87
No dic jou nal o wo king li e s udies Volume 7 ❚ Numbe 2 ❚ June 2017
Time o Displacemen as a P edic o o Re-employmen
❚❚ A ja Jolkkonen
Senio Resea che , Uni e si y o Eas e n Finland, Ka elian Ins i u e, Finland
❚❚ Pe i Kois inen1
Eme i us P o esso , Uni e si y o Tampe e, Facul y o Social Sciences, Finland
❚❚ A ja Ku inen
Senio Resea che , Uni e si y o Eas e n Finland, Ka elian Ins i u e, Finland
❚❚ Liudmila Lipiäinen
Resea che , Uni e si y o Tampe e, Wo k Resea ch Cen e (WRC), Finland
❚❚ Tapio Nummi
Uni e si y Lec u e , Uni e si y o Tampe e, Facul y o Na u al Sciences, Finland
❚❚ Pekka Vi anen
Docen , Uni e si y o Tampe e, Facul y o Social Sciences, Finland
ABSTRACT
P e ious s udies add essing plan closu es sugges ha he eemploymen o wo ke s a ies acco ding
o he economic cycle a he ime o displacemen . Howe e , he e a e ew s udies on he selec i i y
o labo demand and eemploymen in he long un. This s udy applied ajec o y analyses in explo -
ing he 7-yea labo ma ke a achmen o wo ke s displaced in plan closu es in Finland du ing he
ecession in 1992, he boom in 1997, and slow economic g ow h in 2003. Resul s o he mul inomial
logis ic eg ession analyses indica ed ha hose displaced du ing he ecession assumed he ajec o y
o s ong employmen signi ican ly less commonly han hose displaced unde be e economic condi-
ions. Mo eo e , a signi ican in e ac ion be ween age and yea o displacemen was ound, whe eas
he gende , educa ion-, and indus y ela ed isks u ned ou o be imeindependen . The s udy con ib-
u es o he sociological unde s anding o economic es uc u ing and how di e en imes ame he
indi idual’s oppo uni ies o gain e-employmen .
Keywo ds
displacemen / economic es uc u ing / job loss / longi udinal da a analysis / e-employmen /
ajec o y analysis
DOI
To be announced
1 Co esponding au ho : E-mail: [email p o ec ed].
88 Time o Displacemen as a P edic o o Re-employmen A ja Jolkkonen e al.
In oduc ion
The ecen his o y o indus ial socie ies can be ead h ough he con inuous es uc-
u ing o economies. This es uc u ing a ec s he de elopmen o labo ma ke s and
he wel a e o socie ies. In hei seminal s udy on deindus ializa ion and unemploy-
men in he OECD (The O ganisa ion o Economic Co-ope a ion and De elopmen )
coun ies o e he las 30 yea s, Kollmeye and Pichle (2013: 786) discuss whe he
deindus ializa ion has been an impo an bu less ecognized cause o he unemploy-
men p oblems expe ienced by a luen coun ies. They ound ha despi e he s eng h-
ening o esea ch in his a ea, i is s ill unclea i he e is a causal ela ion be ween dein-
dus ializa ion and unemploymen . One o he unanswe ed opics is how he in e play
o economic shocks and s uc u al labo ma ke adjus men may a ec employmen , in
pa icula he longe - e m employmen ajec o ies o edundan wo ke s.
P e ious esea ch on e-employmen a e plan closu es sugges s ha losing one’s
job is mo e damaging du ing a ecession han du ing an economic boom (Fa be , 2005,
2011; F ede iksen & Wes e ga d-Nielsen, 2007; Ga ell e al., 2010; Hall e al., 1995;
Jolkkonen e al., 2014; Kle ze , 1998). Fu he mo e, hose who g adua e in imes o
ecession o a e ma ched o he w ong indus y in hei i s job a e much mo e ulne -
able o business cycle a ia ion. In addi ion, ins i u ional ac o s a ec he employmen
oppo uni ies o displaced wo ke s, and policy di e ences be ween coun ies can lead
o la ge di e ences in employmen ou comes (Cu le e al., 2014; Houle & Auden ode,
1995; Hu unen e al., 2011; Liu e al., 2016; Ne ke e al., 2016).
E idence om p e ious Finnish popula ion-wide s udies sugges s ha s uc u al
ac o s and economic cycles a ec he e-employmen and income de elopmen o dis-
placed wo ke s (Appelq is , 2007; Ve ho, 2008). Those displaced du ing he ecession
o he ea ly 1990s su e ed om a subs an ial and pe sis en loss o ea nings, and he
e ec o displacemen was e y he e ogeneous and s onges a he lowe end o he
ea nings scale (Ko keamäki & Kyy ä, 2014). S udies applying ajec o y analyses also
demons a e di e en ia ion in he e-employmen o displaced wo ke s; he sha e o
hose who assume he ajec o y o s able employmen was qui e high (Jolkkonen e al.,
2014), al hough no as high as he co esponding sha e o employees who e ained hei
posi ions ollowing downsizings (Jolkkonen e al., 2017).
Despi e he undispu ed p og ess o he esea ch, he e a e s ill many ques ions o
be answe ed abou he ac o s a ec ing he long- e m employmen o displaced wo k-
e s. Ou s anding ques ions conce n how he s a e o he business cycle a he ime o
displacemen is ela ed o he employmen ajec o ies in he long un, and i his ela-
ionship depends on indi idual ac o s. I is also impo an o sepa a e employees expe-
iencing plan closu es om hose expe iencing downsizings, and o ake in o accoun
a ia ion in he s abili y o he employees’ ca ee s p io o displacemen .
The p esen s udy ex ends ac oss a pe iod o 20 yea s and co e s h ee di e -
en kinds o economic cycle in Finland. Rep esen a i e egis e based da a allow us o
o e come he limi a ion o s udies based on single plan closu es and use longi udinal
esea ch se ings and ad anced s a is ics ha encompass he s uc u al and indi idual
ac o s con ibu ing o he labo ma ke in eg a ion o displaced wo ke s. The main
objec i e is o e eal i he displaced wo ke s’ isk o alling on o di e en employmen
ajec o ies depends on he ime o plan closu e, aking in o accoun hei age, gende ,
le el o educa ion, and he indus y o he p e ious wo kplace. In addi ion, we analyze
No dic jou nal o wo king li e s udies Volume 7 ❚ Numbe 2 ❚ June 2017 89
i he impo ance o unemploymen as he eason o alling on o ajec o ies o poo
employmen depends on ime.
The esul s clea ly e i y ha he in eg a i e capaci y o he labo ma ke is s ong,
as e en in 1992, a a ime o deep ecession, hal o he displaced employees we e assigned
o he g oup o con inuously s ong employmen , and in 1997 and 2003, he co e-
sponding p opo ions we e abou 80%. The e-employmen was also selec i e, p o id-
ing ewe oppo uni ies o olde wo ke s o all on o a s ong employmen ajec o y;
mo eo e , he e was an in e ac ion be ween age and ime o displacemen , indica ing
ha olde wo ke s displaced du ing he ecession had pa icula ly poo oppo uni ies.
By con as , he associa ions o gende , educa ion, and indus y wi h e-employmen
we e ime-independen , indica ing ha pe manen changes in he indus y s uc u es
and associa ed labo o ce demands do no explain he esul .
The s eng hs o his esea ch lie in quali y o he da a and in no el me hodological
solu ions. Mo eo e , we a gue ha Finland is an in e es ing case o s udies o economic
es uc u ing, because as an open and expo o ien ed economy, i has been sensi i e o
ex e nal shocks and labo ma ke es uc u ings ha ha e led o majo layo s, which
in u n ha e necessi a ed employmen policy e o ms. Fu he mo e, Finland p o ides
an en i onmen conduci e o in ensi e social scien i ic esea ch in o he eemploymen
op ions o displaced wo ke s.
Li e a u e e iew and heo e ical amewo k
Al eady back in he 1970s, Pe e Ba h p esen ed an in e es ing pe spec i e on economic
cycles and labo ma ke adjus men . Acco ding o his heo e ical o mula ions and ime
se ies analysis o displacemen a es, he a gued ha di e en i ms may espond di e -
en ly o iden ical changes in hei own p oduc ma ke when gene al labo ma ke condi-
ions di e . The employmen a e o displaced wo ke s is highe when he displacemen
akes place in imes o economic expansion han in imes o ecession, ye he demand
o labo a ies conside ably by i m and indus y ype. The selec i e demand o labo
may lead o a si ua ion whe e he labo demand is no a ge ed a hose who los hei
jobs, because hei skills do no ma ch he needs o he labo ma ke (Ba h, 1971: 452).
This is in acco dance wi h Lilien’s (1982) ‘sec o al shi ’ hesis, acco ding o which luc-
ua ions in demand ac oss sec o s accoun o a subs an ial ac ion o he a ia ion in
unemploymen and he declining sec o s o he economy ypically inc ease unemploy-
men a es because he ealloca ion o wo ke s ac oss dispa a e indus ies occu s slowly.
The imes o ecession a e ha de o job-seeke s bu easie o ec ui ing i ms because
he e a e mo e job applican s (Chia ini & Piselli, 2001; Kollmeye & Pichle , 2013: 786;
Ne ke e al., 2016).
The human capi al heo y has been commonly applied o s udies o job displace-
men . F om a human capi al pe spec i e, he dis ibu ion o employmen , ea nings, and
income a e displacemen can be ied di ec ly o di e ences in le els o educa ion and
aining (Becke , 1975; Koebe & W igh , 2006; O mis on, 2014). Human capi al is
o en di ided in o wo ca ego ies: gene al, which is widely applicable o many employ-
e s, and speci ic, which can be unique o a ce ain employe o job (Kle ze , 1998). Gen-
e al human capi al can be measu ed by he le el o educa ion and i is ans e able ac oss
employe s. In addi ion o educa ion, he e is e idence ha gende , age, and he indus y
90 Time o Displacemen as a P edic o o Re-employmen A ja Jolkkonen e al.
o he p e ious employe a e impo an ac o s o e-employmen a e job loss. How-
e e , i has no been s udied how business cycles and indi idual ac o s oge he in lu-
ence he de elopmen o employmen ajec o ies in he long e m.
Ga ell e al. (2010) used Swedish da a om 1986 o 2002 o s udy cyclical pa e ns
o job and wo ke lows: hey ound ha job ealloca ion is coun e cyclical. Howe e ,
when es ima ing co ela ions o g oups acco ding o educa ion, job ealloca ion was
coun e cyclical o he lowes educa ed and acyclical o he highe educa ed g oups.
Con a y o job lows, wo ke lows we e acyclical wi h a s ong p ocyclical pa e n o
hi ing a es, and he e was an acyclical pa e n in sepa a ion a es (Ga ell e al., 2010).
Fu he mo e, ins i u ional ac o s and he s uc u e and unc ion o he local labo ma -
ke a ec mass layo s and he e-employmen oppo uni ies o displaced wo ke s (Fa be ,
2011, 2005; F ede iksen & Wes e gaa dNielsen, 2007; Jolkkonen & Ku inen, 2009).
Exis ing s udies on job sepa a ion and ealloca ion p o ide e idence ha many o
hose who a e laid o by hei long e m employe su e income losses e en i hey a e
e-employed. Ko keamäki and Kyy ä (2014) disco e ed ha displaced wo ke s su e
om a subs an ial and pe sis en loss o ea nings, women a e subjec o la ge losses,
and he e ec o displacemen is e y he e ogeneous and s onges a he lowe end o
he ea nings scale. Displacemen no only causes la ge losses in expec ed ea nings bu
also aises he ins abili y o u u e ea nings. Ve ho (2008) compa ed he ea nings o
men who los hei jobs due o a plan closu e du ing he Finnish ecession o he ea ly
1990s wi h he ea nings o hose who emained in employmen . A e a 7-yea ollow-
up, he ound a signi ican ea nings penal y o hose who we e unemployed as well as o
hose whose employmen eco e ed s eadily a e he ecession. The lowe incomes we e
explained by he educ ion o employmen mon hs and a lowe wage le el.
When examining he sho - and long- e m e ec s o displacemen on wo ke s’ employ-
men and ea nings in No way, Hu unen e al. (2011) ound ha , on a e age, displace-
men inc eases he p obabili y o lea ing he labo o ce by 30%. The d opou a e om
he labo o ce among he olde wo ke s was pa icula ly high in he i s yea s ollowing
displacemen . The long- e m e ec was he g ea es o middleaged wo ke s (aged 45–57),
who we e he mos likely o emain ou side he labo o ce 7 yea s a e displacemen .
Fa be (2011) used he Displaced Wo ke s Su ey (DWS) o in es iga e he inci-
dence and consequences o job loss in he USA om 1981 o 2009. Acco ding o his
esul s, i was clea ha hose who los hei jobs du ing he 2007–2009 pe iod eco ded
a highe a e o job loss and g ea e di icul ies in eemploymen when compa ed o
o he pe iods o economic de elopmen . Nea ly one in six wo ke s epo ed ha ing
los a job in he 2007–2009 pe iod, and hese indi iduals also epo ed e y low a es
o eemploymen , di icul y in inding ull- ime employmen , and subs an ial ea nings
losses. Fa be disco e ed ha he e we e s ong di e ences in pos displacemen labo
o ce s a us by age. P ime-aged (25–54 yea s o age) wo ke s who los hei job had he
s onges a achmen o he labo o ce, while olde wo ke s (55–64 yea s o age) we e
subs an ially mo e likely o be ou o he labo o ce. Ne e heless, all age g oups su -
e ed du ing he 2007–2009 ecession (Fa be , 2011: 10–13).
O mis on (2014) ex ended he scope o ques ions and asked why displaced wo ke s
lose so much alue in he labo ma ke . Using ep esen a i e occupa ional da a om
US Depa men o Labo and he DWS su eys, he ound ha displaced wo ke s ha e
g ea e di icul ies in imes o ecession han economic g ow h, bu he esul s sugges ha
a link be ween occupa ion-speci ic human capi al ans e abili y and pos displacemen
No dic jou nal o wo king li e s udies Volume 7 ❚ Numbe 2 ❚ June 2017 91
ea nings upon e-employmen is independen o he s a e o he mac o-economy (O m-
is on, 2014: 375).
S e ens (1997) examined he long- e m wage and ea nings losses o displaced wo k-
e s using he Panel S udy o Income Dynamics. She disco e ed ha much o he pe sis-
en e ec o job loss can be explained by addi ional job losses in he yea s ollowing an
ini ial displacemen , and ha hose losing jobs due o downsizing a e somewha mo e
likely o expe ience addi ional displacemen s compa ed o hose losing hei job due o
plan closu e.
P e ious esea ch on e-employmen a e mass layo s has shown ha he isk o
ending up unemployed is highes among young wo ke s and lowes among he oldes
wo ke s. Howe e , i is easie o young people o become e-employed, whe eas olde
wo ke s o en become long- e m unemployed o mo e ou side he labo ma ke (Elia-
son & S o ie, 2009; Hall e al., 1995; Jolkkonen & Ku inen 2009; Ko keamäki &
Kyy ä, 2014; Ve ho, 2008). Olde wo ke s lose mo e han hei younge colleagues in
layo s because hei e-employmen op ions end o be limi ed and hey end o expe i-
ence conside able wage losses. On he o he hand, olde wo ke s and hose wi h long
wo k his o ies a e less likely o be displaced because layo s a e o en based on he ule
o ‘las in, i s ou ’ (Boockmann & S e es, 2007; Kuhn, 2002). I has also been shown
ha highly educa ed and whi e-colla wo ke s ha e be e eemploymen oppo uni ies
han he unskilled and less educa ed, despi e he ac ha displacemen among educa ed
wo ke s is inc easing (Fa be , 2011; Jolkkonen e al., 2012; Kle ze , 1998; Kuhn, 2002;
O mis on, 2014). The esul s o Nedelkoska e al. (2015) om Ge many show ha a
quali ica ion misma ch is pa icula ly common among well-quali ied employees, a leas
du ing he i s yea s o e-employmen .
The e ec s o gende and leng h o wo k his o y a e mo e complex han hose o
age and educa ion. In many coun ies, men a e mo e likely o be displaced han women,
bu in ega d o ca ee s abili y and ea nings, women lose e en mo e han men (Kuhn,
2002). Mo e equen ly han men, women mo e ou side he labo o ce in o domes ic
wo k, end up seeking u he educa ion, o change occupa ion i hey canno ind a new
job in he ield (Fa be , 2011; Jolkkonen & Ku inen 2009). E idence om Denma k
shows ha he e a e no gende di e ences in job sepa a ion p obabili ies o employees
in simila wo kplaces, bu he e is a clea di e ence in job s abili y because, mo e o en
han men, women wo k in jobs ha a e mo e insecu e, and he ansi ion models o
men and women a e di e en . Women also mo e commonly mo e om employmen
o unemploymen o ou o he labo o ce, and hey a e less likely o make job- o-job
ansi ions (F ede iksen, 2008: 915).
S udies p o e ha employe s a e e y selec i e in e ms o age, educa ion, gende ,
and social cha ac e is ics when ec ui ing new labo , bu , in he case o hei o me
wo ke s, he c i e ia o ec ui men a e mo e lexible (Kling all, 1998). I seems ha
displaced wo ke s who lose hei jobs in plan closu es ha e a be e epu a ion wi h
employe s han hose who lose hei jobs in wo k o ce educ ions because, i s , employ-
e s belie e ha he eason o dismissal was in gene al economic, and second, employe s
assume ha awkwa d indi idual cha ac e is ics and low p oduc i i y we e no he cause
o dismissal (Jolkkonen e al., 2012). These indings a e impo an when se in he wide
con ex o he es uc u ing and ealloca ion o labo .
Taken oge he , his e iew o he exis ing li e a u e clea ly demons a es ha e-
employmen depends on s uc u al, ins i u ional, and indi idual ac o s. The ime o

92 Time o Displacemen as a P edic o o Re-employmen A ja Jolkkonen e al.
displacemen and he ex en and ype o es uc u ing explains he di e en ia ion o
he e-employmen oppo uni ies o displaced wo ke s. In imes o ecession and eco-
nomic g ow h, he c i e ia o ec ui men and he selec i i y o in lows and ou lows
om unemploymen a y. The e a e, howe e , sho comings in he esea ch e idence.
Fo ins ance, many s udies pool employees in hei analyses i espec i e o he ype o
displacemen (i.e., plan closu e o downsizing o single layo s), hus disca ding an
impo an p edic o o u u e employmen . This is he eason why we concen a e on
plan closu es in his s udy. Ano he impo an bu seldom con olled o p edic o is he
s abili y o he wo ke ’s con ac in he se ice o he plan ha has closed; in he p esen
s udy, ela i ely s able employmen is a p e equisi e o inclusion in he sample. Mo e-
o e , we cons uc an employmen ela ed indica o o labo ma ke a achmen and use
s a is ical me hods ha ha e no been applied be o e in his domain.
Based on economic heo ies and p e ious s udies on he eemploymen o displaced
wo ke s, he p esen s udy aims o p o ide suppo o he hypo hesis ha labo ma -
ke in eg a ion is weake when he displacemen occu s du ing an economic ecession
han du ing a pe iod o eco e y and g ow h. Mo eo e , leaning on he sec o al shi
hypo hesis, we assume ha his weakness o in eg a ion is due o pe manen changes in
indus y s uc u es and he associa ed selec i i y o labo o ce demand. We expec o
ind ha he less educa ed and hose wo king in p ima y and seconda y p oduc ion a e
mo e ulne able o weak in eg a ion when displaced du ing a ecession.
socioeconomic con ex
To disco e he e ec o s uc u al ac o s and cyclical and empo al ime, we pooled he
da a o wo ke s displaced due o plan closu es du ing h ee di e en ypes o business
cycle, co esponding o he yea s 1992, 1997, and 2003. The cyclical de elopmen o
he economy is usually illus a ed by he indica o s o g oss domes ic p oduc ion (GDP),
employmen a e, and unemploymen a e (Fa be , 2011; F ede iksen & Wes e gaa d
Nielsen, 2007; Ko keamäki & Kyy ä, 2014). These indica o s show ha 1992, 1997,
and 2003 each ep esen a di e en mac oeconomic cycle (see Fig. 1).
In 1992, Finland was in an excep ionally deep ecession wi h high unemploymen .
Be ween 1991 and 1993, he unemploymen a e inc eased om 8% o almos 20%.
In 1997, he unemploymen a e was s ill abou 16%, bu i dec eased o 11% in 2003.
The yea 1997 was a ime o economic boom a e he ea lie ecession. The speed o
Finland’s economic g ow h slowed a e 2001, al hough he demand o labo and he
employmen a e we e s ill inc easing and he numbe o plan s and hei employees
we e a a highe le el in 2003 han du ing 1992 and 1997. Howe e , s uc u al changes,
such as he apid de elopmen o echnology and changes in occupa ional s uc u e,
we e cha ac e is ic o he beginning o he 2000s (Asplund e al., 2012; S a is ical Yea -
books o Finland, 1990–2013). Al hough 2003 was cha ac e ized by slow economic
g ow h, he e was demand o labo , because in spi e o he closu e o businesses, new
i ms we e also c ea ed.
Be ween 1992 and 2010, he e we e se e al changes in he pension and social
secu i y schemes ha pos poned an exi om he labo o ce (e.g., a pension e o m
in 2005 and changes in unemploymen bene i s). The unemploymen exi ou e was
based on unemploymen compensa ion. I consis ed o unemploymen bene i schemes
No dic jou nal o wo king li e s udies Volume 7 ❚ Numbe 2 ❚ June 2017 93
Figu e 1: Employmen , unemploymen , and economic de elopmen in Finland, 1990–2011.
Sou ces: S a is ics Finland, Annual na ional accoun s; S a is ics Finland, S a is ical Yea book o Finland,
1990–2013; S a is ics Finland, S a Fin da abase; Minis y o Employmen and he Economy, Annual
Employmen Se ice S a is ics 1991–2013; Au ho s’ calcula ions.
and an unemploymen pension. A so-called ‘unemploymen pension ube’ was a ailable
be ween 1986 and 2005, which g an ed an au oma ic ex ension o he ea nings- ela ed
unemploymen allowance o he olde unemployed who we e app oaching e i emen
(Jä ne el , 2010; Kyy ä, 2015).
da a and me hods
da a
We used Finnish Longi udinal Employe –Employee Da a (FLEED,1 S a is ics Finland),
which comp ises a andom sample o one- hi d o he whole popula ion aged 15–70
94 Time o Displacemen as a P edic o o Re-employmen A ja Jolkkonen e al.
li ing in Finland be ween 1988 and 2011. D awing on se e al adminis a i e egis e s,
he FLEED da abase includes in o ma ion on he subjec s’ indi idual cha ac e is ics,
income, and employmen his o y. The e is also a a iable indica ing hei labo ma ke
s a us a he end o each yea and, o he employed, an iden i y code o he plan o
company whe e he wo kplace is loca ed. Using he code, i is possible o iden i y he
employees o a gi en plan om he sample. As he da a o e codes o all plan s ha
exis in Finland a he end o each yea , i is possible o de e mine which plan s ha e
exi ed he ma ke s du ing he p e ious yea . As we we e in e es ed in plan closu es
in 1992, 1997, and 2003, we de ined a plan as closed i i exis ed o a 3yea pe iod
p eceding one o hese yea s bu did no exis du ing he nex 3yea pe iod. Mo eo e ,
he plan had o be ep esen ed in he FLEED da abase by a leas one employee a he
end o 1991, 1996, o 2002, as applicable. Plan s o ewe han 10 employees we e no
included. Changes in owne ship, me ge s, and ou sou cings whe e mo e han 70% o
he pe sonnel mo ed in o he se ice o a new employe we e no conside ed closu es.
The s udy sample consis s o indi iduals who had been employed a he plan o
3 o mo e yea s be o e he closu e and du ing hese yea s had no been unemployed o
pa icipa ed in labo ma ke policy measu es (c . Ko keamäki & Kyy ä, 2008: 7). Thus,
he s udy conce ns indi iduals wi h a s ong labo ma ke s a us p io o he displace-
men . Those who mo ed ab oad o died du ing he closu e yea we e excluded, as we e
hose who in he p eceding yea we e younge han 20 o olde han 60.
In addi ion o gende , age, oca ional educa ion (classi ied as none, seconda yle el,
lowe e ia yle el, and highe e ia y le el), and indus y (p ima y p oduc ion, second-
a y p oduc ion, se ices), we ex ac ed om he FLEED da abase wo a iables indi-
ca ing he indi iduals’ labo ma ke a achmen du ing he 7yea pe iod ollowing he
displacemen . The numbe o mon hs in employmen du ing each calenda yea was
used as he esponse a iable in he ajec o y analyses, and unemploymen du ing he
Table 1 Gende , educa ion, age, and indus y o he displaced employees acco ding o he yea o
plan closu e
Yea o displacemen
1992 1997 2003
Numbe o plan s (N = 985) (N = 1242) (N = 3339)
433 149 532
Gende
Male 650 (66%) 572 (46%) 1468 (44%)
Female 335 (34%) 670 (54%) 1871 (56%)
Voca ional educa ion
None 425 (43%) 426 (34%) 782 (23%)
Seconda yle el 355 (36%) 513 (41%) 1214 (36%)
Lowe e ia yle el 189 (19%) 271 (22%) 946 (28%)
Highe e ia yle el 16 (2%) 32 (3%) 397 (12%)
Mean age (SD) 40.6 (9.8) 43.6 (8.4) 44.5 (9.3)
SD: s anda d de ia ion
No dic jou nal o wo king li e s udies Volume 7 ❚ Numbe 2 ❚ June 2017 95
yea was moni o ed wi h a dicho omous a iable indica ing i he e had been a spell o
mo e han 3 mon hs wi hou wo k.
A o al o 5566 indi iduals we e eligible o he s udy sample (see Table 1). The sub-
sample om 1992 was maledomina ed, wi h a mean age 3 yea s lowe compa ed o he
subsamples om 1997 and 2003. The p opo ion o hose wi h a leas lowe e ia yle el
educa ion g ew om 21% in 1992 o 40% in 2003. Few employees wo ked in he p ima y
indus y; he di ision be ween he seconda y indus y and se ices was abou hal and
hal in 1992, while in 1997 and 2003, abou h ee ou hs o employees wo ked in he
se ices sec o .
s a is ics
T ajec o y analysis, o la en class g ow h analysis (Nagin, 1999; 2005), was applied
in de e mining he a ious eemploymen pa e ns o he displaced employees. This
me hod equi es ha he indi iduals be ollowed up o e ime so he e a e measu e-
men s om di e en ime poin s. The pu pose o his is o ind clus e s o indi iduals
ha , wi h espec o he measu ed p ope y, assume he same kind o de elopmen .
Fo example, he e migh be one g oup wi h a pa e n o changing alues and ano he
g oup wi h unchanged alues. S a ing om he assump ion ha he e a e such di e -
en de elopmen al pa e ns ‘hidden’ in he da a, ajec o y analysis is able o iden i y a
p ope se o pa e ns and hei shapes so ha each indi idual is assigned o he mos
app op ia e pa e n. An impo an assump ion is ha he measu emen s wi hin a a-
jec o y g oup a e aken o be independen . The s a is ical p ocedu e in ol es sea ching
o he numbe o ajec o y g oups whe e he p obabili y o each indi idual being clas-
si ied in a pa icula ajec o y g oup is as high as possible, while i is hoped ha he
p obabili y o ha same indi idual belonging o any o he g oup is as low as possible.
This solu ion is ob ained by ying ou di e en models, compa ing di e en s a is ical
c i e ia, and isually in e p e ing he eme ging ajec o ies (Jones e al., 2001; Jung &
Wick ama, 2008; Nagin, 2005). Finally, he ajec o ies should ul ill he common sense
c i e ion, ha is, hei in e p e a ion should i he subs an ial na u e o he p ope y o
in e es , which in ou s udy was employmen , measu ed as mon hs in a calenda yea
o e 7 successi e yea s.
T ajec o y analysis es ima es he model pa ame e s on he basis o he me hod o
maximum likelihood. In his s udy, he es ima es we e calcula ed i e a i ely using he
Expec a ion and Maximiza ion (EM) algo i hm (Demps e e al., 1977). Bayesian in o -
ma ion c i e ia (BIC) we e conside ed in choosing he bes solu ion— ha is, he op imal
numbe o ajec o y g oups.
A e p oducing he ajec o y g oups, we applied mul inomial logis ic eg ession
analyses o compa e he closu e yea s wi h espec o he dis ibu ion o indi iduals
in o di e en g oups. The analyses we e made i s wi h c ude igu es, hen he indi-
idual ac o s (gende , age, educa ion, indus y) we e in oduced o he model sepa-
a ely, and he inal model includes all ac o s. Since he ajec o y g oups a e based on
p obabili ies— hey a e no ixed cons uc s— he eg ession analyses we e weigh ed
wi h he pos e io p obabili y o belonging o he chosen (i.e., he mos likely) ajec-
o y g oup.
102 Time o Displacemen as a P edic o o Re-employmen A ja Jolkkonen e al.
plan be o e he closing yea . Thi d, ins ead o s udying he yea ly samples sepa a ely, we
pooled he da a o he 3 yea s. This enabled s a is ical es ing o he e ec o he yea
o displacemen and he in e ac ion o he yea and he indi idual cha ac e is ics o he
wo ke s wi h he subsequen 7-yea employmen ajec o y g oups.
One s eng h o his s udy is i s me hodological c edibili y. The selec ion o di -
e ing business cycles was based on a di e se se o mac oeconomic indica o s, he
pionee ing applica ion o mode n s a is ics in he domain u ned ou o be use ul,
and he ocus on a pa icula g oup o displaced wo ke s p o ided a solid basis o
he in e p e a ion o u u e employmen pa hs. The s ic inclusion c i e ia a e, on he
o he hand, a limi a ion, as he sample did no include indi iduals om he sho - e m
and a ypically con ac ed bu e wo k o ce ha los jobs in he plan closu es. Had
ou sample included a andom ci izens who we e unemployed a a gi en poin in ime,
he ajec o y solu ion may ha e been simila , bu he pe cen ages o indi iduals en e -
ing he ajec o ies o s ong employmen would ce ainly ha e been smalle . Fo his
eason, we would like o ancho he s udy o he con ex o plan closu e esea ch, and
ecommend cau ion in gene alizing he indings. On he o he hand, he esul s abou
‘case Finland’ can be gene alized in e na ionally o plan closu es in simila economic
and ins i u ional con ex s. As o he backg ound ac o s, we we e limi ed o he con-
en s o he FLEED da abase; ce ain impo an cha ac e is ics—such as indica o s
heal h and wo k abili y, o ins ance—we e no a ailable.
We selec ed he yea s 1992, 1997, and 2003 as baseline ime poin s o he ollow-
up. The choice o yea s was based on mac oeconomic indica o s like na ional GDP and
he employmen and unemploymen a es. The selec ion wo ked well, and he indings
e i ied ha cyclical ime ma e s and explains a lo o he a ia ion in he employmen
pa e ns o displaced wo ke s, and his is clea especially when compa ing 1992 wi h
1997 and 2003. Howe e , i is wo h no ing ha he di e ences be ween 1997 and
2003 we e less p onounced han we expec ed on he basis o he e iden changes in he
labo ma ke p econdi ions sugges ed by he mac oeconomic indica o s.
Conclusions
This s udy is unique bo h wi h espec o he sample—employees who we e s ongly
a ached o he wo kplace be o e i s closu e—and wi h espec o he me hods— a-
jec o y analysis o he employmen and subsequen eg ession analyses. The e o e, he
esul s canno be conside ed s ic ly in e ms o e i ying o alsi ying he indings o
ea lie Finnish o in e na ional s udies. Some compa isons a e, howe e , possible. The
excep ionali y o ou esul s o 1992 is consis en wi h Ga ell e al. (2010) and Fa be
(2011), who s udied he cyclical de elopmen o job ealloca ion, bu —in e ms o he
in eg a i e capaci y, selec i i y o demand, and de elopmen o employmen ajec o ies
in he long un—ou indings b ing impo an new ligh o he opic o he es uc u ing
and adap a ion o he labo ma ke .
While es ing ou hypo hesis wi h a eg ession analysis, he indi idual cha ac e -
is ics (age, gende , educa ion) we e conside ed o be con ounding ac o s. The associa-
ions be ween yea o displacemen and ajec o y emained p ac ically unchanged when
gende and educa ion we e added o he s a is ical model, whe eas adding age o he
model changed he RRR conside ably. In o de o es he sec o al shi hypo hesis,

No dic jou nal o wo king li e s udies Volume 7 ❚ Numbe 2 ❚ June 2017 103
we con olled he sec o o indus y and ound ha pe manen changes in he indus y
s uc u es and associa ed labo o ce demands do no explain he esul .
In e ms o policies and social ou comes, ou esul s demons a e he s ong ein-
eg a i e capaci y o he labo ma ke , as h ee ou hs o he sample was assigned o
he ajec o y g oup o con inuously s ong employmen . Di e ences in labo ma ke
a achmen be ween 1997 and 2003 we e mino , which pe haps e lec he ac ha new
jobs we e c ea ed and he demand o labo was ela i ely s ong in 2003, despi e he
slow g ow h in he economy.
The Finnish and No dic labo ma ke policy emphasis is no so much on compen-
sa ing o acqui ed bene i s ( he isk o income loss) bu a he —and abo e all—on
suppo ing wo ke s who ha e expe ienced a mass lay-o in e ms o e-employmen
and ein eg a ion in o he labo ma ke . As o whe he na ional lexicu i y models can
wi hs and global economic isks and whe he i is al oge he possible o p o ec he
labo o ce om economic isks wi hou in e na ional ag eemen s and social dialogue,
hese emain open ques ions.
The p incipal analysis o ou s udy ga e c edence o pos -hoc analyses ha sugges
he ins i u ions o he wel a e s a e do con ibu e o employmen ajec o ies, bu his
issue emains la gely a subjec o u he esea ch. The esul s o his s udy may also
open up in e es ing pe spec i es o compa a i e esea ch ac oss he No dic coun ies,
which ha e simila i ies in e ms o employmen policy and he ins i u ional se ings o
he wel a e s a e.
Acknowledgmen s
The au ho s would like o hank he Finnish Wo k En i onmen Fund o hei unding
o his esea ch.
Appendix A
Figu e A1: The e ec o age, con olling o gende and le el o educa ion, on he p obabili y o
alling on o he ajec o y o eco e ing employmen o wo ke s displaced in 1992, 1997, and 2003.
No e: Logi scale on he y-axis. The dashed lines show he 95% con idence in e al.
104 Time o Displacemen as a P edic o o Re-employmen A ja Jolkkonen e al.
e e ences
Appelq is , J. (2007). Wage and ea nings loses o displaced wo ke s in Finland. Discussion
Pape s 422, Helsinki: VATT Ins i u e o Economic Resea ch.
Asplund, R., Kauhanen, A. and Mali an a, M. (2012). Työ eh ä ien ja palkkojen dynamiikka
(Dynamics o asks and wages). Helsinki: Elinkeinoelämän u kimuslai os.
Becke , G. S. (1975). Human Capi al: Theo e ical and Empi ical Analysis wi h Special Re e -
ence o Educa ion, 2nd edi ion: Columbia Uni e si y P ess.
Ba h P. S. (1971). Time se ies analysis o layo a es. Jou nal o Human Resou ces 6, Fall,
pp. 448–65.
Boockmann, B. and S e es, S. (2007). Senio i y and job s abili y: A quan ile eg ession. Ap-
p oach using ma ched employe -employee da a. ZEW Discussion Pape no. 07–014,
Mannheim: Cen e o Eu opean Economic Resea ch.
Chia ini B. and Piselli, P. (2000). Unemploymen , Wage P essu e and Sec o al Shi s: Pe ma-
nen and Tempo a y Consequences o In e sec o al Shocks. Jou nal o Policy Modeling
22 (7): 777–799.
Cu le D., Huang, W. and Lle as-Muney, A. (2014). When does educa ion ma e ? The p o-
ec i e e ec o educa ion o coho s g adua ing in bad imes. Wo king Pape no. 20156,
Camb idge, MA: Na ional Bu eau o Economic Resea ch.
Demps e , A., Lai d, N. and Rubin, D. (1977). Maximum likelihood es ima ion o incom-
ple e da a ia he EM algo i hm. Jou nal o he Royal S a is ical Socie y 39: 1–38.
Eliason, M. and S o ie, D. (2009). Job loss is bad o you heal h – Swedish e idence on
cause-speci ic hospi aliza ion ollowing in olun a y job loss. Social Science & Medicine
68: 1396–1406.
Fa be , H. S. (2005). Wha do we know abou job loss in he Uni ed S a es? E idence om he
displaced wo ke s su ey 1984–2004. Wo king Pape 498, Indus ial Rela ions Sec ion,
P ince on Uni e si y.
Fa be , H. S. (2011). Job Loss in he g ea ecession: His o ical pe spec i e om he displaced
wo ke s su ey 1984–2010, Wo king Pape 564, Indus ial Rela ions Sec ion, P ince on
Uni e si y.
Figu e A2: The e ec o age, con olling o gende and le el o educa ion, on he p obabili y o
alling on o he ajec o y o immedia e exi om employmen o wo ke s displaced in 1992, 1997,
and 2003.
No e: Logi scale on he y-axis. The dashed lines show he 95% con idence in e al.
No dic jou nal o wo king li e s udies Volume 7 ❚ Numbe 2 ❚ June 2017 105
F ede iksen, A. and Wes e gaa d-Nielsen, N. (2007). Whe e did hey go? Modelling ansi-
ions ou o jobs. Labou Economics 14: 811–828.
F ede iksen, A. (2008). Gende di e ences in job sepa a ion a es and employmen s abili y:
New e idence om employe -employee da a. Labou Economics 15: 915–937.
Ga ell, M., Jans, A-C. and Pe sson, H. (2010). The impo ance o educa ion o he eallo-
ca ion o labo : E idence om Swedish employe -employee da a 1986–2002. Labou
Economics 17: 206–214.
Hall, R. E., Fa be , H. and Hal iwange , J. (1995). Los Jobs. B ookings Pape s on Economic
Ac i i y 1995: 221–273. 25 h Anni e sa y issue.
Houle, M. and Van Auden ode, M. (1995). Job displacemen , wages and unemploymen du-
a ion in Canada. Labou Economics 2: 77–91.
Hu unen, K., Møen, J. and Sal anes K. G. (2011). How Des uc i e is c ea i e des uc ion?
E ec s o job loss on job mobili y, wi hd awal and income. Jou nal o he Eu opean Eco-
nomic Associa ion 9: 840–870.
Jolkkonen, A. and Ku inen, A. (2009). Jous a uus ja u allisuus i isanomis ilan eissa –
Tapaus u kimus Pe los Oyj:n uo annon lope amises a Joensuu seudulla (Flexibili y
and secu i y in a lay-o si ua ion – A case s udy o he closing down o he Pe los Oyj
LTD in he Joensuu egion). Työ- ja elinkeinominis e iön julkaisuja. Työ ja y i äjyys 72.
Helsinki: Edi a Publishing.
Jolkkonen, A., Kois inen, P. and Ku inen, A. (2012). Reemploymen o displaced wo ke s –
The case o a plan closing on a emo e egion in Finland. No dic Jou nal o Wo king
Li e S udies 2: 81–100.
Jolkkonen, A., Kois inen, P., Ku inen, A., Lipiäinen, L., Nummi, T. and Vi anen, P, (2014).
Koh alokkaa olosuh ee ja yksilöllise ekijä : oimipaikan lakkau amisen seu auksena
yönsä mene äneiden kiinni yminen yöma kkinoille. (C i ical ci cums ances and indi-
idual ac o s: labo ma ke a achmen o displaced wo ke s a e job loss due o he
plan closu e). Kansan aloudellinen aikakauski ja 4: 492–515.
Jolkkonen, A., Kois inen, P., Ku inen, A., Lipiäinen, L., Nummi, T. and Vi anen, P, (2017).
Labou ma ke a achmen ollowing majo wo k o ce downsizings: a compa ison o
displaced and e ained wo ke s. Wo k, employmen and socie y (in p ess).
Jones, B. L., Nagin, D. S. and Roede , K. (2001). A SAS p ocedu e based on mix u e mod-
els o es ima ing de elopmen al ajec o ies. Sociological Me hods & Resea ch 29:
374–393.
Jung, T., and Wick ama K. A. S. (2008). An in oduc ion o la en class g ow h analysis and
g ow h mix u e modeling. Social and Pe sonali y Psychology Compass 2: 302–317.
Jä ne el , N. (2010). Educa ion and Longe Wo king Li es. A longi udinal s udy on educa ion
di e ences in he la e exi om wo king li e o olde employees in Finland. Finnish Cen-
e o Pensions, S udies 2010: 1.
Kle ze , L. G. (1998). Job displacemen . The Jou nal o Economic Pe spec i es 12: 115–136.
Kling all, M. (1998). Fö e agens ek y e ingsme ode – hu a be slösä å job. A be ma knad
& A be sli 4: 291–303.
Koebe , C. and W igh , D. W. (2006). Gende di e ences in he eemploymen s a us o
displaced wo ke s human capi al as signal ha mi iga e e ec s o bias. The Jou nal
o Socio-Economics 35 (2006): 780–796, doi: h p://dx.doi.o g/10.1016/j.socec.2005.
11.036.
Kollmeye , C. and Pichle , F. (2013). Is Deindus ializa ion Causing High Unemploymen
in A luen Coun ies? E idence om 16 OECD Coun ies, 1970–2003. Social Fo ces
(2013) 91 (3): 785–812. doi: h p://dx.doi.o g/10.1093/s /sos184.
Ko keamäki, O. and Kyy ä, T. (2008). A dis ibu ional analysis o displacemen cos s in an
economic dep ession and eco e y. Discussion Pape s 465, Helsinki: VATT Ins i u e o
Economic Resea ch.
106 Time o Displacemen as a P edic o o Re-employmen A ja Jolkkonen e al.
Ko keamäki, O. and Kyy ä, T. (2014). A dis ibu ional analysis o ea nings losses o displaced
wo ke s in an economic dep ession and eco e y. Ox o d Bulle in o Economics and
S a is ics 76: 565–588.
Kuhn, P. J. (2002). Summa y and Syn hesis.In Losing wo k, mo ing on: In e na ional pe -
spec i es on wo ke displacemen , ed. Pe e J. Kuhn. Kalamazoo: W. E. Upjohn Ins i u e
o Employmen Resea ch.
Kyy ä, T. (2015). Ea ly e i emen policy in he p esence o compe ing exi pa hways: E i-
dence om pension e o ms in Finland.Economica 82: 46–78.
Lilien D. (1982). Sec o al shi s and cyclical unemploymen .Jou nal o poli ical economy 90:
4, 777–93.
Liu, K., Sal anes, K. G. and Sø ensen, E. Ø. (2016). Good skills in bad imes: Cyclical skill
misma ch and he long- e m e ec s o g adua ing in a ecession.Eu opean Economic Re-
iew 84 (2016): 3–17. doi: h p://dx.doi.o g/10.1016/j.eu oeco e .2015.08.015.
Nagin, D. S. (1999). Analyzing de elopmen al ajec o ies: A semipa ame ic, g oup-based
app oach.Psychological Me hods 4: 139–177.
Nagin, D. S. (2005). G oup-based modeling o de elopmen . Camb idge: Ha a d Uni e si y
P ess.
Ne ke F, O o A. and Hidalgo C. (2016). The mobili y o displaced wo ke s: How he local
indus y mix a ec s job sea ch s a egies. Wo king pape s No. 71. Cen e o In e na ion-
al De elopmen a Ha a d Uni e si y.
Nedelkoska, L., Ne ke F. and Wiede hold S. (2015). Skill Misma ch and he Cos s o Job Dis-
placemen . Re ie ed om: h ps://www.aeaweb.o g/con e ence/2015/ e ie e.php?pd-
id=1264
Minis y o Employmen and he Economy, Annual Employmen Se ice S a is ics 1991–2013;
Re ie ed om: h p://www. em. i/en/wo k/employmen _se ice_s a is ics/s a is ical_
publica ions/s a is ical_summa y_-annuals
O mis on, R. (2014). Wo ke displacemen and occupa ion-speci ic human capi al. Wo k and
Occupa ions, 41: 350–384. doi: h p://dx.doi.o g/10.1177/0730888414531500.
S a is ics Finland, Annual na ional accoun s 1991–2013. Re ie ed om: h p://www.s a . i/
il/ p/index_en.h ml
S a is ics Finland, S a is ical Yea book o Finland, 1990–2013. Re ie ed om: h p://www.
s a . i/ up/ki jas o_ ie o/ ilas ollise _ uosiki ja .h ml
S a is ics Finland, S a Fin da abase
S e ens, A. H. (1997). Pe sis en e ec s o job displacemen : The impo ance o mul iple job
losses.Jou nal o Labo Economics 15: 165–188.
Ve ho, J. (2008). Sca s o ecession: he long- e m cos s o he Finnish economic c isis. Wo k-
ing pape 9. IFAU – ins i u e o Labou Ma ke Policy E alua ion, Uppsala.
Wood, S. N. (2006). Gene alized Addi i e Models – An In oduc ion wi h R. Boca Ra on:
CRC P ess.
No es
1 The con en desc ip ion o he combined Finnish Longi udinal Employe -Employee Da a (FLEED) is
a ailable online a h p://www.s a . i/ up/mik oaineis o /me_ku aus_henkilo_en.pd .