Forecasting the Unforecastable? Estimating Investment Returns and Risk

Building sophisticated, automated investment strategies – the kind that power modern robo-advisors – isn't magic. Under the hood, these systems rely heavily on forecasts about how different assets might behave in the future. Two critical inputs drive their portfolio construction engines: expected returns (how much we think an asset will appreciate) and risk, typically measured by variance (how volatile an asset is) and covariance (how assets move together).

But here's the rub: the future is uncertain. Estimating these inputs accurately is one of the most challenging aspects of quantitative finance. Let's dive into some common approaches and their nuances.

The Elusive Expected Return: More Art Than Science?

If we knew exactly how much stocks or bonds would return next year, investing would be easy! Since we don't, we have to estimate. Here are a few ways practitioners approach this:

1. Looking in the Rear-View Mirror: Historical Averages