Economic narrative indices are predominantly English-centric; 84% of sentiment-based forecasting research focuses on developed economies. We present BENI Global 10, the first multilingual economic news corpus spanning 10 languages across 7 language families and 5 economic regions: Bangla (Bangladesh), Hindi (India), Turkish (Turkey), Indonesian (Indonesia), Portuguese (Brazil), Arabic (Egypt), Vietnamese (Vietnam), Filipino (Philippines), Swahili (Kenya), and Urdu (Pakistan). The corpus contains 522,397 economically relevant articles filtered from 2.8M raw documents using 25-32 translated keywords per language. We provide: (1) a reproducible streaming pipeline with checkpoint-resume for low-resource environments, (2) per-language schema-normalized Parquet files with economic relevance labels, (3) a temporally synced cross-lingual index covering 2018-2024, and (4) comparative analysis revealing systematic differences in how economic narratives are framed across Global South regions. Inter-annotator agreement reaches kappa > 0.70 across all languages. The complete dataset, code, and annotation guidelines are publicly released for research use.

翻译：经济叙事指数主要集中于英语，84%基于情感预测的研究聚焦于发达经济体。我们提出BENI Global 10，这是首个覆盖7个语系、5个经济区域中10种语言的多语言经济新闻语料库：孟加拉语（孟加拉国）、印地语（印度）、土耳其语（土耳其）、印度尼西亚语（印度尼西亚）、葡萄牙语（巴西）、阿拉伯语（埃及）、越南语（越南）、菲律宾语（菲律宾）、斯瓦希里语（肯尼亚）和乌尔都语（巴基斯坦）。该语料库从280万份原始文档中通过每语言25-32个翻译关键词筛选出522,397篇经济相关文章。我们提供：（1）面向低资源环境的带断点续传的可复制流式处理管线；（2）含经济相关性标签的每语言模式归一化Parquet文件；（3）覆盖2018-2024年的时间同步跨语言索引；（4）揭示南半球地区经济叙事框架系统性差异的对比分析。所有语言的标注者间一致性均达到kappa>0.70。完整数据集、代码及标注指南已公开供研究使用。