Macroeconomic factors influencing UK household loan losses
| Date | 09 November 2012 |
| Published date | 09 November 2012 |
| Pages | 385-401 |
| DOI | https://doi.org/10.1108/13581981211279345 |
| Author | Minh T.H. Dinh,Andrew W. Mullineux,Peter Muriu |
| Subject Matter | Accounting & finance |
Macroeconomic factors
influencing UK household
loan losses
Minh T.H. Dinh
Department of Accounting and Finance, Nha Trang University,
Khanh Hoa, Vietnam, and
Andrew W. Mullineux and Peter Muriu
Department of Accounting and Finance, University of Birmingham,
Birmingham, UK
Abstract
Purpose – The purpose of this paper is to investigate the effects of macroeconomic factors on secured
and unsecured household loans from UK banks.
Design/methodology/approach – The approach uses Vector auto-regression models to test the
relationship between macroeconomic factors such as interest rates, house prices, unemployment rates,
disposable income and bank write-offs to discern the main factors which could impact on banks’ losses.
Findings – This paper identifies several macroeconomic factors that influence loan losses. The
influence however depends on the type of arrears. Changes in house prices, interest rates and
unemployment rates have a significant impact on secured loans. There is however, minimal impact on
unsecured loans. Unemployment stands out as the major factor thatinfluences both mortgage and credit
card arrears. The estimated results show that the main factors impacting on credit cards are disposable
income and unemploymentrates, while changes in interest rates have no impact on credit card write-offs.
Originality/value – This paper’s value lies in providing methods by which commercial banks could
manage household loans better by reducing the effects of macroeconomic factors.
Keywords VAR model, Interestrates, House price, Unemployment rates,Disposable income,
Bank write-offs,Loans, Credit, United Kingdom
Paper type Research paper
I. Introduction
Prior to the 2007 financial crisis, there was a sustained increase in household debt,
especially mortgage and credit cards borrowing (see Figures A3 and A4 in Appendix 5).
Household loans accounted for 20 per cent of the aggregate balance sheet as at the end
of 2005 in the major UK banks. Although growth in household loans contributed to
higher household consumption by making funds available to buy durable goods and
finance education, it neverthe less exacerbated over-indebt edness. Considerable
defaults in repayment of loans were recorded in the years 2007-2009. For example,
the number of write-offs of household lending increased leading to a decline in profits
for UK banks (see Figure A5 in Appendix 5). Moreover, falls in house prices close to
the end of 2007 (see Figure A6 in Appendix 5) also led to a significant reduction in the
values of properties owned by the borrowers. Households’ loans have therefore become
riskier because property values which secured the loans have plummeted to lower
levels relative to the loans. UK banks’ response has been to restrict loans to households
to reduce risk (see Figure A7 in Appendix 5).
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1358-1988.htm
UK household
loan losses
385
Journal of Financial Regulation and
Compliance
Vol. 20 No. 4, 2012
pp. 385-401
qEmerald Group Publishing Limited
1358-1988
DOI 10.1108/13581981211279345
Declining house prices in 2007 and 2008 impacted on particular borrowers with high
loan to value (LTV) ratios and buy-to-let (BTL) landlords (Bank of England, 2008).
Figure A6 (see Appendix 5), shows that on average house prices fell from £200,623 in
2007:4 to £158,359 in 2009:1 which was a reduction of about 21 per cent. LTV was still
too high during the financial crisis period (see Figure A8, Appendix 5). In 2008, LTV
ranged between 90 and 100 per cent, or even higher than 100 per cent. Hence, the
amounts of loans were higher than the value of properties. The banks were therefore
exposed to higher credit particularly when the market value of properties declined.
Although there was a slight restriction on LTV above 90 per cent in 2009, there still
were some mortgage loans in which LTV was above 100 per cent (Bank of England,
2009). The “together and together connections products” launched by Northern Rock
offered customers total borrowing of up to 125 per cent LTV[1].
“Over time, banks progressively took on more credit risk by lending to, for example,
households with high loan to income (LTI) ratios” (Bank of England, 2008; see also
Figure A9 in Appendix 5). Higher LTI ratios, particularly for new borrowers, implies
that income gearing is more sensitive (Hancock and Wood, 2004). The effect of
recession could also have contributed to lower expected future income[2] (Berry and
Williams, 2009, p. 199). This paper is therefore, timely because of the important
macroeconomic and financial implications that household borrowing has on the overall
economy. It contributes to the existing literature by providing central banks
supervisors with forecasts for macro stress tests based on prevailing macroeconomic
conditions. This is also important for financial stability monitoring.
The remainder of this study proceeds as follows: in the next section, we review the
related empirical literature. Section III describes the research methodology. In
Section IV we describe data and measurements of our variables of interest. In Section V
we present the empirical results. Conclusions and policy suggestions are offered in the
final section.
II. Literature review
There is a vast literature that has tested the relationship between macroeconomic
factors and the probabilities of default on the repayment of household debts. Much of
the literature is however on credit scoring methods. Literature documents two main
strands of empirical modelling of loan losses. The first strand comprises of models that
describe the probability distribution of loan losses that a bank may incur. The aim of
these empirical models is to describe the unconditional probability distribution of loan
losses. As such, majority of these models do not directly show the link between the
economic cycle and the amount of loan. Models in this category include Crouhy et al.
(2000) and Wilson (1997).
The second strand of empirical models comprises of those that seek to provide a
functional relationship between banks’ loan losses and macroeconomic conditions.
Models in this category include Whitley et al. (2004) and Hoggarth et al. (2005),
cross-country studies such as Bikker and Hu (2002) and Pesola (2001, 2005). The model
presented in this paper belongs to this category.
Hoggarth et al. (2005) employed the VAR approach to test UK fragility by
examining the relationship between macroeconomic variables and loan write-offs.
They modelled loan write-offs as a function of household disposable income,
unemployment rate, household income gearing, and other macroeconomic factors such
JFRC
20,4
386
To continue reading
Request your trial