What shapes public support for trade?
The factors that shape public opinion on trade are typically hard to pin down. Earlier studies often found that popular attitudes largely reflect economic self-interest and align with the predictions of standard trade models. More recent survey-based research, however, argues that views on trade are shaped primarily by ideological, social, and cultural factors. Understanding what drives public perceptions of trade, particularly in low- and middle-income countries, can help policymakers better communicate and design trade policy. Yet popular attitudes towards economic issues are typically unobservable, making them hard to analyze in concrete settings.
In a paper published in the Review of Economic Studies, EGC affiliate Diana Van Patten and coauthor Esteban Méndez exploit a unique 2007 referendum in Costa Rica on whether the government should ratify the Central America Free Trade Agreement (CAFTA) with the United States. The authors obtained official records from voting boards – typically classrooms where voters vote – and linked them to unique national voter identifiers and employer-employee data, an innovative approach that breaks new ground in the use of voting data while still preserving voter anonymity. Using a model that explains voting outcomes based on voters’ income, spending patterns, and political ideology, Van Patten and Méndez find that economic factors played a salient role in determining the referendum’s outcome.
Economic factors accounted for 7% of the observed variation in voting patterns – whether a particular voting board supported CAFTA – driven by the expected effects on voters’ cost of living and on their incomes through employer firms.
A voter’s employer was more significant in explaining voting outcomes than the sector in which the voter worked.
Voting boards where voters were politically aligned with pro-trade parties were more sensitive to the economic factors associated with the referendum.
Economic fundamentals were nearly as important as political alignment in accounting for variation in voting behavior.
Exploiting a unique Costa Rican referendum
Elected officials often place significant weight on public attitudes towards trade policy, making it important to understand the relationship between economic factors and public sentiment. Analyzing this relationship can also inform research on how the gains from trade are distributed across different segments of the population. However, since public opinion on economic issues like trade is typically unobservable, this topic has historically been difficult to study.
Addressing this challenge, Van Patten – an EGC affiliate and Assistant Professor of Economics at Yale University's School of Management – and Méndez, a researcher at the Central Bank of Costa Rica, use an innovative approach by exploiting a unique event. In 2007, Costa Rica became the first developing country to hold a national referendum on whether to ratify a free trade agreement – CAFTA with the United States. While CAFTA is a regional agreement, the role of the United States was central to public debate in Costa Rica, given that U.S. trade accounts for about 45% of all Costa Rican imports and exports. The agreement proposed substantial tariff reductions and provisions to liberalize previously government-controlled markets. Voter participation in the referendum was high, with 59% of adult citizens casting a ballot.
As a single-question referendum with broad voter participation, the 2007 CAFTA vote offered a rare opportunity to study public perceptions of trade. Voters were alphabetically assigned to local voting boards that were both small (about 500 voters each) and numerous (nearly 5,000 nationwide). For each voting board, Van Patten and Méndez obtained referendum results along with unique national identifiers for all voters assigned to that board. Using these identifiers, they matched voters to their employers and constructed measures of firm-level trade exposure using customs records, balance sheets, and firm-to-firm transaction data. This mapping enabled the authors to analyze how referendum outcomes relate to the trade exposure of voters’ employers – while preserving voter anonymity by aggregating results at the voting-board level.
How income, expenditure, and politics affect trade opinions
Van Patten and Méndez examined individuals’ likelihood of voting for CAFTA through three main channels: expected effects on their income, expected effects on their consumption costs, and their political alignment.
For the income channel, the authors first looked at voters’ employer firms and estimated the expected change in firm profits if CAFTA were to pass, based on a given firm’s exposure to trade. Any expected increase or decrease in firm profits can be expected to lead to increased or decreased worker incomes. They found that working for a firm expected to benefit from CAFTA increased the share of votes in favor of ratification at a given voting board.
For the expenditure channel, they examined the goods that voters typically consumed and estimated how prices would change under CAFTA, depending on whether the prices of those goods were expected to fall due to cheaper imports. Using detailed survey data, they linked individual voters to household consumption baskets based on characteristics like income, occupation, location, gender, age, and marital status. They found that voters whose consumption was expected to become cheaper were more likely to support the agreement.
To assess political alignment, Van Patten and Méndez accessed voting board data from Costa Rica’s 2006 presidential election. Because voting board assignments remained relatively unchanged, they were able to link 2006 support for a pro-FTA party to voting behavior in the 2007 referendum. They found that the share of politically pro-FTA voters increased the share of pro-ratification votes at a given voting board.
Overall, the study showed that economic fundamentals play an important role in shaping public perceptions towards free trade. Through the income channel, the expected costs and benefits to employers from free trade were shown to directly impact the share of votes for CAFTA, while expected price changes impacted votes through the expenditure channel. Economic factors were found to be almost as significant as political ideology, an especially relevant finding for closely contested referendums and elections.
Julia Luckett
Diana Van Patten and coauthor Esteban Méndez at the 2023 Kuznets Conference.
Implications for policy
In an era where trade is highly politicized and salient in public debate, this research helps shed light on the underlying determinants of public attitudes around the topic.
Van Patten and Méndez emphasize that a clearer understanding of what shapes public opinion towards trade policy can help economists support policy makers in their efforts to communicate the consequences of policy decisions and in designing trade policies that both improve welfare and sustain popular support.
“Nowadays, trade policy has become highly politicized in many countries," said Méndez. "We have seen a resurgence of support for tariffs, trade barriers, and more protectionist agendas. Our work sheds light on the determinants of public attitudes toward trade, and we expect it to help explain why economically efficient policies are not always politically viable.”
Research Summary by Reem Abdo Kahin.