GARCH QUANT
@garchquant.bsky.social
Methods-first research on GARCH-family models, volatility dynamics, and robust financial econometrics for quant and risk professionals.
Two questions for governed volatility forecasting:
When does a forecast difference actually change the decision?
When can model disagreement warn of fragility?
New research + two video explainers
SSRN 7422079 / 7423900
YouTube
www.youtube.com/watch?v=riep...
www.youtube.com/watch?v=1ch0...
When does a forecast difference actually change the decision?
When can model disagreement warn of fragility?
New research + two video explainers
SSRN 7422079 / 7423900
YouTube
www.youtube.com/watch?v=riep...
www.youtube.com/watch?v=1ch0...
Different Volatility Forecasts, Same Portfolio? Decision Equivalence Explained
YouTube video by GARCH QUANT
www.youtube.com
September 14, 2026 at 5:06 AM
Two questions for governed volatility forecasting:
When does a forecast difference actually change the decision?
When can model disagreement warn of fragility?
New research + two video explainers
SSRN 7422079 / 7423900
YouTube
www.youtube.com/watch?v=riep...
www.youtube.com/watch?v=1ch0...
When does a forecast difference actually change the decision?
When can model disagreement warn of fragility?
New research + two video explainers
SSRN 7422079 / 7423900
YouTube
www.youtube.com/watch?v=riep...
www.youtube.com/watch?v=1ch0...
New research from the GARCH Institute: AI, algorithmic trading, and market microstructure.
60 studies. Three debates—liquidity, price discovery, and stability. One emerging reality: the algorithmic ecology.
papers.ssrn.com/sol3/papers....
60 studies. Three debates—liquidity, price discovery, and stability. One emerging reality: the algorithmic ecology.
papers.ssrn.com/sol3/papers....
Artificial Intelligence, Algorithmic Trading, and Financial Market Microstructure: A Systematic Literature Review
The past two decades have witnessed a structural transformation in financial market microstructure: human traders have been displaced by algorithmic agents oper
papers.ssrn.com
August 20, 2026 at 6:33 AM
New research from the GARCH Institute: AI, algorithmic trading, and market microstructure.
60 studies. Three debates—liquidity, price discovery, and stability. One emerging reality: the algorithmic ecology.
papers.ssrn.com/sol3/papers....
60 studies. Three debates—liquidity, price discovery, and stability. One emerging reality: the algorithmic ecology.
papers.ssrn.com/sol3/papers....
As AI moves from hype to infrastructure, the bottleneck shifts from chips to power. Our new working paper tracks how this physical constraint shows up in cross‑asset risk between semiconductors and energy.
SSRN: ssrn.com/abstract=708...
SSRN: ssrn.com/abstract=708...
The Physical Limits of Compute: Dynamic Risk Contagion Networks Between Semiconductor and Energy Infrastructure Assets Under the AI Super Cycle
As generative AI transitions into large-scale commercialization and hyperscale data center construction, the binding constraint on compute expansion has shifted
ssrn.com
July 18, 2026 at 2:15 PM
As AI moves from hype to infrastructure, the bottleneck shifts from chips to power. Our new working paper tracks how this physical constraint shows up in cross‑asset risk between semiconductors and energy.
SSRN: ssrn.com/abstract=708...
SSRN: ssrn.com/abstract=708...
From a modeling perspective, the interesting part of China’s ETS is the gap between variance and tail channels.
Full-sample GJR-GARCH-X deltas are small and insignificant, yet Asymmetric Slope CAViaR-X finds a strong carbon tail-risk effect in the power sector’s 1% VaR.
ssrn.com/abstract=699...
Full-sample GJR-GARCH-X deltas are small and insignificant, yet Asymmetric Slope CAViaR-X finds a strong carbon tail-risk effect in the power sector’s 1% VaR.
ssrn.com/abstract=699...
Carbon Price Uncertainty and Stock Market Volatility Transmission: A GJR-GARCH-X and DCC Approach with Evidence from China's ETS
This paper examines whether carbon price uncertainty transmits to stock market volatility in China's high-emission sectors following the launch of the national
ssrn.com
July 12, 2026 at 1:24 PM
From a modeling perspective, the interesting part of China’s ETS is the gap between variance and tail channels.
Full-sample GJR-GARCH-X deltas are small and insignificant, yet Asymmetric Slope CAViaR-X finds a strong carbon tail-risk effect in the power sector’s 1% VaR.
ssrn.com/abstract=699...
Full-sample GJR-GARCH-X deltas are small and insignificant, yet Asymmetric Slope CAViaR-X finds a strong carbon tail-risk effect in the power sector’s 1% VaR.
ssrn.com/abstract=699...
Carbon price uncertainty and stock volatility in China’s ETS, using GJR-GARCH-X, DCC-GARCH, and CAViaR-X on power, steel, and cement.
The interesting part is a dormant variance channel but an active tail-risk channel in the power sector.
SSRN: ssrn.com/abstract=699...
The interesting part is a dormant variance channel but an active tail-risk channel in the power sector.
SSRN: ssrn.com/abstract=699...
Carbon Price Uncertainty and Stock Market Volatility Transmission: A GJR-GARCH-X and DCC Approach with Evidence from China's ETS
This paper examines whether carbon price uncertainty transmits to stock market volatility in China's high-emission sectors following the launch of the national
ssrn.com
July 10, 2026 at 6:08 PM
Carbon price uncertainty and stock volatility in China’s ETS, using GJR-GARCH-X, DCC-GARCH, and CAViaR-X on power, steel, and cement.
The interesting part is a dormant variance channel but an active tail-risk channel in the power sector.
SSRN: ssrn.com/abstract=699...
The interesting part is a dormant variance channel but an active tail-risk channel in the power sector.
SSRN: ssrn.com/abstract=699...
More complexity in model governance is not automatically more accurate.
The S&P 500 results make that caution concrete.
substack.com/@garchquant/...
The S&P 500 results make that caution concrete.
substack.com/@garchquant/...
S&P 500 Volatility Forecasting and the Limits of Dynamic Model Choice
Loss Sensitivity, Tail-Risk Evaluation, and Why DMA/DMS Do Not Automatically Win
substack.com
July 2, 2026 at 12:03 PM
More complexity in model governance is not automatically more accurate.
The S&P 500 results make that caution concrete.
substack.com/@garchquant/...
The S&P 500 results make that caution concrete.
substack.com/@garchquant/...
Same models, same sample, same rolling design—different loss function, different winner.
That is the most useful takeaway from this paper.
substack.com/@garchquant/...
That is the most useful takeaway from this paper.
substack.com/@garchquant/...
EUR/USD Volatility Forecasting: When the “Best” GARCH Depends on the Metric
Evidence from GARCH, GJR-GARCH, and EGARCH Under Multiple Loss Functions
substack.com
July 2, 2026 at 12:01 PM
Same models, same sample, same rolling design—different loss function, different winner.
That is the most useful takeaway from this paper.
substack.com/@garchquant/...
That is the most useful takeaway from this paper.
substack.com/@garchquant/...
A new paper examines how volatility model rankings vary with the loss function used for forecast evaluation.
In the current sample, EGARCH leads under MSE, QLIKE, and joint VaR–ES loss, whereas GJR-GARCH performs relatively better under MAE.
SSRN：
papers.ssrn.com/sol3/papers....
In the current sample, EGARCH leads under MSE, QLIKE, and joint VaR–ES loss, whereas GJR-GARCH performs relatively better under MAE.
SSRN：
papers.ssrn.com/sol3/papers....
Loss Function Sensitivity, Volatility Forecasting, and Dynamic Model Choice: Evidence from Foreign Exchange and Equity Markets
This paper examines how volatility model rankings depend on the choice of loss function when the true conditional variance is unobservable and forecast evaluati
papers.ssrn.com
June 30, 2026 at 6:30 AM
A new paper examines how volatility model rankings vary with the loss function used for forecast evaluation.
In the current sample, EGARCH leads under MSE, QLIKE, and joint VaR–ES loss, whereas GJR-GARCH performs relatively better under MAE.
SSRN：
papers.ssrn.com/sol3/papers....
In the current sample, EGARCH leads under MSE, QLIKE, and joint VaR–ES loss, whereas GJR-GARCH performs relatively better under MAE.
SSRN：
papers.ssrn.com/sol3/papers....
Distributional assumptions remain central to GARCH-based risk measurement. Gaussian innovations can understate tail risk when returns have heavy tails.
garchquant.substack.com/p/why-distri...
garchquant.substack.com/p/why-distri...
Why Distributional Assumptions Still Matter in GARCH-Based Risk Measurement
How VaR, heavy tails, and non-normal returns turn innovation distributions from a technical detail into a modeling decision.
garchquant.substack.com
June 22, 2026 at 9:59 AM
Distributional assumptions remain central to GARCH-based risk measurement. Gaussian innovations can understate tail risk when returns have heavy tails.
garchquant.substack.com/p/why-distri...
garchquant.substack.com/p/why-distri...
GARCH Quant’s latest post examines why asymmetric volatility modeling still matters. The difference in impulse responses between symmetric and asymmetric GARCH is shown clearly.
open.substack.com/pub/garchqua...
open.substack.com/pub/garchqua...
Why Asymmetry Still Matters in Volatility Modeling
How GJR-GARCH, EGARCH, and related specifications remain central when negative shocks move volatility differently from positive ones.
open.substack.com
June 19, 2026 at 4:55 AM
GARCH Quant’s latest post examines why asymmetric volatility modeling still matters. The difference in impulse responses between symmetric and asymmetric GARCH is shown clearly.
open.substack.com/pub/garchqua...
open.substack.com/pub/garchqua...
A selective narrative review on contemporary GARCH developments is now available. Core message: extension, not replacement.
open.substack.com/pub/garchqua...
open.substack.com/pub/garchqua...
GARCH Is Not Obsolete — It Has Simply Entered a More Complex Phase
A map of recent GARCH research across asymmetry, tail risk, dependence, robust estimation, and hybrid forecasting architectures.
open.substack.com
June 18, 2026 at 1:34 AM
A selective narrative review on contemporary GARCH developments is now available. Core message: extension, not replacement.
open.substack.com/pub/garchqua...
open.substack.com/pub/garchqua...
New synthesis from GARCH Quant: GARCH modeling is evolving through asymmetric responses, robust estimation, dynamic dependence, and hybrid architectures.
GARCH Quant shares a new review: Recent research extends GARCH rather than replaces it. Five key directions are discussed.
open.substack.com/pub/garchqua...
open.substack.com/pub/garchqua...
GARCH Is Not Obsolete — It Has Simply Entered a More Complex Phase
A map of recent GARCH research across asymmetry, tail risk, dependence, robust estimation, and hybrid forecasting architectures.
open.substack.com
June 18, 2026 at 1:32 AM
New synthesis from GARCH Quant: GARCH modeling is evolving through asymmetric responses, robust estimation, dynamic dependence, and hybrid architectures.
GARCH Quant shares a new review: Recent research extends GARCH rather than replaces it. Five key directions are discussed.
open.substack.com/pub/garchqua...
open.substack.com/pub/garchqua...
GARCH Is Not Obsolete — It Has Simply Entered a More Complex Phase
A map of recent GARCH research across asymmetry, tail risk, dependence, robust estimation, and hybrid forecasting architectures.
open.substack.com
June 18, 2026 at 1:31 AM
GARCH Quant shares a new review: Recent research extends GARCH rather than replaces it. Five key directions are discussed.
open.substack.com/pub/garchqua...
open.substack.com/pub/garchqua...
A clear structural framework for identifying where real constraints lie in complex supply chains. Worth reading:
open.substack.com/pub/amorecap...
open.substack.com/pub/amorecap...
Why the SpaceX Supply Chain Should Be Studied Through a Bottleneck Map
The point is not to list SpaceX beneficiaries, but to map the bottlenecks that will decide system throughput and profit distribution.
open.substack.com
June 16, 2026 at 7:12 AM
A clear structural framework for identifying where real constraints lie in complex supply chains. Worth reading:
open.substack.com/pub/amorecap...
open.substack.com/pub/amorecap...
Bottlenecks are not only disruptions. They are structural diagnostics.
open.substack.com/pub/amorecap...
open.substack.com/pub/amorecap...
What Bottlenecks Really Reveal
Why shortages are only the visible symptom of deeper dependencies.
open.substack.com
June 15, 2026 at 6:40 AM
Bottlenecks are not only disruptions. They are structural diagnostics.
open.substack.com/pub/amorecap...
open.substack.com/pub/amorecap...
New paper on CSI 300 5-min data. HARQ + Markov-switching GJR-GARCH improves volatility forecasts. Return predictability concentrates in low-vol regimes, with value mainly from careful implementation.
Full analysis:
open.substack.com/pub/garchqua...
Full analysis:
open.substack.com/pub/garchqua...
Volatility Forecasting Meets Machine Learning: State-Dependent Strategies in the Chinese Equity Market
A Two-Stage Framework on CSI 300 High-Frequency Data (arXiv:2606.09478)
open.substack.com
June 12, 2026 at 7:13 AM
New paper on CSI 300 5-min data. HARQ + Markov-switching GJR-GARCH improves volatility forecasts. Return predictability concentrates in low-vol regimes, with value mainly from careful implementation.
Full analysis:
open.substack.com/pub/garchqua...
Full analysis:
open.substack.com/pub/garchqua...
Working notes on implementing a five-layer framework inspired by Medallion. Shared some practical adjustments we made along the way.
I’ve been working on a five-layer framework inspired by Medallion, with adjustments in regime representation and position sizing under uncertainty.
These are working notes rather than final claims.
Full observations here:
open.substack.com/pub/garchqua...
These are working notes rather than final claims.
Full observations here:
open.substack.com/pub/garchqua...
Notes on Implementing a Five-Layer Framework Inspired by Medallion
Implementation Notes on Signal Detection, Regime Modeling, and Position Sizing
open.substack.com
June 11, 2026 at 8:43 AM
Working notes on implementing a five-layer framework inspired by Medallion. Shared some practical adjustments we made along the way.
On risk.
Volatility is not just a number to forecast — it is the primary input for robust position sizing, drawdown control, and regime-aware risk management.
This is where modelling translates into durable, repeatable quantitative decision making.
Volatility is not just a number to forecast — it is the primary input for robust position sizing, drawdown control, and regime-aware risk management.
This is where modelling translates into durable, repeatable quantitative decision making.
June 3, 2026 at 4:12 PM
On risk.
Volatility is not just a number to forecast — it is the primary input for robust position sizing, drawdown control, and regime-aware risk management.
This is where modelling translates into durable, repeatable quantitative decision making.
Volatility is not just a number to forecast — it is the primary input for robust position sizing, drawdown control, and regime-aware risk management.
This is where modelling translates into durable, repeatable quantitative decision making.
On signals:
All output derived directly from volatility state estimation and cross-asset cycle behaviour.
No price action pattern hunting, no overfitted heuristics. Every signal is rooted in calibrated, out-of-sample tested model output.
Structure over intuition.
All output derived directly from volatility state estimation and cross-asset cycle behaviour.
No price action pattern hunting, no overfitted heuristics. Every signal is rooted in calibrated, out-of-sample tested model output.
Structure over intuition.
June 2, 2026 at 6:42 AM
On signals:
All output derived directly from volatility state estimation and cross-asset cycle behaviour.
No price action pattern hunting, no overfitted heuristics. Every signal is rooted in calibrated, out-of-sample tested model output.
Structure over intuition.
All output derived directly from volatility state estimation and cross-asset cycle behaviour.
No price action pattern hunting, no overfitted heuristics. Every signal is rooted in calibrated, out-of-sample tested model output.
Structure over intuition.
GARCH QUANT is now on Bluesky.
Methods-first research on GARCH-family models, volatility dynamics, and robust financial econometrics, with occasional shorter notes on models, signals, and risk.
Methods-first research on GARCH-family models, volatility dynamics, and robust financial econometrics, with occasional shorter notes on models, signals, and risk.
June 1, 2026 at 3:42 AM
GARCH QUANT is now on Bluesky.
Methods-first research on GARCH-family models, volatility dynamics, and robust financial econometrics, with occasional shorter notes on models, signals, and risk.
Methods-first research on GARCH-family models, volatility dynamics, and robust financial econometrics, with occasional shorter notes on models, signals, and risk.