Improving the Short-term Forecast of World Trade During the Covid-19 Pandemic Using Swift Data on Letters of Credit
IMF Working Papers, November 13, 2020
Source details
- Canonical URL
- Improving the Short-term Forecast of World Trade During the Covid-19 Pandemic Using Swift Data on Letters of Credit
Other formats
Bibliographic details
- Authors: Benjamin Carton, Nan Hu, Joannes Mongardini, Kei Moriya, Aneta Radzikowski
- Published: November 13, 2020
- Series: IMF Working Papers
- DOI: https://doi.org/10.5089/9781513561196.001
Summary
- Objective: Improve the short-term forecast of international trade during the Covid-19 pandemic.
- Data inputs used: SWIFT messages on letters of credit, crude oil prices, and new export orders of manufacturing Purchasing Managers’ Index (PMI).
- Approach: A horse race comparing linear regressions and machine-learning algorithms for the world and 40 large economies.
- Main high-level result: Forecasts based on linear regressions often outperform those based on machine-learning algorithms, confirming the linear relationship between trade and its financing through letters of credit.
Methodology and Scope
- Models compared: linear regression models versus machine-learning algorithms.
- Geographic coverage: world aggregate and 40 large economies.
- Key predictors emphasized: SWIFT letters of credit messages, Brent crude oil price (crude oil prices), new export orders of manufacturing PMI.
Key Findings
- Linear regressions frequently produce more accurate short-term trade forecasts than machine-learning approaches in the sample considered.
- The empirical results support a linear relationship between trade and its financing via letters of credit.
- The combination of SWIFT letters of credit data with crude oil prices and manufacturing export orders (PMI) improves short-term trade forecasting performance.
Subjects and Keywords
- Subject: Exports, Imports, International trade, Oil prices, Prices, Trade balance, Trade finance
- Keywords: Africa, Asia and Pacific, Baltics, Brent crude oil price, Exports, Global, Imports, letter of credit, linear regression forecast, machine learning, Merchandise trade, Oil prices, SWIFT, trade advance, Trade balance, Trade finance, trade forecast, trade message, world trade, World trade sample, WP
IMF Working Paper — Improving the Short-term Forecast of World Trade During the Covid-19 Pandemic Using Swift Data on Letters of Credit (Working Paper No. 2020/247, November 13, 2020).
Content in this bundle
- Working Paper