last minute, the "bridge" gets clogged, and the price of crossing (the permit price) spikes.
Heterogeneous Information (The "Insider Traders"):
Not everyone knows the same things. Some companies are "smart shoppers" who have inside information about future shortages or price changes. They trade actively based on this knowledge. Others are just trying to meet their legal requirements.
- The Result: The "smart shoppers" trade frequently and seem to know when prices are going to go up or down before the rest of the market does.

The Big Problem: The "April Rush"

In the European Union's version of this system (EU ETS), there is a hard deadline: April 30th. By this date, every company must hand over their permits.

Because of the "Lazy Shopper" problem, many companies wait until the very last moment to buy what they need. This creates a massive, predictable rush in April.

The Analogy: Imagine a sale at a store that ends at 5:00 PM. If everyone waits until 4:55 PM to buy, the store runs out of stock, and the remaining items become incredibly expensive because the "bridge" (intermediaries) is overwhelmed.
The Finding: The paper finds that April is consistently the most expensive month to buy permits. Companies that wait until the last minute end up paying a huge "April Premium." The authors estimate this costs delayed buyers about €5 billion extra over the study period. It's like paying double for a ticket because you waited until the concert started to buy it.

The "Smart" Traders and Predictable Prices

The paper also looked at who is actually making money. They found that a small group of companies (about 10%) trade huge amounts of permits. These are the "frequent traders."

The Finding: When these frequent traders buy permits, prices tend to go up in the following months. When they sell, prices tend to go down.
The Analogy: It's like having a group of people who know a storm is coming. They buy umbrellas today. Because they are buying so many, you can predict that umbrellas will be more expensive tomorrow. Their trading activity acts as a crystal ball for future prices.

The Surprising Twist: Frictions Don't Just Add Up

The most interesting part of the paper is how these obstacles interact. You might think that if you fix one problem (like making it easier to trade), the market gets 50% better. If you fix two problems, it gets 100% better.

The paper says: No, it's not that simple.

The Dampening Effect: If you make it easier for intermediaries to handle traffic (fixing the "bridge"), the price doesn't drop as much as you'd expect. Why? Because the "Lazy Shoppers" see that it's easier to trade, so they decide to wait even longer until the deadline, thinking they can still get a deal. Their new laziness cancels out some of the benefit of the better bridge.
The Amplification Effect: Conversely, if trading is already very hard and expensive, making the "bridge" slightly better has a bigger impact because the "Lazy Shoppers" are finally forced to move.

The "Joint Fix" Surprise:
The authors ran a simulation: What if we fixed both problems at once?

Fixing the bridge alone lowered costs by 38%.
Fixing the laziness alone lowered costs by 20%.
Fixing both together lowered costs by exactly 50%.
This proves that the two problems interact in a complex way. You can't just add their effects together; they change each other's behavior.

The Solution: Stagger the Deadline

The paper suggests a simple policy change to fix the "April Rush" without changing the total number of permits or the total pollution allowed.

The Idea: Instead of everyone having to hand in their permits on April 30th, split the companies into four groups.

Group A hands in permits in April.
Group B hands them in May.
Group C in June.
Group D in July.

The Result: This spreads the traffic out. The "bridge" never gets clogged. The "Lazy Shoppers" have more time to trade without panic.

The Payoff: The authors calculate this simple change would reduce the extra "April Premium" paid by companies by about 42%. It saves money for everyone without letting anyone pollute more.

Summary

This paper tells us that the carbon market is inefficient not because the rules are bad, but because companies are lazy, the "bridges" are narrow, and some people have inside info. These factors create a predictable price spike every April. The solution isn't necessarily to build bigger bridges, but to stagger the deadlines so everyone doesn't rush at the same time.

Technical Summary: Trading Frictions in Dynamic Cap-and-Trade Markets

Problem and Motivation
This paper investigates the effectiveness of cap-and-trade systems as a mechanism for correcting climate externalities when markets are subject to multiple, interacting trading frictions. While canonical Pigouvian theory suggests that optimal corrective instruments (taxes or cap-and-trade) can be implemented in closed form, the presence of multiple frictions complicates the design and analysis of such policies. The authors argue that the joint effect of frictions on prices and quantities is not merely additive; interactions can either dampen or amplify price responses, a phenomenon rooted in the classic work of Lipsey and Lancaster (1956) on the second best. The specific focus is the European Union Emissions Trading System (EU ETS), where the authors seek to understand how slow participation, limited intermediation, and heterogeneous information interact to shape market outcomes, specifically regarding participation rates, surrender-month price premiums, and return predictability.

Methodology
The paper employs a three-pronged approach: theoretical modeling, quantitative calibration, and empirical testing using a comprehensive dataset.

Theoretical Framework:
- Two-Period Model: The authors first develop a stylized two-period model to isolate the basic mechanisms. This model incorporates fixed participation costs, quadratic intermediation costs (representing limited balance-sheet capacity), and private information. It establishes that firms trade early only if the expected price premium in the surrender month exceeds the cost of early participation.
- Continuous-Time Model: The core theoretical contribution is a dynamic stochastic model where operators endogenously choose their market-access intensity (modeled as a Poisson arrival rate) at a convex cost. Intermediaries absorb net order flow with a quadratic cost, creating affine price impact. A subset of operators receives private signals about future scarcity.
- Equilibrium Characterization: The model solves for a unique equilibrium surrender-month premium. A key analytical result is that endogenous access choices create a feedback loop: a higher expected premium incentivizes higher access intensity, which reduces residual terminal demand, thereby dampening the premium itself.
Data and Quantitative Calibration:
- Dataset: The empirical analysis utilizes the universe of 2.7 million transactions and compliance records from the European Union Transaction Log (EUTL) spanning February 2005 to September 2021 (covering EU ETS Phases I through IV).
- Calibration: The authors calibrate the model using observed trading frequencies to infer effective market-access intensities ( $\lambda$ ) and observed April return premia to infer the surrender-window return-impact sensitivity ( $\varphi$ ). The model-implied residual terminal demand is validated against observed April net purchases.
Empirical Strategy:
- The authors test three specific predictions derived from the model: (1) persistent non-trading due to high access costs; (2) concentrated purchasing and high returns in the surrender month (April) due to residual demand pressure; and (3) return predictability driven by informed, frequent traders.

Key Contributions and Results

Endogenous Access and Dampening: The paper characterizes access choices in closed form and proves the uniqueness of the equilibrium premium. A central finding is that endogenous access dampens the price response to changes in intermediary capacity. Specifically, if intermediary return impact is halved, the equilibrium premium falls by only ~38% rather than the 50% predicted by a static model. This is because the lower premium reduces the incentive for firms to access the market early, increasing residual terminal demand and partially offsetting the policy gain.
Non-Additive Interactions: The interaction between frictions is non-additive. While halving both access costs and intermediary impact jointly reduces the premium by exactly 50% (due to scale invariance), the individual effects are not additive. Furthermore, the paper demonstrates an amplification channel: when access costs are high, the dampening effect of endogenous access is weaker, making the equilibrium premium more sensitive to changes in intermediary capacity.
Empirical Facts:
- Slow Participation: Approximately 40% of regulated operators do not trade in a given year, a rate that persists even in later phases of the EU ETS. This non-trading is not explained by initial allocations matching compliance needs; rather, it reflects persistent gaps where firms choose low access intensity.
- Surrender-Month Premium: Trading is highly concentrated in April, the surrender deadline. Net purchases in April are roughly ten times larger than in other months, driving a systematic average return premium of ~10%. The authors estimate the "April premium" paid by delayed buyers at approximately €5 billion per cycle.
- Informative Flow: About 10% of firms trade volumes significantly exceeding their compliance needs. The net flow of these frequent traders (Operator Flow Imbalance) positively predicts future allowance returns up to 12 months ahead, with the predictive power concentrated among frequent traders. This suggests these operators possess informational advantages regarding future scarcity.
Policy Counterfactuals: Using the calibrated model, the authors evaluate market-design interventions. Staggering surrender deadlines across four equally sized groups is predicted to reduce the surrender-month premium by ~59% and the implied April premium paid by delayed buyers by ~42%, without altering aggregate emissions or the cap.

Significance and Claims
The paper claims to provide the first theoretical and quantitative characterization of how multiple trading frictions interact in a dynamic cap-and-trade setting. It moves beyond static analyses to show that policy interventions targeting a single friction (e.g., improving intermediary capacity) may yield smaller-than-expected price effects due to endogenous behavioral responses (dampening). Conversely, the interaction of frictions can amplify price sensitivity under certain conditions.

The authors position their work as a bridge between climate economics, asset pricing, and market microstructure. They argue that understanding these frictions is essential for evaluating the real-world effectiveness of cap-and-trade systems, as the secondary market mechanism determines how compliance obligations are reallocated and how scarcity information is incorporated into prices. The findings suggest that market design improvements, such as staggering compliance deadlines, could significantly reduce the financial costs of compliance (the "April premium") without compromising environmental integrity. The paper does not claim to offer a full structural estimation of all EU ETS primitives but rather uses reduced-form objects to quantify the magnitude of frictions and the direction of counterfactual outcomes.