International Funding Cost, Mortgage Rates and Cash Rate Cycle Relationship: Evidence in the Context of Australia Quynh Chau Pham Benjamin Liu Eduardo.

1

International Funding Cost, Mortgage Rates and Cash Rate Cycle Relationship: Evidence in the

Context of Australia

Quynh Chau PhamBenjamin LiuEduardo Roca

Griffith Business SchoolDepartment of Accounting Finance and Economics

November 11, 2014

2

Contents:

• Introduction• Related Literature & Hypotheses• Data & Econometric Methods• Result discussion• Conclusion

3

1. Introduction

Mortgage rates

Cash rate Domestic funding

costs

Offshore funding

costs

4Source: Retrieved from Leon Berkelmans and Andrew Duong (2014)

5

Source: Retrieved from Steward, Robertson and Heath (2013)

1. Introduction

Banks’ wholesale DebtShare of total funding, 2011

Source: Retrieved from Leon Berkelmans and Andrew Duong (2014)

6

1. IntroductionGap in the Literature• In Australia, numerous studies on cash rate pass-through on

mortgage rates (Karamujic, 2011; Lim, 2001; Smales, 2012; Valadkhani, 2013, 2014; Valadkhani & Anwar, 2012).

• Recently, several authors have analysed the trends in the funding and lending behaviour of Australian banks (Bailey et al. (2012), Deans & Stewart (2012), Stewart, Robertson & Heath (2013) and Berkelman & Duong (2014)).

• None of them has attempted to incorporate the impact of offshore funding costs on mortgage rates

7

1• Lenders’ funding cycle differs from the cash rate cycle

2• Proliferated attention of policymakers, publicity towards the cash rate

cycle; and lenders’ pricing behaviour,

3• The heavy reliance of Australian lenders on international financial

markets,

4• The limitation in the scarce literature on influence of offshore funding

costs on mortgage rates in Australian market.

1. Introduction

Aim of the StudyIn particular, this study aims to answer two key questions:

• Whether international funding costs have affected mortgage rates? • Has the effect of the cash-rate cycle on mortgage rates reduced by the

increasing presence of offshore funding?

Motivations of the study

8

Contributions

Contributions of the study

1- the offshore funding cost influence on

mortgage rate

2. Using both nominal & effective

mortgage rates

3-ARDL approach

application

1. Introduction

9

2. Related Literature

Cash rate pass-through •An incomplete pass-through of base rate on mortgage rate in UK (Hofmann and Mizen (2004)), in Hong Kong and Singapore (Chong, 2010; Chong, Liu, & Shrestha, 2006), in NZ (Liu et al., 2008)• In U.S, both complete and incomplete pass-through of FFR ((Payne, 2006a; Payne, 2006b, 2007)

Cash rate pass-through

• Lowe (1995) find incomplete pass-through on mortgage rates

• Valadkhani & Anwar (2012), Valadkhani (2013) & (2014); Valadkhani & Bollen (2013) show complete make-up from cash rate to mortgage rates.

Offshore funding cost impact •Concerns on the increasing trend of offshore funding: e.g. Herald Sun, 2014). Bailey et al. (2012), Deans & Stewart (2012), Stewart, Robertson & Heath (2013) and Berkelman & Duong (2014)

10

2. Research Hypotheses

H01: •The changes in cash rate still directly affect our sampled rates of the standard adjusted mortgages in both the short-term and the long term.

H02: •There are positive responses of the mortgage rates of the banks to changes in international funding costs

H03: •The effect of the cash-rate cycle on mortgage rates has been reduced by the increasing presence of offshore funding.

11

3. Data and Econometric Methods3.1 Sample and Data

Sample No. of Obs.

Observations of standard adjustable mortgages 6,850

No. of groups: 1. Whole banking sector 2. Major group 3. Foreign group 4. Regional group

4

Types of Variables Sources

Dependent variables:Mortgage interest rates: 1. Nominal rate (WRATE) 2. Effective rate (WAAPR)

Cannex’s survey of Australian lenders

Independent variables:1. Cash rate (CR) Table F01 of RBA database

2. International funding costs (AUDLIBOR)Australian spot exchange rate

DataStream

12

3 Data and Econometric Methods3.2 Estimation methods– Unit root tests: ADF, PP and KPSS – The Autoregressive Distributed Lag (ARDL) model

developed by Pesaran et al. (2001)– The Granger causality tests carried out under the

VECM & VAR– The impulse response analysis (IRA) of VAR – The Dynamic Conditional Correlation (DCC) model

developed by Engle (2002)

13

3. Data and Econometric Methods3.3 Empirical Models

Δyt = γ0 + Σγ1i(L)Δyt + Σ∂1j(L)Δx1t + Σδ1k(L)Δx2t + φECTt-1 + μjt (3)

Δyt = α0 + Σγ2i(L)Δyt + Σ∂2j(L)Δx1t + Σδ2k(L)Δx2t + ɛjt (4)

Δy1t = β0 + Σ βiΔyt-i + ΣγjΔx1t-j + ΣδkΔx2t-k + θ0yt-1 + θ1x1t-1 + θ2x2t-1 + e1t (1)

Δy2t = β0 + Σ βiΔyt-i + ΣγjΔx1t-j + ΣδkΔx2t-k + ω0yt-1 + ω1x1t-1 + ω2x2t-1 + e2t (2)

Notes: Y1t is nominal mortgage rate (WRATE)Y2t is effective mortgage rate (WAAPR)X1 is cash rate (CR)X2 is offshore funding costs (AUDLIBOR)

14

3. Data and Econometric MethodsFigure 1. Cash Rate, Mortgage Rate and AUDLIBOR Movements

15

3. Data and Econometric MethodsTable 2: Unconditional Correlations

Variables 1 2 3

Panel A: Whole sector 1 WAAPR 1.00

2 CR 0.57 1.00 3 AUDLIBOR 0.19 0.67 1.00

Panel B: Major group

1 WAAPR 1.00 2 CR 0.38 1.00 3 AUDLIBOR 0.10 0.68 1.00

Panel C: Foreign group

1 WAAPR 1.00 2 CR 0.42 1.00 3 AUDLIBOR 0.15 0.68 1.00

Panel D: Regional group 1 WAAPR 1.00 2 CR 0.56 1.00 3 AUDLIBOR 0.18 0.67 1.00

16

3. Data and Econometric MethodsTable 3: Unit Root Test Result

Description ADF PP KPSS on level 1st diff on level 1st diff on level 1st diff

Panel A: Independent series AUDLIBOR -1.20 -16.98*** -1.08 -17.13*** 1.18*** 0.19 CR -2.91** -5.34*** -2.11 -31.58*** 0.33 0.08 Panel B: Weekly mortgage series Major Group WAAPR -3.72*** -4.68*** -2.75 -26.32*** 0.464** 0.12 Foreign Group WAAPR -2.35 -27.80*** -2.50 -27.61*** 0.53** 0.14 Regional Group WAAPR -3.21** -22.76*** -3.41** -26.39*** 0.44* 0.09 The whole sector WAAPR -2.86* -10.04*** -3.38** -28.21*** 0.58** 0.10

17

4. Empirical results4.1. Long-term relationship

Table 4: The ARDL test for Cointegration under bivariate systemDependent Variables

Forcing variable

F-stat.

Conclusion (H0)

Panel A: Whole sector

WAAPR CR 16.007a Rejected

AUDLIBOR 13.254a Rejected

Panel B: Major group

WAAPR CR 8.522a Rejected

AUDLIBOR 8.685a Rejected

Panel C: Foreign group

WAAPR CR 6.708a Rejected

AUDLIBOR 6.847a Rejected

Panel D: Regional group

WAAPR CR 2.558c Fail to reject

AUDLIBOR 10.028a Rejected

18

4. Empirical results4.1. Long-term relationship

Table 5: Granger noncausality test for a long-term relationship

Long-run causality ECM

Long-run adj. speed

Long-run causality decision

Long-run coefficients

Equations t-statistic X1 X2

Panel A: Whole sector

Y -5.660** 0.078** Y ← X1 0.661***

Y -5.148** 0.060** Y← X2

0.159***

Panel B: Major group

Y -4.117** 0.070** Y ← X1 0.695***

Y -4.163** 0.064** Y← X2

0.186***

Panel C: Foreign group

Y -3.656** 0.063** Y ← X1 0.639***

Y -3.692** 0.059** Y← X2

0.164***

Panel D: Regional group

Y

No

Y -4.462** 0.063** Y← X2

0.172***

19

4. Empirical results4.2. Short-term relationshipTable 6: Granger noncausality test for a short-term relationship

Equations

Short-run causality (F-statistic)

Causality decision

X1 X2

Panel A: Whole sector    

Y 18.018*** Y ← X1

Y 6.061*** Y← X2

Panel B: Major group

Y 2.648** Y← X1

Y 3.009*** Y← X2

Panel C: Foreign group

Y 3.899*** Y← X1

Y 0.244 No

Panel D: Regional group

Y 19.336*** Y← X1

Y 6.111*** Y← X2

20

4. Empirical results4.3. Time Varying RelationshipsFigure 2. Dynamic Conditional Correlations of the whole banking sector

21

5. ConclusionThe ADRL, VEC & VAR granger causality, IRA test results determine:• The cash rate still continues to drive mortgage

rates (Hypothesis 1);• International funding costs significantly affect

mortgage rates both in long & short term (Hypothesis 2);

• The DCC test result confirms hypothesis 3: the linkage between the cash rate and mortgage rates has indeed weakened since 2006

22

• Research Feedback • Q & A