Skip to content
All library documents

Quantitative Approaches to Tactical Asset Allocation

Article Quant Q&A · Author: Tal Fishman
Read the original

Summary

The discussion presents several quantitative ideas for changing portfolio exposures across assets or building active portfolios. Momentum-based timing is introduced as a way to switch between major asset classes. Other suggestions include combining valuation and momentum signals, using a factor model with a risk model for a constrained long-only portfolio, and constructing a global minimum-variance portfolio by estimating covariance rather than expected returns as well.

Additional approaches mentioned are Markov regime-switching models and relative strength. One response describes ranking assets by the slope of the log ratio between each asset class and a composite, then assigning them to groups that determine overweight or underweight positions. These are brief pointers and summaries, not comparable evaluations or complete implementation instructions. The discussion does not specify data choices, rebalancing rules, transaction costs, or out-of-sample evidence, and it mixes tactical allocation with broader active portfolio methods.

Key ideas

  • Momentum can guide decisions about when to rotate among major asset classes.
  • Value and momentum signals can be combined to favor assets that are both relatively cheap and strong.
  • Minimum-variance allocation focuses on estimating covariance rather than expected returns alone.
  • Regime-switching models and relative strength are also proposed as tactical allocation approaches.
  • Ranking the slope of an asset-to-composite log ratio can inform overweight and underweight groupings.

Tags

Multi-assetMomentumFactor investingPortfolio constructionRisk management
Full text
# What are some of the major quantitative approaches to tactical asset allocation?


# What are some of the major quantitative approaches to tactical asset allocation?












Note: This question was written for the weekly topic challenge.

Many of you who deal with asset allocation will probably already be familiar with Mebane Faber's Timing Model, based on one of SSRN's most popular papers of all time, A Quantitative Approach to Tactical Asset Allocation. The crux of his approach is to apply the momentum approach to the decision of when to switch between major asset classes.

What are some of the other major contributions to market timing? What other approaches do they use?

## Answer by Kyle Balkissoon (score 2, accepted)

https://quant.stackexchange.com/a/3045

Not Purely Tactical however some of this should answer your question:

I included the minimum variance portfolio as a "active" (with lots of rebalancing)

Value and Momentum: http://pages.stern.nyu.edu/~lpederse/papers/ValMomEverywhere.pdf A tl;dr: long undervalued stocks (book/market) that have strong up momentum, short overvalued stocks (book/market) that have downwards momentum.

130/30 New Long Only: http://math.nyu.edu/faculty/avellane/Lo13030.pdf Uses a factor model to estimate expected returns, BARRA risk. You can find implementation of this in R.

Global Minimum Variance Portfolio: tons of papers on this topic a "recent" one http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1549949, essentially a portfolio constructed to minimize variance (or downside volatility), you need to estimate the covariance matrix, as opposed to having to estimate returns and the covariance matrix.

## Answer by Ram Ahluwalia (score 2)

https://quant.stackexchange.com/a/3601

Here's an approach by Kritzman et al on Tactical Asset Allocation using Markov regime switching models. Here's another approach that uses relative strength in TAA.

## Answer by Suminda Sirinath S. Dharmasena (score -3)

https://quant.stackexchange.com/a/2937

I look at the slope of the natural log of the ratio between the asset class and composite, i.e., slope(ln(asset/composite)). Put this into 5 baskets and decide to over weight / under weight accordingly.

As for references there there any many but I can give the following two pointers: - Publication by AI Investors - Symmys (much of it is also in ssrn, but of course this has a broader scope.)

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.