Hibbeln, Ma in; Me zle , Ral ; Os e kamp, We ne
A icle — Published Ve sion
No on he same page: comp ehensibili y o MBS
in es men p ospec uses
Re iew o De i a i es Resea ch
P o ided in Coope a ion wi h:
Sp inge Na u e
Sugges ed Ci a ion: Hibbeln, Ma in; Me zle , Ral ; Os e kamp, We ne (2025) : No on he same
page: comp ehensibili y o MBS in es men p ospec uses, Re iew o De i a i es Resea ch, ISSN
1573-7144, Sp inge US, New Yo k, NY, Vol. 28, Iss. 2,
h ps://doi.o g/10.1007/s11147-025-09213-8
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/323690
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h p://c ea i ecommons.o g/licenses/by/4.0/
Vol.:(0123456789)
Re iew o De i a i es Resea ch (2025) 28:9
h ps://doi.o g/10.1007/s11147-025-09213-8
No on hesame page: comp ehensibili y o MBS
in es men p ospec uses
Ma inHibbeln1 · Ral Me zle 1· We ne Os e kamp1
Accep ed: 20 Ap il 2025 / Published online: 26 June 2025
© The Au ho (s) 2025
Abs ac
We in es iga e whe he o igina o s o mo gage-backed secu i ies lowe he com-
p ehensibili y o disclosu e documen s o ob usca e low secu i y quali y. U ilizing
commonly used complexi y p oxies p ima ily based on he olume o he in es men
p ospec us, we ind no impac on he pe o mance o Eu opean mo gage-backed
secu i ies. In con as , ocusing on he comp ehensibili y as a combina ion o p o-
spec us olume and ex eadabili y, we ind ha o igina o s ob usca e low secu i y
quali y by lowe ing he comp ehensibili y o he in es men p ospec us, esul ing in
g ea e de aul s and lowe e u ns. In es o s only pa ially p ice hese dimensions
o complexi y: Since he inancial c isis, in es o s ha e been demanding a signi i-
can isk p emium o p ospec us olume bu no o ex eadabili y. These indings
unde sco e he impo ance o assessing comp ehensibili y o disclosu e documen s
beyond adi ional complexi y measu es and highligh po en ial implica ions o
o igina o s, in es o s and egula o s in communica ing, unde s anding and mi iga -
ing he isks associa ed wi h complex inancial ins umen s.
Keywo ds Complexi y· Disclosu e· P ospec us comp ehensibili y· Secu i y
design· Tex analysis
JEL Classi ica ion D82· D83· G01· G12· G14· G23· M41
We hank Fabian Rendchen o his suppo wi h he da a collec ion. We a e g a e ul o help ul
commen s om ou discussan s Ch is ine Bangsgaa d and Johannes K iebel. We also hank he
pa icipan s o he Pa is Financial Managemen Con e ence 2022, he Banking Resea ch Wo kshop
Müns e 2022, and he Ca di Fin ech Con e ence 2022 o help ul commen s. This wo k was
suppo ed by he Ge man Resea ch Founda ion (DGF) [g an numbe 434130478].
* Ma in Hibbeln
ma [email p o ec ed]
Ral Me zle
[email p o ec ed]
1 Me ca o School o Managemen , Uni e si y o Duisbu g-Essen, Lo ha s . 65, 47057Duisbu g,
Ge many
M.Hibbeln e al.
9 Page 2 o 37
1 In oduc ion
“A 200-page p ospec us ha nobody e e eads and a one-page lye wi h a po o
gold on i and a lo o ine p in ”—Ma in Whea ley, Chie Execu i e o he Secu i-
ies and Fu u es Commission o Hong Kong, abou disclosu e documen s o complex
de i a i e inancial p oduc s (May 28, 2010).
Complex secu i iza ions expe ienced a signi ican su ge in de aul s du ing he
inancial c isis o 2007/08 (G i in e al., 2014). The empi ical li e a u e inds e i-
dence o a complexi y channel in p e-c isis U.S. secu i iza ion ma ke s: O igina-
o s s a egically inc eased complexi y o ob usca e low secu i iza ion quali y, wi h
“complexi y” commonly quan i ied using olume-based p oxies de i ed om he
in es men p ospec us and he numbe o anches (Fu ine, 2014; Ghen e al.,
2019).
In con as , we ocus on a dis inc dimension o complexi y: he comp ehensi-
bili y o he in es men p ospec us, as i is he main disclosu e documen and hus
se es as he main in o ma ion ansmission ool o he o igina o owa d in es o s.
We measu e p ospec us comp ehensibili y as he combina ion o p ospec us olume
and ex eadabili y. While a ecen s and o he li e a u e s udies he ex eadabili y
o co po a e disclosu e documen s in isola ion (e.g., Hwang & Kim, 2017; Li, 2008;
Lo e al., 2017), we con end ha ha d- o- ead bu sho in es men p ospec uses
should s ill be comp ehensible o in es o s (and ice e sa). Howe e , o long
p ospec uses, low ex eadabili y may be a majo obs acle o ully unde s anding
he secu i iza ion’s legal e ms. We con ibu e o he li e a u e by showing ha low
p ospec us comp ehensibili y is associa ed wi h g ea e de aul s and lowe e u ns
on he Eu opean mo gage-backed secu i y (MBS) ma ke . Ye , p ospec us comp e-
hensibili y is only pa ially p iced a e he inancial c isis: While in es o s demand
a signi ican isk p emium o p ospec us olume, hey do no demand any isk p e-
mium o he eadabili y o he ex .
We in e p e he poo e pe o mance o MBS wi h less comp ehensible p ospec-
uses as a s a egic a emp by o igina o s o ex ac alue om in es o s by ob us-
ca ing he low quali y o he sold secu i ies. This in e p e a ion is con ingen on wo
p emises: Fi s , he e mus exis some in es o s who a e no sophis ica ed enough o
ully unde s and he legal e ms se ou in some o he pa icula ly incomp ehensible
p ospec uses. Al hough mos MBS in es o s a e ins i u ions, we belie e his condi-
ion is me o se e al easons. Mos impo an ly, he e is a ia ion in he sophis ica-
ion o ins i u ional in es o s, which is dependen on he esou ces ha hese ins i u-
ions alloca e o analyzing complex secu i iza ions. E en la ge ins i u ions ha do
no lack he necessa y expe ise o p ope ly assess he secu i iza ion quali y may
ha e op ed no o spend hese esou ces, o e - elying on he in la ed a ings o c edi
a ing agencies (Co al, 2009). This iew aligns wi h Ma in Whea ley’s assessmen ,
hen Chie Execu i e o he Secu i ies and Fu u es Commission o Hong Kong,
ega ding he disclosu e documen s o complex de i a i es: “A 200-page p ospec us
ha nobody e e eads and a one-page lye wi h a po o gold on i and a lo o ine
p in ” (Cheng, 2010). To his poin , many Eu opean ins i u ions became insol en
o equi ed bailou s due o se e ely unde es ima ing he isks associa ed wi h hei
No on hesame page: comp ehensibili y o MBS in es men … Page 3 o 37 9
pu chased secu i iza ions ( o name a ew: Ge man Hypo Real Es a e, Swiss UBS,
and UK Royal Bank o Sco land).
Second, he e mus exis some o igina o s who a e awa e ha some in es o s lack
he necessa y sophis ica ion o he willingness o ully assess he secu i iza ion qual-
i y, especially i he p ospec us is incomp ehensible, and o igina o s a e willing o
ex ac alue om in es o s by designing less comp ehensible p ospec uses. No e
ha hese o igina o s may no (only) choose o ac i ely dec ease p ospec us comp e-
hensibili y o low-quali y secu i iza ions bu could also e ain om making e o s
o inc ease p ospec us comp ehensibili y o low-quali y secu i iza ions, hus no
p uning he p ospec us leng h and lea ing he ex s uc u e in a mo e complex s a e.
While we belie e ha his second condi ion is me , we make ex ensi e e o s o ule
ou al e na i e explana ions o ou main esul s. We he e o e es wo al e na i e
in e p e a ions sugges ing ha low p ospec us comp ehensibili y is me ely inciden-
al. Howe e , we ind no e idence linking low p ospec us comp ehensibili y o he
ex-an e iskiness o he unde lying loans, he complexi y o he secu i iza ion s uc-
u e, o loan complexi y. We he eby ule ou ha low p ospec us comp ehensibili y
is jus a byp oduc o complex secu i iza ion s uc u es o loans.
The complexi y channel can be heo e ically mo i a ed h ough he exis ence
o asymme ic in o ma ion wi hin secu i iza ion ma ke s. As a esul o asymme -
ic in o ma ion, o igina o s sys ema ically selec ed poo ly pe o ming loans in o
secu i iza ions (Jiang e al., 2014; K uge , 2018; Pu nanandam, 2011), educed
sc eening e o s o loans which we e o igina ed o be sold ( o he o igina e- o-
dis ibu e model, see Keys e al., 2010; Pu nanandam, 2011), and moni o ed hese
loans less han hose held on he o igina o ’s balance shee (Be nd & Gup a, 2009;
Mian & Su i, 2009; Wang & Xia, 2014). In he con ex o asymme ic in o ma ion,
he deg ee o secu i iza ion complexi y ep esen s he associa ed sea ch cos s ha
an in es o needs o pay o assess he quali y-adjus ed p ice o he secu i y (Ghen
e al., 2019). Theo e ical models show ha s a egically sh ouding nega i e a ib-
u es o p oduc s can be he op imal selle beha io i unawa e buye s exis in he
ma ke (Gabaix & Laibson, 2006). Fu he mo e, a educ ion in he quali y o goods
is no obse ed by buye s i sea ch cos s ou weigh he buye ’s u ili y o he good
(Ellison & Ellison, 2009). E en hough complexi y is no necessa ily a ea u e o
low-quali y p oduc s, p oduc s a e mo e likely o be bo h mo e complex and o low
quali y i he demand o hem is high, which was, e.g., he case o MBS p io o he
inancial c isis (As iyan e al., 2022). Empi ical s udies o he p e-c isis U.S. MBS
ma ke ha e ound e idence o a s a egic ob usca ion o poo secu i iza ion quali y
h ough he complexi y channel: by aising in es o s’ sea ch cos s o low-quali y
secu i iza ions, o igina o s aimed o ex ac alue om in es o s. In es o s did no
an icipa e his beha io , as hey did no demand a isk p emium o buying hese
mo e complex and iskie secu i iza ions (Fu ine, 2014; Ghen e al., 2019).
We speci ically chose o s udy MBS because hey a e a p ime example o com-
plex inancial p oduc s ha a e no easily unde s ood by in es o s, making hem
pa icula ly suscep ible o s a egic ob usca ion and oppo unis ic beha io by
o igina o s in he con ex o asymme ic in o ma ion. MBS also play a c ucial
ole in global inancial ma ke s, se ing as a key in es men ehicle o ins i u-
ions and a i al sou ce o liquidi y o mo gage lende s. Agg ega e annual
M.Hibbeln e al.
9 Page 4 o 37
issuance olumes o Eu opean MBS consis en ly exceed 100 billion Eu os,
unde lining hei economic impo ance. Fu he mo e, ou ocus on he Eu opean
ma ke allows us o in es iga e he impac o egula o y in e en ions on p o-
spec us comp ehensibili y wi hin a dis inc egula o y en i onmen . Speci ically,
he ecen in oduc ion o he EU Secu i iza ion Regula ion (EUSR) p o ides a
unique se ing o assess how measu es aimed a enhancing simplici y, anspa -
ency, and s anda diza ion a ec he comp ehensibili y o p ospec uses.
We con ibu e o he esea ch on complex inancial secu i ies by examining
in de ail he complexi y o he associa ed disclosu e documen s. Speci ically, we
s udy he complexi y o secu i iza ion in es men p ospec uses in h ee dimen-
sions: ype o complexi y ( ex eadabili y and p ospec us comp ehensibili y),
pe iod (p e- s. pos -c isis), and egula o y se ing (EU). In his con ex , we aim
o answe h ee esea ch ques ions: (I) Did o igina o s s a egically inc ease com-
plexi y in Eu opean MBS be o e he inancial c isis o ob usca e low secu i iza-
ion quali y (“complexi y channel”), and was his beha io an icipa ed and he e-
o e p iced by in es o s? (II) Did o igina o s and in es o s change hei beha io
a e he inancial c isis due o lea ning e ec s? (III) Do o igina o s ob usca e
secu i iza ion quali y beyond he commonly used complexi y measu es by lowe -
ing p ospec us comp ehensibili y, and is his beha io an icipa ed and he e o e
p iced by in es o s?
(I) Analogous o he indings o Fu ine (2014) and Ghen e al. (2019) o he
p e-c isis U.S. ma ke , we in es iga e whe he he complexi y channel also
exis ed on he Eu opean MBS ma ke p io o he inancial c isis. Build-
ing on his, we e i y whe he in es o s an icipa ed and he e o e p iced
he complexi y channel in Eu opean MBS. Al hough in es o s could no
obse e he exis ence o he complexi y channel ex-an e, we can check om
an ex-pos pe spec i e whe he hei p icing o complexi y is consis en wi h
o igina o s using complexi y o ob usca ion. This is an impo an ques ion
because he U.S. and Eu opean secu i iza ion ma ke s a e undamen ally
di e en in e ms o hei de elopmen s and ma ke s uc u e. Mos no ably,
o da e, i emains unclea o wha ex en o igina o s used he o igina e- o-
dis ibu e model in he Eu opean ma ke . Albe azzi e al. (2015) ind ha
I alian mo gage loans pe o med be e i hey we e secu i ized, which hey
explain h ough o igina o s’ epu a ional conce ns, con adic ing he con-
ce ns o he o igina e- o-dis ibu e model.
(II) P e ious s udies ha e no ye in es iga ed whe he he complexi y channel
pe sis ed and whe he in es o s changed hei p icing beha io o he pe iod
a e he inancial c isis due o lea ning e ec s. In pa icula , because o he
ex ensi e de aul s o complex secu i iza ions in he U.S. du ing he c isis,
in es o s migh ha e ecognized he high isk o hese p oduc s and inc eased
he isk p emium hey equi e o holding hem. An icipa ing his beha io o
in es o s, o igina o s migh ha e, in u n, e ained om using complexi y
o ob usca ion.
No on hesame page: comp ehensibili y o MBS in es men … Page 5 o 37 9
(III) To measu e complexi y, p e ious s udies ha e commonly elied on p oxies
ha accoun o he complexi y o bo h (a) he secu i ized loan pool and
(b) he secu i iza ion s uc u e. These p oxies a e mos ly de i ed om he
olume o he in es men p ospec us (Fu ine, 2014; Ghen e al., 2019).
While he olume o he p ospec us has been used equen ly, he comp e-
hensibili y o he p ospec us has no ye been in es iga ed. This is su p ising
conside ing ha he p ospec us is he main disclosu e documen and hus
he main in o ma ion ansmission ool o he o igina o owa d in es o s.
O igina o s migh ha e speci ically lowe ed p ospec us comp ehensibili y
o ob usca ion, as his is ela i ely cheap compa ed o inc easing s uc u ing
o loan pool complexi y. To his poin , i is ex-an e no clea i a ia ion in
p ospec us comp ehensibili y is an ac i e decision o he o igina o o a he
simply an inciden al byp oduc o mo e gene al secu i iza ion complexi y.
While he e is no esea ch conce ning he comp ehensibili y (as he combina-
ion o olume and ex eadabili y) o in es men p ospec uses, a ecen s and o
he li e a u e in es iga es he ex eadabili y o co po a e disclosu e documen s,
mos ly annual epo s. These s udies ind ha less eadable and he eby less ans-
pa en annual epo s a e ela ed o lowe i m ea nings, highe ea nings ola il-
i y, and lowe i m alua ions by in es o s, sugges ing ha i ms seek o ob usca e
nega i e in o ma ion while in es o s an icipa e and he eby p ice his beha io
(Hwang & Kim, 2017; Li, 2008; Lo e al., 2017). Consis en wi h hese indings,
Bonsall and Mille (2017) demons a e ha less eadable inancial disclosu es a e
linked o less a o able c edi a ings, highe a ing disag eemen , and inc eased
cos s o deb , unde sco ing he signi ican in luence o disclosu e eadabili y on
c edi ma ke ou comes. Rega ding secu i iza ions, E an e al. (2017) ind ha
egula o ily-en o ced anspa ency equi emen s lead o highe -quali y loans
being chosen o secu i iza ion. Con i ming hese esul s, Billio e al. (2023) ind
ha he new EU Secu i iza ion Regula ion (EUSR) adop ed in 2017 is associa ed
wi h signi ican ly imp o ed c edi quali y.
Apa om ou s udy, he e exis wo o he a icles s udying he con en s o
secu i iza ion p ospec uses: Zhang e al. (2023) show o U.S. secu i iza ions
issued be o e he inancial c isis ha ex ual con en s in he isk- ac o sec ion
p edic subsequen losses and we e no p iced by in es o s. Debene e al. (2021)
p o ide a i s in es iga ion o he ex eadabili y o secu i iza ion in es men
p ospec uses; hey ind ha low ex eadabili y impai s he abili y o in es o s
and c edi a ing agencies o co ec ly assess isk and induces g ea e seconda y
ma ke p ice ola ili y. Howe e , hese s udies nei he in es iga e i o igina o s
lowe ex eadabili y o ob usca ion no examine he le el o in es o p icing in
ela ion o ex eadabili y.
Me hodologically, we i s adop Ghen e al. (2019) p ocedu e ega ding da a
i ems, a iable cons uc ion, and sample selec ion o ensu e ha any di e ences
in ou p e-c isis esul s a e a ibu able o ma ke di e ences be ween he U.S.
and Eu opean esiden ial MBS (RMBS) ma ke s a he han o me hodologi-
cal di e ences. Thus, we ob ain he same Bloombe g da a i ems bu o Eu o-
pean RMBS issued be ween 2003 and 2020. We cons uc one in ensi e measu e
M.Hibbeln e al.
9 Page 6 o 37
(De aul ) and one ex ensi e measu e (In e nal Ra e o Re u n) o secu i y pe o -
mance. We measu e complexi y p icing se by in es o s wi h he C edi Sp ead
o e he 3-mon h EURIBOR. Then, we eg ess hese a iables on deal-le el com-
plexi y, oge he wi h a ious con ol a iables and ixed e ec s o a ing, yea
o issuance, coun y, and o igina o . In addi ion o p oxying complexi y by com-
mon measu es, we use Fog Index o measu e ex eadabili y, which is an es ab-
lished p oxy o assessing he eadabili y o inancial documen s (Bushee e al.,
2018; Dye e al., 2017; Li, 2008; Lo e al., 2017). We u he exploi egula o ily
induced c oss-sec ional a ia ion in deal complexi y s emming om he EUSR:
On he deal-le el, we use he deals’ p ospec uses and documen s o iden i y he
le el o adhe ence he espec i e deal has owa ds 48 unique egula ion ea u es
ela ing o ei he simplici y, anspa ency, o s anda diza ion (STS).
We p o ide b oad insigh s in o he exis ence, pe sis ence, in es o an icipa ion,
egula o y in luences on, and ype o he complexi y channel on he Eu opean RMBS
ma ke . When measu ing secu i iza ion complexi y wi h p oxies commonly used in
he empi ical li e a u e, we do no ind any e idence in a o o o igina o s using com-
plexi y o ob usca ion. In con as , low p ospec us comp ehensibili y is signi ican ly
associa ed wi h g ea e secu i y de aul s and lowe secu i y e u ns, sugges ing ha
o igina o s ob usca ed low secu i iza ion quali y by lowe ing p ospec us comp ehensi-
bili y. The e ec s a e economically meaning ul: Fo a one s anda d de ia ion dec ease
in eadabili y, he annual e u ns a e on a e age 18 basis poin s lowe when conside -
ing a p ospec us wi h a e age p ospec us leng h. Fo a p ospec us whe e he p ospec-
us leng h is one s anda d de ia ion abo e he a e age, his e ec inc eases o 34 basis
poin s. This “p ospec us comp ehensibili y channel” does no pe sis o secu i ies
issued a e he inancial c isis, indica ing ha o igina o s expec ed in es o s o ei he
e ain om buying o demand a la ge isk p emium o secu i ies wi h low p ospec us
comp ehensibili y. Howe e , he p ospec us comp ehensibili y channel is only pa ially
p iced pos -c isis: In es o s demand a signi ican isk p emium o 24 basis poin s o
an addi ional 100 pages in p ospec us leng h, which speaks in a o o lea ning e ec s
based on he documen ed complexi y channel o U.S. secu i iza ions; in con as , hey
do no demand a isk p emium o ex eadabili y and comp ehensibili y o he in es -
men p ospec us. Recen egula o y in e en ions, mos p ominen ly in he o m o he
EUSR, ha e no signi ican e ec on enhancing p ospec us comp ehensibili y, e en
hough educing deal complexi y was one o he egula ions’ majo goals sugges ing
ha p ospec us comp ehensibili y is a dis inc o m o complexi y no ye add essed by
egula o s.
2 S uc u e o heEu opean MBS ma ke
We s udy he Eu opean RMBS ma ke , which is simila o he U.S. RMBS ma -
ke in many espec s. A ypical deal ollows he pay- h ough concep , meaning ha
p io i ized senio anches ecei e incoming cash lows om he unde lying loans
be o e subo dina ed junio anches. Ano he simila i y is ha anches a e usually
a ed by mo e han one c edi a ing agency, excep o he lowes - anked equi y
anches, which a e o en e ained by he o igina o s. In es o s in hese ma ke s a e
No on hesame page: comp ehensibili y o MBS in es men … Page 7 o 37 9
p edominan ly ins i u ions such as banks, insu ance companies, o unds. Simila
o he U.S. ma ke , he deal lead manage s a e la ge in es men banks, which also
include U.S. banks.
Howe e , he e a e also undamen al di e ences be ween he Eu opean and U.S.
RMBS ma ke s in e ms o ma ke s uc u e and his o ical ma ke de elopmen .
Rega ding he ma ke s uc u e, he e is li le go e nmen pa icipa ion in he Eu o-
pean secu i iza ion ma ke , compa ed o he high ele ance o go e nmen -spon-
so ed en i ies in he Uni ed S a es (Al unbas e al., 2009). Fu he mo e, secu i ized
loans on a e age a e less isky in Eu ope, as subp ime lending does no occu a he
same scale as in he Uni ed S a es, and i is no common o g an no-documen a ion
loans. Finally, in ou sample o Eu opean RMBS, he e is almos always a single
loan pool unde lying all anches simul aneously a he han mul iple loan pools
unde lying di e en se ies o anches, as is o en he case o U.S. RMBS deals
(Ghen e al., 2019). In his ega d, Eu opean RMBS a e less complex. Rega ding
his o ical ma ke de elopmen , in he Uni ed S a es, house p ices, mo gage deb ,
and secu i iza ion olume inc eased simul aneously in he un-up o he inancial
c isis (Le i in & Wach e , 2012). In con as , he p e-c isis housing boom in Eu ope
was p ima ily unded by co e ed bonds, whe e he unde lying loans emained on
he o igina o ’s balance shee , wi h o igina o s ha ing an obliga ion o epay in es-
o s. MBS we e gene ally less ele an o unding han in he Uni ed S a es, and
he o igina e- o-dis ibu e model was no a majo con ibu o o he c isis in Eu ope
(Wach e , 2015).
Rega dless o hese di e ences, he gene al exponen ial g ow h o he Eu opean
secu i iza ion ma ke was simila o ha o he U.S. ma ke , albei he Eu opean
ma ke lagged behind. While he U.S. p i a e-label RMBS ma ke collapsed om
i s peak a o e $500 billion issuance olume in 2007 o app oxima ely $250 billion
in 2008 (Ghen e al., 2019), he Eu opean MBS ma ke eached i s peak in 2008
a almos €400 billion and collapsed o app oxima ely €240 billion in 2009.1 The
g ow h o secu i iza ion ma ke s be o e he c isis was a ibu ed o a high demand
o sa e secu i ies om ins i u ional in es o s (Al unbas e al., 2009), which was
pa ly due o in es o s being a ing-cons ained, such as banks and insu ance compa-
nies ha a e capi al-cons ained in hei isk- aking h ough a ing-based egula ion.
A e he inancial c isis, Eu opean and U.S. egula o s iden i ied an o e eliance
o in es o s on a ings in secu i iza ion ma ke s (Co al e al., 2009). Since hen, se -
e al egula ions ha e been implemen ed in bo h ma ke s. Mos no ably, in he con-
ex o secu i iza ion complexi y, in 2014 and 2015, he Eu opean Banking Au ho -
i y (EBA) and he Basel Commi ee on Banking Supe ision (BCBS) p oposed he
concep o simple, anspa en , and s anda dized (STS) secu i iza ions o discus-
sion (EBA, 2014; BCBS, 2015). The STS concep was implemen ed h ough he
EU Secu i iza ion Regula ion (EUSR), which was adop ed in Decembe 2017 and
implemen ed in Janua y 2019, wi h he goal o e i alizing he secu i iza ion ma ke
(EU, 2017). To achie e his goal, in es o s’ isk assessmen should be imp o ed by
mi iga ing in o ma ion asymme y and educing deal complexi y. The EUSR con-
ains legally binding minimum equi emen s, including basic STS ea u es, as well
1 See Figu eC.1 in he Online Appendix.
M.Hibbeln e al.
9 Page 8 o 37
as op ional equi emen s o deals o ob ain he STS label. Fi s e idence shows ha
in es o s end o ocus on he new quali y label ins ead o he secu i y design, bu
he la e is mo e impo an o o igina o s’ beha io and ul ima ely o he unde ly-
ing loan pe o mance (Hibbeln & Os e kamp, 2024).
We also exploi egula o ily induced c oss-sec ional a ia ion in deal complex-
i y s emming om he EUSR: On he deal-le el, we use he deals’ p ospec uses
and documen s o iden i y he le el o adhe ence he espec i e deal has owa ds 48
unique egula ion ea u es ela ing o ei he simplici y, anspa ency, o s anda di-
za ion (STS). We hen show ha he EUSR had no signi ican e ec on enhancing
p ospec us comp ehensibili y, e en hough educing deal complexi y was one o he
egula ions’ majo goals. This sugges s ha p ospec us comp ehensibili y is a dis-
inc o m o complexi y, which is no signi ican ly ela ed o secu i iza ion and loan
pool complexi y and is no ye add essed by egula o s.
3 Hypo heses
The e a e se e al equilib ium conside a ions unde lying ou subsequen hypo hesis
de elopmen . On a pe ec capi al ma ke wi h ull in o ma ion, o igina o s can-
no ob usca e low secu i iza ion quali y h ough he s a egic inc ease o complex-
i y, he eby also making he p icing o complexi y unnecessa y. De ia ing om his
pe ec capi al ma ke assump ion, we assume ha o igina o s ha e he oppo uni y
o ob usca e low secu i iza ion quali y h ough s a egically inc easing complexi y.
This ob usca ion beha io yields hem a posi i e NPV, as hey ex ac alue om he
in es o s in a ze o-sum ype o game. Howe e , his alue ex ac ion only wo ks i
in es o s do no conside high complexi y as a signal o ob usca ion beha io , and
he eby do no ( ully) p ice complexi y. I he in es o s simply ecognize he pos-
sibili y o complexi y being used o s a egic ob usca ion, hey will always p ice
complexi y ega dless o he o igina o ac ually using i o ob usca ion. This is
because we assume in es o s ex-an e canno dis inguish when complexi y is s a egi-
cally used and when i is jus inciden al.
Speci ically, we de elop hypo heses ela ed o ou h ee main esea ch ques ions:
(I) Did o igina o s s a egically inc ease complexi y in Eu opean RMBS be o e
he inancial c isis o ob usca e low secu i iza ion quali y (“complexi y channel”),
and was his beha io an icipa ed and he e o e p iced by in es o s? (II) Did o igi-
na o s and in es o s change hei beha io a e he inancial c isis due o lea ning
e ec s? (III) Do o igina o s ob usca e secu i iza ion quali y beyond he commonly
used complexi y measu es by lowe ing p ospec us comp ehensibili y, and is his
beha io an icipa ed and he e o e p iced by in es o s?
(I) Fo he p e-c isis U.S. secu i iza ion ma ke , he e is e idence o o igina o s
s a egically ob usca ing poo secu i iza ion quali y h ough inc eased complexi y.
By aising in es o s’ sea ch cos s o low-quali y secu i iza ions, o igina o s aimed
o ex ac alue om in es o s (Fu ine, 2014; Ghen e al., 2019). In es o s did no
an icipa e his use o complexi y o ob usca ion, meaning ha hey did no demand
a highe isk p emium o mo e complex secu i iza ions. This in es o p icing beha -
io is no di ec ly ela ed o o igina o s ac ually using complexi y o ob usca ion,
No on hesame page: comp ehensibili y o MBS in es men … Page 15 o 37 9
a iables a e e y high: he longes p ospec us is 725 pages long, and he highes num-
be o Glossa y Te ms is o e 1000.
Table 1 Summa y s a is ics o
deal complexi y
This able epo s he summa y s a is ics o ou deal-le el complex-
i y a iables. Uni o obse a ion is deals. File Size is measu ed in
megaby es. Va iable de ini ions a e gi en in Table8
obs mean sd min p25 p50 p75 max
Deal T anches 625 4.5 2.9 1 3 4 6 31
Pagesmpool 625 17.7 11.5 1 10 15 22 72
Pageswa e all 625 17.3 8.0 2 12 16 20 64
Glossa y Te ms 625 315 158 0 223 290 406 1017
File Size (MB) 625 1.6 1.2 0.2 0.9 1.3 2.0 13.9
To al Pages 625 223 82 56 174 214 256 725
Fog Index 625 22.0 1.3 16.9 21.2 22.1 22.9 25.4
Table 2 T anche pe o mance and c edi sp ead summa y s a is ics
This able p esen s anche-le el summa y s a is ics o he pe o mance a iables De aul and IRR, as
well as he p icing a iable C edi Sp ead. Uni o obse a ion is anches. Va iable de ini ions a e gi en
in Table8. The p esen ed a ing ca ego ies e e o he a ing a he ime o anche issuance
Obs Mean sd p5 p25 p50 p75 p95
Panel A: De aul
AAA 817 1.84
AA 391 1.53
A 329 3.04
BBB 284 8.10
To al 1821 2.97
Panel B: IRR
AAA 817 1.63 2.35 0.05 0.64 1.69 2.55 3.92
AA 391 1.79 1.24 0.30 1.07 1.62 2.46 3.35
A 329 2.19 1.34 0.28 1.34 1.98 2.93 4.44
BBB 284 3.20 2.10 1.02 1.93 2.50 3.90 7.20
To al 1821 2.01 2.03 0.13 1.09 1.86 2.77 4.70
Panel C: C edi Sp ead
AAA 817 65.3 56.2 7.0 19.7 53.7 95.0 155.0
AA 391 131.7 99.6 17.0 40.8 115.0 200.0 300.0
A 329 149.2 124.5 24.0 50.0 110.0 201.9 400.0
BBB 284 204.2 179.8 45.0 78.3 150.0 253.6 600.3
To al 1821 116.4 118.2 10.0 35.0 80.1 150.3 375.0
M.Hibbeln e al.
9 Page 16 o 37
The highes Fog Index is 25.4. Using he o iginal in e p e a ion o Fog Index, his
would imply ha he eade is equi ed o ha e 25.4yea s o o mal educa ion o unde -
s and his p ospec us in he i s eading. This in e p e a ion is no pa icula ly use ul in
ou con ex , so we will e ain om in e p e ing Fog Index in ha way and ins ead ocus
on he dis ibu ion o he a iable. The mean o Fog Index (22.0) is sligh ly highe han
he means obse ed in compa able inance and accoun ing s udies (e.g., Dye e al.,
2017; Li, 2008; Lo e al., 2017; Lough an & McDonald, 2014), which epo a e -
age Fog Index alues o inancial documen s anging om 18 o 21. While Fog Index
does no g ea ly a y o e ime, we obse e a no able c oss-sec ional a ia ion. Fig-
u e2 exempli ies his obse a ion: Panel A p esen s he subsec ion “Cash Collec ion
A angemen s” o a p ospec us wi h a low measu ed eadabili y (Fog Index = 25.4).
Panel B p esen s simila in o ma ion o a p ospec us wi h high a measu ed eadabili y
(Fog Index = 18.2).
Table2 epo s he summa y s a is ics o ou pe o mance measu es De aul and
IRR as well as C edi Sp ead. Compa ed o U.S. RMBS, we obse e a e De aul s:
Only 1.8% o AAA- a ed anches de aul ed, whe eas Ghen e al. (2019) epo a 42%
de aul a e o AAA- a ed anches. Fo ini ially BBB- a ed anches, he de aul a e
was ela i ely high a 8.1%, conside ing ha his is in es men -g ade, al hough i is s ill
conside ably lowe han o BBB- a ed U.S. RMBS (97%). A look a he IRR con i ms
his pic u e o ela i ely mode a e Eu opean RMBS losses. The mean IRR inc eases
along wi h wo se ini ial a ings. This indica es ha isk did no ma e ialize o he ex en
ha i was p iced, which is in s a k con as o U.S. RMBS, as hose secu i ies had
sh inking a e age IRRs wi h wo se ini ial a ings.
4.4 Me hodology
To es ou hypo heses ega ding he complexi y channel (H1a, H2a), we eg ess
ou pe o mance measu es De aul and IRR on common complexi y measu es, as
well as a ious con ols and ixed e ec s:
We es ima e he p obabili y ha anche i, which was issued a ime , de aul s
un il he poin in ime T, which is he ime o De aul measu emen (Feb ua y 2021).
The e o e, T– gi es he du a ion o obse a ion in which a anche de aul would be
eco ded by us. To model he p obabili y o De aul , we use he linea p obabili y
model.12 We use model (3) o es ima e he con inuous a iable IRR by OLS eg es-
sions. We clus e s anda d e o s a he deal le el.
(2)
P
(De aul i, ,T=1|xi,j, )=𝛼+Complexi y
�
j,
𝛽+Con ols
�
i,j,
𝛾+𝜓 +𝜓c+𝜓 +𝜓
o
(3)
E
(IRRi, ,T|xi,j, )=𝛼+Complexi y
�
j,
𝛽+Con ols
�
i,j,
𝛾+𝜓 +𝜓c+𝜓 +𝜓
o
12 We choose o model he p obabili y o De aul ia he linea p obabili y model because o ou ex en-
si e use o ixed e ec s. Since ou De aul a iable has a mean o app oxima ely 3%, i.e., 97% o he
anches in ou sample did no de aul , i can be conside ed a mild case o “ a e e en s da a” (King and
Zeng, 2001). Timoneda (2021) in es iga es he s a is ical p ope ies o a ious bina y dependen a iable
models in he case o a e e en s da a using Mon e Ca lo simula ions. They conclude ha he linea p ob-
No on hesame page: comp ehensibili y o MBS in es men … Page 17 o 37 9
On he igh -hand side o ou eg ession equa ions, we include he common com-
plexi y measu es desc ibed in Sec .4.2, which a y a deal le el j. Fu he mo e, we
include Deal Volume and Excess Sp ead as con ol a iables ha a y a he deal
le el j. We also include con ols ha a y a he anche le el i: In addi ion o he
Subo dina ion o he anche, we also con ol o a dummy a iable ha equals one
i a leas wo c edi a ing agencies issued di e ging a ings a he ime o anche
Fig. 2 Tex eadabili y examples o di e en p ospec uses. This igu e p esen s exce p s om he sec ion
“Cash Collec ion A angemen s” o wo di e en in es men p ospec uses wi h Fog Index 25.4 (Panel
A) e sus 18.2 (Panel B)
abili y model ou pe o ms o he models such as logi in his ega d. This inding is in ui i e conside ing
ha he logis ic model d ops obse a ions i he e is no a ia ion in he bina y dependen a iable wi hin
one o he ixed e ec g oupings. Ne e heless, as a obus ness check, we es ima e he de aul p obabili y
wi h a p obi eg ession and ob ain simila esul s.
Foo no e 12 (Con inued)
M.Hibbeln e al.
9 Page 18 o 37
issuance (Disag ee Ra ing).13 We u he con ol o he C edi Sp ead o e he
3-mon h EURIBOR measu ed in basis poin s. All igh -hand side a iables a e
measu ed a he ime o deal issuance . Finally, we include a ious ixed e ec s,
pa icula ly o he yea o issuance ψ , coun y o colla e al ψc, a ing a issuance
ψ ,14 and o igina o ψo.15 We p o ide de ailed a iable de ini ions in Table8.
To es ou complexi y p icing hypo heses (H1b, H2b), we apply a simila eg es-
sion model, excep ha C edi Sp ead is no longe among he con ol a iables:
To es ou hypo heses ega ding he p ospec us comp ehensibili y channel (H3a,
H3b, H4a, and H4b), we es ima e simila models as Eqs.(2)–(4):
In hese models, we change he measu emen o complexi y o include bo h
dimensions o p ospec us comp ehensibili y, To al Pages and Fog Index, sepa a ely,
as well as combined in an in e ac ion e m. To al Pages is an ex ensi e measu e o
p ospec us comp ehensibili y, as he p ospec us leng h inc eases he ime i akes
an in es o o ully unde s and he de ails o he deal. Fog Index, as a ex ead-
abili y p oxy, can be in e p e ed as an in ensi e measu e o p ospec us comp ehen-
sibili y, which aises he ime he in es o s need o unde s and one gi en page o
he p ospec us. We include he in e ac ion e m To al Pages
×
Fog Index o quan-
i y he e o equi ed by an in es o o unde s and he p ospec us. While any gi en
page o a p ospec us wi h a high Fog Index migh be di icul o unde s and, i will
no equi e g ea e o o unde s and i i he p ospec us is sho . Con e sely, e en
long p ospec uses can be ela i ely easy o unde s and i he language used is clea
(4)
E
(C edi Sp eadi, |xi,j, )=𝛼+Complexi y
�
j
,
𝛽+Con ols
�
i
,
j
,
𝛾+𝜓 +𝜓c+𝜓 +𝜓
o
(5)
P (De aul
i, ,T
=1|x
i,j,
)=𝛼+To alPages
j,
𝛽
1
+Fog
j,
𝛽
2
+
To alPages ×Fogj, 𝛽3+Con ols�
i
,
j
,
𝛾+𝜓 +𝜓c+𝜓 +𝜓
o
(6)
E(IRR
i, ,T
|x
i,j,
)=𝛼+To alPages
j,
𝛽
1
+Fog
j,
𝛽
2
+
To alPages ×Fogj, 𝛽3+Con ols�
i
,
j
,
𝛾+𝜓 +𝜓c+𝜓 +𝜓
o
(7)
E(C edi Sp ead
i,
|x
i,j,
)=𝛼+To alPages
j,
𝛽
1
+Fog
j,
𝛽
2
+
To alPages ×Fogj, 𝛽3+Con ols�
i,j,
𝛾+𝜓 +𝜓c+𝜓 +𝜓
o
14 Ra ing a issuance is de ined as he bes a ing ha one o he h ee a ing agencies issued. Disag ee
Ra ing, he e o e, indica es a wo se a ing being issued by ano he a ing agency.
15 We ex end he con ol a iables o Ghen e al. (2019) o include o igina o ixed e ec s. We a ion-
alize his choice in Appendix D o ou Online Appendix. When eg essing ou complexi y a iables
solely on o igina o ixed e ec s, we obse e R2s o up o 80%, indica ing ha complexi y mos ly a ies
be ween o igina o s.
13 Disag ee Ra ing co esponds o Ghen e al. (2019)’s a iable disag ee anche. As we lack he dis inc-
ion o di e en loan g oups, we canno cons uc Ghen e al. (2019)’s con ol a iable c osscolla e ali-
za ion. We u he do no con ol o he o al yea ly issuance olume o he lead manage (lead o ) bu
ins ead con ol o o igina o ixed e ec s.
No on hesame page: comp ehensibili y o MBS in es men … Page 19 o 37 9
and concise. Howe e , i bo h ex eadabili y is low and p ospec us leng h is high,
unde s anding he p ospec us would be a majo obs acle e en o sophis ica ed
in es o s.
5 Main esul s
5.1 Pe o mance andp icing o common complexi y measu es
Fo all subsequen analyses, we spli he sample in o a p e-c isis and a pos -c isis
sample ( anches issued be ween 2003 and 2007 and hose issued be ween 2008 and
2020). In Table3, we epo he eg essions o ou ex ensi e and in ensi e pe o -
mance measu es De aul and IRR on he common complexi y a iables and analyze
he p icing o common complexi y h ough C edi Sp ead.
Fo MBS issued be o e he inancial c isis, we hypo hesize ha o igina o s ob us-
ca ed poo secu i iza ion quali y by inc easing complexi y (H1a). I his we e he
case, we would expec mo e complex MBS o pe o m wo se, e en a e con ol-
ling o po en ial con ounding ac o s ha a ec bo h secu i iza ion pe o mance and
complexi y. We es his hypo hesis in Table3, Panel A, columns 1–6. When we
measu e pe o mance by using ou ex ensi e measu e o De aul (columns 1–3), we
ind almos no e idence in a o o his complexi y channel: None o ou common
complexi y measu es a e signi ican ly ela ed o highe anche De aul s p e-c isis,
excep o a sligh posi i e coe icien o To al Pages (column 1). When we ins ead
measu e pe o mance by using he in ensi e measu e o IRR (columns 4–6), we s ill
do no obse e a signi ican ela ionship be ween mos o ou common complexi y
a iables and pe o mance. The only common complexi y a iable associa ed wi h
lowe e u ns is To al Pages, which we will in es iga e in de ail in he nex sec ion.
Taken oge he , ou esul s o De aul and IRR do no suppo he exis ence o he
complexi y channel in he p e-c isis pe iod (H1a) when p oxying complexi y wi h
common measu es.
Disag ee Ra ing is associa ed wi h signi ican ly wo se pe o mance p e-c i-
sis (columns 1–6), as Ra ing a Issuance ixed e ec s only include he bes a ing
assigned. I a di e en a ing agency issued a wo se a ing, he De aul p obabili y
inc eases by mo e han 6 pe cen age poin s, indica ing he use ulness o such in o -
ma ion o de aul p edic ion. Beyond he c edi a ing, in es o s did no inco po-
a e use ul in o ma ion in hei p icing decision: Highe C edi Sp ead is no signi i-
can ly associa ed wi h an inc eased De aul p obabili y (columns 1–3). Fu he mo e,
an addi ional basis poin o C edi Sp ead is associa ed wi h 0.7 basis poin s highe
IRR. The e o e, he isk p iced h ough he sp ead has no ully ma e ialized, as
would be indica ed by a coe icien o 0.
Nex , we check i in es o s an icipa ed he o igina o s using complexi y o
ob usca ion and hus equi ed a isk p emium o complexi y in he p e-c isis pe iod
(H1b). We canno con i m his hypo hesis based on columns 7–9, as he e is almos
no signi ican ela ion be ween C edi Sp ead and any o ou common complex-
i y a iables. Fo Glossa y Te ms, we e en ind a weak nega i e ela ion. F om an
M.Hibbeln e al.
9 Page 20 o 37
ex-pos pe spec i e, he lack o common complexi y p icing was he co ec deci-
sion, as mo e complex MBS did no pe o m wo se p e-c isis.
Subsequen ly, we ocus on anches issued du ing 2008–2020 (“pos -c isis”). We
es i o igina o s use he complexi y channel pos -c isis, despi e pe haps ea ing
being punished by in es o s h ough inc eased isk p emia (H2a). When es ima ing
ou eg essions o De aul and IRR on ou common complexi y a iables o he
pos -c isis sample, we ind li le e idence o wo se pe o mance o complex MBS.
Though he coe icien s o Pagesmpool a e s a is ically signi ican o bo h De aul
(column 10) and IRR (column 13), he economic signi icance is low, wi h an addi-
ional page o Pagesmpool educing annual e u ns by jus 1.3 bps. The only no able
excep ion is To al Pages, which emains signi ican ly nega i ely ela ed o IRR (col-
umn 14). Excep o To al Pages, we canno con i m H2a (columns 10–15).
Conce ning he pos -c isis in es o p icing o complexi y, in es o s may ha e
become inc easingly complexi y-a e se, hus s a ing o equi e a isk p emium o
mo e complex MBS (H2b). E en hough we ind li le e idence ha o igina o s used
complexi y o quali y ob usca ion in Eu ope, in es o s may s ill apply he lesson
hey lea ned on he U.S. ma ke , whe e he complexi y channel did exis p e-c isis (.,
2019). Mo eo e , egula o s communica ed ha he high complexi y o secu i iza-
ions and hei poo pe o mance du ing a c isis a e ela ed, which may ha e lead
in es o s becoming mo e complexi y-a e se (EBA, 2014). We ind ha in line wi h
H2b, common complexi y is gene ally p iced pos -c isis: A one s anda d de ia ion
inc ease in PC1 esul s in a 23bp inc ease in C edi Sp ead (p < 1%, see column
18). Complexi y ela ed o he numbe o Glossa y Te ms is p iced o a la ge ex en ,
wi h 100 addi ional e ms inc easing he sp ead by app oxima ely 15 bps (column
16). To al Pages is s a is ically signi ican in model (17), wi h a 100-page inc ease in
p ospec us leng h associa ed wi h a 24 bps inc ease in C edi Sp ead. O e all, his
p o ides e idence in a o o in es o s s a ing o p ice common complexi y a e
he inancial c isis (H2b). Ex-pos , his p icing decision is no jus i ied by pe o -
mance, conside ing ha nei he p e- no pos -c isis, he common complexi y was
ela ed o signi ican ly wo se MBS pe o mance. S ill, om an ex-an e equilib ium
pe spec i e, in es o s canno dis inguish i complexi y is jus inciden al o a s a e-
gic ob usca ion beha io o he o igina o . In pa icula , only because he o igina o s
did no use common complexi y measu es o ob usca ion in he pas , his does no
mean ha hey will no s a using i a some poin . The e o e, his p icing beha io
o common complexi y by in es o s can be a ionalized.
Con a y o he p e-c isis esul s, he coe icien o Disag ee Ra ing is highly sig-
ni ican and posi i e pos -c isis (columns 16–18). In es o s now ake de ia ing a -
ings in o conside a ion and demand a isk p emium o app oxima ely 43 bps, which
is a ional conside ing ha anches his o ically pe o med wo se i a de ia ing a -
ing agency assigned a wo se a ing a anche issuance.
In summa y, Eu opean o igina o s did no signi ican ly inc ease common com-
plexi y measu es o ob usca e poo RMBS quali y, ei he be o e o a e he inan-
cial c isis, which is di e en om indings o p e-c isis issues in he U.S. The only
excep ion is he a iable To al Pages, whe e we ind some indica ions o o igina o s
using p ospec us leng h o ob usca e low secu i iza ion quali y. While in es o s did
no equi e a isk p emium o common complexi y be o e he c isis, hey s a ed
No on hesame page: comp ehensibili y o MBS in es men … Page 21 o 37 9
Table 3 De aul , IRR, and C edi Sp ead eg essions on common complexi y measu es
Panel A: P e-c isis issuances
De aul IRR C edi Sp ead
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Deal T anches − 0.001
(− 0.238)
0.015
(0.373)
4.116+
(1.845)
Pagesmpool 0.003
(1.299)
− 0.006
(− 0.721)
0.616
(1.410)
Pageswa e all − 0.000
(− 0.097)
− 0.006
(− 1.046)
− 0.083
(− 0.172)
Glossa yTe ms − 0.030
(− 1.319)
− 0.091
(− 1.239)
− 10.230*
(− 2.190)
File Size − 0.016
(− 1.616)
0.057
(1.489)
− 2.730
(− 1.256)
To al Pages 0.099+
(1.798)
0.058
(1.563)
− 0.302*
(− 2.011)
− 0.379**
(− 2.715)
− 0.567
(− 0.046)
− 14.711
(− 1.424)
PC1 0.003
(0.130)
− 0.063
(− 0.550)
− 4.111
(− 0.734)
Disag eeRa ing 0.067**
(3.214)
0.064**
(3.071)
0.064**
(3.063)
− 0.264**
(− 2.722)
− 0.254**
(− 2.765)
− 0.250**
(− 2.693)
5.796
(0.846)
4.372
(0.644)
4.611
(0.679)
C edi Sp ead − 0.008
(− 0.253)
− 0.006
(− 0.178)
− 0.007
(− 0.218)
0.007***
(3.894)
0.007***
(3.923)
0.007***
(3.929)
Obse a ions 607 607 607 607 607 607 607 607 607
Adjus ed R20.226 0.228 0.226 0.347 0.350 0.346 0.450 0.448 0.447
Con ols Yes Yes Yes Yes Yes Yes Yes Yes Yes
Ra ing FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
Yea Issue FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
Coun y FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
O igina o FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
M.Hibbeln e al.
9 Page 22 o 37
Table 3 (con inued)
Panel B: Pos -
c isis issuances
De aul IRR C edi Sp ead
(10) (11) (12) (13) (14) (15) (16) (17) (18)
Deal T anches 0.008+
(1.952)
− 0.001
(− 0.039)
2.696
(1.183)
Pagesmpool 0.002*
(2.551)
− 0.013*
(− 2.079)
0.117
(0.363)
Pageswa e all 0.004
(1.633)
0.003
(0.400)
− 1.362
(− 1.553)
Glossa yTe ms − 0.021*
(− 2.005)
0.019
(0.363)
15.389**
(3.106)
File Size − 0.005
(− 0.712)
0.112
(1.387)
− 1.668
(− 0.379)
To al Pages − 0.021
(− 1.153)
− 0.010
(− 0.680)
− 0.297+
(− 1.906)
− 0.266*
(− 2.087)
10.741
(0.893)
24.135*
(2.075)
PC1 0.003
(0.259)
− 0.175+
(− 1.919)
22.771**
(2.862)
Disag eeRa ing 0.010
(0.725)
0.013
(0.923)
0.014
(0.948)
− 0.109
(− 0.922)
− 0.135
(− 1.199)
− 0.127
(− 1.111)
42.359***
(4.585)
43.829***
(4.674)
42.983***
(4.605)
C edi Sp ead − 0.005
(− 0.907)
− 0.006
(− 1.030)
− 0.007
(− 1.088)
0.011***
(16.479)
0.011***
(16.324)
0.011***
(16.439)
Obse a ions 1214 1214 1214 1214 1214 1214 1214 1214 1214
Adjus ed R20.219 0.197 0.197 0.565 0.565 0.565 0.662 0.657 0.659
Con ols Yes Yes Yes Yes Yes Yes Yes Yes Yes
Ra ing FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
Yea Issue FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
No on hesame page: comp ehensibili y o MBS in es men … Page 23 o 37 9
Table 3 (con inued)
Panel B: Pos -
c isis issuances
De aul IRR C edi Sp ead
(10) (11) (12) (13) (14) (15) (16) (17) (18)
Coun y FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
O igina o FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
This able shows OLS es ima es om eg essing De aul , IRR, and C edi Sp ead on common complexi y a iables and con ols. IRR is measu ed in pe cen age; C edi
Sp ead is measu ed in basis poin s. Uni o obse a ion is anches. Panel A epo s he esul s o anches issued in 2003–2007 (“p e-c isis”); Panel B epo s he esul s
o anches issued in 2008–2020 (“pos -c isis”). Va iable de ini ions a e gi en in Table8. S anda d e o s a e clus e ed a he deal le el. s a is ics a e gi en in pa en he-
ses. ***, **, *, and + deno e s a is ical signi icance a he 0.1%, 1%, 5%, and 10% le els, espec i ely
M.Hibbeln e al.
9 Page 24 o 37
p icing i a e wa d. This may be due o in es o s becoming complexi y-a e se due
o lea ning e ec s. This may be as hey (1) obse ed ex ensi e de aul s o complex
U.S. secu i ies du ing he inancial c isis, and (2) go communica ed by egula o s
ha he high complexi y o secu i iza ions and hei poo pe o mance du ing a c i-
sis a e ela ed. This p icing beha io , hough ex-pos inconsis en wi h ealized isk,
can be a ionalized om an ex-an e equilib ium pe spec i e, as in es o s o a gi en
secu i iza ion canno obse e i i s high complexi y is s a egically used o ob usca-
ion o simply inciden al.
5.2 Pe o mance andp icing o p ospec us comp ehensibili y
In his sec ion, we analyze whe he o igina o s used p ospec us comp ehensibili y,
which is a combina ion o p ospec us leng h and ex eadabili y pe page, o ob us-
ca e poo secu i iza ion quali y and whe he low comp ehensibili y was p iced by
in es o s. As in he p e ious sec ion, we spli he sample in o a p e-c isis sample
(2003–2007) and a pos -c isis sample (2008–2020). In Table4, we epo he eg es-
sions o ou ex ensi e and in ensi e pe o mance measu es De aul and IRR on he
p ospec us comp ehensibili y a iables and analyze he p icing o p ospec us com-
p ehensibili y h ough C edi Sp ead. Fo hese analyses, we demean Fog Index and
To al Pages o acili a e in e p e a ion.
Fo MBS issued be o e he inancial c isis, we hypo hesize ha o igina o s
lowe ed ex eadabili y in combina ion wi h inc easing p ospec us leng h o
ob usca e poo secu i iza ion quali y (H3a). I his was he case, we would expec
less comp ehensible MBS o pe o m wo se. We es his hypo hesis in Table4,
Panel A, columns 1 and 2. The coe icien s a e signi ican ly posi i e a he 5%
le el (column 3). Due o he demeaning, he coe icien o Fog Index gi es he
ma ginal e ec ha Fog Index has on De aul o a p ospec us wi h an a e age
numbe o pages: an inc ease o one uni in Fog Index co esponds o an inc ease
o 5.2 pe cen age poin s in De aul p obabili y o a p ospec us wi h an a e age
o 223 pages. Fo a p ospec us whe e To al Pages is one s anda d de ia ion abo e
he a e age (223 + 82 = 305 pages), his e ec inc eases o 5.2 + 5.6 × 0.82 = 9.8
pe cen age poin s, which is o high economic signi icance. Con e sely, he e ec
o To al Pages on De aul is also s a is ically and economically signi ican :
A 100-page inc ease in To al Pages is associa ed wi h a 7.7 pe cen age poin
inc ease in De aul p obabili y o a p ospec us wi h an a e age Fog Index. These
esul s p o ide he i s e idence in a o o he exis ence o he p e-c isis p o-
spec us comp ehensibili y channel (H3a). This inding is u he suppo ed by he
eg ession o IRR on p ospec us comp ehensibili y: In column 2, o a one s and-
a d de ia ion inc ease in Fog Index, IRR is 0.135 × 1.3 = 0.18 pe cen age poin s
lowe when conside ing a p ospec us wi h a e age To al Pages. Fo a p ospec us
whe e he numbe o pages is one s anda d de ia ion abo e he a e age, his e ec
inc eases o –0.135 × 1.3 – 0.153 × 0.82 × 1.3 = –0.34 pe cen age poin s. The ma -
ginal e ec o To al Pages on IRR is e en s onge : Fo he a e age Fog Index
o 22, a 100-page inc ease in To al Pages is associa ed wi h a 0.44 pe cen age
poin dec ease in annual e u ns. In o al, his p o ides s ong e idence in a o o
No on hesame page: comp ehensibili y o MBS in es men … Page 31 o 37 9
canno obse e hei pe o mance o e hei en i e li espan. De aul s o la ge p inci-
pal losses ha lowe he IRR migh only occu in la e pe iods o he anche li espan.
To accoun o his measu emen e o , we cons uc he a iable IRR 5Y, which is
de ined as he in e nal a e o e u n du ing a 5-yea holding pe iod a e anche issu-
ance. T anches o which we canno obse e he ull i e-yea pe iod o cash lows
( hose issued a e 1s o May 2016) a e d opped om ou sample. We epo he
esul s in Tables E.2.1 and E.2.2. Panel A o Table E.2.1 only p o ides e idence in
a o o he p e-c isis complexi y channel o a iables ela ed o p ospec us leng h
(Glossa y Te ms and To al Pages). In con as , and con i ming ou p e ious esul s,
we ind ha low p ospec us comp ehensibili y is nega i ely ela ed o anche IRR 5Y
(Panel A o Table E.2.2). The coe icien magni udes emain almos iden ical o hose
epo ed in he baseline IRR eg ession (Table4, column 2).
Thi d, we check i ou esul s a e obus o an al e na i e eadabili y measu e.
Ini ially, we use he Fog Index, as i is an es ablished p oxy o measu e he ead-
abili y o inancial documen s (Bushee e al., 2018; Dye e al., 2017; Li, 2008; Lo
e al., 2017). As an al e na i e, we compu e he Smog Index, which is compu ed
using only he ac ion o complex wo ds (wo ds wi h h ee syllables o mo e scaled
by he numbe o sen ences) as an inpu . Ou esul s emain comple ely obus o
his change. Pa icula ly ega ding he p e-c isis p ospec us comp ehensibili y chan-
nel, we now obse e ha an inc ease o one uni in Smog Index co esponds o an
inc ease o 8.5 pe cen age poin s in De aul p obabili y o a p ospec us wi h an
a e age o 223 pages (column 3 o Table E.3).
Fou h, we include yea -qua e ixed e ec s ins ead o yea ixed e ec s. We con-
i m he non-exis ence o he common complexi y channel, and we s ill ind s ong
e idence in a o o he p e-c isis p ospec us comp ehensibili y channel (Table
E.4.1 and Table E.4.2).
Fi h, we e un all ou De aul eg essions using he p obi model ins ead o he
linea p obabili y model, con i ming ou p e-c isis esul s.16
Six h, we epea ou eg essions ega ding ou common complexi y measu es
(Sec . 5.1) wi hou including o igina o ixed e ec s o allow o a di ec com-
pa ison wi h Ghen e al.’s (2019) esul s ega ding p e-c isis issues. Ou esul s
la gely emain simila o ou main esul s. The e is s ill li le e idence in a o
o he common complexi y channel, bo h p e-c isis and pos -c isis (see Table F.1
and Table F.2). The excep ions a e he a iables To al Pages and Glossa y Te ms,
which a e associa ed wi h highe De aul and lowe IRR in he p e-c isis pe iod
(columns 5 and 7 in bo h Tables F.1 and F.2). Simila ly, we obse e no p icing o
complexi y o he p e-c isis pe iod (see Table F.3, Panel A). Fo pos -c isis p ic-
ing, we con i m ou p e ious esul s, al hough he coe icien o PC1 dec eases o
15, indica ing ha a one s anda d de ia ion inc ease in complexi y is associa ed
wi h a 15 bps inc ease in C edi Sp ead when aking a ia ion be ween o igina o s
in o accoun (see Table F.3, column 16).
16 As ou De aul a iable has e y low a ia ion pos -c isis (only 6 o 1257 anches de aul ed), we
canno un p obi eg essions o he pos -c isis pe iod. The p e-c isis p obi esul s a e a ailable upon
eques .
M.Hibbeln e al.
9 Page 32 o 37
7 Conclusion
We in es iga e whe he o igina o s o Eu opean RMBS lowe he comp ehensi-
bili y o disclosu e documen s o ob usca e low secu i iza ion quali y. Relying
on commonly used, mainly olume-based complexi y p oxies, we ind no e i-
dence ha o igina o s inc ease complexi y o ob usca e low secu i iza ion qual-
i y. Howe e , we a gue ha he combina ion o ex eadabili y (which meas-
u es he di icul y o eading a gi en page) and p ospec us leng h is decisi e
o measu ing he le el o di icul y in eading and unde s anding disclosu e
documen s. Implemen ing his measu e o comp ehensibili y, we ind e idence
o o igina o s ob usca ing low secu i iza ion quali y. The e ec s a e economi-
cally meaning ul: Fo a one s anda d de ia ion dec ease in eadabili y, he
annual e u ns a e on a e age 18 basis poin s lowe when conside ing a p o-
spec us wi h a e age p ospec us leng h. Fo a p ospec us whe e he p ospec us
leng h is one s anda d de ia ion abo e he a e age, his e ec inc eases o 34
basis poin s. Al hough in es o s did no an icipa e o igina o s using complexi y
o ob usca ion p io o he inancial c isis, hey changed hei p icing beha -
io he ea e , demanding a isk p emium o up o 23 basis poin s o a one
s anda d de ia ion inc ease in common complexi y. The “p ospec us comp e-
hensibili y channel”, howe e , is only pa ially p iced: While in es o s demand
a signi ican isk p emium o p ospec us leng h, hey do no demand any isk
p emium o low ex eadabili y. The ecen egula o y in e en ion in he o m
o he EUSR has no signi ican e ec on enhancing p ospec us comp ehensibil-
i y, e en hough educing deal complexi y was one o he egula ion’s decla ed
goals, sugges ing ha p ospec us comp ehensibili y is a dis inc o m o secu-
i iza ion complexi y.
Ou esul s ha e impo an implica ions o o igina o s, in es o s, and egula o s.
O igina o s should y o a oid designing long p ospec uses, as p ospec us leng h
and he numbe o e ms in he glossa y a e associa ed wi h a s ong inc ease in
C edi Sp ead. In addi ion, in es o s should conside no only he leng h o he p o-
spec us bu also i s ex eadabili y because o igina o s migh be ying o ob usca e
low-quali y secu i iza ions h ough bo h dimensions o comp ehensibili y. Fo in es-
o p o ec ion, egula o s should also conside explici ly enhancing p ospec us com-
p ehensibili y. On a e age, he p ospec us leng h mo e han doubled be ween 2003
and 2019 o app oxima ely 300 pages wi hou imp o ing he measu ed eadabili y
pe page, which highligh s ha he low comp ehensibili y o MBS in es men p o-
spec uses is s ill a majo obs acle.
Appendix
Table7p esen s he dis inc ion be ween ou “common complexi y” and “p ospec-
us comp ehensibili y” measu es. The common complexi y a iables a e mos ly
olume-based measu es de i ed om he in es men p ospec us. Fo p ospec us
No on hesame page: comp ehensibili y o MBS in es men … Page 33 o 37 9
comp ehensibili y, we ocus on he join e ec o ex eadabili y and p ospec us
leng h. P ecise a iable de ini ions a e gi en in Table8
Table 7 Types o complexi y
Type o complexi y Desc ip ion
Panel A: Common complexi y measu es
P ospec us olume P oxies de i ed om he p ospec us, ha a emp o measu e
he complexi y in he secu i ized loan pool and he secu-
i iza ion s uc u e. These measu es include he numbe o
pages ( o al and o speci ic sec ions), he numbe o glossa y
e ms, and he ile size o he p ospec us
Numbe o anches Numbe o anches in a deal as a p oxy o complexi y in he
secu i iza ion s uc u e
Panel B: Tex eadabili y and p ospec us comp ehensibili y
Tex eadabili y Tex eadabili y o an a e age page in he in es men p ospec-
us o a gi en deal, measu ed by Fog index (o Smog index).
A highe Fog index indica es a lowe ex eadabili y
Tex eadabili y × P ospec us leng h The in e ac ion o ex eadabili y and p ospec us leng h
(measu ed by he numbe o pages o he in es men
p ospec us) can be a ac o ha in luences p ospec us com-
p ehensibili y beyond he indi idual measu es. Fo example,
i he in es men p ospec us is e y sho , ins i u ional
in es o s should ha e ela i ely li le ouble in es ing he
associa ed sea ch cos s o unde s and he con en s, e en i
he p ospec us ex eadabili y is low. Howe e , o long p o-
spec uses, low ex eadabili y may be a majo obs acle o
ully unde s anding he secu i iza ion’s legal e ms. A highe
alue o he in e ac ion e m Fog#To alPages indica es a
lowe comp ehensibili y
M.Hibbeln e al.
9 Page 34 o 37
Table 8 Va iable de ini ions
Va iable De ini ion
Panel A: Pe o mance ( anche o deal le el)
De aul Bina y a iable ha is equal o one i he cu en a ing (measu ed in
Feb ua y 2021) signals a anche de aul o he e a e any p incipal
losses g ea e han ze o eco ded in Bloombe g. An S&P a ing o
CCC + o lowe , a Moody’s a ing o Caa1 o lowe , o a Fi ch a ing
o CCC o lowe is de ined as an indica o o de aul
IRR In e nal a e o e u n based on he cash lows owa d he anche un il
Ap il 2021. The anche is assumed o ha e been bough a pa and
any emaining p incipal le ou s anding was paid back in ull in
June 2021. IRR is measu ed in pe cen age
IRR 5Y In e nal a e o e u n based on he cash lows ha he anche ecei ed
un il 5yea s a e anche issuance. The anche is assumed o ha e
been bough a pa and any emaining p incipal le ou s anding was
paid back in ull 5yea s a e anche issuance. IRR 5Y is measu ed
in pe cen age
Colla e al Loss Sha e Deal-le el p incipal losses (as obse ed in Feb ua y 2021) di ided by
Deal Volume a ime o deal issuance, measu ed in pe cen age
Panel B: P icing ( anche le el)
C edi Sp ead C edi sp ead o e he 3-mon h EURIBOR measu ed a anche issu-
ance in basis poin s
Panel C: Complexi y (deal le el)
Deal T anches The o al numbe o anches wi hin a deal
Pagesmpool The numbe o pages in he p ospec us desc ibing he pool o he
unde lying mo gage loans
Pageswa e all The numbe o pages in he p ospec us desc ibing he cash low allo-
ca ion o he anches (wa e all mechanism)
Glossa y Te ms The o al numbe o e ms in he glossa y o he p ospec us
File Size The ile size o he p ospec us measu ed in megaby es
To al Pages The o al numbe o pages o he p ospec us
PC1 Fi s p incipal componen o he a iables: Deal T anches, Pagesm-
pool, Pageswa e all, Glossa y Te ms, File Size, To al Pages. Fac o
loadings a e as ollows: To al Pages (0.55), Glossa y Te ms (0.51),
Pagesmpool (0.41), File Size (0.38), Deal T anches (0.32) and
Pageswa e all (0.14). PC1 is s anda dized o a s anda d de ia ion
o one
Fog Index Linea combina ion o he a e age numbe o wo ds pe sen ence and
he ac ion o complex wo ds ( ac ion o wo ds wi h h ee syllables
o mo e):
Fog Index = (a e age sen ence leng h + ac ion o complex
wo ds) × 0.4
Smog Index Smog Index is calcula ed by:
Smog Index
=1.043 ×
√
complex wo ds ×30
numbe o sen ences
+
3.1291 ,
whe e a “complex wo d” is de ined as a wo d wi h h ee syllables o
mo e
STS Fea u es The o al numbe o simplici y, anspa ency, and s anda diza ion
(STS) ea u es ha a speci ic deal adhe es o. Takes a maximum
alue o 48 i a deal adhe es o all 48 STS ea u es
No on hesame page: comp ehensibili y o MBS in es men … Page 35 o 37 9
Supplemen a y In o ma ion The online e sion con ains supplemen a y ma e ial a ailable a h ps:// doi.
o g/ 10. 1007/ s11147- 025- 09213-8.
Funding Open Access unding enabled and o ganized by P ojek DEAL. This s udy was unded by
Deu sche Fo schungsgemeinscha (G an no.: 434130478) by Ma in Thomas Hibbeln.
Open Access This a icle is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License,
which pe mi s use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any medium o o ma , as long
as you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e
Commons licence, and indica e i changes we e made. The images o o he hi d pa y ma e ial in his
a icle a e included in he a icle’s C ea i e Commons licence, unless indica ed o he wise in a c edi line
o he ma e ial. I ma e ial is no included in he a icle’s C ea i e Commons licence and you in ended
use is no pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will need o ob ain pe mis-
sion di ec ly om he copy igh holde . To iew a copy o his licence, isi h p://c ea i ecommons.o g/
licenses/by/4.0/.
Re e ences
Albe azzi, U., E amo, G., Gambaco a, L., & Salleo, C. (2015). Asymme ic in o ma ion in secu i iza-
ion: An empi ical assessmen . Jou nal o Mone a y Economics, 71, 33–49.
Al unbas, Y., Gambaco a, L., & Ma ques-Ibanez, D. (2009). Secu i iza ion and he bank lending chan-
nel. Eu opean Economic Re iew, 53, 996–1009.
As iyan, V., Foa a, D., & Vanasco, V. (2022). The good, he bad and he complex: P oduc design wi h
impe ec in o ma ion. Ame ican Economic Jou nal: Mic oeconomics, 15, 187–226.
BCBS, and IOSCO. (2015). C i e ia o Iden i ying Simple, T anspa en and Compa able Secu i isa ions.
Basel Commi ee on Banking Supe ision and In e na ional O ganiza ion o Secu i ies Commis-
sions. A ailable online: h ps:// www. bis. o g/ bcbs/ publ/ d441. pd .
Table 8 (con inued)
Va iable De ini ion
Simplici y/T anspa ency
Fea u es
The o al numbe o simplici y ( anspa ency) ea u es ha a speci ic
deal adhe es o. Takes a maximum alue o 17 (5) i a deal adhe es
o all 17 (5) simplici y ( anspa ency) ea u es
Panel D: Con ols ( anche le el)
Subo dina ion The ac ion o he deal olume ha can de aul be o e he anche
su e s i s losses
Disag ee Ra ing Bina y a iable ha is equal o one i a leas wo c edi a ing agen-
cies issued di e ging a ings a he ime o anche issuance
Ra ing The bes a ing ha one o he h ee a ing agencies (S&P, Fi ch, and
Moody’s) assigned a he ime o anche issuance
Panel E: Con ols (deal le el)
Deal olume Volume o he deal measu ed a he ime o deal issuance in billions
o Eu os
Excess sp ead The excess coupon ha he loans unde lying he deal pay minus he
alue-weigh ed i s a e age coupon ha he secu i ies wi hin he
deal pay
Yea o issuance Yea in which he deal was issued
Coun y o colla e al Coun y in which he unde lying mo gage loans we e o igina ed
O igina o O igina o o he unde lying mo gage loans
M.Hibbeln e al.
9 Page 36 o 37
Be nd , A., & Gup a, A. (2009). Mo al haza d and ad e se selec ion in he o igina e- o-dis ibu e model
o bank c edi . Jou nal o Mone a y Economics, 56, 725–743.
Billio, M., Du ou , A., Sega o, S., & Va o o, S. (2023). Complexi y and he de aul isk o mo gage-
backed secu i ies. Jou nal o Banking and Finance, 155, 1–19.
Bonsall, S. B., & Mille , B. P. (2017). The impac o na a i e disclosu e eadabili y on bond a ings and
he cos o deb . Re iew o Accoun ing S udies, 22, 608–643.
Bozanic, Z., K a , P., & Tille , A. (2023). Quali a i e disclosu e and c edi analys s’ so a ing adjus -
men s. Eu opean Accoun ing Re iew, 32, 779–807.
Bushee, B. J., Gow, I. D., & Taylo , D. J. (2018). Linguis ic complexi y in i m disclosu es: Ob usca ion
o in o ma ion? Jou nal o Accoun ing Resea ch, 56, 85–121.
Cheng, J. (2010). The ine p in is enla ged, bu will in es o s ead i ? The Wall S ee Jou nal, May 31.
A ailable online: h ps:// www. wsj. com/ a ic les/ SB100 01424 05274 87039 57604 57527 25102 63503
010.
Co al, J., Ju ek, J., & S a o d, E. (2009). The economics o s uc u ed inance. Jou nal o Economic
Pe spec i es, 23, 3–25.
Debene , J., Fenne , A., Klein, P., Ongena, S. (2021). Linguis ic complexi y in ABS p ospec uses—E i-
dence om eu opean secu i iza ion da a. Wo king Pape .
Dye , T., Lang, M., & S ice-Law ence, L. (2017). The e olu ion o 10-K ex ual disclosu e: E idence
om la en di ichle alloca ion. Jou nal o Accoun ing and Economics, 64, 221–245.
EBA. (2014). EBA Discussion Pape on Simple, S anda d and T anspa en Secu i isa ions. Eu opean
Banking Au ho i y. A ailable online: h ps:// www. eba. eu opa. eu/ egul a ion- and- policy/ secu i isa
ion- and- co e ed- bonds/ echn ical- ad ice- on- simple- s and a d- and- ans pa en - secu i isa ions.
Ellison, G., & Ellison, S. F. (2009). Sea ch, ob usca ion, and p ice elas ici ies on he in e ne . Econome -
ica, 77, 427–452.
E an, A., Loumio i, M., & Wi enbe g-Moe man, R. (2017). Enhancing loan quali y h ough anspa -
ency: E idence om he Eu opean cen al bank loan le el epo ing ini ia i e. Jou nal o Accoun -
ing Resea ch, 55, 877–918.
EU. (2017). Regula ion (EU) 2017/2402 Laying Down a Gene al F amewo k o Secu i isa ion and C ea -
ing a Speci ic F amewo k o Simple, T anspa en and S anda dised Secu i isa ion. Eu opean Union.
A ailable online: h ps:// eu - lex. eu opa. eu/ legal- con e n / EN/ TXT/? u i= CELEX% 3A320 17R24 02.
Fu ine, C. H. (2014). Complexi y and loan pe o mance: E idence om he secu i iza ion o comme cial
mo gages. Re iew o Co po a e Finance S udies, 2, 154–187.
Gabaix, X., & Laibson, D. (2006). Sh ouded a ibu es, consume myopia, and in o ma ion supp ession in
compe i i e ma ke s. Qua e ly Jou nal o Economics, 121, 505–540.
Ghen , A. C., To ous, W. N., & Valkano , R. I. (2019). Complexi y in s uc u ed inance. Re iew o Eco-
nomic S udies, 86, 694–722.
G i in, J., Lowe y, R., & Sa e o, A. (2014). Complex secu i ies and unde w i e epu a ion: Do epu a-
ble unde w i e s p oduce be e secu i ies? Re iew o Financial S udies, 27, 2872–2925.
Gunning, R. (1952). The echnique o clea w i ing. McG aw-Hill: 36–37.
Hibbeln, M. T., & Os e kamp, W. (2024). Simple is simply no enough— ea u es e sus lables o com-
plex inancial secu i ies. Re iew o De i a i es Resea ch, 27, 113–150.
Hwang, B.-H., & Kim, H. H. (2017). I pays o w i e well. Jou nal o Financial Economics, 124,
373–394.
Jiang, W., Nelson, A. A., & Vy lacil, E. (2014). Secu i iza ion and loan pe o mance: ex an e and ex pos
ela ions in he mo gage ma ke . Re iew o Financial S udies, 27, 454–483.
Keys, B. J., Mukhe jee, T., Se u, A., & Vig, V. (2010). Did secu i iza ion lead o lax sc eening? E idence
om subp ime loans. The Qua e ly Jou nal o Economics, 125, 307–362.
King, G., & Zeng, L. (2001). Logis ic eg ession in a e e en s da a. Poli ical Analysis, 9, 137–163.
K uge , S. (2018). The e ec o mo gage secu i iza ion on o eclosu e and modi ica ion. Jou nal o
Financial Economics, 129, 586–607.
Le i in, A. J., & Wach e , S. M. (2012). Explaining he housing bubble. Geo ge own Law Jou nal, 100,
1177–1258.
Li, F. (2008). Annual epo eadabili y, cu en ea nings, and ea nings pe sis ence. Jou nal o Accoun -
ing and Economics, 45, 221–247.
Lo, K., Ramos, F., & Rogo, R. (2017). Ea nings managemen and annual epo eadabili y. Jou nal o
Accoun ing and Economics, 63, 1–25.
Lough an, T., & McDonald, B. (2014). Measu ing eadabili y in inancial disclosu es. Jou nal o
Finance, 69, 1643–1671.
No on hesame page: comp ehensibili y o MBS in es men … Page 37 o 37 9
Mian, A., & Su i, A. (2009). The consequences o mo gage c edi expansion: E idence om he U.S.
mo gage de aul c isis. The Qua e ly Jou nal o Economics, 124, 1449–1496.
Pu nanandam, A. (2011). O igina e- o-dis ibu e model and he subp ime mo gage c isis. Re iew o
Financial S udies, 24, 1881–1915.
SEC. (1998). A plain english handbook. A ailable online: h ps:// www. sec. go / pd / handb ook. pd
Timoneda, J. C. (2021). Es ima ing g oup ixed e ec s in panel da a wi h a bina y dependen a iable:
How he LPM ou pe o ms logis ic eg ession in a e e en s da a. Social Science Resea ch, 93, A i-
cle 102486.
Wach e , S. M. (2015). The housing and c edi bubbles in he Uni ed S a es and Eu ope: A compa ison.
Real Es a e Economics, 47(S1), 37–42.
Wang, Y., & Xia, H. (2014). Do lende s s ill moni o when hey can secu i ize loans? The Re iew o
Financial S udies, 27, 2354–2391.
Zhang, H. H., Zhao, F., & Zhao, X. (2023). Neglec ed isks in he communica ion o esiden ial mo -
gage-backed secu i ies o e ings. Jou nal o Finance, 79, 129–172.
Publishe ’s No e Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in published maps
and ins i u ional a ilia ions.
