Price transmission research is crucial for understanding market efficiency and resource allocation. While extensive literature exists on agricultural price transmission, particularly within domestic markets, cross-border linkages, especially in the international beef market, remain understudied. This gap is significant given disruptions to meat value chains caused by events like BSE and COVID-19. Traditional error correction models (ECMs) are insufficient for capturing time-variant price linkages. This study uses a GARCH-DCC model to analyze the dynamic correlations between international and regional beef prices across nine net-importing countries. The authors also incorporate structural break dummy variables into their DCC model, a methodological innovation, to enhance the reliability of their estimations. This approach allows for a more accurate depiction of market connectivity intensity over time, including the impact of major economic events.

The existing literature on agricultural price transmission primarily focuses on domestic markets and utilizes error correction models (ECMs), which fail to capture time-varying linkages. The DCC approach, while frequently employed in financial markets, has seen limited use in agricultural price transmission research, especially in the beef sector. Although some studies analyze price co-movements in the cereal sector using DCC, research on international beef market linkages remains scarce. Previous studies on beef price co-movement have considered structural breaks, but none have incorporated them into time-varying correlation analyses using DCC. This study addresses this gap by employing GARCH-DCC models with structural break dummy variables to analyze the relationships between international and regional beef prices across nine importing countries.

The study focuses on nine net beef-importing countries (Azerbaijan, Georgia, Japan, Kazakhstan, Kyrgyzstan, Tajikistan, Tunisia, UK, and USA). Monthly international and local retail beef prices (in USD) from January 2006 to May 2020 were collected from various sources. The X-13-ARIMA method was used to eliminate seasonal fluctuations, and monthly returns were calculated as the first differences of logarithmic prices. Stationarity was tested using ADF and KPSS unit root tests, along with Zivot-Andrews and Perron tests to account for structural breaks. To examine Granger causality, the authors applied Hong's (2001) non-uniform weighting CCF to the standardized residuals and squared standardized residuals from three types of GARCH models (GARCH, EGARCH, and GJR-GARCH). The best model was selected based on the Bayesian Information Criterion (BIC). Bai and Perron's (2003) test was used to identify structural breaks in the mean and variance equations. A bivariate GARCH model with DCC was then used to estimate the time-varying conditional correlations between international and domestic beef prices. Four DCC specifications (DCC, A-DCC, G-DCC, and AG-DCC) were considered, with the best model selected based on the BIC. Finally, an AR(1) model with dummy variables was used to analyze the impact of structural breaks on the dynamic correlations.

The study's key findings are threefold: 1. **Structural Changes:** Local retail beef prices in Azerbaijan, Georgia, Japan, Kazakhstan, Kyrgyzstan, Tajikistan, the UK, and the USA showed structural changes in their mean or variance, primarily occurring after the 2007-2009 global food crisis. This suggests a delayed impact of global events on domestic markets. 2. **Granger Causality:** International prices unidirectionally Granger-caused regional prices (both in mean and volatility) in Georgia, Tajikistan, and the USA. No country exhibited price or price-volatility transmission from regional to international markets. This indicates a stronger influence of global prices on these three specific countries. 3. **Dynamic Conditional Correlations (DCCs):** Volatility linkages between global and local beef markets were generally weak. However, price volatility exhibited closer synchronization around the 2008 global food crisis, reflecting the increased market interconnectedness during periods of significant global economic turmoil. The DCC estimates range from -0.099 (JPN) to 0.210 (KYR). The results of Bai and Perron's structural breaks test indicate the presence of breaks for most countries, especially during the 2007-2008 food and financial crisis. Analysis using an AR(1) model with dummy variables suggests that structural breaks substantially influence the dynamic correlations between international and domestic beef prices.

The findings highlight the complex relationship between global and local beef markets. While generally weak, the influence of international prices is evident in the unidirectional Granger causality observed in several countries, notably Georgia, Tajikistan, and the USA. The impact of the 2007-2009 global food crisis was significant, as evidenced by increased volatility synchronization and structural breaks in the DCCs. The results suggest that while self-sufficiency may not be directly correlated with market connectivity, the responsiveness of local markets to international shocks varies significantly across countries, and the weak linkage in the beef sector might be attributed to the flexible adjustment of beef supply. The absence of reverse causality suggests that domestic markets do not significantly influence international beef prices.

This study offers valuable insights into international beef price volatility transmission and its dynamics. The application of GARCH-DCC models with structural break dummy variables provides a more nuanced understanding of market linkages. The results suggest that while global prices significantly influence domestic markets in some cases, these connections are generally weak, especially outside of periods of significant global crises. Future research could explore the role of other factors (e.g., policy interventions, consumer behavior, and specific supply chain characteristics) to explain the variations in price volatility transmission across countries.

The study's limitations include the focus on a specific set of nine beef-importing countries, which might not be fully representative of global dynamics. The use of retail prices rather than wholesale or farm-gate prices could potentially influence the results, and the model relies on certain assumptions (e.g., the normality of residuals) that may not fully capture the complexity of real-world market conditions. Further investigations could use alternative econometric techniques or broaden the scope of countries considered for a more comprehensive understanding.