Credit Risk - Time Series

Overview


A simple way to estimate credit loss is to aggregate credit losses to a set of periods (such as yearly) and then to apply the tools of time series analysis.

Sample loss data as a time series. The loss is stated as a rate, which when multiplied by the total portfolio value gives the loss for the period.


Basic Statistics


in general, loss data will need to assumed to be stationary so that one can simply apply a set of statistics. (see moments)

  • Average (Mean)
  • Standard Deviation

Implementation


The following sample code utilizes the moments library


let mt = await import('/lib/statistics/moments/v1.0.0/moments.mjs');

let average = mt.average(data.map(p=>p.rate));
let stdDev = Math.sqrt(mt.variance(data.map(p=>p.rate)));