Jose Menchero is Head of Portfolio Analytics Research at Bloomberg and the principal architect of the MAC3 Global Equity Risk Model. He is one of the most prolific researchers in commercial equity risk modeling, with over 30 publications in leading practitioner journals. Inducted into the Performance and Risk Hall of Fame in 2014.

Background

  • PhD in Theoretical Physics, UC Berkeley
  • BS in Aerospace Engineering, University of Colorado at Boulder
  • CFA Charterholder

Before entering finance, Menchero was Professor of Physics at the University of Rio de Janeiro. His physics training informs his approach to risk modeling — the MAC3 paper reads more like a physics monograph than a typical finance whitepaper, with its emphasis on dimensional reduction, eigenvalue problems, and estimation theory.

Career

PeriodRoleOrganization
~2018-presentHead of Portfolio Analytics ResearchBloomberg
2015Scientific Advisory BoardMcKinley Capital Management
2015Founder & CEOMenchero Portfolio Analytics Consulting
2007-2014Managing Director, Global Head of ResearchMSCI (Barra)
~1999-2007Director of ResearchThomson Financial / Thomson Reuters
1997-1999Professor of Physics (Quantum Theory of Solids)University of Rio de Janeiro, Brazil
~1997PhD Theoretical PhysicsUC Berkeley
EarlierBS Aerospace EngineeringUniversity of Colorado Boulder

Key contributions

Risk model methodology

At MSCI, Menchero was responsible for building the next-generation suite of Barra equity risk models, including GEM2 (Global Equity Model). At Bloomberg, he led the development of the MAC3 suite, introducing several innovations:

  • Industry and country beta factors: replacing binary (0/1) exposures with time-series betas, increasing explanatory power and reducing spurious correlations
  • PCA Shrinkage for covariance estimation: blending sample and PCA correlation matrices to produce well-conditioned matrices (first introduced in MAC2, 2016)
  • Cross-Sectional Volatility (CSV) scaling: feedback loop using bias statistics for rapid regime detection
  • Finite-sample adjustment: correcting the traditional double-counting of specific risk in factor portfolios
  • Term structure of risk: daily-frequency estimation for multiple prediction horizons (daily through long-term)
  • Inverse residual variance regression weights: replacing the traditional square-root-of-market-cap scheme

Performance and risk attribution

Menchero made foundational contributions to return and risk attribution methodology:

  • Multiperiod Arithmetic Attribution (Financial Analysts Journal, 2004): framework for linking single-period attribution across multiple periods
  • Custom Factor Attribution (Financial Analysts Journal, 2008, with Poduri): attributing returns to custom factor definitions
  • X-Sigma-Rho formula (Journal of Portfolio Management, 2011, with Davis): decomposing risk contribution as exposure x volatility x correlation. Provides a clean link between return attribution and risk attribution.
  • Pitfalls in Risk Attribution (with Davis): identifying common errors in risk decomposition

Covariance estimation and portfolio optimization

  • Eigen-Adjusted Covariance Matrices (2011, with Wang and Orr): adjusting eigenvalues of sample covariance matrices to reduce estimation error
  • Improving Risk Forecasts for Optimized Portfolios (Financial Analysts Journal, 2012, with Wang and Orr): showing that factor model estimation error leads to underforecasting risk of optimized portfolios
  • Portfolio Optimization with Noisy Covariance Matrices (Journal of Investment Management, 2019, with Ji): demonstrating that noise mitigation makes MVO practically superior to naive diversification
  • Correlation Shrinkage: Implications for Risk Forecasting (Journal of Investment Management, 2020, with Li): showing that the time-series method for integrated models systematically underestimates cross-asset correlations by up to 70%

Cross-sectional volatility

  • Decomposing Cross-Sectional Volatility (2010, with Morozov): framework for understanding dispersion of stock returns as a measure of market regime
  • Capturing Equity Risk Premia (2010, with Morozov): defining “optimized” approaches for constructing minimum-risk portfolios with unit factor exposure

Research themes

Menchero’s work centers on a consistent set of problems:

  1. How to estimate covariance matrices reliably when the number of assets far exceeds the number of time periods
  2. How to decompose portfolio risk and return into interpretable factor contributions
  3. How to build risk models that work for portfolio construction, not just risk forecasting (the key lesson: a good risk model must produce a well-conditioned covariance matrix)
  4. How to detect and respond to regime changes in real time (CSV scaling, daily updates)

In this collection

Not yet downloaded

  • Portfolio Optimization with Noisy Covariance Matrices (JOIM 2019, with Ji)
  • Global Equity Risk Modeling / GEM2 (Springer 2010, with Morozov, Shepard)
  • Correlation Shrinkage: Implications for Risk Forecasting (JOIM 2020, with Li)

Sources

  • MAC3 Global Equity Risk Model (File)
  • Risk Contribution Is Exposure Times Volatility Times Correlation: Decomposing Risk Using the X-Sigma-Rho Formula (File, DOI)
  • Eigen-Adjusted Covariance Matrices (File, URL)
  • Decomposing Cross-Sectional Volatility (File, DOI)
  • Characteristics of Factor Portfolios (File, URL)
  • Why Traditional Risk Models Overstate Factor Risk (File)
  • Multiperiod Arithmetic Attribution (File, DOI)
  • Custom Factor Attribution (File, DOI)
  • Improving Risk Forecasts for Optimized Portfolios (File, DOI)
  • Capturing Equity Risk Premia (File, DOI)