Using Unemployment Measures to Classify Business Cycle Stages

Summary

The document proposes classifying the business cycle as early, mid, or one of two late stages using unemployment indicators. Its qualitative scheme considers whether unemployment is rising, near a peak, falling, or low, along with comparisons between the current level and a recent three-month average. It also suggests using the unemployment level’s percentile within a three- or seven-year history to distinguish stages.

The author is considering z-scores as a way to turn these observations into a predictive rule, but provides no implementation, test results, or evidence that the proposed thresholds identify cycle stages reliably. The table gives directional hypotheses rather than a validated model, and some categories allow unemployment to move in more than one direction. A researcher would need to define the indicators and stage labels precisely, then assess the rule against a suitable historical reference. The document leaves open whether z-scores or another calculation would be more useful.

Key ideas

• The proposed framework uses unemployment direction and level to label business cycle stages.
• It compares unemployment with a recent three-month average and historical percentiles over several years.
• The author considers z-scores but does not establish that they predict cycle stages well.
• The stage definitions are qualitative hypotheses that require precise specification and evaluation.

Tags

# How to predict what stage of business cycle we are currently in based off of unemployment indicators


# How to predict what stage of business cycle we are currently in based off of unemployment indicators












I am trying to predict what part of the business cycle (Early, Mid, Late 1, Late 2) we are currently in by looking at unemployment indicators.

Qualitatively, I've reasoned that:

| . | Early | Mid | Late 1 | Late 2 |
| Unemployment | Rising | Peak/Falling | Low/Falling | Low/(Steady Decline or Rising) |
| Change |  |  | level < prior 3m avg | level > prior 3m avg |
| z-score |  | level > 50th %tile(3yr or 7yr) | level < 50th %ile (3y or 7y) | <=25th %ile |

I am now trying to express this through code... I believe by taking the z-score I can try to predict what the current state of the business cycle we are in. I am unsure if there is a better calculation than z-score to predict this.

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.