{
  "title": "Benchmark Priors Revisited: On Adaptive Shrinkage and the Supermodel Effect in Bayesian Model Averaging",
  "publication": "IMF Working Papers, September 1, 2009",
  "sourceUrl": "https://www.imf.org/en/publications/wp/issues/2016/12/31/benchmark-priors-revisited-on-adaptive-shrinkage-and-the-supermodel-effect-in-bayesian-model-23292",
  "canonical": "https://www.imf.org/en/publications/wp/issues/2016/12/31/benchmark-priors-revisited-on-adaptive-shrinkage-and-the-supermodel-effect-in-bayesian-model-23292",
  "overlayPath": "/en/publications/wp/issues/2016/12/31/benchmark-priors-revisited-on-adaptive-shrinkage-and-the-supermodel-effect-in-bayesian-model-23292/index.md",
  "summary": "Default prior choices fixing Zellner's g are predominant in the Bayesian Model Averaging literature, but tend to concentrate posterior mass on a tiny set of models.",
  "sections": [
    {
      "heading": "Key findings",
      "content": "- Default prior choices fixing Zellner's g are predominant in the Bayesian Model Averaging literature but tend to concentrate posterior mass on a tiny set of models.\n- The paper demonstrates a \"supermodel effect\" whereby fixed-g priors produce posterior distributions heavily concentrated on few models.\n- A hyper-g prior is proposed to address the supermodel effect by providing data-dependent shrinkage that adapts posterior model distributions to data quality.\n- An application to determinants of economic growth identifies several covariates whose robustness differs considerably from previous results based on default fixed-g priors."
    },
    {
      "heading": "Analytical contributions",
      "content": "- Complements existing work on the hyper-g-prior by providing posterior expressions essential to:\n  - fully Bayesian analysis, and\n  - sound numerical implementation.\n- Emphasizes that adaptive (data-dependent) shrinkage alters posterior model distributions relative to fixed-g benchmark priors."
    },
    {
      "heading": "Methodology and experiments",
      "content": "- Analytical derivations of posterior expressions for the hyper-g-prior are provided to support implementation and interpretation.\n- A simulation experiment is used to illustrate the implications of prior choice for posterior inference, highlighting differences between fixed-g and hyper-g priors."
    },
    {
      "heading": "Implications for empirical practice",
      "content": "- Researchers using Bayesian Model Averaging should be aware that default fixed-g priors can concentrate posterior mass on a small subset of models (the supermodel effect), potentially overstating model robustness.\n- Employing a hyper-g prior yields shrinkage that adapts to data quality, which can change posterior model weights and the assessed robustness of covariates.\n- Reassessment of empirical findings (for example, determinants of economic growth) may be warranted when moving from fixed-g to hyper-g priors.\n\nIMF Working Paper — Benchmark Priors Revisited: On Adaptive Shrinkage and the Supermodel Effect in Bayesian Model Averaging (Martin Feldkircher and Stefan Zeugner, September 1, 2009).\n\n---\n\n Content in this bundle\n\n- wp09202\n  - wp09202 (Markdown version){rel=\"alternate\" type=\"text/markdown\"}\n  - wp09202 (PDF){rel=\"external\" type=\"application/pdf\"}\n\n---\n\nSource: https://www.imf.org/en/publications/wp/issues/2016/12/31/benchmark-priors-revisited-on-adaptive-shrinkage-and-the-supermodel-effect-in-bayesian-model-23292"
    }
  ],
  "bullets": [
    "[Markdown version](/en/publications/wp/issues/2016/12/31/benchmark-priors-revisited-on-adaptive-shrinkage-and-the-supermodel-effect-in-bayesian-model-23292/index.md)",
    "[Structured JSON version](/en/publications/wp/issues/2016/12/31/benchmark-priors-revisited-on-adaptive-shrinkage-and-the-supermodel-effect-in-bayesian-model-23292/index.json)",
    "[Bundle manifest](/en/publications/wp/issues/2016/12/31/benchmark-priors-revisited-on-adaptive-shrinkage-and-the-supermodel-effect-in-bayesian-model-23292/bundle-manifest.json)",
    "Authors: Martin Feldkircher, Stefan Zeugner",
    "Published: September 1, 2009",
    "Series: IMF Working Papers",
    "DOI: https://doi.org/10.5089/9781451873498.001",
    "Default prior choices fixing Zellner's g are predominant in the Bayesian Model Averaging literature but tend to concentrate posterior mass on a tiny set of models.",
    "The paper demonstrates a \"supermodel effect\" whereby fixed-g priors produce posterior distributions heavily concentrated on few models.",
    "A hyper-g prior is proposed to address the supermodel effect by providing data-dependent shrinkage that adapts posterior model distributions to data quality.",
    "An application to determinants of economic growth identifies several covariates whose robustness differs considerably from previous results based on default fixed-g priors.",
    "Complements existing work on the hyper-g-prior by providing posterior expressions essential to:",
    "Emphasizes that adaptive (data-dependent) shrinkage alters posterior model distributions relative to fixed-g benchmark priors.",
    "Analytical derivations of posterior expressions for the hyper-g-prior are provided to support implementation and interpretation.",
    "A simulation experiment is used to illustrate the implications of prior choice for posterior inference, highlighting differences between fixed-g and hyper-g priors.",
    "Researchers using Bayesian Model Averaging should be aware that default fixed-g priors can concentrate posterior mass on a small subset of models (the supermodel effect), potentially overstating model robustness.",
    "Employing a hyper-g prior yields shrinkage that adapts to data quality, which can change posterior model weights and the assessed robustness of covariates.",
    "Reassessment of empirical findings (for example, determinants of economic growth) may be warranted when moving from fixed-g to hyper-g priors.",
    "**_wp09202**"
  ],
  "related": [
    {
      "title": "_wp09202",
      "role": "document",
      "sourceUrl": "https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09202.pdf",
      "summary": {
        "path": "/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09202.pdf.md",
        "mime": "text/markdown"
      },
      "binary": {
        "path": "/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09202.pdf",
        "mime": "application/pdf"
      }
    }
  ],
  "alternates": {
    "markdown": "/en/publications/wp/issues/2016/12/31/benchmark-priors-revisited-on-adaptive-shrinkage-and-the-supermodel-effect-in-bayesian-model-23292/index.md",
    "json": "/en/publications/wp/issues/2016/12/31/benchmark-priors-revisited-on-adaptive-shrinkage-and-the-supermodel-effect-in-bayesian-model-23292/index.json",
    "bundleManifest": "/en/publications/wp/issues/2016/12/31/benchmark-priors-revisited-on-adaptive-shrinkage-and-the-supermodel-effect-in-bayesian-model-23292/bundle-manifest.json"
  },
  "generatedAtUtc": "2026-09-17T09:59:23.841Z"
}