Compulsory*Military*Service*and*Sibling*Externalities*on* …of*Contents*! Introduction* 1*...
Transcript of Compulsory*Military*Service*and*Sibling*Externalities*on* …of*Contents*! Introduction* 1*...
!
!
!
Compulsory*Military*Service*and*Sibling*Externalities*on*
the*High*School*Dropout*Rate*
Evidence(from(Switzerland(!
Husam!El)Tarifi!
!
Bachelor(of(Economics!in!Economics!and!Finance!
!Declaration*
*
I!declare!that!this!thesis!has!not!been!submitted!anywhere!for!award,!is!my!own!
work!and,!to!the!best!of!my!knowledge,!holds!no!content!that!has!been!published!
by!or!is!the!intellectual!right!of!another,!except!where!due!acknowledgement!is!
made.!!
!
*
*
*
*
Date:! 22/10/2012! Signature:! _______________________!
!
Acknowledgments*!Of(my(family(
!
I!would!like!to!thank!my!father!for!advising!me!to!study!economics,!choosing!this!
major!been!one!of!the!best!decisions!I’ve!made.!Thanks!to!my!mother!for!being!the!
most!amazing!person!I!know,!and!special!thanks!to!my!brother!for!the!sibling!
support!which!I!find!to!be!so!important!in!this!thesis.!
!
Of(the(Honours(Cohort(
!
Thanks!to!Ashna!“Ash”!Taneja!and!Axelle!“Axe”!Marjolin!for!putting!up!with!the!
nicknames.!!
!
Thanks!to!David!and!Marcus!for!being!two!of!the!most!brilliant!people!I!know!and!
thanks!to!Will,!Carlos!and!Cassie!for!being!just!plain!awesome.!
!
Thanks!to!Rakan!and!Suchita!for!the!constant!stream!of!econo)memes.!Thanks!to!
Adam!and!Peter!for!being!constantly!hilarious.!
!
Thanks!to!Richie!and!Jen!for!being!so!inimitable,!Jason!for!being!the!man,!and!James!
for!being!the!nicest!guy!I!know.!
!
Of(the(Faculty(
!
Thanks!to!Denise,!your!econometrics!class!was!one!of!the!best!I’ve!ever!taken;!it!
ultimately!led!me!to!choose!applied!work!for!my!thesis.!!
!
Special!thanks!goes!to!Stanley!for!putting!up!with!my!shenanigans!and!to!Valentyn!
for!giving!me!that!little!bit!of!extra!time!to!put!all!of!this!together.!!
!
Lastly,!I!am!forever!grateful!to!my!supervisor,!Marian,!for!all!the!assistance!with!this!
thesis,!you!have!been!invaluable.! !
!
!
*
*
*
*
*
*
Abstract*
!!
This!paper!examines!the!effect!of!an!individual!being!conscripted!on!their!siblings’!
propensity!to!dropout!from!high!school.!!
!
Using!a!Swiss!panel!dataset,!this!thesis!merges!the!two!literatures!of!‘sibling!
externalities’!and!‘compulsory!military!service!effects’!to!build!on!both.!It!establishes!
a!clearer!causality!relationship!in!both!fields!by!observing!sibling!level!effects!as!
opposed!to!broader!peer!effects.!
!
This!paper!finds!negative!effects!on!the!high!school!completion!rates!of!the!female!
siblings!of!conscripts,!while!male!siblings!are!unaffected!overall.!Examining!this!in!
detail!shows!heterogeneous!effects;!females!in!less!wealthy!households!experience!
much!stronger!and!more!damaging!effects!while!their!wealthier!counterparts!
experience!no!discernable!effect.!These!effects!are!found!to!be!largely!due!to!
reductions!in!peer!support,!which!result!from!older!brothers!leaving!the!household!
when!conscripted.!
* *
Table*of*Contents*
!
Introduction* 1*
Literature*Review* 5*
Compulsory(Military(Service(and(Labour(Outcomes( 5(
Compulsory(Military(Service(and(Education(Externalities( 7(
Siblings(and(Education(Externalities( 8(
Institutional*Setting* 11*
Education( 11(
Conscription( 11(
Causal(Mechanisms( 12(
Data* 14*
SHP(Dataset( 14(
Sample(Construction( 14(
SLFS(dataset( 15(
Summary(Statistics( 16(
Empirical*Strategy* 18*
Causality( 18(
Fixed(Effects(Regressions( 18(
2S2SLS!Generation!Procedure( 19(
The(Instrument( 21(
Results* 22*
Robustness*Checks* 32*
Falsification( 32(
Individual(V.S.(Household(Fixed(Effects( 33(
Panel(Attrition( 35(
Sibling(Support(Robustness( 37(
Conclusion* 39*
Summary(of(Results( 39(
Implications( 39(
Limitations(and(Future(Research( 40(
Appendix* 42*
!
List*of*Tables*
! !
Table!1:!Summary!of!the!Literature! 6!
Table!2:!SHP!Summary!Statistics! 17!
Table!3:!Females!aged!20!and!above! 23!
Table!4:!Males!aged!20!and!above! 25!
Table!5:!Females!aged!15!to!19! 27!
Table!6:!Family!incomes!for!females!aged!20!and!above! 28!
Table!7:!Family!incomes!for!males!aged!20!and!above! 29!
Table!8:!Emotional!support!for!females!aged!18!and!above!in!low/middle)income!
families! 30!
Table!9:!Females!aged!20!and!above!in!different!households! 33!
Table!10:!Individual!V.S.!household!level!fixed!effects! 34!
Table!11:!Females!aged!20!and!above,!1999!households! 36!
Table!12:!Individuals!aged!18!and!above,!by!primary!language! 38!
!
! 1!
Introduction!
!
Every! year,! over!20,000!young!men! called!up! for!mandatory!military! service! in!
Switzerland.! ! It! has!been!well! established! that! compulsory!military! service! (CMS)!
tends! to! have! negative! effects! on! those! conscripted! in! terms! of! their! labour! and!
educational! outcomes.! Little! is! known! however! about! the! externalities! associated!
with!CMS.!
!
Today,! compulsory! military! service! is! present! in! many! OECD! countries;! these!
include! Denmark,! Austria,! Finland,! Greece,! Israel,! South! Korea,! Singapore,! Turkey!
and!Switzerland.!!
!
Many! countries! have! recently! begun! examining! suspending! CMS! laws,! which! are!
often! legacies! of! the! first! and! second! World! Wars.! Germany! and! Sweden,! for!
example,! have! ended! CMS! in! 2011! and! 2010,! respectively.! In! Switzerland,! a!
referendum! is! likely! to! occur! in! the! near! future! to! decide!whether! to!maintain! or!
repeal!CMS!laws.!
!
In!this!thesis,!I!build!on!a!very!new!externalities!literature!by!examining!the!effects!
associated!with!male!CMS!on!the!siblings!of!conscripts.!
!
There!are!many!studies!examining!the!effects!of!CMS!on!the!conscripts!themselves!
(see! Table! 1! and! Appendix! 1A).! These! have! been! conducted! on! a! vast! range! of!
countries,! and! while! some! effects! are! country! specific,! most! papers! find! negative!
effects!on!education!and!future!earnings.!
!
Some!of!these!papers!employ!women!as!a!control!group!with!the!assumption!that,!if!
they!are!exempt! from!service! (true!everywhere! in! the!OECD!with! the!exception!of!
Israel),! they! cannot! be! affected!by! it.!However,! if! there! exist! any! externalities,! the!
findings!in!these!papers!may!be!misestimated.!
!
As!far!as!I!know!there!is!only!one!other!study!examining!the!externalities!associated!
with! CMS! as! a! central! question;! Cipollone! and! Rosolia! (2007)! find! negative! peers!
effects!at!the!town!level!as!a!result!of!CMS!in!Italy.!A!number!of!studies!(Buonanno!
! 2!
2007;!Maurin!and!Xenogiani!2007;!Grenet,!Hart!et!al.!2011)!only!examine!the!effects!
of!peer!service!as!a!robustness!check1,!they!all!find!no!educational!or!earning!based!
externalities!from!the!repeal!CMS!policies!in!the!U.K.!and!in!France.!!
!
It! is! plausible! that! country! level! differences! (such! as! CMS! enforcement! or! more!
general!cultural!factors)!account!for!most!of!these!differences!in!results.!2!I!employ!a!
more!micro! level!approach!by!observing!household! level!effects! (specifically! those!
relating!to!siblings).! !This!allows!the!investigation!of!clearer!causality!relationships!
as!opposed!to!broad!peer!effects.!Many!of!these!relationships!are!robust!to!countryW!
level! heterogeneity! and! therefore! hint! at! likely! outcomes! in! countries! outside! of!
Switzerland.!!!
!
In!this!thesis,!I!also!attempt!to!study!sibling!composition!and!sibling!presence!effects!
by!estimating!the!effect!of!a!male!sibling!leaving!home.!I!do!this!by!employing!CMS!
as!an!exogenous!shock!to!male!sibling!presence.!
!
These! relationships! are! important! to! examine,! as! any! peer! effects/externalities!
present!involve!much!of!society,!possibly!inform!social!policy,!and,!as!suggested!by!
Butcher!and!Case!(1994),! !may!provide!a!practical! instrument!for!use! in!education!
related!analyses.!
!
The!siblings’!externalities! literature! finds!ambiguous!results,! especially! in! the!area!
concerning! sibling!gender! composition.!Butcher!and!Case! (1994)! find! that! females!
with!all!male!siblings!attain!higher!levels!of!education!than!their!counterparts!with!a!
least!one! sister.!However,!Hauser!and!Kuo! (1998)!and!Kaestner! (1997),! in! similar!
studies,! find! no! effect! of! sibling! gender! composition! on! education.! I! examine! the!
presence! of! shortWterm! composition/presence! effects! through! CMS! and! argue! that!
the!causal!mechanisms!are!likely!to!be,!at!least!partially,!generalizable!to!the!longer!
term/overall!cases!examined!in!this!literature.!
!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!That!is,!their!focus!lies!on!estimating!the!effect!of!CMS!on!the!conscripts!themselves.!They!all!run!the!same!analysis!with!women!with!the!reasoning!that,!if!women!are!exempt!from!service!and!no!effect!is!found,!the!assumptions!and!methodologies!are!sound.!Clearly,!externalities/peer!effects!are!assumed!not!to!exist.!2!The!methodologies!employed!by!the!two!sets!of!papers!are!also!vastly!different.!Cipollone!&!Rosolia!(2007)!employ!a!very!convincing!instrumental!variable!approach,!while!the!papers!that!only!examine!women!as!a!robustness!check!all!employ!regression!discontinuity!design!techniques!to!assess!the!abolishment!of!CMS!laws.!
! 3!
There! are! a! number! of! causal! channels! through!which! the! peer! effects! of! CMS! on!
siblings!can!occur.!These!factors!often!act!in!opposite!directions;!as!such!the!overall!
effect!seems!ambiguous.! !On!the!one!hand,!males!completing!their!service!not!only!
leave!home!(resulting!in!a!relaxing!of!household!financial!constraints!as!spending!on!
them! is! reduced)! but! also! receive! an! income! from! the! Swiss! government! for! the!
length! of! their! service! (meaning! they! can! actually! contribute! to! the! household’s’!
income).!For!poorer! families,! it! is! therefore!possible! that! this! resultant! relaxing!of!
familial!financial!constraints!can!have!a!positive!effect!on!sibling!education!as!more!
money! is! spent! on! tutoring! or! other! forms! of! educational! support.! On! the! other!
hand,! when! siblings! are! away! for! relatively! long! periods! of! time,! there! results! a!
decrease!in!sibling!support!factors!such!as!homework!help1!and!emotional!support.!
This! support! channel! is! therefore! likely! to! act! in! the! opposite! direction! to! the!
financial!channel.!
!
In!this!thesis,! I!employ!individualWlevel! fixed!effects!regressions!with!a!twoWsample!
twoWstage! least!squares! instrumental!variable!approach! to!control! for!a!number!of!
endogeneity! and! selection! factors.! I! find! that,! overall;! there! are! strong! negative!
educational! externalities! for! females! associated! with! compulsory!military! service.!
Specifically,!I!find!that!on!average,!females!are!4%!more!likely!to!drop!out!when!an!
older!male!sibling!is!conscripted.!
!
Upon!dissecting!this!effect,!I!find!that!the!femaleWbased!effects!are!not!discernable!in!
wealthy!households2!but!are!very!clear!(and!stronger!than!the!overall!case)! in! less!
wealthy! families.! I! also! find! that! much! of! this! low/middle! class! effect! can! be!
explained! by! decreases! in! sibling! support! resulting! from! a! sibling! temporarily!
leaving! the! household! due! to! CMS.! This! effect! seems! to! outweigh! any! financial!
effects,!which!are!expected!to!act!in!the!opposite!direction!(i.e.,!result!in!a!reduction!
in!the!propensity!to!dropout).!
!
Unlike! females,! the! younger! male! siblings! of! conscripts! tend! to! experience! no!
discernable! overall! effect,! however,! when! restricting! the! analysis! to! less! wealthy!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!To!ensure!the!causal!direction!is!appropriate,!that!is,!a!sibling!being!conscripted!affects!an!individual’s!dropout!propensity!and!not!the!other!way!around,!I!only!examine!individuals!with!older!male!siblings.!This!means!the!homework!assistance!channel!is!much!more!promising.!2!Wealthy!households!are!defined!as!households!earning!more!than!150,000!SFr!per!year,!less!wealthy!households!are!those!that!earn!less.!
! 4!
households,! I! find!that! they!experience!a!decrease! in! their!dropout!propensity!as!a!
result!of!a!sibling!being!conscripted.!There!does!not!seem!to!be!any!evidence!of!the!
sibling!support!effect!for!males.!The!effect!for!lower/middleWclass!males!is!likely!the!
result!of!a!decrease!in!household!financial!constraints;!however,!this!is!not!explicitly!
tested!for.!
!
I!conduct!a!number!of!alternative!tests!to!ensure!the!robustness!of!the!results;!these!
include! a! falsification! exercise,! a! test! for! attritionWbased! effects,! and! a! robustness!
check!for!the!sibling!support! finding.!All!of!tests!support!the!findings!of!this!study.
! 5!
Literature-Review!
-
Compulsory.Military.Service.and.Labour.Outcomes.
-
A! large! number! of! studies! have! examined! the! direct! effects! of! CMS! on! labour!
outcomes;!these!largely!focus!on!wage!and!educational!attainment!differences.!!
!
As!in!the!brief!overview!provided!in!Table!1,!the!effects!of!CMS!vary!with!countryW
level! heterogeneity.! In!most! examined! countries,! conscripted! individuals! generally!
exhibit! lower! levels!of!education;! this! is!especially! true! for!countries!that!maintain!
CMS! deferment! options! for! those! enrolled! in! higher! education.! The! tendency! for!
draft!avoidance!behaviours! to!affect!schooling!achievement! is!most!notably!shown!
by!Angrist!and!Krueger!(1992),!where!the!authors!use!the!VietnamWera!draft!lottery!
as!an!instrument!to!measure!the!returns!to!schooling.!
!
Of!the!15!papers!examining!the!effects!of!CMS!on!employment!or!education!in!Table!
1,! only! 3! find! positive! outcomes! for! conscripts.! These! results! are! largely! due! to!
country!specific!properties!such!as!the!GI!Bill!for!the!U.S.A.!
!
The!majority!of!this!literature!is!very!recent!and!EuropeWcentric.!This!is!largely!due!
to!the!recent!trend!across!European!countries!to!end!CMS.!
!
A! few!of! the! studies! surveyed!attempt! to!measure! the! labour!outcomes!of!CMS!by!
employing!women!as!the!control!group.!These!include!Card!and!Lemieux!(2001),!Di!
Pietro!(2009)!and!Paloyo!(2010).!It!is!important!to!note!that!if!peer!externalities!are!
present,!these!results!may!be!misestimated.1!
!
Two!papers,!Card!and!Cardoso!(2011)!and!Di!Pietro!(2009)!find!heterogeneity!in!the!
effects!of!CMS.!Card!and!Cardoso!(2011)!find!those!with!lower!levels!of!education!in!
Portugal!benefit! from!CMS,!while! those!with!higher! levels!of! education!experience!
no!discernable!effect.!Di!Pietro!(2009)!in!an!Italian!analysis,!however,!finds!that!less!
advantaged! individuals! obtain! lower! levels! of! education! as! a! result! of! CMS! while!
advantaged!ones!obtain!higher!levels.!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!Whether!these!results!are!under!or!over!estimated!depends!on!the!direction!of!the!link!between!military!service!and!peer!outcomes.!A!positive!link!implies!systematic!underestimation.!
! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!This!p
apers!e
mploys!a!cro
ss0country!approach!
Table&1:&Summary&of&the&literature&
Papers&Examining&the&D
irect&Effects&of&Compulsory&M
ilitary&Service&(CMS)&on&Earnings&
Paper!
Country!
Deferment!due!
to!Higher!
Education!
Peacetim
e!
Effect!o
f!CMS!
Albrecht,!E
din!et!al.!(1
999)!
Sweden!
No!
Yes!
Conscrip
ts!earned!higher!w
ages!th
an!those!that!avoided!CMS!
Angrist!a
nd!Chen!(2011)!
U.S.A.!
Yes!
No!
Negligible!wage!diffe
rences!over!th
e!long!term!
Angrist!a
nd!Krueger!(1
992)!
U.S.A.!
Yes!
No!
Higher!education!deferments!re
sulted!in!higher!w
ages!fo
r!those!that!avoided!CMS!
Buonanno!(2007)!
U.K.!
Yes!
Yes!
Conscrip
ts!earned!less!th
an!exempt!co
horts!
Card!and!Cardoso!(2011)!
Portu
gal!
Yes!
Yes!
Improved!prospects!fo
r!conscrip
ts!with!only!prim
ary!education,!no!effect!fo
r!
conscrip
ts!with!higher!le
vels!o
f!education!
Galiani,!R
ossi!e
t!al.!(2
011)!
Argentina!
Yes!
Yes/No!
Conscrip
ts!showed!lower!la
bour!fo
rce!particip
ation!and!marginally!lower!w
age!
rates!in
!both!peacetim
e!and!wartim
e!
Grenet,!H
art!e
t!al.!(2
011)!!
U.K.!
Yes!
Yes!
No!effect!
Imbens!and!van!der!Klaauw!
(1995)!
Netherlands!
Yes!
Yes!
Conscrip
ts!earned!less!th
an!those!that!avoided!CMS!
Paloyo!(2010)!
Germany!
No!
Yes!
No!effect!
Papers&Examining&the&D
irect&Effects&of&Compulsory&M
ilitary&Service&(CMS)&on&Educational&Attainm
ent&
Paper!
Country!
Deferment!due!
to!Higher!
Education!
Peacetim
e!
Effect!o
f!CMS!
Angrist!a
nd!Chen!(2011)!
U.S.A.!
Yes!
No!
Higher!education!levels!fo
r!veterans!due!to!GI!Bill!
Bauer,!B
ender!et!al.!(2
012)!
West!G
ermany!
No!
No!
No!effect!
Buonanno!(2007)!
U.K.!
Yes!
Yes!
Higher!education!deferments!re
sulted!in!slig
htly!higher!education!levels!fo
r!those!
that!avoided!CMS!!
Card!and!Lemieux!(2001)!
U.S.A.!
Yes!
No!
Higher!education!deferments!re
sulted!in!higher!education!levels!fo
r!those!that!
avoided!CMS!
Cipollone!and!Rosolia!(2007)!
Italy!
Yes!
Yes!
Conscrip
ts!were!less!e
ducated!than!exempt!co
horts!
Di!Pietro!(2009)!
Italy!
Yes!
Yes!
CMS!abolitio
n!is!a
ssociated!with!higher!educational!outco
mes!fo
r!advantaged!
individuals!b
ut!lo
wer!outco
mes!fo
r!less!a
dvantaged!individuals!
Grenet,!H
art!e
t!al.!(2
011)!
U.K.!
Yes!
Yes!
Draft!a
voidance!behaviours!re
sulted!in!a!sm
all!b
ut!positiv
e!effect!o
n!education!
levels!
Keller,!P
outvaara!et!al.!(2
010)!
N/A1!
0!Yes!
Countrie
s!with!CMS!tended!to!have!lower!le
vels!o
f!higher!education!demand!
Maurin!and!Xenogiani!(2
007)!
France!
Yes!
Yes!
Draft!a
voidance!behaviours!re
sulted!in!a!positiv
e!effect!o
n!education!levels!
! 7!
The! findings! of! Card! and! Cardoso! (2011)! are! much! more! believable;! the! panel!
dataset! and! the! structural! methodology! they! employ! seem! to! control! for!
endogeneity! factors! much! more! effectively! than! the! non?panel! difference! in!
difference!approach!employed!by!Di!Pietro!(2009).!It!is!likely!that!not!controlling!for!
higher!education!deferment!is!confounding!Di!Pietro’s!results.!A!similar!bias!effect!is!
also!seen! in!Buonanno!(2007)!and! is!discussed! in! the!Compulsory!Military!Service!
and!Education!Externalities!subsection!below.!
!
I! build! on! this! heterogeneous! effect! analysis! in! this! thesis! by! examining! the!
externalities!in!a!wealth!context.!!
!
In!a!macro?level!cross?country!analysis,!Keller,!Poutvaara!et!al.!(2010)!find!that!CMS!
in! OECD! countries! is! associated! with! a! negative! effect! on! the! demand! for! higher!
education.!This!effect!is!shown!to!increase!in!intensity!and!significance!as!duration!
and!intensity!of!enforcement!increases.!!!
!
Compulsory*Military*Service*and*Education*Externalities!
!
The!direct!effects!literature’s!main!focus!is!on!males!eligible!to!be!conscripted.!There!
are!few!papers!examining!the!educational!externalities!associated!with!CMS.!As!far!
as!I!know,!there!is!only!one!published!paper!that!attempts!to!measure!peer!effects!as!
a!central!question;!Cipollone!and!Rosolia!(2007).!
!
Cipollone! and!Rosolia! (2007)! are! the! first! to! link! CMS! to! educational! peer! effects.!
The!authors!exploit!a!partial!CMS!exemption!in!the!1980s,!caused!by!an!earthquake!
in! Southern! Italy! as! a! very! convincing! instrument! to! isolate! the! effect! of! CMS! on!
schooling! attainment.! The! authors! find! that! the! exemption! raised! the! high?school!
graduation! rates!of! boys!by!over!2%.!When!examining! the!high?school! graduation!
rates!of!girls,!they!find!a!similar!positive!effect!(an!increase!of!1%)!despite!them!not!
being!required!(or!able)!to!serve!in!the!military.!
!
Buonanno!(2007)!and!Grenet,!Hart!et!al.!(2011)!both!examine!the!effects!of!ending!
CMS!in!the!United!Kingdom.!Buonanno!(2007)!finds!that!males!who!served!for!two!
years! in! the! National! Service! earned,! on! average,! 5%! less! than! their! exempt!
! 8!
counterparts.!He!also!finds!that!exempt!cohorts!accumulated,!on!average,!0.25!years!
of! additional! education.! Grenet,! Hart! et! al.! (2011)! builds! on! Buonanno! (2007)! to!
observe! no! long?term!wage! effects! that,!when! controlling! for! conscript! deferment!
(due!to!apprenticeships!or!university).!
!
Both!of! these!papers!test!the!robustness!of! their!results!by!estimating!the!effect!of!
repealing! CMS! laws! on! women! with! the! hypothesis! that! no! effects! should! be!
observed!due! to!women!being!exempt! from!service.!They! find!no!effect!on! female!
education!or!wages.!In!a!similar!analysis,!Maurin!and!Xenogiani!(2007)!also!find!no!
effect.!
!
The! differences! in! these! sets! of! results! may! be! due! to! differences! in! CMS! policy!
and/or!broad!country?level!differences!such!as!stronger!of!CMS!enforcement!in!Italy!
(Cipollone!and!Rosolia!2007).!
!
In! this! paper! I! employ! a! focus! on! sibling! peer! effects! to! build! on! the! literature!
examining!the!effects!of!CMS!on!peer!educational!outcomes.!I!focus!on!sibling!effects!
to!establish!a!clearer!causality!relationship!and!observe!a!local!effect!(as!opposed!to!
more!broad!peer!effects)!that!can!be!more!robust!to!country!level!heterogeneity.!
!
Siblings*and*Education*Externalities*
!
A! large! number! of! studies! examine! the! effects! of! externalities! on! educational!
outcomes.!The!siblings!externalities!sub?field! is! itself!vast!and!ranges! from!studies!
that!examine!the!balance!between!genetic!and!environmental!factors!in!determining!
educational!outcomes!(Behrman!and!Taubman!1989;!Plug!and!Vijverberg!2003)1!to!
studies! that! examine! the! relative! impact! of! siblings! against! that! of! friends! and!
classmates!(Duncan,!Boisjoly!et!al.!2001)2.!
!
The!subset!examining!gender!and!composition!based!factors!is!most!relevant!to!this!
paper.!!
!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!These!typically!find!an!approximately!even!split!between!genetic!factors!and!environmental!ones.!2!Sibling!and!family!based!factors!are!more!correlated!with!achievement!and!delinquency!measures!than!neighbour,!friend,!or!classmate!factors.!
! 9!
Black,!Devereux!et!al.! (2005)!examine! the!effects!of! family!size!and!birth!order!on!
children’s! educational! outcomes.! The! authors! find! negative! family! size! effects! on!
educational!outcomes.!However,!when!controlling!for!birth!order,!family!size!effects!
are! found! to! be! insignificant.! The! authors! also! find! that! a! higher! birth! order! is!
associated!with!lower!levels!or!education!and!earnings.!
!
Butcher!and!Case!(1994)!examine!the!effects!of!sibling!quantity!and!sex!composition!
on! educational! attainment! with! U.S.A.! data.! They! find! that,! when! controlling! for!
family! size,!men!were!unaffected!by! the! sex! composition!of! their! siblings.!Women!
raised! with! all! male! siblings,! however,! received! more! education! on! average! than!
women!with!at! least!one!sister.!Butcher!and!Case!(1994)!also!find!that!educational!
outcomes!tend!to!be!lower!for!individuals!with!larger!numbers!of!siblings.!
!
Kaestner! (1996)!and!Hauser!and!Kuo! (1998)!contest! the! results! found!by!Butcher!
and! Case! (1994).! Both! papers! find! no! evidence! of! general! correlations! between!
sibling!gender!and!educational!attainment.!Kaestner!(1997),!however,!does!find!that!
black!adults!who!have!sisters!tend!to!have!greater!levels!of!education.!!
!
A!number!of!papers!examine!these!effects!for!developing!countries.!Qureshi!(2012)!
finds! that! investing! in! older! sisters’! education! is! associated! with! positive!
externalities! for! the! education! of! younger!male! siblings.! The! author! suggests! that!
this!is!likely!due!to!the!childcare!responsibilities!associated!with!older!sisters.!!Garg!
and! Morduch! (1998)! find! that,! due! to! a! preference! for! males! and! constrained!
resources,!children!with!all! female!siblings!have!health! indicators!that!are!25?40%!
higher!than!for!children!with!all!male!siblings.!Morduch!(2000)!finds!variable!sibling!
composition! effects! by! country.1!Bommier! and! Lambert! (2004)! find! sibling! rivalry!
for!human!capital!investments!is!not!exclusively!the!result!of!constrained!resources.!
!
Fletcher,!Seltzer!et!al.!(2012)!examine!the!effect!of!sibling!death.!They!find!negative!
effects!on!educational!attainment,!and!show!that!sisters!are!more!affected!than!their!
brothers.!
!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!This!suggests!that!the!discrepancy!between!Butcher!and!case!(1994)!and!Kaestner!(1996)!and!Hauser!and!Kuo!(1998)!could!be,!at!least!partially,!due!to!country!level!effects.!
! 10!
Many!of!these!findings!suggest!sibling!support!factors!are!at!play,!but!none!attempt!
explicitly! measure! these! effects.! ! I! hypothesise! that! if! these! are! found! to! be!
significant! in! Switzerland,! a! country! known! for! relatively! distant! familial!
relationships,!then!this!effect!should!be!highly!generalizable!to!other!countries.!
!
In!this!paper,!I!expand!on!this!literature!(simultaneously!with!the!CMS!literature)!by!
focusing! on! the! outcomes! relating! to!male! siblings!moving! out! from! home! due! to!
CMS.! I! argue! that,!when! controlling! for! a!number!of! variables! and! individual?level!
fixed!effects,!the!act!moving!out!is!an!exogenous!event,!and!as!such!enables!a!clearer!
analysis! of! these! sibling! effects! –! at! least! in! the! short?term! period! over! which! a!
sibling!is!conscripted.!
!
! !
! 11!
Institutional!Setting*
!
Education!
!
The!Swiss!dropout!rate!is!approximately!12%!and!is!one!of!the!lowest!in!the!world.!
!
In!Switzerland,!the!government!provides!primary!education!for!free,!and!almost!all!
children!attend!these!public!schools.!After!primary!school,!students!enter!secondary!
education!where!they!are!more!likely!to!pay!for!tuition.!
!
The!minimum!age! at!which! an! individual! can!dropout! from!school! is! 15!years! (or!
after!9!years!of!schooling).!This!means!that! individuals!must!spend!some!time!in!a!
secondary!school!before!having!the!option!of!dropping!out.!
!
Conscription!
!
At!age!20,!Swiss!males!are!required!by!law!to!undertake!CMS.!Approximately!sixty!
percent!serve! in! the!military!and!10%!opt! for!civil! service.!Conscripts!are!rejected!
on!grounds!of!psychological!and!intelligence!tests,!as!well!as!on!grounds!of!general!
health1.!Those!deemed!unfit! for!service!pay!an!additional!3%!of! income!tax! for!10!
years.!These!selection!effects!are!addressed!in!the!Empirical!Strategy!section!below.!
!
All!military!conscripts!live!in!the!barracks!of!their!military!unit!for!the!length!of!their!
service.!Service!typically!lasts!for!at!least!18!to!21!weeks.!Those!in!certain!divisions!
as! well! as! those! seeking! higher! ranks! accept! longer! service! lengths.! Division!
selection!is!typically!the!result!of!a!recruitment!officer’s!allocation.!It!is!unusual!for!a!
conscript!to!serve!in!the!same!canton!as!the!one!he!resides!in.!
!
Deferment! options! in! Switzerland! are! very! limited.! Those! in! civilian! service! serve!
50%!longer!than!those!in!the!military!who!are!instead!called!back!every!year!for!3?
week!refresher!courses.!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!Of!those!that!don’t!serve,!approximately!40%!are!rejected!on!grounds!of!physical!health,!40%!on!grounds!of!psychological!reasons,!and!20%!for!a!combination!of!the!two.!
! 12!
Due! to! the!highly! limited!number!of!officially! recognised! civilian! service! locations,!
many! individuals! opting! for! public! service! have! to! move! out! to! perform! their!
service1.!
!
All!conscripts!are!paid!for!their!service.!Those!with!jobs!earn!80%!of!the!income!that!
they!would!have!otherwise!missed!out!on,!on!top!of!a!basic!salary!of!4!to!30!SFr!per!
day.!Students!and!the!unemployed!receive!a!fixed!compensation!of!62!SFr!on!top!of!
the!basic!salary.!Special!circumstances!such!as!being!a!parent!allow!for!the!provision!
of!further!income.!
!
Causal*Mechanisms*
!
This!institutional!setting!implies!a!number!of!causal!streams!through!which!a!sibling!
may!be!affected.!It!is!not!clear!in!which!direction!the!overall!effect!can!be!expect!to!
act!due!to!the!number!of,!often!conflicting,!channels.!These!include:!
!
A*financial*factor:!The!fact!that!conscripts!are!sent!away!from!home!(meaning!they!
do! not! burden! their! household! with! everyday! expenses)! and! are! paid! for! their!
service,! reduces! the! financial! burden! a! household! has.! This! reduction! in! financial!
constraints! allows! families! to! spend! more! on! the! education! of! younger! siblings!
through!private!tutoring,!or!better!schools!that!are!likely!to!be!more!expensive.!This!
causal! stream! is! more! likely! to! result! in! positive! impacts! for! the! siblings! of!
conscripts!that!are!in!poorer!households.!
!
A* sibling* support* factor:! The! fact! that! conscripts! are! made! to! leave! their! home!
suggests! that! there! may! be! reductions! in! sibling! support! factors! such! as! the!
homework!help!older!brothers!provide!for!their!younger!siblings,!and!the!emotional!
support! afforded.! These! factors! may! be! more! likely! to! have! a! stronger! affect! for!
females.!!
!
Other!factors!resulting!from!a!sibling!being!conscripted!may!also!be!at!play,!such!as!
more!general!peer!effects!or!even!a!resulting!increase!in!housework.!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!Both!service!options!therefore!force!individuals!to!move!out,!and!thus!allow!the!measurement!of!sibling!externalities!by!treating!conscription!as!an!exogenous!shock!when!controlling!for!a!range!of!selection!factors.!
! 13!
!
All!of!these!impacts!are!unlikely!to!differ!between!those!that!have!a!sibling!join!the!
military!service!and!those!that!have!a!sibling!join!the!civilian!service!as!the!financial!
and!sibling?presence!factors!likely!at!play!are!the!same!across!the!two!groups.!
! !
! 14!
Data!
!
SHP*Dataset*
!
The!principal!dataset!in!use!is!the!Swiss!Household!Panel!(SHP).!The!SHP!is!a!yearly!
panel! that! features!12!waves! from!1999!to!2010!and!consists!of!9,018!households!
and!22,708!individuals.!All!individuals!in!a!household!are!surveyed!either!directly!or!
by! proxy.! Questions! in! the! panel! concern! a! range! of! socioeconomic! and! family!
outcomes.!Data!collection!initiated!with!a!sample!of!5,074!households!(with!12,931!
household! members).! In! 2004! a! second! sample! of! 2,538! households! (with! 6,569!
members)!was!added.!
!
Sample*Construction*
!
I! first!separate!out!males!and!females,!creating!a!sample!for!each!gender.! I!do!this,!
instead!of! creating! a! joint! sample! and! employing! gender! interaction! terms,! purely!
because! the! separate! sample! approach! is! easier! to! work! with! when! applying! the!
sample!manipulations!discussed!below.!
!
For!each!gender!sample,!I!use!real!values!instead!of!nominal!ones!to!guarantee!the!
income!measures!used!are!consistent!across!the!large!time!span!of!the!dataset.!
!
To! ensure! an! appropriate! comparison! group,! I! exclude! all! individuals! that! did!not!
have! at! least! one!male! sibling! of! age! 20! (the! required! age! of! service)! at! the! time!
when!they!have!the!option!of!dropping!out!open!to!them!(i.e.,!are!between!the!ages!
of!15!and!19).! I! also!exclude!non?Swiss! citizens,! as! their! siblings!are!not! called!up!
for,!or!even!allowed!to!complete,!compulsory!military!service!(CMS).!
!
To! ensure! that! the! causal! relationship! flows! in! the! direction! intended,! that! is,! the!
instance! of! a! sibling! being! conscripted! is! followed! by! and! affects! the! target!
individual’s!decision!on!whether!to!drop!out,!and!not!the!other!way!around,!I!drop!
all!observations!that!feature!the!target!individual!being!older!than!all!their!brothers.!
This!makes!sure! that! that! the!choice! for!dropping!out!of!high!school!occurs!either!
contemporaneously!with!or!just!after!a!brother!can!be!conscripted.!!
! 15!
!
From! here! I! drop! all! target! individuals! under! the! age! of! 20! (the! age! where!
individuals! should! have! completed! their! schooling)! such! that! the! number! of! older!
brothers!conscripted!can!be!compared!with!an!outcome!variable!examining!whether!
an!individual!has!completed!their!schooling!or!dropped!out.!
!
I! also! construct! an! ancillary! sample! (for! each! of! the! genders)! where! I! attempt! to!
examine! the! act! of! dropping! out! itself! at! the! time! a! sibling! is! conscripted! (i.e.,! a!
contemporaneous! effect).! I! do! this! by! restrict! these! samples! to! target! individuals!
aged!between!15! and!19! –! the! range!where! they! have! the! option! of! dropping! out!
from!high!school,!the!key!metric!in!my!analysis.!
!
Upon!examining!the!sample!in!detail,!I!find!that,!due!to!the!way!the!question!is!asked!
in!the!survey1,!only!1%!of!Swiss!males!at!age!20!reported!that!they!were!conducting!
their!service.!This!is!significantly!underreported.2!
!
As!the!variable!is!unsalvageable,!I!drop!it!from!my!sample!and!generate!a!predicted!
variable! examining! the! rate! of! conscription.! The! procedure! I! employ! to! do! this!
involves! the! use! of! a! second! dataset! and! a! two?sample! two?stage! least! squares!
approach!to!be!discussed!in!the!Empirical!Strategy!section!below.!Briefly,!it!involves!
predicting! the! probability! of! service! through! a! number! of! characteristics! in! one!
dataset!and!then!imputing!the!variable!in!the!second!dataset.3!
!
SLFS*dataset*
!
The!ancillary!dataset!used! for! the! imputation!procedure! is! the!Swiss!Labour!Force!
Survey!(SLFS).!Both!the!SLFS!and!the!SHP!strictly!adhere!to!international!definitions!
to! allow! for! comparisons! with! OECD! and! EU! data,! this! means! the! two! are! highly!
compatible.!!
!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!Individuals!are!queried!on!their!working!status!and!then!asked!if!they!missed!work!or!are!currently!not!working!due!to!CMS.!The!focus!on!employment!suggests!the!variable!is!significantly!biased!as!well!as!underreported.!2!Recall!that!approximately!60%!of!males!serve!in!the!military!and!10%!in!the!civilian!service.!3!As!will!be!discussed!in!the!Empirical!Strategy!section,!selecting!appropriate!variables!in!the!first!stage!can!allow!for!the!imputed!variable!to!also!be!instrumented!(therefore!reducing!bias).!
! 16!
The!SLFS!is!a!cross?sectional!dataset!that!is!conducted!on!a!yearly!basis!between!the!
years!of!1991!and!20091.!!
!
The! SLFS! targets! the! permanent! resident! population! aged! 15! and! older! and! is!
conducted! by! the! Federal! Statistical! Office! of! Switzerland.! In! total,! the! dataset!
contains!105,000!interviews.!
!
The!relevant!questions!on!conscription!in!this!survey!are!not!present!prior!to!2003.!I!
therefore!focus!my!analysis!on!the!years!that!follow!(in!both!datasets).!
!
The! SLFS! dataset! groups! military! service! and! civilian! service! together.! This! is!
unlikely!to!be!an!issue!because,!as!discussed!in!the!Institutional!Setting!section,!both!
imply!the!same!causal!mechanisms,!as!such;!the!outcomes!should!be!very!similar.!!
!
Summary*Statistics!
!
As!a!result!of!the!sample!construction!procedure,!the!number!of!individuals!drops!to!
174!(with!509!observations)!in!the!female!sample!and!209!(with!600!observations)!
in!the!male!sample.!
!
The!summary!statistics!for!the!key!male!and!female!samples!can!be!seen!in!Table!3!
below.!
!
The! mean! dropout! rate,! household! income,! and! the! percentage! of! individuals! in!
rural! areas! are! all! nationally! representative! averages! corroborated! with! census!
statistics!from!Switzerland’s!Federal!Statistical!Office.!
! !
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!From!2010!onwards,!the!survey!is!conducted!on!a!continuous!basis.!
! 17!
!
Table!2:!SHP!Summary!Statistics!
!Males! Females!
Variables! Mean! Standard!Deviation! Mean! Standard!
Deviation!
Dropout! 0.125! 0.337! 0.!087! 0!.284!Age! 22.7! 2.42! 22.4! 2.22!
No.!of!Siblings!Conscripted16! 0.931! 0.594! 1.34! 0.939!Rural17! 0.155! 0.362! 0!.171! 0.378!
Household!Income18! 136,521! 117,421! 129,214! 80,377!No.!of!Brothers! 1.418! 0.629! 2.045! 1.01!
Individuals! 209! !! 174! !!Observations! 600!
!509! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
16!Imputed!17!Rural=1,!Urban/Suburban=0!18!2005!SFr!!
!
The!high!average!number!of!brothers! is!an!expected!result!of! the!sample!selection!
criteria’s!requirement!of!at!least!one!older!male!sibling!per!household.!
!
The! results! across! genders! are! very! similar,! and! as! such,! the! summary! statistics!
suggest!that!both!gender!specific!samples!are!nationally!representative.!
!
!
! !
! 18!
Empirical!strategy!
!
I! employ! a! number! of! strategies! to! address! three! key! concerns:! how! I! ensure! a!
causal! relationship,! how! I! minimise! bias/endogeneity! and! how! I! correct! the! data!
problems!resulting!from!the!under?reported!conscription!variable.!
!
Causality!
!
As! the!age!of!conscription! is!greater! than!the!age!of!high!school!graduation,!under!
normal! circumstances,! an! individual! with! a! younger! brother! cannot! have! their!
schooling! affected!when! that! brother! undertakes! his! service.! This! is! because! they!
would!likely!have!already!graduated!when!their!younger!brother!is!conscripted.! In!
this!case,!any!measurable!effect!would!largely!be!the!result!of!an!individual!dropping!
out! affecting! their! brothers! likelihood! of! service;! this! is! not! what! I! intend! to!
measure.!
!
As!briefly!pointed!out! in! the!Data! section,! I! ensure! that!any!measured!change! in!a!
target!individual’s!propensity!to!dropout!is!the!result!of!a!sibling!being!conscripted,!
and!not! the! result! of! reverse! causality,! by!only! examining! individuals! that!have! at!
least!one!older!brother.!This! ensures! that! that! the! choice! for!dropping!out!of!high!
school! occurs! either! contemporaneously! with! or! just! after! a! brother! can! be!
conscripted.!!
!
Causality,! however,! cannot!be! ensured!with!out!minimising! endogeneity.!As!noted!
earlier,! Swiss! conscription! is! subject! to! many! selection! effects,! as! such,! I! employ!
fixed! effects! regression! techniques! as!well! as! an! instrumental! variable! strategy! to!
control!for!these!and!argue!causality.!
!
Fixed*Effects*Regressions*
!
The! factors! determining! conscript! admission! (ability,! physical! fitness,! personality,!
and! political! views)! are! largely! fixed! at! the! individual! level,! especially! when!
considering!the!limited!time!scale!I!employ.!
!
! 19!
While! the! fact! that! I! examine! externalities! means! the! factors! leading! to! a! sibling!
being!conscripted!are! less! likely!to!be!the!factors!attributed!to!educational!success!
(therefore!affording!some! level!of!exogeneity),! it! is!notable! that!some!of! the!major!
determinants! of! educational! attainment! feature! high! levels! of! within?family!
correlations!(ability,!race,!etc.).!These!are!largely!fixed!and!are!similar!to!the!factors!
determining!conscript!admission.!
!
I!utilise!individual?level!fixed!effects!regressions!to!control!for!these!factors1!as!the!
individual?level! fixed!effects! should!be!a!more! conservative!model! and!encompass!
any!fixed!effects!at!the!household?level2.! Individual!fixed!effects!for!example!better!
control!for!the!number!of!siblings,!which!may!vary!in!a!household?level!fixed!effects!
model.!
!
To!control!for!broader!peer?level!effects!that!may!make!an!individual!more!likely!to!
avoid!CMS3,!I!also!include!canton?level!fixed!effects!dummies!in!every!regression.!
!
These!two!layers!of!fixed!effect!controls!should!result!in!more!exogenous!estimates!
for!the!effect!of!a!sibling!being!conscripted!on!an!individual’s!likelihood!of!dropping!
out!of!high!school.!
!
2S2SLS*Generation*Procedure*
!
As!outlined!in!the!Data!section!above,! the!variable!of! interest! in!the!SHP!dataset! is!
significantly! underreported! and! cannot! be! used! in! a! regression! analysis! without!
giving! rise! to! substantial! levels! of! bias.! The! SHP! conscription!variable! is! therefore!
abandoned!for!a!generated/predicted!alternative.!
!
To!predict!the!conscription!variable,!I!employ!a!two?sample!two?stage!least!squares!
(2S2SLS)!approach!where!the!chief!goal!is!to!generate!a!new!variable!as!opposed!to!
minimising!the!endogeneity!of!a!variable!through!the!use!of!an!instrument.1!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!It!is!very!important!to!control!for!the!factors!that!may!lead!individuals!to!attempt!to!intentionally!fail!an!entry!exam;!the!individual?level!fixed!effects!regression!implicitly!controls!for!the!political!factor!(among!others)!most!likely!to!determine!this.!2!For!certainty,!this!is!subject!to!a!robustness!test!in!the!Robustness!Tests!section!of!this!paper.!3!The!percentage!of!males!completing!service!varies!immensely!between!cantons.!Naturally,!more!densely!populated!city!cantons!such!as!Zurich!or!Geneva!experience!lower!levels!of!service!than!other!cantons.!
! 20!
!
The!generation!procedure!I!employ!can!be!split!into!3!main!steps:!
!
Step*1:!Upon!ensuing!that!the!two!datasets!feature!very!similar!variable!means!and!
distributions,!I!regress,!using!ordinary!least!squares,!the!conscription!variable!in!the!
SLFS! dataset! (containing! clean! information! on! service! rates)! against! a! number! of!
characteristics! that! are! present! in! both! datasets.! This! is! the! first! stage! regression,!
and!can!be!represented!as:!
!
N"=!!!+!!!γ!+!!!δ + ε!!
Where! N! is! the! conscription! variable,!!! !is! the! constant! term,!γ !is! a! vector! of!characteristics! present! in! both! datasets,!!!!is! the! vector! of! coefficients! for! these!characteristics,!δ!is!an!instrument,!!!!the!coefficient!for!that!instrument!and!ε!is!the!error!term.!
!!
Step*2:! I!use! the!estimated!coefficients,! represented!by!!!,!!!!and!!!,! from!the! first!stage!regression!to!compute!the!predicted!values!in!the!SHP!dataset.!This!generates!
the!variable!of!interest!in!the!panel!dataset,!therefore!allowing!the!use!of!the!panel!
techniques!outlined!earlier!to!minimise!endogeneity.!
!
Step*3:! I! utilise! a! linear! fixed! effects! regression! to! estimate! the! effect! of! having! a!
sibling!conscripted!on!the!high!school!dropout!rate.!This!can!be!represented!by!the!
equation:!
!
D"=!N* + !!γ"+"τ!!
Where! D! is! the! dropout! dummy,! N*! is! an! imputed! N,! and!τ!is! the! time?varying!component!of!the!error!term.!
!
We!can!see!that,! if!appropriate!variables!are!used!in!the!first!stage!regression,! this!
2S2SLS!procedure!can!be!used!to!simultaneously!generate!as!well!as!instrument!the!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!As!will!be!discussed!later,!selecting!suitable!variables!in!the!first!stage!regression!can!ensure!the!generated!variable!is!also!exogenous.!
! 21!
service! variable.! This! adds! another! layer! to! the! fixed! effects! strategies! outlined!
above!to!help!further!minimise!endogeneity.!
!Due!to!the!use!of!a!predicted!variable!in!the!second!stage!regression,!the!computed!
errors!in!the!must!be!adjusted.!A!number!of!papers!examine!this!second?stage!error!
adjustment! process,! these! include! Inoue! and! Solon! (2010),! Murphy! and! Topel!
(2002),!Angrist!and!Krueger!(1995)!and!Hole!(2006).!
!
Of! the!valid!options!available,! including!a!minimum!distance!estimator!and!a!delta!
method! approach,! I! follow! the! bootstrap! procedure! employed! by! Björklund! and!
Jäntti!(1997)!for!its!simplicity.!
!
The*Instrument*
!
Switzerland!has! recently!experienced!a! lower! rate!of! service! than! in!past!decades.!
This! rate! has! historically! centred! around! 80%! but! has! declined! to! approximately!
60%!with!the!introduction!of!civilian!service!options,!a!more!selective!military,!and!
a!lower!level!of!support!for!CMS.!This!rate!of!service!has!been!relatively!stable!over!
the!past!decade.!
!
In!my! first! stage! regression,! I! utilise! the! average! rate! of! service! to! instrument! the!
sibling! service! variable.! This! instrument! is! employed! by! Bedard! and! Deschenes!
(2006)! to! assess! the! long! term! health! effects! associated! with! military?induced!
smoking!in!World!War!II.!
!
Note!that!the!strength!of!this!instrument!increases!with!the!number!of!years/cross?
sections! used;! an! f?test! of! 12.13! demonstrates! that! it! surpasses! the! rule! of! thumb!
threshold! for! instrument! strength! advocated!by! Stock! and!Watson! (2003),!when! I!
only!utilise!the!7!years!of!available/compatible!data!in!the!SHP!and!SLFS.1! !
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!The!first!stage!regression!output!can!be!found!in!Appendix!2A.!
! 22!
Results!
!
In* this* section* I* find* that,*overall,* there* is*a* strong*negative* shock* to* the*high*school*
completion* rate* of* females* when* a* male* sibling* is* conscripted.* Upon* dissecting* this*
effect,*I*find*that*less*wealthy*households*tend*to*experience*stronger*effects*and*that*
decreases* in* sibling* support* seem* to* explain* much* of* the* resulting* femalesNspecific*
increases*in*dropout*rates.*
!
I!begin!by!examining!females!that!have!had!the!opportunity!to!complete!high!school!
(that! is,! are! above! the! age! of! 19)! and! evaluate! the! effect! of! having! male! siblings!
conscripted!on!their!probability!of!dropping!out!of!high!school.!
!
As!discussed!previously,!I!only!examine!individuals!who!have!at!least!one!older!male!
sibling!to!ensure!appropriate!effect!directionality.!
!
I!control!for!a!number!of!factors!in!the!regressions!below,!these!include!the!size!of!
the! household! and! the! number! of! brothers! that! are! liable! for! service! –! these! are!
included! to! avoid! estimating! sibship! size! effects! as! part! of! the! ‘number!of! siblings!
conscripted’!variable.!
!
I!also!control!for!the!number!of!brothers!permanently!leaving!a!household!as!those!
at!an!age! likely! to!be!conscripted!are!also!at!an!age!where! they!are!more! likely! to!
move!out!–!this!ensures!the!externalities!associated!with!the!minimum!of!5!months!
of! service!are!estimated! separately! from!other! circumstances!where!a! sibling!may!
leave!the!household.!
!
Lastly,!I!include!canton?level!fixed!effects!dummies!(in!all!of!the!regressions)!as!well!
as! individual! level! fixed! effects! to! control! for! a! range! of! factors,! such! as! political!
leanings,!that!may!bias!results.!Note!that!the!22!canton!dummies!are!excluded!from!
the!Tables!below!and!can!be!found!in!the!Appendix!section.!
!
As!we!see!in!the!tables!below,!many!of!these!controls!drop!out!when!controlling!for!
individual?level!fixed!effects.!!
! 23!
Recall! from! the! Empirical! Strategy! section! that! I! employ! a! 2S2SLS! IV! strategy.! All!
Tables!in!this!section!contain!second!stage!regressions,!the!first!stage!output!can!be!
found!in!Appendix!2A.!
!
A! Hausman! test1!suggests! that! the! use! of! a! fixed! effects! model! is! not! necessary,!
possibly!due!to!the!effectiveness!of!the!instrument.!
!
In! the! regressions!shown! in!Table!3,!we!observe! that,! as!expected,! the!coefficients!
for! age! are! insignificant! as,! by! age! 20,! individuals! should! have! already! completed!
their!schooling!or!dropped!out!of!high!school.!!
!
!
Table!3:!Females!aged!20!and!above2!
!OLS3! RE! FE!
Dependant!Variable!(Dropout=1)! (1)! (2)! (3)!
No.!of!Siblings!Conscripted! 0.0209***! 0.0231**! 0.0434**!
!(0.00779)! (0.00939)! (0.0194)!
Age! ?0.0459! ?0.0419! ?0.0602!
!(0.0352)! (0.0316)! (0.0366)!
Age!Squared! 0.000868! 0.000800! 0.00119!
!(0.000749)! (0.000671)! (0.000770)!
No.!of!Siblings! ?0.531! ?0.471! ?!
!(0.411)! (0.370)!
!Siblings!Left!Household! ?0.0172**! ?0.0179! ?!
!(0.00840)! (0.0112)!
!Brothers!Left!Household! 0.000220! 0.000826! ?!
!(0.0151)! (0.0198)!
!No.!of!Brothers! 0.569! 0.506! ?!
!(0.410)! (0.369)!
!Constant! 0.520! 0.462! 0.666!
!(0.410)! (0.368)! (0.435)!
Observations! 509! 509! 509!R?squared! 0.284! 0.2815! 0.0265!
No.!of!Individuals! ?!! 174! 174!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!P?value!=!0.7575,!therefore!we!cannot!reject!the!null!that!the!random!effects!model!is!consistent.!2!The!canton!dummies!for!these!regressions!are!shown!in!appendix!3A.!Standard!errors!are!bootstrapped!and!are!in!parentheses.!Asterisks!denote!p?values:!***!p<0.01,!**!p<0.05,!*!p<0.1!3!OLS!denotes!ordinary!least!squares,!RE!denotes!random!effects,!and!FE!denotes!fixed!effects!at!the!individual!level.!
! 24!
Table!3,!column!3!shows!that,!in!the!most!conservative!model,!having!more!siblings!
conscripted,! the! key! variable! in! the! analysis,! is! associated!with! an! increase! in! the!
dropout!rate!by!approximately!4%.!This!is!significant!at!the!5%!level.!!
!
Note! that! the!number!of! brothers! a! female!has! is! associated!with!no! effect! on! the!
probability! of! dropping! out! from! high! school.! Initially,! this! seems! to! support! the!
findings!of!Kaestner!(1997)!and!Hauser!and!Kuo!(1998)!that!there!is!no!evidence!for!
sibling!gender!composition!effects!(therefore!contesting!the!findings!of!Butcher!and!
Case! (1994)),! however,! it! is! possible! that! the! effect! found! by! Butcher! and! Case!
(1994)! is,! at! least! partially,! reflected! in! the! variable! measuring! the! number! of!
siblings!conscripted–!this!is!examined!in!detail!below.!
!
Table!4!follows!from!Table!3!by!employing!males!as!the!target!group.!!
!
As! in! the! female! specific! results,! a!Hausman! test1!suggests! a! fixed! effects!model! is!
not!needed.!!
!We!observe! that! the!key!variable! in! the!analysis,! the! coefficient! for! the!number!of!
siblings!conscripted!variable!is!insignificant!and!close!to!zero.!This!is!vastly!different!
to!the!female!case!and!therefore!suggests!that!the!causal!channels!are!more!likely!to!
involve! sibling! support! factors! than! financial! ones.! This! means! that! the! sibling!
composition! effects! identified! by! Butcher! and! Case! (1994)! are! likely,! at! least!
partially,! reflected! in! the! finding! that!males! leaving! to! complete! their! CMS! have! a!
detrimental!effect!on!their!sisters’!education.!
! !
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!P?value!=!0.7575,!therefore!we!cannot!reject!the!null!that!the!random!effects!model!is!consistent.!
! 25!
!Table!4:!Males!aged!20!and!above1!
!OLS! RE! FE!
Dependant!Variable!(Dropout=1)! (1)! (2)! (3)!
No.!of!Siblings!Conscripted! 0.000378! ?0.00476! ?0.0108!
!(0.00564)! (0.00691)! (0.0101)!
Age! ?0.0167! ?0.0190! ?0.0199!
!(0.0237)! (0.0190)! (0.0203)!
Age!Squared! 0.000313! 0.000374! 0.000403!
!(0.000497)! (0.000400)! (0.000424)!
No.!of!Siblings! 0.0101! 0.000228! ?!
!(0.00751)! (0.0131)!
!Siblings!Left!Household! ?0.00317! ?0.00187! ?!
!(0.0130)! (0.0209)!
!Brothers!Left!Household! ?0.00623! ?0.0104! ?!
!(0.0145)! (0.0231)!
!No.!of!Brothers! 0.00893! 0.0155! ?!
!(0.0110)! (0.0174)!
!Constant! 0.234! 0.244! 0.260!
!(0.278)! (0.226)! (0.240)!
Observations! 600! 600! 600!R?squared! 0.0366! 0.0703! 0.0012!
No.!of!Individuals! ?!! 209! 209!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!The!canton!dummies!for!these!regressions!are!shown!in!appendix!3B.!Standard!errors!are!in!parentheses.!Asterisks!denote!p?values:!***!p<0.01,!**!p<0.05,!*!p<0.1!!
!
From! the! analysis! so! far,! it! is! clear! that! there! exist! some! negative! externalities!
associated!with!CMS!for! females.!This!result!corroborates!the! findings!of!Cipollone!
and!Rosolia!(2007)!and!suggests!that!the!papers! in!the!CMS!literature!that!employ!
women! as! a! control! group,! namely! Card! and! Lemieux! (2001)! and! Paloyo! (2010),!
may!underestimate!the!effects!associated!with!CMS!on!males.!
!
In! order! to! further! probe! into! the! sibling! externalities! literature,! I! conduct! an!
analysis!of!the!contemporaneous!effects!associated!with!CMS!–!this! is!necessary!as!
the!sibling?support!hypothesis!should!involve!a!drop!in!support!and!therefore!result!
in!an!increase!in!the!dropout!propensity!at!the!time!a!sibling!is!conscripted.!
!
! 26!
In! Table! 5,! I! attempt! to! observe! the! act! of! dropping! out! itself! by! focussing! on!
individuals!that!are!able!to!do!so.!I!therefore!focus!my!analysis!on!those!between!the!
ages!of!15!(the!minimum!school!leaving!age!in!Switzerland)!and!19!(the!age!where!
most!individuals!should!have!completed!high!school).!
!
In! this! contemporaneous! effect! analysis,! I! use! a! dummy! variable! that! identifies!
whether! a! sibling! is! conscripted! in! the! current! year! as! opposed! to! examining! the!
number!of!siblings!conscripted.!
!
It!is!notable!that,!in!this!contemporaneous!case,!a!Hausman!test,!yielding!a!p?value!of!
0.000,!shows!that!individual!level!fixed!effects!are!necessary!to!reduce!bias.!This!is!
likely! due! to! the! fact! that! the! instrument! is! less! effective! in! the! contemporaneous!
case!as!it!requires!a!larger!time!range.!This!may!also!help!explain!the!high!errors!for!
the!key!(instrumented)!variable!in!the!OLS!and!random?effects!regressions!in!Table!
5!below.!!
!
The!age!coefficients!in!the!most!conservative!model,!the!fixed!effects!regression,!are!
highly!significant!and!imply!that!an!individual!is!most!likely!to!dropout!at!the!age!of!
16.!
!In! table! 5,! we! observe! that,! as! before,! having! a! sibling! leave! to! undertake! CMS!
negatively!impacts!the!probability!of!completing!high!school.!!
! !
! 27!
!
Table!5:!Females!aged!15!to!191!
!OLS! RE! FE!
Dependant!Variable!(Dropout=1)! (1)! (2)! (3)!
Sibling!Conscripted! ?0.00242! ?0.00242! 0.112**!
!(0.0241)! (0.0241)! (0.0459!)!
Age! 1.307***! 1.307***! 1.475***!
!(0.300)! (0.300)! (0.311)!
Age!Squared! ?0.0392***! ?0.0392***! ?0.0452***!
!(0.00887)! (0.00887)! (0.00924)!
No.!of!Siblings! 10.75***! 0.0998***! ?!
!(2.522)! (0.0270)!
!Siblings!Left!Household! ?0.126***! ?0.126***! ?!
!(0.0265)! (0.0265)!
!Brothers!Left!Household! 0.0876**! 0.0876**! ?!
!(0.0422)! (0.0422)!
!No.!of!Brothers! ?10.69***! ?0.0467! ?!
!(2.522)! (0.0389)!
!Constant! ?10.65***! ?10.65***! ?11.85***!
!(2.523)! (2.523)! (2.618)!
Observations! 612! 612! 612!R?squared! 0.302! 0.188! 0.0379!
No.!of!Individuals!!
233! 233!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!The!canton!dummies!for!these!regressions!are!shown!in!appendix!3C.!Standard!errors!are!bootstrapped!and!are!in!parentheses.!Asterisks!denote!p?values:!***!p<0.01,!**!p<0.05,!*!p<0.1!!
!
An!identical!analysis!is!conducted!for!men!(see!Appendix!3D);!this!shows,!as!in!the!
non?contemporaneous!case,!no!discernable!effect!of!having!a!sibling!conscripted.!
!
I!now!move!on!to!further!analyse!the!causal!channels!leading!to!the!increase!in!the!
dropout!rate!for!females.!
!
To!observe!if!there!are!any!wealth!or!income!effects,!I!segment!the!sample!in!Table!3!
into! 2! groups,! an! upper! class! and! a! middle/lower?class2,! and! conduct! individual?
level!fixed!effects!regressions,!the!results!of!which!are!shown!in!Table!6!below.!
!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!2!!The!upper!class!is!defined!as!households!earning!more!that!150,000!SFr,!while!those!in!the!lower/middle?class!are!defined!as!households!earning!less!than!that.!!
! 28!
!
Table!6:!Family!incomes!for!females!aged!20!and!above1!
!The!Lower/Middle!Class! The!Upper!Class!
Dependant!Variable!(Dropout=1)! (1)! (2)!
No.!of!Siblings!Conscripted! 0.0586**! 0.0238!
!(0.0208)! (0.0340)!
Age! 0.00324! ?0.0803!
!(0.0426)! (0.0549)!
Age!Squared! ?0.000104! 0.00161!
!(0.000889)! (0.00117)!
Constant! ?0.118! 0.931!
!(0.506)! (0.621)!
Observations! 284! 225!R?squared! 0.0701! 0.0017!
No.!of!Individuals! 127! 103!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!The!canton!dummies!for!these!regressions!are!shown!in!appendix!3E.!Standard!errors!are!bootstrapped!and!are!in!parentheses.!Asterisks!denote!p?values:!***!p<0.01,!**!p<0.05,!*!p<0.1!Individual!level!fixed!effects!are!employed!for!both!regressions.!!
!
This! analysis! demonstrates! that! some! income?related! effects! are! present! when!
siblings!are!conscripted,!as!females!in!less!wealthy!families!experience!a!significant!
increase!of!5%! in! their!propensity! to!dropout!of!high! school!while! those! in! richer!
families!experience!no!discernable!effects.!
!
As!in!Table!7!below,!measurable!income!related!effects!also!exist!for!males!(and!are!
significant!at! the!5%!level),!however! it! is!notable!that! the!male!specific!coefficient,!
unlike! that! for! females,! is! negative,! that! is,! having! a! sibling! conscripted! benefits!
males!in!terms!of!a!lower!dropout!rate.!
!
One! possible! hypothesis! that! explains! the! discrepancy! between! genders! is! that!
females!experience!support!effects!that!are!correlated!with!family!income!levels.!To!
investigate! this!hypothesis! I! split! the! lower/middle?class! females! into! two!groups:!
those! that! identified! themselves! as! having! experienced! a! reduction! in! emotional!
support! from! their! family! and! those! that! identified! themselves! as! having!
experienced!either!no!change!or!an!increase!in!support.! !
! 29!
!
Table!7:!Family!incomes!for!males!aged!20!and!above1!
!The!Lower/Middle!Class! The!Upper!Class!
Dependant!Variable!(Dropout=1)! (1)! (2)!
No.!of!Siblings!Conscripted! ?0.0420**! 0.00923!
!(0.0173)! (0.0156)!
Age! ?0.0113! ?0.0435!
!(0.0329)! (0.0333)!
Age!Squared! 0.000250! 0.000854!
!(0.000684)! (0.000700)!
Constant! 0.177! 0.535!
!(0.393)! (0.389)!
Observations! 302! 296!R?squared! 0.0016! 0.013!
No.!of!Individuals! 148! 131!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!The!canton!dummies!for!these!regressions!are!shown!in!appendix!3F.!Standard!errors!are!bootstrapped!and!are!in!parentheses.!Asterisks!denote!p?values:!***!p<0.01,!**!p<0.05,!*!p<0.1!Individual!level!fixed!effects!are!used!for!both!regressions.!!
!
Note!that!due!to!the!limited!size!of!the!sample,!I!reduce!the!age!cut?off!to!18!years!in!
the! following! regressions.! I! argue! that! this! is! unlikely! to! increase! bias! by! much!
because,!as!shown!previously,! individuals! tend!to!dropout!at! the!age!of!16,! that! is,!
earlier!rather!than!later.!
! !
! 30!
!
Table!8:!Emotional!support!for!females!aged!18!and!above!in!low/middleLincome!families1!
!Decrease!in!Emotional!Support! No!Change/Increase!in!Emotional!Support!
Dependant!Variable!(Dropout=1)! (1)! (2)!
No.!of!Siblings!Conscripted! 0.214**! ?0.113**!
!(0.0771)! (0.0452)!
Age! 0.0493! ?0.159!
!(0.213)! (0.100)!
Age!Squared! ?0.00157! 0.00334!
!(0.00465)! (0.00215)!
Constant! ?0.707! 2.095*!
!(2.343)! (1.120)!
Observations! 117! 262!R?squared! 0.387! 0.338!
No.!of!Individuals! 91! 145!! ! !
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!The!canton!dummies!for!these!regressions!are!shown!in!appendix!3G.!Standard!errors!are!bootstrapped!and!are!in!parentheses.!Asterisks!denote!p?values:!***!p<0.01,!**!p<0.05,!*!p<0.1!Individual!level!fixed!effects!are!used!for!both!regressions.!!
!
Here! we! see! that,! in! middle! or! low?income! households,! those! who! experience! a!
decrease! in! emotional! support! as! a! result! of! a! sibling! being! conscripted! are!much!
more! likely! to! dropout,! while! those! who! experience! no! change! or! an! increase! in!
emotional!support!are!less!likely!to!do!so.!Both!of!these!findings!are!significant!at!the!
5%!level.!
!
A! comparison!of! emotional! support! in!wealthier! families! (Appendix!3H)! shows!no!
discernable!effect,!suggesting!that!this!form!of!support!may!not!be!as!critical!in!these!
households!as! it! is! in! less!wealthy!ones.!This! link! is!upheld!by!a!number!of!papers!
examining! support! across! family! incomes! (Wight,! Botticello! et! al.! 2006;! North,!
Holahan!et!al.!2008).!
!
Due!to!the!limited!sample!sizes,!it!is!not!possible!to!separate!the!effect!of!increased!
support!from!the!effect!of!increased!income!in!this!analysis.!
!
In!summary,!although!it!seems!clear!that!some!income!effects!are!present,!they!seem!
to!be!outweighed!by! the! support!effects! for! females! (as! seen! in!Table!3,! analysing!
! 31!
the!overall! effects).!Males!do!not! seem! to!be!affected!by! these! support! effects! and!
only! benefit! from! the! income! effects! when! in! the! lower! end! of! the! wealth!
distribution.!
!
It! seems! very! plausible! that! the! gender! composition! effects! found! by!Butcher! and!
Case!(1994)!are,!at!least!partially,!a!manifestation!of!sibling!support!effects.!
! !
! 32!
Robustness!Checks!
!
In!this!section!I!perform!four!robustness/sensitivity!tests! to!ensure!the!results!are!
consistent!and!not!of!a!spurious!nature.!The!first!of!these!is!a!falsification!exercise,!
the!second!an!examination!of!the!individual?level!fixed!effects!assumption,!the!third!
an! analysis! of! the! base! year,! and! the! fourth! a! robustness! check! for! the! sibling!
support!finding.!
!
Falsification!
!
The! first! test,! the! falsification!exercise,! shown! in!Table!9!below! involves!assessing!
the!effect!of!a!sibling!being!conscripted!if!that!sibling!lives!in!a!different!household!
(due!to!parental!divorce,!for!example).!If!the!externality!relationship!is!not!spurious,!
there!should!be!no!discernable!effect!to!having!a!brother!conscripted,!as!household!
level!effects!should!be!largely!contained!within!the!household!(i.e.,!a!sibling!already!
outside!the!households!cannot!affect!wealth!or!support!factors)!
! !
! 33!
!Table!9:!Females!aged!20!and!above!in!different!households1!
!OLS2! RE! FE!
Dependant!Variable!(Dropout=1)! (1)! (2)! (3)!
No.!of!Siblings!Conscripted! ?0.00512! ?0.00512! 0.0140!
!(0.00666)! (0.00666)! (0.0286)!
Age! ?0.0283! ?0.0283! ?0.0694!
!(0.0307)! (0.0307)! (0.0495)!
Age!Squared! 0.000533! 0.000533! 0.00140!
!(0.000655)! (0.000655)! (0.00105)!
No.!of!Siblings! 0.0215**! 0.0215**! ?!
!(0.00937)! (0.00937)!
!Siblings!Left!Household! ?0.00337! ?0.00337! ?!
!(0.0140)! (0.0140)!
!Brothers!Left!Household! ?0.0165! ?0.0165! ?!
!(0.0169)! (0.0169)!
!No.!of!Brothers! 0.0219*! 0.0219*! ?!
!(0.0130)! (0.0130)!
!Constant! 0.386! 0.386! 0.795!
!(0.356)! (0.356)! (0.568)!
Observations! 280! 280! 280!R?squared! 0.2012! 0.2012! 0.0412!
No.!of!Individuals! ?! 109! 109!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!The!canton!dummies!for!these!regressions!are!shown!in!appendix!4A.!Standard!errors!are!bootstrapped!and!are!in!parentheses.!Asterisks!denote!p?values:!***!p<0.01,!**!p<0.05,!*!p<0.1!2!OLS!denotes!ordinary!least!squares,!RE!denotes!random!effects,!and!FE!denotes!fixed!effects!at!the!individual!level.!!!We! observe! that! the! coefficient! for! the! number! of! siblings! conscripted! variable! is!
insignificant!and!close! to! zero!across!all! three!models! in!Table!9.!This! falsification!
exercise! show! that! the! causal! effects! found!are! robust! and! therefore! suggests! that!
the!results!in!Table!3!are!unlikely!to!be!spurious.!
!
Individual*V.S.*Household*Fixed*Effects*
*
As!outlined! in! the!Empirical! Strategy! section,! I! utilise! individual?level! fixed! effects!
over! household?level! fixed! effects! because! I! argue! that! the! individual?level! effects!
largely! encompass! the! household! level! effects! and! are!more! conservative! because!
they!implicitly!control!for!factors!that!household?level!fixed!effects!do!not.!I!now!test!
! 34!
the! significance! of! this! assumption!by! comparing! the! two! schemes! for! each!of! the!
genders.!
!Table!10:!Individual!V.S.!household!level!fixed!effects1!
!Females! Males!
!Ind.!FE2! HH!FE! Ind.!FE! HH!FE!
Dependant!Variable!(Dropout=1)! (1)! (2)! (3)! (4)!
No.!of!Siblings!Conscripted! 0.0434**! 0.0313*! ?0.0108! ?0.00726!
!(0.0194)! (0.0167)! (0.0101)! (0.0129)!
Age! ?0.0602! ?0.0462! ?0.0199! ?0.0347!
!(0.0366)! (0.0337)! (0.0203)! (0.0255)!
Age!Squared! 0.00119! 0.000921! 0.000403! 0.000694!
!(0.000770)! (0.000712)! (0.000424)! (0.000532)!
No.!of!Siblings! ?! 0.0121! ?! 0.00487!
!!! (0.521)! !! (0.0658)!
Siblings!Left!Household! ?! ?0.0406! ?! 0.0117!
!!! (0.254)! !! (0.133)!
Brothers!Left!Household! ?! ?0.0298! ?! ?0.00946!
!!! (0.365)! !! (0.245)!
No.!of!Brothers! ?! 0.0529! ?! ?0.00902!
!!! (0.344)! !! (0.177)!
Constant! 0.666! 0.557! 0.260! 0.434!
!(0.435)! (0.443)! (0.240)! (0.510)!
Observations! 509! 509! 600! 600!R?squared! 0.063! 0.695! 0.009! 0.459!
No.!of!Individuals! 174! ?! 209! ?!No.!of!Households! ?! 153!! ?! 177!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!The!canton!dummies!for!these!regressions!are!shown!in!appendix!4B.!Standard!errors!are!bootstrapped!and!are!in!parentheses.!Asterisks!denote!p?values:!***!p<0.01,!**!p<0.05,!*!p<0.1!2!Ind.!FE!denotes!individual?level!fixed!effects,!HH!FE!denotes!household?level!fixed!effects!!!Note!that!the!variables!for!the!number!of!brothers!and!the!number!of!sisters!do!not!
drop! out! completely! in! the! household?level! fixed! effects! regressions.! The! two!
regressions! show! highly! similar! regression! outcomes,! in! terms! of! both! the!
coefficients! and! the! errors.! One! key! difference! is! in! the! number! of! siblings!
conscripted! variable! for! females.! The! error! coefficients! here! are! slightly! higher! in!
the!household?level!case,!being!significant!at!the!10%!level!(and!just!insignificant!at!
5%).! Note! that,! however,! the! coefficients! for! both! are! within! each! other’s! 95%!
confidence!intervals.!
!
! 35!
I!therefore!conclude!that,!due!to!the!very!small!differences!between!the!two!models!
and!the! fact! that! the! individual?level! fixed!effects!model!seems!to!control! for!more!
(i.e.,!is!more!conservative),!the!individual?level!fixed!effects!model!is!an!appropriate!
choice.!
!
Panel*Attrition*
!
I!now!move!on!to!examine!whether!any!attrition!in!the!panel!could!have!affected!the!
results.!Recall!that!I! largely!base!the!analysis!on!households!that!began!completing!
the!survey!in!2004.!
!
I! test! for! any! attrition! effects! by! recreating! this! analysis! with! households! that!
undertake! the! first! of! their! panel! surveys! in! 1999,! while! still! examining! only! the!
years!of!2003!to!2009.!This!larger!time!range!means!this!sample!is!much!more!likely!
to!feature!attrition!related!effects!than!the!sample!employed!in!the!results!section.!
! !
! 36!
!Table!11:!Females!aged!20!and!above,!1999!households1!
!OLS! RE! FE!
Dependant!Variable!(Dropout=1)! (1)! (2)! (3)!
No.!of!Siblings!Conscripted! 0.0247! 0.0317*! 0.0575*!
!(0.0151)! (0.0180)! (0.0291)!
Age! ?0.0665! ?0.0757! ?0.105!
!(0.0648)! (0.0609)! (0.0724)!
Age!Squared! 0.00122! 0.00145! 0.00210!
!(0.00139)! (0.00130)! (0.00155)!
No.!of!Siblings! ?0.0112! ?0.00953!!
!(0.0200)! (0.0192)!
!Siblings!Left!Household! ?0.0190! ?0.0249**!!
!(0.0213)! (0.0105)!
!Brothers!Left!Household! 0.0754***! 0.0726***!!
!(0.0286)! (0.0268)!
!No.!of!Brothers! ?0.0122! 0.799!!
!(0.0386)! (0.699)!
!Constant! 0.819! 0.730! 1.166!
!(0.702)! (0.749)! (0.823)!
Observations! 280! 280! 280!R?squared! 0.400! 0.374! 0.072!
No.!of!Individuals!!
90! 90!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!The!canton!dummies!for!these!regressions!are!shown!in!appendix!4C.!Standard!errors!are!bootstrapped!and!are!in!parentheses.!Asterisks!denote!p?values:!***!p<0.01,!**!p<0.05,!*!p<0.1!!!In!the!table!above!we!can!see!that!most!of! the!control!coefficients! follow!the!same!
sign!and!significance!levels!as!those!in!Table!3.!For!the!key!variable!in!the!analysis,!
the! number! of! siblings! conscripted,!we! also! see! a! very! similar! coefficient! (see! the!
most!conservative!model!in!column!3!in!the!table!above),!however,!we!also!observe!
a!drop!in!significance!(with!all!three!coefficients!revolving!around!the!10%!level!of!
statistical!significance).!
!
This! suggests! that! there!may!be!some!small!attrition!related!effects! in! the!sample.!
This!means! it! is! important! to!use!households! that!began! completing! the! survey! in!
later!years!to!ensure!these!effects!are!minimised.!
! !
! 37!
Sibling*Support*Robustness!
!
I! now!move! on! to! the! final! test,! this! involves! testing! the! robustness! of! the! sibling!
support!results!by!replacing!the!subjective!measure!of!sibling!support!with!one!that!
is!less!so;!I!compare!those!who!speak!Italian!with!who!do!not!(i.e.,!French!or!German!
speakers).!
!
This! measure! is! employed! because! a! number! of! studies! find! that! those! from! an!
Italian! background! feature! much! closer! family! ties! (and! therefore! rely! more! on!
familial! support! factors)! than! those! from! French! or! German! backgrounds! (Reher!
2005;!Giuliano!2007),!this!means,!when!a!sibling!is!conscripted,!they!are!much!more!
likely! to! experience! a! loss! in! familial! support.! Those! from! French! or! German!
backgrounds!would!rely!more!on!other!forms!of!peer!based!support,!such!as!friends,!
and! should! not! experience! as! large! a! drop! in! support! from! a! sibling! leaving! to!
complete!CMS.!
!
Below,!I!employ!the!same!structure!as!that!of!Table!8,!replacing!the!sibling!support!
factor!with!a!language/cultural!one.!
! !
! 38!
!
!Table!12:!Individuals!aged!18!and!above,!by!primary!language1!
!Italian! French/German!
Dependant!Variable!(Dropout=1)! (1)! (2)!
No.!of!Siblings!Conscripted! 0.341! ?0.0409*!
!(0.291)! (0.0241)!
Age! 0.0109! ?0.0580!
!(1.267)! (0.0491)!
Age!Squared! ?0.000448! 0.00125!
!(0.0259)! (0.00107)!
Constant! ?0.558! 0.753!
!(14.53)! (0.552)!
Observations! 16! 456!R?squared! 0.890! 0.202!
No.!of!Individuals! 7! 199!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1!The!canton!dummies!for!these!regressions!are!shown!in!appendix!2A.!Standard!errors!are!bootstrapped!and!are!in!parentheses.!Asterisks!denote!p?values:!***!p<0.01,!**!p<0.05,!*!p<0.1!!
!
Of!the!three!official!languages,!Italian?speaking!households!are!by!far!the!minority!in!
Switzerland;! this! is! reflected! in! the! very! small! sample! in! Table! 11,! column! 1.! I,!
therefore,!cannot!make!any!reasonable!inference!for!Italian?speaking!households.!!
!
French!and!German!speakers,!however,!make!up!the!majority!in!Switzerland!and,!as!
can!be!seen!in!column!2!of!table!11!above,!the!coefficient!is!negative!(i.e.,!females!in!
lower! income! French! and! German! speaking! houses! benefit! in! terms! of! a! lower!
dropout!propensity)!and!is!significant!at!the!10%!level.!
!
This!table!also!seems!to!help!(partially)!explain!the!discrepancy!between!Cipollone!
and!Rosolia!(2007)!and!Maurin!and!Xenogiani!(2007)!as!differences!between!French!
and!Italian!levels!of!sibling!support.!
!
This!seems! to!corroborate! the! finding! that!peer!support! is!a!strong! factor!because!
the! strong,! positive! effect! of! support! seems! to! diminish! as! the! coefficient! changes!
sign.! This! suggests! that! income! related! factors! are! more! at! play! for! French! and!
German!speakers!in!lower!income!households.!
! 39!
Conclusion!
!
This!thesis!has!two!main!goals,!to!estimate!the!externalities!associated!with!CMS!on!
a!micro!(familial)!level!that!may!be!more!robust!to!country!level!heterogeneity!and!
to!examine!the!sibling!level!externalities!on!education.!!
!
*Summary*of*Results!
!
In!this!thesis,!I!find!that,!overall;!there!are!strong!negative!educational!externalities!
associated! with! compulsory! military! service.! Specifically,! I! find! that! females!
experience! higher! high?school! dropout! propensities!when! an! older!male! sibling! is!
conscripted.!
!
Upon!breaking!down!this!effect,!I!find!that!the!female?based!effects!are!strongest!in!
less! wealthy! families.! I! find! that! much! of! this! can! be! explained! by! decreases! in!
sibling! support! resulting! from! a! sibling! temporarily! leaving! the! household.! This!
effect! seems! to! outweigh! any! financial! effects,! which! are! expected! to! act! in! the!
opposite!direction!(i.e.,!result!in!a!reduction!in!the!propensity!to!dropout).!
!
Males! in! less!wealthy! households! seem! to! experience! a! decrease! in! their! dropout!
propensity!as!a!result!of!a!sibling!being!conscripted.!There!does!not!seem!to!be!any!
evidence!of!the!sibling!support!effect!for!males.!
!
Implications*
!
These!findings!have!a!range!of!implications!for!conscription!policy.!
!
Card! and! Cardoso! (2011)! find! that! CMS! has! a! positive! effects! for! the! poor;! this! is!
attributed!to!gains!in!human!capital!that!may!not!be!otherwise!obtained!due!to!the!
poor’s’!lower!schooling!prospects.!With!this!thesis,!I!extend!this!analysis!to!examine!
the! externalities! associated! with! those! in! less! wealthy! households.! I! find! a!
segregating!effect!where!males!benefit!when!a!sibling! is!conscripted!while! females!
are! hurt.! This! finding,! as! well! as! the! overall! finding! that! conscription! may! hurt!
females! (who!are!not! liable! for! service),! suggests! the! externalities! associated!with!
! 40!
CMS! may! be! deemed! damaging! to! society! and! Informs! the! policy! debate! in!
Switzerland.!
!
As! these! findings! employ! a! household! level! approach,! they! should! be! highly!
generalizable! to! nations! with! similar! cultures! and! conscription! regimes,! most!
notably! Austria.! The! fact! that! sibling! support! factors! are! so! powerful,! even! in!
Switzerland!where!family!ties!are!less!prominent!than!other!countries,!suggests!that!
a! part! of! these! findings! may! be! generalizable! to! other! OECD! countries! that! also!
maintain!CMS!policies.!!
!
The!result!that!social!support!effects!account!for!much!of!the!negative!outcomes!for!
women,!however,!does!not!seem!to!have!clear!policy!implications.!It!is!possible!that!
a!system!where!conscripts!are!not!sent!away! to! live! in!barracks!may!reduce! these!
support!effects,!however,!it!is!not!certain!whether!the!increased!travel,!among!other!
changes,! would! hurt! those! conscripted! more! than! it! helps! reduce! the! negative!
externalities.!!
!
Limitations*and*Future*Research!
!
One!major! omission! in! this! analysis! is! the! effect! of! females! being! conscripted! on!
their!siblings’!educational!outcomes.!As!Switzerland!does!not!conscript!females,!this!
effect!is!not!measurable,!as!such;!the!gender!effects!I!measure!only!show!one!side!of!
the! sibling! relationship.! Israel! is! the! only! country! in! the! OECD! that! conscripts!
females,! depending! on! the! availability! of! suitable! data;! future! research! may! be!
conducted!here.!
!
I!do!not!explicitly!examine!the!‘financial!constraints’!causal!relationship!as!the!data!
on!household!level!income!is!not!of!a!fine!enough!grain!to!do!so!effectively.!Similarly,!
as! the! effects! of! conscription! seem! to! differ! across! income! groups,! a! quantile!
regression! approach! may! be! appropriate! to! examine! these! distributional! effects,!
however,! cleaner! data! must! be! obtained.! This! is! an! option! for! future! research! is!
suitable!data!can!be!found.!
!
! 41!
The!finding!of!a!segregating!effect!where!males!benefit!when!a!sibling!is!conscripted!
while! females! are! hurt! can! be!magnified! if,! as! in! Card! and! Cardoso! (2011),!males!
from!poorer!backgrounds!reap!benefits!from!being!conscripted.!More!research!must!
be!done!to!assess! this!effect! in! full!by!moving!beyond!the!household! level! focus!of!
this!thesis,!into!a!broader!approach!similar!to!the!country!specific!studies!employed!
in!the!direct?effects!literature!(see!Tables!1!and!Appendix!1A).!
!Finally,!since!no!paper!seems!to!have!examined!the!direct!effects!of!CMS!on!the!
conscripts!themselves!for!Switzerland!(as!in!Table!1),!this!would!to!be!a!relatively!
important!area!to!examine!given!the!high!policy!relevance.! !
! 42!
Appendix!!Appendix*1A:*Other*CMS*Effects!!
Papers!Examining!the!Direct!Effect!of!CMS!on!Health!and!Crime!
Paper! Country! Effect!
Bedard!and!Deschenes!(2006)!
U.S.A.! Strong!negative!effect!on!veteran!health!due!to!cigarette!subsidies!
Dobkin!(2009)! U.S.A.! No!effect!on!health!
Autor!et!al.!(2011)! U.S.A.! Negative!effect!on!veteran!health!
Galliani!et!al.!(2011)! Argentina! Concripts!more!likely!to!commit!crimes;!small!negative!effect!on!employment!and!earnings!
!
! 43!
Appendix*2A:*Example*First*Stage*Regression!!
!OLS!
!Conscription!Variable! Coefficient! Standard!Error!Age! 0.649***! 0.130!
Age!Squared! ?0.0152***! 0.00302!Yearly!Percentage!of!Conscripts! 1.069***! 0.312!
Berne! 0.0715**! 0.0288!Lucerne! 0.0391! 0.0353!
Uri! 0.00683! 0.0823!Schwytz! 0.167***! 0.0576!Obwald! 0.140! 0.0893!Nidwald! 0.000624! 0.0937!
Zoug! 0.0865! 0.0614!Fribourg! 0.0462! 0.0406!Soleure! 0.0738! 0.0467!B?Leville! 0.148**! 0.0619!
B?Campagne! 0.0436! 0.0453!Schaffhouse! 0.127! 0.0969!Appenzelle! 0.0849! 0.0823!Appenzelli! ?0.0265! 0.0953!Saintgall! 0.0758**! 0.0344!Grisons! 0.0584! 0.0498!Argovie! 0.0546*! 0.0329!
Thurgovie! 0.0943**! 0.0475!Tessin! 0.172***! 0.0303!Vaud! ?0.0114! 0.0314!Valais! 0.103**! 0.0433!
Neuchatel! ?0.00734! 0.0623!Constant! ?6.960***! 1.386!
Observations! 4,248!!R?squared! 0.021!!!
!! !
! 44!
Appendix*3A:*Canton*Dummies,*Females*Aged*20*and*Above!!
!OLS! RE! FE!
Dependant!Variable!(Dropout=1)! (1)! (2)! (3)!
Berne! 0.00449! 0.00594! ?0.0533!
!(0.0171)! (0.0216)! (0.139)!
Lucerne! 0.0230! 0.0377! 0.350***!
!(0.0206)! (0.0257)! (0.106)!
Schwytz! 0.00534! 0.00245! ?0.0641!
!(0.0208)! (0.0269)! (0.154)!
Obwald! ?0.0239! ?0.0116! ?!
!(0.0486)! (0.0587)!
!Nidwald! 0.0298! 0.0248! ?!
!(0.0353)! (0.0523)!
!Zoug! 0.0315! 0.0319! ?!
!(0.0443)! (0.0515)!
!Fribourg! 0.00189! 0.00723! ?0.0413!
!(0.0257)! (0.0342)! (0.191)!
Soleure! ?0.00773! ?0.00892! ?!
!(0.0247)! (0.0348)!
!B?Leville! 0.00267! 0.00988! ?0.0164!
!(0.0299)! (0.0379)! (0.120)!
B?Campagne! ?0.00619! 0.000507! ?0.00249!
!(0.0250)! (0.0328)! (0.151)!
Appenzelle! 0.00377! 0.00921! ?!
!(0.0404)! (0.0528)!
!Saintgall! ?0.0132! ?0.00957! 0.00402!
!(0.0221)! (0.0273)! (0.0825)!
Grisons! 0.119***! 0.140***! 0.360***!
!(0.0420)! (0.0446)! (0.0885)!
Argovie! 0.0153! 0.0139! ?0.0167!
!(0.0181)! (0.0224)! (0.0870)!
Thurgovie! 0.00348! 0.00833! 0.0179!
!(0.0295)! (0.0384)! (0.117)!
Tessin! 0.268***! 0.273***! ?!
!(0.0271)! (0.0345)!
!Vaud! 0.0200! 0.0189! ?0.0764!
!(0.0176)! (0.0222)! (0.165)!
Valais! 0.0258! 0.0243! ?0.0684!
!(0.0242)! (0.0297)! (0.171)!
Neuchatel! 0.00719! 0.00986! ?0.0511!
!(0.0215)! (0.0255)! (0.183)!
!!
! 45!
Appendix*3B:*Canton*Dummies,*Males*Aged*20*and*Above**
!OLS! RE! FE!
Dependant!Variable!(Dropout=1)! (1)! (2)! (3)!
Berne! ?0.0203! ?0.0274! ?!
!(0.0129)! (0.0212)!
!Lucerne! ?0.0160! ?0.0189! 0.00240!
!(0.0130)! (0.0216)! (0.0357)!
Uri! ?0.0271! ?0.0293! ?!
!(0.0242)! (0.0397)!
!Schwytz! ?0.0131! ?0.0121! ?!
!(0.0249)! (0.0394)!
!Obwald! ?0.0101! ?0.0176! ?!
!(0.0218)! (0.0308)!
!Zoug! ?0.0390! ?0.0458! ?!
!(0.0375)! (0.0517)!
!Fribourg! ?0.0185! ?0.0217! ?!
!(0.0163)! (0.0313)!
!Soleure! ?0.0328! ?0.0376! ?!
!(0.0371)! (0.0581)!
!B?Leville! ?0.0232! ?0.0216! ?!
!(0.0295)! (0.0435)!
!B?Campagne! ?0.0166! ?0.0240! ?!
!(0.0188)! (0.0323)!
!Schaffhouse! 0.148***! 0.0751**! ?0.0138!
!(0.0343)! (0.0370)! (0.0557)!
Appenzelle! ?0.00721! ?0.0103! ?!
!(0.0471)! (0.0787)!
!Appenzelli! ?0.0170! ?0.0274! ?!
!(0.0809)! (0.0911)!
!Saintgall! ?0.0151! ?0.0153! 0.00506!
!(0.0132)! (0.0198)! (0.0505)!
Grisons! 0.0617**! 0.0736*! ?!
!(0.0248)! (0.0389)!
!Argovie! ?0.0202! ?0.0234! ?0.00425!
!(0.0130)! (0.0208)! (0.0674)!
Thurgovie! ?0.0205! ?0.0220! ?!
!(0.0315)! (0.0408)!
!Tessin! ?0.0144! ?0.0172! ?0.00744!
!(0.0193)! (0.0269)! (0.0480)!
Vaud! ?0.0131! ?0.0184! ?!
!(0.0155)! (0.0266)!
!Valais! ?0.0124! ?0.0176! ?!
!(0.0198)! (0.0351)!
!Neuchatel! ?0.0163! ?0.0143! ?!
!(0.0201)! (0.0354)!
!
! 46!
Appendix*3C:*Canton*Dummies,*Females*Aged*15*to*19**
!OLS! RE! FE!
Dependant!Variable!(Dropout=1)! (1)! (2)! (3)!
Berne! ?0.129**! ?0.129**! ?!
!(0.0547)! (0.0547)!
!Lucerne! ?0.0831! ?0.0831! ?!
!(0.0637)! (0.0637)!
!Uri! ?0.308! ?0.308! ?!
!(0.218)! (0.218)!
!Schwytz! ?0.0382! ?0.0382! ?!
!(0.0922)! (0.0922)!
!Obwald! ?0.169! ?0.169! ?!
!(0.139)! (0.139)!
!Zoug! 0.105! 0.105! ?!
!(0.141)! (0.141)!
!Fribourg! 0.455***! 0.455***! 0.200!
!(0.0982)! (0.0982)! (0.465)!
Soleure! 0.148*! 0.148*! ?!
!(0.0836)! (0.0836)!
!B?Leville! ?0.157! ?0.157! ?!
!(0.112)! (0.112)!
!B?Campagne! ?0.187*! ?0.187*! ?!
!(0.112)! (0.112)!
!Schaffhouse! ?0.0966! ?0.0966! ?!
!(0.210)! (0.210)!
!Appenzelle! ?0.137! ?0.137! ?!
!(0.114)! (0.114)!
!Saintgall! ?0.0936! ?0.0936! ?!
!(0.0646)! (0.0646)!
!Grisons! 0.285**! 0.285**! ?!
!(0.113)! (0.113)!
!Argovie! ?0.0629! ?0.0629! ?!
!(0.0606)! (0.0606)!
!Thurgovie! ?0.0250! ?0.0250! ?!
!(0.0800)! (0.0800)!
!Tessin! 0.0878! 0.0878! ?!
!(0.0980)! (0.0980)!
!Vaud! 0.0974*! 0.0974*! ?!
!(0.0576)! (0.0576)!
!Valais! 0.0787! 0.0787! ?!
!(0.0863)! (0.0863)!
!Neuchatel! 0.103! 0.103! ?!
!(0.0635)! (0.0635)!
!!! !
! 47!
Appendix*3D:*Males*Aged*15*to*19**
!OLS! RE! FE!
Dependant!Variable!(Dropout=1)! (1)! (2)! (3)!
No.!of!Siblings!Conscripted! ?0.00737! 0.00648! ?0.00909!
!(0.0329)! (0.0325)! (0.0461)!
Age! 1.342***! 1.278***! 1.227***!
!(0.303)! (0.260)! (0.283)!
Age!Squared! ?0.0402***! ?0.0382***! ?0.0364***!
!(0.00897)! (0.00772)! (0.00845)!
Berne! ?0.0866! ?0.0545! ?!
!(0.0534)! (0.0610)!
!Lucerne! ?0.0772! ?0.0622! ?!
!(0.0643)! (0.0743)!
!Uri! 0.0311! 0.0563! ?!
!(0.105)! (0.125)!
!Schwytz! 0.0431! 0.0214! ?!
!(0.136)! (0.157)!
!Obwald! ?0.189! ?0.160! ?!
!(0.157)! (0.197)!
!Zoug! ?0.190**! ?0.159! ?!
!(0.0936)! (0.114)!
!Fribourg! 0.0673! 0.0625! ?!
!(0.0957)! (0.109)!
!Soleure! ?0.0770! ?0.0125! ?!
!(0.123)! (0.144)!
!B?Leville! ?0.161! ?0.135! ?!
!(0.136)! (0.163)!
!B?Campagne! ?0.110! ?0.113! ?!
!(0.0870)! (0.0994)!
!Schaffhouse! 0.0267! 0.0451! ?!
!(0.0967)! (0.116)!
!Appenzelle! ?0.161! ?0.148! ?!
!(0.172)! (0.204)!
!Saintgall! ?0.0668! ?0.0552! ?!
!(0.0617)! (0.0726)!
!Grisons! ?0.0224! ?0.0262! ?!
!(0.0736)! (0.0862)!
!Argovie! ?0.0383! ?0.0385! ?!
!(0.0560)! (0.0646)!
!Thurgovie! 0.0474! 0.0434! ?!
!(0.0964)! (0.116)!
!Tessin! 0.364***! 0.329***! ?!
!(0.0940)! (0.114)!
!Vaud! 0.209***! 0.202***! ?!
!(0.0635)! (0.0765)!
!
! 48!
Valais! ?0.00842! 0.00789! ?!
!(0.0833)! (0.0929)!
!Neuchatel! ?0.0592! ?0.0604! ?!
!(0.0817)! (0.0916)!
!No.!of!Siblings! 0.0337! 0.0243! ?!
!(0.0264)! (0.0317)!
!Siblings!Left!Household! ?0.0422! ?0.0319! ?!
!(0.0446)! (0.0513)!
!Brothers!Left!Household! 0.00375! ?0.0178! ?!
!(0.0537)! (0.0619)!
!No.!of!Brothers! 0.0511! 0.0542! ?!
!(0.0378)! (0.0454)!
!Constant! ?11.08***! ?10.57***! ?10.07***!
!(2.541)! (2.177)! (2.367)!
Observations! 685! 685! 685!R?squared! 0.130! 0.130! 0.058!
No.!of!Individuals! ?! 279! 279!*!Appendix*3E:*Canton*Dummies,*Family*Incomes*for*Females*Aged*20*and*Above**
!The!Lower/Middle!Class! The!Upper!Class!
Dependant!Variable!(Dropout=1)! (1)! (2)!
Berne! ?0.0598! ?!
!(0.107)!
!B?Leville! ?0.00139! ?!
!(0.0913)!
!B?Campagne! 0.0256! ?!
!(0.115)!
!Saintgall! ?0.00340! ?!
!(0.0661)!
!Grisons! 1.055***! ?!
!(0.0832)!
!Argovie! ?0.00258! ?!
!(0.0665)!
!Vaud! ?0.0658! ?!
!(0.128)!
!Valais! ?0.0717! ?!
!(0.137)!
!Neuchatel! ?0.0127! ?!
!(0.141)!
!Schwytz! ?! ?0.0103!
! !(0.0575)!
Fribourg! ?! 0.00777!
! !(0.0583)!
*
! 49!
*Appendix*3F:*Canton*Dummies,*Family*Incomes*for*Males*Aged*20*and*Above**
!The!Lower/Middle!Class! The!Upper!Class!
Dependant!Variable!(Dropout=1)! (1)! (2)!
Lucerne! ?0.00386! ?!
!(0.0709)!
!Schaffhouse! ?0.0428! ?!
!(0.0594)!
!Argovie! ?0.00106! ?!
!(0.0703)!
!Tessin! ?0.0188! ?!
!(0.0706)!
!**Appendix*3G:*Canton*Dummies,*Emotional*Support*for*Females*Aged*18*and*Above*in*Low/Middle*Income*Families**
!
Decrease!in!Emotional!Support!
No!Change/Increase!in!Emotional!Support!
Dependant!Variable!(Dropout=1)! (1)! (2)!
Berne! ?! 0.106!
! !(0.255)!
Bleville! ?0.0108! 0.0631!
!(0.197)! (0.165)!
Appenzelle! ?! ?0.0679!
! !(0.206)!
Saintgall! ?! 0.0463!
! !(0.156)!
Grisons! ?! 0.826***!
! !(0.170)!
Argovie! ?! ?0.0163!
! !(0.165)!
Thurgovie! ?! 0.0436!
! !(0.242)!
Vaud! ?! ?0.0465!
! !(0.120)!
Neuchatel! ?! ?0.173!
! !(0.190)!
*** *
! 50!
Appendix*3H:*Emotional*Support*for*Females*Aged*18*and*Above*in*High*Income*Families**
!
Decrease!in!Emotional!Support!
No!Change/Increase!in!Emotional!Support!
Dependant!Variable!(Dropout=1)! (1)! (2)!
No.!of!Siblings!Conscripted! 0.0258! 0.112!
!(0.0773)! (0.0756)!
Age! ?0.157! ?0.426**!
!(0.168)! (0.201)!
Age!Squared! 0.00325! 0.00905*!
!(0.00394)! (0.0046)!
Berne! ?0.00475! ?!
!(0.248)!
!Fribourg! ?0.0108! ?!
!(0.273)!
!Schwytz! ?! ?0.256!
! !(0.2)!
Constant! 1.85! 4.743**!
!(1.696)! (2.072)!
Observations! 203! 107!R?squared! 0.094! 0.253!
No.!of!individuals! 109! 78!** *
! 51!
Appendix*3I:*Emotional*Support*for*Males*Aged*18*and*Above*in*Low/Middle*Income*Families**
!
Decrease!in!Emotional!Support!
No!Change/Increase!in!Emotional!Support!
Dependant!Variable!(Dropout=1)! (1)! (2)!
No.!of!Siblings!Conscripted! 0.101! ?0.0392!
!(0.332)! (0.0343)!
Age! ?0.230! ?0.136*!
!(0.314)! (0.0749)!
Age!Squared! 0.00445! 0.00268*!
!(0.00647)! (0.00160)!
Lucerne! ?! 0.0872!
! !(0.145)!
Argovie! ?! ?0.0475!
! !(0.163)!
Tessin! ?! ?0.0348!
! !(0.142)!
Saintgall! ?! 0.183!
! !(0.344)!
Constant! 2.788! 1.731**!
!(3.680)! (0.860)!
Observations! 75! 240!R?squared! 0.056! 0.091!
Number!of!idpers! 60! 129!*** *
! 52!
Appendix*4A:*Canton*Dummies,*Females*Aged*20*and*Above*in*Different*Households**
!OLS! RE! FE!
Dependant!Variable!(Dropout=1)! (1)! (2)! (3)!
Berne! 0.0100! 0.0100! ?0.00761!
!(0.0145)! (0.0145)! (0.0952)!
Lucerne! 0.00404! 0.00404! 0.325***!
!(0.0168)! (0.0168)! (0.0946)!
Schwytz! 0.0180! 0.0180! ?0.0157!
!(0.0189)! (0.0189)! (0.111)!
Obwald! 0.0115! 0.0115! ?!
!(0.0571)! (0.0571)!
!Nidwald! 0.0229! 0.0229! ?!
!(0.0425)! (0.0425)!
!Zoug! 0.0111! 0.0111! ?!
!(0.0280)! (0.0280)!
!Fribourg! 0.0211! 0.0211! ?!
!(0.0217)! (0.0217)!
!Soleure! 0.0131! 0.0131! ?!
!(0.0301)! (0.0301)!
!B?Leville! 0.0207! 0.0207! ?0.000173!
!(0.0371)! (0.0371)! (0.0801)!
B?Campagne! 0.00610! 0.00610! ?!
!(0.0225)! (0.0225)!
!Appenzelle! ?0.0173! ?0.0173! ?!
!(0.0257)! (0.0257)!
!Saintgall! 0.00208! 0.00208! 0.0101!
!(0.0169)! (0.0169)! (0.0731)!
Grisons! 0.141***! 0.141***! 0.331***!
!(0.0283)! (0.0283)! (0.0792)!
Argovie! 0.0168! 0.0168! ?!
!(0.0151)! (0.0151)!
!Thurgovie! 0.00333! 0.00333! 0.0222!
!(0.0275)! (0.0275)! (0.102)!
Tessin! 0.0182! 0.0182! ?!
!(0.0213)! (0.0213)!
!Vaud! 0.00875! 0.00875! ?0.0287!
!(0.0154)! (0.0154)! (0.122)!
Valais! 0.0151! 0.0151! ?0.0120!
!(0.0208)! (0.0208)! (0.137)!
Neuchatel! 0.00822! 0.00822! ?0.00618!
!(0.0154)! (0.0154)! (0.0514)!
** *
! 53!
Appendix*4B:*Canton*Dummies,*Individual*V.S.*Household*Fixed*Effects*for*Individuals*Aged*20*and*Above**!
!Females! Males!
!Ind.!FE! HH!FE! Ind.!FE! HH!FE!
Dependant!Variable! (1)! (2)! (3)! (4)!Berne! ?0.0533! ?0.0118! ?! 0.0142!
!(0.139)! (0.114)!
!(0.0992)!
Lucerne! 0.350***! 0.347***! 0.00240! ?0.00187!
!(0.106)! (0.103)! (0.0357)! (0.141)!
Uri! ?! ?! ?! 0.00356!
! ! ! !(0.121)!
Schwytz! ?0.0641! ?0.0161! ?! ?0.00293!
!(0.154)! (0.128)!
!(0.159)!
Obwald! ?! ?0.00212! ?! ?0.00341!
! !(0.248)!
!(0.147)!
Zoug! ?! ?0.00875! ?! 0.0107!
! !(0.248)!
!(0.259)!
Fribourg! ?0.0413! ?0.00299! ?! 0.00396!
!(0.191)! (0.173)!
!(0.0901)!
Soleure! ?! ?0.0170! ?! 0.00932!
! !(0.164)!
!(0.106)!
B?Leville! ?0.0164! ?0.0121! ?! ?0.00403!
!(0.120)! (0.116)!
!(0.133)!
B?Campagne! ?0.00249! ?0.00180! ?! 0.0118!
!(0.151)! (0.147)!
!(0.0796)!
Schaffhouse! ?! ?! ?0.0138! ?0.0128!
! ! !(0.0557)! (0.0730)!
Appenzelle! ?! 0.0220! ?! 0.00734!
! !(0.240)!
!(0.125)!
Appenzelli! ?! ?! ?! 0.00997!
! ! ! !(0.139)!
Saintgall! 0.00402! 0.00471! 0.00506! 0.00978!
!(0.0825)! (0.0801)! (0.0505)! (0.0662)!
Grisons! 0.360***! 0.353***! ?! 1.25e?05!
!(0.0885)! (0.0859)!
!(0.105)!
Argovie! ?0.0167! ?0.0120! ?0.00425! ?0.00795!
!(0.0870)! (0.0846)! (0.0674)! (0.0884)!
Thurgovie! 0.0179! 0.0113! ?! ?0.0158!
!(0.117)! (0.111)!
!(0.138)!
Tessin! ?! ?0.0172! ?0.00744! ?0.00848!
! !(0.248)! (0.0480)! (0.0573)!
Vaud! ?0.0764! ?0.0279! ?! ?0.00331!
!(0.165)! (0.143)!
!(0.0894)!
Valais! ?0.0684! ?0.0200! ?! 1.25e?05!
!(0.171)! (0.150)!
!(0.105)!
Neuchatel! ?0.0511! ?0.00903! ?! ?0.00741!
!(0.183)! (0.164)!
!(0.168)!
! 54!
Appendix*4C:*Canton*Dummies,*Females*Aged*18*and*Above,*1999*Households***
!OLS! RE! FE!
Dependant!Variable!(Dropout=1)! (1)! (2)! (3)!
Berne! 0.0545! 0.0414! ?0.0632!!! (0.0433)! (0.0524)! (0.220)!
Lucerne! 0.0600*! 0.0732*! 0.350**!!! (0.0337)! (0.0421)! (0.142)!
Schwytz! ?0.0539! ?0.0403! ?!!! (0.0618)! (0.0740)!
!Obwald! ?0.0823! ?0.0746! ?!!! (0.0708)! (0.0903)!
!Nidwald! 0.0912*! 0.0664! ?!!! (0.0533)! (0.0811)!
!Zoug! 0.0708! 0.0592! ?!!! (0.0730)! (0.0927)!
!Fribourg! 0.0625! 0.0480! 0.0244!!! (0.0389)! (0.0498)! (0.154)!
Soleure! 0.0258! 0.0167! ?!!! (0.0342)! (0.0454)!
!B?Leville! 0.00679! 0.0113! ?0.0168!!! (0.0423)! (0.0539)! (0.123)!
B?Campagne! 0.0713! 0.0603! ?!!! (0.0436)! (0.0555)!
!Appenzelle! 0.00985! 0.0130! ?!!! (0.0531)! (0.0672)!
!Saintgall! ?0.0286! ?0.0299! 0.0113!!! (0.0408)! (0.0509)! (0.113)!
Grisons! 0.0581! 0.118*! 0.364***!!! (0.0576)! (0.0612)! (0.118)!
Argovie! 0.0651**! 0.0492! ?0.0165!!! (0.0318)! (0.0386)! (0.165)!
Thurgovie! 0.0699*! 0.0640! ?!!! (0.0414)! (0.0546)!
!Tessin! 0.380***! 0.379***! ?!!! (0.0427)! (0.0526)!
!Vaud! 0.0653**! 0.0486! ?0.0184!!! (0.0327)! (0.0393)! (0.0680)!
Valais! 0.0850**! 0.0621! ?!!! (0.0396)! (0.0480)!
!Neuchatel! 0.0509! 0.0406! 0.0111!!! (0.0355)! (0.0419)! (0.128)!
** *
! 55!
Appendix*4D:*Canton*Dummies,*Individuals*Aged*18*and*Above,*by*Primary*Language*!
!Italian! French/German!
Dependant!Variable!(Dropout=1)! (1)! (2)!
Berne! ?! 0.0191!
! !(0.204)!
B?Leville! ?! ?0.0292!
! !(0.175)!
B?Campagne! ?! ?0.0495!
! !(0.223)!
Appenzelle! ?! ?0.0149!
! !(0.167)!
Saintgall! ?! 0.0237!
! !(0.124)!
Grisons! ?! 0.942***!
! !(0.139)!
Argovie! ?! ?0.0234!
! !(0.126)!
Thurgovie! ?! 0.0231!
! !(0.176)!
Vaud! ?! 0.0169!
! !(0.244)!
Valais! ?! 0.0391!
! !(0.262)!
Neuchatel! ?! ?0.0236!
! !(0.272)!
Fribourg! 0.00567! ?!
!(0.360)!
!! !
! 56!
!References!
!
Albrecht,! J.!W.,! P.?A.! Edin,! et! al.! (1999).! "Career! Interruptions! and! Subsequent!Earnings:! A!Reexamination!Using! Swedish!Data."! The! Journal! of!Human!Resources!34(2):!294?311.!
Angrist,! J.! D.! and! S.! H.! Chen! (2011).! "Schooling! and! the! Vietnam?Era! GI! Bill:!Evidence! from! the! Draft! Lottery."! American! Economic! Journal:! Applied!Economics!3(2):!96?118.!
Angrist,!J.!D.!and!A.!B.!Krueger!(1992).!"Estimating!the!Payoff!to!Schooling!Using!the! Vietnam?Era! Draft! Lottery."! National! Bureau! of! Economic! Research!Working!Paper!Series!No.!4067.!
Angrist,! J.! D.! and! A.! B.! Krueger! (1995).! "Split?Sample! Instrumental! Variables!Estimates! of! the! Return! to! Schooling."! Journal! of! Business! &! Economic!Statistics!13(2):!225?235.!
Bauer,! T.! K.,! S.! Bender,! et! al.! (2012).! "Evaluating! the! labor?market! effects! of!compulsory!military!service."!European!Economic!Review(Forthcoming):!1?16.!
Bedard,!K.!and!O.!Deschenes!(2006).!"The!Long?Term!Impact!of!Military!Service!on! Health:! Evidence! from! World! War! II! and! Korean! War! Veterans."!American!Economic!Review!96(1):!176?194.!
Behrman,! J.! R.! and! P.! Taubman! (1989).! "Is! Schooling! "Mostly! in! the! Genes"?!Nature?Nurture! Decomposition! Using! Data! on! Relatives."! Journal! of!Political!Economy!97(6):!1425?1446.!
Björklund,!A.!and!M.!Jäntti!(1997).!"Intergenerational!Income!Mobility!in!Sweden!Compared! to! the!United!States."!The!American!Economic!Review!87(5):!1009?1018.!
Black,! S.! E.,! P.! J.! Devereux,! et! al.! (2005).! "The!More! the!Merrier?! The! Effect! of!Family! Composition! on! Children's! Education."! The! Quarterly! Journal! of!Economics!120(2):!669?700.!
Bommier,! A.! and! S.! Lambert! (2004).! "Human! capital! investments! and! family!composition."!Applied!Economics!Letters!11(3):!193?196.!
Buonanno,!P.! (2007).! "Long?term!effects!of! conscription:! lessons! from! the!UK."!University!of!California,!Berkeley!and!University!of!Bergamo,!Mimeo.!
Butcher,! K.! F.! and! A.! Case! (1994).! "The! Effect! of! Sibling! Sex! Composition! on!Women's! Education! and! Earnings."! The!Quarterly! Journal! of! Economics!109(3):!531?563.!
Card,! D.! and! A.! R.! Cardoso! (2011).! "Can! Compulsory! Military! Service! Raise!Civilian! Wages?! Evidence! from! the! Peacetime! Draft! in! Portugal."!American!Economic!Journal:!Applied!Economics(Forthcoming).!
Card,! D.! and! T.! Lemieux! (2001).! "Going! to! College! to! Avoid! the! Draft:! The!Unintended!Legacy!of!the!Vietnam!War."!The!American!Economic!Review!91(2):!97?102.!
Cipollone,!P.!and!A.!Rosolia!(2007).!"Social!Interactions!in!High!School:!Lessons!from!an!Earthquake."!The!American!Economic!Review!97(3):!948–965.!
Di!Pietro,!G.!(2009).!"Changes!in!the!Extent!of!Inequality!in!Access!to!University!Education!in!Italy."!Rivista!di!Politica!Economica!98(1):!175?213.!
! 57!
Duncan,!G.! J.,! J.!Boisjoly,! et!al.! (2001).! "Sibling,!Peer,!Neighbor,! and!Schoolmate!Correlations! as! Indicators! of! the! Importance! of! Context! for! Adolescent!Development."!Demography!38(3):!437?447.!
Fletcher,!J.,!M.!M.!Seltzer,!et!al.!(2012).!"A!Sibling!Death!in!the!Family:!Common!and!Concequential."!Demography(Forthcoming):!1?47.!
Galiani,!S.,!M.!A.!Rossi,!et!al.!(2011).!"Conscription!and!Crime:!Evidence!from!the!Argentine!Draft!Lottery."!American!Economic!Journal:!Applied!Economics!3(2):!119?136.!
Garg,! A.! and! J.!Morduch! (1998).! "Sibling! rivalry! and! the! gender! gap:! Evidence!from! child! health! outcomes! in!Ghana."! Journal! of! Population!Economics!11(4):!471?493.!
Giuliano,! P.! (2007).! "Living! Arrangements! in! Western! Europe:! Does! Cultural!Origin!Matter?"!Journal!of!the!European!Economic!Association!5(5):!927?952.!
Grenet,! J.,! R.!A.!Hart,! et! al.! (2011).! "Above! and!beyond! the! call.! Long?term! real!earnings! effects! of! British! male! military! conscription! in! the! post?war!years."!Labour!Economics!18(2):!194?204.!
Hauser,!R.!M.!and!H.?H.!D.!Kuo!(1998).!"Does!the!Gender!Composition!of!Sibships!Affect! Women's! Educational! Attainment?"! The! Journal! of! Human!Resources!33(3):!644?657.!
Hole,! A.! R.! (2006).! "Calculating! Murphy?Topel! variance! estimates! in! Stata:! A!simplified!procedure."!Stata!Journal!6(4):!521?529.!
Imbens,!G.!and!W.!van!der!Klaauw!(1995).!"Evaluating!the!Cost!of!Conscription!in!the!Netherlands."! Journal! of!Business!&!Economic! Statistics!13(2):! 207?215.!
Inoue,!A.!and!G.!Solon!(2010).!"Two?Sample!Instrumental!Variables!Estimators."!Review!of!Economics!and!Statistics!92(3):!557?561.!
Kaestner,! R.! (1996).! "Are! Brothers! Really! Better?! Sibling! Sex! Composition! and!Educational! Achievement! Revisited."! National! Bureau! of! Economic!Research!Working!Paper!Series!No.!5521.!
Kaestner,! R.! (1997).! "Are! brothers! really! better?! Sibling! sex! composition! and!educational! achievement! revisited."! The! Journal! of! Human! Resources!32(2):!250?284.!
Keller,! K.,! P.! Poutvaara,! et! al.! (2010).! "Does! a! Military! Draft! Discourage!Enrollment! in!Higher!Education?"!FinanzArchiv:!Public!Finance!Analysis!66(2):!97?120.!
Maurin,! E.! and!T.! Xenogiani! (2007).! "Demand! for! Education! and! Labor!Market!Outcomes:! Lessons! from! the! Abolition! of! Compulsory! Conscription! in!France."!The!Journal!of!Human!Resources!42(4):!795?819.!
Morduch,! J.! (2000).! "Sibling!Rivalry! in!Africa."!The!American!Economic!Review!90(2):!405?409.!
Murphy,! K.!M.! and! R.! H.! Topel! (2002).! "Estimation! and! Inference! in! Two?Step!Econometric! Models."! Journal! of! Business! &! Economic! Statistics! 20(1):!88?97.!
North,! R.! J.,! C.! J.! Holahan,! et! al.! (2008).! "Family! Support,! Family! Income,! and!Happiness:!A!10?Year!Perspective."! Journal! of! Family!Psychology!22(3):!475?483.!
! 58!
Paloyo,! A.! R.! (2010).! Compulsory! Military! Service! in! Germany! Revisited,!Rheinisch?Westfälisches! Institut! für! Wirtschaftsforschung,! Ruhr?Universität!Bochum,!Universität!Dortmund,!Universität!Duisburg?Essen.!
Plug,!E.!and!W.!Vijverberg!(2003).!"Schooling,!Family!Background,!and!Adoption:!Is! It!Nature!or! Is! It!Nurture?"! Journal!of!Political!Economy!111(3):!611?641.!
Qureshi,!J.!A.!(2012).!"Additional!Returns!to!Investing!in!Girls’!Education:!Impact!on!Younger!Sibling!Human!Capital."!University!of!Chicago,!Mimeo.!
Reher,! D.! (2005).! Family! Ties! in! Western! Europe.! Strong! Family! and! Low!Fertility:!A!Paradox?!G.!Zuanna!and!G.!Micheli,!Springer!Netherlands.!14:!45?76.!
Stock,!J.!H.!and!M.!W.!Watson!(2003).!Introduction!to!Econometrics.!Boston,!MA,!Addison?Wesley.!
Wight,!R.!G.,!A.!L.!Botticello,!et!al.!(2006).!"Socioeconomic!Context,!Social!Support,!and! Adolescent! Mental! Health:! A! Multilevel! Investigation."! Journal! of!Youth!and!Adolescence!35(1):!109?120.!
!!