Rocke bie, Duane; Eas on, S ephen T.
A icle
Go e nmen unding alloca ions o uni e si ies and
he business cycle: An analysis o Canada's p o incial
go e nmen s
Economies
P o ided in Coope a ion wi h:
MDPI – Mul idisciplina y Digi al Publishing Ins i u e, Basel
Sugges ed Ci a ion: Rocke bie, Duane; Eas on, S ephen T. (2024) : Go e nmen unding alloca ions
o uni e si ies and he business cycle: An analysis o Canada's p o incial go e nmen s, Economies,
ISSN 2227-7099, MDPI, Basel, Vol. 12, Iss. 4, pp. 1-11,
h ps://doi.o g/10.3390/economies12040094
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/329020
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/4.0/
Ci a ion: Rocke bie, Duane, and
S ephen Eas on. 2024. Go e nmen
Funding Alloca ions o Uni e si ies
and he Business Cycle: An Analysis
o Canada’s P o incial Go e nmen s.
Economies 12: 94. h ps://doi.o g/
10.3390/economies12040094
Academic Edi o s: Ma hias Beck and
And ew Wa e son
Recei ed: 27 Ma ch 2024
Re ised: 12 Ap il 2024
Accep ed: 15 Ap il 2024
Published: 16 Ap il 2024
Copy igh : © 2024 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license (h ps://
c ea i ecommons.o g/licenses/by/
4.0/).
economies
A icle
Go e nmen Funding Alloca ions o Uni e si ies and he
Business Cycle: An Analysis o Canada’s
P o incial Go e nmen s
Duane Rocke bie 1,* and S ephen Eas on 2
1
Depa men o Economics, Uni e si y o Le hb idge, 4401 Uni e si y D i e, Le hb idge, AB T1K3M4, Canada
2Depa men o Economics, Simon F ase Uni e si y, 8888 Uni e si y D i e, Bu naby, BC V5A1S6, Canada;
[email p o ec ed]
*Co espondence: [email p o ec ed]
Abs ac : Canada’s uni e si ies each ecei e an annual ope a ing g an om hei p o incial go -
e nmen o pa ially inance ope a ing expenses. This pape es ima es he sensi i i y o p o incial
ope a ing g an s o he business cycle by disen angling he e ec s o p ocyclical income on go e n-
men e enue and he coun e cyclical e ec on s uden demand by u ilizing an economic eg ession
model composed o h ee equa ions. Ou panel da a include he o al eal ope a ing g an paid
o all uni e si ies wi hin a p o ince, o al s uden en olmen , eal pe capi a go e nmen e enue,
and eal pe capi a g oss domes ic p oduc o Canada’s en p o inces o e he 1992–2019 sample
pe iod. The esul s con i m ha eal pe capi a go e nmen e enues a e p ocyclical and ha ull- ime
equi alen s uden en olmen s a e coun e -cyclical. The o al eal ope a ing g an is only weakly
associa ed wi h cyclical changes in p o incial go e nmen e enue. Ins ead, he o al eal ope a ing
g an is mainly de e mined by coun e cyclical changes in s uden demand. This pa ially o se s he
po en ial educ ion in unding o uni e si ies du ing an economic down u n. P o incial go e nmen s
in Canada can smoo h he o al alloca ion o e he business cycle by adjus ing o he expendi u es
and using deb inancing. Ou esul s sugges hey do his o some ex en , bu no enough o a oid a
ne educ ion in eal ope a ing g an s du ing an economic down u n.
Keywo ds: business cycle; ope a ing g an ; highe educa ion; Canada
1. In oduc ion
Highe educa ion p o ides signi ican bene i s o he Canadian economy and he
cul u al ab ic o he na ion. Economis s e e o hese bene i s as public goods. Public goods
ha e he p ope y ha he indi idual ( he s uden in his case) canno cap u e all he bene i s
om p o iding he goods– he e a e many spill-o e bene i s ecei ed by he public o
which hey do no pay. In ecogni ion o he huge public bene i s highe educa ion con e s
on socie y, highe educa ion is subsidized by he Canadian axpaye . P io o he 1970s,
uni e si y subsidies we e la gely p o ided by he ede al go e nmen and he emphasis
was o expand he uni e si y sys em in Canada, which was ca ied ou apidly (Elb eck
2015). The ede al go e nmen passed his esponsibili y o he p o inces and e i o ies in
he ea ly 1970s by p o iding much expanded ans e paymen s o highe educa ion. Up o
he ea ly 1990s, p o inces and e i o ies emphasized inc easing en olmen s, and unding
was la gely based on ull- ime equi alen s (FTEs). Today, he p io i ies o p o incial and
e i o ial go e nmen s ha e changed imp o ing accessibili y, a he han en olmen s,
esul ing in he p omo ion o some colleges o deg ee-g an ing s a us.
Recen economic e en s ha e ad e sely a ec ed many o he p o incial and e i o ial
go e nmen ’s abili y o aise e enue, pa icula ly ola ile commodi y p ices and he
gene al slowdown o economic g ow h in he Canadian economy. This pape ocuses on
he sensi i i y o spending on highe educa ion o hese economic e en s wi h emphasis
Economies 2024,12, 94. h ps://doi.o g/10.3390/economies12040094 h ps://www.mdpi.com/jou nal/economies
Economies 2024,12, 94 2 o 11
on Canada’s p o incial uni e si ies. We do no add ess he dis ibu ion o spending o
di e en ins i u ions wi hin each p o ince— ha is usually based on a unding o mula
ha uses FTEs and o he key pe o mance indica o s as inpu s, as well as disc e iona y
spending. Ins ead, we ocus on he o al p o incial e enue as i is in luenced by cyclical
and andom luc ua ions in he economy. The o al ope a ing g an alloca ed o each
uni e si y in ou sample is hen a po ion o he o al p o incial e enue. We do no include
capi al expendi u es o o he special expendi u es. By es ima ing he elas ici y o he eal
g ow h in he o al ope a ing g an o he eal g ow h in p o incial e enue, we can es o
cyclical and andom componen s and es o smoo hing he alloca ion o e he business
cycle. O cou se, he abili y o smoo h spending o highe educa ion om e enues also
depends upon go e nmen p io i ies, and in pa icula , he s a e o he go e nmen su plus
o de ici and any public deb ou s anding. We also inco po a e a poli ical spending cycle
by including a iables o elec ions and poli ical ideology.
The e exis s li e a u e on he sensi i i y o highe educa ion spending in he Uni ed
S a es o economic e en s; howe e , we could ind no such s udy o Canada. Humph eys
(2000) es ima ed he sensi i i y o he eal g ow h a es o highe educa ion spending by
U.S. s a es o he eal g ow h a es o s a e GDP and ound i o ha e an elas ici y o 1.17
by es ima ing a eg ession o s a e educa ion app op ia ions pe FTE on eal GDP pe
capi a. He also ound a la ge elas ici y when he economy was in ecession ha when i
was in an expansion. The es ima ed elas ici y a ises om a educed o m equa ion since
s a e highe educa ion app op ia ions a e no a di ec unc ion o eal GDP pe capi a.
Ins ead, luc ua ions in he economy igge luc ua ions in go e nmen e enue ha
esul in luc ua ions in highe educa ion app op ia ions. We add ess his sho coming
by employing a mul iple equa ion model o his p ocess. Kane e al. (2005) also ound a
p ocyclical associa ion o s a e app op ia ions using he s a e unemploymen a e as a
measu e o business cycle swings.
E en i a new go e nmen is elec ed, many U.S. s a es ha e adop ed ax and expendi-
u e limi a ions (TELs), e ec i ely ying he hands o u u e go e nmen s by placing an
uppe limi on he abili ies o he go e nmen o aise speci ic ax e enues and conduc
di e en ypes o expendi u es ( hese a e no a ea u e o Canadian go e nmen s o ou
knowledge). A chibald and Feldman (2006,2008) es ima e he elas ici y o highe educa ion
spending o a ious economic ac o s in s a es ha ha e adop ed TELs e sus s a es ha
ha e no , bu ind no signi ican di e ence in he elas ici y alue, sugges ing ha highe ed-
uca ion spending is alued mo e highly by s a e go e nmen s han o he ypes o spending.
A balanced budge equi emen (BBR) is a mo e s ingen s a e iscal policy ha equi es he
s a e o ma ch spending o e enue and hus no accumula e any new public deb . Albe a
is he only Canadian p o ince o e i o y o adop a BBR, bu se e al U.S. s a es ha e done
so. Se na and Ha is (2014) examined s a e unding o highe educa ion and ound no
associa ion be ween s a es wi h BBRs and hose wi hou , sugges ing ha he p esence o
a BBR is la gely igno ed when conside ing highe educa ion spending. We include he
p e ious yea ’s p o incial budge de ici as a cons ain in alloca ing e enue o highe
educa ion ope a ing g an s.
P i a e uni e si ies a e much mo e p e alen in he Uni ed S a es han Canada; hence,
hei a ailabili y could impac he decisions o s a e go e nmen s o und public highe
educa ion ins i u ions. Thelin (2004) ound ha s a es wi h mo e p i a e uni e si ies
gene ally had la ge elas ici ies han hose wi hou . The me hod h oughou hese pape s
is o eg ess o al s a e app op ia ions o highe educa ion di ec ly on eal GDP pe capi a,
a me hod we hope o imp o e upon.
Ou model has se e al in e es ing esul s. P o incial go e nmen e enue is s ongly
in luenced by cyclical changes in p o incial GDP, bu p o incial ope a ing g an s espond
only weakly o changes in p o incial go e nmen e enue, esul ing in a weak o e all
cyclical e ec . P o incial go e nmen s do espond o changes in s uden demand measu ed
by FTEs ha beha e coun e cyclically. This is he main d i ing o ce behind changes o he
ope a ing g an .
Economies 2024,12, 94 3 o 11
2. Highe Educa ion Funding in Canada
All Canadian p o inces ecei e a ans e paymen om he ede al go e nmen ha
is a ge ed owa ds heal h ca e, educa ion, and social se ices. These ans e s a e unded
h ough ede al income ax and hus o e no speci ic ax bu den o he p o inces. P o inces
also collec e enue h ough he b oad ca ego ies o axa ion (income ax, sales ax and
co po a e ax), use ees, na u al esou ce oyal ies, and e enues om gene al go e nmen
ope a ions. Running a budge de ici is also a o m o e enue as i is used o inance alued
expendi u es (excluding unding deb se ice paymen s), al hough i is a bu den o u u e
axpaye s. Tax e enues can be pa icula ly sensi i e o he business cycle.
P o incial unding o he ope a ing g an has expe ienced episodes o inc eases and
dec eases. Figu e 1displays he o al p o incial ope a ing g an s ela i e o o al p o incial
GDP. The 1990s expe ienced a signi ican d op in his measu e o unding only o see
inc eased unding in he nex decade. Funding has again dec eased since 2010.
Economies 2024, 12, x FOR PEER REVIEW 3 o 12
only weakly o changes in p o incial go e nmen e enue, esul ing in a weak o e all
cyclical effec . P o incial go e nmen s do espond o changes in s uden demand meas-
u ed by FTEs ha beha e coun e cyclically. This is he main d i ing o ce behind changes
o he ope a ing g an .
2. Highe Educa ion Funding in Canada
All Canadian p o inces ecei e a ans e paymen om he ede al go e nmen ha
is a ge ed owa ds heal h ca e, educa ion, and social se ices. These ans e s a e unded
h ough ede al income ax and hus offe no speci ic ax bu den o he p o inces. P o -
inces also collec e enue h ough he b oad ca ego ies o axa ion (income ax, sales ax
and co po a e ax), use ees, na u al esou ce oyal ies, and e enues om gene al go -
e nmen ope a ions. Running a budge de ici is also a o m o e enue as i is used o
inance alued expendi u es (excluding unding deb se ice paymen s), al hough i is a
bu den o u u e axpaye s. Tax e enues can be pa icula ly sensi i e o he business
cycle.
P o incial unding o he ope a ing g an has expe ienced episodes o inc eases and
dec eases. Figu e 1 displays he o al p o incial ope a ing g an s ela i e o o al p o in-
cial GDP. The 1990s expe ienced a signi ican d op in his measu e o unding only o see
inc eased unding in he nex decade. Funding has again dec eased since 2010.
Figu e 1. P o incial ope a ing g an /p o incial GDP o 10 Canadian p o inces, 1992–2019. Sou ce:
CAUBO da abase and CANSIM.
On a io is Canada’s la ges p o ince bo h in e ms o popula ion and GDP. I also
boas s he la ges numbe o uni e si ies o any p o ince (17). On a io adop ed a o mula-
based unding model o he dis ibu ion o he o al alloca ion o e enue o he ope a ing
g an in 1985 (Highe Educa ion Quali y Council o On a io 2015). The model is complex
bu in i s basic o m, each uni e si y is awa ded a numbe o basic income uni s (BIUs)
based on en olmen and he ypes o p og ams offe ed. A uni e si y ha offe s mo e ex-
pensi e p og ams o se ice, such as sciences, medicine, enginee ing, e c., will ecei e
mo e BIUs a a gi en le el o en olmen han one ha offe s much ewe o hese p o-
g ams. En olmen g ow h pu s p essu e on he numbe o BIUs o inc ease as well; how-
e e , mo e unding will be p o ided only i BIUs inc ease by o e 3% using a i e-yea
mo ing a e age o a oid p o iding an incen i e o apidly inc ease en olmen me ely o
ob ain mo e unding. A po ion o he o al monies made a ailable by he go e nmen in
0.003
0.0035
0.004
0.0045
0.005
0.0055
0.006
0.0065
1992 1994 1996 1998 2000 20022004 2006 2008 2010 2012 2014 2016 2018
Ope a ing g an / GDP
Figu e 1. P o incial ope a ing g an /p o incial GDP o 10 Canadian p o inces, 1992–2019. Sou ce:
CAUBO da abase and CANSIM.
On a io is Canada’s la ges p o ince bo h in e ms o popula ion and GDP. I also
boas s he la ges numbe o uni e si ies o any p o ince (17). On a io adop ed a o mula-
based unding model o he dis ibu ion o he o al alloca ion o e enue o he ope a ing
g an in 1985 (Highe Educa ion Quali y Council o On a io 2015). The model is complex
bu in i s basic o m, each uni e si y is awa ded a numbe o basic income uni s (BIUs)
based on en olmen and he ypes o p og ams o e ed. A uni e si y ha o e s mo e
expensi e p og ams o se ice, such as sciences, medicine, enginee ing, e c., will ecei e
mo e BIUs a a gi en le el o en olmen han one ha o e s much ewe o hese p og ams.
En olmen g ow h pu s p essu e on he numbe o BIUs o inc ease as well; howe e ,
mo e unding will be p o ided only i BIUs inc ease by o e 3% using a i e-yea mo ing
a e age o a oid p o iding an incen i e o apidly inc ease en olmen me ely o ob ain
mo e unding. A po ion o he o al monies made a ailable by he go e nmen in i s
annual budge is di ided by he o al numbe o BIUs o de e mine he dolla alue o a
BIU. Funds made a ailable o capi al p ojec s a e no included in he o mula.
The o al p o incial ope a ing g an is subjec o changes in poli ical ideology and iscal
eali ies. The Conse a i e go e nmen o Michael Ha is educed he o al ope a ing g an
a ailable o uni e si ies in he On a io p o incial budge beginning in 1995. Coinciding
wi h his was he de egula ion o ui ion ees wi h he objec i e o educing he sha e o
uni e si y e enues a ising om he ope a ing g an . Figu e 2displays a ma ked decline
Economies 2024,12, 94 4 o 11
in he eal ope a ing g an pe FTE o he p o ince’s 17 majo uni e si ies ha con inued
un il 2003 when a Libe al go e nmen was elec ed. Ne e heless, by 2013–14, he p o incial
ope a ing g an composed only 27% o he CAD 13.1 billion in o al uni e si y e enue,
wi h ui ion accoun ing o 38%, a la ge dec ease om he 70% sha e o e enue by he
ope a ing g an in 1991. An accessibili y und was added o he ope a ing g an in 2001 o
enable he inc eased en olmen s due o he elimina ion o he high school g ade 13. This
und has con inued oday. A small amoun (4%) o he ope a ing g an is alloca ed based
on pe o mance indica o s submi ed by each ins i u ion.
Economies 2024, 12, x FOR PEER REVIEW 4 o 12
i s annual budge is di ided by he o al numbe o BIUs o de e mine he dolla alue o
a BIU. Funds made a ailable o capi al p ojec s a e no included in he o mula.
The o al p o incial ope a ing g an is subjec o changes in poli ical ideology and
iscal eali ies. The Conse a i e go e nmen o Michael Ha is educed he o al ope a -
ing g an a ailable o uni e si ies in he On a io p o incial budge beginning in 1995. Co-
inciding wi h his was he de egula ion o ui ion ees wi h he objec i e o educing he
sha e o uni e si y e enues a ising om he ope a ing g an . Figu e 2 displays a ma ked
decline in he eal ope a ing g an pe FTE o he p o ince’s 17 majo uni e si ies ha
con inued un il 2003 when a Libe al go e nmen was elec ed. Ne e heless, by 2013–14,
he p o incial ope a ing g an composed only 27% o he CAD 13.1 billion in o al uni e -
si y e enue, wi h ui ion accoun ing o 38%, a la ge dec ease om he 70% sha e o e -
enue by he ope a ing g an in 1991. An accessibili y und was added o he ope a ing
g an in 2001 o enable he inc eased en olmen s due o he elimina ion o he high school
g ade 13. This und has con inued oday. A small amoun (4%) o he ope a ing g an is
alloca ed based on pe o mance indica o s submi ed by each ins i u ion.
Figu e 2. Real ope a ing g an pe FTE o 17 majo uni e si ies in On a io, CAD millions, 1992–
2019. Sou ce: CAUBO da abase.
Some o Canada’s o he nine p o inces use a o mula me hod o de e mine he sha e
o he o al monies alloca ed o uni e si y ope a ing g an s o each ins i u ion. The me h-
ods diffe in de ail wi h some p o inces, such as Albe a, placing mo e emphasis on e-
sponding o ma ke demand o p og ams han con inuing unding o exis ing p og ams
(Fle che 2020). B i ish Columbia does no use a unding o mula—inc eases in ope a ing
g an s a e ied o he p e ious yea ’s ope a ing g an wi h disc e iona y inc eases o s u-
den demand (Minis y o Ad anced Educa ion and Skill T aining 2022). Quebec u ilized
a unding o mula ha included 538 ac o s up o 2018, educing o 107 ac o s a e wa d
(Venne 2018). Tui ion ee inc eases a e egula ed in all p o inces excep On a io (Ba aka
2023).
Canada’s o he nine p o inces expe ienced simila declines o On a io in unding o
uni e si ies a e he 2008 inancial c isis. All p o incial go e nmen s suffe ed a decline
in eal pe capi a e enue in 2009, he la ges being Saska chewan a −21.4%. The a e age
decline ac oss all en p o inces measu ed −5.01% in 2009, esul ing in a decline in eal
ope a ing g an s pe FTE o all en p o inces. Fo una ely, mos o he p o inces e u ned
o a le el exceeding he 2008 le el by 2013. On a io was an excep ion. The po en ial nega-
i e effec on ope a ing g an s caused by he 2008 inancial c isis was pa ly offse by he
4000
4500
5000
5500
6000
6500
7000
7500
8000
8500
9000
19921994199619982000200220042006200820102012201420162018
Figu e 2. Real ope a ing g an pe FTE o 17 majo uni e si ies in On a io, CAD millions, 1992–2019.
Sou ce: CAUBO da abase.
Some o Canada’s o he nine p o inces use a o mula me hod o de e mine he sha e
o he o al monies alloca ed o uni e si y ope a ing g an s o each ins i u ion. The me hods
di e in de ail wi h some p o inces, such as Albe a, placing mo e emphasis on esponding
o ma ke demand o p og ams han con inuing unding o exis ing p og ams (Fle che
2020). B i ish Columbia does no use a unding o mula—inc eases in ope a ing g an s a e
ied o he p e ious yea ’s ope a ing g an wi h disc e iona y inc eases o s uden demand
(Minis y o Ad anced Educa ion and Skill T aining 2022). Quebec u ilized a unding
o mula ha included 538 ac o s up o 2018, educing o 107 ac o s a e wa d (Venne
2018). Tui ion ee inc eases a e egula ed in all p o inces excep On a io (Ba aka 2023).
Canada’s o he nine p o inces expe ienced simila declines o On a io in unding o
uni e si ies a e he 2008 inancial c isis. All p o incial go e nmen s su e ed a decline in
eal pe capi a e enue in 2009, he la ges being Saska chewan a
−
21.4%. The a e age
decline ac oss all en p o inces measu ed
−
5.01% in 2009, esul ing in a decline in eal
ope a ing g an s pe FTE o all en p o inces. Fo una ely, mos o he p o inces e u ned
o a le el exceeding he 2008 le el by 2013. On a io was an excep ion. The po en ial
nega i e e ec on ope a ing g an s caused by he 2008 inancial c isis was pa ly o se
by he coun e cyclical inc ease in pos -seconda y en olmen o 39,000 s uden s ac oss
Canada expe ienced in 2009 (The Canadian P ess 2009). P o incial go e nmen s espond
o highe en olmen by inc easing ope a ing g an s as a commi men o main ain pos -
seconda y educa ion.
The app oach used in his pape sides eps he e ec o di e ing unding o mulae on
ope a ing g an s by ocusing solely on he o al ope a ing g an alloca ed by each p o ince—
ope a ing g an s o indi idual ins i u ions a e agg ega ed. Since he alloca ion o each
ins i u ion is o mulaic in many cases, i is he sensi i i y o he o al p o incial alloca ion
o ope a ing g an s o economic cycles ha is wo hy o s udy.
Economies 2024,12, 94 5 o 11
3. Poli ical Fac o s
P e ious esea ch using Canadian p o incial go e nmen accoun s da a has ocused
on wo main hemes: de e mining i p o incial expendi u es a e ied o a poli ical cycle,
and e iciencies in p o incial spending. The poli ical cycle model has wo compe ing e ec s
(Hibbs 1977). Any uling poli ical pa y may wish o inc ease i s expendi u es in an elec ion
yea o inc ease i s chances o e-elec ion ( he elec o al cycle); howe e , a igh -wing uling
poli ical pa y may wish o p ac ice aus e i y in he non-elec ion yea s o signal i s ideology
i a majo i y o o e s a e pe cei ed o ag ee wi h i ( he pa isan cycle). The empi ical
e idence has been suppo i e using Canadian p o incial da a. Dickson and Yu (1997),
Pe y e al. (1999), Kneebone and McKenzie (2001), and Tellie (2006) ound suppo o he
poli ical cycle hypo hesis. Se le is and A xen iou (1998) ound no e idence o a poli ical
cycle using da a o he Canadian ede al go e nmen s. Poli ical conside a ions could en e
how p o inces choose o aise e enues and spend hem. The ma ginal poli ical cos s, in
e ms o he educed p obabili y o e-elec ion, we e ound o be highes o di ec axes
by Landon and Ryan (1997), while a g ea e eliance on indi ec axes (use ees, licenses,
e c.) was ound o ha e a much lowe ma ginal cos . Go e nmen spending on goods
and se ices was he only spending ca ego y o ha e a high ma ginal cos . Abizadeh
and G ay (1992) ound ha he ideology o he uling p o incial poli ical pa y had no
in luence on p o incial economic g ow h. Dickson (2009) ound ha ede al go e nmen
spending pe capi a was highe in p o inces ha demons a ed loyal y o he uling ede al
go e nmen pa y.
Elec ion cycles and poli ical ideologies could in luence he sha e o On a io go e n-
men e enue alloca ed o he o al uni e si y ope a ing g an . We accoun o poli ical
and pa isan cycle e ec s using he same me hod as Pe y e al. (1999). An elec ion and
pos -elec ion yea s a e coded wi h bina y independen a iables (Elec and Pos elec). Pe y
e al. (1999) cons uc ed h ee bina y independen a iables, one o he yea p io o an
elec ion, one o he mid-poin o an elec ion cycle, and one o he yea ollowing an
elec ion. The pa isan e ec is measu ed using he powe index (Powe ) de ailed in Pe y
e al. (1999) ha akes on he maximum alues o 1 o a igh go e nmen and
−
1 o a
le go e nmen .
4. Me hodology
The eg ession model’s pu pose is o cap u e he unding p ocess om changes in
p o incial income o changes in p o incial go e nmen e enue o changes in he o al
ope a ing g an each yea . P e ious s udies ha e modelled his p ocess in a single eg ession
equa ion ha ypically includes cu en and lagged alues o income—a educed o m
equa ion om an unspeci ied sys em o equa ions. As such, he con ibu ion o each
componen o he sys em canno be iden i ied; ins ead, he single eg ession equa ion yields
coe icien s ha a e mong els con aining he unknown sys em coe icien s. We seek o
imp o e his app oach by speci ying a eg ession model composed o ou equa ions ha
allow o he es ima ion o he deepe coe icien s in he mong els.
e enue =α1+β1gdp +β2gdp −1+e1(1)
g an =α2+δ1 e enue +δ2 e enue −1+δ3 e +θ1elec ion +θ2pos elec ion +θ3powe +e2(2)
e =α3+π1gdp +π2gdp −1+e3(3)
g an =γo+γ1gdp +γ2gdp −1+γ3gdp −2+θ1elec ion +θ2pos elec ion +θ3powe + 1(4)
While i is e y unlikely ha he p o incial wo king age popula ion (15–70 yea s)
has a s ong associa ion wi h he p o incial business cycle, he same migh no be a gued
o o al FTEs. The ex ensi e li e a u e ends o ind ha uni e si y en olmen beha es
coun e cyclically, possibly due o he lack o employmen oppo uni ies. Cha les e al.
(2018), Sie e sen (2016), Ba and Tu ne (2015), and Johnson (2013) a e ecen e e ences.
Economies 2024,12, 94 6 o 11
Ca d and Lemieux (2000) ound a coun e cyclical e ec using Canadian da a. Equa ion (3)
cap u es he possible coun e cyclical beha io o en olmen s using cu en and one-pe iod-
lagged eal pe capi a GDP g ow h. The g ow h a e o he cu en and one-yea -lagged
p o incial unemploymen a e (S a is ics Canada Table 14-10-0078-01) was con empla ed o
inclusion in (3) bu was no . The g ow h in he eal ope a ing g an is no di ec ly a unc ion
o he g ow h a e o he unemploymen a e; since he unemploymen a e is no e enue,
i is necessa y o include a i h equa ion ha ela es eal GDP g ow h o he g ow h a e o
he unemploymen a e. Howe e , i is no clea wha he di ec ion o causa ion is be ween
he wo. I hey a e simul aneously de e mined, his c ea es endogenei y issues equi ing
some o m o ins umen al a iable (IV) es ima ion, an unnecessa y complexi y. Second, he
g ow h in he ope a ing g an is clea ly a unc ion o he g ow h in eal incomes as hese a e
axed o gene a e go e nmen e enue. The unemploymen a e is no axed. In addi ion,
when (3) is es ima ed wi h he g ow h a es o he cu en and lagged unemploymen a e,
he s ong co ela ion wi h eal GDP g ow h ( =
−
0.704) ende s he coe icien s
π1
and
π2
s a is ically insigni ican . The unemploymen a e also has measu emen issues ha makes
i s inclusion in he model ques ionable (Humph eys 2000).
Subs i u ing Equa ions (1) and (3) in o (2) gi es he educed o m eg ession in Equa-
ion (4), whe e
γo=α2+(δ1+δ2)α1+δ3α3
,
γ1=β1δ1+δ3π1
,
γ2=β2δ1+β1δ2+δ3π2
,
γ3=β2δ2
, and
1=e2 +δ1e1 +δ2e1 −1+δ3e3
. Equa ion (4) is es ima ed o es he
c oss-equa ion es ic ions o he alidi y o Equa ions (1)–(3). The s uc u e o he sys em
o Equa ions (1)–(3) p esen s no appa en issues o endogenei y.
The o al p o incial ope a ing g an is measu ed as he sum o he ope a ing g an s
o each p o incial uni e si y lis ed in he appendix o e he sample pe iod 1992–2019.
The 2020–2021 yea s we e excluded due o he COVID-19 pandemic when spending on
heal hca e domina ed dis o ed go e nmen unding p io i ies. The sample was limi ed o
77 public ins i u ions ha each ecei ed o e CAD 1 million in ope a ing g an s each yea .
A lis o hese ins i u ions is a ailable om he co esponding au ho upon eques . Annual
ope a ing g an s we e ob ained om he Canadian Associa ion o Uni e si y Business O i-
ce s (CAUBO) inancial da abase (ob ained by eques om h ps://www.caubo.ca/). To al
p o incial go e nmen e enue was ob ained om he da abase main ained by Kneebone
and Wilkins (2016) (h ps://www.policyschool.ca/publica ion-ca ego y/ esea ch-da a/ ac-
cessed on 21 Sep embe 2023). P o incial GDP alues we e ob ained om S a is ics Canada
(Table 36-10-0222-01). The o al ope a ing g an , p o incial go e nmen e enue, and GDP
we e de la ed using he p o incial consume p ice index (2002 = 100,
Table 18-10-0005-01
).
The o al numbe o FTE s uden s o he sample uni e si ies was collec ed o scale he
o al ope a ing g an ac oss p o inces (Table 37-10-0011-01). Pa - ime s uden s we e
assigned a weigh o 0.5 in he o al. The eal p o incial go e nmen e enue and eal
income we e di ided by he p o incial popula ion aged 15–70 o p oduce pe capi a alues
(
Table 17-10-0005-01
). The comple e da ase includes he eal ope a ing g an pe FTE
(g an ), eal p o incial e enue pe capi a ( e enue), eal p o incial GDP pe capi a (gdp),
and each measu ed as a g ow h a e. A dummy a iable o an elec ion yea (elec ion),
a dummy a iable o he yea ollowing an elec ion yea (pos elec ion), and a a iable
measu ing he pa isan e ec o a igh -wing go e nmen (powe ) a e also included.
Real p o incial e enue is eg essed on cu en and one-pe iod-lagged eal p o incial
g oss domes ic p oduc in (1) as a la ge po ion o hese e enues a e di ec and indi ec ax
collec ions. The a e age ax collec ions as a sha e o e enue anged om a low o 31.4%
o Albe a o a high o 66.8% o Quebec be ween 1992 and 2019. The o e all p o incial
a e age was 43.4%. A po ion o ede al ans e paymen s a e de e mined by a o mula
ha is p ocyclical. The emaining e enues a e esou ce e enues and e enues ob ained
om ees o go e nmen se ices ha a e cap u ed in he in e cep and e o e m. The
o al eal p o incial ope a ing g an o i s uni e si ies is unded by cu en and one-pe iod-
lagged eal p o incial e enue in (2) o accoun o he lag be ween e enues ecei ed and
budge alloca ions ha ypi y he go e nmen budge ing p ocess; howe e , a po ion o
he o al ope a ing g an can be unded wi h de ici inancing and/o educing spending
Economies 2024,12, 94 7 o 11
on o he go e nmen accoun s. This decision is posi ed o be a unc ion o inc eases in
s uden demand, ep esen ed by ull- ime equi alen s (
e
), he elec ion cycle (
elec ion
and pos elec ion ), and he ideological posi ion o he p o incial go e nmen ( powe ).
5. Time Se ies S a iona i y
Al hough wo a iables may appea o be s ongly associa ed, hey may no ac ually
be i a spu ious eg ession e o is commi ed. G ange and Newbold (1974) demons a ed
ha wo andom walk se ies ha ha e no meaning ul associa ion can be s ongly co ela ed,
gene a ing a eg ession model wi h good i . The p oblem can be add essed by de e mining
he deg ee o s a iona i y o each a iable using a uni oo es . I bo h se ies a e s a iona y
a e i s -di e encing (I(1)), hey can be eg essed in i s -di e ences and in e p e ed in
he same manne as a eg ession model in le els. The panel uni oo es s ha assume a
common oo (Le in e al. 2002 o LLC) and indi idual c oss-sec ion oo s (Im e al. 2003
o IPS) a e used he e. The null hypo hesis in bo h he LLC and IPS es s is ha he se ies
is I(1). Table 1p esen s he esul s o hese es s ha can be pe o med using an in e cep
and end a iable o an in e cep only. The IPS panel es ails o ejec a uni oo o each
a iable. Based on hese uni oo es esul s, using he g ow h a es o g an , e enue,gdp,
and e is deemed app op ia e.
Table 1. Resul s o LLC and IPS panel uni oo es s. * Indica es s a is ical signi icance a 95%
con idence.
g an e enue gdp e
LLC es
In e cep 0.294 0.210 −2.186 * 1.061
In e cep and end −0.315 0.872 −0.153 −0.910
IPS es
In e cep 2.385 1.783 0.778 3.910
In e cep and end 0.349 −0.279 1.081 −1.273
6. Resul s
The sys em o Equa ions (1)–(4) was es ima ed using leas squa es wi h a co ec ed
co a iance ma ix o he e oskedas ici y wi hin each p o ince. Se ial co ela ion was no
ound in any o he p o inces in any o he equa ions. Fixed e ec s we e ini ially used o
accoun o any ime-in a ian di e ences ac oss he p o inces; howe e , he s a is ical
signi icance o he ixed e ec s was ound only o (3), likely due o he a iables being
measu ed in g ow h a es a he han in le els. The e o e, ixed e ec s we e omi ed
om Equa ions (1) and (2). C oss-sec ional dependence occu s when he e o s ac oss
he p o inces a e co ela ed in one o he equa ions. This can occu when all p o inces
expe ience he same shock o an unobse ed a iable. In he case o highe educa ion,
inno a ions o ypes o p og ams and how educa ion was deli e ed since 1992 could be
a common shock. Igno ing c oss-sec ional dependence can esul in un eliable es ima es
and e en inconsis en es ima o s. I he numbe o co ela ed c oss-sec ion uni s is small,
he p oblem can be minimal (Chudik and Pesa an 2015). C oss-sec ional dependence
could no be ejec ed o Equa ions (1)–(4) using he Pesa an CD es , which is app op ia e
when he numbe o c oss-sec ions is less han he numbe o ime pe iods. As a esul , a
panel SUR model was es ima ed o each equa ion ha u ilizes he a iance–co a iance
ma ix o e o s in a wo-s ep es ima ion. The panel size is N= 260 obse a ions due o
he one-pe iod-lagged gdp a iable. Including addi ional lags o gdp did no o e any
imp o emen s in he es ima ed esul s. The eg ession esul s appea in Table 2below.
Economies 2024,12, 94 8 o 11
Table 2. Reg ession esul s o Equa ions (1)–(4), N= 260 (1992–2019). No e ha p o incial GDP and
FTEs a e annual g ow h a es. * Indica es s a is ical signi icance a 95% con idence.
Equa ion (1) Equa ion (2) Equa ion (3) Equa ion (4) Equa ion (5)
Cons an 0.547 −0.140 1.771 * 0.374 −0.139
gdp 0.565 * −0.102 * 0.058
gdp −1−0.109 0.001 0.171 *
gdp −20.113 *
e enue −0.019
e enue −10.091 *
e 0.204 *
elec ion 1.490 * 1.475 * 1.264 *
pos elec ion 1.541 * 1.486 * 1.415 *
powe −0.017 −0.177 −0.011
e enue xDe 0.035
e enue xD −0.185 *
e enue −1xDe 0.098 *
e enue −1xD 0.004
e xDe 0.193 *
e xD 0.236 *
Adjus ed R20.252 0.139 0.288 0.188 0.142
N260 260 260 260 260
We ocus on he s a is ically signi ican eg ession coe icien s in Table 2. The esul s
o (1) sugges a s ong posi i e associa ion be ween g ow h in p o incial go e nmen
e enue and GDP g ow h (bo h eal pe capi a), which is no su p ising as a signi ican
po ion o e enue is collec ed om axes o all p o inces. The elas ici y is 0.843 e alua ed
a he means, sugges ing an inelas ic esponse; howe e , no all e enues a e ela ed o
GDP (go e nmen businesses, esou ce e enues, e c.). Jumping o he esul s o (3),
he coe icien o
−
0.102 (elas ici y =
−
0.193) sugges s a small, coun e cyclical e ec on
en olmen g ow h ha ag ees wi h he esul s ound in p e ious esea ch. Equa ion (2)
is he mos ele an o ou s udy. The g ow h in he eal o al p o incial go e nmen
ope a ing g an is posi i ely associa ed wi h g ow h in eal o al e enue pe capi a when
FTE g ow h is held cons an . The elas ici y alue is 0.115, sugges ing a e y inelas ic
esponse—go e nmen s ha e o he spending p io i ies han highe educa ion, such as
heal h ca e, social se ices, and in as uc u e. Holding e enue g ow h cons an , he
g ow h in he o al ope a ing g an posi i ely esponds o g ow h in FTEs wi h a coe icien
o 0.204 (elas ici y = 0.329). Go e nmen s se ice an inc easing numbe o s uden s when
e enue is no g owing wi h a la ge ope a ing g an , bu no by he amoun necessa y
o main ain a cons an ope a ing g an pe FTE. The g ow h in he o al ope a ing g an
inc eases signi ican ly du ing an elec ion yea and in he yea ollowing, an amoun equal o
app oxima ely 1.5% in each case. This is posi i e e idence o an elec ion cycle in Canada’s
p o inces wi h ega ds o highe educa ion spending. Howe e , he e is no e idence o
a pa isan cycle in highe educa ion spending due o he s a is ical insigni icance o he
powe coe icien .
Equa ion (4) sugges s a signi ican inelas ic esponse o he eal g ow h in he ope a ing
g an o eal pe capi a GDP g ow h, wi h a o al alue o 0.283 (elas ici y = 0.347) using he
one and wo-pe iod-lagged coe icien s.
The app op ia eness o es ima ing (4) is e alua ed by es ing he ou es ic ions
sugges ed by Equa ions (1)–(3). Each was es ed indi idually using a Wald es a e
es ima ing Equa ions (1)–(3) as a sys em o equa ions using ixed e ec s in Equa ion (3).
None o he es ic ions we e ejec ed a 95% con idence.
Ho:γo=α2+(δ1+δ2)α1+δ3α3Fail o ejec (p=0.5396)
Ho:γ1=β1δ1+δ3π1Fail o ejec (p=0.8979)
Ho:γ2=β2δ1+β1δ2+δ3π2Fail o ejec (p=0.2321)
Ho:γ3=β2δ2Fail o ejec (p=0.1465)
