Estimating Portfolio Volatility from Weighted Asset Returns

Summary

The discussion considers a portfolio allocated across stocks, high-yield bonds, and commodities, and asks how to estimate its volatility. One response notes that fixed-income returns need to account for bond characteristics such as duration and spread duration; it also describes combining successive single-period returns multiplicatively to obtain a period return. This highlights why a bond yield correlation alone may not capture all relevant price behavior.

Another approach is to calculate the portfolio’s return series directly using its fixed asset weights, then apply a chosen volatility estimator to that series. This avoids estimating and combining a separate correlation matrix and permits methods such as a rolling window or GARCH. The excerpt offers conceptual guidance rather than a worked calculation or comparison of estimator performance. Its direct portfolio-return method assumes the stated weights remain fixed, and the text does not discuss rebalancing, changing exposures, or how to estimate the individual asset returns.

Key ideas

• Portfolio volatility can be estimated from a time series of portfolio returns formed using asset weights.
• Applying volatility methods to portfolio returns avoids separately modeling asset correlations.
• Bond returns require attention to duration and spread effects, not only yield correlations.
• The discussion does not compare volatility estimators using empirical results.

Tags

# Portfolio volatility - Real life application


# Portfolio volatility - Real life application












Given that a portfolio consists of Stock=USD 30, High-yield bonds(duration=5 years,spread duration=5 years) =USD 40 , Commodity = USD 30.

## Answer by Vitomir (score 1)

https://quant.stackexchange.com/a/45151

Taking the correlation between the bond yield and the stock may miss some information, such as duration, rolldown.

For calculating the returns of a Fixed income product over a period of time = PROD(1+r)-1

where r is the single period return

## Answer by JPN (score 1)

https://quant.stackexchange.com/a/45178

You actually don't need to worry about correlations. Just calculate the portfolio value over time using your fixed weights `[.3, .4, .3]`.

The advantage of this approach is;

- You don't have to worry about correlations.

- You can use any volatility method you want (rolling window, GARCH, etc).

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.