Skip to content
All library documents

Konno Mean Absolute Deviation Portfolio Optimization with Linear Programming

Article Quant Q&A · Author: purbani

Summary

The document describes implementing Konno’s mean absolute deviation portfolio optimization model as a linear program in a spreadsheet. It identifies the LP Simplex method as the solver approach, presenting MAD as an alternative risk measure for portfolio construction. The referenced method is associated with a paper on reducing the Konno–Yamazaki model, though the document does not provide the spreadsheet itself or walk through its variables and constraints.

It states that, when the underlying returns are strictly multivariate normal, the MAD approach can be shown to be equivalent to standard Markowitz mean-variance optimization. This equivalence is conditional on the stated distributional assumption; the document gives no empirical comparison, numerical example, or discussion of behavior when returns depart from normality. It therefore offers a concise pointer to a modeling approach rather than a complete implementation guide.

Key ideas

  • Konno’s mean absolute deviation model can be formulated as a linear programming portfolio optimization problem.
  • The described spreadsheet approach uses LP Simplex methods.
  • Under strictly multivariate normal returns, the document says MAD optimization is equivalent to Markowitz mean-variance optimization.
  • The source provides no detailed spreadsheet steps or evidence beyond the stated theoretical equivalence.

Tags

Full text
# How to implement Konno's Mean-Absolute Deviation Portfolio Optimization Model using LP methods in Excel


# How to implement Konno's Mean-Absolute Deviation Portfolio Optimization Model using LP methods in Excel












Konno proposed a LP method for portfolio optimization using the Mean Absolute Deviation (MAD)

## Answer by purbani (score 1)

https://quant.stackexchange.com/a/16922

This spreadsheet shows how to implement Konno's Mean Absolute Deviation (MAD) Portfolio Optimization in Excel using LP Simplex methods.

For strictly multivariate normal underlyings, the method can be shown to be equivalent to the standard Mean Variance Optimization method of Markowitz et al.

The method is based on the paper Further Reduction of the Konno-Yamazaki Mean-Absolute Deviation Portfolio Optimization Model by Mike Fox.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.