Cross Sectional Mispricing

Quick Reference

Property	Value
Dimension	regime
Category	market_regime
Version	v0.9.0 (Beta)
Output Column	mispricing_score

Cross-sectional mispricing score - measures relative mispricing of an asset compared to its peer group

Formula

CDM Inputs

Column	CDM Table	Description
asset_return	cdm_*	CDM source table
peer_returns	cdm_*	CDM source table

Parameters

Parameter	Type	Default	Description
epsilon	float [0.0, 1.0]	1e-10	Small constant to prevent division by zero

Output

Column: mispricing_score

Deviation from cross-sectional mean return (z-score)

Market Intuition & Trading Rationale

Cross-sectional mispricing measures an asset's return deviation from its peer group: (asset_return - peer_mean) / (peer_std + ε). This is the z-score of the asset's return relative to its cross-section. A positive z-score means the asset outperformed its peers (potentially overvalued). A negative z-score means it underperformed (potentially undervalued). This is the foundation of statistical arbitrage and pairs trading strategies.

Usage Cases

• Mean reversion across peers: Go long assets with negative z-scores (underperformed, expected to catch up) and short assets with positive z-scores (overperformed, expected to revert). This is the classic long-short equity stat-arb strategy.
• Pairs trading entry: When the z-score of the spread between two cointegrated assets is extreme, enter a convergence trade. Exit when the z-score returns to zero.

YAML Definition

name: cross_sectional_mispricing
description: Cross-sectional mispricing score - measures relative mispricing of an
  asset compared to its peer group
category: market_regime
version: v0.9.0 (Beta)
dimension: regime
status: Pre-release
required_inputs:
  - asset_return
  - peer_returns
output_column: mispricing_score
output_description: Deviation from cross-sectional mean return (z-score)
parameters:
  epsilon:
    type: float
    description: Small constant to prevent division by zero
    required: false
    default: 1.0e-10
    constraints:
      min: 0.0
      max: 1.0
steps:
  - primitive: TRANSFORM
  op: array_mean
  params:
    array: peer_returns
  output: peer_mean
  - primitive: TRANSFORM
  op: array_std
  params:
    array: peer_returns
  output: peer_std
  - primitive: TRANSFORM
  op: sub
  params:
    a: asset_return
    b: peer_mean
  output: deviation
  - primitive: TRANSFORM
  op: add
  params:
    a: peer_std
    b: $epsilon
  output: denom
  - primitive: TRANSFORM
  op: div
  params:
    a: deviation
    b: denom
  output: mispricing_score