normal.test {PDtoolkit}

Multi-period predictive power test

Description

normal.test performs multi-period testing of PD model predictive power. This procedure can be applied on the level of the rating grade as well on the portfolio level.

Usage

normal.test(pdc, odr, alpha = 0.05)

Arguments

pdc

Numeric vector of calibrated probabilities of default (PD).

odr

Numeric vector of observed default rates.

alpha

Significance level of p-value for implemented tests. Default is 0.05.

Value

The command normal.test returns a data frame with estimated difference between odr and pdc, test statistics, standard error of the test statistics, selected significance level, p-value of test statistics and finally the test results.

References

Basel Committee on Banking Supervision (2005). Studies on the Validation of Internal Rating Systems, working paper no. 14.

Examples

set.seed(678)
normal.test(pdc = rep(0.02, 5), 
odr = runif(5, 0.02, 0.03))