Trade uncertainty and credit supply
Threats of restrictive trade policy and geopolitical risks are ending the period of stable international trade the global economy experienced during the past couple of decades. How does this trade uncertainty spill over to credit supply? This is the question Correa et al. analyze in this paper using a difference-in-differences estimation framework. They examine the spikes in trade uncertainty that occurred between 2018 and 2019 and combine them with detailed data on loan origination by U.S. banks to domestic and foreign borrowers collected through the Federal Reserve. They further augment the data set with bank and firm financial information. The authors identify three principal effects:
- An increase in trade uncertainty is associated with a larger credit contraction at the bank-firm level for banks with a greater ex-ante share of loans to firms in sectors that face higher ex-post trade uncertainty.
- The contraction in credit supply is larger for banks with business models that support global trade and for banks that face larger financial frictions.
- Firm characteristics also affect how banks adjust their lending behavior in the face of trade uncertainty. More specifically, banks contract lending more to firms that are more dependent on imported intermediate goods for production, highlighting a potential channel through which shocks to global value chains may affect domestic credit market conditions.
A simple aggregation based on our estimates shows that credit originations while controlling for loan demand, dropped by about 0.5 percentage points (per quarter) for banks exposed to trade uncertainty in 2018 and 2019—a material effect given that the average growth rate during 2015–2017 was 4.2%. The banks in the 75th percentile of the exposure to trade uncertainty distribution had credit growth rates that were 3.9 percentage points lower than banks in the 25th percentile—also a notable difference.


Trade Uncertainty and U.S. Bank Lending
Authors: Ricardo Correa, Linda Goldberg, Julian di Giovanni, Camelia Minoiu
From: Federal Reserve Board of Governors, Federal Reserve Bank of New York, Federal Reserve Bank of Atlanta
Nowcasting trade data with machine learning
Chinn et al. put forth a new framework on how to generate timely information on trade volumes, ahead of the Dutch Centraal Plan Bureau (CPB) data releases which take place roughly eight weeks after month end. Toward this goal they use machine learning, distinguishing between tree-based methods (random forest, gradient boosting) and their regression-based counterparts (macroeconomic random forest, gradient linear boosting). The authors adopt a three-step approach for forecasting with machine learning and large datasets. The approach works sequentially:
- Step 1: a pre-selection technique identifies the most informative predictors among our dataset of 600 variables;
- Step 2: selected variables are summarized and orthogonalized into a few factors;
- Step 3: factors are used as explanatory variables in the regression of world trade, using machine learning techniques.
This three-step approach also outperforms workhorse benchmarks, such as a PCA-OLS model, an elastic net, or a dynamic factor model. Finally, on top of high accuracy, the approach is flexible and can be extended seamlessly beyond world trade.
Nowcasting World Trade with Machine Learning: a Three-Step Approach
Authors: Menzie D. Chinn, Baptiste Meunier, and Sebastian Stumpner
From: University of Wisconsin-Madison, Banque de France, ECB