riskvalidation 0.8.0__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- riskvalidation/__init__.py +30 -0
- riskvalidation/aggregation/__init__.py +11 -0
- riskvalidation/aggregation/band.py +151 -0
- riskvalidation/aggregation/bounds.py +191 -0
- riskvalidation/aggregation/copulas.py +241 -0
- riskvalidation/aggregation/model_risk.py +133 -0
- riskvalidation/aggregation/simple.py +147 -0
- riskvalidation/ccr/__init__.py +13 -0
- riskvalidation/ccr/collateral.py +175 -0
- riskvalidation/ccr/cva.py +193 -0
- riskvalidation/ccr/exposure.py +192 -0
- riskvalidation/ccr/hw1f.py +426 -0
- riskvalidation/ccr/swaps.py +172 -0
- riskvalidation/core/__init__.py +1 -0
- riskvalidation/core/build.py +57 -0
- riskvalidation/core/hashing.py +48 -0
- riskvalidation/core/policy.py +130 -0
- riskvalidation/core/registry.py +65 -0
- riskvalidation/core/result.py +117 -0
- riskvalidation/core/suite.py +78 -0
- riskvalidation/engines/__init__.py +2 -0
- riskvalidation/engines/low_default.py +170 -0
- riskvalidation/engines/pd_curve.py +529 -0
- riskvalidation/engines/vasicek.py +72 -0
- riskvalidation/fairness/__init__.py +10 -0
- riskvalidation/fairness/_checks.py +115 -0
- riskvalidation/fairness/conditional.py +201 -0
- riskvalidation/fairness/inference.py +64 -0
- riskvalidation/fairness/metrics.py +261 -0
- riskvalidation/fairness/plants.py +240 -0
- riskvalidation/fairness/policy.py +97 -0
- riskvalidation/fairness/replication.py +39 -0
- riskvalidation/fairness/situation.py +307 -0
- riskvalidation/generators/__init__.py +68 -0
- riskvalidation/generators/_base.py +145 -0
- riskvalidation/generators/counts.py +232 -0
- riskvalidation/generators/drift.py +107 -0
- riskvalidation/generators/framingham.py +124 -0
- riskvalidation/generators/paths.py +349 -0
- riskvalidation/generators/ratings.py +186 -0
- riskvalidation/generators/scores.py +347 -0
- riskvalidation/harness/__init__.py +46 -0
- riskvalidation/harness/exact.py +137 -0
- riskvalidation/harness/performance.py +74 -0
- riskvalidation/harness/rates.py +103 -0
- riskvalidation/harness/run.py +197 -0
- riskvalidation/irrbb/__init__.py +9 -0
- riskvalidation/irrbb/behavioural.py +206 -0
- riskvalidation/irrbb/curves.py +286 -0
- riskvalidation/irrbb/eve_nii.py +245 -0
- riskvalidation/irrbb/shocks.py +171 -0
- riskvalidation/irrbb/simulation.py +203 -0
- riskvalidation/irrbb/slotting.py +184 -0
- riskvalidation/lgd_ead/__init__.py +8 -0
- riskvalidation/lgd_ead/_checks.py +50 -0
- riskvalidation/lgd_ead/ccf.py +228 -0
- riskvalidation/lgd_ead/downturn.py +64 -0
- riskvalidation/lgd_ead/floors.py +114 -0
- riskvalidation/lgd_ead/lgd_models.py +297 -0
- riskvalidation/lgd_ead/workout.py +162 -0
- riskvalidation/liquidity/__init__.py +9 -0
- riskvalidation/liquidity/_checks.py +78 -0
- riskvalidation/liquidity/cbc.py +158 -0
- riskvalidation/liquidity/deposits.py +151 -0
- riskvalidation/liquidity/ladder.py +119 -0
- riskvalidation/liquidity/reverse.py +141 -0
- riskvalidation/liquidity/runoff.py +220 -0
- riskvalidation/liquidity/scenarios.py +272 -0
- riskvalidation/market/__init__.py +8 -0
- riskvalidation/market/caviar.py +239 -0
- riskvalidation/market/conventions.py +83 -0
- riskvalidation/market/evt.py +111 -0
- riskvalidation/market/frtb.py +175 -0
- riskvalidation/market/garch.py +204 -0
- riskvalidation/market/gas_fz.py +219 -0
- riskvalidation/market/historical.py +96 -0
- riskvalidation/market/parametric.py +78 -0
- riskvalidation/market/scoring.py +145 -0
- riskvalidation/ml/__init__.py +11 -0
- riskvalidation/ml/_checks.py +67 -0
- riskvalidation/ml/_logistic.py +82 -0
- riskvalidation/ml/cbpe.py +91 -0
- riskvalidation/ml/challenger.py +73 -0
- riskvalidation/ml/conformal.py +116 -0
- riskvalidation/ml/dataquality.py +130 -0
- riskvalidation/ml/effects.py +67 -0
- riskvalidation/ml/explanations.py +125 -0
- riskvalidation/ml/generalisation.py +79 -0
- riskvalidation/ml/leakage.py +99 -0
- riskvalidation/ml/plants.py +425 -0
- riskvalidation/ml/policy.py +182 -0
- riskvalidation/ml/robustness.py +286 -0
- riskvalidation/ml/seeds.py +106 -0
- riskvalidation/oprisk/__init__.py +18 -0
- riskvalidation/oprisk/_checks.py +60 -0
- riskvalidation/oprisk/aggregate.py +375 -0
- riskvalidation/oprisk/distributions.py +758 -0
- riskvalidation/oprisk/scenarios.py +342 -0
- riskvalidation/oprisk/sla.py +80 -0
- riskvalidation/oprisk/sma.py +169 -0
- riskvalidation/policies/__init__.py +4 -0
- riskvalidation/policies/default.json +273 -0
- riskvalidation/portfolio/__init__.py +16 -0
- riskvalidation/portfolio/asrf_fit.py +283 -0
- riskvalidation/portfolio/copula.py +129 -0
- riskvalidation/portfolio/creditrisk_plus.py +232 -0
- riskvalidation/portfolio/euler.py +126 -0
- riskvalidation/portfolio/granularity.py +135 -0
- riskvalidation/portfolio/importance.py +277 -0
- riskvalidation/portfolio/portfolio.py +309 -0
- riskvalidation/provisions/__init__.py +8 -0
- riskvalidation/provisions/_checks.py +100 -0
- riskvalidation/provisions/ecl.py +233 -0
- riskvalidation/provisions/local_cl.py +360 -0
- riskvalidation/provisions/local_es.py +157 -0
- riskvalidation/provisions/overlays.py +102 -0
- riskvalidation/provisions/reconciliation.py +157 -0
- riskvalidation/provisions/staging.py +211 -0
- riskvalidation/provisions/term_structure.py +199 -0
- riskvalidation/regulatory/__init__.py +14 -0
- riskvalidation/regulatory/ba_cva.py +138 -0
- riskvalidation/regulatory/frtb.py +345 -0
- riskvalidation/regulatory/irb.py +314 -0
- riskvalidation/regulatory/lcr_nsfr.py +369 -0
- riskvalidation/regulatory/output_floor.py +45 -0
- riskvalidation/regulatory/sa_ccr.py +486 -0
- riskvalidation/regulatory/standardised.py +348 -0
- riskvalidation/stress/__init__.py +9 -0
- riskvalidation/stress/capital.py +476 -0
- riskvalidation/stress/paths.py +108 -0
- riskvalidation/stress/reverse.py +179 -0
- riskvalidation/stress/satellites.py +268 -0
- riskvalidation/stress/scenario_models.py +274 -0
- riskvalidation/transitions/__init__.py +0 -0
- riskvalidation/transitions/em.py +120 -0
- riskvalidation/transitions/embedding.py +221 -0
- riskvalidation/transitions/estimators.py +213 -0
- riskvalidation/transitions/intervals.py +306 -0
- riskvalidation/transitions/mobility.py +57 -0
- riskvalidation/transitions/ttc.py +171 -0
- riskvalidation/validation/__init__.py +103 -0
- riskvalidation/validation/_common.py +57 -0
- riskvalidation/validation/aggregation.py +425 -0
- riskvalidation/validation/calibration.py +757 -0
- riskvalidation/validation/ccr.py +530 -0
- riskvalidation/validation/discrimination.py +314 -0
- riskvalidation/validation/fairness.py +632 -0
- riskvalidation/validation/ifrs9.py +693 -0
- riskvalidation/validation/irrbb.py +389 -0
- riskvalidation/validation/lgd_ead.py +469 -0
- riskvalidation/validation/liquidity.py +647 -0
- riskvalidation/validation/market.py +807 -0
- riskvalidation/validation/ml.py +958 -0
- riskvalidation/validation/oprisk.py +357 -0
- riskvalidation/validation/portfolio.py +353 -0
- riskvalidation/validation/rating_system.py +245 -0
- riskvalidation/validation/stability.py +256 -0
- riskvalidation/validation/stress.py +552 -0
- riskvalidation/validation/transitions.py +386 -0
- riskvalidation-0.8.0.dist-info/METADATA +136 -0
- riskvalidation-0.8.0.dist-info/RECORD +164 -0
- riskvalidation-0.8.0.dist-info/WHEEL +5 -0
- riskvalidation-0.8.0.dist-info/licenses/LICENSE +21 -0
- riskvalidation-0.8.0.dist-info/top_level.txt +1 -0
|
@@ -0,0 +1,170 @@
|
|
|
1
|
+
"""Most prudent estimation of PDs for low-default portfolios (Pluto and Tasche).
|
|
2
|
+
|
|
3
|
+
Grades are ordered best first. The PD of grade i is estimated as an upper confidence bound at level gamma under the
|
|
4
|
+
most prudent assumption that every grade from i downward has the same PD, so the obligors and defaults of grades i to
|
|
5
|
+
k are pooled (Pluto and Tasche 2005, arXiv cond-mat/0411699, sections 2 and 3). With N pooled obligors and D pooled
|
|
6
|
+
defaults, independent defaults, the bound p solves
|
|
7
|
+
|
|
8
|
+
P(Binomial(N, p) <= D) = 1 - gamma, (zero defaults: p = 1 - (1 - gamma)^(1/N)),
|
|
9
|
+
|
|
10
|
+
the Clopper-Pearson upper bound. With asset correlation rho in the one-factor model (section 4, eq. 4.1) the binomial
|
|
11
|
+
probability is averaged over the systematic factor, computed here by Gauss-Hermite quadrature:
|
|
12
|
+
|
|
13
|
+
1 - gamma = integral of phi(y) P(Binomial(N, p(y)) <= D) dy, p(y) = N((G(p) - sqrt(rho) y) / sqrt(1 - rho)).
|
|
14
|
+
|
|
15
|
+
The bounds are monotone across grades by construction. The paper notes that the choice of gamma is arbitrary and
|
|
16
|
+
that levels below 95% "seem more appropriate"; the level is the caller's.
|
|
17
|
+
"""
|
|
18
|
+
|
|
19
|
+
from __future__ import annotations
|
|
20
|
+
|
|
21
|
+
import math
|
|
22
|
+
from collections.abc import Sequence
|
|
23
|
+
|
|
24
|
+
import numpy as np
|
|
25
|
+
from scipy import optimize, stats
|
|
26
|
+
from scipy.special import ndtr, ndtri
|
|
27
|
+
|
|
28
|
+
PLUTO_TASCHE = (
|
|
29
|
+
"Pluto and Tasche (2005), Estimating probabilities of default for low default portfolios, arXiv cond-mat/0411699"
|
|
30
|
+
)
|
|
31
|
+
|
|
32
|
+
_NODES, _WEIGHTS = np.polynomial.hermite_e.hermegauss(256)
|
|
33
|
+
_WEIGHTS = _WEIGHTS / math.sqrt(2.0 * math.pi)
|
|
34
|
+
|
|
35
|
+
|
|
36
|
+
def _pooled(obligors: Sequence[int], defaults: Sequence[int]) -> tuple[np.ndarray, np.ndarray]:
|
|
37
|
+
n = np.asarray(obligors, dtype=int)
|
|
38
|
+
d = np.asarray(defaults, dtype=int)
|
|
39
|
+
if n.shape != d.shape or n.ndim != 1 or len(n) == 0:
|
|
40
|
+
raise ValueError("obligors and defaults are one number per grade")
|
|
41
|
+
if np.any(n <= 0) or np.any(d < 0) or np.any(d > n):
|
|
42
|
+
raise ValueError("each grade needs obligors > 0 and 0 <= defaults <= obligors")
|
|
43
|
+
# grade i pools grades i..k (best first, so the pool grows towards the best grade)
|
|
44
|
+
return np.cumsum(n[::-1])[::-1], np.cumsum(d[::-1])[::-1]
|
|
45
|
+
|
|
46
|
+
|
|
47
|
+
def _independent_bound(n: int, d: int, gamma: float) -> float:
|
|
48
|
+
if d >= n:
|
|
49
|
+
return 1.0
|
|
50
|
+
return float(stats.beta.ppf(gamma, d + 1, n - d))
|
|
51
|
+
|
|
52
|
+
|
|
53
|
+
def _correlated_bound(n: int, d: int, gamma: float, rho: float) -> float:
|
|
54
|
+
target = 1.0 - gamma
|
|
55
|
+
|
|
56
|
+
def tail(p: float) -> float:
|
|
57
|
+
cond = ndtr((ndtri(p) - math.sqrt(rho) * _NODES) / math.sqrt(1.0 - rho))
|
|
58
|
+
return float(np.sum(_WEIGHTS * stats.binom.cdf(d, n, cond))) - target
|
|
59
|
+
|
|
60
|
+
lo, hi = 1e-12, 1.0 - 1e-12
|
|
61
|
+
return float(optimize.brentq(tail, lo, hi, xtol=1e-15, rtol=1e-12, maxiter=500))
|
|
62
|
+
|
|
63
|
+
|
|
64
|
+
def most_prudent_pd(obligors: Sequence[int], defaults: Sequence[int], gamma: float, rho: float = 0.0) -> dict:
|
|
65
|
+
"""Most prudent PD bounds per grade (best first) at confidence ``gamma``, independent or with asset correlation
|
|
66
|
+
``rho``. Returns the bounds and the pooled counts each one rests on."""
|
|
67
|
+
if not 0.0 < gamma < 1.0:
|
|
68
|
+
raise ValueError("gamma must lie in (0, 1)")
|
|
69
|
+
if not 0.0 <= rho < 1.0:
|
|
70
|
+
raise ValueError("rho must lie in [0, 1)")
|
|
71
|
+
big_n, big_d = _pooled(obligors, defaults)
|
|
72
|
+
if rho == 0.0:
|
|
73
|
+
bounds = [_independent_bound(int(n), int(d), gamma) for n, d in zip(big_n, big_d, strict=True)]
|
|
74
|
+
else:
|
|
75
|
+
bounds = [_correlated_bound(int(n), int(d), gamma, rho) for n, d in zip(big_n, big_d, strict=True)]
|
|
76
|
+
return {
|
|
77
|
+
"pd": bounds,
|
|
78
|
+
"pooled_obligors": [int(v) for v in big_n],
|
|
79
|
+
"pooled_defaults": [int(v) for v in big_d],
|
|
80
|
+
"gamma": gamma,
|
|
81
|
+
"rho": rho,
|
|
82
|
+
"reference": PLUTO_TASCHE,
|
|
83
|
+
}
|
|
84
|
+
|
|
85
|
+
|
|
86
|
+
def scale_bounds(bounds: Sequence[float], obligors: Sequence[int], target: float) -> dict:
|
|
87
|
+
"""Section 5, equations 5.1 and 5.2: one factor K for every grade, such that the obligor-weighted average of the
|
|
88
|
+
scaled bounds equals ``target``. Scaling keeps the bounds' ordering and their ratios (the "relative" PDs the most
|
|
89
|
+
prudent principle derives) and gives up their interpretation as confidence bounds."""
|
|
90
|
+
b = np.asarray(bounds, dtype=float)
|
|
91
|
+
n = np.asarray(obligors, dtype=float)
|
|
92
|
+
if b.shape != n.shape or b.ndim != 1 or np.any(n <= 0) or np.any(b <= 0):
|
|
93
|
+
raise ValueError("bounds and obligors are one positive value per grade")
|
|
94
|
+
if not 0.0 < target < 1.0:
|
|
95
|
+
raise ValueError("the scaling target must lie in (0, 1)")
|
|
96
|
+
k = target * float(n.sum()) / float(np.sum(b * n))
|
|
97
|
+
return {"pd": [float(v) for v in k * b], "k": k, "target": float(target),
|
|
98
|
+
"reference": f"{PLUTO_TASCHE}, section 5, equations 5.1 and 5.2"}
|
|
99
|
+
|
|
100
|
+
|
|
101
|
+
def most_prudent_pd_scaled(obligors: Sequence[int], defaults: Sequence[int], gamma: float, rho: float = 0.0, *,
|
|
102
|
+
target: str | float = "upper_bound") -> dict:
|
|
103
|
+
"""The most prudent bounds scaled by section 5: to the portfolio's observed default rate (``"central_tendency"``,
|
|
104
|
+
Tables 9 and 10; needs at least one default), to the most prudent bound of the whole portfolio at the same level,
|
|
105
|
+
which is the best grade's bound (``"upper_bound"``, Tables 11 and 12, the paper's proposal), or to a given PD."""
|
|
106
|
+
raw = most_prudent_pd(obligors, defaults, gamma, rho)
|
|
107
|
+
n, d = np.asarray(obligors, dtype=float), np.asarray(defaults, dtype=float)
|
|
108
|
+
if target == "central_tendency":
|
|
109
|
+
if d.sum() <= 0:
|
|
110
|
+
raise ValueError("scaling to the central tendency needs at least one default (section 5)")
|
|
111
|
+
level = float(d.sum() / n.sum())
|
|
112
|
+
elif target == "upper_bound":
|
|
113
|
+
level = raw["pd"][0]
|
|
114
|
+
elif isinstance(target, (int, float)) and not isinstance(target, bool):
|
|
115
|
+
level = float(target)
|
|
116
|
+
else:
|
|
117
|
+
raise ValueError('target is "central_tendency", "upper_bound" or a PD in (0, 1)')
|
|
118
|
+
scaled = scale_bounds(raw["pd"], obligors, level)
|
|
119
|
+
return {**raw, "unscaled": raw["pd"], "pd": scaled["pd"], "k": scaled["k"], "target": level,
|
|
120
|
+
"target_kind": target if isinstance(target, str) else "given",
|
|
121
|
+
"reference": f"{PLUTO_TASCHE}, sections 3 to 5"}
|
|
122
|
+
|
|
123
|
+
|
|
124
|
+
def _multiperiod_tail(p: float, n: int, k: int, rho: float, factors: np.ndarray) -> float:
|
|
125
|
+
"""(6.8b): P[at most k of n obligors default within the period], averaged over the simulated factor paths."""
|
|
126
|
+
cond = ndtr((ndtri(p) - math.sqrt(rho) * factors) / math.sqrt(1.0 - rho))
|
|
127
|
+
pi = 1.0 - np.prod(1.0 - cond, axis=1) # (6.8a): an obligor's probability to default within the period
|
|
128
|
+
return float(np.mean(stats.binom.cdf(k, n, pi)))
|
|
129
|
+
|
|
130
|
+
|
|
131
|
+
def most_prudent_pd_multiperiod(obligors: Sequence[int], defaults: Sequence[int], gamma: float, *, rho: float,
|
|
132
|
+
theta: float, years: int, n_sim: int = 100_000, seed: int = 0) -> dict:
|
|
133
|
+
"""Section 6: most prudent bounds from a cohort observed for ``years`` years, with ``defaults`` the defaults of
|
|
134
|
+
each grade over the whole period. The systematic factors of the years are standard normal with correlation
|
|
135
|
+
theta^|s - t| (6.5b); obligor i defaults in year t when sqrt(rho) S_t + sqrt(1 - rho) xi_it first falls below
|
|
136
|
+
Phi^{-1}(p) (6.1 to 6.4), so p is a long-run one-year PD. Given the factors an obligor defaults within the period
|
|
137
|
+
with probability 1 - prod_t (1 - p(S_t)) (6.7, 6.8a), and the bound is the largest p with
|
|
138
|
+
P[at most k defaults] >= 1 - gamma (6.6, 6.8b). The expectation over the factors is a Monte Carlo average over
|
|
139
|
+
``n_sim`` paths drawn with ``seed``, the same paths for every p (so the bound is a smooth function of p)."""
|
|
140
|
+
if not 0.0 < gamma < 1.0:
|
|
141
|
+
raise ValueError("gamma must lie in (0, 1)")
|
|
142
|
+
if not 0.0 < rho < 1.0:
|
|
143
|
+
raise ValueError("rho must lie in (0, 1)")
|
|
144
|
+
if not 0.0 <= theta < 1.0:
|
|
145
|
+
raise ValueError("theta must lie in [0, 1)")
|
|
146
|
+
if years < 1 or n_sim < 1000:
|
|
147
|
+
raise ValueError("years >= 1 and n_sim >= 1000")
|
|
148
|
+
big_n, big_d = _pooled(obligors, defaults)
|
|
149
|
+
t = np.arange(years)
|
|
150
|
+
corr = theta ** np.abs(np.subtract.outer(t, t)).astype(float)
|
|
151
|
+
factors = np.random.default_rng(seed).standard_normal((n_sim, years)) @ np.linalg.cholesky(corr).T
|
|
152
|
+
bounds = []
|
|
153
|
+
for n, d in zip(big_n, big_d, strict=True):
|
|
154
|
+
if d >= n:
|
|
155
|
+
bounds.append(1.0)
|
|
156
|
+
continue
|
|
157
|
+
bounds.append(float(optimize.brentq(lambda p, n=int(n), d=int(d): _multiperiod_tail(p, n, d, rho, factors)
|
|
158
|
+
- (1.0 - gamma), 1e-12, 1.0 - 1e-12, xtol=1e-15, rtol=1e-12)))
|
|
159
|
+
return {
|
|
160
|
+
"pd": bounds,
|
|
161
|
+
"pooled_obligors": [int(v) for v in big_n],
|
|
162
|
+
"pooled_defaults": [int(v) for v in big_d],
|
|
163
|
+
"gamma": gamma,
|
|
164
|
+
"rho": rho,
|
|
165
|
+
"theta": theta,
|
|
166
|
+
"years": int(years),
|
|
167
|
+
"n_sim": int(n_sim),
|
|
168
|
+
"seed": int(seed),
|
|
169
|
+
"reference": f"{PLUTO_TASCHE}, section 6, equations 6.1 to 6.8",
|
|
170
|
+
}
|