Research Grants 23/01728-0 - Previsão, Finanças comportamentais - BV FAPESP
Advanced search
Start date
Betweenand

Econometric modeling and forecasting in high dimensional models

Grant number: 23/01728-0
Support Opportunities:Research Projects - Thematic Grants
Field of knowledge:Applied Social Sciences - Economics - Quantitative Methods Applied to Economics
Principal Investigator:Pedro Luiz Valls Pereira
Grantee:Pedro Luiz Valls Pereira
Host Institution: Escola de Economia de São Paulo (EESP). Fundação Getúlio Vargas (FGV). São Paulo , SP, Brazil
Pesquisadores principais:
Luiz Koodi Hotta ; Marcelo Fernandes
Associated researchers:Alexandre Rubesam ; André Barbosa Oliveira ; Andreza Aparecida Palma ; Bruno Tebaldi de Queiroz Barbosa ; Carlos Cesar Trucios Maza ; Diogo de Prince Mendonça ; Eduardo Fonseca Mendes ; Emerson Fernandes Marçal ; Jéfferson Augusto Colombo ; Mauricio Enrique Zevallos Herencia ; Pedro Luiz Paolino Chaim ; Ricardo Buscariolli Pereira ; Vitor Borges Monteiro

Abstract

This project deals with new methodologies for econometric modelling and forecasting in high dimensional and/or frequency setting, as well as mixed-data frequencies setting. One of the challenges in econometrics in high dimensional problems is to deal with complex world, including structural changes, using sparce vector autoregressive models and dense (factors). Applications of these methodologies include: estimation of covariance matrix using robust methods and statistical learning; dynamic factor model; asset pricing using statistical learning techniques; forecasting economic and financial time series using mixed-data frequencies and regularization; construction of vintage for economic and financial time series; modelling cryptofinance; exploits cross-sectional information to improve forecasts of stock index returns and volatility at the high frequency; employs textual data from major newspapers to predict stock volatility at B3; aims to assess the risk-adjusted performance of semivariance-managed portfolios for a very broad set of anomaly portfolios; statistical analysis of curve time series in the high dimensional; use nonparametric factor models to identify clusters of time series in high dimension vectors; production frontier estimation to incorporate environmental conditioning variables, allow for the dimension increase of the covariate set without suffering from the curse of dimensionality; incorporate time series models in the study. Results will be published in international journals with selective editorial policies and shall be presented in scientific conferences. Forming human resources is one of our main concerns. For this effect we will supervise pos-docs, undergraduate research assistantships, master theses and PhD dissertations. Seminars will be held regularly, in which the presentation of results, exchange of ideas, and where prospective talents may get acquainted with these research areas. In order to bring a serious push in this area in Brazil we intend to start a "Summer or Winter School" which will reunite every year international and Brazilian guest speakers, the project team, advanced undergraduate and graduate students. Two intermediate short courses will be held as well as a workshop where the state of the art and open problems will be discussed, and new and/or established scientific collaborations will happen. (AU)

Articles published in Agência FAPESP Newsletter about the research grant:
More itemsLess items
Articles published in other media outlets ( ):
More itemsLess items
VEICULO: TITULO (DATA)
VEICULO: TITULO (DATA)

Scientific publications
(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)
MARCAL, EMERSON FERNANDES. Testing rational expectations in a cointegrated VAR with structural change. INTERNATIONAL REVIEW OF FINANCIAL ANALYSIS, v. 95, p. 15-pg., . (23/01728-0)

Please report errors in scientific publications list using this form.
X

Report errors in this page


Error details: