The two files are:

  • distribute.R: illustrates the use of the csv data file.
  • distribute.csv.gz: the data file.

    From memory, here are the fields:

    yyyymmddnow
    date
    call based data (.c)
    meanivolday.c
    average
    p95.c,n05.c,Nd.c,Nat0.c,ivol.at.K0.c
    N is number of options used in calc. p95.c and n05.c are probably the options sitting at the 5th and 95th percentile for the at-the-money options (which are counted as such if they are within 2% of the current strikeprice).
    ivol.at.Km16.c,ivol.at.Km16b.c,code.c
    put based data (.p)
    meanivolday.p
    p95.p,n05.p,Nd.p,Nat0.p,ivol.at.K0.p
    ivol.at.Km16.p,ivol.at.Km16b.p,code.p
    x,logsigma64
    historical volatility
    Most of the data is superfluous (and my description is from memory, so it may be faulty), but the interesting series is ivol.at.Km16.p, which can be viewed as left-tail disaster risk.

The data was created for Ivo Welch. The (Time-Varying) Importance of Disaster Risk. Financial Analysts Journal 72-5, Sep/Oct 2016. doi: 10.2469/faj.v72.n5.3.


/professional/tailrisk/ Last modified: