Risk-Free Rate in Backtests: How We Handle Cash Returns on BackAlpha

Published June 15, 2025

When simulating a trading strategy, it's important to consider not just what happens when you buy stocks — but also what happens when you don't. On BackAlpha, we incorporate a risk-free rate to simulate interest earned on idle cash. This allows for more accurate modeling when your strategy is partially or fully in cash.

What Is the Risk-Free Rate?

The risk-free rate is the theoretical return you'd get on cash with zero risk. In real life, this is typically approximated by short-term government bonds or cash sweep programs at brokerages. On BackAlpha, we use practical estimates based on real-world brokerage data — such as the cash sweep program offered by Robinhood, which currently pays around 4% annually for uninvested cash.

Why Does It Matter?

Many trading strategies don't keep your capital fully invested at all times. For example, a crossover strategy may hold cash between buy and sell signals. If that cash earns interest, the performance of the strategy improves — especially in high-rate environments.

Ignoring the return on idle cash would understate performance and misrepresent real-world behavior. Including the risk-free rate gives a fairer view of how capital grows, even when not actively in the market.

How BackAlpha Applies the Risk-Free Rate

In our crossover strategies, cash positions automatically accrue interest based on the selected risk-free rate. You can specify this value directly when running a simulation using the built-in form. For example, setting a risk-free rate of 4% will apply 4% annual interest to any uninvested cash throughout the backtest.

Interest compounds daily while capital is in cash
The rate can be adjusted interactively using the form
Higher rates have larger impact on strategies with longer cash periods

Try it yourself by exploring a Golden Cross simulation with 4% risk-free rate — or use the form to set your own assumptions and see how cash returns shape your performance over time.