The United States agricultural industry is one of the sectors hit hardest by recent trade wars. The United States is the world’s second largest exporter of agricultural products selling, on average, more than 20 percent of its output to foreign markets. Nearly 50 percent of some crops, including wheat, cotton and soybeans, are sold abroad. Because growers and food processors sell such a significant portion of their goods elsewhere, they have become easy targets for retaliation with tariffs and other measures in trade disputes with foreign partners such as China, Canada and the European Union.
These retaliatory actions have caused sharp declines in agricultural exports, prompting the federal government to respond with bailouts to farmers. Farmers are set to receive $28 billion in aid packages so far, but growers say these payments do not cover all of their losses. What is the economic cost of a trade war on agriculture?
Dr. Sandro Steinbach, assistant professor in the Department of Agricultural and Resource Economics, is attempting to answer that question. With fellow economist Colin Carter, a distinguished professor of agricultural and resource economics at the University of California, Davis, Steinbach is creating tools to precisely measure the impact these retaliatory trade policies have on the United States’ agricultural exports. The researchers received a $485,000 grant from the United States Department of Agriculture’s National Institute of Food and Agriculture to study the issue.
“Our research project will enhance the understanding of a highly relevant foreign trade issue that is of vital importance for the future of agriculture in the United States,” says Steinbach. “Because trade measures have the potential to disturb foreign trade significantly, they are a risk factor for the satisfaction of food and fiber needs. Trade disputes have the potential to strongly impact the viability of farm operations in the US.”
The agricultural sector differs from other industries, like manufacturing, in unique ways that prompted the researchers to shed light on the question of how trade disputes affect agricultural exports, says Steinbach.
Trade disputes involving agriculture are usually more widespread than those involving manufacturing. The agricultural industry is affected by the storability and longevity of products, crop choices that are made well in advance and standards of sanitation and other regulations for plants and crops. These factors make agricultural producers less able to act quickly to avert the effects of trade disputes compared to manufacturing or many other industries. Steinbach says that while economists have looked at aspects of trade dispute effects on agriculture and have started evaluating different trade policy measures, the research is far from conclusive. Studying trade policy is challenging due to the lack of reliable benchmarks, so the researchers are gathering historical data that can provide answers to the economic costs of trade disputes.
To begin this research, Steinbach is collecting data on all agricultural exports going back to the 1990s. This data will include categorizations that take into account product characteristics, customs information and the nature and timing of retaliatory trade disputes, focusing on monthly changes over the last thirty years. Machine learning techniques are employed to statistically study this data for patterns. The goal is to create comparison units, called counterfactuals, that can be used as benchmarks to evaluate trade policy effects and understand trade flows. Rather than selecting a single product for comparison, this approach allows researchers to study multiple features of each good to create amalgamated products with the most appropriate characteristics.
“We want to find out what would have happened to foreign trade without the retaliatory policies taking effect,” says Steinbach. “We’re looking for similarities in the pre-treatment period, before the effect took place, to create counterfactuals.”
The researchers are studying more than 8,000 commodities, focusing on finding patterns in the period before the product was affected by a retaliatory trade policy. Their investigative window looks at the four years prior to the implementation of the trade measure and then studies the three following years to ensure they capture the full effect of the policy.
“In data science, the more information you can integrate, the better it is,” says Steinbach. “Machine learning approaches let us more easily weigh certain attributes and they help us find underlying characteristics. For example, the algorithms might find that garlic is similar to steel. The volume is different, the value is different, but the movement of the function is similar in the pre-treatment period.”
Steinbach says the research will not only determine how agriculture in the United States has been affected by current trade disputes, but will provide a tool to predict how the agricultural industry could be affected by future policy decisions. The research could also be used to assess aid for farmers.
“Some effects are likely being strongly underestimated,” says Steinbach. “With soybeans, the farmers who haven’t been able to export to China have turned to sell them mostly on the domestic market, where prices are lower. The counterfactual evaluation is about knowing the potential foreign sales you’re missing.”
“Trade policies can have considerable impact on the decision to allocate resources and entail substantial consequences for the well-being of farms and our society as a whole. This research advances the knowledge on consequences of trade disputes for agricultural exporters in the United States. Our research will enhance the understanding of a highly relevant foreign trade issue that is of vital importance for the future of agriculture in the United States.”