Skip to content
All library documents

Assessing Factor Loading Stability in Asset Pricing Models

Article Quant Q&A · Author: Si Chen

Summary

The document raises a research question about measuring the stability of factor loadings in models such as Fama–French or Carhart. It frames loading stability as a criterion for evaluating alternative ways to construct a risk factor, using a possible climate or carbon risk factor as an example. The focus is on whether assets’ estimated sensitivities to a proposed factor remain relatively consistent over time.

No measurement method, empirical study, or results are supplied; the text asks whether relevant research exists. It therefore identifies a factor-design concern rather than recommending a specific stability test. Any assessment would need to define what stability means, including the period, estimation approach, and acceptable degree of change. The document does not specify these choices or discuss how stable loadings should be weighed against other properties of a useful risk factor.

Key ideas

  • Factor loading stability is proposed as one criterion for evaluating constructed risk factors.
  • The question refers to Fama–French and Carhart style factor models.
  • Climate or carbon risk is given as an example of a factor that might be constructed.
  • The document provides no test procedure, empirical evidence, or definition of acceptable stability.

Tags

Full text
# how to measure the stability of factor loadings?


# how to measure the stability of factor loadings?












Is there any research on how to measure the stability of factor loadings, for example from Fama French/Carhart type factor models?

The goal is to test different ways to construct a risk factor, such as for example a risk factor for climate/carbon risk. One criterion is that the risk factor produces loadings which are relatively stable.

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.