Current unemployment variance decomposition and consequences of using proxies

ABSTRACT

We offer a decomposition for the variance of the current unemployment rate that not only measures the contributions of labour market flows but also of the approximation error embedded in other decompositions that use surrogates for the current rate. Using data for the United States and Brazil, the results for the latter show significant differences in the flows’ contributions and non-negligible distortions of approximation errors when the variance of the current (instead of the proxy) rate is decomposed; for the U.S., no substantial changes are detected.

KEYWORDS:

• Labour market flows
• unemployment
• variance decomposition
• proxy

JEL CLASSIFICATION:

• E24
• J64

Disclosure statement

No potential conflict of interest was reported by the author (s).

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

¹ See Moreira, Foguel, and Corseuil (2020) for the details of the derivation.:

² As it is well known, the contributions of the components of the decompositions can be obtained by regressing the time series of each component against the time series of the unemployment rate of interest.

³ The contributions of the initial conditions (${\beta }_0^h$ and γ${ }_0^h$) decline monotonically with $h$. Our results are normalized with respect to ${\beta }_0^h$ and γ${ }_0^h$. Results without the normalization exhibit the same pattern. Full results are presented in an online Appendix.