Measuring prior sensitivity and prior informativeness in large Bayesian models

Müller, Ulrich K.

doi:10.1016/j.jmoneco.2012.09.003

Cited by 55 publications

(51 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…At the same time, this prior has a strong influence on the empirical performance of the model, as discussed in Del Negro and Schorfheide (2013). Closely related, Müller (2011) derives an analytical approximation for the sensitivity of posterior means to shifts in prior means and finds evidence that the stickiness of prices and wages is driven substantially by the priors.…”

Section: The Smets-wouters Modelmentioning

confidence: 85%

Sequential Monte Carlo Sampling for DSGE Models

Herbst

Schorfheide

2013

FEDS

View full text Add to dashboard Cite

We develop a sequential Monte Carlo (SMC) algorithm for estimating Bayesian dynamic stochastic general equilibrium (DSGE) models, wherein a particle approximation to the posterior is built iteratively through tempering the likelihood. Using three examples-an artificial statespace model, the Smets and Wouters (2007) model, and Schmitt-Grohé and Uribe's (2012) news shock model-we show that the SMC algorithm is better suited for multimodal and irregular posterior distributions than the widely-used random-walk Metropolis-Hastings algorithm. We find that a more diffuse prior for the Smets and Wouters (2007) model improves its marginal data density and that a slight modification of the prior for the news shock model leads to important changes in the posterior inference about the importance of news shocks for fluctuations in hours worked. Unlike standard Markov chain Monte Carlo (MCMC) techniques, the SMC algorithm is well suited for parallel computing.

show abstract

Section: The Smets-wouters Modelmentioning

confidence: 85%

Sequential Monte Carlo Sampling for DSGE Models

Herbst

Schorfheide

2013

FEDS

View full text Add to dashboard Cite

show abstract

“…(Emphasis added. )similar to Müller's (2012) measure of prior sensitivity for Bayesian models, which allows readers to adjust reported results to better reflect their own priors.…”

Section: Relationship To Prior Literaturementioning

confidence: 99%

Measuring the Sensitivity of Parameter Estimates to Estimation Moments*

Andrews

Gentzkow

Shapiro

2017

The Quarterly Journal of Economics

209

View full text Add to dashboard Cite

We propose a local measure of the relationship between parameter estimates and the moments of the data they depend on. Our measure can be computed at negligible cost even for complex structural models. We argue that reporting this measure can increase the transparency of structural estimates, making it easier for readers to predict the way violations of identifying assumptions would affect the results. When the key assumptions are orthogonality between error terms and excluded instruments, we show that our measure provides a natural extension of the omitted variables bias formula for nonlinear models. We illustrate with applications to published articles in several fields of economics.

show abstract

“…First, following the suggestion in Mueller [2012], we shifted the prior mean of each of the standard deviations for the trends by one (prior) standard deviation and examined the change in the posterior mean. Table 3.…”

Section: The Information In the Datamentioning

confidence: 99%

What Can the Data Tell Us About the Equilibrium Real Interest Rate?

Kiley

2015

financeandeconomicsa

View full text Add to dashboard Cite

The equilibrium real interest rate (r * ) is the short-term real interest rate that, in the long run, is consistent with aggregate production at potential and stable inflation. Estimation of r * faces considerable econometric and empirical challenges. On the econometric front, classical inference confronts the "pile-up" problem. Empirically, the co-movement of output, inflation, unemployment, and real interest rates is too weak to yield precise estimates of r * . These challenges are addressed by applying Bayesian methods and examining the role of several "demand shifters", including asset prices, fiscal policy, and credit conditions. We find that the data provide relatively little information on the r * data-generating process, as the posterior distribution of this process lies very close to its prior. This result contrasts sharply with those for the trend growth or natural rate of unemployment processes. Second, credit spreads are very important for the estimated links between output and interest rates and hence for estimates of r * . Estimates of r * that account for this range of considerations are more stable than other estimates, with r * at the end of 2014 equal to approximately 1-1/4 percent.

show abstract

Measuring prior sensitivity and prior informativeness in large Bayesian models

Cited by 55 publications

References 35 publications

Sequential Monte Carlo Sampling for DSGE Models

Sequential Monte Carlo Sampling for DSGE Models

Measuring the Sensitivity of Parameter Estimates to Estimation Moments*

What Can the Data Tell Us About the Equilibrium Real Interest Rate?

Contact Info

Product

Resources

About