Engel’s approach as a tool for estimating consumer behaviour

15 2, XX, 2017 Economics DOI: 10.15240/ ul/001/2017-2-002 In oduc ion Consume ’s p e e ences and weal h (income, esou ces) a e iewed as exogenous quan i ies in he neoclassical heo y. Me hodological indi idualism, a ional beha iou , equilib ium and pe ec in o ma ion o consume a e he u he ea u es o his concep . Simple p ecau ions a e he necessa y condi ion o model applica ion. Me hodological indi idualism assume, ha he p inciples o he indi idual subjec s beha iou a e he mos impo an de e mining ac o s o he model unc ioning. These subjec s ac acco ding o maximizing o he o al u ili y in he gi en condi ions. The concep o equilib ium is s a ic. The pe ec in o ma ion p econdi ion need no o be necessa y ul i lled because he con empo a y concep s o he consume ’s beha iou unc ion in he condi ions o isk and unce ain y (Agui e So elo & Block, 2014). Ins i u ional concep s o consump ion a e de i ed om he T. Veblen’s heo y o he leisu e class consump ion (G zega, 2015). The p e e ences a e in l uenced by he posi ion o he consume subjec in he social hie a chy and by ac i i y o i ms. Acco ding o J. Galb ai h, i ms can pa ially c ea e he desi es o he consume s. No only he p e e ences, esou ces, and he indi idual beha iou , bu also social ins i u ions a e impo an ac o s o consump ion in he ins i u ionalism concep s. Consump ion is iewed also as a social beha iou . Psychological aspec s o he consume subjec s’ beha iou a e also impo an . In he mode n concep o consump ion, he e mus be iewed all aspec s o i . Consume beha iou is based on he decision-making o indi iduals spending hei own esou ces (i.e. ime, money and e o s) in o de o ob ain he i ems associa ed wi h consump ion No o ný and Duspi a (2014); Ho áko á (2015). This o m o beha iou in ol es he easons why, when, whe e, how o en and wha people buy, how o en hey use he pu chased i ems, how hey e alua e hem a e he pu chase, and in which way hese ac o s in l uence hei u u e pu chases Š édl, Soukup and Se e o á (2013); Wanninayake W.M.C. Banda a (2014). Ma ke ing is so much mo e han c ea ing a ca chy ph ase o a jingle people will sing o days. Unde s anding consume beha io is a i al aspec o ma ke ing. Consume beha io is acco ding o S á ko á e al. (2006) he s udy o how people make decisions abou wha hey buy, wan , need o ac in ega ds o a p oduc , se ice, o company. This issue was s udied by K ishna and Schwa z (2014) in de ail and in ela ion o psychology aspec s. I is c i ical o unde s and consume beha io o know how po en ial cus ome s will espond o a new p oduc o se ice (Bačík, Ga u o á, & Fedo ko, 2015). I also helps companies iden i y oppo uni ies ha a e no cu en ly me (Ga u o á, Vagaso a, & Ko ac, 2016). Th ough esea ch and obse a ion, se e al models ha e been de eloped ha help u he explain why consume s make decisions, including black box, pe sonal a iables, and complex models. The black-box model is based on ex e nal s imulus- esponse, meaning some hing igge s he consume o make buying decisions ha a e in l uenced by many ac o s, including ma ke ing messages, sampling, p oduc a ailabili y, p omo ions, and p ice (Michalski, 2008; And ejo ska & Banocio a, 2014; Raiso a & Du co a, 2014). When in l uenced by he pe sonal- a iable model, consume s make decisions based on in e nal ac o s. These in e nal ac o s may include pe sonal opinions, belie sys ems, alues, adi ions, goals, o any o he in e nal mo i a o (Dúb a ská, Mu a, Ko ulič, & No o ný, 2015). The complex model o consume beha io conside s bo h in e nal and ex e nal a iables. ENGEL’S APPROACH AS A TOOL FOR ESTIMATING CONSUMER BEHAVIOUR Rena a Benda-P okeino á, Kamil Dobeš, Ladisla Mu a, Ján Buleca EM_2_2017.indd 15EM_2_2017.indd 15 14.6.2017 9:29:1514.6.2017 9:29:15 16 2017, XX, 2 Ekonomie Human ood choice is a complex phenomenon in l uenced by a wide ange o ac o s (Hes e al., 2008; Cza niewski, 2014). The heo y o planned beha iou o e s a means o ying o unde s and he oles o some o hese ac o s. E idence will be p esen ed o he inclusion o measu es o sel -iden i y and mo al obliga ion. P oblems wi h he measu emen o pe cei ed con ol and possible means o o e coming hem will also be discussed. The choice o oods is an a ea o conce n o many people in ol ed in he p oduc ion and dis ibu ion o oods, and o hose conce ned wi h nu i ion and heal h educa ion. Rela i ely li le is known abou how and why people choose he oods ha cons i u e hei die s o abou how hei choices can be in l uenced in an e ec i e way. Like any complex human beha iou , ood choice will be in l uenced by many ac o s. The e a e a numbe o models in he li e a u e, which seek o delinea e he e ec s o likely in l uences. Pique as-Fiszman and Jaege (2016) concen a e on he ques ion why consume s associa e ce ain emo ions wi h ood/be e ages and meal occasions. This opic is impo an as people’s ood- ela ed beha iou and a i udes a e shaped by hese associa ions. Howe e , many such models a e simply ca alogues o he likely in l uences. Few o hem p esen any indica ion o he likely mechanisms o ac ion o he mul i ude o ac o s iden i i ed, no do hey quan i y he ela i e impo ance o ac o s o allow any quan i a i e es s which a e p edic i e o ood choice. Al hough such models can be use ul in poin ing o he a iables o conside in s udies in his a ea, hey do no p o ide a amewo k o quan i a i e modelling o ood choice beha iou . The ac o s in l uencing ood choice a e ca ego ized as hose ela ed o he ood, o he pe son making he choice and o he ex e nal economic and social en i onmen wi hin which he choice is made (Hildenb and, 1994; Vi gle o a, Dobes, & Voj o ic, 2016). Some chemical and physical p ope ies o he ood will be pe cei ed by he pe son in e ms o senso y a ibu es, e.g. l a ou , ex u e o appea ance. Howe e , pe cei ing hese senso y a ibu es in a pa icula ood does no necessa ily mean ha a pe son will o will no choose o consume ha ood. I is he pe son’s liking o ha a ibu e in ha pa icula ood which will de e mine whe he o no he ood is chosen. O he componen s in he oods will ha e e ec s upon he pe son, e.g. educing hunge , and he lea ning o he associa ion be ween he senso y a ibu es o a ood and i s consequences appea s o be a majo mechanism by which p e e ences de elop. Ma ke ing and economic a iables, as well as social, cul u al, eligious o demog aphic ac o s a e also likely o be e y impo an . This pape applies a beha io al app oach o consump ion o connec di e ences in sa ia ion pa e ns be ween inna e needs wi h sys ema ic changes in consump ion expendi u es when income ises. I is conjec u ed ha , a he le el o agg ega e expendi u e da a, hese di e ences ansla e in o di e en income elas ici ies o demand o g oups o goods and se ices ha a e likely o be consumed o se e hose needs. This way, i should be able o explain a good deal o he di e ences in he shapes o Engel cu es o he unde lying goods and se ices (Lewbel, 2008; Lades, 2013). We illus a e in he pape he echnique o Engel cu e compu a ion on he household’s income and expendi u e da a o igina ed om he S a is ical O i ce o he Slo ak Republic. Resul s o he me hod will be p esen ed as a se o elas ici ies, which p o ide deepe knowledge abou consume ’s beha io o he Slo ak households. 1. Theo e ical Backg ound o Consume Beha io Fac o s ha in l uence he pu chasing beha io o consume s a e e y signi i can o businesses, because based on hese ac o s, i is possible o ocus he ade policy, which hen could lead o be e business esul s (e.g. an inc ease in sales olume, ma ke sha e, dis ibu ion o a po olio o clien s, e c.). E e y human is a consume , and o he o he consume s ep esen s he key ac o s ha in l uence hei pu chasing beha iou . Theo y and p ac ice nowadays lacks g ea e links be ween businesses and consume s, based on segmen a ion by app op ia e segmen a ion c i e ia. Mo e de ailed desc ip ion o segmen a ion c i e ia highligh s (Koudelka, 2005). In many a eas o law such connec ions, including esea ch ha would acili a e company decisions is inadequa e. The e a e a numbe o esea ch p ojec s ela ed o li es yle, d ess, smoking, ea ing (Šol és & Ga u o á, 2014), bu esea ch including models and ac o s EM_2_2017.indd 16EM_2_2017.indd 16 14.6.2017 9:29:1514.6.2017 9:29:15 17 2, XX, 2017 Economics in l uencing pu chasing beha iou consume s in many manu ac u ing indus ies is missing. The issue o pu chasing beha io o consume s was discussed by se e al p ominen au ho s (Ko le & A ms ong, 2004; Solomon, Ma shall, & S ua 2006; S á ko á e al., 2006; and o he s). The assump ion is ha he consume will ac a ionally and will maximize he o e all bene i ha is associa ed wi h a combina ion o goods and se ices ha accompany he p oduc . Acco ding o Hes e al. (2008), howe e , i is limi ed by he consume when making decisions, especially by he i nancial esou ces. In many i elds linking businesses and consume s, including esea ch hei beha io ha would acili a e decision o company managemen is insu i cien . One o o he app oaches o he analysis o consume beha iou is also moni o ing beha iou when he e is a change in in o ma ion cos . This may a y ac oss consume s, which is he case when consume s who sea ch o in o ma ion a e “bounded a ional” – making andom e o s due o limi ed a en ion (Manzini & Ma io i, 2014) o “ a ionally ina en i e” – op imally alloca e a en ion (Ma ějka & McKay, 2015). Following hese app oaches, i is also impo an o s udy how o op imally in o m a consume abou he alua ions o mul iple goods when a consume can lea n abou ewe goods han he/she can consume (Saak, 2016). Ma ke o ien a ion is an app oach based on ma ke ing concep ion, in oduced by D ucke (1954). Al hough he p inciple o ma ke o ien a ion has been known o six y yea s, de ailed analysis only began a he beginning o he 1990s. In 1990, he Jou nal o Ma ke ing published he esul s o wo s udies sponso ed by he Ma ke ing Science Ins i u e. Those s udies we e made by Kohli and Jawo ski (1990) and by Na e and Sla e (1990). These wo pieces o esea ch had a subs an ial impac on he o ma ion o a new app oach o ma ke ing and subsequen s udies on ma ke o ien a ion and ook he ini ial s eps owa ds c ea ing a de i ni ion o ma ke o ien a ion. The de i ni ion o Kohli and Jawo ski (1990) was o ally di e en o hose by Na e and Sla e (1990). The de i ni ion by Kohli and Jawo ski (1990) pe cei es ma ke o ien a ion as an implemen a ion, namely as an ins umen o s a egic decision-making. Al e na i ely, he de i ni ion o Na e and Sla e (1990) pe cei es ma ke o ien a ion om a business lea ning and co po a e cul u e poin o iew No o ný and Duspi a (2014). O he app oach by Š édl, Soukup and Se e o á (2013) and by Pe éz e al. (2013) we e applied in mode n ma ke economy whe e he supply exceeds demand, and he impo ance o he “consume ’s beha iou in he ma ke analysis” con inuously inc eases. Consume ’s p e e ences a e iewed as exogenous a iables in a neoclassical heo y. They a e no explained in a amewo k o he concep bu a e iewed as gi en ones. Heikkinen (2015) s udied decline and g ow h in a dynamic equilib ium model wi h p e e ence he e ogenei y. Simula ions sugges ha decline igge ed by olun a y simplici y inc eases equilib ium wel a e unde ex e nali ies in consump ion and leisu e. Me hodological indi idualism, a ional beha io , equilib ium, and pe ec in o ma ion abou he consume s a e he nex ea u es o his concep . Simple p ecau ions a e a necessa y condi ion o he model applica ion. Me hodological indi idualism means ha p inciples o indi idual subjec s beha iou a e he mos impo an de e mining ac o s o model unc ioning. These subjec s abide by he p inciple o o al u ili y maximizing in gi en condi ions. The concep o equilib ium is s a ic. Pe ec in o ma ion p ecau ions doesn’ need o be necessa ily ul i lled, because con empo a y concep s o consume ’s beha io unc ion in he condi ions o isk and unce ain y. Ins i u ional concep s o consump ion a e de i ed om Veblen’s heo y o leisu e class consump ion (He ns ein, 1997). The e is also a s ong ma ke ing in l uence. Sub amanyam and Kuma aswamy (1981) conside he e ec o ma ke ing policies and condi ions on demand. I is wo h o men ion a ew sma ma ke ing ools in l uencing cus ome beha iou . Online cus ome beha io is a new phenomenon and i is a e y impo an case o i m´s ma ke ing (Rypako a, Mo a ciko a, & K izano a, 2015). K amolis and Kopecko a (2014) deal wi h his kind o sma ma ke ing ool (P oduc placemen ) which clea ly in l uences cus ome s. Indeed, he e a e no only luxu y goods, (K amolis & D abko a, 2012) which a e p omo ed by means o p oduc placemen . This ma ke ing ac i i y has a simple pu pose – o connec b ands wi h app op ia e en e ainmen pa ne s o maximize b and ele ance –- which leads o s eng hening consume s’ a i udes o p omo ed p oduc s. This sma ma ke ing whispe ings can build he p e e ences wan ed. EM_2_2017.indd 17EM_2_2017.indd 17 14.6.2017 9:29:1514.6.2017 9:29:15 18 2017, XX, 2 Ekonomie Mo eo e , p e e ences a e in l uenced by he posi ion o a consume subjec in he social hie a chy. No only p e e ences, esou ces and indi idual beha io , bu also social ins i u ions ep esen he impo an ac o s o consump ion in ins i u ionalis s’ concep s. Consump ion is also iewed as social beha io . Ul ima ely psychological aspec s o consume subjec ’s beha io a e also impo an . The con empo a y ela ionship be ween b and pe sonali y and consume pe sonali y has become a esea ched issue in ecen yea s and i was s udied by Bane jee (2016). The au ho ´s esul shows buying decision, pe sonali ies o p omo ed p oduc s and co po a e b ands in l uencing consume s´ p e e ences. Engel cu es desc ibe he change o expendi u e (o demand) o di e en commodi ies as a unc ion o income. I is common o speak abou „c oss-sec ion Engel cu es“ i one conside s expendi u es o a popula ion o households ( o example, all households in a la ge coun ies (Bece a Alonso, And oniceanu, & Geo gescu, 2016)) a he han single indi iduals. Since he wo k o Engel (1857), c oss-sec ion Engel unc ion has been a majo esea ch in economies and econome ics. Engel cu es ha e been ex ensi ely s udied o es ablish empi ical ela ionships be ween expendi u e and income. Engel’s law s a es, ha he budge sha es o he commodi y ood dec ease wi h an income. Gene ally, Engel unc ions can be used o classi y commodi ies in o luxu ies, necessi ies and in e io goods. The mos known classical s udy o c oss-sec ion Engel cu es is ha o P ais and Hou hakke (1955). Impo an objec i e is he de e mina ion o budge elas ici ies o di e en commodi ies om he co esponding Engel cu es. In demand analysis, he s uc u e o c oss- sec ion Engel cu es has been exploi ed o build up demand sys em. Wo king (1943) p oposes he pa ame ic model m(x) = θ1x + θ2x logx o he c oss-sec ion Engel unc ion m(x) o some commodi y. He e, x deno es income, he θ1, θ2 a e unknown pa ame e s, which a e o be es ima ed om gi en da a. Dea on and Muellbaue (1980) conside his model as an impo an mo i a ion o hei amous „AIDS – almos ideal demand sys em“. When doing he analysis o ood demand, he model o AIDS was used in wo ks o Benda P okeino a and Hano a (2016), Ulubasoglu e al. (2016) and Haya e al. (2016). The i nding o Lewbel (1993), elling ha a be e pa ame ic model migh be w i en in he o m m(x) = θ1x + θ2x log x + θ3x (log x)2 is o en in e p e ed as a jus i i ca ion o he quad a ic almos ideal demand sys em o Blundell e al. (1993), which gene alizes he esul s o Dea on and Muellbaue (1980). These app oaches a e based on he idea o eco e abili y o indi idual beha io om agg ega e da a. In he p esen con ex his co esponds o he assump ion ha he s uc u e o c oss-sec ion Engel cu es e l ec s he s uc u e o indi idual ones. The same a gumen was used by Hä dle and Je ison (1991), whe e a ce ain a iance o es ima ed c oss-sec ion Engel cu es o e ime we e used o d aw conclusions on indi idual beha io . Engel cu es also play a ole in he heo e ical economics. Hildenb and (1994), and Kahneman (1997) p o ides condi ions on he s uc u e o hese cu es which ensu e he law o demand. The e is hus s ong empi ical and heo e ical in e es o analyze he c oss-sec ion Engel unc ion o eal popula ions. A p e equisi e o any economic in e p e a ion is a easonable es ima ion o hese cu es om gi en c oss- sec ion da a abou household expendi u es and income. S a is ically, c oss-sec ion Engel cu es a e eg ession unc ion, m(x) equals o he condi ional expec a ion o expendi u e o a commodi y gi en he income le el x. The classical app oaches in eg ession analysis a e based on pa ame ic models, i.e. m(x) = g(x, θ) wi h a known unc ion g and an unknown pa ame e θ. The as majo i y o s udies o Engel cu es a e based on such pa ame ic models. Apa om he Wo king model, he e a e many u he pa ame ic app oaches (double loga i hmic, semiloga i hmic, e c.). In any pa ame ic model he unc ional shape o he es ima e is al eady gi en by an assump ion. The quali y o he esul ing es ima o depends hea ily on he co ec ness o his speci i ca ion. I a model is co ec , good es ima o s o he pa ame e s and he cu es can usually be ob ained. Howe e , i he model is poo ly speci i ed, he esul ing es ima o is no e en consis en . Since he economic heo y doesn’ yield any in o ma ion abou a co ec pa ame ic model, he e is hus eason o conside he al e na i e app oaches. EM_2_2017.indd 18EM_2_2017.indd 18 14.6.2017 9:29:1514.6.2017 9:29:15 19 2, XX, 2017 Economics 2. Ma e ial and Me hods Acco ding o F iedman’s hypo hesis he consump ion o indi iduals o households depends on p income, i does no depend on he cu en income o economic uni as Keynes insis ed. Pe manen income can be explained as an a e age income, which an indi idual expec s o ea n du ing his li e using his labo and capi al. Šimso á and Reisso á (2016) deal wi h he opic o expec ed income o selec ed g oups o inhabi an s, e.g. uni e si y s uden s. Acco ding o Gombos, Kiss and Z a a (2016) i depends on household wel a e (human capi al, educa ion, heal h included), in e es a e, and expec ed job aca ions. The simila app oach was desc ibed by Engel. Engel cu es desc ibe he change o expendi u e (o demand) o di e en commodi ies as a unc ion o income om he heo e ical iewpoin . 2.1 Engel Cu es and he Income Elas ici y o Demand The gene al connec ion be ween he shape o an Engel cu e and he income elas ici y o he espec i e good can easily be illus a ed. I an Engel cu e o good i is exp essed in e ms o he expendi u es qi spen on i depending on he households’ income y, he slope o he i ed cu e dqi/dy can indi ec ly be used o de i e i′s income elas ici y o demand. As a good’s income elas ici y is de i ned by he ela i e change in qi (dqi/qi) di ided by he ela i e change in y (dy/y), i can be es ima ed by eg essing loga i hm o qi on loga i hm o y Lewbel (2008). Using expendi u e sha es (wi=qi/ywi=qi/y), ins ead o he expendi u es spen on i, acili a es mo e di ec ly an in e ence o he income elas ici y o demand om he cu e’s slope. While an inc easing slope ep esen s mo e han p opo ional expendi u es, hus luxu ies, a dec easing slope poin s o less han p opo ional expendi u es wi h ising income, i.e., necessi ies. Al hough Engel cu es usually show conside able nonlinea i ies Lewbel (2008), budge sha e Engel cu es acili a e o eadily obse e he income elas ici y o demand o he unde lying g oup o goods and se ices. 2.2 A Mo i a ional App oach o Consume Beha io Engel’s app oach o consump ion ne e solely ocused on ood expendi u es. In ac , Engel’s o iginal con ibu ion was mean o de e mine and measu e household wel a e by Chai and Mone a (2010). He s a ed by ca ego izing expendi u e i ems acco ding o he unde lying “wan s” hey sa is y. The impo ance o hese wan s was subsequen ly assessed by he empi ically de i ed expendi u e pa e ns. Engel’s wo k hus essen ially ocused on a beha io al ounda ion o he necessi y-luxu y axonomy. In a e ised e sion o his o iginal con ibu ion, Engel (1895) made e en mo e explici , ha he mo i a ion o human ac ion, and hus also consump ion beha io , is oo ed in he sa is ac ion o uni e sally-sha ed needs (Chai & Mone a, 2010). To build upon Engel’s legacy, his pape en iches he classi i ca ion o needs wi h cu en scien i i c knowledge on he na u e o consume needs and how hey a e sa is i ed. Howe e , mo i a ional unde pinnings o economic beha io in gene al and consume beha io in pa icula a e a ely add essed in economics. Among he exis ing wo ks, wo di e en explana o y app oaches we e iden i i ed by Wi (2010). While in he u ili a ian hedonic app oach he explana ion e e s o he mo i es o seeking pleasu e and a oiding pain, non- hedonis ic a ian s ocused on he mo i a ing powe ha dep i ed needs and wan s ha e o consump ion ac i i ies. This pape closely connec s o he la e app oach. A beha io al-need based– in e p e a ion o he consump ion mo i a ion was o e ed by Wi (2001; 2010) o ac ually heo e ically explain he necessi y-luxu y dis inc ion ha becomes isible h ough he shape o Engel cu es. The heo y pos ula es an in ima e ela ionship be ween human biological and cul u al e olu ion in he sense ha cul u al de elopmen is based upon as well as cons ained by inna e beha io al disposi ions and cogni i e lea ning abili ies, which ha e eme ged du ing human phylogeny. Hence, he heo y ocuses on he explana ion o long- un economic change om a biological and psychological pe spec i e (Wi , 2008). The heo y o he lea ning consume (Wi , 2001; 2010) emphasizes he ole o human needs as ul ima e mo i es o consump ion beha io . The heo y dis inguishes be ween gene ically de e mined-inna e-needs and bo h cul u e and socializa ion speci i c acqui ed wan s, ha esul om p ocesses o associa i e lea ning (classical and ope an condi ioning) EM_2_2017.indd 19EM_2_2017.indd 19 14.6.2017 9:29:1614.6.2017 9:29:16 20 2017, XX, 2 Ekonomie and social-cogni i e lea ning. The heo y holds he a emp o elie e o educe dep i a ion o a limi ed numbe o inna e needs, consequen ly an inc ease in sa ia ion le el o hese needs, is one majo mo i a ion o consume. Dep i a ion is hus seen o in insically mo i a e consume s o ac which c ea es a ewa ding expe ience. Needs a e he con ingencies unde which dep i a ion occu s. Al hough needs can be mani old, in his con ex only he subse o uni e sally-sha ed “basic” needs i ele an . Among hese a e ha o sleep, o some hing o d ink, o some hing o ea , o main aining body empe a u e, o physical ac i i y, o s a us ecogni ion, o o senso y a ousal Wi (2011). An impo an cha ac e is ic o hese needs is ha hei sa is ac ion by an ac ion e ec s a p ima y ein o cemen in he sense o ins umen al o ope an condi ioning (S addon & Ce u i, 2003). This, o cou se, has implica ions o he alloca ion o esou ces. Imagine indi iduals o be equipped wi h a i xed se o inna e needs ha show di e en need dep i a ion s a es. I indi iduals alloca e hei beha io o ob ain a educ ion o hei dep i a ion s a es, he likelihood ha a pa icula ac i i y is chosen o e ano he one depends on i s ela i e con ibu ion o educing dep i a ion. Indi iduals a e acco dingly assumed o shi hei beha io o hose al e na i es which p o ide he highes a e age ewa d. The alloca ion o income o consump ion ca ego ies esul s as p opo ional o he co esponding need dep i a ion s a es. Unde ein o cemen lea ning, he equency dis ibu ion o e ac ions con e ges o a s a e sa is ying he so called “ma ching law” He ns ein (1997), which seems o be a good app oxima ion o beha io d i en by mainly basic need dep i a ion s a es and ha is cha ac e ized by low le els o cogni i e in e en ion. Among he uni e sally-sha ed needs, a u he dis inc ion ela es o he unde lying sa ia ion p ope ies hey show Wi (2001; 2010). On he one hand, he e a e basic needs which ollow homeos a ic ea u es, i.e., dep i a ion can be educed ela i ely easily up o he empo a y sa ia ion poin once ising income allows o a su i cien inc ease in he co esponding consump ion expendi u es pe pe iod o ime. The mo i a ion o consume is, hen, empo a ily educed o emo ed. Sa is ac ion o hese needs depends mainly on he in insic alue o he co esponding goods and se ices. Examples a e he homeos a ic needs unde lying o ea ing, d inking, sleeping, and he main enance o body empe a u e. On he o he hand, he e a e also basic needs whe e homeos a ic ea u es a e absen , and whe e i is he e o e di i cul , i no impossible, o educe he a e age dep i a ion o ze o. Typically, hese a e needs whose sa ia ion le el is de i ned in ela i e e ms, like he need o a ousal and o social ecogni ion. Despi e in e pe sonal sou ces o a iance, which can be expec ed due o indi idual cogni i e and condi ioning lea ning p ocesses, i can be conjec u ed, ha sha ed inna e needs exe some sys ema ic e ec s on beha io , ha a e isible a he le el o he popula ion means, i.e., a he le el o agg ega e consume expendi u es. As needs di e wi h espec o he amoun o spending ha is necessa y o each sa ia ion, his di e ence can be expec ed o become ele an wi h ising eal income Wi (2001; 2011). Being able o spend mo e, consume s should be able o app oach he sa ia ion le el o some needs as e han he sa ia ion le el o o he needs. Thei consump ion mo i a ion is no equally upheld and hei spending should hus no expand equally. Di e ences in he income elas ici y o demand o he p oduc s ha se e he di e en needs should exp ess his di e en ial sa ia ion e ec . A o mal model o he di e en ial sa ia ion e ec s and i s implica ions o Engel cu es has ecen ly been pu o wa d by Lades (2013). Taken oge he , he beha io al app oach o consump ion sugges s connec ing he di e ences in sa ia ion pa e ns be ween inna e needs wi h sys ema ic changes in consump ion expendi u es when income ises. I is conjec u ed ha , a he le el o agg ega e expendi u e da a, hese di e ences ansla e in o di e en income elas ici ies o demand o p oduc s o g oups o goods and se ices ha a e likely o be consumed o se e hose needs. The beha io al app oach o consump ion should hus be able o explain a good deal o he di e ences in he shapes o Engel’s cu es and hus in he income elas ici ies o he unde lying goods and se ices. A double-log speci i ca ion o he Engel’s unc ion has been chosen in o de o es ima e he expendi u e elas ici y o households by using an economic s a us o households’ head a wo k o each household by child en pe EM_2_2017.indd 20EM_2_2017.indd 20 14.6.2017 9:29:1614.6.2017 9:29:16 21 2, XX, 2017 Economics pe son. The da a se we e ob ained om he Slo ak S a is ical o i ce and consis s o he yea ly obse a ions: money incomes o p i a e household by economic s a us o household head a wo k pe pe son and mon h (€); money incomes o p i a e household by child en pe pe son and mon h (€); money expendi u es o households by using and economic s a us o household head a wo k pe pe son and mon h (€); money expendi u es o households by aim o using and numbe o child en (€). The household Budge Su ey o he Slo ak S a is ical o i ce was used o he pe iod 2004-2014. The use o he household le el da a o e s he po en ial o iche da ase ha may o e an addi ional insigh in o he unde lying economic ela ionships. A Double-Log Speci i ca ion has p o en he mos app op ia e way o es ima ing he expendi u e elas ici y o demand; i gene a es mo e ealis ic expendi u e elas ici ies. The gene al model can be w i en as ollow: ln wj =αj + βj ln y +ηj (1) whe e: wj is he a e age annual pe capi a expendi u e sha e o ood g oup j, αj, and βj a e he es ima ed coe i cien s, y is he a e age o al pe capi a income calcula ed as he a e age annual o al pe capi a expendi u e, and ηj is he dis u bance e m. As poin ed ou be o e, he de i a ion o he Engel’s unc ion assumes cons an p ices. Equa ion is es ima ed o each o he 13 ood g oups o each household by child and by economic s a us o household head. We ha e o include dummy a iables. We ha e supposed he e a e di e ences in he elas ici y o each ood g oup among se e al ypes o households. Households’ Classi i ca ion Analysis o he income elas ici y was de e mined in 2,418 Slo ak households. Di e en ypes o households we e classi i ed based on he economic s a us o he household head (household o employees, sel -employed household, household o pensione s and o he s). Fo be e o ien a ion i is essen ial o cha ac e ize he household ypes as de i ned by he S a is ical O i ce o he Slo ak Republic: Economic s a us o household head was cha ac e ized by his/he employmen s a us. Employee: a pe son, who wo ks o a public o p i a e employe , and who ecei es compensa ion in he o m o wage o sala ies, including pensione s wi h income om employmen and co-ope a i e membe s. Sel -employed pe son: wo ks in hei own business, including pensione s wi h income om own business. Pensione : a pe son wi hou job, who ecei es e i ed pension, wi h no income om job o business; in household could li e also o he pe sons wi hou old-age pension. O he s: all pe sons, which we e no included he ein be o e (unemployed, ma e ni y o pa e ni y lea e, s uden s, e c.). 3. Resul s and Discussion 3.1 Analysis o he Income Elas ici y o Demand Quan i y demanded sensi i i y o changes in income, ce e is pa ibus, was measu ed by he coe i cien o income elas ici y. The income elas ici y o demand is he a io o he pe cen age change in quan i y demanded p oduc X o pe cen age change in income, and shows how many pe cen will change he demand o p oduc X i he e is 1 pe cen change in consume income. Fo no mal goods he income elas ici y is posi i e. As we deepen ou classi i ca ion o goods o essen ial goods, he 1% change in consume ’s income caused less han 1% change in demand o he goods. Because he sha e o he goods on a o al income dec eases, he a e age p opensi y o consume is dec easing and ue: 0 ≤ Eid ≤ 1. Analysis o he income elas ici y was in es iga ed o each ype o households by numbe o dependen child en and o households by economic s a us o he household head. F om he Tab. 1 i could be iden i i ed ha he calcula ed elas ici y we e highly signi i can . All ood g oups ha e explained a iabili y o e 90%. Almos e e y elas ici y eached he posi i e alue, excep hese o mea and oils and a s. This means, ha all he ood g oups wi hin he in e al 0 ≤ β ≤ 1 a e he Slo ak households wi hou child en needs. 1% change in he consume ’s income esul s in a change in demand o he good which is less han 1%. EM_2_2017.indd 21EM_2_2017.indd 21 14.6.2017 9:29:1614.6.2017 9:29:16 22 2017, XX, 2 Ekonomie The highes alue o he elas ici y was ound in case o ege ables, including po a oes. I we ha e inc eased he e enue by 1%, he demand changed by 0.66%. So d inks, co ee, ea, cocoa, and mine al wa e ha e ela i ely simila le els o elas ici y a he le el o 0.5%. Fish, milk, cheese, and eggs showed he elas ici y a ound 0.2%. The consump ion o hese p oduc s is ela i ely li le a ec ed by he change o income. We ha e ocused on nega i e elas ici y mainly. Ve y sensi i e commodi y was he mea . F om a nu i ional poin o iew his is an impo an ood, and in i nancial e ms, i ep esen s one o he mos expensi e ood o he Slo ak households. This ac was indica es by he elas ici y alue -0.08. 1% change in income will su ely esul s in change o consume p e e ences o mea by -0.08%. Fa s & oils ha e e en mo e nega i e elas ici y alue, eaching -0.22% only. In case o households wi hou child en he needs o goods such as mea , oil & a s we e only in e io , demand o hem dec eases wi h an inc easing income. Fig. 1 shows he di e ences in e i emen elas ici y o households acco ding o numbe o child en. The e is a signi i can di e ence be ween households wi h child en and households wi hou hem. The numbe o child en pe household is no he mos impo an ac o o a change in e i emen elas ici y. The analysis o a iance among he g oups o households s udied shows ha he null hypo hesis H0 canno be ejec ed: All popula ion Household wi hou child en Household wi h 1 child Household wi h 2 child en Household wi h 3 and mo e child en αβP ob R2 EID αβ EID αβ EID αβ EID p ob R2 B ead & ce eals -1.04 0.55 1.39E-15 0.982 necessa y goods -1.38 0.75 necessa y goods -1.38 0.81 necessa y goods -1.38 0.93 necessa y goods 1.23E-17 0.96 Mea 1.28 -0.08 7.67E-16 0.983 in e io goods 0.43 0.37 necessa y goods 0.43 0.26 necessa y goods 0.43 0.22 necessa y goods 1.77E-15 0.93 Fish -2.18 0.21 4.96E-18 0.991 necessa y goods -2.48 0.5 necessa y goods -2.48 0.4 necessa y goods -2.48 0.35 necessa y goods 2.24E-16 0.94 Milk, cheese & eggs 0.06 0.29 9.93E-12 0.949 necessa y goods -0.77 0.55 necessa y goods -0.77 0.54 necessa y goods -0.77 0.54 necessa y goods 7.07E-14 0.91 Oil & a s 0.13 -0.22 7.33E-14 0.972 in e io goods -0.21 0.23 necessa y goods -0.21 0.1 necessa y goods -0.21 0.1 necessa y goods 2.86E-09 0.8 F ui s -1.95 0.3 5.44E-13 0.964 necessa y goods -1.67 0.44 necessa y goods -1.67 0.36 necessa y goods -1.67 0.29 necessa y goods 2.51E-12 0.88 Vege ables incl. po a oes -3.3 0.66 1.17E-10 0.932 necessa y goods -3.6 0.92 necessa y goods -3.6 0.88 necessa y goods -3.6 0.94 necessa y goods 1.70E-11 0.87 Suga , jam, honey, chocola e & con ec ione y 0.19 0.01 6.01E-13 0.964 necessa y goods 0.23 0.2 necessa y goods 0.23 0.14 necessa y goods 0.23 0.1 necessa y goods 1.25E-09 0.81 Co ee, ea & cocoa -3.11 0.5 2.48E-14 0.975 necessa y goods -3.22 0.7 necessa y goods -3.22 0.67 necessa y goods -3.22 0.66 necessa y goods 1.52E-15 0.93 Mine al wa e s, juices & o he non-alcoholic d inks -2.18 0.59 6.77E-11 0.936 necessa y goods -3.07 0.68 necessa y goods -3.07 0.68 necessa y goods -3.07 0.58 necessa y goods 1.73E-20 0.97 Alcoholic be e ages & obacco -1.86 0.47 4.88E-14 0.973 necessa y goods -1.86 0.58 necessa y goods -1.86 0.42 necessa y goods -1.86 0.29 necessa y goods 1.63E-16 0.95 Alcoholic be e ages -2.38 0.35 7.13E-17 0.987 necessa y goods -2.47 0.58 necessa y goods -2.47 0.54 necessa y goods -2.47 0.46 necessa y goods 9.65E-15 0.92 Sou ce: own calcula ions Legend: α-in e cep , β- eg ession coe i cien , EID-income elas ici y o demand, p ob.-p obabili y alue, R2-coe i cien o de e mina ion Tab. 1: Elas ici y o households acco ding o numbe o child en pe pe son and mon h in 2004-2014 (in €) EM_2_2017.indd 22EM_2_2017.indd 22 14.6.2017 9:29:1614.6.2017 9:29:16 23 2, XX, 2017 Economics wi h ega d o elas ici y is equal (p- alue = 0.14738). The e o e, he e is no s a is ically signi i can di e ence among he households acco ding o numbe o child en. Analysis o he ela ionship be ween models o consume demand based on mic o and agg ega e da a we e p o ided by Blundell e al. (1993). Acco ding o hei i ndings i is impo an o es ablish he p esence o nonlinea i y in he mic o-le el Engel’s cu es and he need o in e ac ions wi h household- speci i c cha ac e is ics, since ei he o hese would ule ou simple linea agg ega ion. In ou sample, pooled o e 10 yea s, we ha e ound s ong e idence o bo h. Simila o ou esul we e he i ndings o he oods u model desc ibed by Chai and Mone a (2010), in which he p ices o co n and ui s & ege ables posi i ely a ec income; he p ices o dai y p oduc s a ec s he income nega i ely. In he wage income model on he o he side he elas ici y o co n was he same as in p e ious models; he elas ici y o dai y p oduc s was signi i can ly nega i e, bu smalle in absolu e alues (-0.58 s. -1.41, and -1.10); and he elas ici y o ui s & ege ables was no longe s a is ically signi i can . The e we e mino di e ences in he es ima es o he demand elas ici y ac oss he models. Income elas ici y was sligh ly abo e he le els o cheese & gible s, and no ably abo e he alues o bo ine (1.43), o he ce eals (1.37), and mu on (1.50). Income elas ici y alls be ween 0.5 and uni y o i sh (0.56), b ead (0.64), eggs (0.66), ice (0.74), yogu (0.75), milk (0.79), poul y (0.87), a s (0.88), dough p oduc (0.95), and oils (0.98). Ou income elas ici y o b ead cons i u es he lowe limi o ea lie es ima es, which ange om 0.65 o 0.85, while income elas ici ies o mea , mea p oduc s, dai y p oduc s, and eggs we e simila o hose epo ed by Hes e al. (2008), Lades (2013), and Wi (2001; 2011). The es ima ion esul s indica e ha he o al pe capi a expendi u e and he household size ha e signi i can ly nega i e es ima es in bo h g oups o households. The es ima e o emi ances is posi i e and signi i can only in he g oup o ich households, so when he emi ances inc ease, households abo e he po e y line inc ease hei sha e o ood consump ion Na e (1990). Households wi h 1 child has he highes alue o he elas ici y o ege ables including po a oes and b ead and ce eals. I we inc eased e enue by 1%, he demand o ege ables inc eased by 0.92% and he demand o b ead and ce eals should be inc eased by 0.75%. The lowes Fig. 1: Box plo o elas ici y o households acco ding o numbe o child en Sou ce: own EM_2_2017.indd 23EM_2_2017.indd 23 14.6.2017 9:29:1614.6.2017 9:29:16