In risk management average daily volume (ADV) is often used to gauge the liquidity of a security. To measure ADV, we calculate the average number of units traded per day over a given period.
As an example, if we want to calculate the ADV of stock XYZ, we could add the volume of shares of XYZ traded each day for the last 30 days, and divide that sum by 30.
You can use any lookback window, but spans that are close to one or three months are popular. Depending on the convention, these might correspond to 20, 21, 60, or 63 days. These can be either business days or weekdays. Weekends are rarely included in ADV calculations.
There are arguments to be made for using both the mean and the median in this calculation. The ambiguous term “average” allows for both possibilities. Daily volume can change dramatically from day to day. Often, there will be outliers, days when there is a significant spike in volume. If there is a spike and you calculate the mean, then your average could end up being significantly greater than the volume on most days. This might lead a portfolio manager to believe that a security was easier to buy or sell than it would be on most days. The median would not be significantly impacted by an isolated volume spike, and more closely reflect the volume seen most days. In the presence of spikes, the median will also be more stable over time. However, if spikes in volume are not uncommon, and a portfolio manager is able to buy or sell over a longer period, then the median might suggest that the total volume available is less than it actually is. There might also be reason to believe that the spike in volume is characteristic of the ultimate volume available. In other words, even if the volume is considerably less on most days, the spike might indicate how much potential volume is available. In practice, both the mean and median are used. There is no consensus on which is more appropriate. Risk managers should be aware of the pros and cons of both approaches.
