FEDS Paper: Linear and Nonlinear Econometric Models versus Machine-Learning Models: Evidence from Realized-Volatility Forecasting(Revised) — QuantCalc Tools Rate & Fee Analysis

FEDS Paper: Linear and Nonlinear Econometric Models versus Machine-Learning Models: Evidence from Realized-Volatility Forecasting(Revised) — QuantCalc Tools Rate & Fee Analysis (QuantCalc Tools Desk Note, 2026-10-31) — This briefing summarizes the supplied source material at https://www.federalreserve.gov/econres/feds/linear-and-nonlinear-econometric-models-against-machine-learning-models.htm. Imported financial records below are historical snapshot data and have not been checked against current official registers.

Answer-First Summary: Rehim Kilic

This paper examines which representations of persistence and nonlinearity are most useful for forecasting realized volatility and whether machine learning adds value beyond econometric models designed for long memory and regime dependence. We compare HAR, ARFIMA, threshold HAR, smooth-transition HAR, and Markov-switching HAR with XGBoost and several neural-network models for the S&P 500 and 40 U.S. equities. Models are evaluated under a baseline information set using realized-volatility history and an extended set of predictors selected by Elastic Net. Forecast performance is assessed using MSFE, MAE, QLIKE, Model Confidence Sets, Diebold–Mariano tests, realized utility, and filtered-historical-simulation VaR and expected shortfall. The results reveal a robust horizon-dependent ranking: Markov-switching HAR performs best at short horizons, ARFIMA generally leads at the monthly horizon, and the five-day horizon is intermediate. Machine-learning models sometimes improve on HAR but do not systematically outperform the broader econometric set. These findings are robust across estimation windows, re-estimation frequencies, richer information sets, log-volatility targets, and individual equities; stock-level results are also robust to an alternative realized-volatility measure. Overall, forecast performance depends primarily on capturing the persistence and nonlinear dynamics most relevant at each horizon.

Imported Financial Source Snapshot (Not Independently Verified)#

The following fields are copied from an imported historical snapshot. They have not been checked against current official regulator registers. Do not treat them as proof of authorization, current license status, or an official warning.

• Confidence (Confidence) — Imported score: 1.0/10 | Imported risk field: unknown | Source license records: [No license record in this snapshot]
• MOST (MOST) — Imported score: 1.0/10 | Imported risk field: unknown | Source license records: [No license record in this snapshot]

Core Analysis (QuantCalc Tools Rate & Fee Analysis)#

Rehim Kilic

This paper examines which representations of persistence and nonlinearity are most useful for forecasting realized volatility and whether machine learning adds value beyond econometric models designed for long memory and regime dependence. We compare HAR, ARFIMA, threshold HAR, smooth-transition HAR, and Markov-switching HAR with XGBoost and several neural-network models for the S&P 500 and 40 U.S. equities. Models are evaluated under a baseline information set using realized-volatility history and an extended set of predictors selected by Elastic Net. Forecast performance is assessed using MSFE, MAE, QLIKE, Model Confidence Sets, Diebold–Mariano tests, realized utility, and filtered-historical-simulation VaR and expected shortfall. The results reveal a robust horizon-dependent ranking: Markov-switching HAR performs best at short horizons, ARFIMA generally leads at the monthly horizon, and the five-day horizon is intermediate. Machine-learning models sometimes improve on HAR but do not systematically outperform the broader econometric set. These findings are robust across estimation windows, re-estimation frequencies, richer information sets, log-volatility targets, and individual equities; stock-level results are also robust to an alternative realized-volatility measure. Overall, forecast performance depends primarily on capturing the persistence and nonlinear dynamics most relevant at each horizon.

Imported Financial Source Snapshot (Not Independently Verified)#

The following fields are copied from an imported historical snapshot. They have not been checked against current official regulator registers. Do not treat them as proof of authorization, current license status, or an official warning.

• Confidence (Confidence) — Imported score: 1.0/10 | Imported risk field: unknown | Source license records: [No license record in this snapshot]
• MOST (MOST) — Imported score: 1.0/10 | Imported risk field: unknown | Source license records: [No license record in this snapshot]

Key Quantitative & Regulatory Takeaways#

• Primary Subject: FEDS Paper: Linear and Nonlinear Econometric Models versus Machine-Learning Models: Evidence from Realized-Volatility Forecasting(Revised)
• Figures & Entities in Supplied Text: no source facts provided
• Provided Source: Source link

Structured Fact & Metric Comparison Table#

Frequently Asked Questions (FAQ)#

What are the exact figures and regulatory details reported in FEDS Paper: Linear and Nonlinear Econometric Models versus Machine-Learning Models: Evidence from Realized-Volatility Forecasting(Revised)?#

The supplied text states: Rehim Kilic

This paper examines which representations of persistence and nonlinearity are most useful for forecasting realized volatility and whether machine learning adds value beyond econometric models designed for long memory and regime dependence. We compare HAR, ARFIMA, threshold HAR, smooth-transition HAR, and Markov-switching HAR with XGBoost and several neural-network models for the S&P 500 and 40 U.S. equities. Models are evaluated under a baseline information set using realized-volatility history and an extended set of predictors selected by Elastic Net. Forecast performance is assessed using MSFE, MAE, QLIKE, Model Confidence Sets, Diebold–Mariano tests, realized utility, and filtered-historical-simulation VaR and expected shortfall. The results reveal a robust horizon-dependent ranking: Markov-switching HAR performs best at short horizons, ARFIMA generally leads at the monthly horizon, and the five-day horizon is intermediate. Machine-learning models sometimes improve on HAR but do not systematically outperform the broader econometric set. These findings are robust across estimation windows, re-estimation frequencies, richer information sets, log-volatility targets, and individual equities; stock-level results are also robust to an alternative realized-volatility measure. Overall, forecast performance depends primarily on capturing the persistence and nonlinear dynamics most relevant at each horizon.

Imported Financial Source Snapshot (Not Independently Verified)#

The following fields are copied from an imported historical snapshot. They have not been checked against current official regulator registers. Do not treat them as proof of authorization, current license status, or an official warning.

• Confidence (Confidence) — Imported score: 1.0/10 | Imported risk field: unknown | Source license records: [No license record in this snapshot]
• MOST (MOST) — Imported score: 1.0/10 | Imported risk field: unknown | Source license records: [No license record in this snapshot]

Has QuantCalc Tools independently verified this tools data?#

No independent verification was performed by this rewrite step. The supplied source link is https://www.federalreserve.gov/econres/feds/linear-and-nonlinear-econometric-models-against-machine-learning-models.htm.

Imported Financial Source Snapshot (Not Independently Verified)#

The following fields are copied from an imported historical snapshot. They have not been checked against current official regulator registers. Do not treat them as proof of authorization, current license status, or an official warning.

• Confidence (Confidence) — Imported score: 1.0/10 | Imported risk field: unknown | Source license records: [No license record in this snapshot]
• MOST (MOST) — Imported score: 1.0/10 | Imported risk field: unknown | Source license records: [No license record in this snapshot]

Quantitative Verification & Analytical Model#