Aid, Lending, and TRIPS

Back to Home

What is National Trade Estimate (NTE)?

Pursuant to the Trade Act of 1974 and the amendments made in 1984, the United States Office of the Trade Representative (USTR) has been vested with the authority to investigate and identify countries that pose significant threats to the fair trade practice with the United States. National Trade Estimate (NTE) reports are the proceedings of these investigations carried out by USTR. While the reports have been published since 1986, the Wayback Machine only has access to the old versions of USTR's website starting from 1995. That being said, this project involves 28 years of data spanning from 1995 to 2022.

NTE report features roughly 60 countries every year, with each chapter dedicated to each country, including major areas of trade-relevant concerns, e.g., import tariffs, export subsidies, investment barriers, technical barriers to trade, and so forth. These categories are not uniform across entire set of reports but differ from administration to administration or even from year to year, although some of them consistently appear throughout the reports. I recategorize them into 15 issue areas that appear most frequently throughout the reports: Import Policies; Export Subsidies; Standards, Labeling and Certification; Government Procurement; Intellectual Property Rights; Services Barriers; Investment Barriers; Anti-competitive Pratices; Technical Barriers to Trade; Sanitary and Phytosanitary Barriers; E-commerce; Barriers to Digital Trade; Agriculture; Trade Remedies; and Other Barriers. For more information on the data, refer to this hyperlinked page.

New Proxy for US IPR Evaluation vis-à-vis Emerging Economies

DeBERTa-v3-large LLM (Laurer et al. 2024)

What is DeBERTa?

DeBERTa-v3-large is BERT-based LLM chosen for stance detection based on recommendations from Burnham (2024) and Laurer (2024).
Offers high accuracy, efficiency, and flexibility, especially in low-label settings.
Excels at textual entailment tasks and outperforms other public models on Hugging Face as of June 2022.
Suitable for few-shot learning, reducing the need for extensive manual annotation.
A -5 to +5 scale is used to capture the U.S. stance toward a country’s IPR regime.
Proxy is based on multiple hypotheses reflecting various aspects of IPR enforcement and policy.
Weighted entailment probabilities (from model outputs) are used to calculate the final score.

Stratification in evaluations?

It is interesting to see that the trend of the DeBERTa score coincides with the rise of China, and the fall of Japanese economy, while EU has been subject to ever-increasing complaints from US pharmaceuticals about drug pricing and reimbursement policies of its member states. Moreover, the long-standing battle between Boeing and Airbus only settled around 2020.

Main Results

*Hover over the plots to see which observations belong to which country-year.

I use 2SLS design to address endogeneity concerns for aid obligations.

The instruments are:

US divided government status (Kersting & Kilby, 2021)
infectious diseases mortality rates (WHO Mortality Database)

Aid and TRIPS

The joint effect (triple interaction term) of US aid obligations, DeBERTa score and regime type is visualized in the figure as a contour plot with scattered points of observations. The plot reveals an interesting pattern. In both autocratic and democratic emerging economies, those that face significant IPR concerns as indicated by lower DeBERTa scores are more likely to sign TRIPS-plus trade agreements with the United States as the amount of aid obligations they receive increases, but the effect is much more pronounced for democracies. Say there are a hypothetical democracy and an autocracy each with DeBERTa score at -2.5. Increasing aid obligation from 18 to 19 results in increasing the predicted probability of signing TRIPS-plus for democracies from [0.20, 0.25) to [0.25, 0.30), but it barely has an effect for autocracies as the predicted probability range is still at [0.15, 0.20).

IFC lending and TRIPS

This figure illustrates the joint effect of DeBERTa score and IFC lending on the probability of signing TRIPS-plus. While the predicted range for autocracies is lower than democracies in general, both DeBERTa score and IFC lending clearly have joint effect on TRIPS-plus for autocracies, but IFC lending shows little to no effect in democracies. More specifically, probability of signing TRIPS-plus increases for autocracies receiving substantial IPR-related concerns (DeBERTa < -2.5) as their private sectors receive more IFC loans. In contrast, signing TRIPS-plus is mostly driven by an increase in the DeBERTa score in democratic developing countries, as mostly vertical contour lines show.

After instrumenting IFC lending on US divided government status and UNGA voting alignment with the United States, it does appear to have some effect in democracy panel as well, compared to the preceding analysis where IFC lending shows almost no effect on probability of signing TRIPS-plus agreements. Compared to autocracies, however, the effect is still less substantial in democracies. Contour lines of the democracy panel show broader gaps than more tightly packed lines of the non-democracy panel, meaning the joint effect of IFC lending and DeBERTa score is stronger in autocracies. Moreover, pitting an autocracy against a democracy in the lower score region (DeBERTa < -2.5) shows that increasing the logged IFC loan amount from 16 to 17.5 ramps up the predicted probabilities interval from [0.40, 0.44) to [0.44, 0.48) for autocracies, whereas the predicted probabilities interval remains the same at [0.40, 0.44).

Google Sites

Report abuse