On the relative importance of demand and supply shocks: a panel VAR study for US states

ABSTRACT

This article is a panel VAR study of demand and supply shocks in the USA using state-level data where structural shocks are decomposed into state idiosyncratic and common components. Decomposition suggests that in all instances, idiosyncratic state shocks rather than common shocks have larger impact and explain most variation in both the state-level unemployment rate and real gross state product. Further, demand shocks are the primary driving force in unemployment rate fluctuations, while both demand and supply are important in output movements to varying degree of impact and importance depending on the use of quarterly or annual data.

KEYWORDS:

• Structural panel VAR
• demand
• supply
• impulse response

JEL CLASSIFICATION:

• E30
• E32
• R11

Acknowledgement

We thank Peter Pedroni for providing us with details and discussion on his estimation methodology.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Lag length is chosen separately in each VAR.

² Programs for the bootstrapping estimation were written in RATS 9.0.

³ A longer span of data can more accurately estimate permanent effects and this may explain the increased explanatory power of supply shocks in the annual data.

⁴ The similarity of results in the article to BQ’s findings using panel state-level data instead of macro data is a reassuring feature for both results here and in BQ.

⁵ Also noteworthy in both quarterly and annual results, the unemployment rate and real GSP respond significantly to supply shocks in expected opposite directions. This is in contrast to BQ and Chen and Netsunajev (2016), where both variables show a statistically significant increase in the short-run.