Portfolio Optimization, Estimation Error, and Factor Portfolios
Summary
The document considers how portfolio construction works in practice, contrasting mean variance optimization and related risk objectives with simpler diversification and factor based approaches. Its central caution is that optimized weights depend on estimated expected returns and covariance inputs, which can be highly uncertain. An illustrative historical estimate for a market investment has a wide standard error, showing how noisy inputs can make an apparently precise allocation unreliable.
The answers describe equal weighting and broad diversification as practical alternatives, and suggest that factor portfolios may be useful where persistent return patterns are supported by research. They also mention resampled efficient frontiers as a way to address sensitivity to estimated inputs. The discussion cautions that optimized portfolios can underperform a simple equal weight benchmark out of sample, and transaction costs can further reduce apparent gains. These are broad observations rather than a universal account of industry practice; the document gives limited evidence and does not specify investor constraints, risk objectives, or implementation details.
Key ideas
- Mean variance portfolio weights can be highly sensitive to estimated returns and covariances.
- Estimation uncertainty can make optimized allocations less reliable than their point estimates suggest.
- Equal weighting and broad diversification are described as practical portfolio approaches.
- Factor portfolios may capture return patterns that conventional optimization does not represent directly.
- Transaction costs and out of sample performance can weaken optimization results.
Tags
Full text
# Portfolio construction in reality? # Portfolio construction in reality? There are various models for portfolio selection in literature, like, - Harry Markowitz (HM) model ( Mean-Variance Model) [well known model] - Konno and Yamazaki (1991) model: minimizes the sum of absolute deviations - Markowitz- semi-variance model (1959) - Speranza- mean-absolute semi-deviation (1993) There are many more such variants. I just want to know what kind of models are used by practitioners in reality to construct portfolio? How they takes into account the assumption of various model when constructing portfolio (You may assume assumption of HM model)? ## Answer by Jean-Paul (score 9, accepted) https://quant.stackexchange.com/a/24418 Portfolio optimalisation depends heavily on the estimation of the moments (and therefore has HUGE estimation uncertainty). Even though it's useful for comparing and analysing different existing strategies, I think practitioners are moving more towards the usage of factor portfolios for the strategies themselves (e.g. Fama-French). Also because the exploitation of such anomalies have been proven to be quite persistent and relatively profitable. To give you an example of the estimation uncertainty that goes together with portfolio optimalisation, a simple plug-in Markowitz regression on the S&P 500 (1997-2006) using the delta method yields a weight of w=0.5 and a standard error of 0.4! So the estimation says that we should invest half our portfolio in the risky asset (S&P 500), with a standard deviation of 0.4. You can imagine that a 95% confidence interval would range all the way from investing nothing in the market to investing everything. So what is really optimal? The estimation uncertainty can be improved using shrinkage methods but you get the picture. Another disadvantage of such portfolio optimalisation is that it doesn't really capture anomalies like factor models can. It's of course possible to combine the optimalisation of weights with the results from factor models to sort of 'get the best of both worlds'. Besides, most portfolio optimalisation models can't beat the 1/N portfolio out of sample. Even if they do beat the benchmark, they often still have transaction costs which need to be corrected for. After correction, you will find again that the models are quite useless and you're better off investing in the 1/N portfolio instead. Like I said before, this is not true for cross-sectional anomalies. It has been shown numerous times that various factor models and predictive variables perform quite well. See for example Fama and French (2008), Campbell and Thompson (2008) and Goyal and Welch (2008). ## Answer by QuantK (score 1) https://quant.stackexchange.com/a/24599 Many pension funds and mutual funds acquire small positions in many stocks, therefore just banking on the main results of the Markowitz framework: diversification. This could also just be seen as a plain 1/N rule: naive diversification. In the limit, this just equals the market portfolio. Alternatively, to overcome the sensitivity in changes in the return/ variance-covariance matrix, you could apply Michaud's method (resampled efficient frontier) method to portoflio construction.
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.