This study investigates the foundational processes and applications of AI language models in the field of economics. It focuses on developing a Korean Small-scale Large Language Model (sLLM) specialized in economics, finance, and fiscal policy, while exploring its practical applications. Using Meta-Llama/ Llama-3.1-8B-Instruct as the base model, the sLLM is developed by applying reliable datasets in economics, finance, and fiscal policy, along with various training methods such as SFT and DPO. Its applicability and versatility are evaluated through its implementation on economic databases. Empirical analysis indicates that the proposed sLLM exhibits enhanced contextual understanding in response to economic queries, with additional improvements observed when integrated with RAG (Retrieval-Augmented Generation). These findings suggest that the model successfully addresses challenges such as translation errors and limited contextual comprehension, laying a foundation for tackling AI sovereignty issues.