About

About the Project

Border Value is an interactive catalog of value chains between Mexico and the U.S., developed by Datawheel thanks to a grant from the Open Society Foundations. It uses data from AIPNET and the Observatory of Economic Complexity (OEC) to identify the most relevant cross-border components.

The goal of Border Value is to illustrate the importance of crossborder value chains for the competitiveness of North America.

Data Sources

Border Value is built from two sources:

1. AI-generated Production Network - AIPNET

   Developed by Fetzer, Lambert, Feld, and Garg (2024), AIPNET is a database created with machine learning to map input-output relationships among more than 5,000 goods

2. The Observatory of Economic Complexity (OEC)

   A leading online platform that integrates, visualizes, and distributes economic data for countries, regions, and firms. The OEC gathers information from custom offices and financial reports for the largest economies in the world, including the latest trade data for Mexico and the U.S.

Methodology

These value chains combine the relationships identified by AIPNET with trade data from the OEC. Each product is connected to its inputs using AIPNET data and displays the flag of the country with the largest net trade according to the most recent OEC data (for example, Mexico’s flag if Mexico exports more of that product to the United States than it imports).

This results in a three-level hierarchical visualization that shows the most relevant production links between Mexico and the U.S.

1. Level 0 - Root or Final Product (HS4)

   This is the product you can select through search. The chain begins with a single 4-digit product from the Harmonized System (HS4).

2. Level 1 - Direct Components (HS6)

   These are products identified as direct inputs by combining the data from AIPNET and the OEC.

   • Products are filtered to maintain sectoral consistency: same HS section or adjacent sections.
   • They are prioritized based on total bilateral trade (U.S. and Mexico), selecting the top 10.

3. Level 2 - Secondary Components (HS6)

   We identify up to 30 secondary components using the following criteria.

   • We avoid products belonging to the same HS4 as the root product to avoid loops.
   • We select up to 3 secondary components per direct component, based on total bilateral trade (for a maximum of 30 nodes).

Building Rules

Each product appears at only one level. The value chain is generated by consistently applying the sources and methodological steps described above.

Border Crossings

This panel shows the crossings in a product’s value chain between the United States and Mexico. We count a crossing each time there is a change in the orientation of the net trade balance. For example, since Mexico exports more cars to the United States than it imports, but imports more processors from the United States than it exports, the link between processors and cars counts as a crossing (U.S. processors in Mexican cars).

Each panel is composed of:

1. Direction: indicates the direction of the crossing (U.S. → Mexico or Mexico → U.S.).

2. Crossings: total number of crossings identified for the network and corresponding direction.

3. Estimated Monthly Flow: estimated monthly flow of exports between both countries for all products in the network. It is calculated using the following formula:

\(F_{est,m}\) = \(\frac{X_{cp'd} + \sum_{p} \frac{X_{cp'd}}{X_{cd}} \cdot X_{cpd}}{m}\)

Where:

• \(p'\): represents the root product of the value chain.

• \(p\): represents the level 1 and 2 products of the value chain.

• \(c\): represents the origin country of the commercial flow.

• \(d\): represents the destination country of the commercial flow.

• \(m\): represents the number of months considered in the average, based on data availability.

• \(X_{cp'd}\): represents the total exports of the root product from the origin country to the destination country.

• \(X_{cpd}\): represents the total exports of the product p from the origin country to the destination country.

• \(X_{cd}\): represents the total exports from the origin country to the destination country.

4. Crossing level: represents the position in the value chain network:

   • Level #1: links between level 0 and level 1 nodes.
   • Level #2: links between level 1 and level 2 nodes.

5. Crossings by level: number of crossings recorded by direction and level.

HomeNetworks