Testing Visa Spending Data as a Predictor of EUR/USD Direction

Summary

The article investigates whether Visa’s monthly Spending Momentum Index series can help classify future EUR/USD direction. It combines headline, discretionary, and non-discretionary spending measures with monthly price data, examines plots and correlations, and trains a neural network with parameter search. The reported validation accuracy is 71% for a forecast based on monthly data. The analysis also uses feature selection and mutual information to assess whether the spending variables add predictive information.

The evidence is mixed. The headline and discretionary series are highly correlated with each other, while feature selection retained only the opening price and performance worsened as more inputs were added. The author therefore cautions that the apparent relationship may be unstable or difficult for the model to capture, and does not claim the data is a reliable signal. The reported accuracy is limited to the article’s validation setup; the document does not establish live trading performance, robustness across periods, or profitability after trading costs.

Key ideas

• The study tests three Visa consumer-spending series as inputs for monthly EUR/USD direction classification.
• A neural network with randomized parameter search reportedly reached 71% validation accuracy.
• The headline and discretionary Visa series are reported as highly correlated.
• Feature selection favored the opening price and excluded the Visa measures, weakening the case for their added value.
• The validation result does not establish a stable relationship or profitable live strategy.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.