arxiv: 2605.12099 · v1 · submitted 2026-05-12 · 📊 stat.ME · q-fin.ST

Recognition: no theorem link

Bayesian Dynamic Modeling of Realized Volatility in Financial Asset Price Forecasting

Mike West, Patrick Woitschig

Pith reviewed 2026-05-13 03:53 UTC · model grok-4.3

classification 📊 stat.ME q-fin.ST

keywords Bayesian dynamic modelsrealized volatilityasset price forecastingleverage effectsdynamic gamma processfinancial time seriesS&P ETFsvolatility feedback

0 comments

The pith

A Bayesian model coupling realized volatility with price dynamics improves equity return forecasts by capturing leverage and feedback effects.

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

The paper develops Bayesian dynamic models that jointly handle asset prices and realized volatility time series for financial forecasting. It combines a dynamic gamma process for volatility, built from high-frequency intraday data, with standard dynamic linear models for daily prices or returns. This setup represents volatility leverage and feedback in reduced form by feeding volatility proxies into the price models. The approach supports simple sequential Bayesian updating and simulation-based forecasting at low computational cost. Empirical work on S&P sector ETFs shows forecast gains over models that omit volatility information and yields insights into how volatility influences price movements.

Core claim

The central claim is that integrating a novel dynamic gamma process model for realized volatility with Bayesian dynamic linear models for prices creates a reduced-form representation of leverage and feedback effects. This coupling of intraday volatility data with daily price series enables efficient conjugate Bayesian filtering, monitoring, and improved forecasting of equity returns, as shown in applications to multiple S&P sector ETFs.

What carries the argument

The dynamic gamma process for realized volatility proxies integrated into conditional Bayesian dynamic linear models for prices, which tracks volatility fluctuations and transmits leverage effects to the price forecasts.

If this is right

Forecast accuracy for asset returns improves relative to models without the volatility component.
The structure scales to multivariate price series for portfolio construction and risk management.
Sequential analysis remains computationally straightforward with negligible added cost.
Contextual understanding of volatility leverage and feedback effects becomes available from the fitted models.

Where Pith is reading between the lines

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

The same reduced-form coupling could be tested on other asset classes such as currencies or commodities to check whether leverage patterns generalize.
Real-time monitoring of the gamma process parameters might support adaptive trading rules that adjust exposure when volatility feedback strengthens.
Adding macroeconomic covariates to the price models could be explored to isolate whether the volatility proxies retain explanatory power.

Load-bearing premise

The reduced-form coupling of realized volatility proxies into the price models accurately captures leverage and feedback without omitted-variable bias or misspecification in the gamma process dynamics.

What would settle it

Out-of-sample tests on the same S&P ETF data showing that forecast errors or log predictive scores from the new models are not materially better than those from standard dynamic linear models that ignore realized volatility.

Figures

Figures reproduced from arXiv: 2605.12099 by Mike West, Patrick Woitschig.

**Figure 2.** Figure 2: Cumulative log Bayes factors over the post-2010 period for each of the 9 ETFs and the [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗

**Figure 3.** Figure 3: Cumulative log Bayes factors over the post-2010 period for 3 of the ETFs: [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗

**Figure 4.** Figure 4: Filtered trajectories of 2 state vector elements [PITH_FULL_IMAGE:figures/full_fig_p014_4.png] view at source ↗

**Figure 5.** Figure 5: As in Figure [PITH_FULL_IMAGE:figures/full_fig_p015_5.png] view at source ↗

**Figure 6.** Figure 6: Filtered trajectories of all 4 elements of the state vector [PITH_FULL_IMAGE:figures/full_fig_p016_6.png] view at source ↗

**Figure 7.** Figure 7: Trajectories of filtered posterior medians of the RVL-DLM contemporaneous regression [PITH_FULL_IMAGE:figures/full_fig_p017_7.png] view at source ↗

read the original abstract

We present a new class of Bayesian dynamic models for bivariate price-realized volatility time series in financial forecasting. A novel dynamic gamma process model adopted for realized volatility is integrated with traditional Bayesian dynamic linear models (DLMs) for asset price series. This represents reduced-form volatility leverage and feedback effects through use of realized volatility proxies in conditional DLMs for prices or returns, coupled with the synthesis of higher frequency data to track and anticipate volatility fluctuations. Analysis is computationally straightforward, extending conjugate-form Bayesian analyses for sequential filtering and model monitoring with simple and direct simulation for forecasting. A main applied setting is equity return forecasting with daily prices and realized volatility from high-frequency, intraday data. Detailed empirical studies of multiple S&P sector ETFs highlight the improvements achievable in asset price forecasting relative to standard models and deliver contextual insights on the nature and practical relevance of volatility leverage and feedback effects. The analytic structure and negligible extra computational cost will enable scaling to higher dimensions for multivariate price series forecasting for decouple/recouple portfolio construction and risk management applications.

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

read the letter

This combines a dynamic gamma process for realized volatility with DLMs for prices in a reduced-form bivariate setup and reports forecast gains on S&P ETFs, but the empirical edge depends on how the benchmarks were chosen. The modeling choice itself is the clearest addition: they keep the volatility side conjugate via the gamma process and feed realized-vol proxies directly into the conditional price DLMs to capture leverage and feedback without needing a full joint specification. That keeps filtering and monitoring simple, and the forecasting step is just forward simulation, which is cheap enough to scale. The empirical part runs the model on several sector ETFs using intraday data for the volatility series and claims better price forecasts plus some readable insights on how volatility effects play out in practice. Those are the parts that could actually get used by someone already working with DLMs or Bayesian time-series tools in finance. The main soft spot is the lack of detail on what exactly the comparison models were and whether the out-of-sample windows or data choices were fixed in advance. If the baselines are just plain DLMs or basic GARCH without any volatility coupling, the reported gains are plausible but not yet convincing as a general improvement. The reduced-form proxy approach also leaves room for omitted factors that jointly move returns and volatility, and it would help to see whether the paper tests sensitivity to that or just presents the point estimates. Overall the technical structure holds together and the application is concrete rather than abstract. This is for quant teams or researchers already using Bayesian dynamic models who want a practical way to bring high-frequency volatility information into daily price forecasts. It is worth sending to peer review because the core construction is coherent and the data exercise is real, even if the referee will likely ask for tighter benchmark reporting and a couple of robustness checks.

Referee Report

2 major / 2 minor

Summary. The paper claims to introduce a new class of Bayesian dynamic models for bivariate price-realized volatility time series in financial forecasting. It integrates a novel dynamic gamma process model for realized volatility with traditional Bayesian dynamic linear models (DLMs) for asset prices or returns. This captures reduced-form volatility leverage and feedback effects via realized volatility proxies in conditional DLMs, with synthesis of higher-frequency data. The approach maintains conjugacy for straightforward sequential filtering, model monitoring, and simulation-based forecasting. Empirical studies on multiple S&P sector ETFs are said to show improved asset price forecasting relative to standard models and to deliver insights on the nature and practical relevance of leverage and feedback effects. The structure is presented as scalable to multivariate settings for portfolio and risk management.

Significance. If the empirical gains prove robust, the work provides a computationally efficient Bayesian framework for fusing intraday volatility information into daily price forecasts while preserving conjugacy and enabling direct simulation. The negligible extra cost and scalability to higher dimensions for decouple/recouple applications represent practical strengths. The reduced-form treatment of leverage/feedback could offer contextual insights useful for risk management if the modeling assumptions hold.

major comments (2)

[§5] §5 (Empirical Studies): The central claim of improved forecasting performance on S&P sector ETFs and insights on leverage/feedback rests on comparisons to 'standard models,' but the manuscript provides insufficient detail on benchmark specifications (e.g., GARCH, HAR, or univariate DLMs), out-of-sample periods, loss functions, or statistical tests of differences. This directly affects evaluation of whether reported gains reflect genuine improvements or post-hoc choices, as highlighted by the stress-test concern.
[§3.2] §3.2 (Reduced-form coupling): The bivariate DLM-gamma construction represents leverage and feedback solely through realized volatility proxies in the conditional DLMs. No explicit checks for omitted-variable bias (e.g., macro or order-flow factors) or sensitivity analyses on the dynamic gamma process specification are described; if these are absent, the forecast gains and interpretive conclusions on leverage/feedback may not be isolated from model artifact.

minor comments (2)

[Abstract and §6] The abstract states 'negligible extra computational cost' but the main text lacks any timing benchmarks or scaling experiments to support this for multivariate extensions.
[§3.1] Notation for the dynamic gamma process parameters (shape, rate, evolution) would benefit from a consolidated table or explicit listing to aid reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed comments on our manuscript. We address each major comment point by point below, indicating where we will revise the paper to improve clarity, transparency, and robustness while preserving the core contributions of the Bayesian dynamic modeling framework.

read point-by-point responses

Referee: [§5] §5 (Empirical Studies): The central claim of improved forecasting performance on S&P sector ETFs and insights on leverage/feedback rests on comparisons to 'standard models,' but the manuscript provides insufficient detail on benchmark specifications (e.g., GARCH, HAR, or univariate DLMs), out-of-sample periods, loss functions, or statistical tests of differences. This directly affects evaluation of whether reported gains reflect genuine improvements or post-hoc choices, as highlighted by the stress-test concern.

Authors: We agree that greater specificity on the empirical setup is needed to allow readers to fully evaluate the reported forecast improvements. In the revised manuscript, we will expand Section 5 with explicit descriptions of all benchmark models (including exact GARCH and HAR specifications, as well as the univariate DLM baselines), the precise out-of-sample periods and rolling-window evaluation scheme, the loss functions employed (MSE, QLIKE, and others), and formal statistical tests such as Diebold-Mariano tests for pairwise forecast accuracy differences. We will also add the requested stress-test robustness checks using alternative market regimes and volatility proxies. These additions will be presented in new tables and text to demonstrate that the gains are not artifacts of post-hoc choices. revision: yes
Referee: [§3.2] §3.2 (Reduced-form coupling): The bivariate DLM-gamma construction represents leverage and feedback solely through realized volatility proxies in the conditional DLMs. No explicit checks for omitted-variable bias (e.g., macro or order-flow factors) or sensitivity analyses on the dynamic gamma process specification are described; if these are absent, the forecast gains and interpretive conclusions on leverage/feedback may not be isolated from model artifact.

Authors: The reduced-form coupling is intentional, as the paper focuses on a computationally conjugate and scalable framework that directly incorporates realized-volatility proxies to capture leverage and feedback effects without requiring a full structural model. However, we acknowledge the value of additional checks. In revision, we will add sensitivity analyses that vary key hyperparameters of the dynamic gamma process (e.g., evolution variance and shape parameters) and report the resulting impact on forecast performance and leverage-effect estimates. We will also include a brief discussion of potential omitted variables such as macro factors, noting that their explicit inclusion lies outside the reduced-form scope but that the current specification remains robust within the class of models that rely on volatility proxies. These additions will help readers assess whether the reported insights are sensitive to modeling choices. revision: partial

Circularity Check

0 steps flagged

No significant circularity; empirical claims rest on external data validation

full rationale

The paper introduces a modeling framework that couples DLMs for prices with a dynamic gamma process for realized volatility to capture leverage/feedback in reduced form. Forecasting proceeds via conjugate sequential updates and direct simulation, with performance evaluated on external S&P sector ETF data. No step equates a prediction to its own fitted inputs by construction, renames a known result, or relies on a self-citation chain for the core result. The derivation is self-contained against standard benchmarks and external time series.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review provides no explicit list of fitted parameters or invented entities; the dynamic gamma process itself is presented as a modeling choice whose distributional assumptions are not detailed here.

pith-pipeline@v0.9.0 · 5471 in / 1149 out tokens · 86481 ms · 2026-05-13T03:53:58.035233+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

156 extracted references · 156 canonical work pages

[1]

High-frequency realized stochastic volatility model , year =

Toshiaki Watanabe and Jouchi Nakajima , journal =. High-frequency realized stochastic volatility model , year =

work page
[2]

Journal of Financial Econometrics , year =

Bollerslev, Tim and Litvinova, Julia and Tauchen, George , title =. Journal of Financial Econometrics , year =

work page
[3]

Forecasting Daily Volatility of Stock Price Index Using Daily Returns and Realized Volatility , year =

Makoto Takahashi and Toshiaki Watanabe and Yasuhiro Omori , journal =. Forecasting Daily Volatility of Stock Price Index Using Daily Returns and Realized Volatility , year =

work page
[4]

Irie and M

K. Irie and M. West , journal =. Bayesian emulation for multi-step optimization in decision problems , year =. doi:10.1214/18-BA1105 , url =

work page doi:10.1214/18-ba1105
[5]

West , journal =

M. West , journal =. Mixture models,. 1993 , pages =

work page 1993
[6]

, publisher =

Taylor, Stephen J. , publisher =. Modelling Financial Time Series , year =

work page
[7]

and Rossi, Peter E

Jacquier, Eric and Polson, Nicholas G. and Rossi, Peter E. , journal =. Bayesian Analysis of Stochastic Volatility Models , year =

work page
[8]

Statistical Aspects of ARCH and Stochastic Volatility , year =

Shephard, Neil , booktitle =. Statistical Aspects of ARCH and Stochastic Volatility , year =

work page
[9]

and Hansen, Peter R

Barndorff-Nielsen, Ole E. and Hansen, Peter R. and Lunde, Asger and Shephard, Neil , journal =. Designing Realized Kernels to Measure the Ex-Post Variation of Equity Prices in the Presence of Noise , year =

work page
[10]

and Bollerslev, Tim and Mikkelsen, Hans Ole , journal =

Baillie, Richard T. and Bollerslev, Tim and Mikkelsen, Hans Ole , journal =. Fractionally Integrated Generalized Autoregressive Conditional Heteroskedasticity , year =

work page
[11]

The Analysis of Stochastic Volatility in the Presence of Daily Realized Measures , year =

Koopman, Siem Jan and Scharth, Marcel , journal =. The Analysis of Stochastic Volatility in the Presence of Daily Realized Measures , year =

work page
[12]

, journal =

Engle, Robert F. , journal =. New Frontiers for. 2002 , number =

work page 2002
[13]

and Gallo, Giampiero M

Engle, Robert F. and Gallo, Giampiero M. , journal =. A Multiple Indicators Model for Volatility Using Intra-Daily Data , year =

work page
[14]

and Kotz, Samuel and Balakrishnan, N

Johnson, Norman L. and Kotz, Samuel and Balakrishnan, N. , publisher =. Continuous Univariate Distributions, Volume 2 , year =

work page
[15]

Realising the Future: Forecasting with High-Frequency-Based Volatility (

Shephard, Neil and Sheppard, Kevin , journal =. Realising the Future: Forecasting with High-Frequency-Based Volatility (. 2010 , number =

work page 2010
[16]

and Shephard, Neil , journal =

Pitt, Michael K. and Shephard, Neil , journal =. Filtering via Simulation: Auxiliary Particle Filters , year =

work page
[17]

Pole, Andy and West, Mike and Harrison, Jeff , publisher =. Applied. 1994 , address =

work page 1994
[18]

Prado, Raquel and Ferreira, Marco A. R. and West, Mike , publisher =. Time Series: Modeling, Computation, and Inference , year =

work page
[19]

Xu , note =

C. Xu , note =. The. 2024 , url =

work page 2024
[20]

, journal =

Mulder, Joris and Pericchi, Luis R. , journal =. The Matrix-. 2018 , number =

work page 2018
[21]

J. R. M. Hosking , journal =. Fractional Differencing , year =

work page
[22]

Granger, C. W. J. and Joyeux, Roselyne , journal =. AN INTRODUCTION TO LONG-MEMORY TIME SERIES MODELS AND FRACTIONAL DIFFERENCING , year =. doi:https://doi.org/10.1111/j.1467-9892.1980.tb00297.x , eprint =

work page doi:10.1111/j.1467-9892.1980.tb00297.x 1980
[23]

Long memory stochastic volatility : A bayesian approach , year =

Ngai Hang Chan and Giovanni Petris , journal =. Long memory stochastic volatility : A bayesian approach , year =. doi:10.1080/03610920008832549 , eprint =

work page doi:10.1080/03610920008832549
[24]

Forecasting Daily Volatility of Stock Price Index Using Daily Returns and Realized Volatility , year =

Makoto Takahashi and Toshiaki Watanabe and Yasuhiro Omori , journal =. Forecasting Daily Volatility of Stock Price Index Using Daily Returns and Realized Volatility , year =. doi:https://doi.org/10.1016/j.ecosta.2021.08.002 , keywords =

work page doi:10.1016/j.ecosta.2021.08.002 2021
[25]

Volatility and quantile forecasts by realized stochastic volatility models with generalized hyperbolic distribution , year =

Makoto Takahashi and Toshiaki Watanabe and Yasuhiro Omori , journal =. Volatility and quantile forecasts by realized stochastic volatility models with generalized hyperbolic distribution , year =. doi:https://doi.org/10.1016/j.ijforecast.2015.07.005 , keywords =

work page doi:10.1016/j.ijforecast.2015.07.005 2015
[26]

Studies of Stock Price Volatility Changes , year =

Black, Fischer , booktitle =. Studies of Stock Price Volatility Changes , year =

work page
[27]

Localized Realized Volatility Modelling , year =

Chen, Ying and Härdle, Wolfgang Karl and Pigorsch, Uta , journal =. Localized Realized Volatility Modelling , year =

work page
[28]

Stochastic volatility model with leverage and asymmetrically heavy-tailed error using

Jouchi Nakajima and Yasuhiro Omori , journal =. Stochastic volatility model with leverage and asymmetrically heavy-tailed error using. 2012 , issn =. doi:https://doi.org/10.1016/j.csda.2010.07.012 , keywords =

work page doi:10.1016/j.csda.2010.07.012 2012
[29]

Leverage, heavy-tails and correlated jumps in stochastic volatility models , year =

Jouchi Nakajima and Yasuhiro Omori , journal =. Leverage, heavy-tails and correlated jumps in stochastic volatility models , year =. doi:https://doi.org/10.1016/j.csda.2008.03.015 , url =

work page doi:10.1016/j.csda.2008.03.015 2008
[30]

Stochastic volatility with leverage: Fast and efficient likelihood inference , year =

Yasuhiro Omori and Siddhartha Chib and Neil Shephard and Jouchi Nakajima , journal =. Stochastic volatility with leverage: Fast and efficient likelihood inference , year =. doi:https://doi.org/10.1016/j.jeconom.2006.07.008 , keywords =

work page doi:10.1016/j.jeconom.2006.07.008 2006
[31]

Measuring and Modeling Risk Using High-Frequency Data , year =

H. Measuring and Modeling Risk Using High-Frequency Data , year =. Applied Quantitative Finance , doi =

work page
[32]

and Hansen, Peter R

Barndorff‐Nielsen, Ole E. and Hansen, Peter R. and Lunde, Asger and Shephard, Neil , journal =. Realized Kernels in Practice: Trades and Quotes , year =

work page
[33]

Combined Parameter and State Estimation in Simulation-Based Filtering , year =

Liu, Jane and West, Mike , editor =. Combined Parameter and State Estimation in Simulation-Based Filtering , year =. Sequential Monte Carlo Methods in Practice , doi =

work page
[34]

Gordon and D.J

N.J. Gordon and D.J. Salmond and A.F.M. Smith , journal =. Novel approach to nonlinear/non-Gaussian Bayesian state estimation , year =. doi:10.1049/ip-f-2.1993.0015 , eprint =

work page doi:10.1049/ip-f-2.1993.0015 1993
[35]

, journal =

Christie, Andrew A. , journal =. The Stochastic Behavior of Common Stock Variances , year =

work page
[36]

and Ng, Victor K

Engle, Robert F. and Ng, Victor K. , journal =. Measuring and Testing the Impact of News on Volatility , year =

work page
[37]

Ameen, J. R. M. and Harrison, P. J. , booktitle =. Normal Discount. 1985 , address =

work page 1985
[38]

Szerszen , institution =

Dobrislav Dobrev and Pawel J. Szerszen , institution =. The Information Content of High-Frequency Data for Estimating Equity Return Models and Forecasting Risk , year =

work page
[39]

The Economic Value of Realized Volatility: Using High-Frequency Returns for Option Valuation , year =

Peter Christoffersen and Bruno Feunou and Kris Jacobs and Nour Meddahi , journal =. The Economic Value of Realized Volatility: Using High-Frequency Returns for Option Valuation , year =

work page
[40]

Leverage Effects in Stochastic Volatility with Time‐Varying Persistence , year =

Nakajima, Jouchi and Omori, Yasuhiro , journal =. Leverage Effects in Stochastic Volatility with Time‐Varying Persistence , year =

work page
[41]

and Hautsch, Nikolaus and Mihoci, Andrija , journal =

Härdle, Wolfgang K. and Hautsch, Nikolaus and Mihoci, Andrija , journal =. Local Adaptive Multiplicative Error Models for High-Frequency Forecasts , year =. doi:https://doi.org/10.1002/jae.2376 , eprint =

work page doi:10.1002/jae.2376
[42]

GENERALIZED AUTOREGRESSIVE SCORE MODELS WITH APPLICATIONS , year =

Creal, Drew and Koopman, Siem Jan and Lucas, André , journal =. GENERALIZED AUTOREGRESSIVE SCORE MODELS WITH APPLICATIONS , year =. doi:https://doi.org/10.1002/jae.1279 , eprint =

work page doi:10.1002/jae.1279
[43]

Engle and Giampiero M

Robert F. Engle and Giampiero M. Gallo , journal =. A multiple indicators model for volatility using intra-daily data , year =. doi:https://doi.org/10.1016/j.jeconom.2005.01.018 , keywords =

work page doi:10.1016/j.jeconom.2005.01.018 2005
[44]

and Scott, James G

Polson, Nicholas G. and Scott, James G. , journal =. On the Half-Cauchy Prior for a Global Scale Parameter , year =

work page
[45]

Prior distributions for variance parameters in hierarchical models (Comment on article by Browne and Draper) , year =

Gelman, Andrew , journal =. Prior distributions for variance parameters in hierarchical models (Comment on article by Browne and Draper) , year =

work page
[46]

and Bollerslev, TIM , journal =

Andersen, TORBEN G. and Bollerslev, TIM , journal =. Heterogeneous Information Arrivals and Return Volatility Dynamics: Uncovering the Long-Run in High Frequency Returns , year =. doi:https://doi.org/10.1111/j.1540-6261.1997.tb02722.x , eprint =

work page doi:10.1111/j.1540-6261.1997.tb02722.x 1997
[47]

A Gamma Filter for Positive Parameter Estimation , year =

Govaers, Felix and Alqaderi, Hosam , booktitle =. A Gamma Filter for Positive Parameter Estimation , year =. doi:10.1109/MFI49285.2020.9235265 , keywords =

work page doi:10.1109/mfi49285.2020.9235265 2020
[48]

Polson and Peter E

Eric Jacquier and Nicholas G. Polson and Peter E. Rossi , journal =. Bayesian Analysis of Stochastic Volatility Models , year =

work page
[49]

Statistics for Long-Memory Processes , year =

Beran, Jan , publisher =. Statistics for Long-Memory Processes , year =. doi:10.1201/9780203738481 , url =

work page doi:10.1201/9780203738481
[50]

Local scale models: State space alternative to integrated GARCH processes , year =

Neil Shephard , journal =. Local scale models: State space alternative to integrated GARCH processes , year =. doi:https://doi.org/10.1016/0304-4076(94)90043-4 , keywords =

work page doi:10.1016/0304-4076(94)90043-4
[51]

Realized GARCH:

Hansen, Peter Reinhard and Huang, Zhuo and Shek, Howard Howan Stephen , month = oct, note =. Realized GARCH:. 2010 , day =. doi:10.2139/ssrn.1533475 , url =

work page doi:10.2139/ssrn.1533475 2010
[52]

Realising the future: forecasting with high-frequency-based volatility (HEAVY) models , year =

Shephard, Neil and Sheppard, Kevin , journal =. Realising the future: forecasting with high-frequency-based volatility (HEAVY) models , year =. doi:https://doi.org/10.1002/jae.1158 , eprint =

work page doi:10.1002/jae.1158
[53]

2006 , number =

Andrew Gelman , journal =. 2006 , number =. doi:10.1214/06-BA117A , keywords =

work page doi:10.1214/06-ba117a 2006
[54]

West , journal =

M. West , journal =. Bayesian forecasting of multivariate time series:. 2020 , pages =

work page 2020
[55]

Carvalho and P

Amir Bashir and Carlos M. Carvalho and P. Richard Hahn and M. Beatrix Jones , journal =. Post-processing posteriors over precision matrices to produce sparse graph estimates , year =

work page
[56]

Carvalho and Hedibert F

Henrique Bolfarine and Carlos M. Carvalho and Hedibert F. Lopes and Jared S. Murray , journal =. Decoupling shrinkage and selection in. 2024 , pages =. doi:10.1214/22-BA1349 , keywords =

work page doi:10.1214/22-ba1349 2024
[57]

Comparative politics and the synthetic control method , year =

Abadie, Alberto and Diamond, Alexis and Hainmueller, Jens , journal =. Comparative politics and the synthetic control method , year =

work page
[58]

Synthetic control methods for comparative case studies:

Abadie, Alberto and Diamond, Alexis and Hainmueller, Jens , journal =. Synthetic control methods for comparative case studies:. 2010 , pages =

work page 2010
[59]

Li and G

K. Li and G. Tierney and C. Hellmayr and M. West , journal =. Compositional dynamic modelling for counterfactual prediction in multivariate time series , year =. doi:10.1002/asmb.2908 , url =

work page doi:10.1002/asmb.2908
[60]

West and P

M. West and P. J. Harrison , journal =. Subjective intervention in formal models , year =

work page
[61]

L. F. Gruber and M. West , journal =. 2016 , pages =. doi:10.1214/15-BA946 , url =

work page doi:10.1214/15-ba946 2016
[62]

L. F. Gruber and M. West , journal =. Bayesian forecasting and scalable multivariate volatility analysis using simultaneous graphical dynamic linear models , year =. doi:10.1016/j.ecosta.2017.03.003 , owner =

work page doi:10.1016/j.ecosta.2017.03.003 2017
[63]

Bayesian forecasting of stock Returns on the

Nelson Kyakutwika and Bruce Bartlett , journal =. Bayesian forecasting of stock Returns on the. 2024 , pages =

work page 2024
[64]

Quantitative Finance , title =

Griveau-Billion, Th. Quantitative Finance , title =. 2021 , pages =

work page 2021
[65]

G. C. Tiao and R. S. Tsay , journal =. Model specification in multivariate time series , year =

work page
[66]

Yoshida and M

R. Yoshida and M. West , journal =. Bayesian learning in sparse graphical factor models via annealed entropy , year =

work page
[67]

A. J. Cron and M. West , booktitle =. Models of random sparse eigenmatrices and. 2016 , editor =

work page 2016
[68]

Lopes, H. F. and C. M. Carvalho , journal =. Factor stochastic volatility with time varying loadings and. 2007 , pages =

work page 2007
[69]

Lopes and Robert E

Hedibert F. Lopes and Robert E. McCulloch and Ruey S. Tsay , journal =. Parsimony inducing priors for large scale state–space models , year =

work page
[70]

Primiceri, G. E. , journal =. Time varying structural vector autoregressions and monetary policy , year =

work page
[71]

Bayesian Analysis , title =

Sylvia Fr. Bayesian Analysis , title =. 2024 , doi =

work page 2024
[72]

Large time-varying parameter

Gary Koop and Dimitris Korobilis , journal =. Large time-varying parameter. 2013 , pages =. doi:https://doi.org/10.1016/j.jeconom.2013.04.007 , url =

work page doi:10.1016/j.jeconom.2013.04.007 2013
[73]

Bayesian Multivariate Time Series Methods for Empirical Macroeconomics , year =

Gary Koop and Dimitris Korobilis , journal =. Bayesian Multivariate Time Series Methods for Empirical Macroeconomics , year =

work page
[74]

Journal of Applied Econometrics , title =

Marta Ba. Journal of Applied Econometrics , title =. 2010 , pages =

work page 2010
[75]

Nakajima and M

J. Nakajima and M. West , journal =. Bayesian dynamic factor models:. 2013 , pages =. doi:10.1093/jjfinec/nbs013 , url =

work page doi:10.1093/jjfinec/nbs013 2013
[76]

Nakajima and M

J. Nakajima and M. West , journal =. Dynamics and sparsity in latent threshold factor models:. 2017 , pages =. doi:10.1214/17-BJPS364 , url =

work page doi:10.1214/17-bjps364 2017
[77]

Nakajima and M

J. Nakajima and M. West , journal =. Bayesian analysis of latent threshold dynamic models , year =. doi:10.1080/07350015.2012.747847 , url =

work page doi:10.1080/07350015.2012.747847 2012
[78]

Nakajima and M

J. Nakajima and M. West , journal =. Dynamic network signal processing using latent threshold models , year =. doi:10.1016/j.dsp.2015.04.008 , url =

work page doi:10.1016/j.dsp.2015.04.008 2015
[79]

C. M. Carvalho and J. E. Lucas and Q. Wang and J. Chang and J. R. Nevins and M. West , journal =. High-dimensional sparse factor modelling--. 2008 , pages =

work page 2008
[80]

Comparative politics and the synthetic control method revisited:

Klößner, Stefan and Kaul, Ashok and Pfeifer, Gregor and Schieler, Manuel , journal =. Comparative politics and the synthetic control method revisited:. 2018 , pages =

work page 2018

Showing first 80 references.