We build on our academic expertise [1,2,3,4,5,6] in analyzing economic networks to aggregate the universe of seaborne, shipment-level import data for the United States. Using these data, we construct a comprehensive set of critical supply disruption indices. The high-quality data and the resulting indices are essential for understanding the evolution of supply disruptions for products critical to the U.S. economy and national defense.
In response to Executive Order 14017, which aims to enhance the resilience of U.S. supply chains, the Department of Commerce’s International Trade Administration (ITA) compiled a list of 2,409 critical products at the HS 6-digit level. We manually reviewed news outlets and scientific literature and narrowed this list to 183 products explicitly identified as critical. Among these, 38 percent fall under the energy sector, 27 percent under critical minerals and materials, 17 percent under information and communication technology, and the remaining 17 percent under public health.
Some products—such as gloves (HS 621600) and disinfectants (HS 380894)—appear critical due to situational media attention (e.g., during COVID-19) but lack long-term systemic importance. To refine the list, we assess each product across four dimensions: (1) national security importance, (2) economic criticality, (3) foreign dependence and domestic production feasibility, and (4) substitutability. Each criterion is rated on a scale from 0 (least critical) to 3 (most critical). Products scoring 2 or higher on either national security or economic criticality are retained. Additionally, products with lower scores in these dimensions but with high foreign dependence or low substitutability are also included.
This refining process is essential for distinguishing temporarily important items from those that are truly critical to long-term U.S. resilience and policy planning. As a result, 152 products remain on the final list.
Additional details are provided in the technical report.
We analyze nearly 200 million transactions covering the universe of U.S. seaborne imports since 2007. The raw data, updated nearly in real time, are based on bills of lading (BoL), which detail the contents of each shipment—including shipper and consignee names and addresses, goods descriptions, vessel names, weights, quantities, and container information. We aggregate the BoL data to the importer–exporter pair level and focus on the ultimate parent companies of U.S. consignees. The dataset is extensively cleaned and refined to ensure consistency across dimensions and robustness to outliers and redaction practices by firms.
In constructing this dataset, we build on our prior research aggregating shipment-level data to study supply chain disruptions [1,6].
General Index Methodology
Our monthly index of supply disruptions reflects the activity of U.S. importer–foreign exporter pairs over time. We define “established” trade pairs as those that traded frequently at some point in our sample. For any given month, we measure the disruption rate for a product as the share of temporarily inactive, established trade pairs among all established pairs that were active in the recent past. The indices are scaled to average zero prior to 2020.
Identifying temporarily inactive trade pairs is particularly challenging toward the end of the sample, as it is not immediately observable whether an inactive pair will resume trade in the future. To address this, we develop an imputation algorithm that leverages the remarkably stable recovery rates of trade pairs across different time horizons. This allows us to construct a timely and near–real-time index of supply disruptions.
Our algorithm includes several tuning parameters—such as the frequency threshold for establishing a trade pair, and the duration required to classify a pair as inactive. We generate a full set of time series spanning all feasible parameter combinations. Each series is de-seasonalized and smoothed using a 3-month rolling average. The final index of supply disruptions is constructed by aggregating across these series.
the underlying data and general methodology are described in more detail in the technical report;
link to U.S. Supply Disruptions Index (SDI).
"Macroeconomic and Asset Pricing Effects of Supply Chain Disasters", by Vladimir Smirnyagin and Aleh Tsyvinski, National Bureau of Economic Research Working Paper 30503, 2022
"Dynamical Structure and Spectral Properties of Input-Output Networks", by Ernest Liu and Aleh Tsyvinski, National Bureau of Economic Research Working Paper 28178, 2023
"Innovation Networks and R&D Allocation", by Ernest Liu and Song Ma, National Bureau of Economic Research Working Paper 29607, 2023
"Shock Propagation Within Multisector Firms", by Jay Hyun, Vladimir Smirnyagin and Ziho Park, 2022
"Industrial Policies in Production Networks", by Ernest Liu, Quarterly Journal of Economics, 2019
"Supply Chain Disruptions, Supplier Capital, and Financial Constraints", by Ernest Liu, Yukun Liu, Vladimir Smirnyagin and Aleh Tsyvisnki.