수산업 관련 언론기사의 활넙치 수요 영향 분석

Abstract

This study identifies key policy and academic issues in the fisheries industry through media analysis and empirically examines how media sentiment influences demand for olive flounder in the seafood market. Specifically, the research addresses the following questions: (1) What core issues are consistently covered in fisheries-related news? (2) What sentiment characteristics do these issues carry? (3) How do these sentiments affect the demand for olive flounder? (4) How do sentiment-driven changes in demand influence social welfare? To address these questions, the study proceeds in four stages. First, Latent Dirichlet Allocation (LDA) topic modeling is applied to 41,483 news articles containing the keyword “fisheries” from 2012 to 2023, identifying six major topics: enhancement of seafood exports and international competitiveness; revitalization of fishing-village infrastructure; perceptions of damage from Japan’s discharge of radioactive wastewater; fostering of young fishers and smart aquaculture; changes in fishery supply structures; and government support for agricultural and fishery disasters. Second, sentiment analysis using a Long Short-Term Memory (LSTM) model is performed separately on articles related to the key fisheries issues and those concerning radioactive-contaminated seafood. The former captures compound sentiments across diverse economic, environmental, and policy concerns, while the latter focuses on a single negative issue that directly affects consumer trust. Although negative sentiment predominates overall, certain topics (e.g., export competitiveness) show a gradual shift toward a positive tone. Notably, sentiment scores vary by narrative tone, even within the same issue. Third, sentiment scores are converted into monthly time-series variables and incorporated into an Autoregressive Distributed Lag– Error Correction Model (ARDL-ECM) to reflect consumer preferences. Olive flounder consumption serves as a proxy for demand. Results demonstrate that media tone regarding specific issues significantly influences consumer behavior: in the key-issue model, positive article counts have a significant short-term impact with a one- to two-period lag; in the radiation-focused model, sentiment scores exert a significant long-term positive effect at the 10% level, and negative article counts produce a significant short-term negative effect with a two-period lag. Price variables have a significant negative long-run effect in both models, consistent with the law of demand. Pork belly prices, used as a substitute, exert significant positive effects, confirming a substitution relationship. Income is significant only in the radiation-related model, implying that olive flounder is a normal good. Finally, price and sentiment elasticities derived from the ARDL-ECM model are used to construct demand curves and estimate changes in consumer and social surplus resulting from shifts in sentiment. Using supply elasticities from prior studies (0.3460 and 0.6390), two scenarios are simulated. A 5%, 10%, and 15% increase in sentiment scores leads to welfare gains of KRW 6.29 billion, 12.75 billion, and 19.40 billion in Scenario 1, and KRW 6.32 billion, 12.88 billion, and 19.68 billion in Scenario 2, respectively. Although both scenarios assume inelastic supply, differences in elasticity alter the magnitude of welfare gains, underscoring the role of supply structure in shaping sentiment-driven outcomes. Empirical results reveal that sentiment-management strategies must be issue-specific. For broad fisheries issues, expanding positive coverage can boost demand after a one- to two-month lag, suggesting a phased “initiation–stabilization–recovery” supply program. By contrast, for negative single issues such as radiation-related events, rapid fact-checking to curb misinformation and sustained delivery of objective safety information are essential to mitigate short-run shocks and improve long-run sentiment. Welfare simulations show that when supply is more inelastic, price increases dominate quantity changes, channeling additional surplus mainly to producers; when supply is less inelastic, larger quantity increases expand consumer surplus. Thus, consumer-relief measures should be prioritized under highly inelastic supply, whereas producer-support schemes suit more elastic conditions. Given the significant cross-market linkage—olive flounder demand responds positively to pork belly prices—an integrated monitoring system that combines media sentiment indices with substitute-market data is recommended to pre-empt demand flight. Establishing threshold-based early-warning triggers for sentiment indices and negative-news counts, linked to automated supply and promotion measures, would further enhance market stability. Methodologically, this study extends fisheries economics by quantifying qualitative exogenous factors—media sentiment—and embedding LSTM-derived sentiment series within an ARDL-ECM, thereby capturing both short- and long-run effects and estimating sentiment-induced welfare changes. In addition, the dual approach— automatic LDA topic extraction for sector-wide issues alongside keyword-focused single-issue analysis—minimizes researcher bias and enhances analytical breadth, providing data-driven evidence for differentiated policy decisions and offering a scalable framework for detecting latent issues in future research.