Department of Statistics, Jahangirnagar University, Savar, Dhaka, Bangladesh.
Department of Statistics, Mawlana Bhashani Science and Technology University, Santosh, Tangail, Bangladesh.
PLoS One. 2023 Apr 7;18(4):e0284179. doi: 10.1371/journal.pone.0284179. eCollection 2023.
The reserve of a country is a reflection of the strength of fulfilling its financial liabilities. However, during the past several years, a regular variation of the total reserve has been observed on a global scale. The reserve of Bangladesh is also influenced by several economic and financial indicators such as total debt, net foreign assets, net domestic credit, inflation GDP deflator, net exports (% of GDP), and imports of goods and services (% of GDP), as well as foreign direct investment, GNI growth, official exchange rate, personal remittances, and so on. Therefore, the authors aimed to identify the nature of the relationship and influence of economic indicators on the total reserve of Bangladesh using a suitable statistical model.
To meet the objective of this study, the secondary data set was extracted from the World Bank's website which is openly accessible over the period 1976 to 2020. Moreover, the model used the appropriate splines to describe the non-linearity. The performance of the model was evaluated by the Akaike information criterion (AIC), Bayesian information criterion (BIC), and adjusted R-square.
The total reserve of Bangladesh gradually increased since 2001, and it reached its peak in 2020 which was 43172 billion US dollars. The data were first utilized to build a multiple linear regression model as a base model, but it was later found that the model has severe multicollinearity problems, with a maximum value of VIF for GNI of 499.63. Findings revealed that total debt, inflation, import, and export are showing a non-linear relationship with the total reserve in Bangladesh. Therefore, the authors applied the Generalized Additive Model (GAM) model to take advantage of the nonlinear relationship between the reserve and the selected covariates. The overall response, which is linearly tied to the net foreign asset in the GAM model, will change by 14.43 USD for every unit change in the net foreign asset. It is observed that the GAM model performs better than the multiple linear regression.
A non-linear relationship is observed between the total reserve and different economic indicators of Bangladesh. The authors believed that this study will be beneficial to the government, monetary authorities also to the people of the country to better understand the economy.
一个国家的储备是其履行财政债务能力的反映。然而，在过去的几年中，全球范围内的总储备量呈现出有规律的变化。孟加拉国的储备也受到总债务、净外国资产、国内信贷净额、通胀 GDP 平减指数、净出口额（占 GDP 的百分比）、货物和服务进口额（占 GDP 的百分比）以及外国直接投资、国民总收入增长、官方汇率、个人汇款等多个经济和金融指标的影响。因此，作者旨在通过合适的统计模型，确定经济指标对孟加拉国总储备的关系和影响的性质。
为了实现本研究的目标，从世界银行的网站上提取了可公开获取的 1976 年至 2020 年期间的二手数据集。此外，该模型使用适当的样条函数来描述非线性。通过赤池信息量准则（AIC）、贝叶斯信息量准则（BIC）和调整后的 R 平方来评估模型的性能。
自 2001 年以来，孟加拉国的总储备逐渐增加，2020 年达到 43172 亿美元的峰值。最初，将数据用于构建多元线性回归模型作为基础模型，但后来发现该模型存在严重的多重共线性问题，GNI 的最大 VIF 值为 499.63。研究结果表明，总债务、通胀、进口和出口与孟加拉国的总储备呈非线性关系。因此，作者应用广义加性模型（GAM）模型来利用储备与选定协变量之间的非线性关系。在 GAM 模型中，与净外国资产线性相关的整体响应，将随净外国资产每单位变化而变化 14.43 美元。观察到 GAM 模型的表现优于多元线性回归。
观察到孟加拉国总储备与不同经济指标之间存在非线性关系。作者认为，这项研究将使政府、货币当局和国家人民受益，以更好地了解经济。
Zotero 插件
