,i #aaf? ,,congress #a/ ,session ,h4 ,r4 #cfbi ,,9 ,,! ,,h\se ,,( ,,repres5tatives ,july #i1 #bjai,mr4 ,lyn* introduc$ ! foll[+ bill2 :i* 0 ref]r$ to ! ,committee on ,f9ancial ,s]vices ,a ,,bill ,to am5d ! ,fair ,cr$it ,report+ ,act to e/abli% cle> ,f$]al ov]sight ( ! developm5t ( cr$it scor+ models by ! ,bureau ( ,consum] ,f9ancial ,protection1 & = o!r purposes4 #a4 ,%ort title ,?is ,act may be cit$ z ! ,cl>ity 9 ,cr$it ,score ,=mation ,act ( #bjai4 #b4 ,f9d+s ,congress f9ds ! foll[+3 7#a7 ,! ,febru>y #bjae report ( ! ,bureau ( ,consum] ,f9ancial ,protection titl$ ,consum] ,voices on ,cr$it ,reports & ,scores f\nd t "s consum]s are reluctant to comp>ison %op = loans & o!r types ( consum] cr$it products out ( fe> t !y w l[] _! cr$it scores by do+ s4 7#b7 ,! ,bureau ( ,consum] ,f9ancial ,protection f\nd t "o ( ! mo/ common b>ri]s = p 9 review+ _! own cr$it reports & %opp+ = ! be/ cr$it t]ms 0 a lack ( und]/&+ ( ! diff]5ces betwe5 s(t & h>d inquiries & :e!r reque/+ a copy ( _! own report w\ld adv]sely impact _! cr$it /&+4 7#c7 ,! ,bureau ( ,consum] ,f9ancial ,protection reveal$ t consum]s ) accurate p]ceptions ( _! cr$itwor?9ess may be bett] equipp$ to %op = favorable cr$it t]ms4 #c4 ,,cfpb ov]sight ( cr$it scor+ models 7a7 ,9 g5]al ,! ,fair ,cr$it ,report+ ,act 7#ae ,u4,s4,c4 #afha et seq47 is am5d$-- 7#a7 by add+ at ! end ! foll[+ new section3 #fcj4 ,cr$it scor+ models 7a7 ,validat$ cr$it scor+ models ,n lat] ?an #a ye> aft] ! date ( ! enactm5t ( ?is section1 ! ,bureau %all issue f9al regulations applicable to any p]son t creates1 ma9ta9s1 or pur*ases cr$it scor+ models us$ 9 mak+ cr$it decisions to e/abli% /&>ds = validat+ ! accuracy & pr$ictive value ( all su* cr$it scor+ models1 bo? be=e release = initial use & at regul> int]vals !reaft]1 = z long z su* cr$it scor+ models are made available = pur*ase or use by su* p]son4 7b7 ,prohibition ,at lea/ once e #b ye>s1 ! ,bureau %all conduct a review ( cr$it scor+ models to det]m9e :e!r ! use ( any p>ticul> factors1 or ! weight or consid]ation giv5 to c]ta9 factors by cr$it scor+ models1 is inappropriate1 includ+ if su* factors d n enhance or contribute to ! accuracy & pr$ictive value ( ! models4 ,^u ! conclusion ( its review1 ! ,bureau may prohibit a p]son describ$ 9 subsection 7a7 f weigh+1 consid]+1 or includ+ c]ta9 factors 91 or mak+ available = pur*ase or use1 c]ta9 cr$it scor+ models or v]sions1 z ! ,bureau det]m9es appropriate4 2 & 7#b7 9 ! table ( cont5ts = su* ,act1 by add+ at ! end ! foll[+ new item3 #fcj4 ,cr$it scor+ models4 4 7b7 ,rulemak+ ,n lat] ?an ! end ( ! #b-ye> p]iod beg9n+ on ! date ( ! enactm5t ( ?is ,act1 ! ,bureau ( ,consum] ,f9ancial ,protection %all issue f9al rules to implem5t ! am5dm5ts made by ?is section4 #d4 ,,cfpb /udy & report to ,congress on ! impact ( non-traditional data 7a7 ,/udy ,! ,bureau ( ,consum] ,f9ancial ,protection %all c>ry out a /udy to assess ! impact 7includ+ ! availability & af=dability ( cr$it & o!r noncr$it decisions1 ! pot5tial positive & negative impacts on consum] cr$it scores1 & any un9t5d$ consequ5ces7 ( us+ traditional model+ te*niques or alt]native model+ te*niques to analyze non-traditional data f a consum] report & ( includ+ non-traditional data on consum] reports on ! foll[+3 7#a7 ,consum]s ) no or m9imal traditional cr$it hi/ory4 7#b7 ,traditionally und]s]v$ communities & populations4 7#c7 ,consum]s resid+ 9 rural areas4 7#d7 ,consum]s resid+ 9 urban areas4 7#e7 ,racial & e?nic m9orities & wom54 7#f7 ,consum]s across v>i\s income /rata1 p>ticul>ly consum]s e>n+ less ?an #abj p]c5t ( ! area m$ian income 7z def9$ by ! ,secret>y ( ,h\s+ & ,urban ,developm5t74 7#g7 ,immigrants1 refugees1 & non-p]man5t resid5ts4 7#h7 ,m9ority f9ancial in/itutions 7z def9$ "u section #cjh7;b7 ( ! ,f9ancial ,in/itutions ,re=m1 ,recov]y1 & ,en=cem5t ,act ( #aihi 7#ab ,u4,s4,c4 #adfc note77 & community f9ancial in/itutions4 7#i7 ,consum]s resid+ 9 f$]ally assi/$ h\s+1 includ+ consum]s receiv+ ,f$]al r5tal subsidies4 7b7 ,additional consid]ations ,9 assess+ impacts "u subsection 7a71 ! ,bureau ( ,consum] ,f9ancial ,protection %all also consid] impacts on-- 7#a7 ! privacy1 security1 & confid5tiality ( ! f9ancial1 m$ical1 & p]sonally id5tifiable in=mation ( consum]s2 7#b7 ! control ( consum]s ov] h[ su* in=mation may or w be us$ or consid]$2 7#c7 ! und]/&+ ( consum]s ( h[ su* in=mation may be us$ or consid]$ & ! ease ) :i* a consum] may decide to re/rict or prohibit su* use or consid]ation ( su* in=mation2 7#d7 pot5tial discrim9atory effects2 & 7#e7 disp>ate outcomes ! use or consid]ation ( su* in=mation may cause4 7c7 ,report ,n lat] ?an #a ye> aft] ! date ( ! enactm5t ( ?is ,act1 ! ,bureau ( ,consum] ,f9ancial ,protection %all issue a report to ! ,committee on ,f9ancial ,s]vices ( ! ,h\se ( ,repres5tatives & ! ,committee on ,bank+1 ,h\s+1 & ,urban ,affairs ( ! ,s5ate conta9+ all f9d+s & det]m9ations1 includ+ any recomm5dations = any legislative or regulatory *anges1 made 9 c>ry+ out ! /udy requir$ "u subsection 7a74 7d7 ,def9itions ,9 ?is section3 7#a7 ,alt]native model+ te*niques ,! t]m alt]native model+ te*niques means /ati/ical & ma!matical te*niques t are n traditional model+ te*niques1 includ+ decision trees1 r&om =e/s1 artificial neutral networks1 ne>e/ neighbor1 g5etic programm+1 & boo/+ algori?ms4 7#b7 ,consum] report ,! t]m consum] report has ! mean+ giv5 su* t]m 9 section #fjc ( ! ,fair ,cr$it ,report+ ,act 7#ae ,u4,s4,c4 #afha;a74 7#c7 ,non-traditional data ,! t]m non-traditional data means data relat$ to telecommunications1 utility paym5ts1 r5t paym5ts1 remittances1 wire transf]s1 data n o!rwise regul>ly includ$ 9 consum] reports issu$ by consum] report+ ag5cies describ$ "u section #fjc7p71 & su* o!r items z ! ,bureau ( ,consum] ,f9ancial ,protection deems appropriate4 7#d7 ,traditional model+ te*niques ,! t]m traditional model+ te*niques means /ati/ical & ma!matical te*niques 7includ+ models1 algori?ms1 l9e> & logi/ic regression me?ods1 & _! outputs7 t are traditionally us$ 9 automat$ und]writ+ processes4