Skip to content
All library documents

Building and Validating Macroeconomic Models for Silver Prices

Article Quant Q&A · Author: Johan Smith
Read the original

Summary

The document outlines how to investigate whether macroeconomic indicators such as growth, inflation, and monetary policy help forecast silver prices. It recommends first considering the economic relationships, timing, and functional form linking those variables to silver. When theory does not specify a reliable model, it suggests comparing empirical approaches by their out-of-sample prediction error.

Candidate methods mentioned include ARIMAX, multivariate regression, and machine-learning models. The response points to prior research on commodity-price forecasting with macroeconomic and financial predictors, alongside studies comparing neural networks and time-series approaches. These references offer starting points for method selection, but the document does not report a silver-specific test, identify a winning model, or provide dataset recommendations and validation details. Any forecast would therefore need its own carefully designed evaluation, including suitable predictor lags and genuinely held-out data; performance on other commodity series does not establish accuracy for silver.

Key ideas

  • Economic reasoning can help identify relevant macroeconomic predictors, their lag structure, and possible nonlinear relationships with silver prices.
  • When theory is unclear, compare candidate models by their error on held-out observations.
  • ARIMAX, multivariate regression, and machine-learning methods are possible empirical approaches.
  • Prior commodity-forecasting research can inform model choices but does not prove performance for silver.
  • The document offers no silver-specific results, concrete dataset guidance, or detailed validation design.

Tags

CommoditiesStatisticsMachine learningBacktesting
Full text
# What Quantitative Methods Best Predict Silver Prices Based on Macroeconomic Indicators?


# What Quantitative Methods Best Predict Silver Prices Based on Macroeconomic Indicators?












I'm seeking guidance on developing a robust quantitative model to predict silver prices using macroeconomic indicators. How can I incorporate variables like GDP growth, inflation rates, and monetary policy changes into a predictive model for silver? Which statistical techniques and models (e.g., ARIMA, VAR, machine learning) have shown success in accurately forecasting precious metal prices in response to macroeconomic shifts? Recommendations on datasets and model validation methods would also be helpful.

## Answer by Sane (score 1)

https://quant.stackexchange.com/a/79643

Developing a robust quantitative model to predict silver prices using macroeconomic indicators requires a careful consideration of the relationships between these indicators and the price of silver. Ideally, you should understand theoretical relationship between silver price and macro variables. In other words, what macro variables affect silver price, what is lag-structure between variables, is the functional form linear or non-linear, among other questions. If "theory" is available, you can empirically estimate the model (estimate parameters of the model).

However, most of the time theoretical relationship is not available. Therefore, you should play around with empirical data and find best model having minimum test error. You can use both econometric (e.g., ARIMAX, or multivariate regression models) or machine learning models.

Have a look at the following paper: Gargano, A., & Timmermann, A. (2014). Forecasting commodity price indexes using macroeconomic and financial predictors. International Journal of Forecasting, 30(3), 825-843. (https://www.sciencedirect.com/science/article/abs/pii/S016920701300099X)

Other papers that might be relevant:

- Wang, J., & Li, X. (2018). A combined neural network model for commodity price forecasting with SSA. Soft Computing, 22, 5323-5333.

- Kohzadi, N., Boyd, M. S., Kermanshahi, B., & Kaastra, I. (1996). A comparison of artificial neural network and time series models for forecasting commodity prices. Neurocomputing, 10(2), 169-181.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

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