國際貨物運量預測及受災害影響之蔓延現象

Abstract

This study aims to develop an approach for the forecasting of cargo volumes which changes rapidly due to interactions and competitions between industries and/or countries. Two models are proposed and demonstrated with case studies. In the first part, we propose a hybrid forecasting model to capture industry shares of cargo import and export between countries changing over time and quickly responsive to the dynamic changes in the industry. The proposed Grey hybrid model is based on Grey forecasting model and Grey residual modifications with Markov-chain sign estimations, together with an industry share transformation technique. The model is used to predict cargo volumes between Taiwan and North American by different industries. The second part of this study explores the contagion effect of cargo volumes between major trading partners after the occurrence of major disasters. It investigates the spatial contagion on trading markets when an industrialized country is stricken by a natural disaster. A methodology is proposed for contagion detection on the cargo trading among the trading partners. The test is based on a comparison of the correlation coefficients of transnational trade linkages in industrial cargo volumes before and after a disaster. Two case studies are performed to demonstrate the performances and applicability of the two proposed models. The first case study investigates the cargo export and import by industry between Taiwan and North American, and the modeling results indicate that the industry share forecasting can smooth out variations in historical data and assist in identifying industrial cargo volume trends such that the proposed hybrid model can rapidly respond to the dynamic changes in the industry. The proposed hybrid model outperforms all other models compared shows good forecasting results and outperforms several other forecasting models. The forecasting results show that, in the near future, the majority of export cargo volumes from Taiwan to North America are non-metallic mineral and machinery industries and the majority of imports from North America to Taiwan are agriculture and metal industries. The findings can be useful for marine carriers in response to the variation in future industrial cargo trends; and further, they can make timely adjustments in scheduling to lessen the impact on operation planning. In the second case study, the contagion of international trading after the Japanese 311 earthquake in 2011 is investigated. The results show that spatial contagion is observed for neighboring countries exporting to Taiwan. The occurrence of contagion effect depends on the industrial market share of the linkages, and the contagion effect disperses over time. More specifically, potential trade relationships between transnational linkages reflecting substitution relationships as a negative correlation sign, which indicates the transfer of industrial cargo volumes from the disaster country to another country. Based on findings, it can provide a useful reference for shippers regarding how long the contagion effect is accumulated on transnational trade linkages by industry; and further, they can understand the trade-off between importing cargo from other neighboring countries to meet demands and put up with dynamic changes on industrial cargo flows in trading networks for a short-term period.

This study aims to develop an approach for the forecasting of cargo volumes which changes rapidly due to interactions and competitions between industries and/or countries. Two models are proposed and demonstrated with case studies. In the first part, we propose a hybrid forecasting model to capture industry shares of cargo import and export between countries changing over time and quickly responsive to the dynamic changes in the industry. The proposed Grey hybrid model is based on Grey forecasting model and Grey residual modifications with Markov-chain sign estimations, together with an industry share transformation technique. The model is used to predict cargo volumes between Taiwan and North American by different industries. The second part of this study explores the contagion effect of cargo volumes between major trading partners after the occurrence of major disasters. It investigates the spatial contagion on trading markets when an industrialized country is stricken by a natural disaster. A methodology is proposed for contagion detection on the cargo trading among the trading partners. The test is based on a comparison of the correlation coefficients of transnational trade linkages in industrial cargo volumes before and after a disaster. Two case studies are performed to demonstrate the performances and applicability of the two proposed models. The first case study investigates the cargo export and import by industry between Taiwan and North American, and the modeling results indicate that the industry share forecasting can smooth out variations in historical data and assist in identifying industrial cargo volume trends such that the proposed hybrid model can rapidly respond to the dynamic changes in the industry. The proposed hybrid model outperforms all other models compared shows good forecasting results and outperforms several other forecasting models. The forecasting results show that, in the near future, the majority of export cargo volumes from Taiwan to North America are non-metallic mineral and machinery industries and the majority of imports from North America to Taiwan are agriculture and metal industries. The findings can be useful for marine carriers in response to the variation in future industrial cargo trends; and further, they can make timely adjustments in scheduling to lessen the impact on operation planning. In the second case study, the contagion of international trading after the Japanese 311 earthquake in 2011 is investigated. The results show that spatial contagion is observed for neighboring countries exporting to Taiwan. The occurrence of contagion effect depends on the industrial market share of the linkages, and the contagion effect disperses over time. More specifically, potential trade relationships between transnational linkages reflecting substitution relationships as a negative correlation sign, which indicates the transfer of industrial cargo volumes from the disaster country to another country. Based on findings, it can provide a useful reference for shippers regarding how long the contagion effect is accumulated on transnational trade linkages by industry; and further, they can understand the trade-off between importing cargo from other neighboring countries to meet demands and put up with dynamic changes on industrial cargo flows in trading networks for a short-term period.