📈 economics

Export Performance under US Tariff Pressure: A Time-Series Study of Bangladesh, 2001–2025

This time-series study of Bangladesh's exports to the US from 2001 to 2025 finds that while oil prices significantly influence export performance, US and competitor tariffs show no statistically significant impact, a result attributed to the limited variation in tariff data prior to 2025 and the model's overall lack of explanatory power.

Original authors: Zihad Alam Talha

Published 2026-10-28

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Zihad Alam Talha

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Trade is often imagined as a simple exchange: a factory in one country makes a shirt, and a store in another country buys it. But the path that shirt takes is rarely a straight line. It is shaped by taxes called tariffs, which are fees charged by a government on goods coming from abroad. When a country raises these fees, the cost of the imported item goes up. In the world of economics, the standard expectation is that if a buyer has to pay more, they will buy less, or they will simply switch to a seller from a different country where the fees are lower. This is the logic behind trade wars and tariff disputes. For a nation like Bangladesh, which relies heavily on selling clothes to the United States, these rules are not just abstract theories; they are the difference between a factory staying open or closing down. The question that has occupied economists and policymakers recently is whether a sudden spike in these American fees actually stops the flow of goods, or if other forces, like the price of oil or the value of money, are the ones really pulling the strings.

A researcher at Jagannath University in Dhaka decided to test this idea by looking at the last twenty-five years of trade data. The goal was to see if changes in the taxes the United States charges on Bangladeshi clothes could explain why the amount of clothing Bangladesh sells goes up or down. To do this, the researcher gathered annual records from 2001 to 2025, a period that includes the recent announcement of new, higher tariffs. The study also looked at other factors that might influence sales, such as the taxes the United States charges on a major competitor, Vietnam, the value of the Bangladeshi currency, the money sent home by workers abroad, and the global price of oil. The researcher treated the data like a set of moving parts, carefully separating the long-term trends from the year-to-year changes to see what actually caused the numbers to shift.

The results of this careful examination were surprising. Despite the common belief that higher taxes would immediately hurt sales, the data did not show a clear link between the new tariffs and a drop in exports. In fact, when the researcher ran the numbers, the changes in the tax rates on Bangladeshi goods had no statistically significant effect on how much was sold. The same was true for the taxes on Vietnamese goods; even though Vietnam is a fierce rival in the clothing market, changes in its tax rate did not seem to push more orders toward Bangladesh. The study found that the only factor that consistently moved the needle was the price of oil. When the cost of a barrel of oil went up, the value of Bangladeshi exports to the United States also tended to rise. This suggests that global demand, which often moves in step with energy prices, might be a stronger driver of trade than the specific taxes applied to the goods.

The researcher was careful to explain why the tariff numbers came up empty. For most of the twenty-five years studied, the tax rate on Bangladeshi goods barely changed at all. It sat at a steady level for two decades, only jumping significantly in the final year of the data. Because the tax rate was so flat for so long, there was very little variation for the study to analyze. It is like trying to figure out how a car accelerates by only looking at the speedometer when the car is parked and then suddenly at the moment it hits the gas; there is not enough data in between to see the pattern. The study also noted that the recent tariff changes are so new that their full impact has not yet played out in the yearly numbers. The data shows that in 2025, even after the tax rate jumped, exports actually grew. This does not prove that the taxes are harmless, but it does prove that the available annual records are not yet detailed enough to measure their true effect.

Ultimately, this study serves as a clear baseline for what we know and what we do not know. It shows that for a long time, the relationship between Bangladeshi exports and American taxes was too quiet to measure. The strong connections that appear when looking at raw numbers disappear once the long-term growth trends are removed, leaving only the oil price as a reliable signal. The researcher concludes that to truly understand the impact of these new trade policies, we will need to wait for more data to accumulate, specifically looking at monthly figures rather than yearly ones. Until then, the story of how tariffs affect Bangladesh's clothing industry remains an open question, waiting for the next chapter of data to be written.

Technical Summary: Export Performance under US Tariff Pressure (2001–2025)

Problem Statement
This study investigates the empirical association between US tariff policies and Bangladesh's export performance to the United States, specifically within the context of the 2025 reciprocal tariff regime. While policymakers attribute recent export volatility to these tariffs, the author notes that Bangladesh simultaneously faces confounding macro-financial variables, including currency depreciation, fluctuating remittance inflows, foreign exchange reserve constraints, and global energy price shocks. The core research problem is distinguishing the specific impact of tariff changes from these other drivers using available annual data. The study addresses a gap in existing literature, which largely focuses on the 2018–2019 US–China trade war or relies on general-equilibrium simulations, leaving a scarcity of time-series evidence for small, garment-dependent economies like Bangladesh.

Methodology
The research employs an Ordinary Least Squares (OLS) time-series framework using annual data from 2001 to 2025 ( $N=25$ ). The dependent variable is Bangladesh's exports to the US (in millions of USD). The explanatory variables include:

Policy Variables: The effective US tariff on Bangladeshi goods and the effective US tariff on a key competitor, Vietnam.
Macroeconomic Controls: Exchange rate (BDT/USD), personal remittances, foreign exchange reserves, domestic inflation, and global oil prices.

The econometric strategy prioritizes stationarity to avoid spurious regression. The author conducted Dickey–Fuller unit root tests, revealing that most series (exports, tariffs, exchange rate, remittances, reserves, oil) possess unit roots in levels. Consequently, the model specification utilizes differenced series:

First differences ( $\Delta$ ) for exports, the Bangladesh tariff, reserves, and oil prices.
Second differences ( $\Delta^2$ ) for the exchange rate and remittances.
Levels for the Vietnam tariff and inflation (which were found to be stationary).

The final regression model includes 23 observations (2003–2025) due to the lag structure required for second differencing. Diagnostic checks included the Breusch–Pagan test for heteroskedasticity, Variance Inflation Factors (VIF) for multicollinearity, and the Durbin–Watson statistic for autocorrelation.

Key Results
The empirical findings indicate that the model has limited explanatory power regarding year-to-year changes in exports:

Model Fit: The overall model is not jointly significant ( $F(7,15) = 1.37, p = 0.287$ ), with an adjusted $R^2$ of 0.11.
Tariff Effects: Contrary to the hypothesis that higher tariffs reduce exports, the coefficient for the US tariff on Bangladeshi goods is positive but statistically insignificant ( $p = 0.642$ ). Similarly, the tariff on the competitor (Vietnam) is insignificant ( $p = 0.983$ ). The author attributes this lack of significance to the data structure: the tariff series remained nearly constant (approx. 15%) for over two decades, with significant variation occurring only in the final three years of the sample.
Macroeconomic Factors: Among the controls, only the change in oil prices is statistically significant ( $p = 0.016$ ), showing a positive association with export changes. The exchange rate, remittances, reserves, and inflation are all statistically insignificant.
Diagnostics: The model passes diagnostic tests for heteroskedasticity, multicollinearity (max VIF = 2.87), and serial correlation (Durbin–Watson = 2.22).

Significance and Claims
The paper positions itself not as a definitive causal analysis, but as a "transparent baseline" for future research. Its primary contributions are:

Empirical Baseline: It provides the first time-series assessment of Bangladesh's export performance under the 2025 US tariff regime, explicitly linking it to trade diversion theories.
Methodological Rigor: It demonstrates the necessity of rigorous unit-root testing and differencing, showing that strong correlations observed in level data (e.g., between remittances and exchange rates) disappear once trends are removed.
Data Limitations: The study highlights that current annual data is insufficient to isolate tariff effects due to the lack of variation in the tariff series prior to 2025.

The author explicitly cautions against interpreting the non-significant tariff result as evidence that tariffs have no effect. Instead, the paper argues that the data simply lacks the variation to detect the effect. The study concludes that credible causal assessment requires higher-frequency (monthly) data and product-level analysis to capture the immediate impact of the 2025 policy shock.

Technical Summary: Export Performance under US Tariff Pressure (2001–2025)

More like this