Generates a value at risk (VaR) quantile based on the Lower Partial Moment ratio.
Arguments
- percentile
numeric [0, 1]; The percentile for left-tail VaR.
- degree
integer;
(degree = 0)gives the interpolated empirical quantile (quantile(x, percentile, type = 7)), which on tied data need not be an exact inverse ofLPM.ratio(0, ., x). Use it for probability statements, such as a VaR exceeded with a given frequency.(degree = 1)gives an area-weighted (severity-weighted) threshold, thepercentile-expectile ofx(Newey and Powell 1987):percentileof the area of deviations, not of the observations, lies below it, so it is not a quantile ofx(for exponential data the degree-1 5% point has about a quarter of the observations below it). Higher degrees weight large deviations more.- x
a numeric vector.
Value
Returns the type-7 empirical quantile (degree = 0), whose strict or inclusive sample tail fraction may differ from percentile because of interpolation or ties, or the threshold with percentile of the directional deviation area below it (degree > 0).
References
Viole, F. and Nawrocki, D. (2013) "Nonlinear Nonparametric Statistics: Using Partial Moments" (ISBN: 1490523995, 2nd edition: https://ovvo-financial.github.io/NNS/book/)