💰 quantitative finance

Market-implied time to transition to a low-carbon economy: a stochastic modelling and inference framework

This paper introduces "Time to Transition" (TtT), a new market-implied metric derived from the greenium term structure, and develops a two-layer stochastic inference framework combining regulatory deadline-constrained models with regime-switching extensions to estimate latent transition timing and ensure consistent parameter identification.

Original authors: Lorenzo Mercuri, Andrea Perchiazzo, Edit Rroji, Ilaria Stefano

Published 2026-05-06

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Lorenzo Mercuri, Andrea Perchiazzo, Edit Rroji, Ilaria Stefano

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 the global economy is a massive ship trying to steer from a stormy, carbon-heavy sea toward a calm, low-carbon harbor. The people on board (investors) know the destination, but they are arguing about when the ship will actually arrive. Some think it will dock by 2030; others think it won't happen until 2041.

This paper introduces a new tool to figure out what the investors really think about that arrival time, not by asking them, but by listening to the "whispers" in the bond market.

Here is a breakdown of the paper's ideas using simple analogies:

1. The "Green Premium" and the "Greenium"

In the bond market, governments and companies can issue "Green Bonds" (to fund eco-friendly projects) and "Brown Bonds" (standard projects). Usually, Green Bonds are slightly cheaper for the issuer to sell because investors are willing to accept a lower return just to support the environment. This price difference is called the Greenium.

Think of the Greenium like a discount coupon.

Short-term coupons: Investors are very excited about the future. They are willing to take a big discount (a large Greenium) for bonds that mature soon because they believe the "green transition" is happening now.
Long-term coupons: As you look further into the future, the excitement fades, or the discount gets smaller.

2. The Core Idea: "Time to Transition" (TtT)

The authors noticed something interesting in the German bond market: The "discount" (Greenium) is often inverted. Investors are willing to take a bigger hit on their returns for short-term bonds than for long-term ones.

Why? Because investors believe the transition to a green economy is urgent and happening soon. But they also worry that if the transition drags on too long, the "green" label won't be worth the sacrifice.

The paper creates a new metric called Time to Transition (TtT).

The Metaphor: Imagine the Greenium is a hill. Right now, the hill is steep (high discount for short-term, low for long-term). The authors define the "Time to Transition" as the moment the hill flattens out completely.
The Signal: When the hill becomes flat, it means investors believe the transition is done. At that point, the "green" label is just a permanent feature, and the discount is the same for all timeframes.
The Goal: By measuring how steep the hill is today, the paper tries to calculate how many years it will take for that hill to flatten. It's like looking at the speed of a car to guess when it will reach the finish line.

3. The Two Models: The "Strict Deadline" vs. The "Wobbly Clock"

To calculate this time, the authors built two mathematical models (like two different types of clocks).

Model A: The Regulatory Deadline (RDCM)

The Analogy: Imagine a strict teacher who says, "The exam is on December 31st, 2030, no matter what."
How it works: This model assumes everyone agrees on a single, fixed date when the transition will be finished. The math predicts that as we get closer to that date, the "Greenium hill" will smooth out.
The Problem: In real life, people don't always agree. Some think the deadline is 2030; others think it's 2040. This model is too rigid.

Model B: The Switching Deadline (SRDCM)

The Analogy: Imagine a group of investors where some are "Optimists" (believing the deadline is 2030) and some are "Skeptics" (believing it's 2041). Every day, the market mood shifts. One day, the Optimists are in charge; the next day, the Skeptics take over.
How it works: This model allows the "deadline" to switch back and forth between different dates based on new news (like a new law or a war). It acts like a filter, constantly re-evaluating: "Right now, does the market think we are on track for 2030, or has the goalpost moved to 2041?"

4. The Experiment: Listening to German Bonds

The authors tested their theory using real data from German government bonds (specifically "twin bonds," which are identical except for their "Green" or "Brown" label). This is a perfect laboratory because the only difference between the two bonds is the label, making the data very clean.

The Result: When they used the simple "Strict Deadline" model, the math didn't fit the real world well (the "residuals" were messy).
The Success: When they used the "Switching Deadline" model, it fit the data perfectly.
The Finding: The model showed that for a long time, the market was split. However, in recent months, the "Optimist" view (transition by 2030) has become less dominant, and the market is increasingly leaning toward the idea that the transition might take longer (pushing the perceived deadline further out).

5. The Mathematical Guarantee

The paper also proves mathematically that if you watch the market long enough and closely enough, you can accurately figure out the specific "speed" at which the market is changing its mind. It's like proving that if you watch a clock long enough, you can eventually tell exactly how fast its gears are turning, even if the clock is a bit wobbly.

Summary

This paper doesn't just measure how much money people are saving for green bonds. Instead, it measures when people think the green transition will actually happen.

The Tool: A new way to read bond prices to guess the "arrival time" of a low-carbon economy.
The Method: Using a flexible model that accounts for the fact that investors change their minds and have different deadlines in mind.
The Takeaway: The market is a living thing that constantly updates its belief about the future. This paper gives us a way to decode those beliefs in real-time.

Technical Summary: Market-Implied Time to Transition to a Low-Carbon Economy

Problem Statement
The transition to a low-carbon economy involves institutional timelines (e.g., the EU's 2030 and 2050 targets) that are perceived differently by market participants. While financial prices reflect beliefs about future states, extracting the specific timing of these transitions from market data is inherently difficult due to the latent nature of the transition date and the noise in observable signals. Existing literature has extensively measured the magnitude of the "greenium" (the yield spread between green and conventional bonds), but less attention has been paid to using the term structure of the greenium to infer when investors believe the transition will effectively materialize. This paper addresses the problem of recovering a market-implied "Time to Transition" (TtT) from the cross-maturity differences in the greenium term structure, specifically investigating under which observation schemes such inference admits a structural interpretation.

Methodology
The authors propose a two-layer analytical framework centered on a new observable object, $X_t$ , defined as the difference between the greenium at two selected maturity nodes (short and long) of twin bonds (bonds identical in all financial characteristics except for their "green" label). The paper posits that $X_t$ follows a mean-reverting process that converges to zero as the transition occurs, at which point the greenium term structure becomes flat.

Regulatory Deadline-Constrained Model (RDCM):
The baseline model assumes a deterministic regulatory deadline $T$ (e.g., 2030). The dynamics of $X_t$ are modeled as a linear stochastic differential equation (SDE) with time-varying drift and diffusion coefficients that depend on the time remaining to the deadline ( $T-t$ ). Crucially, the model incorporates a structural time point $\tau < T$ . Before $\tau$ , coefficients are constant; after $\tau$ , they decay as functions of $T-t$ , forcing the process to hit the terminal condition $X_T = 0$ .
- Inference: The model utilizes an exact Gaussian bridge likelihood. Because the process is conditioned on a terminal value of zero, the authors derive a telescopic representation of the log-likelihood, allowing for tractable calibration via Maximum Likelihood Estimation (MLE) on a fixed observation grid.
Switching Regulatory Deadline-Constrained Model (SRDCM):
To address heterogeneity in perceived deadlines, the authors extend the RDCM to a regime-switching framework. The perceived deadline is treated as a latent state governed by a discrete-time Hidden Markov Model (HMM). Between observation dates, the process follows a specific RDCM dynamics associated with the active regime (e.g., a 2030 deadline vs. a 2041 deadline).
- Inference: Calibration is performed using an Expectation-Maximization (EM) algorithm adapted for this structure. The forward-backward recursions are modified to account for the fact that the latent regime at time $t_{i-1}$ governs the bridge dynamics over the interval $(t_{i-1}, t_i]$ . This allows for the filtering of posterior probabilities across competing transition scenarios.
Asymptotic Identification (Fixed-Horizon Infill):
The paper investigates the consistency of estimating the diffusion parameters under a fixed-horizon infill asymptotic scheme (where the number of observations $n \to \infty$ within a fixed time window $[t_0, l]$ ). Using $M_n$ -contrast functions and consistency theory for M-estimators, the authors analyze which parts of the parameter space are identifiable.
- Key Finding: Identification depends on whether the observation window overlaps the "decay region" ( $t > \tau$ ). If the window lies entirely before $\tau$ , the post-decay parameters (which govern the shape of convergence) are not jointly identifiable; only the pre-decay parameters are consistently estimable.

Key Results

Empirical Application: The framework was applied to daily data from the German sovereign twin-bond market (September 2021 – January 2025).
- A single-regime RDCM (assuming a 2031 deadline) failed to capture the data dynamics, as evidenced by non-normal residuals and a rejected Kolmogorov-Smirnov test.
- A two-regime SRDCM (comparing a 2031 deadline against a 2041 deadline) provided a significantly better fit (lower AIC) and residuals consistent with a normal distribution.
- The filtering results indicated that for most of the observation period, particularly in late 2024, the market did not perceive the 2030 transition deadline as realistic, with probability mass shifting toward the later 2041 scenario.
Theoretical Identification: The paper establishes that under fixed-horizon infill, the diffusion parameters associated with the "visible" part of the trajectory (the region where the process is active) can be consistently estimated. However, parameters governing the convergence profile (active only after $\tau$ ) remain unidentified if the sample does not extend into the post- $\tau$ region.

Significance and Contributions
The paper makes two primary contributions:

A New Inferential Object: It introduces the "Time to Transition" (TtT) as a market-implied measure derived from the cross-maturity greenium difference. Unlike static measures of green premia, TtT is forward-looking and captures the collective investor expectation of when the transition risk is neutralized.
A Tractable Inference Framework: It develops a rigorous stochastic modeling and inference framework (RDCM and SRDCM) that combines exact Gaussian bridge likelihoods with regime-switching filtering. This allows for the operational monitoring of transition timing on daily grids.
Structural Identification Results: The paper provides necessary and sufficient conditions for the consistent estimation of diffusion parameters in this setting. It clarifies that structural identification of competing transition dates is only possible when the observation window captures the specific dynamics of the convergence phase (post- $\tau$ ).

The authors position the German twin-bond market as a "disciplined laboratory" for this inference, as the identical financial characteristics of the bonds isolate the informational content of the green label. The work shifts the focus from measuring the level of the green premium to the stochastic inference of latent temporal labels, providing a foundation for a broader theory of market-implied transition timing.