Bubble Detection with Application to Green Bubbles: A Noncausal Approach

Alain Hecq; Aryan Manafi Neyazi; Francesco Giancaterini; Joann Jasiak

arxiv: 2505.14911 · v2 · submitted 2025-05-20 · 💰 econ.EM

Bubble Detection with Application to Green Bubbles: A Noncausal Approach

Francesco Giancaterini , Alain Hecq , Joann Jasiak , Aryan Manafi Neyazi This is my paper

Pith reviewed 2026-05-22 13:30 UTC · model grok-4.3

classification 💰 econ.EM

keywords bubble detectionnoncausal autoregressive processesgreen bubblesrenewable energystationary processesexplosive episodestime series analysis

0 comments

The pith

Bubbles can be detected as intrinsic features of stationary price processes using mixed causal and noncausal autoregressive models.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces a bubble detection method that relies on mixed causal and noncausal autoregressive processes to identify explosive episodes and measure their duration. It treats bubbles as normal parts of strictly stationary nonlinear price dynamics instead of temporary nonstationary explosions. The approach is demonstrated on green bubbles in renewable energy investments. A reader would care because this reframes bubble analysis as a property of ongoing stationary behavior rather than an anomaly that must eventually burst.

Core claim

By using the tail process representation of mixed causal and noncausal autoregressive processes during explosive episodes, the authors derive a bubble indicator that detects bubbles and quantifies their length inside a framework where prices remain strictly stationary.

What carries the argument

mixed causal and noncausal autoregressive processes together with their tail process representation during explosive episodes

If this is right

The indicator can be used to detect and time green bubbles in renewable energy asset prices.
Bubble duration becomes measurable without first assuming the price series is nonstationary.
The same procedure applies to any price series that can be modeled as a mixed causal-noncausal process.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The method might be tested on historical episodes such as housing or tech bubbles to see whether they register as stationary features.
If successful, regulators could monitor ongoing price series for bubble signals without waiting for an eventual crash.
The approach opens the possibility that some bubbles are predictable components of normal market dynamics rather than rare breaks.

Load-bearing premise

Prices are assumed to follow a strictly stationary process in which any bubble forms as an ordinary part of the nonlinear dynamics.

What would settle it

Applying the indicator to price data from a widely accepted bubble episode and finding that it fails to flag the episode or measure its length would challenge the method.

read the original abstract

This paper introduces a new approach for bubble detection based on mixed causal and noncausal autoregressive processes and their tail process representation during an explosive episode. Departing from traditional definitions of bubbles as nonstationary and temporarily explosive processes, we adopt a perspective in which prices are assumed to follow a strictly stationary process, with the bubble considered an intrinsic component of its nonlinear dynamics. The proposed approach provides a bubble indicator for detecting bubbles and measuring their duration. We implement our strategy to investigate the phenomenon called the "green bubble" in the field of renewable energy investment.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper reframes bubbles as tail behavior inside a strictly stationary mixed causal-noncausal AR process and applies the resulting indicator to green energy prices.

read the letter

The paper's main contribution is to model bubbles inside a strictly stationary process by using mixed causal and noncausal autoregressions, with the explosive episode captured through the tail process representation. This lets them define a bubble indicator and duration measure without invoking temporary nonstationarity. The green-bubble application to renewable energy investments is a straightforward illustration of the method on recent data. The conceptual shift is clean and the noncausal component supplies a natural way to generate the forward-looking dependence that can look explosive in samples while preserving overall stationarity. That part is worth seeing in full. The soft spot is the mapping step. The indicator needs to isolate actual bubble intervals rather than just flagging stretches of heavy tails or high volatility that any stationary process with fat tails can produce. Without explicit derivations or simulation checks on how sensitive the detected start and end dates are to lag order and tail index, it is hard to judge whether the method adds separation power beyond existing tools. The paper would be stronger with direct comparisons to standard explosive-root tests on the same series. This is for readers already comfortable with noncausal AR models who want to try a stationary alternative for bubble work. Someone focused on green finance or alternative detection methods could extract the application, but the core technical claim requires the full derivations to evaluate. I would send it to referees so they can verify whether the tail-process indicator holds up under the stationarity constraint.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces a bubble detection approach based on mixed causal and noncausal autoregressive processes and their tail process representation during explosive episodes. It departs from conventional definitions by modeling prices as strictly stationary processes in which bubbles form an intrinsic component of the nonlinear dynamics rather than temporary explosive nonstationary regimes. The paper derives a bubble indicator for detection and duration measurement and applies the method to investigate green bubbles in renewable energy investment.

Significance. If the tail-process indicator can be shown to isolate economically meaningful bubble intervals rather than generic heavy-tail fluctuations, the stationary framework would provide a useful alternative to explosive-root tests, with direct relevance to policy questions around speculative episodes in green assets.

major comments (2)

[Introduction and §2 (model setup)] The central modeling choice—strict stationarity of the price process with bubbles generated by the noncausal component—requires explicit verification that the resulting tail process produces identifiable, temporary explosive episodes that remain integrable. This assumption is load-bearing for the claim that the indicator distinguishes bubbles from ordinary tail fluctuations; without a derivation or simulation establishing uniqueness and robustness to lag order and tail index, the indicator risks flagging high-volatility periods already present in any heavy-tailed stationary process.
[§3–4 (indicator construction) and §5 (empirical application)] The mapping from tail-process exceedances to bubble start and end dates (described in the abstract and presumably formalized in §3–4) must be accompanied by Monte Carlo evidence or theoretical bounds showing that the dates are robust to reasonable choices of lag orders and tail-index estimators. Absent such checks, the duration measure cannot be treated as reliable for the green-bubble application.

minor comments (2)

[§2] Clarify the precise definition of the mixed causal-noncausal AR specification and the tail-process representation to ensure readers can replicate the indicator construction.
[§5] Specify the data sources, sample period, and variable definitions used for the renewable-energy price series in the green-bubble application.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed comments, which help clarify the presentation of our stationary noncausal framework for bubble detection. We respond to each major comment below and indicate the revisions we will implement.

read point-by-point responses

Referee: [Introduction and §2 (model setup)] The central modeling choice—strict stationarity of the price process with bubbles generated by the noncausal component—requires explicit verification that the resulting tail process produces identifiable, temporary explosive episodes that remain integrable. This assumption is load-bearing for the claim that the indicator distinguishes bubbles from ordinary tail fluctuations; without a derivation or simulation establishing uniqueness and robustness to lag order and tail index, the indicator risks flagging high-volatility periods already present in any heavy-tailed stationary process.

Authors: The mixed causal-noncausal autoregressive specification is constructed to be strictly stationary by design, with the noncausal roots generating the temporary forward-looking explosive behavior that appears locally as a bubble while the process remains globally integrable under the maintained tail-index conditions. We agree that an explicit verification of these properties strengthens the distinction from generic heavy-tail volatility. In the revised manuscript we will add a new subsection to §2 that (i) derives the tail-process representation during explosive episodes and confirms integrability, and (ii) reports Monte Carlo experiments demonstrating that the resulting bubble indicator is robust to lag-order selection and tail-index estimation and does not merely flag ordinary heavy-tailed fluctuations. revision: yes
Referee: [§3–4 (indicator construction) and §5 (empirical application)] The mapping from tail-process exceedances to bubble start and end dates (described in the abstract and presumably formalized in §3–4) must be accompanied by Monte Carlo evidence or theoretical bounds showing that the dates are robust to reasonable choices of lag orders and tail-index estimators. Absent such checks, the duration measure cannot be treated as reliable for the green-bubble application.

Authors: We concur that robustness of the dating procedure is essential for credible duration measurement in the renewable-energy application. The revised version will therefore augment §3–4 with a Monte Carlo study that evaluates the sensitivity of detected start and end dates to plausible ranges of lag orders and alternative tail-index estimators. The design of the simulations will be calibrated to the persistence and tail behavior observed in the green-asset series analyzed in §5, thereby directly supporting the reliability of the empirical duration estimates. revision: yes

Circularity Check

0 steps flagged

No circularity: bubble indicator derived from established mixed causal-noncausal AR tail-process properties without reduction to inputs or self-citations.

full rationale

The paper defines its bubble indicator via the tail-process representation of a mixed causal-noncausal autoregressive process under the maintained assumption of strict stationarity. This modeling choice is presented as an alternative perspective rather than a redefinition of the target quantity, and the abstract supplies no equations showing that the indicator is obtained by fitting a parameter to the same data it is later used to detect. No load-bearing self-citation, uniqueness theorem imported from the authors' prior work, or ansatz smuggled via citation is visible in the provided text. The derivation therefore remains self-contained against external benchmarks for the mixed AR framework and does not reduce the claimed indicator to a fitted input or self-referential definition.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only abstract available; no free parameters, axioms, or invented entities are specified in the provided text.

pith-pipeline@v0.9.0 · 5627 in / 971 out tokens · 33674 ms · 2026-05-22T13:30:18.118556+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Foundation/AbsoluteFloorClosure.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Departing from traditional definitions of bubbles as nonstationary and temporarily explosive processes, we adopt a perspective in which prices are assumed to follow a strictly stationary process, with the bubble considered an intrinsic component of its nonlinear dynamics.
IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

The tail process ... X_h = c_{h+N}/c_N X_0 ... P[N=h] = ψ^{-h α} / [1/(1-ϕ^α) + 1/(1-ψ^α) - 1]

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.