Volatility, a fickle measure of riskiness

Volatility, a fickle measure of riskiness

by Pininvest Analysis

pinindices - blog on pininvest.com

  • 0 constituents
  • 0.0% 1y performance
  • 0.0% volatility
Check the investment theme exit_to_app

Volatility is difficult to apprehend

  • as a measure of risk
  • and as an average


As a statistical measure

Volatility is a measure of dispersion of returns of a stock, updated daily on the pininvest database

Widely fluctuating returns over time create uncertainty regarding a stock’s value – increasing the asset’s risk compared to stocks with more steady returns

Comparing top and bottom segments of Total Market 1 200 Index

Consequently, at any point in time, the absolute volatility of a stock, measured in %, gains significance relative to the volatility of other assets, making a stock more or less risky compared to others


As an average

To complicate matters, the volatility of returns of a stock between day 1 and day 2 carries very little information – in fact none – about the risk of the stock relative to others

It will be understood that rich information can only be extracted from a sample of successive daily returns: the larger the sample, the more reliable the information about actual ‘riskiness’ will become …

On pinindices, the volatility charts are always calculated over a time frame of 6 months, which delivers approx. 125 return data points

But the need of a large data ‘pool’ must be balanced against the ‘ghost’ effect of a major shift in volatility which will impact volatility for months following the event – even though daily measures of volatility return in more normal ranges

Datavariance 1 200 index with sharp volatility run-up on Feb. 1, slowly fading 

The effect can be seen clearly on the chart of the volatility run-up in February ’18, a brutal increase which only ‘fades’ slowly in the average volatility measures

On pinindices, the ‘ghost’ effect of major volatility shifts is being rebalanced to some extent by assigning overweight to the most recent data in the series, which ‘fades’ the return volatilities of the more distant past. The statistical calculation is an exponential weighted moving average (EWMA)