Kasra Pourkermani,
Volume 16, Issue 0 (Summer and Fall 2021 2021)
Abstract
Baltic Dry Index (BDIY:IND) is daily reported by Baltic Exchange. The index is a benchmark for the prices of ship chartering contracts which is a proxy for the maritime economy however the calendar anomalies of BDIY:IND have not yet been researched. This article investigates the day of week effects on BDIY:IND returns from 2014-03 to 2020-03. In this study, GARCH models were used to investigate the calendar effect on stock returns, and the Bootstrapping GARCH Regression is used to obtain the results with higher reliability. Regarding the correlation of time-based observations, the standard Bootstrap method does not apply to time series data; thus, the Bootstrap procedure based on resampling of GARCH's regression model residues is used in the present study. Based on a bootstrapping asymmetric GJR-GARCH approach, the results indicate that the Monday returns are significantly positive, which is in contrast with the usual findings in stock markets. It means the parties involved in shipping markets can still use information analysis as means to obtain further returns. The monetary figures of ship chartering contracts involve quite a large sum of money depends on movement of Baltic Dry Index hence having a knowledge of its behavior is vital for making smarter decisions for investors, shipowners and shipbrokers.
Kasra Pourkermani,
Volume 18, Issue 0 (5-2023)
Abstract
Purpose – Baltic Dry Index (BDI) is shipping freight-cost index which is reported daily by Baltic Exchange. The index is a benchmark for the prices of ship chartering contracts which is a proxy for the maritime economy, BDI is heavily used by financial traders to predict the world economy, the volatility forecast has an important implication for all the investors and hence in this paper the daily forecast performance of different models is evaluated.
Research methodology – The daily forecast performance of conditional and unconditional volatility of 12 long memory GARCH-type models based on the root-mean-square error (RMSE) is evaluated. Because all return series were skewed and fat-tailed, each conditional volatility model was estimated under a skewed Student distribution.
Findings – According to the idea that the accuracy of Value-at-Risk (VaR) estimates was sensitive to the adequacy of the volatility model used, the result showed that the 250-day moving average models, exponential smoothing, and (component GARCH) CGARCH function better than other models based on RMSE standard. The results of hybrid models such as Dibold-Mariano statistics showed that there was no significant difference between the predictive power of 250 days moving average (MA250) and CGARCH.
Practical implications – BDI was widely regarded as a benchmark for the world economy by traders and hedge fund managers.
Originality/Value – we examine the science of volatility prediction in BDI which has not been performed before.