Abstract
This paper evaluates agricultural trade creation and diversion effects of the most important free trade agreements (FTAs). Trade creation and diversion effects are estimated using a Poisson Pseudo-Maximum-Likelihood (PPML) estimator with various fixed effects to deal with heteroskedasticity and zero trade observations. The analysis finds that PPML estimation is preferred to OLS and the estimated impacts of FTAs are different if zero trade observations are considered. The ASEAN-China preferential trade agreement, EU-15, EU-25, and Southern African Development Community agreements have generated large increases in agricultural trade among their members.
Generated Summary
This paper investigates the impacts of Free Trade Agreements (FTAs) on agricultural trade creation and diversion. The study employs a Poisson Pseudo-Maximum-Likelihood (PPML) estimator, addressing heteroskedasticity and zero trade observations, with various fixed effects to account for different factors. The analysis focuses on the ASEAN-China preferential trade agreement, EU-15, EU-25, and Southern African Development Community agreements. The research explores the dynamic patterns of agricultural trade creation and diversion over time, using a three-way fixed effects gravity model to mitigate endogeneity issues. The findings are compared with those obtained from Ordinary Least Squares (OLS) estimations. The study’s scope includes an examination of how FTA impacts vary over time, recognizing that most FTAs have transition periods and that trade effects may not be immediately apparent. The study’s methodology involves the analysis of the effects of FTAs for agricultural trade creation and trade diversion, which is done through the use of the PPML estimator, that avoids issues that can arise from the logarithmic specification in the presence of heteroskedasticity.
Key Findings & Statistics
- The number of zero trade observations is 12,343, which is nearly 32% of the sample.
- The membership in FTAs generally increases agricultural and food trade. For instance, agricultural trade among NAFTA members increased by 145% in 1995–2004.
- The results support food trade diversion for members of the Caribbean Community and Common Market, the Central American Common Market, the Andean Community, and the Common Market for Eastern and Southern Africa (COMESA).
- The coefficient for exporter GDP is 0.60, 0.24, and 0.10 for time but no country fixed effect, time and importer and exporter country fixed effect, and time, bilateral country pair fixed effect respectively.
- The coefficient for importer GDP is 0.88, 0.52, and 0.67 for time but no country fixed effect, time and importer and exporter country fixed effect, and time, bilateral country pair fixed effect respectively.
- The coefficients for the ASEAN-China PTAs are: 0.98, 0.40, and 0.35 for time but no country fixed effect, time and importer and exporter country fixed effect, and time, bilateral country pair fixed effect respectively.
- The coefficients for the EU-15 are: -0.26, -0.08, and -0.07 for time but no country fixed effect, time and importer and exporter country fixed effect, and time, bilateral country pair fixed effect respectively.
- The coefficients for the EU-25 are: -0.80, 0.13, and 0.45 for time but no country fixed effect, time and importer and exporter country fixed effect, and time, bilateral country pair fixed effect respectively.
- The coefficients for the NAFTA are: -0.06, 0.09, and 0.07 for time but no country fixed effect, time and importer and exporter country fixed effect, and time, bilateral country pair fixed effect respectively.
- The coefficients for the SADC are: 1.77, 1.58, and 0.98 for time but no country fixed effect, time and importer and exporter country fixed effect, and time, bilateral country pair fixed effect respectively.
- The number of observations is 38,880, with 7,482 bilateral observations
- The analysis shows that the import creation for SADC comes only in 2002 and 2007.
- There was significant export diversion in 2005 and 2007, and significant import diversion in 1996, 1999, and 2002 for COMESA.
- The EU-15 has significant import diversion for only two of the early years, 1996 and 2002, and there is no significant export diversion.
- The analysis shows no trade creation from NAFTA for any year, yet there is significant export diversion from 1999 to 2007.
- There is significant import diversion for NAFTA in 1996 and 1999.
- The SADC appears to be the model FTA in that there is significant trade creation in 2002 and 2005, and significant import creation in 2002 and 2007.
Other Important Findings
- The PPML estimation is preferred to OLS, and the estimated impacts of FTAs differ when zero trade observations are considered.
- The effects of FTAs are sensitive to the specification of fixed effects and vary over time.
- The ASEAN-China preferential trade agreement, EU-15, EU-25, and Southern African Development Community agreements have generated large increases in agricultural trade among their members.
- The study finds clear trade creation dynamic effects from the EU-15 and SADC.
- There is export creation for COMESA and NAFTA in 1996, but this disappears and even becomes export diversion in later years.
- These same agreements show significant import diversion in early years, which disappears in later years.
- The results show that FTA effects vary over time.
- The paper finds limited evidence that FTAs have led to multilateral lowering of trade barriers for agricultural products.
- The results would attribute the increases in trade between the United States and Mexico and the United States and Canada to GDP growth and the nature of the bilateral relationships that these countries had over the observation period. NAFTA evidently did not enhance these relationships because they were already close.
Limitations Noted in the Document
- The study acknowledges that the impact of FTAs may not be immediately apparent, as most FTAs have a transition period.
- The study does not include agricultural trade creation and trade diversion impacts and does not include observations with no trade.
- The analysis is limited to agricultural trade and does not consider the broader economic impacts.
- The study does not address the complex nature of food regulations, which may contribute to trade diversion, as seen in the EU.
- The results for NAFTA are not consistent with prior research.
- The model with time-varying fixed effects was not estimated due to the inability to identify some parameters of interest.
Conclusion
The study underscores the significant impacts of free trade agreements (FTAs) on agricultural trade. Key findings highlight the success of agreements like the ASEAN-China PTAs, EU-15, EU-25, and SADC in boosting agricultural trade among member countries. The analysis reveals complexities, such as the EU-15’s trade diversion and the dynamic nature of FTA effects, which can vary over time. The study’s use of the PPML estimator, which avoids problems from using the logarithmic specification in the presence of heteroskedasticity and allows zero trade observations in the analysis, is particularly noteworthy. The research also finds that the impact of FTAs is sensitive to the choice of fixed effects and the need to account for the time dimension. The study suggests that while FTAs can create trade, they can also lead to diversion, especially when agricultural tariffs remain high for non-member countries. Therefore, efforts to lower agricultural tariffs within the World Trade Organization (WTO) remain important to maximize the benefits of global free trade. The EU-15 has significant import diversion for only two of the early years, 1996 and 2002, and there is no significant export diversion. The findings also highlight the dynamic effects of FTAs, with impacts changing over time, and the importance of considering these changes in future analyses. It emphasizes that despite trade creation, there may also be trade diversion, depending on the agreement and the time frame considered. The study reinforces the idea that the benefits of trade agreements are complex and can vary significantly depending on the specific circumstances.