Estimating Portfolio Beta When Assets Have Short Histories

Summary

The document addresses how to estimate a portfolio’s beta when some holdings, such as recently listed companies, lack enough historical returns for reliable individual beta estimates. It considers estimating beta by regressing portfolio returns on index returns, after reallocating the weights of assets with missing histories among the holdings with longer records.

The response describes using a proxy security with a longer price history to stand in for an asset with limited data, potentially adjusting the proxy to better reflect the target company. This can make individual beta estimates usable in a weighted portfolio calculation. The document does not compare this approach empirically with portfolio-level regression or explain how to choose or adjust proxies. Proxy selection is identified as an open judgment call, so results depend on whether the substitute represents the asset’s risk sufficiently well.

Key ideas

• A proxy security can provide historical data for an asset with a limited trading record.
• Proxy-based beta estimates may support a weighted calculation of portfolio beta.
• Selecting a suitable proxy is an open problem and may require adjustments.
• The document does not establish that proxy estimates outperform portfolio-level regression.

Tags

# Constructing Portfolio Beta


# Constructing Portfolio Beta












Suppose I have a portfolio with securities with different history. Say some securities have 15-20 years of history and some are like Uber or Lyft, which has limited history. There are assets with 1/2/3/5/ etc years of history as well.

How would you go about calculating portfolio Beta? Calculating individual betas for Uber/Lyft etc would be impossible. Hence calculating portfolio beta by taking weighted sum of individual security betas would, in my opinion, be problematic.

I am thinking about creating portfolio value for, say, 10 years, by reassigning weights of securities missing proportionally to other securities. And then calculate portfolio Beta by regressing portfolio returns against the Index returns.

Is that incorrect?

## Answer by KaiSqDist (score 1)

https://quant.stackexchange.com/a/81849

"Hence calculating portfolio beta by taking weighted sum of individual security betas would, in my opinion, be problematic." - This is solvable.

In industry, it is common to use a proxy, that is - securities that can be used as a substitute for others that have problems such as limited price histories. For example, you could make a case for using a FAANG stock historical price time series as a proxy for a tech start-up WITH appropriate adjustments.

Not a perfect answer, considering that the selection of an appropriate proxy is quite an open question.

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.