Hedging Cryptos with Bitcoins Futures

Name of Quantlet: 'Data and Codes of the Paper Hedging Cryptos with Bitcoins Futures'

Published in: 'Hedging Cryptos With Bitcoin Futures'

Description: 'This repository archives the data and codes that produce the results of paper Hedging Cryptos with Bitcoin Futures.'

Keywords: 'Optimal hedge ratio, risk measures, BTC future, BTC, futures contract'

Author: 'Anonymous for double blind peer reviewed'

Submitted:   '6 Jan 2022'

Generate Results Files

To generate the results files, one needs to run the following python scripts with the same configuration file as input

1. src/calibration.py: calibrates copula using the training data stated in the configuration file,
2. src/get_OHR.py: numeraically search for the optimal hedge ratios according the exponential risk measure, expected shortfall, value-at-risk with parameters stated in the configuration file,
3. src/get_HEs.py: generate results from testing data and calculate hedging effectiveness.

The python scripts should be run in the exact order above to generate correct results. The procedure can generate results of all the copulae except the NIG factor copula. Codes in Mathematica for NIG factor copula is avaliable upon request.

One can generate results with other datasets, following are the steps to do so prepare the data according to the format and file structure in processed_data prepare a configuration json file Run the three python scripts above using the configuration file in the second step, e.g.

1. calibration.py --config configs/BBT_future_Tiingo_ltc_MM.json
2. get_OHR.py --config configs/BBT_future_Tiingo_ltc_MM.json
3. get_HEs.py --config configs/BBT_future_Tiingo_ltc_MM.json

Configuration file is a json file with fields describe as follows

• data_name: the name of the dataset in the processed_data folder, e.g. BBT_future_BITW20
• spot_name: column name of the spot’s log return in each data file, e.g. log return BITW20
• future_name: column name of the futures’ log return in each data file, e.g. log return future
• calibration_method: MM stands for method of moments; MLE stands for maximum likelihood
• q_arr: the quantile level used by MM. This field will not affect MLE, but it cannot be empty.
• k_ERM: Risk aversiveness of exponent risk measure being used as loss function while searching for the optimal hedge ratio
• q_arr_ES: quantile level of expected shortfall being used as loss function
• q_arr_VaR: quantile level of value-at-risk being used as loss function
• h_Clip: range of optimal hedge ratio

Generate Plots in the Paper

The figures and tables in the paper are generated from the Jupyter Notebook files in Pynotebooks/figures and Pynotebooks/tables respectively.