[Paper Review] Unveiling the Impact of Macroeconomic Policies: A Double Machine Learning Approach to Analyzing Interest Rate Effects on Financial Markets

Anoop Kumar, Suresh Dodda+2 more|arXiv (Cornell University)|Mar 31, 2024

Monetary Policy and Economic ImpactEconomics, Econometrics and Finance3 citations

TL;DR

This study applies a Double Machine Learning (DML) framework combining gradient boosting and causal inference to analyze how U.S. Federal Reserve interest rate changes affect fixed income and equity mutual fund returns from 1986 to 2021. It finds that a 1% rise in interest rates reduces actively managed fund returns by 11.97%, while passive funds show lower sensitivity, demonstrating ML's power in identifying causal macroeconomic effects on financial assets.

ABSTRACT

This study examines the effects of macroeconomic policies on financial markets using a novel approach that combines Machine Learning (ML) techniques and causal inference. It focuses on the effect of interest rate changes made by the US Federal Reserve System (FRS) on the returns of fixed income and equity funds between January 1986 and December 2021. The analysis makes a distinction between actively and passively managed funds, hypothesizing that the latter are less susceptible to changes in interest rates. The study contrasts gradient boosting and linear regression models using the Double Machine Learning (DML) framework, which supports a variety of statistical learning techniques. Results indicate that gradient boosting is a useful tool for predicting fund returns; for example, a 1% increase in interest rates causes an actively managed fund's return to decrease by -11.97%. This understanding of the relationship between interest rates and fund performance provides opportunities for additional research and insightful, data-driven advice for fund managers and investors

Motivation & Objective

• To examine the causal impact of U.S. Federal Reserve interest rate changes on mutual fund returns from 1986 to 2021.
• To compare the performance of gradient boosting and linear regression within a Double Machine Learning (DML) framework for causal inference.
• To investigate differential sensitivity of actively versus passively managed funds to interest rate shocks.
• To provide data-driven insights for fund managers and investors on interest rate risk exposure.

Proposed method

• The study employs the Double Machine Learning (DML) framework to estimate causal effects of interest rate changes on fund returns, enabling robust inference in high-dimensional settings.
• It uses gradient boosting (e.g., XGBoost) and linear regression models as machine learning estimators within the DML framework to predict fund returns.
• The DML approach enables control for confounding variables by first estimating nuisance parameters (e.g., conditional mean and propensity scores) using ML methods.
• The method separates the estimation of the causal effect from the estimation of the conditional expectation, reducing bias in high-dimensional data.
• The analysis covers 36 years of monthly data on fixed income and equity mutual funds, distinguishing between active and passive management styles.
• Causal inference is validated through double-robustness properties, ensuring consistent estimates even if one of the nuisance models is misspecified.

Experimental results

Research questions

• RQ1 How do U.S. Federal Reserve interest rate changes causally affect the returns of fixed income and equity mutual funds between 1986 and 2021?
• RQ2 Do actively managed mutual funds exhibit greater sensitivity to interest rate changes than passively managed funds?
• RQ3 Which machine learning model—gradient boosting or linear regression—provides more accurate and robust causal estimates within the DML framework for this financial application?
• RQ4 To what extent do macroeconomic policy shifts, particularly interest rate changes, drive systematic risk in mutual fund portfolios?

Key findings

• A 1% increase in U.S. interest rates leads to a statistically significant decrease of 11.97% in the return of actively managed mutual funds.
• Passively managed funds show lower sensitivity to interest rate changes, supporting the hypothesis that they are less exposed to rate risk compared to actively managed funds.
• Gradient boosting outperformed linear regression in predicting fund returns within the DML framework, demonstrating superior predictive accuracy and robustness.
• The DML framework successfully isolated the causal effect of interest rate changes on fund returns, even in the presence of high-dimensional confounders.
• The study confirms that machine learning techniques, when combined with causal inference, can yield reliable and actionable insights in financial econometrics.
• The results provide empirical support for using Double Machine Learning in analyzing macro-financial linkages, especially in non-linear and high-dimensional contexts.