Compression in monetary user costs in the aftermath of the financial crisis: implications for the Divisia M4 monetary aggregate

ABSTRACT

Differences in Divisia and simple-sum money arise from appropriate weighing mechanisms in Divisia, which rely on information on the user cost of monetary assets. We show convergence in the growth rate of Divisia M4 and its simple-sum counterpart beginning in early 2009, shortly after the collapse in the Federal Funds rate. This phenomenon results from compression in user costs.

KEYWORDS:

• Divisia
• money aggregates
• financial crisis

JEL CLASSIFICATION:

• E30
• E31
• E65

Acknowledgements

We want to thank the editors and an anonymous referee for their prompt and helpful comments. We are also grateful to William Barnett for illuminating feedback. The views expressed herein are solely those of the authors and do not represent the views of the Center for Financial Stability. All standard disclaimers apply.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Barnett, Keating, and Kelly (2008) find that simple-sum M2 significantly overstates the monetary aggregate.

² See Basu and Wang (2013), Belongia and Ireland (2014), and Keating et al. (2014) for treatments of Divisia user costs in DSGE models.

³ Even broader aggregates than DM4 – augmented to include the services of credit card transactions services – are to be made available by the CFS in the future. While this article finds that various user costs on monetary assets have been compressed toward zero by the Federal Reserve’s distortionary interventions in money markets, user costs for credit card services remained much higher than the compressed user costs on monetary assets during this period (Barnett and Su 2014). Thus, this augmented DM4 measure will be less vulnerable to the compression revealed in this article.

⁴ Thus, the difference in information content must lie with information on the user cost. If one were to compare the broadest measure extant for simple sum (M2) and the broadest for Divisia (DM4), one might see important differences in their growth rates. However, it would be impossible to reconcile how much of the difference is explained by a different bundle of monetary assets and how much is explained by the appropriate weighing scheme.

⁵ Inspection of Fig. 4 shows growth rates in DM4 and M4 generally differ up to the Financial Crisis. Figures 1–3 suggest a substantial compression in user costs occurs in the aftermath of the Financial Crisis concurrent with the convergence in growth rates of DM4 and M4.

⁶ Distance is computed as and and denote compression (where the upper envelope of user costs decreases) and decompression (increasing distance of the upper envelope relative to mean) in user costs, respectively. Numbers underneath the regressions denote correlation estimates and t-statistics.

⁷ Coefficients were qualitatively and quantitatively robust to substituting Z in (5) with a band of ranges in user costs each period or with the un-normalized upper envelope of all user costs per period.