|
|
 |
| |
|
|
|
Statistical Analysis of the BitCoin and South African Rand Exchange Rates Risks when the Tails are Somewhat Heavy |
|
|
|
PP: 1411-1429 |
|
|
|
doi:10.18576/jsap/130501
|
|
|
|
Author(s) |
|
|
|
Delson Chikobvu,
Thabani Ndlovu,
|
|
|
|
Abstract |
|
|
| In the work of the global financial crisis, in which incorrect and normal distribution-based risk models were
used, this paper employs some selected statistical parent distribution models to better quantify and compare the financial riskiness of BitCoin and the Rand; both measured against the United States Dollar. The positive returns (gains) and negative returns (losses) are fitted separately to selected statistical distributions with varying degrees of tail heaviness, hence providing for various levels of risk. Riskiness is measured using the Value at Risk and Expected Shortfall. Relatively heavy-tailed models like: the exponential, Weibull, and the Burr distributions are considered and are non-normal, a feature that is common in financial, and currency data. The Anderson-Darling test is used to confirm the goodness-of-fit of the distributions. Using the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC), BitCoin gains are confirmed to follow the heavy-tailed Burr distribution, while the losses follow the lighter-tailed exponential distribution. In the case of the Rand, both gains and losses follow the Weibull distribution with tails that are even lighter than the Exponential distribution tail. BitCoin has a greater upside and downside risk hence rendering it riskier than the Rand. The derived information helps investors in BitCoin and the South African Rand understand the comparative risk they are exposed to when they convert their investments from the Rand to invest in BitCoin. Risk managers can use the
approach and the results to compute risk-adjusted capital requirements.
|
|
|
|
|
|
|
|
