BlockChainPublic

This is a replication package for "Security-Cost Efficiency of Competing Proof-of-Work Cryptcurrencies".

Install

1. Install the latest version of R, RStudio, git, and aws cli. Install C/C++ compilers. In Windows, you can install Rtools. You also need to install python by anaconda. This project uses reticulate::use_condaenv to call python functions from R for running some process in the simulation.

2. Clone this git repository to your local environment by running the following in the terminal:

   git clone git@github.com:kohei-kawaguchi/BlockChainPublic.git
   

3. As you clone the git repository, double click the BlockChainpublic.Rproj to open it as a project in RStudio. Every code is written relative to this project root folder. By starting it as an RStudio project, you can always start with this projet root folder as the current working directory.

4. Test whether you can clean and rebuild the project as an R package by clicking the clean and rebuild command from the build tab.

1. If it succeeds, the build console will show the following message. You may be required to install some R packages, such as RcppEigen and Rcpp.

1. You can load the library by library(BlockChainPublic) to call R functions defined in R/ folder. C++ functions defined in src/ folder is exported to R functions by Rcpp.

2. Download cleaned/ folder and output/ folder from aws s3 bucket by running the following command in the terminal:

   aws s3 cp s3://blockchain-kawaguchi-gkh3xcs8ag69utjzohufkwwq3f5uyapn1b-s3alias/cleaned cleaned --recursive
   aws s3 cp s3://blockchain-kawaguchi-gkh3xcs8ag69utjzohufkwwq3f5uyapn1b-s3alias/output output --recursive
   

Folder Structure

The folder structure is

• cleaned: Store cleaned data.
• output: Store transformed data nd other outputs of the analysis.
• main: Store code for cleaning, transforming, and analyzing the data.
• report: Store reporting Rmd documents.
• R: Store R function definitions.
• src: Store C++ function definitions.
• module: Store Python function definitions.

The files DESCRIPTION and NAMESPCE describe the meta data of this package. It does not contain the raw data, because the raw data is too large. All analysis can be replicated based on the cleaned data. image folder stores image files for this readme document.

Code Structure

Main Files

• The main folder contains execution files. The missing numberings are for scraping, data cleaning, and unpublished analysis. Because we do not include the raw data due to its file size, the current code is self-contained by themselves.

• The file main/4_make_data_blockheight.R transforms the cleaned data for the following analysis. It combines the blockchain, reward, and exchange rate to make a currency-algo-blockheight-level data. Then, it combines the currency-algo-blockheight-level data of SHA-256 currencies to make a epoch-currency-level data and save it as output/epoch_currency_sha256.rds.

• The file main/5_estimate_arrival_rate_sha256.R estimates the hash supply function using the epoch-currency-level data output/epoch_currency_sha256.rds. The file runs various specifications that are not used in the paper. The main specification we use is saved as output/estimate_arrival_rate_after_bsv_local.rds.

• The file main/8_estimate_exogenous.R estimates the exchange rate process and saves the result as output/rate_estimate.rds.

• The file main/9_1_simulate_reduced_btc_halving.R uses the data output/epoch_currency_sha256.rds, the estimate of the exchange rate process output/rate_estimate.rds, and the estimate of the hash supply function output/estimate_arrival_rate_after_bsv_local.rds to simulate the minimg market after the third BTC halving.

• One can change the DAA setting by changing the variable setting. The choice is actual, original_original_cw144, original_original_original, cw144_cw144_cw_144, original_asert_cw144, original_asert_asert, asert_asert_asert, cw144_asert_asert, asert_asert_cw144, and cw144_original_original. Except for actual, they represent the DAAs of BTC, BCH, and BSV.

• The code is supposed to run batch from the command line. As one runs the code in the terminal as follows, then the $SETTING is passed to the file as the argument.

  Rscript main/9_1_simulate_reduced_btc_halving.R $SETTING
  

• The first few lines of main/9_1_simulate_reduced_btc_halving.R read the argument and pass to setting as follows

  args <- commandArgs(trailingOnly = TRUE)
  if (length(args) > 0) {
    setting <- args[1]
  } else {
    setting <- "cw144_original_original"
  }
  

• The files main/9_2_simulate_reduced_bch_halving.R and main/9_3_simulate_reduced_bsv_halving.R are for the simulatino after the third BCH and BSV halving. The files main/10_1_simulate_reduced_btc_halving_96paths.R, main/10_2_simulate_reduced_bch_halving_96paths.R, and main/10_3_simulate_reduced_bsv_halving_96paths.R run the same simulations for 96 paths.

• The simulation files require the anaconda distribution of python 3.9. At the beginning of the code, it calls the anaconda distribution of python using recitulate as follows. If the if clause is true, it setup an anaconda environment named "blockchain". If the environment is already defined, it calls it in the else clause.

  if (!("blockchain" %in% reticulate::conda_list()$name)) {
    # set up python environment and install pyblp
    reticulate::conda_create(
      envname = "blockchain",
      forge = TRUE,
      python_version = "3.9"
    )
  } else {
    # load conda environment and load pyublp
    reticulate::use_condaenv("blockchain", required = TRUE)
    asert <- 
      reticulate::import_from_path(
        module = "asert",
        path = "module"
      )
  }
  

Reporting Files

• report/summarize_cleaned_data.Rmd is a code book of the cleaned data. It describes the definition of each variable of each cleaned data set.

• report/analyze_sha256.Rmd compares the time series of SHA-256 currencies' data.

• report/describe_exchange.Rmd examines the distribution of the exchange rate.

• report/describe_simulator.Rmd checks whether the our simulator can replicate the data, especially for the difficulty adjustment.

• report/summarize_arrival_rate_rdd.Rmd reports the SHA-256 around the halving and run a regression discontinuity analysis, and report/arrival_rate.Rmd reports the estimation result of the hash supply function.

• report/summarize_simulate_reduced_halving.Rmd reports the simulation results.

• report/summarize_security_cost.Rmd summarizes the simulation results from the security-cost perspective.

Table and Figures

• Table 1 and 2 are generated in report/summarize_cleaned_data.Rmd.
• Figure 2 is generated in report/analyze_sha256.Rmd.
• Figures 3, 4, A1, A2, A3, and A4, and Table 3, A1, and A2, are generated in report/summarize_arrival_rate_rdd.Rmd.
• Table 4 is generated in report/summarize_arrival_rate.Rmd.
• Table 5 is generated in report/describe_exchange.Rmd.
• Figure 5, 6, and 7 are generated in report/summarize_simulate_reduced_halving.Rmd.
• Table 9, 10, A4, A5, and A6, are generated in report/summarize_security_cost.Rmd.