Check Accuracy


Function Check Accuracy within
Yield Curve with keys Yield Curve Check Accuracy keys returns a diagnostics table that compares the zero rates Zerᴮ and forward rates Fwdᴮ produced with the Curve Booster facility against the corresponding exact zero rates Zerᴺᴮ and forward rates Fwdᴺᴮ produced without it as of v supplied maturities T₁, T₂, ..., Tᵥ under u fixed scenarios SC₁, SC₂, ..., SCᵤ of market quote shifts.
The superscript indicates Boost and the ᴺᴮ superscript indicates No Boost.
The description below corresponds to the table that is generated when
Full Report = TRUE
A much smaller, summary table may be also produced if Full Report = FALSE

Each row - indexed by i below - corresponds to a specific scenario SCᵢ, i = 1, 2, ..., u of market quote shifts that is described below in detail.
The first column contains labels that indicate the scenario applying on each row.
Then follow 6 groups of columns - each consisting of v columns - that report respectively:
Group 1: The zero rate discrepancies in bp (defined below) δZer(SCᵢ,Tⱼ)
Group 2: The forward rate discrepancies in bp (defined below) Fwdᴮ(SCᵢ,Tⱼ)
Group 3: The approximated zero rates Zerᴮ(SCᵢ,Tⱼ) produced when the boost facility is applied
Group 4: The exact zero rates Zerᴺᴮ(SCᵢ,Tⱼ) produced when the boost facility is not applied
Group 5: The approximated forward rates Fwdᴮ(SCᵢ,Tⱼ) produced when the boost facility is applied
Group 6: The exact forward rates Fwdᴺᴮ(SCᵢ,Tⱼ) produced when the boost facility is not applied

In particular:
The rate discrepancies δZer(SCᵢ,Tⱼ) and δFwd(SCᵢ,Tⱼ) are simply the differences in basis points between the two corresponding rates.
Formally:
δZer(SCᵢ,Tⱼ) = 10000[DZerᴮ(SCᵢ,Tⱼ) - DZerᴺᴮ(SCᵢ,Tⱼ)], i = 1, 2, ..., u and j = 1, 2, ..., v
and
δFwd(SCᵢ,Tⱼ) = 10000[DFwdᴮ(SCᵢ,Tⱼ) - DFwdᴺᴮ(SCᵢ,Tⱼ)], i = 1, 2, ..., u and j = 1, 2, ..., v

The scenarios SCᵢ, i = 1, 2, ..., u are constructed as follows:
The first scenario with i = 1 (top row) corresponds to a global flat shift, whereby all market quotes are parallelly shifted upwards by the same amount δ specified as input (in bps) to this function.
If the referenced curve contains links to K exogenous curves, the next row would correspond to a local flat shift, whereby only the referenced curve is parallelly shifted upwards by δ, with the following K rows dedicated to similar local flat shifts affecting only the corresponding exogenous curves.
The succeeding rows correspond to sub-local flat shifts whereby only the market quotes included in a particular group of instruments represented by an object of type
Yield Curve Input undergo an upwards parallel shift, while all other market quotes stay fixed.
First the instrument groups contained in the referenced curve are considered, followed by those contained in the linked exogenous curves, if any.
Finally, all remaining rows are dedicated to single market quote shifts, similar to a by-bucket DV01 output.
First, the referenced curve is considered so that each and every of its market quotes is independently shifted upwards by δ, while all other quotes stay fixed.
This process produces N rows, where N is the total number of market quotes involved either directly or indirectly in the construction of the referenced curve.
Then rows are added that correspond to a similar by-bucket treatment of the K linked exogenous curves, if any.