2017 SIdE Summer School of Econometrics
Even this year the SIdE Summer School of Econometrics will be two-week long.
The Summer School will be in July 2017.
New deadline for applications: May 31st, 2017
For information on enrollment, fees, etc, see the links at the end of this page
First week topic: Non-Parametric Bayesian Models for Big Data and Macro/FInance
Speakers:
Dates: from July 10th through July 14th, 2017
Syllabus: See below or download syllabus for first week
Second week topic: High-Dimensional Econometrics
Speakers:
Dates: from July 17th through July 21st, 2017
Syllabus: See Below or download syllabus for second week
**********************************************************************************
FIRST WEEK:
**********************************************************************************
Non-Parametric Bayesian Models for Big Data and Macro/Finance
Tamara Broderick (MIT)
Syllabus
Description
Nonparametric Bayesian methods make use of infinite-dimensional mathematical structures to allow the practitioner to learn more from their data as the size of their data set grows. What does that mean, and how does it work in practice? In this tutorial, we'll cover why machine learning and statistics need more than just parametric Bayesian inference. We'll introduce such foundational nonparametric Bayesian models as the Dirichlet process and Chinese restaurant process and touch on the wide variety of models available in nonparametric Bayes. Along the way, we'll see what exactly nonparametric Bayesian methods are and what they accomplish.
Prerequisites
* Know what a prior, likelihood, and posterior are.
* Know how to use Bayes' Theorem to calculate a posterior for both discrete and continuous parametric distributions.
* Understand what a generative model is.
* Have a basic idea of what Gibbs sampling is and when it is useful (at least check out the Wikipedia article in advance).
Joshua Chan (University of Technology Sydney)
Syllabus
Course Description: Bayesian econometric methods are increasingly popular in em-
pirical macroeconomics. In particular, flexible models that allow for non-Gaussian dis-
tributions and time variation in coefficients and volatility are now widely used among
macroeconomists. The overarching purpose of this course is to bring you to the research
frontier so that you are prepared to do research in Bayesian macroeconometrics.
This course first provides an overview of Bayesian theory and computations. It then
gives a brief review of the linear regression and the Gibbs sampler. Some flexible varia-
tions of the linear regression will then be introduced, along with various more sophisti-
cated MCMC algorithms. We will then dive into a few state-of-the-art macroeconometric
models, including unobserved components models, time-varying parameter models and
stochastic volatility models.
Course notes: The course notes are available at http://joshuachan.org/notes_BayesMacro.html
Course outline: We will cover the following topics in this course:
1. Overview of Bayesian econometrics: Bayesian theory and computations
2. Linear regressions: Gaussian and t errors, moving average errors, independence-
chain Metropolis-Hastings, Griddy-Gibbs
3. Mixture models: scale mixture of normals, finite mixture of normals
4. Linear state space models: unobserved components models, time-varying parameter
models, precision-based samplers
5. Nonlinear state space models: stochastic volatility model, stochastic volatility in
mean, auxiliary mixture sampler
**********************************************************************************
SECOND WEEK:
**********************************************************************************
High-dimensional econometrics
Mehmet Caner (OSU) and Anders Bredahl Kock (Aarhus)
Course description: Recent years have seen a massive increase in the availability of large data sets. In this course we will cover some of the techniques that have been developed to analyze such data sets. Particular attention will be given to precise estimation, variable selection, and hypothesis testing. We will also see how to implement
the some of the techniques in R.
Topics covered
1. The Lasso and some of its asymptotic properties when the dimension of
the model is .xed. Knight and Fu [2000].
2. Introduction to the oracle property. The adaptive Lasso as an example of
an estimator possessing the oracle property. Zou [2006].
3. The adaptive Lasso for instrumental variable selection. Caner and Fan
[2015] .
4. The adaptive elastic net for generalized method of moments. Caner and
Zhang [2014].
5. Upper bounds on the estimation error in the `1-norm and variable selection
via thresholding. Lounici [2008].
6. Finite sample oracle inequalities for the Lasso when the number of variables
is larger than the sample size. Bickel et al. [2009], Buhlmann and
Van De Geer [2011], Caner and Kock [2014].
7. Oracle inequalities and inference in high-dimensional VAR models. Kock
and Callot [2015]. Some results from applications to large macroeconomic
data sets. Callot and Kock [2014].
8. Uniformly valid inference in high-dimensional models when the number
of variables is larger than the number of parameters. van de Geer et al.
[2014] and Caner and Kock [2014].
9. How to use the glmnet and the lars package in R to implement Lasso-type
estimators.
The recent book van de Geer [2016] may also be useful.
References
Peter J Bickel, Ya'acov Ritov, and Alexandre B Tsybakov. Simultaneous analysis
of lasso and dantzig selector. The Annals of Statistics, pages 1705{1732,
2009.
Peter Buhlmann and Sara Van De Geer. Statistics for high-dimensional data:
methods, theory and applications. Springer Science & Business Media, 2011.
Laurent AF Callot and Anders B Kock. Oracle e.cient estimation and forecasting
with the adaptive lasso and the adaptive group lasso in vector autoregressions.
Essays in Nonlinear Time Series Econometrics, page 238, 2014.
Mehmet Caner and Qingliang Fan. Hybrid generalized empirical likelihood estimators:
Instrument selection with adaptive lasso. Journal of Econometrics,
187(1):256{274, 2015.
Mehmet Caner and Anders Bredahl Kock. Asymptotically honest con.dence
regions for high dimensional parameters by the desparsi.ed conservative lasso.
arXiv preprint arXiv:1410.4208, 2014.
Mehmet Caner and Hao Helen Zhang. Adaptive elastic net for generalized
methods of moments. Journal of Business & Economic Statistics, 32(1):30{
47, 2014.
Keith Knight and Wenjiang Fu. Asymptotics for lasso-type estimators. Annals
of Statistics, pages 1356{1378, 2000.
Anders Bredahl Kock and Laurent Callot. Oracle inequalities for high dimensional
vector autoregressions. Journal of Econometrics, 186(2):325{344, 2015.
Karim Lounici. Sup-norm convergence rate and sign concentration property of
lasso and dantzig estimators. Electronic Journal of Statistics, 2:90{102, 2008.
Sara van de Geer, Peter Buhlmann, Yaacov Ritov, and Ruben Dezeure. On
asymptotically optimal con.dence regions and tests for high-dimensional
models. The Annals of Statistics, 42(3):1166{1202, 2014.
Sara A van de Geer. Estimation and testing under sparsity. Springer, 2016.
Hui Zou. The adaptive lasso and its oracle properties. Journal of the American
Statistical Association, 101(476):1418{1429, 2006.
Organizer (on behalf of SIDE): Juri Marcucci
Venue: Bank of Italy Sadiba Center, Perugia, Italy
Some useful links:
How to reach Perugia (PDF in Italian)
How to reach Perugia (PDF in English)
If you have any further questions, please send an email to "Società Italiana di Econometria - SIdE" at info@side-iea.it