Home >> Trading Strategy Evaluator >> Optimization

Applying advanced optimization techniques to your portfolio selection enables you to truly push the envelope in balancing the risk and return from your investments. It allows you to select a desirable combination of risk and return, knowing that the methodology presents you with a combination that displays the lowest attainable risk for the given return and the highest attainable return for the given level of risk.

To bring the benefits of optimization to you, Zacks has partnered with Infanger Investment Technology, LLC, a successful investment advisory firm. IIT has packaged as a software tool the core functionality of the advanced portfolio optimization methods it has developed and utilized over the last 10 years. This software product is called IITPortf, and its use has been integrated into Zack’s Trading Strategy Evaluator (TSE).

Usage Modes

The TSE/IITPortf combination supports two complementary modes of operation:

• TSE can execute a complete back-test process to evaluate the performance of an optimization strategy configured by user choices in TSE and IITPortf. In this process, TSE sequentially prepares inputs and dispatches the IITPortf optimizer for each period of the back-test. Performance reports are produced just as they are for the other strategies that TSE can evaluate via back-test.
  
  
• In "production" mode, TSE prepares the inputs for a single optimization, typically used to define a target portfolio for current trading using an optimization configuration that performed well in a back-test.

Optimization Model

The portfolio optimization problem constructed and solved by IITPortf is based on well-established and familiar mean-variance techniques, as pioneered by Markowitz. The so-called efficient frontier at the heart of the Markowitz model is precisely the envelope of risk-return combinations referred to above. In finding a point on this frontier (Sharpe Ratio), a simple switch in IITPortf allows the user to choose either to:

• optimize a weighted combination of expected return and risk (variance), or
  
  
• maximize expected return subject to an upper bound on risk.
  
  

In implementation, the IITPortf core exploits an advanced problem formulation and licensed optimization software that together permit the very fast computation of optimal portfolios, even for a large universe of several thousand assets.

Risk and Return

The central inputs to the optimization are predictions of asset returns (typically called alphas) and measures of the covariance of asset returns.

• A ZRS user can employ licensed tools to develop a set of alphas based on a custom alpha built by the user or an already proven one such as the Zacks Indicator.
  
  
• This model is generated on the fly using a so-called internal factor model derived from a principal component analysis of the historical returns. Simple parameters in IITPortf control the number of internal factors extracted from the data.
  
  

Risk Factor Exposures

The user may specify any number of risk factors for purposes of controlling the exposure of the optimized portfolio to that risk factor. These factors are distinct from the internal factors referred to above, which are statistically derived from asset returns alone. Examples of risk factors are market beta and earnings-to-price ratio. ZRS tools can be used to deposit items in the database that represent the individual exposure of each asset to the risk factor (called a factor loading). Once in the database, risk factors are readily chosen in TSE for export to the optimization. For each exported risk factor, lower and/or upper bounds on the portfolio-weighted exposure can be specified in IITPortf, where the bounds may be specified in absolute terms or relative to the exposure of a benchmark portfolio.

Similar to the treatment of risk factors, the sector composition of the optimized portfolio can be bounded, either in absolute terms or relative to the composition of a benchmark portfolio.

Optimization Control

A number of other user choices affect the final form of the optimization problem.

• A simple switch in IITPortf controls whether the variance is to be measured in absolute terms or as a tracking error relative to a benchmark portfolio (configured in and exported by TSE).
  
  
• Transaction costs specified in TSE are automatically exported to IITPortf, wherein the optimization considers the negative impact of transaction cost on effective portfolio return.
  
  
• In the case of revising an existing portfolio, the optimization can consider an upper bound on turnover specified in the IITPortf interface.
  
  
• Buy and sell rules configured in TSE are automatically exported to IITPortf in the form of stipulations that any existing position must be sold off, that any existing position may be held but not increased, or that the asset may be bought or sold as justified by the optimization.
  
  
• The IITPortf interface provides for specifying an upper bound to be applied uniformly to all assets, where the bound can be in absolute terms or relative to the holding in the benchmark portfolio.
  
  
• An upper bound can be specified in IITPortf on the number of active positions in the optimized portfolio.
  
  

For a Trial, demonstration or product information, please Request Product Information or call Tim Nyland, Marketing Manager at 800-767-3771 ext. 9455 or email at tnyland@zacks.com.