Win e s, John V.
A icle
Di e ences in employmen ou comes o college own
s aye s and lea e s
IZA Jou nal o Mig a ion
P o ided in Coope a ion wi h:
IZA – Ins i u e o Labo Economics
Sugges ed Ci a ion: Win e s, John V. (2012) : Di e ences in employmen ou comes o college own
s aye s and lea e s, IZA Jou nal o Mig a ion, ISSN 2193-9039, Sp inge , Heidelbe g, Vol. 1, pp. 1-22,
h ps://doi.o g/10.1186/2193-9039-1-11
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/92228
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/2.0/
ORIGINAL ARTICLE Open Access
Di e ences in employmen ou comes o college
own s aye s and lea e s
John V Win e s
Co espondence:
[email p o ec ed]
Depa men o Economics,
Uni e si y o Cincinna i and IZA, PO
Box 210371, Cincinna i, OH
45221-0371, USA
Abs ac
New college g adua es mus choose whe he o s ay in he geog aphic a ea whe e
hey comple ed hei deg ee o mo e o a new loca ion o begin hei ca ee s. This
pape classi ies 41 U.S. me opoli an a eas as “college owns”and in es iga es
di e ences in employmen ou comes be ween college g adua es who s ay in he
college own whe e hey ob ained hei deg ee and college g adua es who lea e
a e comple ing hei deg ee. We ind ha college own s aye s expe ience less
a o able employmen ou comes along mul iple dimensions. On a e age, s aye s
ea n lowe annual and hou ly wages and wo k in less educa ed occupa ions.
JEL codes: J61, R23
Keywo ds: Mig a ion, Human capi al, Highe educa ion, College owns, Wages
1. In oduc ion
In es men s in highe educa ion ha e impo an bene i s o indi iduals because human
capi al inc eases p oduc i i y, wages, and employmen p ospec s (Psacha opoulos and
Pa inos 2004)
1
. Howe e , indi idual employmen ou comes and bene i s o educa ion
a e inhe en ly ied o loca ion decisions. A e comple ing hei educa ion, indi iduals
ha e o decide whe e o begin hei ca ee s. Many college g adua es will choose o
begin hei ca ee s in di e en geog aphic a eas han whe e hey a ended college, bu
some will emain in he same a ea as whe e hey a ended college (Hu man and
Quigley 2002; G oen 2004; Oos e beek and Webbink 2011). The e a e bo h bene i s
and cos s in deciding whe he o mig a e and ela i e employmen p ospec s likely
play an impo an ole.
Mig a ion decisions a e o en iewed as human capi al in es men s (Sjaas ad 1962).
In his amewo k indi iduals maximize hei own expec ed u ili y and choose o li e
in he loca ion ha o e s he highes expec ed u ili y. Thus, upon comple ing he
deg ee, a college g adua e will mo e o a new a ea i i gi es he g ea e expec ed u ili y
han he place whe e she ea ned he deg ee. The expec ed u ili y ha an a ea o e s
will depend on a numbe o ac o s including wages and employmen oppo uni ies,
he local cos o li ing, consume ameni ies and quali y o li e, and he as es and p e-
e ences o he indi idual making he mig a ion decision (Cla k and Hun e 1992;
F anklin 2003). Chen and Rosen hal (2008) sugges ha young college g adua es a e
o en mos s ongly a ac ed o la ge me opoli an a eas wi h high wages and a s ong
labo ma ke . Whisle e al. (2008) a gue ha cul u al and ec ea ional ameni ies also
© 2012 Win e s; licensee Sp inge . This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion
License (h p://c ea i ecommons.o g/licenses/by/2.0), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium,
p o ided he o iginal wo k is p ope ly ci ed.
Win e s IZA Jou nal o Mig a ion 2012, 1:11
h p://www.izajom.com/con en /1/1/11
play an impo an ole in college g adua e mig a ion decisions. Fu he mo e, Be y and
Glaese (2005) and Waldo (2009) sugges ha educa ed indi iduals a e a ac ed o
a eas ha al eady ha e la ge numbe s o educa ed people.
While some a eas a e ce ainly be e han o he s a a ac ing and e aining ecen
college g adua es, he op imal loca ion will no be he same o all indi iduals. Indi i-
duals ha e di e en skill se s and di e en p e e ences o a ious loca ions. Fu he -
mo e, some g adua es may de elop human capi al ha is speci ic o he loca ion whe e
hey ea ned hei college deg ee (DaVanzo 1983; K upka 2009; Win e s 2011). Fo
example, hey may de elop ne wo ks wi h p o esso s and local employe s ha enhance
hei employmen p ospec s in he a ea. Al e na i ely, ecen g adua es may ha e buil
s ong iendships and de eloped a as e o local ameni ies ha help ie hem o he
a ea whe e hey a ended college. These p e e ences o he place whe e hey a ended
college may e en incline some ecen g adua es o accep lowe paying jobs o s ay in
he a ea ha hey ha e g own o lo e.
To da e he e has been e y li le esea ch examining how employmen ou comes o
ecen g adua es a e a ec ed by s aying in he a ea whe e hey a ended college. P e i-
ous esea che s ha e examined he e ec s o mig a ion on income and employmen o
a ious g oups (e.g., Axelsson and Wes e lund 1998; Yankow 2003; Boheim and Taylo
2007; Anil, Sjoquis , and Wallace 2010; Blackbu n 2010a,b), bu ha e no explici ly
examined he e ec s o s aying in a college own o ecen g adua es. Many ecen
college g adua es migh be be e able o ind gain ul employmen by eloca ing o a
di e en labo ma ke and o cou se many g adua es do lea e he a ea whe e hey
comple ed hei educa ion. Recen g adua es who s ay in he a ea whe e hey ea ned
hei deg ee may pay a p ice o doing so, and his p ice is likely o be especially la ge
in smalle labo ma ke s ha p oduce la ge numbe s o college g adua es, i.e., “college
owns.”
This pape seeks o ill an impo an gap in he esea ch li e a u e by in es iga ing
di e ences in employmen ou comes be ween ecen college g adua es who s ay in he
college own whe e hey ob ained hei deg ee and hose who lea e a e comple ing
hei deg ee. Using decennial census da a, we classi y 41 U.S. me opoli an a eas as
college owns and examine di e ences in he p obabili y o employmen , annual wage
income, hou ly wages, annual hou s wo ked, and a measu e o he educa ion le el o a
wo ke ’s occupa ion. The esul s sugges ha , on a e age, college own s aye s ea n
lowe annual and hou ly wages and wo k in less educa ed occupa ions. Thus, i appea s
ha college own s aye s gene ally expe ience wo se employmen ou comes han hose
who lea e. These esul s con inue o hold when we employ a di e ence in di e ences
app oach ha ea s ecen g adua es om la ge me opoli an a eas as a con ol g oup
and when we con ol o di e ences in he cos o li ing ac oss labo ma ke s.
2. Empi ical app oach and da a
This pape examines wage and employmen di e ences be ween ecen college g a-
dua es who s ay in he college own whe e hey ea ned hei deg ee and hose who
lea e a college own a e hei educa ion is comple e. We ocus on ecen college g a-
dua es who ea ned hei deg ees in ela i ely small college owns because i is hese
g adua es who a e likely o ha e he g ea es di icul y inding a good job in he a ea
Win e s IZA Jou nal o Mig a ion 2012, 1:11 Page 2 o 17
h p://www.izajom.com/con en /1/1/11
whe e hey comple ed hei deg ee. La ge me opoli an a eas like New Yo k and Los
Angeles ha e hicke labo ma ke s and gene ally o e be e employmen p ospec s
o ecen g adua es and a e be e able o e ain ecen college g adua es (Whisle
e al. 2008). This pape uses mic oda a om he 2000 Census a ailable om he
IPUMS (Ruggles e al. 2008) o classi y 41 U.S. me opoli an a eas as college owns. We
de ine a me opoli an a ea as a college own based on he pe cen age o adul s in he
a ea who a e en olled in college. Mo e speci ically, an a ea is conside ed a college own
i he sha e o he popula ion age 18 and olde en olled in college is mo e han one
s anda d de ia ion g ea e han he mean ac oss all me opoli an a eas. Fo he yea
2000, he mean sha e o adul s en olled in college ac oss me opoli an a eas was 0.092,
and he s anda d de ia ion ac oss me opoli an a eas was 0.050. Thus an a ea is consi-
de ed a college own i i s sha e en olled in college exceeds 0.142. The 41 me opoli an
a eas mee ing his c i e ion and included in his s udy a e lis ed in Table 1.
Admi edly, his de ini ion o college owns is a ai ly es ic i e one, and a case could
be made ha many a eas no included should be igh ully conside ed as college owns.
The p esen de ini ion is in ended o limi he analysis o a eas whe e colleges and
uni e si ies ha e a ela i ely la ge in luence. Fu he mo e, he 41 a eas mee ing his
c i e ion a e also ela i ely small me opoli an a eas ha ing a mean popula ion o
203,919. Again, i is hese smalle a eas in which ecen g adua es may su e he mos
om s aying in he a ea whe e hey wen o college. Ideally, we migh also like o
include non-me opoli an college owns, bu un o una ely he Census public use
mic oda a do no allow sepa a e iden i ica ion o geog aphic a eas wi h popula ions less
Table 1 The 41 "college own" me opoli an a eas included in he s udy
Albany-Co allis-Lebanon, OR CSA Kalamazoo-Po age, MI CBSA
Ames-Boone, IA CSA La C osse, WI-MN CBSA
A hens-Cla ke Coun y, GA CBSA La aye e-F ank o , IN CSA
Bellingham, WA CBSA Lansing-Eas Lansing-Owosso, MI CSA
Blacksbu g-Ch is iansbu g-Rad o d, VA CBSA Las C uces, NM CBSA
Blooming on, IN CBSA Law ence, KS CBSA
Blooming on-No mal, IL CBSA Lewis on, ID-WA CBSA
Champaign-U bana, IL CBSA Lincoln, NE CBSA
Cha lo es ille, VA CBSA Logan, UT-ID CBSA
Cheyenne, WY CBSA Lubbock-Le elland, TX CSA
Chico, CA CBSA Madison-Ba aboo, WI CSA
College S a ion-B yan, TX CBSA Mo gan own, WV CBSA
Columbia, MO CBSA Muncie, IN CBSA
Columbus-Aubu n-Opelika, GA-AL CSA P o o-O em, UT CBSA
Flags a , AZ CBSA San Luis Obispo-Paso Robles, CA CBSA
Gaines ille, FL CBSA San a Ba ba a-San a Ma ia-Gole a, CA CBSA
G and Fo ks, ND-MN CBSA S a e College, PA CBSA
G een ille, NC CBSA Tallahassee, FL CBSA
Ha iesbu g, MS CBSA Tuscaloosa, AL CBSA
Iowa Ci y, IA CBSA Waco, TX CBSA
I haca-Co land, NY CSA
No e: Me opoli an a eas a e classi ied as "college owns" i he sha e o he adul popula ion (age 18 and olde ) en olled
in college is mo e han one s anda d de ia ion g ea e han he mean ac oss all me opoli an a eas.
Win e s IZA Jou nal o Mig a ion 2012, 1:11 Page 3 o 17
h p://www.izajom.com/con en /1/1/11
han 100,000. Consequen ly, he lowes le el o iden i iable geog aphy, he PUMA,
lumps e y small a eas oge he wi h o he nea by a eas ha o en span la ge land
masses. This makes i ha de o iden i y whe he ecen college g adua es emain in
non-me opoli an college owns because hey migh s ill be in he same PUMA bu
ha e mo ed o a new loca ion a conside able dis ance om whe e hey a ended
college. These geog aphic iden i ica ion p oblems also a ec a ew o he me opoli an
a eas included in his s udy, hough o a lesse ex en han would be ue o
non-me opoli an a eas.
A e de ining he 41 college own me opoli an a eas examined in his s udy, we nex
wish o conside di e ences in employmen ou comes be ween ecen g adua es who
s ay in hese college owns and hose who lea e a e comple ing hei deg ee. We
de ine ecen college g adua es as pe sons be ween he ages o 23 and 27 wi h a leas a
ou -yea college deg ee. We do no obse e when an indi idual comple ed he deg ee,
so ou de ini ion o ecen is based on age. College g adua es ages 23–27 a he ime o
he census we e ages 18–22 i e yea s be o e he census and we e likely o ha e been
en olled in college bu no ye ha e ea ned a bachelo ’s deg ee. The main esul s p e-
sen ed below a e quali a i ely obus o al e ing he age ange o include pe sons ages
28 and 29 a he ime o he census o exclude pe sons age 23 a he ime o he census.
Ou ini ial analysis also es ic s he sample o ecen college g adua es who esided in
one o he 41 college owns i e yea s p io o he census. Un o una ely, he e may be
some mis epo ing o p e ious esidence o ecen college g adua es. Some ecen
g adua es li ing in a college own i e yea s p io may ins ead epo hei esidence
i e yea s ago as hei pa en s’ esidence. I is unclea how much o in which di ec ion
he esul ing sampling and measu emen e o would a ec he esul s.
The sample also excludes wo g oups o ecen g adua es. Fi s , we exclude pe sons
who a e s ill en olled in highe educa ion in 2000. Since hese people a e s ill in school
i does no make sense o compa e hei employmen ou comes o hose who ha e
comple ed hei schooling. We also exclude all pe sons bo n ou side he U.S. since
many o hese only come o he U.S. o an educa ion and lea e a e comple ing hei
deg ee. As a p ac ical ma e , he esul s below a e quali a i ely obus o including
ecen college g adua es who a e cu en ly en olled and o eign bo n g adua es who
a e s ill in he U.S. in 2000, bu he esul s p esen ed do no include hese g oups o e-
cen g adua es. A e es ic ing he sample along hese dimensions we a e le wi h
11,237 indi iduals.
We nex de ine he bina y a iable, S aye , which is equal o one i he ecen g a-
dua e is s ill in he college own whe e she esided i e yea s p io and ze o o he wise.
We hen use linea eg ession o es ima e di e ences be ween s aye s and lea e s o
se e al employmen ou comes con olling o di e ences in indi idual cha ac e is ics,
Xi.e., we es ima e eg essions o he ollowing o m:
Y¼θS aye þXβþεð1Þ
The indi idual cha ac e is ics include sepa a e dummy a iables o he highes
deg ee ea ned, age, whe he he indi idual is emale, Black, Hispanic, Asian, O he
Nonwhi e, ma ied, o has child en. One un o una e limi a ion o he cu en pape is
ha he e may be unobse able di e ences be ween s aye s and lea e s ha also sepa-
a ely a ec employmen ou comes
2
. Fo example, lea e s may be mo e ambi ious and
Win e s IZA Jou nal o Mig a ion 2012, 1:11 Page 4 o 17
h p://www.izajom.com/con en /1/1/11
ca ee d i en han s aye s. I d i e and ambi ion also imp o e employmen p ospec s,
hen ailing o obse e hese will induce a nega i e co ela ion be ween employmen
ou comes and s aying. Se e al p e ious s udies ha e a emp ed o accoun o whe he
mig an s a e gene ally a sel -selec ed g oup (e.g., Axelsson and Wes e lund 1998;
Yankow 2003; Boheim and Taylo 2007; Blackbu n 2010). This usually in ol es ying
o iden i y a a iable ha a ec s he p obabili y o mig a ion bu does no sepa a ely
a ec employmen ou comes such as p oximi y o amily o housing enu e in he p e-
ious pe iod. The census da a used in his s udy do no o e in o ma ion on ei he o
hese and we a e hus unable o examine whe he mig a ion is endogenous o ecen
college g adua es who ea ned hei deg ee in a college own
3
. Ins ead, we examine he
sensi i i y o he esul s o employing a di e ence in di e ences (DD) echnique ha
ea s ecen g adua es om la ge me opoli an a eas as a con ol g oup. Mo e de ails
a e discussed la e 4.
This pape conside s i e employmen ou comes: 1) whe he an indi idual is
employed a he ime o he census, 2) he log o annual wage income ea ned in he
p e ious yea , 3) he log a e age hou ly wage in he p e ious yea compu ed as annual
wages di ided by he numbe o hou s wo ked in he p e ious yea , 4) he log o hou s
wo ked in he p e ious yea , and 5) he pe cen age o wo ke s in he same occupa ion
wi h a bachelo ’s deg ee o highe o employed ecen g adua es. This i h ou come is
in ended o cap u e whe he s aye s a e mo e likely o ake jobs o which hey a e
o e educa ed. The hypo hesis o his pape is ha college own s aye s may expe ience
wo se employmen ou comes han hose who lea e along a numbe o dimensions.
Summa y s a is ics o he dependen and independen a iables o he ull sample a e
epo ed in Table 2. A ew hings a e wo h men ioning. Fi s , 92.3 pe cen o he
sample is employed, e lec ing he high a es o labo o ce pa icipa ion and low a es
o unemploymen o young college g adua es as well as he ela i ely s ong labo ma -
ke in 2000. Second, only 19.4 pe cen o ecen g adua es who li ed in a college own
i e yea s p io a e s ill in he college own whe e hey comple ed hei deg ee. The
la ge majo i y, he e o e, ha e ei he mo ed back o a p e ious loca ion o mo ed on o
a new loca ion.
Table 3 p esen s he mean alues o he dependen and independen a iables sepa-
a ely o college own s aye s and lea e s and also epo s he di e ences in he means
and whe he he di e ences a e s a is ically signi ican ly di e en om ze o. This p o-
ides a i s look a a numbe o di e ences be ween hose who s ay in a college own
a e comple ing a deg ee and hose who lea e. Acco ding o hese aw di e ences,
s aye s pe o m signi ican ly wo se han lea e s in all i e o he employmen ou comes
ha we s udy. S aye s ha e lowe employmen a es, ea n lowe annual and hou ly
wages, wo k ewe annual hou s, and wo k in less educa ed occupa ions. Howe e , he
able also shows ha he e a e impo an di e ences be ween s aye s and lea e s in
indi idual cha ac e is ics. S aye s a e signi ican ly less likely o hold a mas e ’s o p o-
essional deg ee, and he e a e some di e ences in age be ween he wo g oups.
Females a e signi ican ly mo e likely han males o s ay in he college own whe e hey
comple ed hei deg ee. Blacks a e signi ican ly mo e likely o s ay han Whi es, bu
Asians a e signi ican ly less likely o s ay han Whi es. S aye s a e also signi ican ly
mo e likely o be ma ied and mo e likely o ha e child en. Being ma ied and ha ing
child en, he e o e, appea o inc ease he cos s o lea ing and ie ecen g adua es o
Win e s IZA Jou nal o Mig a ion 2012, 1:11 Page 5 o 17
h p://www.izajom.com/con en /1/1/11
he a ea whe e hey comple ed hei deg ee. Because o hese signi ican di e ences in
indi idual cha ac e is ics, we canno accu a ely assess he e ec o s aying on employ-
men ou comes simply by aking di e ences in means. Ins ead, we need o use mul i-
a ia e eg ession echniques o con ol o he indi idual cha ac e is ics. The nex
sec ion p esen s he esul s o eg essing he i e employmen ou comes on he a iable
S aye and he indi idual cha ac e is ics.
3. Empi ical esul s
3.1 Di e ences be ween college own s aye s and lea e s
Reg ession esul s o di e ences be ween college own s aye s and lea e s a e
p esen ed in Table 4. The i s column epo s esul s o he p obabili y o employ-
men es ima ed ia a linea p obabili y model. The esul s a e quali a i ely obus o
es ima ing p obi and logi models. The es ima ed coe icien o S aye is small and
no s a is ically di e en om ze o. In o he wo ds, he esul s sugges ha he e is no
di e ence in he p obabili y o employmen o s aye s and lea e s. Recen college g a-
dua es om college owns ha e an equal p obabili y o inding employmen ega dless
o whe he hey s ay in he a ea whe e hey ea ned hei deg ee o mo e o a di e en
a ea. No e ha his is in con as o wha was obse ed in Table 3, con i ming he
impo ance o using mul i a ia e eg ession echniques o con ol o indi idual cha ac-
e is ics. Only a ew o he addi ional a iables ha e s a is ically signi ican e ec s on
Table 2 Summa y s a is ics
Va iable Mean S d. De . Min Max
Employed 0.923 0.267 0 1
Log Annual Wage 10.108 0.760 2.303 12.757
Log Hou ly Wage 2.622 0.556 −3.219 7.058
Log Annual Hou s 7.487 0.558 2.079 8.546
% o Occupa ion wi h Bachelo 's o Highe 0.504 0.278 0.014 1
S aye 0.194 0.396 0 1
Mas e 's Deg ee 0.099 0.299 0 1
P o essional Deg ee 0.032 0.177 0 1
Doc o al Deg ee 0.004 0.064 0 1
Age 24 0.176 0.381 0 1
Age 25 0.236 0.425 0 1
Age 26 0.246 0.430 0 1
Age 27 0.237 0.425 0 1
Female 0.553 0.497 0 1
Black 0.042 0.201 0 1
Asian 0.015 0.123 0 1
Hispanic 0.031 0.173 0 1
O he Nonwhi e 0.013 0.114 0 1
Ma ied 0.395 0.489 0 1
Has Child en 0.141 0.348 0 1
No es: The sample includes 11,237 indi iduals who esided in one o 41 college owns in 1995, who we e no longe in
school and had ea ned a leas a bachelo 's deg ee by 2000, we e be ween he ages o 23 and 27 in 2000, and we e
bo n in he U.S.
Win e s IZA Jou nal o Mig a ion 2012, 1:11 Page 6 o 17
h p://www.izajom.com/con en /1/1/11
he p obabili y o employmen o ecen college g adua es who li ed in a college own
i e yea s p io . Pe sons who ha e ea ned a mas e ’s deg ee a e signi ican ly mo e likely
o be employed han hose who only ha e a bachelo ’s deg ee wi h an es ima ed coe i-
cien o 0.019. P o essional deg ees and doc o al deg ees, howe e , do no ha e a s a is-
ically signi ican e ec on employmen . Simila ly, age does no ha e a signi ican e ec
on he p obabili y o employmen o ecen g adua es. Females a e signi ican ly less
likely o be employed han males wi h a coe icien o −0.034, bu he e a e no signi i-
can di e ences based on ace/e hnici y. Ma iage has no signi ican e ec on employ-
men , bu ha ing child en signi ican ly educes he p obabili y o employmen wi h a
ela i ely la ge coe icien o −0.172.
The second column o Table 4 epo s esul s o he log o annual income om
wages and sala ies. Fi s no e ha his eg ession and he ollowing eg essions ha e
ewe obse a ions han in he i s column because some indi iduals did no wo k in
he p e ious calenda yea . Taking logs o ces us o d op hose wi h ze o wages and
hou s because he log o ze o is no de ined. Tu ning o he esul s, S aye has a s a is-
ically signi ican nega i e e ec wi h a coe icien o −0.267. This con i ms ha ecen
g adua es who s ay in he college own whe e hey comple ed hei deg ee on a e age
accep much lowe paying jobs han hose who lea e. Con e ing logs o pe cen ages,
he coe icien es ima e sugges s ha college own s aye s ea n oughly 30 pe cen
lowe wages han hose who lea e. This is qui e a p ice o pay o s aying and helps
explain why so many ecen g adua es lea e college owns a e comple ing hei
deg ees. Se e al addi ional a iables also ha e impo an e ec s on log annual wages.
In e es ingly, holde s o p o essional deg ees ac ually ea n signi ican ly lowe annual
Table 3 Di e ences in means o s aye s and lea e s
Va iable S aye Mean Lea e Mean Di e ence
Employed 0.911 0.925 −0.014**
Log Annual Wage 9.874 10.164 −0.290***
Log Hou ly Wage 2.458 2.660 −0.202***
Log Annual Hou s 7.416 7.503 −0.087***
% o Occupa ion wi h Bachelo 's o Highe 0.481 0.544 −0.063***
Mas e 's Deg ee 0.069 0.106 −0.037***
P o essional Deg ee 0.026 0.034 −0.008**
Doc o al Deg ee 0.003 0.004 −0.002
Age 24 0.175 0.177 −0.002
Age 25 0.214 0.241 −0.027***
Age 26 0.222 0.251 −0.030**
Age 27 0.264 0.230 0.034***
Female 0.584 0.545 0.039**
Black 0.058 0.038 0.019***
Asian 0.008 0.017 −0.009***
Hispanic 0.030 0.031 −0.002
O he Nonwhi e 0.013 0.013 −0.001
Ma ied 0.422 0.389 0.032*
Has Child en 0.216 0.123 0.094***
No e: *Signi ican a 10%; **Signi ican a 5%; ***Signi ican a 1%.
Win e s IZA Jou nal o Mig a ion 2012, 1:11 Page 7 o 17
h p://www.izajom.com/con en /1/1/11
wages han hose wi h only a bachelo ’s deg ee wi h a coe icien o −0.132. Howe e ,
only 3.2 pe cen o he sample has a p o essional deg ee and hey a e likely o ha e
ea ned ha deg ee e y ecen ly and ha e e y li le wo k expe ience in hei new job.
Simila ly, mas e ’s deg ees and doc o al deg ees do no signi ican ly a ec annual wages
o his g oup. Age has a e y s ong e ec on annual wages wi h olde wo ke s ea ning
signi ican ly mo e. Females ea n signi ican ly lowe annual wages han males wi h a
coe icien o −0.208, and Blacks ea n signi ican ly lowe annual wages han Whi es
wi h a coe icien o −0.082. Howe e , Asians, Hispanics and O he Nonwhi es do no
ea n signi ican ly di e en annual wages han Whi es. Ma ied pe sons ea n
Table 4 Reg ession esul s o employmen ou comes
P obabili y o
Employmen
Log
Annual
Wage
Log
Hou ly
Wage
Log
Annual
Hou s
% o Occupa ion wi h
Bachelo 's o Highe
S aye 0.001 −0.267*** −0.209*** −0.058*** −0.052***
(0.007) (0.020) (0.014) (0.015) (0.008)
Mas e 's Deg ee 0.019** −0.023 0.084*** −0.106*** 0.157***
(0.008) (0.026) (0.020) (0.019) (0.009)
P o essional Deg ee 0.011 −0.132** 0.044 −0.176*** 0.364***
(0.013) (0.053) (0.041) (0.041) (0.016)
Doc o al Deg ee 0.038 0.065 0.199* −0.134 0.258***
(0.029) (0.103) (0.104) (0.096) (0.048)
Age 24 −0.003 0.190*** 0.007 0.184*** 0.002
(0.010) (0.032) (0.024) (0.026) (0.011)
Age 25 −0.001 0.401*** 0.108*** 0.294*** 0.015
(0.009) (0.031) (0.022) (0.025) (0.010)
Age 26 0.003 0.479*** 0.155*** 0.324*** 0.002
(0.010) (0.031) (0.023) (0.025) (0.010)
Age 27 0.011 0.560*** 0.175*** 0.386*** −0.007
(0.010) (0.031) (0.023) (0.025) (0.011)
Female −0.034*** −0.208*** −0.115*** −0.092*** 0.015***
(0.005) (0.015) (0.011) (0.011) (0.006)
Black −0.016 −0.082* −0.001 −0.082** −0.022
(0.017) (0.042) (0.030) (0.033) (0.015)
Asian −0.008 0.120 0.158** −0.038 0.028
(0.022) (0.073) (0.065) (0.057) (0.021)
Hispanic −0.019 −0.012 0.052 −0.064* −0.004
(0.019) (0.043) (0.034) (0.036) (0.016)
O he Nonwhi e −0.032 −0.062 0.025 −0.087* −0.065***
(0.023) (0.065) (0.054) (0.053) (0.024)
Ma ied 0.006 0.080*** 0.053*** 0.027** 0.030***
(0.005) (0.015) (0.012) (0.011) (0.006)
Has Child en −0.172*** −0.263*** −0.031* −0.232*** −0.015
(0.012) (0.030) (0.018) (0.025) (0.010)
Obse a ions 11237 10700 10700 10700 10368
No e: He e oskedas ici y obus s anda d e o s in pa en heses.
*Signi ican a 10%; **Signi ican a 5%; ***Signi ican a 1%.
Win e s IZA Jou nal o Mig a ion 2012, 1:11 Page 8 o 17
h p://www.izajom.com/con en /1/1/11
coe icien o −0.025, bu con olling o he cos o li ing does no meaning ully a ec
he main esul s o he p obabili y o employmen . CollegeTown is again signi ican
wi h a coe icien o 0.011, bu S aye and he in e ac ion e m a e again s a is ically in-
signi ican . The esul o he in e ac ion e m again sugges s ha s aying in a college
own a e g adua ing does no a ec he p obabili y o employmen .
The esul s o con olling o he cos o li ing in he log annual wage DD eg ession
a e epo ed in he second column o Table 7. As expec ed, Mean log en has a signi i-
can ly posi i e e ec on log annual wages wi h a coe icien o 0.389. This esul sug-
ges s ha a one pe cen inc ease in housing en s esul s in a oughly 0.4 pe cen
inc ease in annual wages o compensa e wo ke s o he highe cos o li ing. Con ol-
ling o he cos o li ing also changes he e ec o S aye on log annual wages. S aye
had a signi ican ly posi i e e ec on log annual wages in Table 6, bu a e con olling
o he cos o li ing he e ec is signi ican ly nega i e wi h a coe icien o −0.018. This
sugges s ha la ge me o s aye s ea n sligh ly lowe annual wages han la ge me o lea-
e s a e con olling o he highe cos o li ing in la ge me o a eas. Con olling o
he cos o li ing, p io esidence in a college own has a small and insigni ican e ec
on annual wages. The in e ac ion o S aye and CollegeTown again has a signi ican ly
nega i e e ec on log annual wages, bu he coe icien o −0.203 is somewha smalle
han he co esponding es ima e in Table 6. The implica ion is ha a e con olling o
he cos o li ing, s aying in a college own educes annual wages o ecen g adua es
by oughly 22.5 pe cen .
The hi d and ou h columns o Table 7 p esen he DD esul s o log hou ly wages
and log annual hou s wo ked con olling o he cos o li ing. Mean log en has a sig-
ni ican ly posi i e e ec on log hou ly wages wi h a coe icien o 0.416 bu an insigni i-
can e ec on log annual hou s, sugges ing ha o ecen g adua es he cos o li ing
a ec s hou ly wages bu no hou s wo ked. S aye again has a signi ican ly posi i e
e ec on log hou ly wages and a signi ican ly nega i e e ec on log annual hou s wi h
coe icien s o 0.018 and −0.036, espec i ely. CollegeTown now has an insigni ican
e ec on bo h log hou ly wages and log annual hou s. The in e ac ion o S aye and
CollegeTown has a s a is ically nega i e e ec on log hou ly wages wi h a coe icien o
−0.175 and an insigni ican e ec on log annual hou s.
The i h column o Table 7 p esen s he esul s o he pe cen o wo ke s in an
occupa ion wi h a bachelo ’s deg ee o highe con olling o he cos o li ing. As seen,
Mean log en has a s a is ically insigni ican e ec on he occupa ional educa ion
measu e and con olling o he cos o li ing has no meaning ul e ec on he main
esul s in he i h column. S aye and CollegeTown again ha e signi ican ly posi i e
e ec s wi h coe icien s o 0.014 and 0.015, espec i ely. The in e ac ion o he wo
again has a signi ican ly nega i e e ec wi h a coe icien o −0.067, sugges ing ha
s aying in a college own o ces ecen g adua es o accep jobs in less educa ed
occupa ions.
4. Conclusion
Indi iduals choose o li e and wo k in he loca ion ha gi es hem he highes possible
u ili y. Many ecen college g adua es p e e o s ay in he a ea whe e hey comple ed
hei deg ee because hey ha e de eloped loca ion-speci ic human capi al ha inc eases
Win e s IZA Jou nal o Mig a ion 2012, 1:11 Page 15 o 17
h p://www.izajom.com/con en /1/1/11
he a ea’s desi abili y. Howe e , s aying in he a ea whe e one a ended college may
wo sen employmen p ospec s, especially o pe sons a ending college in ela i ely
small college owns. Using decennial census mic oda a o he U.S., his pape inds
ha ecen g adua es who s ay in he college owns whe e hey ea ned hei deg ees
pay a conside able p ice o doing so. They ea n lowe annual and hou ly wages and
accep jobs in less educa ed occupa ions. Fu he mo e, he magni udes o hese e ec s
a e o en qui e la ge. The nominal di e ence in annual wages be ween college own
s aye s and lea e s is oughly 30 pe cen . E en a e con olling o he cos o li ing
and ea ing ecen g adua es om la ge me o a eas as a con ol g oup in an a emp
o accoun o selec ion e ec s, s aying in a college own s ill educes annual wages by
oughly 22.5 pe cen . Some ecen g adua es s ay in he a ea because o local a ach-
men s ha hey ha e de eloped, bu doing so o en equi es hem o ake lowe paying
jobs han hey could ind elsewhe e.
Endno es
1
Resea che s ha e sugges ed ha socie y u he bene i s because college educa ed
wo ke s inc ease p oduc i i y, wages, and employmen p ospec s o o he wo ke s
(Rauch 1993; Mo e i 2004, Win e s 2012). The ex e nal bene i s o human capi al also
a ac addi ional wo ke s and cause highly educa ed me opoli an a eas in he U.S. o
expe ience g ea e in-mig a ion and popula ion g ow h (Glaese , Scheinkman and
Shlei e 1995; Simon and Na dinelli 2002; Win e s 2011).
2
An addi ional ca ea is ha ad anced educa ion, ma i al s a us and ha ing child en
may be simul aneously de e mined wi h employmen ou comes. Howe e , emo ing
hese a iables om he analysis does no change he quali a i e esul s p esen ed
below.
3
E en i his in o ma ion we e a ailable, i would be unlikely o ully alle ia e endo-
genei y conce ns since ew young college s uden s a e homeowne s and p oximi y o
amily while in college may i sel be co ela ed wi h unobse ed cha ac e is ics ha
a ec employmen p ospec s.
4
An addi ional issue sugges ed by p e ious li e a u e is ha employmen di e ences
be ween mo e s and s aye s may be a ec ed by he amoun o ime since he mo e.
Un o una ely, we do no know exac ly when people mo e and a e unable o o e e i-
dence on how he e ec s e ol e o e ime.
5
G oss en s a e measu ed o include ce ain u ili ies paymen s. Fo hese pu poses,
a eas a e de ined as me opoli an a eas whe e hey exis and s a e non-me opoli an
a eas o a eas no pa o a me o a ea.
Compe ing in e es s
The IZA Jou nal o Mig a ion is commi ed o he IZA Guiding P inciples o Resea ch In eg i y. The au ho decla es ha
he has obse ed hese p inciples.
Acknowledgemen s
The au ho hanks anonymous e e ees and Edi o Amelie Cons an o help ul commen s.
Responsible edi o : Amelie Cons an
Recei ed: 5 July 2012 Accep ed: 5 Decembe 2012
Published: 28 Decembe 2012
Re e ences
Anil B, Sjoquis DL, Wallace S (2010) The e ec o a p og am-based housing mo e on employmen : HOPE VI in A lan a.
Sou h Econ J 77:138–160
Win e s IZA Jou nal o Mig a ion 2012, 1:11 Page 16 o 17
h p://www.izajom.com/con en /1/1/11
Axelsson R, Wes e lund O (1998) A panel s udy o mig a ion, sel -selec ion and household eal income. J Popul Econ
11:113–126
Be y CR, Glaese EL (2005) The di e gence o human capi al le els ac oss ci ies. Pap Reg Sci 84:407–444
Blackbu n ML (2010a) The impac o in e nal mig a ion on ma ied couples’ea nings in B i ain. Economica 77:584–603
Blackbu n ML (2010b) In e nal mig a ion and he ea nings o ma ied couples in he Uni ed S a es. J Econ Geog 10:87–111
Böheim R, Taylo MP (2002) F om he da k end o he s ee o he b igh side o he oad? he wage e u ns o
mig a ion in B i ain. Labou Econ 14:99–117
Chen Y, Rosen hal SS (2008) Local ameni ies and li e-cycle mig a ion: do people mo e o jobs o un? J U ban Econ
64:519–537
Cla k DE, Hun e WJ (1992) he impac o economic oppo uni y, ameni ies, and iscal ac o s on age-speci ic mig a ion.
J Reg Sci 32:349–365
DaVanzo J (1983) Repea mig a ion in he Uni ed S a es: who mo es back and who mo es on? Re Econ S a 65:552–559
F anklin RS (2003) Mig a ion o he young, single and college educa ed: 1995 o 2000 Census 2000 special epo s,
CENSR-12 Washing on. US Go e nmen P in ing O ice, DC
Glaese EL, Scheinkman JA, Shlei e A (1995) Economic g ow h in a c oss-sec ion o ci ies. J Mone Econ 36:117–143
G oen JA (2004) The e ec o college loca ion on mig a ion o college-educa ed labo . J Econom 121:125–142
Hu man D, Quigley JM (2002) The ole o he uni e si y in a ac ing high ech en ep eneu ship: a Silicon Valley ale.
Ann Reg Sci 36:403–419
K upka DJ (2009) Loca ion-speci ic human capi al, loca ion choice and ameni y demand. J Reg Sci 49:833–854
Mo e i E (2004) Es ima ing he social e u n o highe educa ion: e idence om longi udinal and epea ed c oss-
sec ional da a. J Econom 121:175–212
Oos e beek H, Webbink D (2011) Does s udying ab oad induce a b ain d ain? Economica 78:347–366
Psacha opoulos G, Pa inos HA (2004) Re u ns o in es men in educa ion: a u he upda e. Educ Econ 12:111–134
Rauch JE (1993) P oduc i i y gains om geog aphic concen a ion o human capi al: e idence om he ci ies. J U ban
Econ 34:380–400
Ruggles S, Sobek M, Alexande T, Fi ch CA, Goeken R, Hall PK, King M, Ronnande C (2008) In eg a ed Public Use
Mic oda a Se ies. Ve sion 4.0. Minneapolis, Minneso a Popula ion Cen e
Simon CJ, Na dinelli C (2002) Human capi al and he ise o Ame ican ci ies, 1900–1990. Reg Sci U ban Econ 32:59–96
Sjaas ad LA (1962) The cos s and e u ns o human mig a ion. J Poli Econ 70:80–93
Waldo BS (2009) Is human capi al accumula ion a sel -p opelling p ocess? compa ing educa ional a ainmen le els o
mo e s and s aye s. Ann Reg Sci 43:323–344
Whisle RL, Waldo BS, Mulligan GF, Plane DA (2008) Quali y o li e and he mig a ion o he college-educa ed: a li e-
cou se app oach. G ow h Chang 39:58–94
Win e s JV (2009) Wages and p ices: a e wo ke s ully compensa ed o cos o li ing di e ences? Reg Sci U ban Econ
39:632–643
Win e s JV (2011) Why a e sma ci ies g owing? who mo es and who s ays. J Reg Sci 51:253–270
Win e s JV (2012) Human capi al ex e nali ies and employmen di e ences ac oss me opoli an a eas o he US. J Econ
Geog Fo hcoming
Yankow JJ (2003) Mig a ion, job change, and wage g ow h: a new pe spec i e on he pecunia y e u n o geog aphic
mobili y. J Reg Sci 43:483–516
doi:10.1186/2193-9039-1-11
Ci e his a icle as: Win e s: Di e ences in employmen ou comes o college own s aye s and lea e s. IZA
Jou nal o Mig a ion 2012 1:11.
Submi you manusc ip o a
jou nal and benefi om:
7 Con enien online submission
7 Rigo ous pee e iew
7 Immedia e publica ion on accep ance
7 Open access: a icles eely a ailable online
7 High isibili y wi hin he fi eld
7 Re aining he copy igh o you a icle
Submi you nex manusc ip a 7 sp inge open.com
Win e s IZA Jou nal o Mig a ion 2012, 1:11 Page 17 o 17
h p://www.izajom.com/con en /1/1/11
