Pith. sign in

REVIEW 3 major objections 4 minor 199 references

Clustered Local Projections for Short and Ultra-Short Time Series -- A Hierarchical Bayesian Framework

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read The paper claims that estimating local projections jointly across a panel of related time series, with units clustered by similarity of their impulse responses, roughly halves impulse-response estimation error for short series while…

desk verdict New combination of hierarchical pooling, sparse mixture, and imputation for short unbalanced panels; simulation gains are real, but the prior-calibration K sensitivity and missing code/base comparisons need addressing. read the letter →

arxiv 2608.04631 v1 pith:23DOS6ZK submitted 2026-08-05 econ.EM stat.ME

classification econ.EMstat.ME MSC 62F1562M1062H30
keywords localprojectionsimpulseresponsefunctionshierarchicalBayesianpoolingsparsefinitemixturesunbalancedpanelsshorttimeseriesinflationdynamicscoveragecorrection
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Local projections are a standard way to trace how an economic shock moves a series forward, but they are data-hungry: each horizon consumes another observation at the end of the sample, so short series yield unreliable responses. This paper claims that estimating the local projections of many related series jointly, with a hierarchical Bayesian model that clusters series by the similarity of their impulse response profiles, lets short series borrow precision from longer ones without forcing every series into the same mold. In simulations calibrated to monthly US macro data, the authors find the model roughly halves the mean absolute error of impulse-response estimates for short series relative to standard series-by-series local projections, while long series' estimates stay essentially unchanged. The same mechanism restores near-nominal interval coverage and extends responses to horizons beyond a short series' own sample length. If the claim holds, impulse-response analysis becomes feasible for the ultra-short series, some with 25 to 60 usable observations, that current practice cannot handle.

What carries the argument

The engine is a sparse finite mixture pooling prior placed directly on the local projection coefficients of the response equation. Each series is assigned, through a categorical allocation $z_i$ that is common across horizons, to one of up to $S=8$ clusters; the horizon-$h$ response is then drawn as $\rho_{i,h} \mid z_i = s \sim N(\mu_{s,h}, \tau_{s,h}^2)$, and the cluster means themselves are pulled toward a common population center $m_h$, so that even a singleton cluster keeps borrowing weakly from the full panel. A sparse Dirichlet prior on the mixture weights, with its concentration estimated from the data, empties redundant components, so the effective number of clusters is learned rather than fixed. Because the prior is conditionally conjugate, the posterior mean of each response remains a transparent weighted average of its own data and its cluster target; and a post-processing sandwich rescaling that replaces the posterior variance with a HAC long-run variance pooled over the same cluster partition converts the posterior credible sets into intervals with correct frequentist coverage.

What would settle it

Re-run the short-sample simulation with the shock loadings left at their unadjusted calibrated size and with unit-specific impulse profiles drawn from a continuum rather than four strongly separated clusters; if the mean absolute error of the short-series estimates falls by less than the claimed factor of two, or the coverage of the nominal 90 percent intervals drops below 90 percent at long horizons, then the cluster-pooling mechanism rather than pooling per se is what fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that the dynamic response of a short time series to an identified shock can be estimated reliably by pooling it with similar series: under a sparse finite-mixture prior, each series' horizon-$h$ response $\rho_{i,h}$ is drawn from a Gaussian centered on the cluster mean of its data-determined group, and the posterior mean becomes a precision-weighted average of the series' own least-squares estimate and its cluster's mean, with the weight on the cross-section increasing automatically as the series' own sample shrinks. For horizons beyond the series' effective sample length, the response is imputed from the cluster's distribution without interpolation. The authors demonstrate the gain in a simulation calibrated to US macro data: relative to naive OLS local projections, the full model that pools response and control coefficients through the mixture reduces mean absolute error on short series ($T$ between 100 and 150 months) by roughly half at short, medium, and long horizons, and by 80 to 90 percent for very short series ($T$ between 25 and 60 months), while tracking the benchmark on long series. They also show that pooling the horizon-implied HAC variance correction at the cluster level restores the empirical coverage of nominal 90 percent intervals from roughly two-thirds to about 95 to 97 percent for short series.

Load-bearing premise

The headline gains rest on the representativeness of the hand-calibrated simulation design: if real unbalanced price panels have weaker shock propagation, less separated impulse-response profiles, or clusters that shift across horizons, the factor-of-two accuracy gain and the restored coverage may not transfer, and a second premise is that series in the same cluster have similar long-run variances, which justifies pooling the coverage correction.

Editorial extensions

If this is right

  • For short series (100 to 150 monthly observations), mean absolute error falls to roughly half the naive-LP level at short, medium, and long horizons; for very short series (25 to 60 observations), the reduction reaches 80 to 90 percent, at horizons where series-by-series estimation is mostly impossible.
  • Series with long samples pay a small price, with mean absolute error rising by about 5 to 7 percent relative to estimating them one by one, so pooling is nearly costless where the own-data signal is strong.
  • The mixture, not pooling alone, is what makes the gain safe: the single-common-response variant shows visible bias, especially at short horizons, whereas the data-determined clusters avoid shrinking across genuinely different dynamics.
  • The cluster-pooled variance correction raises the coverage of nominal 90 percent intervals for short series from about two-thirds at long horizons to 95 to 97 percent, making the intervals usable for inference.
  • In the application, the clusters sort 43 price and survey series into six or seven groups that separate consumer prices, final- and intermediate-demand producer prices, and survey measures, and the estimated responses show headline prices moving more than core prices, goods more than services, and the least-processed goods most.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • An implicit consequence the authors do not stress: the same machinery transfers to any unbalanced panel built around a common identified shock, such as firm-level, regional, or cross-country responses, but the cluster structure must be re-learned for each application, so the gain is conditional on the panel actually containing groups of similar series.
  • The paper's distinction between the precision of the cluster-mean response and the precision of individual member responses suggests a practical diagnostic: users should treat the cluster mean as the reliable object, and wide member dispersion as a warning that the pooled response for a given series is not trustworthy.
  • A testable extension would combine the cross-sectional pooling with across-horizon smoothing of the impulse response, which the authors flag as future work; if the two mechanisms compound, ultra-short panels could reach horizons beyond one sample length with still tighter intervals.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper proposes a hierarchical Bayesian estimator for local projections in panels of time series that may be short, very short, and unbalanced. The response coefficients (and optionally the control coefficients) are pooled across units through a sparse finite mixture; cluster membership is common across horizons, cluster means are shrunk toward a population center, and the number of active clusters is learned through a Dirichlet concentration parameter. An ex post Müller-style correction pools HAC variances at the cluster level. Simulation results based on a FRED-MD-calibrated DGP report MAE reductions of roughly 40 to 55 percent for short series relative to unit-by-unit OLS LP, with comparable performance on long series; the application to 43 US price series estimates clusters of responses to supply-chain and oil shocks. The main contribution is a practical framework for impulse-response estimation in short and ultra-short panels, with a transparent simulation design.

Significance. If the results hold, the paper fills a genuine gap: standard LPs are data-hungry, and the proposed framework allows short series to borrow information from longer related series, including at horizons beyond the short series' own sample, without interpolation and with propagated uncertainty. The simulation study is unusually transparent about the DGP adjustments and about the benchmark's limitations, and the comparison across nested estimators, loss functions, and coverage measures is informative. However, the headline quantitative gains are demonstrated only for a DGP whose four-cluster structure is known to the prior-calibration procedure, and the very-short benchmark comparison is not apples-to-apples. With additional robustness analysis, the contribution would be solid; in current form the central quantitative claims are conditional on untested choices.

major comments (3)
  1. [Sections 3.1, 3.2, and 4.1] The prior scales b0 and bB in Section 3.2 are calibrated from a preliminary four-group K-means partition, and the DGP in Section 3.1 has exactly four clusters with strongly separated factor loadings (sigma_lambda = 0.05, cross-loadings N(0, 0.05^2)). The simulation's prior calibration is therefore handed the true number of clusters before the data are used, and Section 4.1 applies the same four-group calibration in the empirical application. No sensitivity to K (for example, K = 2, 6, or 8) or to the resulting prior scales is reported. This is load-bearing: the central claim that the baseline specification improves MAEs by roughly half in short samples could reflect favorable prior calibration rather than the mixture's own cluster learning. Please add a sensitivity analysis varying K in the calibration and report whether the Table 1 ratios and Table 2 coverage rates are stable, and also report the effective number of clusters under alternative calibration choices.
  2. [Section 3.3, Table 1, and footnote 2] In the very-short block, the pooled estimators' MAEs are computed over all data-informed unit-horizon pairs, while the naive LP raw MAE in the denominator is computed only over the 19 percent of pairs where naive LP is computable. The relative gains of 0.24, 0.16, and 0.10 in the baseline row therefore compare populations of different difficulty, and the statement in Section 3.3 that pooling reduces MAE by about 80 to 90 percent for very short series is not identified from this table. Please report, for the very-short design, the MAE ratios on the common support where naive LP is computable, and separately report the absolute MAE of the pooled estimators over all data-informed pairs, with a clear statement that the latter has no naive-LP benchmark.
  3. [Section 3.1] The DGP is calibrated with three explicit adjustments (first-order persistence increased by five percent, innovation scales halved, and shock loadings tripled) and a four-cluster design with sigma_lambda = 0.05. These choices make the shock signal large and the cluster structure favorable, so the simulation's gains may not transfer to panels with weaker shock propagation, less separated IRF profiles, or no true cluster structure. Please add robustness simulations that vary sigma_lambda (for example, 0.1 and 0.2), the shock-loading multiplier (for example, 1 and 2), and a DGP with no cluster separation, and report whether the ranking of estimators and the MAE ratios change. I recognize that the adjustments are disclosed; the request is for sensitivity evidence, since the abstract's general claim of substantial improvement extends beyond the single calibrated DGP.
minor comments (4)
  1. [Section 2.2, Eq. (2)] The display for the posterior mean uses rho_i,h on the left and tau^2_{i,h} for the posterior variance, while tau^2_h denotes the prior variance; please clarify the distinction between prior and posterior variances in the notation or in the surrounding text.
  2. [Section 3.2] The prior scales are said to be recalibrated in every replication from that replication's own long-series OLS estimates; please clarify whether this is a fully empirical-Bayes procedure and discuss any potential double use of the evaluation data when the long series are also part of the performance assessment.
  3. [Section 4.2, Figure 7] The caption states that the hierarchy 'does not shrink' the within-cluster dispersion of individual member responses, but the posterior allocation and cluster means affect how individual draws are centered; please rephrase to distinguish the prior on individual deviations from the posterior quantities actually displayed.
  4. [General] The paper does not include a data or code availability statement; providing replication code and instructions for obtaining the shock series would strengthen reproducibility, especially because the Gibbs sampler details in Appendix B are central to the application.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central simulation gains are benchmarked against an external OLS-LP estimator and the true impulse responses, not against the model's own fitted values.

full rationale

The paper's central claim is evaluated in a simulated DGP calibrated to FRED-MD data, where mean absolute error is computed against the true impulse response coefficients ρ_{i,h}. This target is external to the estimator, so the headline improvements (e.g., relative MAEs of 0.55, 0.51, and 0.47 for short series under the joint ρ, β SFM pool in Table 1) are not forced by construction. The prior scales in Section 3.2 are calibrated from a preliminary four-group K-means partition, and the application in Section 4.1 uses the same style of data-calibrated prior computed on long series; this is a mildly data-dependent empirical-Bayes feature, but it does not define the quantity being evaluated or guarantee the outcome, because the benchmark naive LP receives no such prior and the comparison is to an independent estimator on the same data. The posterior mean formula in Section 2.2 is a standard precision-weighted average used to motivate borrowing strength across series, not a step that assumes the simulation conclusion. Self-citations (Clark 1999; Huber, Krisztin, and Pfarrhofer 2023; Huber, Matthes, and Pfarrhofer 2024) are background literature and are not load-bearing for the paper's central results. The alignment between the calibrated K=4 K-means prior and the DGP's true four-cluster structure is a robustness or external-validity concern, not a circularity: the paper does not derive its accuracy gains from the calibration, and no equation reduces to its own inputs.

Assumptions & free parameters 13 free parameters · 6 assumptions · 0 invented entities

The central simulation claims depend on several hand-set DGP calibration choices, most notably the tripled shock loadings, halved innovations, and strongly separated four-cluster design, which make the problem favorable to pooling and clustering. The model itself introduces no physical entities; its latent clusters and hierarchical priors are statistical constructs. The most consequential modeling axioms are the common-horizon cluster allocation, the assumption of similar long-run variances within clusters, and the data-calibrated priors.

free parameters (13)
  • DGP shock loading multiplier = 3
    Section 3.1 scales calibrated shock loadings by a factor of three so the monetary policy shock has important quantitative effects; this directly raises the signal the method can recover.
  • DGP innovation scale halving = 0.5
    Section 3.1 halves fitted factor innovation scales so the shock-driven component is not corrupted by idiosyncratic factor noise; this increases signal-to-noise.
  • DGP persistence inflation = 5 percent increase in a_g1
    Section 3.1 raises calibrated first-order persistence for real activity and employment to match empirical persistence; this changes how hard horizons are to estimate.
  • DGP factor loading dispersion = sigma_lambda = 0.05
    Section 3.1 sets small dispersion of loadings within clusters, making clusters tight and pooling favorable.
  • DGP idiosyncratic noise scale = 0.05 * median(sigma_g)
    Section 3.1 sets small idiosyncratic noise, keeping clusters separable.
  • DGP cluster design = 4 clusters, 20 series each
    Section 3.1 generates 80 series in four well-separated clusters, a structure favorable to the sparse finite mixture.
  • Very-short sample length range = U{25,...,60}
    Section 3.1 defines the ultra-short regime; at these lengths naive LP is computable on only 19% of data-informed pairs, complicating the benchmark comparison.
  • Mixture upper bound S = 8
    Section 2.5 fixes S=8 as a generous upper bound; if the true number of clusters exceeds 8, the model underfits. The application finds 6-7 active clusters, close to the bound.
  • Dirichlet concentration prior = Gamma(1,200), prior mean 0.005
    Section 3.2 and 4.1 set the prior that controls how aggressively redundant mixture components are emptied out.
  • Prior shapes a0 and aB = 2.5
    Sections 3.2 and 4.1 set weakly informative inverse-gamma shapes for within- and between-cluster variances.
  • Population center scale c = 100
    Section 4.1 sets m_h ~ N(0, c*s_y^2) with c=100, a weakly informative center prior.
  • Data-calibrated prior scales b0 and bB = b0=0.083, bB=0.356 (supply-chain); b0=0.063, bB=0.431 (oil)
    Section 4.1 computes prior mean scales from the dispersion of long-series OLS estimates under a preliminary four-group K-means partition; this uses the same data that are later pooled.
  • Lag length and horizon = p=12, H=24 in simulation; p=4, H=35 in application
    Section 3.1 and 4.1 choose lag length and maximum horizon; these determine effective sample sizes T_i,h and which horizons are data-informed or imputed.
assumptions (6)
  • domain assumption The identified structural shock w_t is observed and common across all M series.
    Section 2.1 defines w_t shared across series; if the shock is measured with error or not common, the LP responses are not identified.
  • ad hoc to paper Series within a latent cluster share similar IRF profiles across all horizons, and cluster membership z_i is common across horizons.
    Section 2.3 assumes one allocation z_i per series over all horizons; if similarity is horizon-specific, pooling toward a cluster mean could bias some horizons.
  • ad hoc to paper Units in the same cluster have similar long-run variances of LP scores, so the HAC correction can be pooled at the cluster level.
    Section 2.6 asserts this to justify the cluster-pooled Muller correction; the authors note it holds when the DGP is VAR-like and units respond alike.
  • ad hoc to paper The data-calibrated prior scales computed from long-series OLS estimates are appropriate for the short series in the panel.
    Sections 3.2 and 4.1 use long-unit dispersion to set prior scales; if short-series heterogeneity differs, the prior may be mis-scaled.
  • domain assumption The Gibbs sampler converges to the target posterior and the label-switching relabeling recovers a meaningful partition.
    Appendix B sketches the sampler but reports no convergence diagnostics; standard MCMC assumptions are maintained.
  • domain assumption The LP projection errors have the MA(h-1) structure and the Muller (2013) sandwich correction restores frequentist coverage.
    Section 2.6 relies on Muller (2013) and the LP literature for the validity of the ex-post correction.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Clustered Local Projections for Short and Ultra-Short Time Series -- A Hierarchical Bayesian Framework." pith.science (2026). https://pith.science/paper/23DOS6ZK

@misc{pith2026260804631,
  author       = {Pith},
  title        = {Pith review of: Clustered Local Projections for Short and Ultra-Short Time Series -- A Hierarchical Bayesian Framework},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/23DOS6ZK}},
  note         = {Machine review of arXiv:2608.04631}
}
read the original abstract

Estimating the dynamic effects of economic shocks in short and very short samples is impeded by a lack of degrees of freedom. We offer a solution based on a Bayesian hierarchical framework for estimating local projection (LP) impulse response functions across a panel of related time series. The framework explicitly accommodates unbalanced panels in which some series are substantially shorter than others, allowing the short series to borrow information from longer ones at horizons where the short series carry little or no own data. Since series might exhibit heterogeneous dynamics, we develop a sparse finite mixture pool that clusters units by similarity of their impulse response profiles. We show in simulations that our approach substantially improves LP estimation accuracy relative to the standard approach if the time series are short while producing similar LPs for longer time series. Using a US price dataset, augmented with survey responses, we find that supply-chain and oil shocks trigger heterogeneous reactions of different price measures, with headline price indices responding more sharply than their core counterparts and goods prices changing more than services prices.

Figures

Figures reproduced from arXiv: 2608.04631 by the authors.

Figure 1
Figure 1. Bias and posterior standard deviation by horizon, FRED-MD calibration. Notes: Rows, top to bottom: very short (T ∈ [25, 60]), short (T ∈ [100, 150]), and long (T = 500) series; y-axis free per panel. Left column: across-unit median of |Er[ˆρ (r) i,h] − ρi,h|. Right column: across-unit median of the posterior standard deviation of ρi,h, taking the median across replications per unit-horizon pair; for naive LP, its Ne… view at source ↗
Figure 2
Figure 2. Loss-minimizing estimator at each (h, ω) pair, FRED-MD calibration. Notes: Panels, left to right: very short, short, and long series. Color and per-cell symbol: the method minimizing the across-unit average loss Lm i,h(ω) at each (h, ω) pair (the symbol identifies the method in grayscale). Each method’s loss is averaged over its data-informed unit-horizon pairs; naive LP competes where it is computable (footnote 2).… view at source ↗
Figure 3
Figure 3. Pairwise loss comparison, FRED-MD calibration. Notes: Each entry: the fraction of units at which method (1) achieves a lower Lm i,h(ω) than method (2), on unit-horizon pairs where both are computable. Rows, top to bottom: very short, short, and long series; columns: the three pairwise comparisons. Darker blue = method (1) wins more often; blank regions in the very-short row are horizons at which naive LP is computab… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Empirical coverage of nominal 0.90 interval by horizon, FRED-MD calibration. Notes: Panels, left to right: very short (T ∈ [25, 60]), short (T ∈ [100, 150]), and long (T = 500) series. Dashed grey line: nominal level. Each method is shown with its headline inference re…
Figure 5
Figure 5. Figure 5: Cluster IRFs in response to the K¨anzig supply-chain shock, pre-COVID sample (1971:01–2019:12). Notes: One panel per active mixture cluster (sorted by size). Cluster posterior median (dark line) with 68%, 80%, and 90% pointwise M¨uller-corrected bands (shaded), and pos…
Figure 6
Figure 6. Figure 6: Cluster IRFs in response to the Baumeister–Hamilton oil supply shock, pre-COVID sample (1975:02–2019:12). Notes: Layout as in [PITH_FULL_IMAGE:figures/full_fig_p032_6.png]
Figure 7
Figure 7. Figure 7: Precision of the cluster-mean IRF against the precision of the individual member responses. Notes: The vertical axis is on a log scale. The vertical distance from the diamond to the points is the within-cluster dispersion of individual responses around the common locat…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

199 extracted references · 72 canonical work pages

  1. [1]

    Journal of Business & Economic Statistics , Volume =

    Frühwirth-Schnatter, Sylvia and Kaufmann, Sylvia , Title =. Journal of Business & Economic Statistics , Volume =. 2008 , Pages =

  2. [2]

    Journal of the American Statistical Association , volume=

    Fr. Journal of the American Statistical Association , volume=. 2001 , publisher=

  3. [3]

    Fast and order-invariant inference in

    Huber, Florian and Koop, Gary , journal=. Fast and order-invariant inference in. 2024 , publisher=

  4. [4]

    Econometrica , volume=

    Deep neural networks for estimation and inference , author=. Econometrica , volume=. 2021 , publisher=

  5. [5]

    2006 , publisher=

    Williams, Christopher KI and Rasmussen, Carl Edward , volume=. 2006 , publisher=

  6. [6]

    The Journal of Machine Learning Research , volume=

    Bayesian nonparametric covariance regression , author=. The Journal of Machine Learning Research , volume=. 2015 , publisher=

  7. [7]

    Hierarchical shrinkage

    Tang, Tao and Mak, Simon and Dunson, David , journal=. Hierarchical shrinkage. 2024 , publisher=

  8. [8]

    Oil, Inflation Expectations, and Household Characteristics: A Nonlinear Heterogeneous Agent

    Baumeister, Christiane and Frank, Pascal and Huber, Florian and Koop, Gary , year =. Oil, Inflation Expectations, and Household Characteristics: A Nonlinear Heterogeneous Agent

Show all 199 references
  1. [9]

    Drifts and volatilities: monetary policies and outcomes in the post

    Cogley, Timothy and Sargent, Thomas J , journal=. Drifts and volatilities: monetary policies and outcomes in the post. 2005 , doi=

  2. [10]

    Were there regime switches in

    Sims, Christopher A and Zha, Tao , journal=. Were there regime switches in. 2006 , doi =

  3. [11]

    Nowcasting in a pandemic using non-parametric mixed frequency

    Florian Huber and Gary Koop and Luca Onorante and Michael Pfarrhofer and Josef Schreiner , keywords =. Nowcasting in a pandemic using non-parametric mixed frequency. Journal of Econometrics , volume =. 2023 , issn =. doi:10.1016/j.jeconom.2020.11.006 , url =

  4. [12]

    Linero , year=

    Antonio R. Linero , year=. Defensive Model Expansion for Robust. 2510.09598 , archivePrefix=

  5. [13]

    Tony Chernis and Niko Hauzenberger and Haroon Mumtaz and Michael Pfarrhofer , year=. A. 2509.04928 , archivePrefix=

  6. [14]

    and Boivin, Jean and Eliasz, Piotr , title =

    Bernanke, Ben S. and Boivin, Jean and Eliasz, Piotr , title =. The Quarterly Journal of Economics , volume =. 2005 , month =. doi:10.1162/0033553053327452 , url =

  7. [15]

    Journal of Computational and Graphical Statistics , volume =

    Gregor Kastner and Sylvia Frühwirth-Schnatter and Hedibert Freitas Lopes , title =. Journal of Computational and Graphical Statistics , volume =. 2017 , publisher =

  8. [16]

    Messner and Georg J

    Manuel Gebetsberger and Jakob W. Messner and Georg J. Mayr and Achim Zeileis. Estimation Methods for Nonhomogeneous Regression Models: Minimum Continuous Ranked Probability Score versus Maximum Likelihood. Monthly Weather Review. 2018. doi:10.1175/MWR-D-17-0364.1

  9. [17]

    Annual Review of Statistics and Its Applications , year=

    Proper scoring rules for estimation and forecast evaluation , author=. Annual Review of Statistics and Its Applications , year=

  10. [18]

    Econometrica , volume=

    Sign restrictions, structural vector autoregressions, and useful prior information , author=. Econometrica , volume=. 2015 , publisher=

  11. [19]

    Journal of Business & Economic Statistics , volume =

    Dynamic causal effects in a nonlinear world: the good, the bad, and the ugly , author=. Journal of Business & Economic Statistics , volume =. 2025 , publisher =

  12. [20]

    Machine learning the macroeconomic effects of financial shocks , journal =

    Niko Hauzenberger and Florian Huber and Karin Klieber and Massimiliano Marcellino , keywords =. Machine learning the macroeconomic effects of financial shocks , journal =. 2025 , issn =. doi:10.1016/j.econlet.2025.112260 , url =

  13. [21]

    Federal Reserve Bank of Dallas Working Paper , volume=

    Nonparametric Local Projections , author=. Federal Reserve Bank of Dallas Working Paper , volume=. 2024 , doi=

  14. [22]

    Artificial Intelligence and Statistics , pages=

    Handling sparsity via the horseshoe , author=. Artificial Intelligence and Statistics , pages=. 2009 , organization=

  15. [23]

    Adaptive shrinkage in

    Huber, Florian and Feldkircher, Martin , journal=. Adaptive shrinkage in. 2019 , publisher=

  16. [24]

    Kastner, Gregor , journal=. Sparse. 2019 , publisher=

  17. [25]

    and Ng, Serena , journal=

    McCracken, Michael W. and Ng, Serena , journal=. 2016 , publisher=

  18. [26]

    McCracken and Serena Ng , title =

    Michael W. McCracken and Serena Ng , title =. Federal Reserve Bank of St. Louis Review , year =. doi:10.20955/r.103.1-44 , url =

  19. [27]

    Review of Economics and Statistics , volume=

    Are the effects of financial market disruptions big or small? , author=. Review of Economics and Statistics , volume=. 2022 , publisher=

  20. [28]

    2018 , doi =

    Functional Approximation of Impulse Responses , journal =. 2018 , doi =

  21. [29]

    Angrist and Oscar Jorda and Guido M

    Joshua D. Angrist and Oscar Jorda and Guido M. Kuersteiner , title =. Journal of Business & Economic Statistics , volume =. 2018 , doi =

  22. [30]

    Journal of Financial Economics , volume=

    Characteristics are covariances: A unified model of risk and return , author=. Journal of Financial Economics , volume=. 2019 , publisher=

  23. [31]

    and Marcellino, Massimiliano , title=

    Carriero, Andrea and Chan, Joshua and Clark, Todd E. and Marcellino, Massimiliano , title=. Journal of Econometrics , year=2022, volume=

  24. [32]

    Bayesian additive regression trees using

    Hern. Bayesian additive regression trees using. Statistics and Computing , volume=. 2018 , publisher=

  25. [33]

    Journal of Business & Economic Statistics , volume=

    Identification of structural Vector Autoregressions by stochastic volatility , author=. Journal of Business & Economic Statistics , volume=

  26. [34]

    Forecasting using variational

    Gefang, Deborah and Koop, Gary and Poon, Aubrey , journal=. Forecasting using variational. 2022 , doi=

  27. [35]

    On theory for

    Ro. On theory for. The 22nd International Conference on Artificial Intelligence and Statistics , pages=. 2019 , organization=

  28. [36]

    Minnesota-type adaptive hierarchical priors for large

    Chan, Joshua CC , journal=. Minnesota-type adaptive hierarchical priors for large. 2021 , publisher=

  29. [37]

    Bayesian

    Carriero, Andrea and Clark, Todd E and Marcellino, Massimiliano , journal=. Bayesian. 2015 , publisher=

  30. [38]

    Journal of Business & Economic Statistics , volume=

    Generalized shrinkage methods for forecasting using many predictors , author=. Journal of Business & Economic Statistics , volume=. 2012 , publisher=

  31. [39]

    2006 , publisher=

    All of Nonparametric Statistics , author=. 2006 , publisher=

  32. [40]

    Gibbs sampling methods for

    Kozumi, Hideo and Kobayashi, Genya , journal=. Gibbs sampling methods for. 2011 , publisher=

  33. [41]

    Journal of Econometrics , volume=

    Adaptive hierarchical priors for high-dimensional vector autoregressions , author=. Journal of Econometrics , volume=. 2019 , publisher=

  34. [42]

    Quadratic

    Sentana, Enrique , journal=. Quadratic. 1995 , publisher=

  35. [43]

    Journal of Business & Economic Statistics , volume=

    Dynamic conditional correlation: A simple class of multivariate generalized autoregressive conditional heteroskedasticity models , author=. Journal of Business & Economic Statistics , volume=. 2002 , publisher=

  36. [44]

    Modelling the coherence in short-run nominal exchange rates: a multivariate generalized

    Bollerslev, Tim , journal=. Modelling the coherence in short-run nominal exchange rates: a multivariate generalized. 1990 , publisher=

  37. [45]

    Journal of Econometrics , volume=

    Stochastic volatility with leverage: Fast and efficient likelihood inference , author=. Journal of Econometrics , volume=. 2007 , doi=

  38. [46]

    Journal of Business & Economic Statistics , volume=

    Bayesian dynamic factor models and portfolio allocation , author=. Journal of Business & Economic Statistics , volume=. 2000 , publisher=

  39. [47]

    A new algorithm for structural restrictions in

    Dimitris Korobilis , keywords =. A new algorithm for structural restrictions in. European Economic Review , volume =. 2022 , issn =. doi:10.1016/j.euroecorev.2022.104241 , url =

  40. [48]

    CEPR Discussion Paper Series , volume =

    Modeling and forecasting macroeconomic downside risk , author=. CEPR Discussion Paper Series , volume =

  41. [49]

    International Economic Review , year =

    Adrian, Tobias and Boyarchenko, Nina and Giannone, Domenico , title =. International Economic Review , year =

  42. [50]

    Journal of Econometrics , volume=

    Bayesian nonparametric vector autoregressive models , author=. Journal of Econometrics , volume=. 2018 , doi =

  43. [51]

    Journal of the American Statistical Association , volume=

    Functional horseshoe priors for subspace shrinkage , author=. Journal of the American Statistical Association , volume=. 2020 , doi =

  44. [52]

    A unifying view of sparse approximate

    Quinonero-Candela, Joaquin and Rasmussen, Carl Edward , journal=. A unifying view of sparse approximate. 2005 , publisher=

  45. [53]

    IEEE Signal Processing Letters , volume=

    A simple sampler for the horseshoe estimator , author=. IEEE Signal Processing Letters , volume=. 2015 , doi =

  46. [54]

    Kastner, Gregor and Huber, Florian , journal=. Sparse. 2020 , publisher=

  47. [55]

    Journal of the American Statistical Association , volume=

    Bayesian approximate kernel regression with variable selection , author=. Journal of the American Statistical Association , volume=. 2018 , doi =

  48. [56]

    Ancillarity-sufficiency interweaving strategy (

    Kastner, Gregor and Fr. Ancillarity-sufficiency interweaving strategy (. Computational Statistics & Data Analysis , volume=. 2014 , doi=

  49. [57]

    Bayes inference in regression models with

    Chib, Siddhartha and Greenberg, Edward , journal=. Bayes inference in regression models with. 1994 , doi =

  50. [58]

    Journal of Business & Economic Statistics , volume=

    Combined density nowcasting in an uncertain economic environment , author=. Journal of Business & Economic Statistics , volume=. 2018 , doi =

  51. [59]

    The Review of Economic Studies , volume=

    Stochastic volatility: likelihood inference and comparison with ARCH models , author=. The Review of Economic Studies , volume=. 1998 , doi =

  52. [60]

    American Economic Review , volume=

    Vulnerable growth , author=. American Economic Review , volume=. 2019 , doi =

  53. [61]

    2022 , journal=

    The term structure of growth-at-risk , author=. 2022 , journal=

  54. [62]

    Machine Learning , volume=

    Random forests , author=. Machine Learning , volume=. 2001 , doi =

  55. [63]

    Statistics & Probability Letters , volume=

    Bayesian quantile regression , author=. Statistics & Probability Letters , volume=. 2001 , publisher=

  56. [64]

    Stochastic model specification search for

    Fr. Stochastic model specification search for. Journal of Econometrics , volume=. 2010 , publisher=

  57. [65]

    Journal of Business & Economic Statistics , volume=

    Inducing sparsity and shrinkage in time-varying parameter models , author=. Journal of Business & Economic Statistics , volume=. 2021 , publisher=

  58. [66]

    2021 , institution=

    Macroeconomic and financial risks: A tale of mean and volatility , author=. 2021 , institution=

  59. [67]

    Journal of Applied Econometrics , year=

    Nowcasting tail risk to economic activity at a weekly frequency , author=. Journal of Applied Econometrics , year=

  60. [68]

    Capturing macroeconomic tail risks with

    Carriero, Andrea and Clark,. Capturing macroeconomic tail risks with. Federal Reserve Bank of Cleveland Working Paper , volume=. 2020 , doi =

  61. [69]

    Federal Reserve Bank of Kansas City Research Working Paper , volume=

    Assessing macroeconomic tail risks in a data-rich environment , author=. Federal Reserve Bank of Kansas City Research Working Paper , volume=. 2019 , doi =

  62. [70]

    Journal of Applied Econometrics , volume=

    Forecasting tail risks , author=. Journal of Applied Econometrics , volume=. 2017 , doi =

  63. [71]

    International Journal of Forecasting , year =

    Laurent Ferrara and Matteo Mogliani and Jean-Guillaume Sahuc , title =. International Journal of Forecasting , year =

  64. [72]

    Journal of Money, Credit and Banking , volume=

    Constructing density forecasts from quantile regressions , author=. Journal of Money, Credit and Banking , volume=. 2012 , doi =

  65. [73]

    International Journal of Forecasting , volume=

    Asymmetry in unemployment rate forecast errors , author=. International Journal of Forecasting , volume=. 2019 , doi =

  66. [74]

    National Bank of Belgium Research Working Paper , volume=

    Quantile-based inflation risk models , author=. National Bank of Belgium Research Working Paper , volume=. 2018 , url=

  67. [75]

    Journal of Business & Economic Statistics , volume =

    Raffaella Giacomini and Ivana Komunjer , title =. Journal of Business & Economic Statistics , volume =. 2005 , publisher =

  68. [76]

    Journal of Financial Economics , volume=

    Systemic risk and the macroeconomy: An empirical evaluation , author=. Journal of Financial Economics , volume=. 2016 , doi =

  69. [77]

    2019 , author =

    Growth in stress , journal =. 2019 , author =

  70. [78]

    2020 , journal=

    The macroeconomy as a random forest , author=. 2020 , journal=

  71. [79]

    Clark and Florian Huber and Gary Koop and Massimiliano Marcellino and Michael Pfarrhofer , title =

    Todd E. Clark and Florian Huber and Gary Koop and Massimiliano Marcellino and Michael Pfarrhofer , title =. Journal of Business & Economic Statistics , volume =. 2024 , doi =

  72. [80]

    Tail forecasting with multivariate

    Clark, Todd E and Huber, Florian and Koop, Gary and Marcellino, Massimiliano and Pfarrhofer, Michael , journal=. Tail forecasting with multivariate. 2023 , doi=

  73. [81]

    2020 , journal=

    How is machine learning useful for macroeconomic forecasting? , author=. 2020 , journal=

  74. [82]

    Can machine learning catch the

    Goulet Coulombe, Philippe and Marcellino, Massimiliano and Stevanovic, Dalibor , year=. Can machine learning catch the. arXiv , volume=

  75. [83]

    , title =

    Kiley, Michael T. , title =. Journal of Money, Credit and Banking , volume =. 2022 , number =

  76. [84]

    International Journal of Forecasting , volume=

    Quantile regression forecasts of inflation under model uncertainty , author=. International Journal of Forecasting , volume=. 2017 , doi =

  77. [85]

    Journal of Business & Economic Statistics , volume=

    Forecasting the distribution of economic variables in a data-rich environment , author=. Journal of Business & Economic Statistics , volume=. 2015 , doi =

  78. [86]

    Are macroeconomic variables useful for forecasting the distribution of

    Manzan, Sebastiano and Zerom, Dawit , journal=. Are macroeconomic variables useful for forecasting the distribution of. 2013 , doi =

  79. [87]

    Asymmetric quantile persistence and predictability: the case of

    Manzan, Sebastiano and Zerom, Dawit , journal=. Asymmetric quantile persistence and predictability: the case of. 2015 , doi =

  80. [88]

    2021 , journal=

    Machine learning advances for time Series forecasting , author=. 2021 , journal=

  81. [89]

    Essays in Honor of M

    James Mitchell and Aubrey Poon and Gian Luigi Mazzi , title =. Essays in Honor of M. Hashem Pesaran: Prediction and Macro Modeling , publisher =

  82. [90]

    James Mitchell and Aubrey Poon and Dan Zhu , title=

  83. [91]

    Journal of Business & Economic Statistics , volume=

    Forecasting inflation in a data-rich environment: the benefits of machine learning methods , author=. Journal of Business & Economic Statistics , volume=. 2021 , doi =

  84. [92]

    Let the tree decide:

    Sofia Velasco , year=. Let the tree decide:. 2506.11551 , archivePrefix=

  85. [93]

    Brookings Papers on Economic Activity , year=

    When is growth at risk? , author=. Brookings Papers on Economic Activity , year=

  86. [94]

    Real-time nowcasting with a

    Carriero, Andrea and Clark,. Real-time nowcasting with a. Journal of the Royal Statistical Society: Series A , volume=. 2015 , doi =

  87. [95]

    2020 , journal=

    Financial variables as predictors of real growth vulnerability , author=. 2020 , journal=

  88. [96]

    and George, Edward I

    Chipman, Hugh A. and George, Edward I. and McCulloch, Robert E. , journal=. Bayesian. 1998 , doi =

  89. [97]

    Addressing

    Carriero, Andrea and Clark,. Addressing. The Review of Economics and Statistics , volume=. 2022 , doi=

  90. [98]

    and Chipman, Hugh A

    Pratola, Matthew T. and Chipman, Hugh A. and George, Edward I. and McCulloch, Robert E. , journal=. Heteroscedastic. 2020 , doi =

  91. [99]

    Common drifting volatility in large

    Carriero, Andrea and Clark,. Common drifting volatility in large. Journal of Business & Economic Statistics , volume=. 2016 , doi =

  92. [100]

    FEDS working paper , number=

    Financial and Macroeconomic Data Through the Lens of a Nonlinear Dynamic Factor Model , author=. FEDS working paper , number=. 2023 , doi =

  93. [101]

    Journal of Applied Econometrics , volume=

    Macroeconomic forecasting performance under alternative specifications of time-varying volatility , author=. Journal of Applied Econometrics , volume=. 2015 , doi =

  94. [102]

    and George, Edward I

    Chipman, Hugh A. and George, Edward I. and McCulloch, Robert E. , journal=. 2010 , doi=

  95. [103]

    Journal of Applied Econometrics , volume=

    Forecast comparisons in unstable environments , author=. Journal of Applied Econometrics , volume=. 2010 , publisher=

  96. [104]

    Raftery , title =

    Tilmann Gneiting and Adrian E. Raftery , title =. Journal of the American Statistical Association , volume =. 2007 , doi =

  97. [105]

    Journal of Business & Economic Statistics , volume=

    Comparing density forecasts using threshold- and quantile-weighted scoring rules , author=. Journal of Business & Economic Statistics , volume=. 2011 , doi =

  98. [106]

    Journal of Business & Economic Statistics , volume=

    Comparing predictive accuracy , author=. Journal of Business & Economic Statistics , volume=. 1995 , doi =

  99. [107]

    manuscript , year=

    Modeling and forecasting macroeconomic downside risk , author=. manuscript , year=

  100. [108]

    Inference in

    Huber, Florian and Rossini, Luca , journal=. Inference in. 2022 , doi =

  101. [109]

    Nowcasting in a pandemic using non-parametric mixed frequency

    Huber, Florian and Koop, Gary and Onorante, Luca and Pfarrhofer, Michael and Schreiner, Josef , journal=. Nowcasting in a pandemic using non-parametric mixed frequency. 2020 , doi =

  102. [110]

    and Reeves, G

    van den Boom, W. and Reeves, G. and Dunson, D. , journal=. Approximating posteriors with high-dimensional nuisance parameters via integrated rotated. 2020 , doi =

  103. [111]

    Advances in Econometrics , volume=

    Panel Vector Autoregressive models: A survey , author=. Advances in Econometrics , volume=

  104. [112]

    Journal of Business & Economic Statistics , volume=

    Modeling regional interdependencies using a global error-correcting macroeconometric model , author=. Journal of Business & Economic Statistics , volume=. 2004 , doi =

  105. [113]

    International Journal of Forecasting , volume=

    Forecasting economic and financial variables with global VARs , author=. International Journal of Forecasting , volume=. doi:10.1198/073500104000000019 , year=

  106. [114]

    Forecasting with global Vector Autoregressive models: A

    Crespo Cuaresma, Jesus and Feldkircher, Martin and Huber, Florian , journal=. Forecasting with global Vector Autoregressive models: A

  107. [115]

    Does joint modelling of the world economy pay off? Evaluating global forecasts from a

    Dovern, Jonas and Feldkircher, Martin and Huber, Florian , journal=. Does joint modelling of the world economy pay off? Evaluating global forecasts from a. 2016 , doi =

  108. [116]

    Density forecasting using

    Huber, Florian , journal=. Density forecasting using. 2016 , doi =

  109. [117]

    International Economic Review , volume=

    Estimating multicountry VAR models , author=. International Economic Review , volume=. 2009 , doi =

  110. [118]

    European Economic Review , volume=

    Model uncertainty in panel Vector Autoregressive model , author=. European Economic Review , volume=. 2016 , doi =

  111. [119]

    Biometrika , volume=

    The horseshoe estimator for sparse signals , author=. Biometrika , volume=. 2010 , doi=

  112. [120]

    and Casella, G

    Park, T. and Casella, G. , journal=. The

  113. [121]

    Journal of the American Statistical Association , volume=

    Dirichlet-Laplace priors for optimal shrinkage , author=. Journal of the American Statistical Association , volume=. 2015 , doi =

  114. [122]

    Real-time forecasting with a mixed-frequency

    Schorfheide, Frank and Song, Dongho , journal=. Real-time forecasting with a mixed-frequency. 2015 , doi =

  115. [123]

    Journal of statistical software , pages=

    Mixed frequency data sampling regression models: the R package midasr , author=. Journal of statistical software , pages=. 2016 , publisher=

  116. [124]

    and Galv

    Clements, Michael P. and Galv. Forecasting. Journal of Applied Econometrics , volume=. 2009 , publisher=

  117. [125]

    Mixed-frequency

    Ankargren, Sebastian and Yang, Yukai , year=. Mixed-frequency

  118. [126]

    A look into the factor model black box: Publication lags and the role of hard and soft data in forecasting GDP

    Marta Ba \'n bura and Gerhard R \"u nstler. A look into the factor model black box: Publication lags and the role of hard and soft data in forecasting GDP. International Journal of Forecasting. 2011

  119. [127]

    Short-term forecasts of euro area real

    Diron, Marie , journal=. Short-term forecasts of euro area real. 2008 , publisher=

  120. [128]

    Real-time density forecasts from

    Clark,. Real-time density forecasts from. Journal of Business & Economic Statistics , volume=. 2011 , doi =

  121. [129]

    Review of Financial Studies , year=

    Forecasting through the Rearview Mirror: Data Revisions and Bond Return Predictability , author=. Review of Financial Studies , year=

  122. [130]

    Real-time forecasting of German GDP based on a large factor model with monthly and quarterly data

    Christian Schumacher and J. Real-time forecasting of German GDP based on a large factor model with monthly and quarterly data. International Journal of Forecasting. 2008

  123. [131]

    Comparing and evaluating

    Geweke, John and Amisano, Gianni , journal=. Comparing and evaluating. 2010 , publisher=

  124. [132]

    Journal of Business & Economic Statistics , volume=

    Macroeconomic forecasting using diffusion indexes , author=. Journal of Business & Economic Statistics , volume=. 2002 , publisher=

  125. [133]

    Ba. Large. Journal of Applied Econometrics , volume=. 2010 , doi=

  126. [134]

    Review of Economics and Statistics , volume=

    The generalized dynamic-factor model: Identification and estimation , author=. Review of Economics and Statistics , volume=. 2000 , publisher=

  127. [135]

    Journal of Monetary Economics , volume=

    Nowcasting: The real-time informational content of macroeconomic data , author=. Journal of Monetary Economics , volume=. 2008 , publisher=

  128. [136]

    Journal of Applied Econometrics , volume =

    D'Agostino, Antonello and Gambetti, Luca and Giannone, Domenico , title =. Journal of Applied Econometrics , volume =. doi:10.1002/jae.1257 , url =. https://onlinelibrary.wiley.com/doi/pdf/10.1002/jae.1257 , year =

  129. [137]

    Handbook of Economic Forecasting , volume=

    Now-casting and the real-time data flow , author=. Handbook of Economic Forecasting , volume=. 2013 , publisher=

  130. [138]

    2018 , publisher=

    Applied economic forecasting using time series methods , author=. 2018 , publisher=

  131. [139]

    Journal of Money, Credit and Banking , volume=

    Data revisions are not well behaved , author=. Journal of Money, Credit and Banking , volume=. 2008 , publisher=

  132. [140]

    Review of Economics and Statistics , pages=

    The use of preliminary data in econometric forecasting , author=. Review of Economics and Statistics , pages=. 1978 , publisher=

  133. [141]

    Data revisions to German national accounts: Are initial releases good nowcasts?

    Till Strohsal and Elias Wolf. Data revisions to German national accounts: Are initial releases good nowcasts?. International Journal of Forecasting. in press. doi:10.1016/j.ijforecast.2019.12.006

  134. [142]

    Journal of Monetary Economics , volume=

    Monetary policy in a data-rich environment , author=. Journal of Monetary Economics , volume=. 2003 , publisher=

  135. [143]

    Journal of Economic Literature , volume=

    Frontiers of real-time data analysis , author=. Journal of Economic Literature , volume=

  136. [144]

    Journal of the American Statistical Association , volume=

    Forecasting output with the composite leading index: A real-time analysis , author=. Journal of the American Statistical Association , volume=. 1991 , publisher=

  137. [145]

    Handbook of Economic Forecasting , volume=

    Forecasting with real-time macroeconomic data , author=. Handbook of Economic Forecasting , volume=. 2006 , publisher=

  138. [146]

    Journal of Econometrics , volume=

    A real-time data set for macroeconomists , author=. Journal of Econometrics , volume=. 2001 , publisher=

  139. [147]

    McCracken and Serena Ng , title =

    Michael W. McCracken and Serena Ng , title =. Journal of Business & Economic Statistics , volume =. 2016 , publisher =

  140. [148]

    Journal of applied econometrics , volume=

    Multivariate GARCH models: a survey , author=. Journal of applied econometrics , volume=. 2006 , publisher=

  141. [149]

    Forni, Mario and Gambetti, Luca and Granese, Antonio and Sala, Luca and Soccorsi, Stefano , journal=. An. 2025 , Month =

  142. [150]

    Journal of Business & Economic Statistics , volume=

    Determining the number of primitive shocks in factor models , author=. Journal of Business & Economic Statistics , volume=. 2007 , doi=

  143. [151]

    and Eisenstat, Eric and Strachan, Rodney W

    Chan, Joshua C.C. and Eisenstat, Eric and Strachan, Rodney W. , journal=. 2020 , doi=

  144. [152]

    , title =

    Baumeister, Christiane and Hamilton, James D. , title =. Econometrica , volume =. doi:10.3982/ECTA12356 , url =

  145. [153]

    The Review of Economic Studies , volume=

    Time varying structural vector autoregressions and monetary policy , author=. The Review of Economic Studies , volume=. 2005 , doi=

  146. [154]

    Journal of Business & Economic Statistics , volume=

    Central bank macroeconomic forecasting during the global financial crisis: The European Central Bank and Federal Reserve Bank of New York experiences , author=. Journal of Business & Economic Statistics , volume=. 2014 , publisher=

  147. [155]

    Primiceri , title=

    Domenico Giannone and Michele Lenza and Giorgio E. Primiceri , title=. The Review of Economics and Statistics , year=2015, volume=. doi:10.1162/rest_a_00483 , abstract=

  148. [156]

    Journal of Applied Econometrics , year=2010, volume=

    Marta Banbura and Domenico Giannone and Lucrezia Reichlin , title=. Journal of Applied Econometrics , year=2010, volume=. doi:10.1002/jae.1137 , url=

  149. [157]

    Journal of Applied Econometrics , year=2013, volume=

    Gary Koop , title=. Journal of Applied Econometrics , year=2013, volume=. doi:10.1002/jae.1270 , url=

  150. [158]

    Carriero, Andrea and Clark,. Large. Journal of Econometrics , year=2019, volume=212, number=1, pages=. doi:10.1016/j.jeconom.2019.04.024 , url=

  151. [159]

    , title=

    West, Kenneth D. , title=. Econometrica , year=1996, volume=

  152. [160]

    ECB Occasional Paper Series , year=2017, volume=

    Rodolfo Arioli and Colm Bates and Heinz Dieden and Ioana Duca and Roberta Friz and Christian Gayer and Geoff Kenny and Aidan Meyler and Iskra Pavlova , title=. ECB Occasional Paper Series , year=2017, volume=

  153. [161]

    EC , month = jan, year =. The

  154. [162]

    The Econometrics of Panel Data

    Marno Verbeek , title=. The Econometrics of Panel Data. Advanced Studies in Theoretical and Applied Econometrics , year=2008, volume=

  155. [163]

    and Eusepi, Stefano and Giannone, Domenico and Qian, Eric and Sbordone, Argia , title =

    Crump, Richard K. and Eusepi, Stefano and Giannone, Domenico and Qian, Eric and Sbordone, Argia , title =. 2025 , journal=

  156. [164]

    Stock and Mark W

    James H. Stock and Mark W. Watson , title =. 2025 , journal=

  157. [165]

    Credit Spreads and Business Cycle Fluctuations , Journal =

    Simon Gilchrist and Egon Zakraj. Credit Spreads and Business Cycle Fluctuations , Journal =. 2012 , Month =

  158. [166]

    Lewis and Egon Zakraj

    Giovanni Favara and Simon Gilchrist and Kurt F. Lewis and Egon Zakraj. Recession Risk and the Excess Bond Premium , year =

  159. [167]

    Hashem and Potter, Simon M

    Koop, Gary and Pesaran, M. Hashem and Potter, Simon M. , title=. Journal of Econometrics , year=1996, volume=74, number=1, pages=

  160. [168]

    Chan, Joshua C. C. and Matthes, Christian and Yu, Xuewen , note =. Large Structural

  161. [169]

    American Economic Journal: Macroeconomics , Volume =

    Kim, Hee Soo and Matthes, Christian and Phan, Toàn , Title =. American Economic Journal: Macroeconomics , Volume =. 2025 , Month =

  162. [170]

    Journal of Business & Economic Statistics , volume =

    Marcelle Chauvet and Jeremy Piger , title =. Journal of Business & Economic Statistics , volume =. 2008 , doi =

  163. [171]

    M\"uller and Mark W

    Ulrich K. M\"uller and Mark W. Watson , title =. Journal of Applied Econometrics , volume =. 2025 , doi=

  164. [172]

    Journal of Business & Economic Statistics , volume =

    Niko Hauzenberger and Florian Huber and Massimiliano Marcellino and Nico Petz , title =. Journal of Business & Economic Statistics , volume =. 2025 , doi =

  165. [173]

    2014 , type =

    Fernald, John , title =. 2014 , type =

  166. [174]

    2026 , journal =

    Marco Schwarzbach , title =. 2026 , journal =

  167. [175]

    Clark and Florian Huber and Gary Koop , journal=

    Todd E. Clark and Florian Huber and Gary Koop , journal=. A Nonparametric Approach to Augmenting a. 2026 , doi =

  168. [176]

    and Miranda-Agrippino, Silvia and Ricco, Giovanni , title =

    Ferreira, Leonardo N. and Miranda-Agrippino, Silvia and Ricco, Giovanni , title =. Review of Economics and Statistics , year =

  169. [177]

    Huber, Florian and Matthes, Christian and Pfarrhofer, Michael , title =

  170. [178]

    Responses to monetary policy shocks in the east and the west of

    Jaroci\'. Responses to monetary policy shocks in the east and the west of. Journal of Applied Econometrics , year =

  171. [179]

    Estimation and inference of impulse responses by local projections , journal =

    Jord\`. Estimation and inference of impulse responses by local projections , journal =. 2005 , volume =

  172. [180]

    Local projections vs

    Li, Dake and Plagborg-M. Local projections vs. Journal of Econometrics , year =

  173. [181]

    Model-based clustering based on sparse finite

    Malsiner-Walli, Gertraud and Fr\". Model-based clustering based on sparse finite. Statistics and Computing , year =

  174. [182]

    Journal of the Royal Statistical Society: Series B (Statistical Methodology) , year =

    Rousseau, Judith and Mengersen, Kerrie , title =. Journal of the Royal Statistical Society: Series B (Statistical Methodology) , year =

  175. [183]

    Mumtaz, Haroon and Piffer, Michele , title =

  176. [184]

    M\". Risk of. Econometrica , year =

  177. [185]

    Computational Economics , year =

    Tanaka, Masahiro , title =. Computational Economics , year =

  178. [186]

    Discussion Papers , year =

    Fortier, Eric , title =. Discussion Papers , year =

  179. [187]

    Review of Economic Studies , year =

    Ke-Li Xu , title =. Review of Economic Studies , year =

  180. [188]

    Florian Huber and Tam. A. The Annals of Applied Statistics , number =. 2023 , doi =

  181. [189]

    and Park, Taeyoung , title =

    van Dyk, David A. and Park, Taeyoung , title =. Journal of the American Statistical Association , year =

  182. [190]

    , title =

    Liu, Jun S. , title =. Journal of the American Statistical Association , year =

  183. [191]

    Journal of Applied Econometrics , volume =

    Piger, Jeremy and Stockwell, Thomas , title =. Journal of Applied Econometrics , volume =. doi:10.1002/jae.70003 , year =

  184. [192]

    NBER Macroeconomics Annual , volume =

    Local Projections or Vector Autoregressions?. NBER Macroeconomics Annual , volume =. 2026 , doi =

  185. [193]

    2024 , doi =

    Bias in Local Projections , journal =. 2024 , doi =

  186. [194]

    , title =

    Clark, Todd E. , title =. The Review of Economics and Statistics , volume =. 1999 , month =

  187. [195]

    , Title =

    Känzig, Diego R. , Title =. American Economic Review , Volume =. 2021 , Month =

  188. [196]

    and Raghavan, Ramya , Title =

    Känzig, Diego R. and Raghavan, Ramya , Title =. 2026 , Month =

  189. [197]

    , Title =

    Baumeister, Christiane and Hamilton, James D. , Title =. American Economic Review , Volume =. 2019 , Month =

  190. [198]

    Journal of Monetary Economics , volume =

    The long and variable lags of monetary policy:. Journal of Monetary Economics , volume =. 2024 , doi =

  191. [199]

    Local Projection Inference Is Simpler and More Robust Than You Think , journal =

    Montiel Olea, Jos. Local Projection Inference Is Simpler and More Robust Than You Think , journal =. 2021 , volume =

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.