Observed Recession Rules and Statistical Significance

Summary

The note distinguishes a recession rule based on observed GDP figures from one that explicitly requires statistical significance. It considers a proposed definition: three consecutive quarters of declining GDP. The example lists three successively lower quarterly growth readings, while noting that the first and third are not statistically significantly different. Under the stated rule, the sequence still qualifies because the rule refers only to the observed values and does not impose a significance test.

If statistical significance is meant to be part of the definition, the criterion must say so and specify a threshold, such as a p-value cutoff. The response clarifies how definitions govern classification; it does not assess whether the proposed recession rule is economically sound or whether GDP estimates carry uncertainty in practice. It also does not provide a statistical test or discuss revisions, measurement error, or other recession indicators. Its lesson is narrowly about applying a rule as written and distinguishing observed changes from inferential claims.

Key ideas

• A rule based only on observed quarterly GDP declines can be satisfied without a significance test.
• Statistical significance affects classification only if the definition explicitly requires it.
• A significance-based rule needs a specified threshold.
• The example clarifies the meaning of the rule but does not validate it as a recession measure.

Tags

# GDP and Statistical Significance


# GDP and Statistical Significance












Let’s say that we are defining a U.S. recession trend as 3 consecutive quarters of GDP decline. The GDP figures below would satisfy this definition. However, these figures are statistics taken from a population with zero sampling error. Q3, as it turns out, is NOT statistically significantly different from Q1. Does that mean the definition of recession is not satisfied?

Q1 – 0.8%

Q2 – 0.7%

Q3 – 0.6%

## Answer by Nik Tuzov (score 1)

https://quant.stackexchange.com/a/29652

It doesn't mean the definition is not satisfied. Apparently, the definition you provided refers just to the observed quantities. If the definition were more involved, e.g. if it were to take statistical significance into account, then it would have specify a significance threshold, like 5% p-value.

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.