Tabel IO Provinsi Bengkulu Tahun 2024

The 2024 Bengkulu Province IO Table data is the result of adjustments and updates conducted using non-survey methods for research purposes. The limited availability of IO Tables makes national-level IO Tables more frequently used than regional IO Tables in both domestic and international research. Compiling IO Tables requires relatively large costs, time, and effort, so the available series of tables typically contain time reference conditions for observations that have long since elapsed since the research date. The non-survey updating method used is the RAS method. Although widely associated with updating input-output tables, the RAS method has deep roots in work in statistics and mathematics, with the key concept being to find a matrix that deviates least from the initial matrix while satisfying marginal constraints. The RAS method is a special case of the matrix adjustment problem known as double proportional balancing. Richard Stone and J. A. C. Brown of the University of Cambridge were the first to formally adapt and apply the double proportional iterative technique to update input-output coefficient matrices in an economic context in 1962. The goal was to adjust the coefficients of the transaction matrix of the input-output table from a base year to a target year when only marginal data (total output and total inputs by sector) for the target year were available. The RAS method is used to update the 2016 Bengkulu Province IO Table with 2024 data. The initial step of the RAS method in this study is to create a hypothetical 2024 IO Table based on the 2016 structure (baseline IO), projected 2024 sectoral output data, and the distribution of 2024 GRDP expenditure (final demand). These data are then balanced using the RAS method, as the initial results are unbalanced. After the 2024 final output and demand projections are distributed following the 2016 structure, an imbalance between total output and total input is likely to emerge, necessitating balancing. The RAS balancing steps begin with the initial technology matrix (from the 2016 IO table) and determine the initial target output (2024 IO table). Then, final demand and primary input are distributed from the old structure. Initialization is carried out by creating a diagonal matrix R for row adjustments (sectoral output) and a diagonal matrix S for column adjustments (sectoral input). Corrections are performed alternately for row and column adjustments over several iterations. The iteration process continues to be repeated until each element in the R and S vectors approaches 1.0000 and the difference between the total sectoral input and output is < 0.0001.2024年明古鲁省（Bengkulu Province）投入产出表（Input-Output Table, 以下简称IO表）数据，是为研究目的采用非调查法进行调整与更新的成果。由于IO表的可获得性有限，国内外研究中对国家级IO表的使用频率往往高于区域级IO表。编制IO表需要耗费较高的成本、时间与人力，因此现有系列IO表的观测时间基准往往距研究时点已相隔多年。 本次研究采用的非调查更新方法为RAS法。尽管RAS法常与IO表更新相关联，但其根源可追溯至统计学与数学领域的研究，核心思想是在满足边际约束的前提下，找到与初始矩阵偏差最小的矩阵。RAS法是被称为双比例平衡的矩阵调整问题的特例。剑桥大学的理查德·斯通（Richard Stone）与J. A. C. 布朗（J. A. C. Brown）于1962年首次将双比例迭代技术正式适配并应用于经济场景下的投入产出系数矩阵更新，其目标是在仅获取目标年份分部门总产出与总投入等边际数据的前提下，将基年IO表的交易矩阵系数调整至目标年份。本次研究使用RAS法，以2024年数据为基准，更新2016年明古鲁省IO表。 本研究中RAS法的初始步骤为：以2016年IO表的结构为基准，结合2024年的分部门产出预测数据以及2024年地区生产总值（Gross Regional Domestic Product, 以下简称GRDP）支出分布（即最终需求），构建假想的2024年IO表。由于初始构建结果存在失衡，后续需通过RAS法完成平衡调整。当以2016年的结构为基准对2024年的总产出与总需求预测进行分配后，总产出与总投入间很可能出现失衡，因此需要进行平衡处理。 RAS法的平衡流程以初始技术矩阵（源自2016年IO表）为基础，并确定初始目标产出（即2024年IO表）。随后，基于原有结构对最终需求与初始投入进行分配。初始化阶段需构建用于行调整（分部门产出）的对角矩阵R，以及用于列调整（分部门投入）的对角矩阵S。随后交替对行、列进行修正，并开展多轮迭代。迭代过程将持续进行，直至R与S向量中的各元素趋近于1.0000，且分部门总投入与总产出的差值小于0.0001。