Skip to content

This repository includes the scripts to replicate the results of my paper entitled "A False Discovery Rate Approach to Optimal Volatility Forecasting Model Selection".

Notifications You must be signed in to change notification settings

hkalager/FDR_buckets

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

39 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MHT_buckets

The codes and files in this repository replicate the results in a manuscript entitled "Profiling Volatility Forecasting Models with Multiple Hypothesis Testing" You can access a free copy of the manuscript through SSRN at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3737477 For any enquiries please get in touch via hassannia@outlook.com

Reproduction:

In order to replicate the results you need to run the m files in MATLAB. The steps are identified in file names as step0_XXX.m, step1_XXX.m et cetra. Each file is accompanied by the necessary guidance within the script.

Dataset:

The datasets provided under folder "Dataset_ETFS" are samples from TAQ at WRDS with no column headings. The provided csv files are for evaulation purposes only and do not bear any official information. It is at the viewer's discretion to interpret the columns. We are not in the capacity to publish the data publicly. You can obtain the data from https://wrds-web.wharton.upenn.edu/wrds/

Once you obtain the data from WRDS you need to prepare a csv file named "TICKER_M5_processed.csv" with the columns below:

  • Date: date as in format DD/MM/YYYY
  • Open: opening price recorded for each day as the first trade after 09:00 ET
  • Close: closing price recorded for each day as the first trade after 16:30 ET
  • OC Return: daily range return calculted as log(Close/Open)
  • OC Return Sq: squared daily range return
  • RV Daily: realized volatility proxied by sum of squared 5-min returns.
  • RQ Daily: realized Quarticity as in Bollerslev, Patton, and Quaedvlieg (2016) https://doi.org/10.1016/j.jeconom.2015.10.007
  • FSI Vol: Office of Finacial Research Financial Stress Index for volatility from https://www.financialresearch.gov/financial-stress-index/

Third-party scripts:

I) The majority of the codes for the GARCH and SV models are from Kevin Sheppard's MATLAB repository at https://www.kevinsheppard.com/code/matlab/mfe-toolbox/ All script in folders distributions, univariate, multivariate, utility, and timeseries are from Kevin Sheppard along with the scripts bsds.m and stationary_bootstrap.m; these codes are provided as is subject to change at all time without previous notice.

II) The scripts performing the Romano, Wolf, and Sheikh (2008)'s k-FWE test is from Michael Wolf's repository with the University of Zurich at https://www.econ.uzh.ch/en/people/faculty/wolf/publications.html#Programming_Code All contents under the folder kfwe is from this source ; these codes are provided as is subject to change at all time without previous notice.

III) Parts of the FDR+/- codes are from the accompanying codes for Bajgrowicz and Scaillet (2012). https://doi.org/10.1016/j.jfineco.2012.06.001

III) The script est_pi0_disc.m is a function estimating the tuning parameter lambda for the estimating pi0 as in Liang (2016). https://doi.org/10.1111/biom.12429

About

This repository includes the scripts to replicate the results of my paper entitled "A False Discovery Rate Approach to Optimal Volatility Forecasting Model Selection".

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages