Miguel Rosiles-Camarena v. Eric Holder, Jr. , 735 F.3d 534 ( 2013 )

In the
United States Court of Appeals
For the Seventh Circuit
____________________
No.  11-­‐‑3086
MIGUEL  A.  ROSILES-­‐‑CAMARENA,
Petitioner,
v.
ERIC  H.  HOLDER,  JR.,  Attorney  General  of  the  United  States,
Respondent.
____________________
Petition  for  Review  of  the  Decision  of  the
Board  of  Immigration  Appeals
____________________
ARGUED  MARCH  26,  2012  —  DECIDED  AUGUST  21,  2013
____________________
Before  EASTERBROOK,  Chief  Judge,  and  BAUER  and  WOOD,
Circuit  Judges.
EASTERBROOK,   Chief   Judge.   Miguel   Rosiles-­‐‑Camarena,   a
citizen   of   Mexico,   was   admitted   to   the   United   States   for
permanent  residence  in  1977,  when  he  was  ten  years  old.  He
did  not  use  his  opportunities  to  become  a  citizen.  Following
his  felony  conviction  for  indecent  solicitation  of  a  minor,  his
permanent-­‐‑residence   status   was   revoked,   and   he   has   been
ordered  removed  to  Mexico.
No.  11-­‐‑3086   2
Rosiles-­‐‑Camarena   is   homosexual   and   HIV   positive.   He
contends  that  gays  are  persecuted  in  Mexico  (at  least  outside
of  cosmopolitan  Mexico  City)  and  that  gays  infected  by  HIV
face   extra   risk.   Although   he   is   not   eligible   for   asylum   (the
deadline  for  seeking  that  relief  expired  long  ago),  he  applied
for   withholding   of   removal   under   8   U.S.C.   §1231(b)(3),   and
relief   under   the   Convention   Against   Torture,   implemented
by  8  C.F.R.  §§  1208.16  to  .18.  To  be  eligible  for  either  benefit,
an  alien  must  show  a  clear  probability  that  persecution  (for
withholding   of   removal)   or   torture   (for   the   Convention)   is
more  likely  than  not  in  the  alien’s  native  country.
The   immigration   judge   and   the   Board   of   Immigration
Appeals   disagree   about   whether   Rosiles-­‐‑Camarena   satisfies
these   requirements.   The   IJ   initially   granted   his   application
for  relief  under  both  the  statute  and  the  Convention,  finding
on   the   basis   of   statistics   and   expert   testimony   that   Rosiles-­‐‑
Camarena  probably  would  be  killed  or  injured  in  Mexico  as
a  result  of  his  sexuality  and  disease.  The  BIA  remanded,  but
the   IJ   adhered   to   his   position   on   remand.   The   BIA   then   re-­‐‑
versed  and,  after  a  remand  (by  consent)  from  this  court,  ad-­‐‑
hered   to   its   position.   The   most   recent   decision   states   that
“[t]he  probability  of  future  harm  is  a  legal  question  that  we
review  de  novo”  and  that,  “[i]n  assessing  the  probability  of
harm  de  novo,  we  may  give  different  weight  to  the  evidence
than  did  the  Immigration  Judge.”  The  BIA  proceeded  to  do
just   that.   It   accepted   all   of   the   IJ’s   findings   of   historical   fact
but   disagreed   with   the   IJ   about   the   risk   implied   by   those
facts.
For   example:   the   IJ   found   that   Rosiles-­‐‑Camarena   is   at
substantial  risk  because  148  persons  were  murdered  in  Mex-­‐‑
ico,  between  1995  and  2006,  because  of  their  sexual  orienta-­‐‑
3   No.  11-­‐‑3086
tion.   But   the   Board   observed   that   this   amounts   to   12   or   13
killings  a  year  in  a  population  exceeding  110  million,  at  least
2%  of  which  is  homosexual,  making  it  unlikely  (a  risk  of  no
more   than   1   in   100,000)   that   any   given   gay   man   would   be
killed  any  given  year.  Expert  testimony  establishing  that  “at-­‐‑
tacks  on  homosexuals  are  frequent”  does  not  show  the  mag-­‐‑
nitude   of   risks,   any   more   than   expert   testimony   that   “auto
accidents   are   frequent”   would   imply   that   a   given   driver
(even   one   in   a   high-­‐‑risk   group,   such   as   men   under   25)   is
more  likely  than  not  to  be  injured.  The  Board  stated  that  the
IJ   did   not   commit   clear   error   in   crediting   the   statistics   and
the   expert’s   testimony   but   added:   “as   atrocious   as   it   is   to
have  12  or  13  such  killings  per  year,  that  fact  does  not  show
a   clear   probability   that   [Rosiles-­‐‑Camarena]   will   be   killed   or
otherwise   persecuted.”   The   Board   treated   the   risk   of   future
harm  as  a  matter  of  legislative  fact,  and  it  took  the  view  that
decisions   on   mixed   (or   “ultimate”)   questions   are   open   to
plenary  decision.  Rosiles-­‐‑Camarena  contends  that  the  Board
made  a  legal  error  by  engaging  in  this  kind  of  review.
He  also  contends  that  the  Board’s  decision  lacks  substan-­‐‑
tial   evidence   in   the   record,   but   we   lack   jurisdiction   to   ad-­‐‑
dress   that   subject.   He   has   been   convicted   of   an   aggravated
felony,  and  as  a  result  8  U.S.C.  §1252(a)(2)(C)  forbids  judicial
review  of  the  removal  decision,  except  to  the  extent  that  the
alien   presents   legal   arguments   (statutory   or   constitutional).
See   8   U.S.C.   §1252(a)(2)(D).   A   contention   that   the   agency’s
decision   is   not   supported   by   enough   evidence   is   not   a   “le-­‐‑
gal”   argument   for   this   purpose.   See   Jiménez   Viracacha   v.
Mukasey,  518  F.3d  511  (7th  Cir.  2008);  Paez  Restrepo  v.  Holder,
610   F.3d   962   (7th   Cir.   2010).   Section   1252(a)(2)(C)   applies   to
applications  for  relief  based  on  §1231(b)(3).  See  Moral-­‐‑Salazar
v.   Holder,   708   F.3d   957   (7th   Cir.   2013).   Although   Moral-­‐‑
No.  11-­‐‑3086   4
Salazar  expresses  a  reservation  for  CAT  claims,  we  need  not
explore  in  this  litigation  what  sort  of  arguments  under  CAT
§1252(a)(2)(C)  allows  us  to  consider.
A  regulation  specifies  the  extent  to  which  the  Board  may
review   or   supplement   factual   decisions   by   immigration
judges.  It  provides:
(i)   The   Board   will   not   engage   in   de   novo   review   of   findings   of
fact   determined   by   an   immigration   judge.   Facts   determined   by
the  immigration  judge,  including  findings  as  to  the  credibility  of
testimony,  shall  be  reviewed  only  to  determine  whether  the  find-­‐‑
ings  of  the  immigration  judge  are  clearly  erroneous.
(ii)   The   Board   may   review   questions   of   law,   discretion,   and
judgment   and   all   other   issues   in   appeals   from   decisions   of   im-­‐‑
migration  judges  de  novo.
…
(iv)  Except  for  taking  administrative  notice  of  commonly  known
facts  such  as  current  events  or  the  contents  of  official  documents,
the  Board  will  not  engage  in  factfinding  in  the  course  of  deciding
appeals.   A   party   asserting   that   the   Board   cannot   properly   re-­‐‑
solve   an   appeal   without   further   factfinding   must   file   a   motion
for  remand.  If  further  factfinding  is  needed  in  a  particular  case,
the  Board  may  remand  the  proceeding  to  the  immigration  judge
or,  as  appropriate,  to  the  Service.
8   C.F.R.   §1003.1(d)(3).   An   argument   that   the   Board   has   ex-­‐‑
ceeded  the  scope  of  review  permissible  under  this  regulation
is  a  legal  one,  for  the  purpose  of  §1252(a)(2)(D).  See  Rotinsulu
v.  Mukasey,  515  F.3d  68,  72  (1st  Cir.  2008).
Matter   of   V–   K–,   24   I&N   Dec.   500   (2008),   on   which   the
Board   relied   here,   concludes   that   §1003.1(d)(3)(i)   does   not
prevent   it   from   disagreeing   with   an   IJ’s   predictions   about
the  likelihood  of  future  harm.  V–  K–  gives  two  principal  rea-­‐‑
sons.  First,  clause  (ii)  authorizes  the  Board  to  “review  ques-­‐‑
5   No.  11-­‐‑3086
tions  of  law,  discretion,  and  judgment”—and  if  the  probabil-­‐‑
ity  of  harm  is  an  issue  of  fact,  it  is  also  one  of  “law”  (to  the
extent   the   Board   must   choose   “how   probable   is   probable
enough?”)  and  of  “judgment”  (because  evaluating  the  prob-­‐‑
ability   of   harm   requires   the   application   of   judgment   to   his-­‐‑
torical   facts).   Second,   the   Board   observed   that   the   explana-­‐‑
tion   issued   with   the   adoption   of   §1003.1(d)(3)   reveals   that
the  resolution  of  a  mixed  question  of  law  and  fact  is  not  itself
a  “fact”  for  the  purpose  of  clause  (i).  The  commentary  stated
that  clause  (ii)  covers  “judgments  as  to  whether  the  facts  es-­‐‑
tablished   by   a   particular   alien   amount   to   ‘past   persecution’
or  a  ‘well  founded  fear  of  persecution.’”  67  Fed.  Reg.  54,878,
54,890  (Aug.  26,  2002).
The  first  circuit  has  held  that  the  approach  articulated  in
V–  K–  is  within  the  Board’s  authority.  See  Rotinsulu,  515  F.3d
at  73;  Sicaju-­‐‑Diaz  v.   Holder,  663  F.3d  1,   5  (1st  Cir.   2011).  The
third   circuit,   by   contrast,   set   aside   V–   K–   on   petition   for   re-­‐‑
view.   Kaplun   v.   Attorney   General,   602   F.3d   260,   269–71   (3d
Cir.   2010)   (CAT).   In   En   Hui   Huang   v.   Attorney   General,   620
F.3d  372,  381–87  (3d  Cir.  2010),  it  applied  Kaplun  to  applica-­‐‑
tions   for   withholding   of   removal.   Four   other   circuits   have
agreed  with  the  third.  See  Hui  Lin  Huang  v.  Holder,  677  F.3d
130   (2d   Cir.   2012);   Turkson   v.   Holder,   667   F.3d   523   (4th   Cir.
2012);  Ridore  v.  Holder,  696  F.3d  907  (9th  Cir.  2012);  Zhou  Hua
Zhu   v.   Attorney   General,   703   F.3d   1303   (11th   Cir.   2013).   The
third  circuit  concluded  that  the  Board  is  entitled  to  adopt  an
independent  view  about  whether  a  potential  harm  identified
by   an   IJ   amounts   to   “persecution”   or   “torture,”   but   that   an
IJ’s  predictions  (which  it  called  the  “present  probability  of  a
future   event”)—such   that   a   particular   harm   is   “likely”
should  an  alien  return  to  his  native  land—are  “facts”  under
No.  11-­‐‑3086   6
clause  (i),  and  the  Board’s  role  is  limited  to  identifying  clear
error  by  the  IJ.
The  Board’s  decision  in  our  case  adds  some  rationales  in
the   course   of   explaining   why   it   finds   Kaplun   and   En   Hui
Huang   unpersuasive.   The   Board’s   principal   concern   is   that
its   legal   views   won’t   have   much   significance   if   all   predic-­‐‑
tions  are  facts.  It  observed  that  “‘predictive’  findings,  partic-­‐‑
ularly  regarding  the  level  of  harm  that  is  likely  to  be  inflict-­‐‑
ed,  may  preordain  resolution  of  the  legal  question  regarding
whether   such   harm   rises   to   the   level   of   persecution   or   tor-­‐‑
ture.”   The   Board   added:   “any   such   predictive   findings   are
likely   to   be   based   on   written   reports   of   country   conditions
over   which   a   trier   of   fact   has   no   particular   expertise   or   ad-­‐‑
vantage,  in  contrast  to  issues  of  credibility,  resolution  of  con-­‐‑
flicting  testimony,  or  questions  of  historical  fact.”
In   other   words,   the   Board   thought   that   the   category   of
predictions  identified  by  the  third  circuit  often  concerns  leg-­‐‑
islative   rather   than   adjudicative   facts.   A   sound   prediction
depends   on   country   conditions,   not   (necessarily)   on   facts
unique   to   the   alien.   For   example,   Rosiles-­‐‑Camarena   has
lived   in   the   United   States   since   he   was   10   and   has   visited
Mexico   only   briefly.   Many   of   the   IJ’s   predictions   concern
conditions   in   Mexico.   The   Board   thinks   it   unacceptable   to
have   one   IJ   assert   that   conditions   in   Mexico   are   horrible,
while   another   deems   them   fine,   and   to   have   both   conclu-­‐‑
sions  immune  to  effective  review.
Immigration  judges  display  substantial  disparity  in  eval-­‐‑
uating  claims  for  asylum  or  withholding  of  removal.  See  Ja-­‐‑
ya   Ramji-­‐‑Nogales,   Andrew   I.   Schoenholtz   &   Philip   G.
Schrag,  Refugee  Roulette:  Disparities  in  Asylum  Adjudication,  60
Stan.  L.  Rev.  295  (2007).  The  Board  thinks  that  it  is  entitled  to
7   No.  11-­‐‑3086
curtail  IJs’  divergent  approaches  and  believes  that  it  can  do
so   by   determining   whether   particular   countries   are,   or   are
not,   hostile   to   particular   political   or   social   groups.   Indeed,
we  have  urged  the  Board  to  make  categorical  decisions.  E.g.,
Banks   v.   Gonzales,   453   F.3d   449,   453–55   (7th   Cir.   2006);   Xiu
Ling  Chen  v.  Gonzales,  489  F.3d  861,  862  (7th  Cir.  2007);  Chun
Hua  Zheng  v.  Holder,  666  F.3d  1064,  1068  (7th  Cir.  2012).  The
Board  fears  that,  under  Kaplun  and  similar  decisions,  every  IJ
may  maintain  a  personal  view  about  the  risks  in  each  nation.
Both   the   Board   and   the   courts   of   appeals   (reviewing   it)
would  have  their  hands  tied.  Rosiles-­‐‑Camarena  would  bene-­‐‑
fit  from  such  a  limit  on  review,  but  aliens  whose  claims  are
denied  by  other  IJs  would  lose;  neither  the  Board  nor  a  court
of  appeals  could  conclude  that  country  conditions  are  more
hazardous  than  the  IJ  found.
The  Board  has  interpreted  §1003.1(d)(3)(i)  in  a  way  that  it
believes  makes  clauses  (i)  and  (ii)  harmonious  by  relying  on
the  longstanding  distinctions  between  adjudicative  and  leg-­‐‑
islative  facts,  and  between  historical  (case-­‐‑specific)  facts  and
those   aspects   of   discretion   or   judgment   that   concern   coun-­‐‑
try-­‐‑wide   conditions—subjects   on   which   the   Board   thinks
that   the   United   States   should   be   able   to   speak   with   one
voice,   rather   than   through   a   cacophony   of   immigration
judges.  The  problem  is  that  the  Board’s  arguments  would  be
better   as   reasons   to   revise   the   regulation   than   as   reasons   to
interpret   it   differently   from   the   similar   language   that   gov-­‐‑
erns  relations  between  federal  trial  and  appellate  courts.
Appellate   courts   are   bound   by   Fed.   R.   Civ.   P.   52(a)(6),
which  says  that  district  judges’  findings  of  fact  must  not  be
set   aside   unless   clearly   erroneous.   See   also   Anderson   v.   Bes-­‐‑
semer  City,  470  U.S.  564  (1985).  Rule  52(a)(6)  may  have  been
No.  11-­‐‑3086   8
the   inspiration   for   §1003.1(d)(3)(i).   The   Rule   does   not   allow
plenary  appellate  review  of  district  judges’  characterizations
based   on   historical   facts.   A   finding   on   an   “ultimate”   fact   (a
conclusion  based  on  the  application  of  legal  rules  to  histori-­‐‑
cal   facts)   often   is   treated   as   one   of   fact.   See,   e.g.,   Pullman-­‐‑
Standard  v.  Swint,  456  U.S.  273  (1982)  (whether  the  plaintiff  is
a  victim  of  “discrimination”  is  a  question  of  fact);  Icicle  Sea-­‐‑
foods,   Inc.   v.   Worthington,   475   U.S.   709   (1986)   (whether   the
plaintiff  is  a  “seaman”  is  a  question  of  fact).  When  a  decision
is   person-­‐‑specific,   there   is   rarely   a   good   reason   for   having
three  judges  substitute  their  views  for  the  assessment  of  the
trial  judge.  See,  e.g.,  Mars  Steel  Corp.  v.  Continental  Bank  N.A.,
880  F.2d  928,  933–34  (7th  Cir.  1989)  (en  banc);  Mucha  v.  King,
792   F.2d   602,   605–06   (7th   Cir.   1986).   When   a   decision   de-­‐‑
pends   on   the   characteristics   of   non-­‐‑litigants,   however,   and
establishes  a  rule  with  broad  scope,  the  appellate  role  can  be
more  substantial.  For  example,  in  American  Needle,  Inc.  v.  Na-­‐‑
tional  Football  League,  560  U.S.  183  (2010),  the  Supreme  Court
gave   no   weight   to   a   district   judge’s   conclusion   that   the   Na-­‐‑
tional   Football   League   should   be   characterized   as   a   single
firm   for   the   purpose   of   antitrust   law.   But   even   that   sort   of
review,   which   concerns   legal   consequences,   differs   from   an
entitlement   to   make   an   independent   appellate   decision   on
facts  just  because  they  are  “legislative”  in  nature.
Kaplun   observed   that   many   predictions   are   facts,   in   the
sense   that   they   rest   on   subsidiary   facts   and   can   be   true   or
false.   It   gave   this   example:   “It   is   likely   that   it   will   take   less
than  3  hours  to  drive  the  100  miles  to  grandmother’s  house
next   week.”   602   F.3d   at   269.   Likewise,   a   medical   prediction
about  whether  a  victim  of  injury  will  recover  is  factual,  even
though   it   rests   on   the   application   of   medical   knowledge   to
subsidiary  facts.  Id.  at  270.  These  illustrations  show  how  per-­‐‑
9   No.  11-­‐‑3086
son-­‐‑specific   circumstances   (adjudicative   facts)   can   give   rise
to   predictions   that   also   are   sensibly   treated   as   facts.   That   is
as  true  when  a  prediction  depends  on  country  conditions  as
when  it  depends  on  what  happened  to  a  particular  alien.  We
therefore  agree  with  Kaplun  and  similar  decisions.
Perhaps  the  Board’s  view  that  it  may  make  independent
decisions  about  predictions  is  a  consequence  of  giving  itself
too  little  leeway  to  find  IJs’  predictions  clearly  erroneous.  A
federal  court  of  appeals  would  be  inclined  to  think  it  a  clear
error,   correctible   under   Rule   52,   for   a   district   judge   to   say
that  a  1-­‐‑in-­‐‑100,000  chance  of  death  meets  a  more-­‐‑likely-­‐‑than-­‐‑
not  burden  of  persuasion.  But  that’s  not  what  the  Board  did.
Instead  it  claimed  a  right  to  substitute  its  judgment  for  that
of  the  IJ  without  finding  a  clear  error.  That  is  a  mistake  under
the   regulation.   A   court   is   limited   to   the   agency’s   stated
grounds  of  decision  and  cannot  enforce  an  order  on  a  basis
that  the  agency  did  not  include  among  its  reasons.  Our  task,
having  corrected  a  legal  error,  is  to  remand  to  the  Board  ra-­‐‑
ther  than  make  our  own  decision.  See,  e.g.,  Gonzales  v.  Thom-­‐‑
as,  547  U.S.  183  (2006).
To  say  that  the  regulation  leaves  the  Board  free  to  declare
an   IJ’s   findings   clearly   erroneous   is   not   at   all   to   say   that   it
would  be  appropriate  for  the  Board  to  do  so  in  this  case.  For
although  we  have  mentioned  so  far  only  the  statistical  risk  of
death   for   homosexuals   as   a   group,   Rosiles-­‐‑Camarena   con-­‐‑
tends  that  he  is  at  greater  risk.  He  is  not  only  gay  and  HIV
positive  but  also  “out”  and  planning  to  live  openly  with  his
partner.   He   contends,   and   the   IJ   found,   that   his   family   has
disowned  him  and  will  not  offer  any  support.  He  adds  that,
because  he  has  lived  in  the  United  States  most  of  his  life  and
does  not  know  contemporary  Mexican  customs,  he  will  find
No.  11-­‐‑3086   10
it  hard  to  avoid  attracting  attention  from  persons  who  might
do  him  harm.  And  he  stresses  that  injuries  (and  deprivations
of   economic   opportunities)   short   of   death   may   amount   to
persecution.   The   question   for   the   Board   on   remand   is   thus
not   whether   aggregate   data   imply   that   Rosiles-­‐‑Camarena   is
likely  to  be  killed,  but  whether  the  IJ  clearly  erred  in  finding
that   he   is   more   likely   than   not   to   be   persecuted.   That   ques-­‐‑
tion  is  for  the  Board  in  the  first  instance;  we  do  not  express
an  opinion  on  it.
The   petition   for   review   is   granted,   and   the   matter   is   re-­‐‑
manded   to   the   Board   for   proceedings   consistent   with   this
opinion.