Alfredo Sosa
Economist specializing in trade policy, supply chains, and applied econometrics
🌐 https://alfredo-sosa-economics.com/
https://www.linkedin.com/in/alfredososa513/
https://orcid.org/0009-0001-2453-2496
✉️ alfredosos@gmail.com 📍 Ann Arbor, Michigan
About
Economist with more than 20 years of experience applying econometrics, data science, and business analytics to real-world economic and operational problems. My work focuses on international trade policy, supply chains, and industrial organization, with particular emphasis on how tariff shocks propagate through U.S. manufacturing supply chains, including a sustained focus on the automotive and commercial vehicle sector.
Current research examines how tariff shocks propagate through global production networks, using detailed trade data, industrial production indicators, and supply-chain measures to analyze impacts on prices, output, and cost structures.
This website presents a portfolio of research papers, data products, and applied analytics tools focused on the economic effects of trade policy and supply chain dynamics.
Research Program
These papers trace a connected research program: an early sectoral application (Paper 2) and a supporting policy dataset (Paper 3) motivated the development of a more rigorous causal design (Paper 1), with firm-level extensions planned next (Paper 4).
1. Where Do Tariff Shocks Propagate? Evidence on Production, Prices, and Shipments from Section 301 China Tariffs
Working Paper, 2026 — Under review at a peer-reviewed journal. Available at SSRN.
This paper estimates the causal effects of the 2018–2019 Section 301 China tariffs on U.S. manufacturing activity. It combines a continuous, time-varying measure of industry tariff exposure with monthly NAICS-4 outcome panels for producer prices (BLS-PPI), real industrial production (FRED-IP), and nominal manufacturers’ shipments (Census M3). The empirical strategy uses a triple-difference design that controls for world, non-China import penetration, helping distinguish tariff exposure from broader import-competition dynamics.
The principal result is a statistically significant decline in real manufacturing activity among more tariff-exposed industries in the FRED-IP panel. The producer-price results are suggestive of pass-through but remain statistically inconclusive under bootstrap inference. In the primary specification, the Census M3 shipment estimates also point to a negative effect, though this result is imprecisely estimated. Taken together, the findings suggest that tariffs imposed on Chinese imports may raise input-cost pressure and weaken real manufacturing output rather than producing a straightforward expansion of domestic activity.
Download Paper (PDF)
View on SSRN
Replication data archived at Zenodo: 10.5281/zenodo.21764320 (files under embargo until August 2, 2027; DOI is citable now).
2. Earlier Vehicle-Sector Study: The Impact of Section 301 Tariffs on U.S. Vehicle Production and Sales
This earlier study examines how the 2018–2019 Section 301 tariffs affected downstream production and sales in the U.S. commercial- and light-vehicle sectors. It uses monthly industry data and an HS6-level, import-weighted tariff-exposure index to measure how strongly different vehicle-related industries were exposed to tariffed Chinese inputs. The analysis combines continuous-treatment difference-in-differences models with event-study specifications to examine both average effects and their timing.
The results indicate that higher tariff exposure is associated with lower production in both commercial- and light-vehicle sectors. Nominal sales outcomes are more mixed across specifications, suggesting that price and revenue responses may not move in parallel with real production. The event-study evidence indicates that the estimated production effects emerge after tariff implementation rather than reflecting clearly divergent pre-treatment trends. This vehicle-focused paper provides an initial sectoral application that motivates the broader manufacturing analysis in the newer working paper. Note: this early specification’s standard errors rely on a static, cross-sectional exposure measure repeated across the panel period, which likely overstates inferential precision; the triple-difference design in Paper 1 above addresses this limitation directly.
3. A Tariff Exposure Dataset for the U.S. Vehicle Supply Chain: Policy Evidence from the 2025–2026 Tariff Expansions
This paper documents the expansion of U.S. tariff policy during 2025–2026 and develops a structured, product-level dataset for measuring tariff exposure across the U.S. vehicle supply chain. It consolidates information from Federal Register notices, administrative actions, and public trade data covering Section 301 expansions, Section 232 measures, antidumping and countervailing-duty actions, and related policy developments.
The project converts fragmented legal and administrative materials into a transparent framework that records the timing, intensity, and sectoral distribution of tariff changes. It also assembles descriptive evidence on production, sales, prices, employment, logistics, and supply-chain conditions during the expansion period. The paper’s central contribution is data infrastructure rather than a completed causal estimate: it creates a documented foundation for future research on how rapidly changing trade policy affects vehicle-related industries, input costs, production networks, and supply-chain adjustment.
4. Firm-Level Evidence on Tariff Exposure and U.S. Manufacturing Outcomes
Future research
This future project will extend the research agenda from industry-level panels to firm-level evidence on how tariff exposure affects production, sales, investment, employment, and performance across heterogeneous U.S. manufacturers. The goal is to identify whether the consequences of trade-policy shocks differ by firms’ import dependence, supply-chain position, export orientation, size, and capacity to adjust sourcing or prices.
Building on the tariff-exposure measures and policy documentation developed in the preceding working papers, the project will link product-level policy variation to firm-level outcomes and implement causal designs such as difference-in-differences. A firm-level approach can help distinguish broad industry effects from the heterogeneous adjustments of individual producers and can provide more direct evidence on mechanisms, including input-cost pressure, sourcing substitution, and trade retaliation. This work is part of a longer-term research program connecting public trade-policy data to measurable manufacturing and supply-chain outcomes.
View all research and working papers
Applied Analytics Projects
Interactive Supply Chain Dashboards
Developed interactive dashboards analyzing trade flows, tariff exposure, production indicators, and supply chain dynamics in the U.S. vehicle sector. These tools integrate data from USITC, FRED, and internal processing pipelines to support real-time analysis of policy and market conditions.
Machine Learning Applications
Developed machine learning models to support forecasting and policy analysis in supply chains and logistics. Applications include demand forecasting, pricing dynamics, and counterfactual analysis of tariff exposure and input costs.