KRED: Korea Research Economic Database for Macroeconomic Research

2025-09-19

AI Summary

To address the fragmentation, inconsistent formatting, and limited accessibility of South Korea’s macroeconomic data, this paper introduces KRED—the first open-source, FRED-MD–style database for Korea—systematically integrating 88 official monthly time series. Through frequency alignment, standardized transformations, and cross-source variable mapping, KRED constructs a comparable, reproducible balanced panel (N = 80, T ≥ 200). Methodologically, we apply principal component analysis to extract four economically interpretable common factors—capturing monetary policy, labor market conditions, real output, and housing demand—which jointly explain approximately 40% of total variance. Leveraging these factors, we construct an interpretable diffusion index that accurately identifies key cyclical turning points, including the 2020 pandemic-induced recession. KRED fills a critical gap in Korea’s macroeconomic empirical research infrastructure, providing standardized, transparent data to support business-cycle analysis, structural model validation, and cross-country benchmarking.

Technology Category

• Data Mining & Knowledge Management: Linked Open Data, Knowledge Graphs & KB Completion
• Machine Learning: Kernel Methods
• Application Domains: Humanities & Computational Social Science

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

We introduce KRED (Korea Research Economic Database), a new FRED MD style macroeconomic dataset for South Korea. KRED is constructed by aggregating 88 key monthly time series from multiple official sources (e.g., Bank of Korea ECOS, Statistics Korea KOSIS) into a unified, publicly available database. The dataset is aligned with the FRED MD format, enabling standardized transformations and direct comparability; an Appendix maps each Korean series to its FRED MD counterpart. Using a balanced panel of 80 series from 2009 to 2024, we extract four principal components via PCA that explain approximately 40% of the total variance. These four factors have intuitive economic interpretations, capturing monetary conditions, labor market activity, real output, and housing demand, analogous to diffusion indexes summarizing broad economic movements. Notably, the factor based diffusion indexes derived from KRED clearly trace major macroeconomic fluctuations over the sample period such as the 2020 COVID 19 recession. Our results demonstrate that KRED's factor structure can effectively condense complex economic information into a few informative indexes, yielding new insights into South Korea's business cycles and co movements.

Problem

• Constructs a unified macroeconomic database for South Korea
• Extracts principal components to explain economic variance
• Creates diffusion indexes tracing macroeconomic fluctuations

Innovation

• Aggregated 88 monthly time series
• Aligned with FRED MD format
• Extracted four principal components via PCA