Model methodsSystemic riskactive

Marginal Expected Shortfall

An institution's average return on days when the system return is in its own 5% left tail.

Registry slug:: mes
Visible surface:: /systemic

Data and implementation

Vintage

38,561 observations for 28 entities from 2021-01-04 through 2026-07-06, measured 2026-08-30

Data tables

data/parquet/systemic_series.parquet

Engine

engine/finweave_engine/layers/systemic/mes.py

Producer

scripts/build_systemic.py

Outputs

data/parquet/systemic_series.parquet

Source: Yahoo Finance equity returns via the argus collector, internal-use input

Method

Within each trailing 252-trading-day window, the engine finds days when the system proxy return is at or below its 5% quantile and averages the institution's aligned returns on those days.

A more negative MES means the institution lost more on the system's tail days. The stored method label is rolling_tail_mean.

Methodology evidence

Path and lines	Evidence
engine/finweave_engine/layers/systemic/mes.py:11-45	Selects system-tail days and returns the mean institution return over those observations.
docs/systemic_methodology.md:135-146	Defines MES, its sign interpretation, rolling window, quantile, method label, and span.
docs/systemic_methodology.md:269-338	Records crisis anchors and independent recomputation validation for the systemic output.

Equations

MESMES = mean of bank returns on days when system return falls below its q-th percentile

Validation

The parquet contains 38,561 MES observations for 28 entities from 2021-01-04 through 2026-07-06.
The documented independent recomputation of JPM MES on 2026-07-06 matches the stored value with zero difference, and the SVB-week cross-section localizes the episode toward affected regional banks.

Limitations

MES is a conditional historical return average, not a probability of distress or failure.
Results depend on the KBE system proxy, the 5% tail threshold, the covered listed-bank universe, and the trailing 252-day window.
The underlying Yahoo-derived equity-return input carries an internal-use, display-aggregates-only licence posture and is not a redistributable raw dataset.

References

Acharya, Pedersen, Philippon, and Richardson (2017), Measuring Systemic Risk, Review of Financial Studies 30(1): 2-47.

Metadata endpoint

Path	Method	Returns	Example
/api/methods/[slug]	GET	Registry metadata, implementation paths, measured vintage, methodology evidence, validation, and limitations. No model observations or parquet contents.	/api/methods/mes