Bayesian Econometrics
Bayesian econometrics is the branch of econometrics that treats every unknown quantity—a coefficient, a variance, a policy parameter, even the choice of model—as a random variable described by a probability distribution.
11 frameworks, from 1946 to today · Econometrics · Economics
- Objective and Reference Priors 1946-Present
- Subjective Bayesian Decision Theory 1954-Present
- Conjugate prior 1960-1990
- Bayesian hierarchical modeling 1972-Present
- Bayesian Nonparametrics 1973-Present
- Monte Carlo integration for Bayesian estimation 1978-1995
- Shrinkage Priors and Bayesian VARs 1980-Present
- Bayesian Model Comparison and Averaging 1990-Present
- Markov Chain Monte Carlo 1990-Present
- Bayesian DSGE Estimation 2000-Present
- Variational Inference and Probabilistic Programming 2010-Present
Overview
Bayesian econometrics is the branch of econometrics that treats every unknown quantity—a coefficient, a variance, a policy parameter, even the choice of model—as a random variable described by a probability distribution1. Where classical (frequentist) econometrics asks how often a procedure would be right across repeated samples, the Bayesian asks what the data, combined with prior knowledge, imply about the quantities of interest12. The answer is a posterior distribution: a probability statement about the parameters given the data actually observed1. What is at stake is how economists reason under uncertainty—whether prior knowledge is a legitimate input, how to forecast, and how to compare rival theories. The history of the field is largely a history of overcoming a computational obstacle: for most of the twentieth century the mathematics of Bayes' rule was simple but the integrals it required were impossible to evaluate, until simulation methods changed everything23.
Bayes' Rule and the Early Case for Subjective Probability
The foundation is Bayes' theorem, published posthumously by Thomas Bayes in 1763 and generalized by Pierre-Simon Laplace, who used it to estimate unknown quantities in astronomy and demography. The rule is a piece of arithmetic: the posterior is proportional to the likelihood (how probable the observed data are under each parameter value) times the prior (what was believed before seeing the data)1. Laplace applied it widely, but the method fell out of favor in the nineteenth and early twentieth centuries as statisticians sought "objective" procedures that did not require a prior.
The modern revival in economics came through decision theory. Leonard Savage's The Foundations of Statistics (1954) gave an axiomatic justification: an agent whose choices obey certain consistency conditions behaves as if she holds a subjective probability distribution and maximizes expected utility4. This made subjective probability respectable for economists, who were already comfortable with expected-utility reasoning. In econometrics the case was pressed by Arnold Zellner, whose textbook An Introduction to Bayesian Inference in Econometrics (1971) became the field's founding reference. Zellner argued that the Bayesian framework handled nuisance parameters, sequential learning, and model uncertainty more coherently than sampling theory, and he built much of the early toolkit himself.
Conjugate Priors and the Age of Closed Forms
Because the required integrals were intractable, early Bayesian econometrics depended on conjugacy: choosing a prior distribution of a mathematical form that combines with the likelihood to yield a posterior of the same family, so the answer can be written down in closed form. The workhorse is the normal–linear model, where a normal prior on regression coefficients paired with a known or inverse-gamma-distributed variance yields a normal posterior. Zellner's work on this "natural conjugate" setup, and the parallel development of the normal-gamma and inverted-Wishart families, let economists compute posteriors by hand or with a desk calculator.
The limitation was severe: conjugate analysis worked only for a narrow class of models, and the priors it required were often convenient rather than defensible. This tension drove the next debate.
The Battle over Priors: Objective, Reference, and Hierarchical
If the prior is arbitrary, the posterior is arbitrary—so a large strand of the field sought priors that let the data "speak." Harold Jeffreys in the 1930s and 1940s proposed improper and invariant priors (rules that give the same answer regardless of how a variable is measured), and in econometrics Zellner and others developed "diffuse" or "noninformative" priors intended to represent ignorance. The difficulty is that ignorance is hard to formalize: a flat prior on one parameter implies a non-flat prior on its transformations, and improper priors can produce improper (non-integrable) posteriors.
José Bernardo (1979) proposed reference priors, designed to maximize the information the data contribute, and James Berger and colleagues developed the theory further through the 1980s and 1990s. A different resolution was hierarchical modeling: instead of fixing a prior, treat its parameters (hyperparameters) as themselves drawn from a higher-level distribution, so the data inform how much prior weight to give5. This multilevel structure, formalized in the 1970s and widely adopted in the social sciences, let economists pool information across units—states, firms, countries—borrowing strength where data are thin. The debate between subjective, objective, and hierarchical priors never fully closed; it remains a live methodological choice.
The Minnesota Prior and Bayesian Vector Autoregressions
A decisive practical success came in macroeconomics. Christopher Sims's vector autoregression (VAR), introduced in 1980, modeled each variable as a function of its own and others' past values—flexible but parameter-heavy, so that unrestricted VARs overfit and forecast poorly. Robert Litterman, working at the Federal Reserve Bank of Minneapolis, proposed in the early 1980s a Minnesota prior: shrink coefficients toward the belief that each variable follows a random walk (its own recent value, with others' lags contributing little), with the strength of shrinkage tuned by a small number of hyperparameters. The result, the Bayesian VAR, forecast better than unrestricted VARs and became a standard tool at central banks. It was an early demonstration that a shrinkage prior—one that pulls estimates toward a simple benchmark—could beat both classical and naive alternatives, an idea that later flowered into regularization across econometrics.
Simulation Breaks the Computational Barrier
The real transformation came from computing1. In the late 1980s and early 1990s, statisticians developed Markov chain Monte Carlo (MCMC): instead of computing the posterior integral directly, draw a long sequence of random values whose long-run distribution is the posterior, then summarize it by averaging. The Metropolis–Hastings algorithm (1953) and the Gibbs sampler (popularized by Geman and Geman in 1984, and brought into statistics by Gelfand and Smith in 1990) made this practical. The Gibbs sampler is especially suited to econometrics because many models decompose into conditional distributions that are conjugate even when the joint posterior is not.
The effect was to dissolve the closed-form constraint5. Nicholas Polson, in a 1990s body of work with various coauthors, showed how to fit models with regime switching, stochastic volatility, and heavy-tailed errors by augmenting the parameter space with latent variables and sampling them. The Bayesian approach now handled models that frequentist methods could barely estimate. A parallel line, Monte Carlo integration more generally—using random draws to approximate integrals—underpinned all of it.
Model Comparison, Averaging, and Diagnostics
Once models could be estimated, the question of which model arose. The Bayesian answer is the posterior odds: the ratio of each model's marginal likelihood (the probability of the data under the model, averaging over its parameters)67. This is elegant but sensitive to priors, and computing marginal likelihoods is hard—leading to approximations such as the Bayesian information criterion and, later, bridge sampling and nested sampling. Bayesian model averaging takes the logic further: rather than pick one model, weight all candidate models by their posterior probability and average their predictions8. Developed in the 1990s (with influential work by Adrian Raftery and colleagues), it addressed the "which regressors?" problem in growth economics and forecasting, and it made explicit that model uncertainty is itself a quantity to be reported. Diagnostics—checking whether a fitted model reproduces features of the data—grew alongside, using posterior predictive checks and convergence measures for MCMC chains.
Bayesian Macroeconomics: DSGE Estimation
The most ambitious application came in macroeconomics. Dynamic stochastic general equilibrium (DSGE) models—microfounded models of the whole economy—were traditionally calibrated rather than estimated, because their likelihoods were intractable. In the 2000s, work by Frank Schorfheide, Jesús Fernández-Villaverde, Juan Rubio-Ramírez, and others showed how to estimate them by Bayesian methods: solve the model, linearize it into a state-space form, evaluate the likelihood with the Kalman filter, and sample the posterior with MCMC. This let macroeconomists bring formal theory and data together, compare competing model specifications, and produce forecasts with error bands. Bayesian DSGE estimation became a standard tool at central banks and in academic macro, and it remains a benchmark against which newer methods are judged.
Nonparametrics and the Modern Toolkit
Two further developments broadened the field. Bayesian nonparametrics replaces fixed finite-dimensional parameters with infinite-dimensional objects—random functions, Dirichlet processes, infinite mixtures—so that model complexity grows with the data. The Dirichlet process, introduced by Thomas Ferguson in 1973, and later developments such as the hierarchical Dirichlet process and Gaussian-process priors, gave economists flexible tools for density estimation, clustering, and heterogeneous-agent modeling without committing to a fixed number of components.
Most recently, variational inference and probabilistic programming have changed how Bayesian models are built and fit9. Variational methods replace sampling with optimization: approximate the posterior by the closest member of a tractable family, trading exactness for speed on large datasets10. Probabilistic programming languages (Stan, PyMC, and others) let a researcher write a model declaratively and have the software choose an inference algorithm91112. These tools have lowered the barrier to entry and pushed Bayesian methods toward the scale of modern economic data.
The Present
Bayesian econometrics today is less a separate school than a set of methods woven through the discipline. MCMC and its descendants remain the default for structural macro and time-series work; hierarchical models are standard in panel and multilevel settings; Bayesian model averaging and shrinkage priors are routine in forecasting and in high-dimensional regression; and variational and probabilistic-programming approaches are spreading as datasets grow. The old debates persist in new form—how much prior information is legitimate, how to validate a model that was fit by simulation, how to scale inference to very large problems—and the field's history is, in large part, the record of successive answers to the question of what a probability statement about an unknown parameter should mean.
References
- Bayesian econometrics (introduction, Basics, History), Wikipedia
- Bayesian statistics, Wikipedia
- Bayesian inference (History), Wikipedia
- Subjective expected utility, Wikipedia
- Prior probability, Wikipedia
- Bayes factor, Wikipedia
- Marginal likelihood, Wikipedia
- Ensemble learning (Bayesian model averaging), Wikipedia
- Probabilistic programming, Wikipedia
- Variational Bayesian methods, Wikipedia
- Stan (software), Wikipedia
- PyMC, Wikipedia
The frameworks of Bayesian Econometrics
Each framework is a way of working the field has used: when it held, and the question it answers.
- 1946-Present
Objective and Reference Priors
What prior lets the data speak for themselves?
- 1954-Present
Subjective Bayesian Decision Theory
How should beliefs be represented and updated?
- 1960-1990
- 1972-Present
Bayesian hierarchical modeling
How can units borrow strength from each other?
- 1973-Present
Bayesian Nonparametrics
How can models avoid fixed finite parameterizations?
- 1978-1995
Monte Carlo integration for Bayesian estimation
How can integrals be approximated by random draws?
- 1980-Present
Shrinkage Priors and Bayesian VARs
How can many coefficients be shrunk toward a benchmark?
- 1990-Present
Bayesian Model Comparison and Averaging
Which model should we believe, and how much?
- 1990-Present
Markov Chain Monte Carlo
How can we sample from an intractable posterior?
- 2000-Present
Bayesian DSGE Estimation
How can formal macroeconomic theory be fitted to data?
- 2010-Present
Variational Inference and Probabilistic Programming
How can posterior inference scale to huge models?
How the ideas connect
Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.
Markov Chain Monte Carlo grew out of Monte Carlo integration for Bayesian estimation
Markov chain Monte Carlo grew out of Monte Carlo integration by replacing independent draws with a dependent chain whose stationary distribution is the posterior. The Gibbs sampler and Metropolis-Hastings made the simulation approach practical for high-dimensional posteriors.
Carried by Stuart Geman; Donald Geman; Alan Gelfand.
Bayesian DSGE Estimation influenced by Markov Chain Monte Carlo
Bayesian DSGE estimation became possible because MCMC could sample the posterior of a linearized state-space model whose likelihood was evaluated by the Kalman filter. Schorfheide and Fernández-Villaverde built directly on MCMC methods.
Carried by Frank Schorfheide; Jesús Fernández-Villaverde; Juan Rubio-Ramírez.
Subjective Bayesian Decision Theory rival of Objective and Reference Priors
Objective and reference priors and subjective Bayesian decision theory offered competing answers to how priors should be chosen. Savage's axiomatic subjective probability (1954) defined itself against the idea that a prior could be derived from ignorance alone.
Carried by Leonard Savage; The Foundations of Statistics (1954).
Conjugate prior influenced by Subjective Bayesian Decision Theory
Subjective Bayesian decision theory made the prior a legitimate input, and conjugate analysis supplied the closed-form machinery for applying it in regression models. Zellner's natural conjugate setup operationalized Savage-style subjective priors for econometric practice.
Carried by Arnold Zellner; An Introduction to Bayesian Inference in Econometrics (1971).
Bayesian hierarchical modeling grew out of Conjugate prior
Hierarchical modeling extended conjugate analysis by putting a prior on the prior's parameters, so that conjugate conditional distributions could be chained across levels. The multilevel structure grew directly out of the closed-form conjugate toolkit.
Carried by Dennis Lindley; Adrian Smith.
Shrinkage Priors and Bayesian VARs influenced by Conjugate prior
The Minnesota prior applied conjugate normal priors to vector autoregressions, shrinking coefficients toward a random-walk benchmark. Litterman's Bayesian VAR was built directly on the normal-linear conjugate framework.
Carried by Robert Litterman; Christopher Sims.
Markov Chain Monte Carlo influenced by Conjugate prior
The Gibbs sampler's usefulness in econometrics rested on decomposing models into conditional distributions that were conjugate even when the joint posterior was not. Conjugate blocks became the building blocks of MCMC algorithms.
Carried by Alan Gelfand; Adrian Smith.
Bayesian DSGE Estimation influenced by Bayesian Model Comparison and Averaging
Bayesian model comparison provided the framework for comparing competing DSGE specifications by marginal likelihood and for averaging across them. DSGE practitioners adopted posterior odds and model averaging as standard tools.
Carried by Frank Schorfheide.
Bayesian Model Comparison and Averaging influenced by Markov Chain Monte Carlo
MCMC made it feasible to sample across model space and to compute posterior odds and marginal likelihoods for many candidate models. Bayesian model averaging and its diagnostics were implemented with MCMC draws.
Carried by Adrian Raftery.
Variational Inference and Probabilistic Programming influenced by Markov Chain Monte Carlo
Variational inference and probabilistic programming were developed partly as scalable alternatives to MCMC for large models, taking over the task of posterior approximation. They inherited the same posterior-inference problem that MCMC had solved by sampling.
Conjugate prior influenced by Objective and Reference Priors
The search for defensible priors fed into conjugate analysis, which often used diffuse or natural conjugate forms to represent weak prior information. The normal-gamma and inverted-Wishart families were adopted partly as tractable objective choices.
Carried by Arnold Zellner.
Shrinkage Priors and Bayesian VARs influenced by Bayesian hierarchical modeling
Hierarchical thinking about hyperparameters shaped how shrinkage strength in Bayesian VARs was treated as an unknown to be estimated or tuned. The Minnesota prior's few hyperparameters controlling shrinkage reflect this multilevel logic.
Carried by Robert Litterman.
Variational Inference and Probabilistic Programming influenced by Bayesian Nonparametrics
Nonparametric priors such as Dirichlet processes became standard targets for variational approximations and probabilistic-programming implementations. The modern toolkit extended inference to the infinite-dimensional models nonparametrics had introduced.
Bayesian Model Comparison and Averaging influenced by Monte Carlo integration for Bayesian estimation
Monte Carlo integration supplied the simulation tools later used to approximate marginal likelihoods and posterior model probabilities. Bayesian model averaging relied on such simulation-based estimates of model evidence.
Carried by Adrian Raftery.
Bayesian DSGE Estimation influenced by Shrinkage Priors and Bayesian VARs
The Bayesian VAR tradition supplied shrinkage priors and forecasting benchmarks that DSGE estimation was compared against and sometimes combined with. Macroeconomic Bayesian methods shared the Minnesota prior's concern with overfitting.
Carried by Christopher Sims.
Beyond the field
Where Bayesian Econometrics's ideas came from in other fields, and what they led to there: the strongest 12 of 36 links.
Objective and Reference Priors revived Inverse Probability in Statistical Inference
Bayes's inverse probability was largely displaced by frequentist methods, and objective prior work deliberately brought it back as a principled way to assign priors. Jeffreys built his invariance-based priors directly on the inverse-probability tradition.
Carried by Harold Jeffreys.
Monte Carlo integration for Bayesian estimation influenced by Monte Carlo Methods in Probability Theory
Monte Carlo integration for Bayesian estimation applies general Monte Carlo sampling to compute posterior integrals that have no closed form. The Metropolis and Gibbs samplers carried the simulation technique into Bayesian computation.
Carried by Metropolis et al. (1953); Stuart Geman; Donald Geman.
Shrinkage Priors and Bayesian VARs influenced by Vector autoregression in Time Series Econometrics
Bayesian VAR shrinkage priors are a technique within time series econometrics' treatment of vector autoregressions. The Minnesota prior and subsequent hierarchical priors were developed as tools for this class of models.
Carried by Robert Litterman.
Shrinkage Priors and Bayesian VARs influenced by Vector autoregression in Econometrics
The shrinkage-prior literature was built to estimate vector autoregressions, which are the object the priors are applied to. Litterman's Minnesota prior was designed specifically for Bayesian VAR estimation.
Carried by Robert Litterman; Thomas Doan.
Shrinkage Priors and Bayesian VARs grew out of Bayesian Econometrics in Econometrics
Shrinkage priors for VARs grew directly out of Bayesian econometrics, applying the Bayesian posterior machinery to vector autoregressions. Litterman's Minnesota prior (1980) is the founding instance of this branch.
Carried by Robert Litterman.
Bayesian Model Comparison and Averaging grew out of Bayesian Econometrics in Econometrics
Bayesian model comparison and averaging is a branch of Bayesian econometrics, using marginal likelihoods and posterior model probabilities within econometric modeling. It grew directly out of the Bayesian econometric tradition.
Markov Chain Monte Carlo influenced by Monte Carlo Methods in Probability Theory
Monte Carlo methods, developed at Los Alamos in the 1940s, are the sampling foundation on which Markov chain Monte Carlo was built. MCMC extends plain Monte Carlo by constructing a Markov chain whose stationary distribution is the target.
Carried by Stanislaw Ulam; Nicholas Metropolis.
Markov Chain Monte Carlo influenced by Monte Carlo and Markov Chain Monte Carlo in Stochastic Processes
Monte Carlo and Markov chain methods in stochastic processes provided the sampling techniques that MCMC assembled into a general inference tool. The Metropolis algorithm and later Gibbs sampling carried this passage.
Carried by Nicholas Metropolis.
Markov Chain Monte Carlo influenced by Markov chain in Markov Processes
Markov chain theory, founded by Andrey Markov in 1906, is the mathematical basis of MCMC algorithms. The Metropolis algorithm and Gibbs sampling are Markov chains constructed to sample from a target distribution.
Carried by Andrey Markov.
Markov Chain Monte Carlo influenced by Markov chain in Stochastic Processes
Markov chains in stochastic processes supply the transition-kernel framework that MCMC relies on. The Metropolis-Hastings algorithm is defined as a Markov chain with a specified stationary distribution.
Carried by Andrey Markov.
Markov Chain Monte Carlo influenced by Markov Chain Monte Carlo in Markov Processes
The Metropolis algorithm (1953) and the theory of Markov chains supplied the sampling machinery that MCMC later generalized. Metropolis-Hastings and Gibbs sampling build directly on this Markov process foundation.
Carried by Nicholas Metropolis; W. K. Hastings.
Bayesian DSGE Estimation influenced by Dynamic Stochastic General Equilibrium in General Equilibrium
DSGE models are the object that Bayesian DSGE estimation estimates; the framework applies Bayesian methods to this model class. The estimation procedure presupposes the DSGE representation.

Bayesian Econometrics is a planet in Nousfarer
Its city has a tower for each of these 11 frameworks. Land, walk in, and climb one concept at a time.
Land on Bayesian EconometricsRelated fields
Fields its ideas are linked to
- Econometrics 6 links
- Bayesian Statistics 5 links
- Time Series Econometrics 3 links
- Probability Theory 3 links
- Time Series Analysis 3 links
- General Equilibrium 2 links
- Markov Processes 2 links
- Stochastic Processes 2 links
- Statistical Inference 2 links
- Machine Learning 1 link
- Experimental Economics 1 link
- Structural Estimation 1 link
More in Econometrics
- Econometrics
- Microeconometrics
- ML Econometrics
- Panel Data Methods
- Structural Estimation
- Time Series Econometrics
See where Bayesian Econometrics sits on the Map of Human Knowledge