Factor Models in Portfolio Construction

While fundamental factor models are used to measure manager performance, managers are using fundamental factor models to calibrate their portfolios. Passive managers may want to use a fundamental factor model to identify factor sensitivities of the benchmark. Then it's easier to replicate with a low tracking error. Active managers can use one for calibrating factor sensitivities of a portfolio according to the risks wanted. Then rules-based active management may require recalibrating the model regularly, as events trigger sensitivity changes.
Suppose that a manager was designing a portfolio to have exposure to only one identified factor. What do you think would work best?
Not really. For example, a sensitivity of 0.2 is larger than another of -1.8, but that setup wouldn't be what you're looking for.
Exactly! That's the goal. A bunch of zeros in your fundamental factor model is kind of like the "all else being equal" assumption. It's convenient to assume, but really helpful to design in a portfolio if you can.
No. That would be a little more like ignoring information rather than using it for specific goals.
Specifying risks in a portfolio is very useful for determining which is appropriate. Some portfolios might have a large sensitivity to business cycle risk, for example. What sort of investor would be in a good position to bear that risk?
Not likely. Banks usually need to keep cyclical risks away from their operations.
That's right. An endowment would have a long-term investment horizon and would be in a much better position than many others to ride out the cyclical risks presented by some portfolios.
Probably not. Someone close to retirement wouldn't be in a good position to wait for a recovery if an economic downturn took place.
This is quite a difference from the world of the capital asset pricing model (CAPM), which suggests that every investor just needs to decide about the weights between the risk-free rate and the single risky asset. How would you best characterize the framework of fundamental factor modeling when compared to the CAPM?
Absolutely. Investors of various types can choose the risks that they are exposed to in this framework. Risks can be shifted and controlled once they are modeled and measured, and risks that are better identified and measured can be better priced. It's not as easy, but it's much more informative. It's a multifactor world. Embrace the complexity.
No. A single-factor model is generally simpler than a multifactor model.
Not really. The CAPM has more assumptions and is a little further from reality in some ways. It's probably the more abstract of the two.
To summarize: [[summary]]
Use a single-factor model
Work toward a sensitivity of 1 for that factor and 0 for others
Ensure that the factor sensitivity of the desired factor is larger than the others
A bank
An endowment
A retail investor near retirement
Richer
Simpler
More abstract
Continue
Continue
Continue
Continue

The quickest way to get your CFA® charter

Adaptive learning technology

5000+ practice questions

8 simulation exams

Industry-Leading Pass Insurance

Save 100+ hours of your life

Tablet device with “CFA® Exam | Bloomberg Exam Prep” app