Trade War Expectations, Product Markets, Leverage, and Markup: Evidence from a Field Experiment

Our intervention is a randomized controlled trial (RCT) with a 2×2 factorial design plus a control group, targeting small businesses in China that focus on export. Participating small export-oriented firms in Guangdong, Zhejiang, and Liaoning provinces are randomly assigned to one of five groups:

Foreign Expert – Optimistic Forecast: Receives a written briefing summarizing optimistic predictions from recognized foreign experts about stable and positive prospects for international trade relations.
Foreign Expert – Pessimistic Forecast: Receives a briefing summarizing pessimistic predictions from foreign experts, emphasizing heightened risks and challenges due to potential trade wars.
Domestic Expert – Optimistic Forecast: Receives a briefing presenting optimistic forecasts from leading domestic (Chinese) experts, highlighting stability and minimal trade war risks.
Domestic Expert – Pessimistic Forecast: Receives a briefing presenting pessimistic domestic expert forecasts, stressing increased uncertainty and risks related to trade conflicts.
Control Group: Receives a neutral briefing unrelated to trade wars or geopolitical risks.
By randomly allocating small export-oriented businesses to these groups, our RCT will causally identify the individual and combined effects of expert origin (foreign vs. domestic) and forecast sentiment (optimistic vs. pessimistic) on firms’ export strategies, pricing/markup decisions, and other key business behaviors in the face of geopolitical uncertainty.
Foreign Expert – Optimistic Forecast: Receives a written briefing summarizing optimistic predictions from recognized foreign experts regarding stable and positive future trade relations.
Foreign Expert – Pessimistic Forecast: Receives a briefing summarizing pessimistic predictions from foreign experts, emphasizing increased trade war risks and challenges.
Domestic Expert – Optimistic Forecast: Receives a briefing with optimistic forecasts from leading domestic (Chinese) experts, highlighting trade stability and minimal risk.
Domestic Expert – Pessimistic Forecast: Receives a briefing with pessimistic domestic expert forecasts stressing increased risks and uncertainties from potential trade conflicts.
Control Group: Receives a neutral briefing unrelated to trade wars or geopolitical risks.
By randomly assigning firms to these groups, our RCT allows us to causally identify the separate and interactive effects of expert origin (foreign vs. domestic) and forecast sentiment (optimistic vs. pessimistic) on firms’ export strategies, pricing/markup decisions, and other key business behaviors under geopolitical uncertainty.

2025-09-17

2025-10-17

Primary Outcomes (End Points):

The key outcome variables of interest in this experiment are:

Export Strategy Adjustments:
Changes in firms’ planned or actual export destinations, diversification of export markets, or modifications in export volumes in response to the information treatments.

Pricing/Markup Decisions:
Adjustments in product prices or markups for exported goods, reflecting changes in perceived risk or market outlook.

Business Confidence and Expectations:
Measures of business owners’ confidence in the future of their export operations, including expectations about sales, profits, and risk over the next 6–12 months.

Investment and Employment Plans:
Reported changes in planned investments (e.g., capital expenditures, technology upgrades) and employment (e.g., hiring or layoffs) related to export activities.

Risk Management Behaviors:
Adoption of risk mitigation strategies, such as hedging currency risks, renegotiating contracts, or seeking new suppliers/customers.

Sample and Setting

We conduct a randomized controlled trial (RCT) involving 5,000 small export-oriented businesses located in three of China’s most prominent wholesale and export markets: Huaqiangbei (Guangdong), Yiwu (Zhejiang), and Wuai (Liaoning). These hubs are selected due to their concentration of small businesses highly dependent on international trade.

Eligible businesses are identified via local commercial registries and market management records. All participating firms are formally registered, have fewer than 50 employees, and derive at least 30% of annual revenue from export activities.

Experimental Details

Randomization and Groups

Firms are stratified by market (Huaqiangbei, Yiwu, Wuai) and sector (e.g., electronics, textiles, general merchandise) to ensure balanced representation. Within strata, firms are randomly assigned (via computer algorithm) to one of five experimental arms:

Foreign Expert – Optimistic Forecast (n ≈ 1,000): Receives a standardized briefing attributed to reputable foreign trade experts, summarizing optimistic projections for China’s export outlook (e.g., stable policies, new opportunities).
Foreign Expert – Pessimistic Forecast (n ≈ 1,000): Receives a briefing from foreign experts emphasizing risks and uncertainties in the global trade environment (e.g., escalating trade frictions).
Domestic Expert – Optimistic Forecast (n ≈ 1,000): Receives an optimistic briefing attributed to prominent domestic experts, highlighting positive prospects for Chinese exporters.
Domestic Expert – Pessimistic Forecast (n ≈ 1,000): Receives a pessimistic briefing from domestic experts, stressing challenges such as trade barriers or policy headwinds.
Control Group (n ≈ 1,000): Receives a neutral briefing, with generic business news unrelated to trade policy or geopolitical risk.
Randomization is implemented using re-randomization to achieve balance on baseline variables, including owner age, gender, education, business age, total assets, and recent export revenue. We repeat random allocations until key variables have maximum t-statistics below 1.25 and average below 0.35 across groups.

Intervention Delivery

Briefings are distributed to business owners/managers via email or secure online platform. Each briefing is of equal length and format, and each is clearly attributed (foreign/domestic expert, optimistic/pessimistic tone). Owners are explicitly told that the information comes from leading experts in international trade.

Data Collection

Baseline Survey: Administered pre-intervention to capture firm characteristics, export patterns, pricing strategies, investment plans, risk perceptions, and owner demographics.
Follow-up Surveys: Conducted at 3, 12, and 24 months post-intervention to track changes in export strategy, pricing, revenue, business confidence, risk management, and investment/employment intentions.
Administrative data (where available) will be used to cross-check self-reported export volumes and business activity.
Primary Outcomes

Adjustments in export strategies (e.g., market diversification, volume changes)
Changes in export pricing or markups
Business confidence and expectations regarding exports
Investment and employment plans related to export operations
Adoption of risk management practices

Randomization is conducted at the individual business level using a computer-generated algorithm. The 5,000 small export-oriented businesses are first stratified by wholesale market location (Huaqiangbei, Yiwu, Wuai) and by business sector (e.g., electronics, textiles, general merchandise) to ensure balanced allocation across key subgroups. Within each stratum, firms are randomly assigned to one of five experimental groups (four treatment arms and one control group) using random numbers generated in Stata.

To achieve optimal balance on important baseline characteristics—including business owner’s age, gender, education, business age, total assets, and most recent export revenue—a re-randomization procedure is used. After an initial random allocation, we calculate the maximum and average t-statistics for differences in means of these variables across groups. If the maximum t-statistic exceeds 1.25 or the average exceeds 0.35, the randomization is repeated with a new random seed. This process continues until these balance criteria are met.

All randomization and balance checks are implemented using pre-specified Stata code, and the process is conducted in a transparent and reproducible manner.

Individual business (firm)

No

5,000 individual businesses (firms)

5,000 individual businesses (firms)

1,000 businesses – Foreign Expert Optimistic Forecast
1,000 businesses – Foreign Expert Pessimistic Forecast
1,000 businesses – Domestic Expert Optimistic Forecast
1,000 businesses – Domestic Expert Pessimistic Forecast
1,000 businesses – Control Group