Search Results
Report
Forming priors for DSGE models (and how it affects the assessment of nominal rigidities)
This paper discusses prior elicitation for the parameters of dynamic stochastic general equilibrium (DSGE) models and provides a method for constructing prior distributions for a subset of these parameters from beliefs about the moments of the endogenous variables. The empirical application studies the role of price and wage rigidities in a New Keynesian DSGE model and finds that standard macro time series cannot discriminate among theories that differ in the quantitative importance of nominal frictions.
Working Paper
Estimating the Missing Intercept
Cross-sectional data have proven to be increasingly useful for macroeconomic research. However, their use often leads to the 'missing intercept' problem in which aggregate general equilibrium effects and policy responses are absorbed into fixed effects. We present a statistical approach to jointly estimate aggregate and idiosyncratic effects within a panel framework, leveraging identification strategies coming from both cross-sectional or time-series settings. We then apply our methodology to study government spending multipliers (Nakamura and Steinsson, 2014) and wealth effects from stock ...
Working Paper
What Do Sectoral Dynamics Tell Us About the Origins of Business Cycles?
We use economic theory to rank the impact of structural shocks across sectors. This ranking helps us to identify the origins of U.S. business cycles. To do this, we introduce a Hierarchical Vector Auto-Regressive model, encompassing aggregate and sectoral variables. We find that shocks whose impact originate in the "demand" side (monetary, household, and government consumption) account for 43 percent more of the variance of U.S. GDP growth at business cycle frequencies than identified shocks originating in the "supply" side (technology and energy). Furthermore, corporate financial shocks, ...
Working Paper
A Unified Framework to Estimate Macroeconomic Stars
This paper develops a semi-structural model to jointly estimate “stars” — long-run levels of output (its growth rate), the unemployment rate, the real interest rate, productivity growth, price inflation, and wage inflation. It features links between survey expectations and stars, time-variation in macroeconomic relationships, and stochastic volatility. Survey data help discipline stars’ estimates and have been crucial in estimating a high-dimensional model since the pandemic. The model has desirable real-time properties, competitive forecasting performance, and superior fit to the ...
Working Paper
Factor Selection and Structural Breaks
We develop a new approach to select risk factors in an asset pricing model that allows the set to change at multiple unknown break dates. Using the six factors displayed in Table 1 since 1963, we document a marked shift towards parsimonious models in the last two decades. Prior to 2005, five or six factors are selected, but just two are selected thereafter. This finding offers a simple implication for the factor zoo literature: ignoring breaks detects additional factors that are no longer relevant. Moreover, all omitted factors are priced by the selected factors in every regime. Finally, ...
Working Paper
Estimating (Markov-Switching) VAR Models without Gibbs Sampling: A Sequential Monte Carlo Approach
Vector autoregressions with Markov-switching parameters (MS-VARs) fit the data better than do their constant-parameter predecessors. However, Bayesian inference for MS-VARs with existing algorithms remains challenging. For our first contribution, we show that Sequential Monte Carlo (SMC) estimators accurately estimate Bayesian MS-VAR posteriors. Relative to multi-step, model-specific MCMC routines, SMC has the advantages of generality, parallelizability, and freedom from reliance on particular analytical relationships between prior and likelihood. For our second contribution, we use SMC's ...
Working Paper
Scenario-based Quantile Connectedness of the U.S. Interbank Liquidity Risk Network
We characterize the U.S. interbank liquidity risk network based on a supervisory dataset, using a scenario-based quantile network connectedness approach. In terms of methodology, we consider a quantile vector autoregressive model with unobserved heterogeneity and propose a Bayesian nuclear norm estimation method. A common factor structure is employed to deal with unobserved heterogeneity that may exhibit endogeneity within the network. Then we develop a scenario-based quantile network connectedness framework by accommodating various economic scenarios, through a scenario-based moving average ...
Working Paper
Estimating (Markov-Switching) VAR Models without Gibbs Sampling: A Sequential Monte Carlo Approach
Vector autoregressions with Markov-switching parameters (MS-VARs) offer dramatically better data fit than their constant-parameter predecessors. However, computational complications, as well as negative results about the importance of switching in parameters other than shock variances, have caused MS-VARs to see only sparse usage. For our first contribution, we document the effectiveness of Sequential Monte Carlo (SMC) algorithms at estimating MSVAR posteriors. Relative to multi-step, model-specific MCMC routines, SMC has the advantages of being simpler to implement, readily parallelizable, ...
Report
Fitting observed inflation expectations
This paper provides evidence on the extent to which inflation expectations generated by a standard Christiano et al. (2005)/Smets and Wouters (2003)?type DSGE model are in line with what is observed in the data. We consider three variants of this model that differ in terms of the behavior of, and the public?s information on, the central banks? inflation target, allegedly a key determinant of inflation expectations. We find that: 1) time-variation in the inflation target is needed to capture the evolution of expectations during the post-Volcker period; 2) the variant where agents have ...
Working Paper
Tempered Particle Filtering
The accuracy of particle filters for nonlinear state-space models crucially depends on the proposal distribution that mutates time t-1 particle values into time t values. In the widely-used bootstrap particle filter this distribution is generated by the state-transition equation. While straightforward to implement, the practical performance is often poor. We develop a self-tuning particle filter in which the proposal distribution is constructed adaptively through a sequence of Monte Carlo steps. Intuitively, we start from a measurement error distribution with an inflated variance, and then ...