Skip to main content

Solvency 2

A granular, offline Solvency II calculation toolkit by RiskDataScience GmbH. First version: 0.5.0 (0.5 release). Repository: rds0001/solvency2-python. Use individual formulas in scripts and notebooks, supply your own data, inspect parameters and correlations, and retain the source and scenario trail of each calculation. A complete insurer model or workflow is not required.

What this makes easier

Solvency II work involves more than producing a single capital number. Analysts need reusable calculations, regulatory tables, controlled assumptions and a way to explain how each result was obtained. This library brings those building blocks together in one Python toolkit:

  • Calculate just what you need. Call a single shock, valuation, aggregation, own-funds or MCR function without setting up an application or insurer model.
  • Use regulatory tables without transcribing them yourself. Inspect included parameters, labelled correlation matrices and supported catastrophe tables, with their source references and explicit rule profiles.
  • Work with your own data and assumptions. Load local curves and supply cashflows, exposures or credit inputs. Change supported parameters and matrix entries per scenario, without modifying installed files or global defaults.
  • Explain and reproduce a result. Calculations return the value, units, formula identifier, sources, context, input hashes and applicable override records—not just an unexplained number.
  • Start from runnable examples. Explore 12,382 fixed reference cases and 227 integration portfolios, then adapt the documented inputs to your own work.
  • Connect related tasks. Use components for technical provisions, market and underwriting risks, counterparty risk, capital resources, groups and reporting in the same toolkit. Compose selected functions into a dependency workflow when useful; direct function calls remain the primary interface.

The scope is 399 calculation functions across 21 domain modules, plus local input, scenario and discovery helpers. Explicit baseline and 2027 rule profiles support version-aware work within their documented method coverage. Reporting resources and checks complement the calculation APIs; they are not an automatic filing service. All user documentation is also available from an installed wheel.

For actuarial analysts, risk teams, researchers and developers, the practical benefit is less repeated formula/table plumbing and a shared, inspectable basis for notebooks, sensitivity analyses, validation and internal tools. The runtime is fully offline: no provider account, hosted engine or telemetry is required. Source code is GPL-3.0-or-later; data attributions are documented separately.

Included scope

The package contains the accepted calculation baseline: 399 calculations, 421 inherited public symbols, 12,382 fixed core reference cases and 227 integration portfolios, plus additional reporting references. It preserves the baseline's explicit limitations. This is not regulatory certification, an internal model, an automatic filing service, or a claim of complete coverage of every legal requirement. See scope and limitations.

Market-data edition: synthetic examples only. No historical provider market archive is bundled. Demo lookups issue a warning and return SYNTHETIC_DEMO; their downstream workflow results retain that label. Supply your own local data for production calculations. Regulatory parameters and correlations are retained.

Installation

Python 3.10 or later is required; Python 3.10 is the tested release environment. PyPI package name: solvency2. For a published release:

python -m pip install solvency2

The import is also solvency2; the GitHub repository remains solvency2-python. Alternatively install from a checkout (pip obtains build tooling if needed):

git clone https://github.com/rds0001/solvency2-python.git
cd solvency2-python
python -m pip install .

For an offline installation, first build or obtain the wheel as described in local build instructions, then install it without an index:

python -m pip install --no-index --no-deps dist/solvency2-0.5.0-py3-none-any.whl

The base runtime has no third-party dependencies. Only designated numerical methods need SciPy (optimization / statistics extras); install it separately from a trusted local wheel collection. Installation is distinct from execution: the library does not download data, contact providers, check for updates or send telemetry. It needs neither an external engine installation nor R.

One calculation

from solvency2 import Context
from solvency2.capital import intangible_risk

context = Context(
    as_of="2025-01-17", known_at="2026-09-23", currency="EUR",
    profile_id="EU-S2-DOCUMENTS-2024-2025-v0.1",
)
result = intangible_risk(100, context=context)
assert result.value == 80
print(result.unit, result.sources, result.status)
audit_record = result.to_dict()

The 80 is a capital charge in EUR, not a valuation of the original asset. Dates and the rule profile are explicit. known_at expresses your chosen information cutoff, not the current clock time. REFERENCE_COMPONENT means a qualified software component result, not supervisory approval.

Your own curve, including dates outside the bundled archive

from solvency2 import Context
from solvency2.data import DiscountCurve, best_estimate_with_curve

context = Context("2028-12-31", "2028-12-31", "EUR", "EU-S2-DOCUMENTS-2027-v0.1")
curve = DiscountCurve(
    curve_id="synthetic-example", as_of=context.as_of, currency="EUR",
    curve_kind="base", source_reference="SYNTHETIC: not market data",
    points=(("2029-12-31", 0.95),),
)
result = best_estimate_with_curve(
    {"2029-12-31": {"expected_inflows": 0, "expected_outflows": 100}},
    curve=curve, runoff_complete=True, cashflow_reference="SYNTHETIC: complete toy runoff",
    context=context,
)
assert result.value == 95
assert result.details["user_curve"]["input_hash"] == curve.input_hash

Load the same factors with load_discount_curve_json(path) or load_discount_curve_csv(path, ...). Explicit annual-effective spot rates and year fractions are also supported. There is no automatic interpolation or fallback to an archived curve. Read input formats.

Change a parameter without editing installed files

from dataclasses import replace
from solvency2 import Context, NumericOverride, intangible_risk, parameter

base = Context("2025-01-17", "2026-09-23", "EUR", "EU-S2-DOCUMENTS-2024-2025-v0.1")
original = parameter("scalars/intangible", context=base)
scenario = replace(base, overrides=(NumericOverride(
    "profile.json", ("scalars", "intangible", "value"), 0.70, "Synthetic sensitivity",
),))
changed = intangible_risk(100, context=scenario)
assert changed.value == 70 and changed.status == "ANALYST_SCENARIO"
assert intangible_risk(100, context=base).value == 80

Matrix entries, supported table cells and JSON scenario batches follow the same principle. Read parameter overrides for paths, automatic matrix symmetry, validation and audit details.

Find the right function

Analyst task Public modules
Valuation, cashflows, best estimate, risk margin valuation, provisions, rates, reinsurance
SCR shocks and aggregation market, counterparty, life, health, nonlife, capital
Capital resources and MCR own_funds, mcr
Group calculations, model inputs and governance evidence groups, models, governance
Reporting components and reconciliation reporting, audit
Context, numeric parameters and synthetic data context, parameters, reference, catalogue
Local curves and scenario batches data, scenarios
Optional composition of existing formulas workflow

All inherited functions are also available directly from solvency2. The function reference contains signatures, method descriptions and fixed reference examples. solvency2.catalogue.api_catalogue() supplies the same navigation as structured data. Recipes show how to combine components; executable scripts and local files are in examples. Installed-wheel users can read the guides entirely offline with from solvency2.catalogue import documentation; print(documentation("DATA")). Use documentation("API") for the full function reference.

Reference data and reproducibility

import solvency2 as s2
from solvency2.catalogue import reference_datasets, reference_dataset

cases = s2.reference_cases()["cases"]
portfolios = s2.reference_portfolios()["portfolios"]
extra = reference_dataset(reference_datasets()[0])
# Explicitly write a new directory you own; existing directories are rejected:
# s2.export_resources("my-reference-export")

Reference expected values are independently specified arithmetic fixtures, not outputs generated to match the implementation. The demonstration market grids are invented from simple constants, not transformations of observed data. Regulatory parameter tables are not synthetic. Select profiles and reporting versions deliberately. Users qualify their own inputs for the selected purpose.

Documentation and local development

License: GPL-3.0-or-later, see LICENSE and NOTICE. Author/contact: RiskDataScience GmbH, riskdatascience@web.de. See third-party attributions and data licenses for the EU regulatory material and EIOPA reporting extracts. Market examples and original library code are by RiskDataScience GmbH. Warranty terms are in LICENSE.

Metadata

Release files for solvency2 0.5.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for solvency2 0.5.0
File Size Uploaded Provenance Actions
solvency2-0.5.0.tar.gz 16.6 MB With provenance Details for solvency2-0.5.0.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for solvency2 0.5.0
File Interpreter ABI Platform Actions
solvency2-0.5.0-py3-none-any.whl Python 3 none any Details for solvency2-0.5.0-py3-none-any.whl

Total release size: 33.3 MB

Release files / solvency2-0.5.0.tar.gz

Download URL solvency2-0.5.0.tar.gz
Size 16.6 MB
Tags Source
SHA-256 checksum
How to use checksums
7930ef1a034becc078e1121d5bd05ac5d9ca0a2f29702d1b4aa0b0b3b371969d
BLAKE2b-256 checksum
How to use checksums
7ece5d8ea9fd84832be4307361840f33ed5790decfecc0a28c525cb3eaddf207
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 6, 2026.

Transparency log

Release files / solvency2-0.5.0-py3-none-any.whl

Download URL solvency2-0.5.0-py3-none-any.whl
Size 16.7 MB
Tags Python 3
SHA-256 checksum
How to use checksums
91c76bd72500d78dfe08d6d7d4135a290d43c97dff957da3e2866eff67c30657
BLAKE2b-256 checksum
How to use checksums
5a68ff0afd28a7ed987e30844b9a9518dccbf7221aece7b134d4388d9bf04e09
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 6, 2026.

Transparency log

Security for solvency2 0.5.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

Both files in this release have been uploaded with provenance metadata.

Select from 2 files

Attestation for solvency2-0.5.0.tar.gz

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 6, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release

0.5.0 This release

2 release files Full provenance

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page