Interest rate pass‐through: Divisia user costs of monetary assets and the federal funds rate

• Published date: 01 April 2018
• Author: Victor J. Valcarcel
• DOI: http://doi.org/10.1002/ijfe.1605
• Date: 01 April 2018

RESEARCH ARTICLE

Interest rate pass‐through: Divisia user costs of monetary

assets and the federal funds rate

Victor J. Valcarcel*

School of Economic, Political and Policy

Sciences, University of Texas at Dallas,

Richardson, TX 75080, USA

Correspondence

Victor J. Valcarcel, School of Economic,

Political and Policy Sciences, University of

Texas at Dallas, Richardson, TX 75080,

USA.

Email: victor.valcarcel@utdallas.edu

JEL Classification: E30; E31; E65

Abstract

Evidence of substantial pass‐through in short‐term rates and other rates of

financial and monetary assets has been typically rejected by the data. This

paper investigates level and volatility transmissions among the federal funds

rate and the user cost of various monetary assets, which include both instru-

ments of public debt (e.g., t‐bills) and private debt (e.g., commercial paper).

Results suggest substantial and time‐varying pass‐through. Higher degrees of

bidirectional pass‐through occur between the federal funds rate to the user

costs of more liquid assets—both in levels and volatilities. Federal funds rate

spillovers propagate faster onto more liquid rates as well. These findings have

important implications for monetary transmission not only across the term

structure but along markets of varying liquidity.

KEYWORDS

federal funds rate,interest rate channel, monetary transmission, user costof money, VAR, volatility

spillovers

1|INTRODUCTION

Modern models that analyse how monetary policy affects

economic activity do not typically involve monetary

aggregates. Instead, the consensus in academic and cen-

tral banking circles is that movements in a specific

short‐term rate, the federal funds rate, constitutes an

informative descriptor of monetary policy.

1

A reductive

characterization suggests that, through open market oper-

ations, the Federal Reserve can control the overnight fed-

eral funds rate, which in turn affords the Fed substantial

influence (some would argue “control”) over other inter-

est rates—particularly interest rates of sovereign debt.

Despite vigorous theoretical claims, evidence of sub-

stantial interest rate pass‐through seems to be rejected

by the data, both during unconventional monetary policy

and more normal conditions. Importantly, a key

assumption of empirical models that investigate interac-

tions among multiple interest rates is the perfect substi-

tutability among the assets included. This assumption is

palatable if the basket of interest rates is limited to sover-

eign debt instruments only. However, the assumption of

perfect substitutability becomes more difficult to justify

when considering other interest rates in financial mar-

kets. Presumably, the (unlikely perfect substitute) yields

of private debt matter most for economic activity.

This paper investigates the pass‐through between

movements in the federal funds rate and user cost rates

of various monetary and financial assets that, by construc-

tion, relax the strong perfect substitution assumption

mentioned earlier. We focus on the user cost of monetary

assets for a number of reasons. There is a rich theoretical

framework that establishes the concept of user costs from

first principles and their relation to monetary aggregation

(see Barnett, 1978 for seminal treatment). Because, gener-

ally, monetary assets are durable enough during the

period from use, their own user costs constitute the price

*

Victor Valcarcel can be reached at victor.valcarcel@utdallas.edu or

972‐883‐6213 (ext. 6213). All standard disclaimers apply.

Received: 23 November 2016 Revised: 13 November 2017 Accepted: 14 December 2017

DOI: 10.1002/ijfe.1605

94 Copyright © 2018 John Wiley & Sons, Ltd. Int J Fin Econ. 2018;23:94–110.wileyonlinelibrary.com/journal/ijfe

of the assets (Belongia & Ireland, 2006). Essentially, the

user cost of a given monetary asset acts as an interest rate

spread—which does not presuppose perfect substitution

—and measures the interest foregone by holding a spe-

cific asset instead of an alternative asset that does not fur-

nish any monetary services and earns a higher

(benchmark) rate of return (Anderson & Jones, 2011).

This paper operates on a few novel dimensions.

Although some recent theoretical research has incorpo-

rated the user cost of monetary aggregates in Dynamic

Stochastic General Equilibrium (DSGE) models, fewer

papers have focused on empirical applications of the user

cost. To our knowledge, this is the first paper that con-

siders the user cost of individual assets—rather than an

exclusive focus on the user cost of monetary aggregates.

Thus, we highlight a strong co‐movement between the

federal funds rate, or shadow federal funds rate, and the

user costs of various monetary assets. This paper presents

volatility profiles of the user costs of monetary aggregates

as well as individual assets and contrasts them with those

of the federal funds rate.

As a preview of results, we find relatively large volatil-

ity spillovers between the federal funds rate and several

user costs. These volatility spillovers are larger between

the federal funds rate and the user cost of narrower mon-

etary aggregates rather than broader aggregates, suggest-

ing closer pass‐through of the volatility of the federal

funds rate to user costs of more liquid assets. This conclu-

sion seems confirmed by results from a second specifica-

tion that suggests that the largest portion of federal

funds rate volatility spills over to the user cost of more liq-

uid assets such as demand deposits. Also, spillovers

exhibit faster propagation onto user costs of more liquid

assets and the spillovers seem bidirectional. We find

strong spillovers in the levels of interest rates. These con-

clusions are confirmed by both fixed and time‐varying

parameter models.

2|THE USER COST OF MONETARY

ASSETS

Since the 1980s, a boon in financial innovation has led to

a massive increase in the variety of financial instruments

and services offered by financial institutions. Often these

institutions provide a range of services whose costs are

not explicitly passed on to their clients and are, instead,

recovered by earning a positive interest rate margin. For

example, commercial banks charge rates on loans that

are typically higher than those charged on other market

securities of comparable risks. Similarly, banks routinely

pay lower rates on deposits than those paid on compara-

ble market securities. Bank customers accept these

nonmarket rates because they value the financial services

rendered by the banks. The nominal value of client ser-

vices—when bank compensation comes via an interest

margin—can be inferred as the interest foregone by

depositors when they accept a lower rate than the yield

on market instruments with the most comparable risk

profile (Basu & Wang, 2013). Thus, the nominal value of

depositor services can be imputed as the product of the

current value of deposit balances and the interest rate

spread. This interest rate spread is often referred to as the

user cost of money. Barnett(1978, 1980) constitutes the orig-

inal works that provide proof of existence, a theoretical

foundation, and coin the term. The user cost of money is

an analogue concept to the well‐known Jorgensonian user

cost of durable consumer goods (Donovan, 1977). The user

cost of money (Ψ)foraspecificasset(n) is a function of the

interest foregone (r

n,t

) by holding asset nin place of an

alternative asset that does not provide any monetary ser-

vices but earns a higher rate of return (R

t

), which is typi-

cally referred to as the benchmark rate.

Ψn;t¼Rt−rn;t

1þRt

:(1)

The user cost of money serves a number of theoretical

functions. For example, monetary aggregation generally

requires weights to be assigned to each component asset

of a Divisia measure of the money supply. These weights

are functions of the user cost of the respective assets. Also,

differences in these user costs can arise in the solution to a

household's optimization problems, which may reflect the

different amounts of monetary services furnished by the

assets (Holmstrong & Tirole, 2011). User costs appear in

some DSGE applications that place money in the utility

function (MIUF) of the household or with a shopping time

mechanism.

2

Basu and Wang (2013), Belongia and Ireland

(2014), and Keating, Kelly, Smith, and Valcarcel (2014) are

examples of recent theoretical constructs that incorporate

the user cost of money. This paper's contribution, however,

focuses on empirical application of the user cost of money,

which far less attention has been devoted to thus far.

3

3|USER COSTS AND MONETARY

AGGREGATES

Monetary aggregation employs information on the user

cost of various assets in order to weigh those assets by

degrees of “moneyness.”There is a tight relationship

between the concept of the user cost and the stock of

money. Let gM2

tdenote the growth rate of M2 stock of

money. According to the Board of Governors of the

Federal Reserve, the growth rate of M2 balances

4

is

given by

VALCARCEL 95