Forecasting Mortgage Securitization Risk Under Systematic Risk and Parameter Uncertainty

DOIhttp://doi.org/10.1111/j.1539-6975.2013.12009.x
Date01 September 2014
Published date01 September 2014
©The Journal of Risk and Insurance, 2013, Vol.81, No. 3, 563–586
DOI: 10.1111/j.1539-6975.2013.12009.x
Forecasting Mortgage Securitization Risk Under
Systematic Risk and Parameter Uncertainty
Daniel R¨
osch
Harald Scheule
Abstract
The global financial crisis exposed financial institutions to severe unexpected
losses in relation to mortgage securitizations and derivatives. This article
finds that risk models such as ratings are exposed to a large degree of
systematic risk and parameter uncertainty. An out-of-sample forecasting
exercise of the financial crisis shows that a simple approach addressing
both issues is able to produce ranges for risk measures consistent with
realized losses. This explains how financial markets were taken by surprise
in relation to realized losses.
Introduction
Securitization ratings applied by credit rating agencies (CRAs) were identified as a
source of the global financial crisis (GFC) (cf. Hull, 2009). Financial markets, which
rely on credit ratings, were surprised by the high levels of impairment rates and large
number of downgrades of seemingly high-quality (e.g., AAA-rated) mortgage-backed
securities (MBSs) from 2007 to 2009.
Table 1 supports this point by comparing the impairment rates for Baa-rated MBSs
with Baa-rated home equity loan securities (HELs). Both MBSs and HELs are securi-
tizations of real-estate collateralized loan portfolios.1Impairment rates for MBSs and
Daniel R¨
osch is at the Department of Statistics, University of Regensburg, 93040 Regensburg,
Germany.Harald Scheule is at the Finance Discipline Group, UTS Business School, University
of Technology, Sydney, PO Box 123, Broadway NSW 2007, Australia. The authors can be con-
tacted via e-mail: daniel.roesch@wiwi.uni-regensburg.de and harald.scheule@uts.edu.au. The
authors would like to thank two anonymous referees and the participants of the Convergence,
Interconnectedness, and Crises: Insurance and Banking conference at Temple University, and
participants of financial seminars at the Leibniz University Hannover, Hong Kong Institute
for Monetary Research, University of Technology, Sydney, University, and The University of
Melbourne. The support by the Hong Kong Institute for Monetary Research, the Centre for
International Finance and Regulation (CIFR, project number E001), and the Thyssen Krupp
foundation is gratefully acknowledged. CIFR is funded by the Commonwealth and NSW Gov-
ernments and supported by other Consortium members (see www.cifr.edu.au).
1MBSs are collateralized by prime mortgages and HEL securities are mostly collateralized by
subprime mortgages.
563
564 The Journal of Risk and Insurance
Table 1
Total Number of Observations and Impairment Rates, MBSs, and HELs, 1997–2009
All Grades Aaa–A Baa Ba B Caa–C
Year NO IR NO IR NO IR NO IR NO IR NO IR
Panel A: MBS
1997 7,938 0.0003 7,405 0.0000 312 0.0032 160 0.0000 61 0.0164
1998 8,078 0.0002 7,507 0.0000 335 0.0000 158 0.0063 74 0.0135 4 0.0000
1999 7,398 0.0000 6,814 0.0000 347 0.0000 159 0.0000 78 0.0000
2000 6,801 0.0006 6,238 0.0000 334 0.0060 147 0.0068 81 0.0123 1 0.0000
2001 6,707 0.0006 6,183 0.0000 303 0.0066 140 0.0000 75 0.0267 6 0.0000
2002 7,612 0.0004 6,982 0.0000 373 0.0054 160 0.0063 92 0.0000 5 0.0000
2003 8,851 0.0003 7,952 0.0000 533 0.0000 222 0.0090 139 0.0072 5 0.0000
2004 8,242 0.0008 7,260 0.0004 554 0.0018 254 0.0039 167 0.0120 7 0.0000
2005 10,422 0.0008 9,179 0.0000 722 0.0028 319 0.0125 195 0.0103 7 0.0000
2006 18,164 0.0004 16,036 0.0000 1,377 0.0007 479 0.0021 264 0.0152 8 0.1250
2007 26,456 0.0039 23,367 0.0001 2,169 0.0235 650 0.0492 260 0.0500 10 0.4000
2008 32,065 0.0990 27,931 0.0346 2,435 0.4427 971 0.6056 581 0.6902 147 0.9592
2009 28,759 0.2072 21,853 0.0800 2,366 0.3352 1,722 0.5064 2,038 0.8837 780 0.9551
All 177,493 0.0242 154,707 0.0089 12,160 0.0637 5,541 0.0929 4,105 0.1336 980 0.2218
Panel B: HEL
1997 1,102 0.0000 1,098 0.0000 2 0.0000 1 0.0000 1 0.0000
1998 1,748 0.0000 1,685 0.0000 51 0.0000 8 0.0000 4 0.0000
1999 2,310 0.0009 2,170 0.0000 105 0.0095 24 0.0000 11 0.0909
2000 2,699 0.0026 2,498 0.0000 152 0.0066 31 0.0645 15 0.0667 3 1.0000
2001 3,079 0.0026 2,816 0.0004 205 0.0098 40 0.0250 17 0.2353 1 0.0000
2002 3,507 0.0029 3,113 0.0000 310 0.0065 63 0.0159 15 0.1333 6 0.8333
2003 4,279 0.0051 3,647 0.0000 545 0.0128 71 0.1127 13 0.3846 3 0.6667
2004 5,666 0.0018 4,539 0.0000 1,024 0.0049 84 0.0476 16 0.0000 3 0.3333
2005 8,997 0.0017 6,808 0.0000 1,939 0.0010 221 0.0226 27 0.2593 2 0.5000
2006 14,411 0.0016 10,493 0.0000 3,225 0.0019 658 0.0106 30 0.2333 5 0.6000
2007 20,217 0.0538 14,357 0.0072 4,530 0.1000 1,225 0.3820 80 0.4875 25 0.9600
2008 21,147 0.2815 14,389 0.1234 3,605 0.4366 1,535 0.7440 1,156 0.8815 462 0.9589
2009 14,380 0.2001 8,349 0.0247 2,524 0.1830 1,274 0.4498 1,155 0.6450 1,078 0.8265
All 103,542 0.0426 75,962 0.0120 18,217 0.0594 5,235 0.1442 2,540 0.2629 1,588 0.6679
Note: This table shows the number of observations (NO) and impairment rate (IR) per rating
category for mortgage-backed securities (MBSs, Panel A) and home equity loan securitizations
(HELs, Panel B) from 1997 to 2009. The number of observed tranches increases over time,
which reflects the growth of these financial instruments during recent years. The impairment
rate increases during the GFC (2008 and 2009) and more generally from rating grades Aaa–A
(Aaa, Aa, and A) to Baa to Ba to B to Caa. HELs include to a large degree subprime mortgage
loans and the impairment risk increased to a larger degree than the one of MBSs.
HELs were well below 10 basis points before the GFC and peaked at 20.7 percent
(MBSs) and 28.2 percent (HELs) during the GFC.
Investors relied on CRA ratings and expected impairment rates similar to the historic
experience for such ratings. Credit ratings reflect expectations about future impair-
ment risk.

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT