Asymmetric adjustments in the spread of lending and deposit rates: Evidence from extended threshold unit root tests

• Author: Mark C. Strazicich,Byung Chul Yu,Junsoo Lee
• Date: 01 November 2013
• DOI: http://doi.org/10.1016/j.rfe.2013.08.002
• Published date: 01 November 2013

Asymmetric adjustments in the spread of lending and deposit

rates: Evidence from extended threshold unit root tests☆

Junsoo Lee

a,1

, Mark C. Strazicich

b,

⁎, Byung Chul Yu

c,2

a

Departmentof Economics, Financeand Legal Studies, Universityof Alabama, Tuscaloosa,AL 35487-0224, United States

b

Departmentof Economics, AppalachianState University,Boone, NC 28608-2051,United States

c

Departmentof InternationalTrade, Dong-A University,Busan, Republic of Korea

abstractarticle info

Articlehistory:

Received 3 August 2012

Receivedin revised form 28 July 2013

Accepted13 August 2013

Availableonline 27 August 2013

JEL classification:

C220

E400

Keywords:

Asymmetriceffects

Thresholdmodels

Unit roottests

Interestrates

In this paper, we test f or asymmetric adjustmen ts in the spread of the U.S. prime l ending rate and 3-month

certificate of deposit rate. In doing so, we extend the pioneering threshold unit root tests of Enders and

Granger (1998) to more flexible models where thedeterministic terms and short-run dynamics, in addition

to the persistent parameters, can differ in two regimes. While some previous works have tested for asym-

metric adjustments in the spread of lending and deposit rates using threshold unit root tests, the determin-

istic terms and short-run dynamicswere assumed to be symmetric, which can lead to bias and less accurate

conclusions if these conditions do not hold. Overall, we find that the spread in lend ing and deposit rates is

stationary but adjustment to the equilibrium is asymmetric. In particular, we find more rapid adjustment

when the spread is narrowing below a threshold level than when widening above this level. Several theo-

retical implica tions are suggested.

© 2013 ElsevierInc. All rights reserved.

1. Introduction

In this paper, we test f or asymmetric adjustments in the spre ad

between the United St ates (U.S.) prime lending rate and 3-m onth

certificate of deposit (CD) rate. Previous research has found evi-

dence o f asymmetric behavior in bank lending and/or deposit rates

(e.g., Hannan, 1994; Scholnick, 1996). However, relatively few papers

have focused on empirically examining the spread between lending

and deposit rates. Exceptions are Thompson (2006) and Nguyen and

Islam (2010),who examin e the spread of lending and deposit rates

in the U.S. and Thailand, respectively. These authors apply the

threshold unit root tests of Enders and Granger (1998) and find

asymmetric adjustments of the spread in each case. Nguyen and

Islam (2010) find asymmetric adjustments that are opposite to

those in Thompson (2006), which they attribute to an “oligopsonis-

tic”relationship between the banks in Thailand and their corporate

customers.

To implement our testin g procedures, we depart from the previ-

ous works by adopting m ore flexible models when warranted by

the data. While previ ous studies have tested for asymmetr ic adjust-

ments in lending and dep osit rates and the spread, certain restri c-

tions in the testing pr ocedures were imposed that could lead to

bias and less accurate conclusions. In part icular, while the previous

studies allowed for possible asymmetric persistent parameters,

they assumed that the d eterministic terms and short-run dynam ics

were symmetric. Howe ver, it is possible that the intercept and

trend terms and/or co efficients on the first-differenced lagge d

terms that correct fo r serial correlation can differ in two re gimes. If

so, these asymmetric terms ca n potentially be included in the tes ting

equation to give more accura te results.

An additional methodological contribution relates to the selec-

tion of the threshold t erm in the unit root test equation. In previo us

studies, the thresh old parameter was assumed to ta ke a zero value in

the unit root test, whi ch might have been necessary to make these

tests more manageabl e in practice. While this assumption mi ght be

reasonable in some ca ses, it is possible that the threshold v alue can

take on a wide range of non-z ero values. In such cases, assuming a

zero threshold value i n the unit root test may lead to less accura te

conclusions. In our exten ded tests we drop this restriction and

allow the threshold t erm to take on a wide range of values. Given

that the critical valu es in the unit root test depend on the thres hold

Reviewof Financial Economics 22 (2013)187–193

☆We thanktwo anonymous refereesfor their helpful commentsand suggestions.

⁎Correspondingauthor. Tel.:+ 1 828 2626124.

E-mailaddresses: jlee@cba.ua.edu (J. Lee), strazicichmc@appstate.edu (M.C. Strazicich),

bcu@dau.ac.kr(B.C. Yu).

1

Tel.: +1 205 348 8978.

2

Tel.: +82 51 200 7441.

1058-3300/$–see front matter © 2013 ElsevierInc. All rights reserved.

http://dx.doi.org/10.1016/j.rfe.2013.08.002

Contents listsavailable at ScienceDirect

Review of Financial Economics

journal homepage: www.elsevier.com/locate/rfe