CV - Job Market Paper  
 

Ganics, Gergely

Job market candidate

Contact information

C. Ramon Trias Fargas, 25-27
Barcelona
Tel. +34 93 542 1191
gergely.ganics@upf.edu

Available for Interviews at :

Simposio de la Asociación Española de Economía (SAEe), December 15-17, Bilbao, Spain

Allied Social Science Associations (ASSA), January 6-8, Chicago, US

 

 

Research interests

Econometrics, Macroeconomics, Forecasting

Placement officer

Filippo Ippolito
filippo.ippolito@upf.edu
 

References

Barbara Rossi
barbara.rossi@upf.edu
Majid Al-Sadoon
majid.alsadoon@upf.edu
Christian Brownlees
christian.brownlees@upf.edu
Geert Mesters
geert.mesters@upf.edu

Research

"Optimal Density Forecast Combinations" (Job Market Paper)
Winner of the UniCredit & Universities Economics Job Market Best Paper Award 2016

How should researchers combine predictive densities to improve their forecasts? I propose consistent estimators of weights that deliver density forecast combinations approximating the true predictive density, conditional on the researcher's information set. Monte Carlo simulations confirm that the proposed methods work well for sample sizes of practical interest. An empirical example of forecasting monthly US industrial production demonstrates that the estimator delivers density forecasts which are superior to well-known benchmarks, such as the equal weights scheme. Specifically, I show that housing permits had valuable predictive power before and after the Great Recession. Furthermore, stock returns and corporate bond spreads proved to be useful predictors during the recent crisis, suggesting that financial variables help with density forecasting in a highly leveraged economy.
Data and Code
Interview on the Barcelona GSE website: BarcelonaGSE Interview

Non-technical summary of the paper: BarcelonaGSE The Voice
 

"Forecasting with DSGE Versus Reduced-Form Models: A Time-Variation Perspective"
The out-of-sample forecasting performance of a leading DSGE model is investigated. First, I demonstrate that, while the model delivers competitive forecasts against a number of statistical models, its predictive ability displays time-variation. Generally, in turbulent times, such as the recent financial crisis, simpler statistical models forecast better. Second, I show that swings in the model's absolute and relative out-of-sample performance are strongly related to its in-sample performance. Specifically, I find that the DSGE model's in-sample fit is highly informative in the early 2000's but the financial crisis deteriorated this link. Third, I find that extending  a DSGE model with financial frictions results in better forecasting performance in times of financial distress but not in other times.
Data and code are too big to be posted, please contact me if you need them

"Confidence Intervals for the Strength of Identification", with Atsushi Inoue and Barbara Rossi

This paper provides a grid-bootstrap method to construct confidence intervals for the strength of identification in instrumental variable models. Monte Carlo simulations show that the method has good small sample size and proper properties. An empirical investigation of the New Keynesian Phillips Curve shows that weak identification is a concern even if one uses a factor model to summarize all the relevant information from a large number of instruments.