We explore two approaches (cohort versus hazard) to measure the probability of investment rate migrations of pension funds in Australia. We also develop validation procedures pertinent to each approach and find that the cohort method is more stable in its forecasts and reports a lesser migration probability to lower investment grades with minimal statistical significance. Conversely, the hazard approach reports a higher migration probability to lower investment grades with statistical significance. This finding has considerable consequences for fund managers as they seek to mitigate any downward trends in their investment appraisals, especially as the cohort approach is the industry’s preferred approach in calculating rating migrations. The fund manager has a choice to make regarding measuring probability investment rate migrations, one between: stability (cohort) or accuracy (hazard).