This is the fifth in a series of articles on Unpacking the Investment Process, which provide clarity about various aspects of the investment process of a professional investor. We now continue our investigation of the Construction Phase of an Investment Process.
Our previous article introduced the many aspects of the Construction Phase involved in the answering the question of How Many?, whilst keeping in mind the overall purpose of converting an Investment View into an Order List.
Portfolio Construction Inputs
Before discussing the various methods employed in Portfolio Construction, it’s probably worth recapping the range of inputs that need to be considered when constructing a portfolio.
- Investment Scope – including Investment Universe, Benchmark and Benchmark Weightings
- Current Portfolio Holdings – including Cash
- Funds Flow – since last re-balance
- Return Expectations – either as an explicit Investment Return, such as +15% over 12 months, or -20% over 3 months, or as an Investment Rating, such as 1-to-5, or A-B-C, or Strong Buy – Buy – Hold – Sell – Strong Sell – Avoid
- Risk Expectations – including Variability Metrics such as volatility, correlation, covariance and beta
- Factor Exposures – per security to Style Factors such as Size, Value and Momentum
- Investment Classification – including by asset class, region, country, currency, sector, industry and size
- Investment Objectives – such as Inflation + 2% annual return over a rolling 3-year horizon, or 3%pa active return with 4%pa tracking error
- Investment Restrictions – including regulatory, concentration, shorting, ethical and instrument restrictions
- Investment Guidelines – such as Investment Ranges, Portfolio Turnover, Security Count and Portfolio Factor Exposure
- Liquidity Metrics – such as daily volume forecasts
- Transaction Costs – such as commission, taxes, bid-ask spread and market impact
Despite there being many possible inputs, most approaches to Portfolio Construction can be grouped by the level of focus on risk and return.
a) Focus on Risk, not Return
Portfolios that focus on risk, but not return, are generally quantitatively constructed. Most of the following examples fall into the (poorly named) Smart Beta category:
- Minimum Variance Portfolio which is a portfolio with the lowest possible volatility, which generally results in a concentration of low volatility assets
- Minimum Correlation Portfolio in which assets with low correlations and volatility relative to other assets in the portfolio have higher weighting
- Maximum Diversification Portfolio which maximizes the Diversification Ratio, which is the weighted average volatility of the portfolio as a proportion of the overall portfolio volatility
- Inverse Volatility Portfolio in which each asset is weighted in inverse proportion to its volatility, with all assets then re-scaled to total 100%
- Risk Parity Portfolio in which each invested asset contributes the same amount of volatility to the portfolio
- Passive Portfolio that targets a very low tracking error without owning every stock in the benchmark index
b) Focus on Return, not Risk
Portfolios which focus on return, but not risk, are more likely to be manually constructed by a Portfolio Manager. Examples of construction techniques include:
- Equal Weighted Portfolio in which assets are selected based on a return expectation or rating and then equally weighted in the portfolio
- Tiered Weighting Approach in which assets are grouped into tiers by return expectation or rating, with assets in the top tier each having the largest portfolio weighting, the second tier having the next largest weighting, and so on. For example: A rated stocks have a portfolio weight of 8%, B rated stocks have a 5% weight and C rated stocks 3%, with all weights then re-scaled to add to 100%.
- Tiered Active Weighting Approach is similar to the tiered approach, with each tier mapped to active weightings
- Return Scaled Approach has a defined relationship between expected return and portfolio weighting, or active weighting, without using tiers. The defined relationship is often linear.
c) Focus on both Risk and Return
Portfolios which focus on both return and risk are typically are more difficult to construct.
Let’s look at an approach often used by Portfolio Managers, likely with the help of spreadsheets.
- Two-step Approach is typically constructed by looking first at return, then at risk. Stocks and weights are generated by using an Equal Weight, Tiered or Scaled approach from section b) above. The second step is to manually adjust weights so the risk of the portfolio meets some desired criteria, such as Portfolio Beta =1 or Portfolio Volatility = Benchmark Volatility.
We now look at an approach used by Portfolio Managers in conjuction with Quant or Risk Analysts, generally relying on a Risk Model and an Optimiser.
- Mean-Variance Approach in which there is a trade-off between expected (mean) return and its associated risks (variance). An optimiser is used to either maximise return for a given level of risk, minimise risk for a given level of return, or a combination of both.
d) Focus on neither Risk nor Return
Some portfolios have no focus on risk nor return, including several well-known types:
- Passive Portfolio that fully replicates the index, such as an S&P 500 tracker portfolio or ETF
- Fixed Allocation Portfolio such as a portfolio of 60% Equities and 40% Bonds
- Factor Tilt Portfolio is similar to a Return Scaled approach, with a defined relationship between a Factor Exposure and portfolio weighting or active weighting
e) Combining different approaches
Finally, two or more of the above approaches may also be used to construct portfolios.
Find your approach
There are many possible ways that portfolios are constructed, as our non-exhaustive list highlights. The important thing is to find the approach that best suits your needs, taking all restrictions and other inputs into account.
Appendix: Unpacking the Investment Process
Earlier articles in our series on the Unpacking the Investment Process follow.