When calculating the portfolio's historical volatility, do I need to factor the volatility of each asset individually and then do a covariance calculation or can I get away by measuring the volatility of the total portfolio return?

My gut feeling says no, but I cant work out why?

  • $\begingroup$ Historical or implied portfolio volatility? $\endgroup$
    – Frido
    Commented Apr 4, 2023 at 16:24
  • $\begingroup$ Historical, edited the question $\endgroup$
    – user1234
    Commented Apr 4, 2023 at 16:37
  • 3
    $\begingroup$ In that case I would just calculate the potfolio vol as a whole. You'll just invite trouble if you calc the individual funds' or stocks' vol and then try to combine them using some covar matrix. But if you want to analyze the contribution of the component VaRs to the portfolio VaR then you of course need to calc the separate vols. $\endgroup$
    – Frido
    Commented Apr 4, 2023 at 17:02
  • $\begingroup$ Something to consider: Portfolios often have rules-based allocations (e.g., large-cap only). If a stock increases in size from mid-cap to large-cap, it may enter the portfolio when it wouldn't be eligible before. In this case, including historical vols of that asset in your portfolio's vol may not be a good idea, as you're including vols from when the stock was a mid-cap and therefore inelegible to be in the portfolio. $\endgroup$
    – Rylan
    Commented Sep 5, 2023 at 10:41
  • $\begingroup$ Both are fine. Everything you measure consists of thousands of factors, you don't consider them when measuring the statistics of the interest. $\endgroup$
    – Arshdeep
    Commented May 1 at 18:11


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.