O/S from Johnson and So (2012) should filter volume by time to expiration

[travlake]

From what I can see, the current code uses all option volume when computing the O/S predictor from Johnson and So (2012). This is not how the original paper calculated option volume.

From Page 268 of the paper "Specifically, OPVOLi,w equals the total volume in option contracts across all strikes for options expiring in the 30 trading days beginning five days after the trade date."

Johnson So OS 2012 (JFE).pdf

[chenandrewy]

Thanks so much Travis for letting us know about this! The issue is resolved here 8c570ae. The key update is this file: https://github.com/OpenSourceAP/CrossSection/blob/8c570ae16fc12e1ab366e74060817b6d557940a1/Signals/pyCode/PrepScripts/OptionMetricsVolume.R, which applies the following filters following your paper:

  # 30 trading days after 5 days after trade date
  query = paste0(
    "select secid, date, exdate, volume
    from optionm.opprcd", year, "
    where cp_flag != 'NaN' and
      exdate - date >= 5 and
      exdate - date <= 5+30*(365/252) 
    limit ", querylimit
  )

[chenandrewy]

Actually fixed here fa75cbe