And he Oscar Goes o...peeeeedrooooo! 1 Bey Agnani and Henry Aray 2 Universiy of Granada November, 2010 Absrac In his aricle we are ineresed in how he producion of Spanish feaure films reacs o an Oscar award. We use ime series daa for he 1953-2008 period and esimae a producion funcion assuming ha he Oscar effec accrues hrough an augmening inpu facor. We consider a lag srucure ha allows he Oscar o have a diminishing effec over ime. In general our evidence suppors a posiive Oscar effec. Neverheless, he only significan Oscar award a he 1% and 5% levels is ha of Talk o Her, which Penélope Cruz euphorically announced wih her celebraed cry. Key words: Movie producion, Oscar awards, Cobb-Douglas Producion Funcion. JEL Classificaion: Z11, C13. 1 The auhors graefully acknowledge financial suppor from he Spanish Minisry of Educaion and Science and he Juna de Andalucia hrough Projecs SEJ2007-62081/ECON and SEJ-2547, respecively. 2 Corresponding Auhor: Universidad de Granada. Faculad de Ciencias Económicas y Empresariales. Dpo. de Teoría e Hisoria Económica. Campus de la Caruja S/N, Granada 18011, España. e-mails: bagnani@ugr.es, haray@ugr.es Tel: +34958241000 ex 20140. Fax: +3495849995.
1 Inroducion A review of he relevan lieraure reveals scan ineres among researchers in moion picure producion wih he excepion of he aricles by Lamson (1970) and Canerberry and Marvasi (2001). However, he same canno be said of he demand for cinema aendance. Since he aricle of Cameron (1986), which sudies cinema demand and supply in he Unied Kingdom, several researchers have examined cinema demand in heir respecive counries. For Spain, he mos imporan aricle is probably ha of Fernandez-Blanco and Baños-Pino (1997) in which he auhors use a coinegraion analysis o esimae he individual demand funcion for cinema. Macmillan and Smih (2001) sudy he case of he Unied Kingdom using a long erm VAR, while more recenly Dewener and Wesermann (2005) analyze he German case using SURE models. All hese papers concenrae on explaining he marke for cinema aendance and mos of hem esimae demand funcions. Unlike he radiional lieraure, we focus on he supply side. Specifically, we are ineresed in esing wha effec awards may have on Spanish movie producion. This aricle is relaed o Agnani and Aray (2010) who used panel daa regression o es he effecs of subsidies and inernaional awards on Spanish movie producion. In fac, hey found ha awards posiively affec he produciviy of he movie producion indusry, while subsidies have no effec. In his aricle, we focus paricularly on he Academy Awards (Oscars) due o he enormous imporance ypically aribued o hem around he world and because hey enable us o bring ogeher o a large exen he influence of all awards ha a movie could win. Before winning an Oscar, a movie has ypically won a large number of awards already. 3 In spie of he imporance normally aribued o Oscar awards in cinemaography, here are no aricles in he relevan lieraure ha quaniaively measure heir impac on movie producion in a counry wih a relaive small indusry such as Spain. 4 Neverheless, he effec of Oscar nominaions and awards on he financial success of a movie has been analyzed by Nelson e al. (2001) and Deucher e al. (2005). Our assumpion is ha winning an Oscar could be inerpreed as a posiive expecaion in general by moion picure producers in his kind of counry. Domesic and foreign demand for Spanish films migh be expeced o rise, which would imply higher expeced profis for domesic producers. Winning an Oscar can also be imporan in ha i could arac foreign invesmen o he domesic indusry. The inernaional rend in movie producion shows an increase in films made by more han one counry. This is poined ou by Hoskins e al. (1997), who highligh he increase in co-producions beween Europe and Canada, and by McCalman (2004), who claims ha higher foreign direc invesmen in he movie producion indusry leads o increased collaboraion beween counries. Our approach consiss of using a ime series of oal feaure film producion and 3 The Oscars are he las awards announced on he cinemaography awards calendar. 4 According o he publicaion Focus 2008 of he European Audiovisual Observaory, Spanish feaure film producion in 2007 accouned for abou 5% of he oal producion of he 27 European Union counries, US, Japan, India and China. 2
esimaing a Cobb-Douglas producion funcion accouning for an augmening inpu facor. We consider ha he impac of an Oscar win on movie producion is neiher saic nor fully persisen, bu vanishes over ime. We herefore inroduce a lag srucure ha allows he Oscar effec o diminish over ime. The res of he paper is organized as follows. A preliminary analysis of he daa is oulined in Secion 2. In Secion 3 we specify he economeric model o be esimaed. The main resuls are presened in Secion 4. Finally, our conclusions are presened in Secion 5. 2 Daa and Preliminary Analysis Our daa come from he Esadísicas de Cine y Audiovisuales (Cinema & Audiovisual Saisics) repor published by he Spanish Minisry of Culure. We concenrae on he period beween 1953 and 2008 and use annual daa on he oal producion of feaure films. Therefore, we include domesic films ha are wholly produced by Spanish firms and films produced joinly wih foreign parners (co-producions). Table 1 shows basic saisics for he series of feaure films. As shown, Spain produces on average 98 films per year wih a deviaion of 35 films. Figure 1 plos he evoluion of Spanish movie producion. I can be observed ha he producion of films has flucuaed considerably over ime. Indeed, here was an imporan drop in he indusry in he 1980s, which was probably relaed o he general economic reforms implemened by he governmen 5 and also due o a specific reform aimed a he cinema indusry inroduced by Pilar Miró, 6 who was in charge of he Dirección General de Cinemaografía (General Direcorae for Cinemaography) from 1982 o 1985. Given ha he main objecive of his reform was o improve movie qualiy, i may have had a negaive effec on he quaniy of movies produced. According o Figure 1, he series of feaure film producion is suspeced o have a uni roo. Therefore, we perform uni roo ess considering he following specificaions Y C Y 1 Log Y C Log Y 1 where and µ and µ are random disurbances. Table 1 shows he Dickey-Fuller ( DF ) es which assumes ha disurbances are iid and he Phillip-Perron ( PP ) es which ypically urns ou o be more powerful since i allows for serial correlaion. All τ -saisics are higher han he criical value -2.93 a he 5% significance level. Therefore, we canno rejec he null hypohesis of exisence of a uni roo. In fac, he series is saionary a aking firs Y is he producion of feaure films a each period, C and C are consans 5 Spain became a member of he European Communiy in 1986. This forced he governmen o underake major economic reforms in he early 1980s in order o mee he requiremens for enry. 6 Pilar Miró was an imporan producer, wrier and direcor in he cinema and elevision indusries. 3
difference. 7 3 The Economeric Model We consider an inpu augmening Cobb-Douglas producion funcion model as follows 8 ( A K ) α ( A L ) β Y = (1) K is he physical capial, L is he labor inpu and A is an augmening inpu facor specified as A A 1 e D 1 B 1 D 2 D 1 is a ( ) B 2 b 3 d 3 (2) where δ is a consan erm. n 1 vecor of dummy variables which akes he value of one in he following year in which he indusry wins an Oscar, and zero oherwise. In he sample period considered, he Spanish moion picure indusry as included here won four Oscars in he caegory of Bes Foreign Language Film and one in he caegory of Bes Wriing, Original Screenplay, so n = 5. The award-winning films are as follows: 9 To Begin Again, José Luis Garci (1983), Bes Foreign Language Film. Belle Epoque, Fernando Trueba (1994), Bes Foreign Language Film. All Abou My Moher, Pedro Almodóvar (2000), Bes Foreign Language Film. Talk o Her, Pedro Almodóvar (2003), Bes Wriing Original Screenplay. The Sea Inside, Alejandro Amenabar (2005), Bes Foreign Language Film. D 2 is a ( m 1) vecor of dummy variables ha akes he value of one from he year following a governmen reform of he cinema indusry, and zero oherwise. Here m = 3 wih he following reforms being considered: The appoinmen of Pilar Miró o he General Direcorae for Cinemaography, which we call he Pilar effec. The legislaion inroduced as Royal Decree 1039/1997, which allowed subsidies o be generalized o mos movies produced in Spain. This decree was superseded by Royal Decree 526/2002, which promoed independen producions by making i compulsory for TV companies o become involved in movie producion. Ac 15/2001, which alleging argumens on culural grounds, guaranees public funds for he producion of movies. We also inroduce a dummy variable, d 3, which akes he value of one from he ime 7 We also carried ou regressions inroducing a linear rend and obained similar resuls. 8 A similar specificaion of he producion funcion in (1) was used by Bergsröm (2000) o sudy he effec of subsidies on he performance of firm produciviy and by Agnani and Aray (2010). 9 Noice ha he year we ake ino accoun is he year in which he award is announced and no he year of he corresponding Oscars even, which is he previous year. 4
ha Privae TV appears in Spain (1990) o he end of he sample period, and zero oherwise. The reason for including ha variable is ha Privae TV operaors have had a very imporan roll in he producion of films in Spain. Finally, ε is a random disurbance. Taking naural logarihm o equaion (1) we obain Log Y Log A Log K Log L In he previous secion we show ha Log ( Y ) has a uni roo. Therefore, in order o avoid spurious regression, we consider Subsiuing (2) in (3) we wrie ΔLog Y ΔLog A ΔLog K ΔLog L (3) ΔLog Y D 1 B 1 D 2 B 2 b 3 d 3 ΔLog K ΔLog L (4) B 1 B 1 B 2 B 2 b 3 b 4 A drawback o he specificaion in equaion (4) is ha he Oscar would have a consan effec, and hen only in he year afer he announcemen: i has no effec for he res of he sample period. A more naural assumpion would be ha he iniial impac diminishes over ime. To overcome his shorcoming, we incorporae a simple lag srucure à la Koyck (1954) ino he equaion (4), which in his specific case allows for a more general model K ΔLog Y D 1 k B 1,k D 2 B 2 b 3 d 3 k 0 ΔLog K ΔLog L, (5) where B λ λ wih λ being he decay rae of he disribued lag k 1, k = B1, k 1 = B1, 0 srucure and whose value falls in he inerval ( 0,1]. In order o esimae he equaion model (5) for values of λ in he inerval ( 0,1], we 5
consruc he vecor of auxiliary variables, follows K k V1 = k = 0λ D1 k and wrie he model as ΔLog Y V 1 B 1,0 D 2 B 2 b 3 d 3 ΔLog K ΔLog L (6) Noice ha we consider a unique lag srucure, K, for he five dummy variables included in D 1 and we know ha he Oscar is awarded on differen daes, hus i is naural o hink ha each dummy variable in D 1 should have is own lag srucure. However, i is sraighforward o see ha wih he specificaion of he dummy variables, whenever we consider a lag srucure ha is equal o or larger han he oal lag srucure of he Oscar for 1983, ha is 24 (2008-1984), we obain he same vecor of he auxiliary variables, V 1. In fac, our model can even be inerpreed as having an infinie lag srucure since we know ha he dummy variables ake he value of zero before he beginning of our sample period. Equaion (6) ness he opposie cases of an Oscar effec only in he period following he announcemen and a fully persisen Oscar effec over ime. Thus, whaever he value of λ, if K = 0, we have he model from he equaion (4) wih V 1 D 1 and B 1,0 = B 1. On he oher hand, if λ = 1, he vecor of auxiliary variables, V 1, becomes a vecor of dummy variables ha ake he value of one from he period following he period in which he indusry wins an Oscar o he end of he sample period, and zero oherwise. In his case, he effec of he Oscar is consan over ime and persiss forever. We have o esimae he parameers in vecors B, B and b 3, δ, α and β, which can be inerpreed as he sensiiviy of ΔLog Y o each explanaory variable in equaion (6). When aemping o perform he economeric esimaion we encouner a problem. Daa for physical capial and labor inpu are no available. Therefore, in order o overcome his problem, we have o rely on proxies and make a very resricive assumpion. We consider ha he physical and human capial grow a he same rae and are equal o he growh rae of he number of firms involved in he producion of a leas one film in each period. Therefore, by leing N be he number of firms, we consider Log 1,0 2 ( N ) = Log( K ) = Log( L ) This is acually a very srong assumpion. However, i is rue ha he larger he number of firms in he indusry, he higher he physical and human capial employed in he producion of films. Therefore, we rewrie equaion (6) as ΔLog Y V 1 B 1,0 D 2 B 2 b 3 d 3 ΔLog N, (7) 4 Esimaion Issues 6
We esimae equaion (7) conrolling for heeroskedasiciy and auocorrelaion by means of a covariance marix à la Newey and Wes (1987) and considering incremens of λ of 0.1 in he inerval ( 0,1]. Therefore, we carry ou 10 regressions and choose he one 2 which provides he lowes sum of squared residuals or he highes R. The resuls are shown in Table 2. According o our crierion, we choose he esimaion wih 0.9. As can be seen, all he coefficiens of he Oscar's dummies are posiive, which sugges ha Oscar awards impac posiively he producion of films. However, he only significan Oscar coefficien a he 1% and 5% significance levels is ha of 2003 for Talk o Her by Pedro Almodóvar. A he 10% significance level, he Oscar of 1983 (To Begin Again) and 1994 (Belle Epoque) are significan. Noice ha he Oscar for Talk o Her apars from being srongly significan respec o he ohers, is effec is also much higher. As we poin ou, winning an Oscar could be inerpreed as a posiive expecaion in general by moion picure producers in a counry wih a relaively small indusry. Demand for Spanish films migh be expeced o rise, which would imply higher expeced profis for domesic producers. Therefore, an Oscar award proves o be an incenive o increase producion for acive firms and also for he enry of new firms. Much more imporanly, an Oscar win affecs he produciviy of he secor since i allows for an increase in oupu, which is no explained by an increase in inpus. Somehow our resuls would sugges an Almodóvar effec on he Spanish moion picure indusry. The popular filmmaker became well known in Hollywood when his film Women on he Verge of a Nervous Breakdown was nominaed for an Oscar in 1988 and All Abou My Moher was awarded an Oscar in 2000. Moreover, he Spanish movie indusry had already been awarded hree previous Oscars. However, i was no unil he Oscar for Talk o Her and he much celebraed and funny announcemen of Penélope Cruz, "Peeeeedroooo", ha he Spanish movie indusry began o be very popular a he inernaional level. Somehow, ha Oscar made i easier for new Spanish alens such as Penélope Cruz, Javier Bardem or Alejandro Amenabar, among ohers, o break ino Hollywood; an even which is assumed o have benefied he Spanish cinema indusry as a whole. Our lag srucure has allowed us o esimae a diminishing Oscar effec over ime. Figure 2 shows he ime-varying Oscar effecs of To Begin Again, Belle Epoque and Talk o Her on Log( Y ), since hey are he only significan effecs. As we can see, given he high value of λ (0.9), here is no a fully persisen Oscar effec, bu i vanishes over ime. In fac, afer five years i is abou a half of he iniial impac, afer en years i is abou a hird of he iniial impac and afer weny years i approaches o zero. The oal effec over ime 1 B k =. In he case of he Oscar effec of Talk o Her is can be calculaed as 0 1, B1,0 ( ) 1 λ 1.1391. Therefore, if we inerpre Log( ) Y as a growh rae, his paricular Oscar could lead o a leas a wofold increase in Spanish movie producion over ime. The oal effecs of To Begin Again and Belle Epoque are 0.6783 and 0.6372, which have no been in any case negligible. The variable collecing he effec of he inpus, ΔLog N, has had a posiive effec on he producion, wih a significan coefficien a he 1 %, as we expeced, due o he larger he number of firms are, he higher he producion of movies is. Table 2 also suggess ha he producion funcion of he Spanish moion picure indusry exhibis consan reurns o scale. This is due o he fac ha we es he hypohesis H : α + β 1 (p-values in parenheses), and are unable o rejec i a he 1% level. A 1 = 7
sronger resul was found by Agnani and Aray (2008) since hey do no even rejec he null hypohesis a he 10% level. Regarding cinema indusry reforms, he Pilar effec and he effec of Ac 15/2001 are negaive a he 1% level, respecively. Recall ha in Secion 2 we poined ou ha he movie producion reform of he early 1980s was aimed a improving he qualiy of films. On he oher hand, following concerns abou he culural imporance of domesic producion, he governmen passed Ac 15/2001 in he early 2000s. Our resuls are herefore very sriking. In spie of he movie indusry being largely financed hrough public funds, he acions aken by he governmen had negaive effecs on movie producion. In fac, wha our resuls sugges is ha here is a negaive effec of legislaion on Log( Y ) accruing hrough produciviy. The inroducion of privae elevision channels has had a posiive effec on he producion of films. As we can see in Table 2, he dummy variable for Privae TV is significan a he 5% level. Privae TV firms have become very imporan supporers of he Spanish film indusry and much more imporan hey have improve he produciviy of he secor. Finally, noe ha our specificaion accouns for abou 70% of he variabiliy of he dependen variable. 5 Conclusions This paper analyzes moion picure producion in Spain. Specifically, we es he effec of awards on movie producion. We use he Oscars since hey are considered o be he mos imporan awards worldwide. We esimae a producion funcion accouning for an augmening inpu facor and considering a lag srucure ha allows Oscars o have a diminishing effec over ime. Our general evidence suppors he exisence of a posiive Oscar effec on he produciviy of Spanish movie producion. The ime-varying Oscar effec suggess ha i vanishes over ime. However, he only Oscar significan a he 1% and 5% levels was he one awarded o Pedro Almodóvar in 2003 for Talk o Her. 8
References Agnani, B. and Aray, H. (2010). Subsidies and Awards in Movie Producion. Applied Economic Leers, 17, 1509-1511. Bergsröm, F. (2000). Capial Subsidies and he Performance of Firms. Small Business Economics, 14, 183-193. Cameron, S. (1986). The Supply and Demand for Cinema Tickes: Some U.K. Evidence. Journal of Culural Economics, 10, 38-62. Canerberry, E. R. and Marvasi, A. (2001). The U.S. Moion Picure Indusry: An Empirical Approach. Review of Indusrial Organizaion, 19, 81-98. Deucher, E., Adjamah, K., and Florian P. (2005). For Oscar Glory or Oscar Money. Journal of Culural Economics, 29, 159-176. Dewener, R. and Wesermann, M. (2005). Cinema Demand in Germany. Journal of Culural Economics, 29, 213-231. Fernández-Blanco, V. and Baños-Pino, J. (1997). Cinema Demand in Spain: A Coinegraion Analysis. Journal of Culural Economics, 21, 57-75. Focus 2008. World Film Marke Trends. European Audiovisual Observaory. Hoskins, C., McFayden, S., Finn, A. and Jackel, A. (1997). Evidence on he Performance of Canada/Europe Co-Producions in Television and Film. Journal of Culural Economics, 21, 129-138. Koyck, L. (1954). Disribued Lag and Invesmen Analysis. Amserdam. Norh Holland. Lamson, R. D. (1970). Measured Produciviy and Price Change: Some Empirical Evidence on Service Indusry Bias, Moion Picure Theaers. Journal of Poliical Economy, 78, 291-305. Macmillan, P. and Smih, I. (2001). Explaining Pos-War Cinema Aendance in Grea Briain. Journal of Culural Economics, 25, 91-108. McCalman, P. (2004). Foreign direc invesmen and inellecual propery righs: evidence from Hollywood's global disribuion of movies and videos. Journal of Inernaional Economics, 62, 107-123. Nelson, R. A., Donihue, R., Waldman, D.M. and Calbraih Wheaon. (2001). Wha's an Oscar Worh? Economic Inquiry, 39, 1-16. Newey, W. and Wes, K. (1987). A Simple, Posiive Definie, Heeroskedasiciy and Auocorrelaion Consisen Covariance Marix. Economerica, 55, 703-708. 9
Table 1. Preliminary Daa Analysis Basic saisics Sample Mean 98.1429 Sandard Error 34.9022 Uni Roo Tess Y C Y 1 DF -1.8655 PP -1.9003 Log Y C Log Y 1 DF -2.2907 PP -2.3335 P 2.93 0.05, T 50 Table 2. Esimaion of he Equaion (7) 0.9 Variable Esimae Sandard Error Consan 0.0136 0.0200 Log( N ) 0.7937*** 0.0882 Oscar83 0.0637* 0.0372 Oscar94 0.0678* 0.0362 Oscar00 0.0272 0.0555 Oscar03 0.1139*** 0.0246 Oscar05 0.0032 0.0215 Pilar effec -0.1169*** 0.0441 RD 1039/1997 0.0346 0.0548 Ac 15/2001-0.1202*** 0.0150 Privae TV 0.0448** 0.0229 R 2 0.6983 H 1 5.4720 (0.0193) *** Significan a 1% level ** Significan a 5% level * Significan a 10% level 10
175 150 125 Producion (unis) 100 75 50 25 0 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008 Years Figure 1. Spanish Feaure Film Producion. 1953-2008 0.15 Talk o Her 0.10 To Begin Again Belle Epoque 0.05 0.00 1984 1987 1990 1993 1996 1999 2002 2005 2008 Figure 2. Time-Varying Oscar Effec 11