1 Studying Device Power: The Festival Effect
Does it, as de Valck’s claims, become easier for a movie which “has been selected for a festival program, screened in a competition or perhaps even honored with an award” to be sold to cinemas and ancillary markets, because of its increased cultural value (2007, p. 38)? Like many film festival researchers introduced in Chapter 1, guidebook authors have been no less assertive of this ‘film festival effect’, writing about festivals possibly making the difference in getting distribution and movie screenings raising the profile of the filmmaker (Erickson et al., 2005, p. 340; Parks, 2007, p. 67; Tuttle, 2006, p. 189). This relationship between movies and film festivals will be explored in this chapter, adding to Chapter 5’s qualitative inquiry about economic-coordination participation by film festivals a systematic analysis of valorization by spectacle and buzz production (Reckwitz, 2017a, pp. 162–171).1
To this end, a random sample of 299 US American-released movies (all in 2006) which were subsequently screened in film festivals (hereafter: festivals) is entered into an intermediation analysis of festival participation’s impact on artistic reputation and commercial success for both art works and artists (movies and their directors). To keep the chapter slim, I have incorporated the presentation of data-related and methodological strategies, specifically the sampling procedure, the variables’ designs, and the choice of statistical techniques in the Appendix at the end of the book. The chapter starts with an overview of film performance studies, which are the relevant methodological reference, and looks at the role film festivals play in those studies. After this I describe five quantitative models which will be used to analyze the effect from film festival participation of movies. The rest of the chapter presents the findings and offers interpretations with respect to the value-adding potential of festivals as intermediaries.
1.1 Film Performance Study
Allègre Hadida’s (2008) comprehensive review of performance studies for cinema reflects the interest by applied sciences to understand more of the determinants of box-office outcomes—the effects of the classical format (see Chapter 3). Among these 135 reviewed studies, published between 1977 and 2006, only five articles combine analysis of artistic with commercial movie performance. The review usefully identifies the major and broad concept groups, i.e., film factors, organizational factors, audience characteristics, and third-party information. Film factors (e.g., artistic reputation) and organizational factors (e.g., track record of a producer) are inputs that can be controlled by industry-side actors, but third-party information, such as critical peer review and ‘word of mouth’ (shared audience opinions) are the work of judgement devices (see Chapter 5).2 In her own study, Hadida argues that artistic and commercial dimensions are distinct, and artistic value can potentially be prioritized regardless of the mutual reinforcement of artistic and commercial values (2008, p. 75). Still, performance studies are hardly interested in artistic award ceremonies (with the exception of the US American Academy Awards). Therefore, performance studies say little about the film festivals as participants in the authenticity regime (Karpik, 2010), even though festivals dominate by volume and diversity the awards architecture and certainly provide valuation in both mainstream and independent production (Baumann, 2007). In more recent years, the performance literature has been enriched by the study of non-festival awards and reviews (e.g., Gemser, Van Oostrum, & Leenders, 2008). Reviews have been studied as to whether they ‘influence’ or ‘predict’ movie success, as art-house audiences pay attention to reviews, which can be negative but not deter them from seeing the movie anyway, whereas mass audiences are assumed to pay no attention to experts as sufficiently credible sources, going instead by what other audience members make of the quality (Gemser, et al., 2008). These studies corroborate Karpik’s formulation of the mega-film and the authenticity regimes (2010). Despite the lacking attention on film festivals, these mainly US-industry focused studies are still informative. Festival operations include a variety of third-party information providers investigated in the literature, and therefore they mediate in specific ways between past prizes and future commercial or artistic success. This hypothesis of the festival effect is consistent with Gemser and colleagues’ finding that the
1.2 Device Impact for Prestigious Film Festivals
The festival effect, as I call it, has been studied by Stephen Mezias and co-authors (2011) with a sample of movies premiered in the A-list festivals of Berlin, Cannes, and Venice during 1996–2005 and including the outcome dimension of audience uptake (measured conventionally in ticket sales) in subsequent theatrical exhibition across European cinema. The study examines whether elite festivals produce attention which will generate further attention and whether this follows a status order within these three film festivals. Similar to Lampel, discussed in Chapter 4, Mezias and colleagues argue that collectively shared perceptions and information will be transformed into ‘product performance’ based on social and reputational resource located in competitive spaces. They conceptualize the effect in terms of ‘superior product identity’ bestowed by prize wins, which in turn will help find bigger audiences in subsequent circulation. They also hypothesize that prize-winning movies in which producers and directors win prizes will have bigger audiences than those with stage actors’ prizes and that the aforementioned festival events’ prestige differentials will attach themselves to the festival prizes won and therefore produce a hierarchy of value in the outcome.3 The study finds no effect from best director-prize and no differential effect on the audience-demand in terms of all prizes won at a festival. It finds that movies winning in festivals, and especially Best-Picture winners, can attract more audiences; nominations at the most prestigious film festival of the world will also attract more audiences in theatres, but there is no status-ordering effect regarding the total prize wins of a movie. Furthermore, when a movie has achieved a number of wins at a festival, the festival prestige (its distinct identity) is secondary to the audience demand.
2 Operationalization and Conjectures
The following analysis seeks to model the effect of festival participation on movies and their artistic producers. Assuming that festival awards and participations are generally influential, my analysis aims to capture status differences of a greater width by a broader sample and, furthermore, introduces measures to gauge the festival effect for different movie types, which I simply call ‘genre’ in the rest of this chapter.4 How is the device quality of the film festival translated into a feasible quantitative model? The current perception among experts, discussed at the end of Chapter 3, is that the circuit has become ‘crowded’ with competition and movie oversupply. Device effects in form of basic attention-signaling mechanisms, as we have to assume in this study, are conceptualized at singular-event and circuit level respectively. Each film festival provides its own competition for reputation. A simple inclusion/exclusion binary works at the most basic level, as permission to screen in front of a public festival audience constitutes already an achievement.
This study cannot operationalize all the complex competitive dynamics emerging from the curations, such as relative positioning of movies in the event program, which exposes films differently to judgement devices (e.g., opening night, premiere, catalog display, interpretive communication, press conferences, seminars for artists, prize competitions, etc.), and which definitely deserve attention in future studies despite the foreseeable compromise on representativeness. This complexity increases in the sense Karpik defines singularities (see Chapter 5), as on the circuit level festival locations and length of run arguably render signals that potentially alter movie identity. One must emphasize that in this particular art world valuation occurs on all ranks of the festival world. The data I collected in preparation of the empirical analysis show that only about half of the annually recorded global cohort of movies in the used database (IMDb, see the Appendix) gets to participate in festivals,
Festival managers have been observed to influence the opportunity structure, making deals with other festivals and industry actors to screen particular movies as well as through the micro dynamics of scheduling and programming of festival screenings and media events, as aforementioned. In the analysis, this complexity has to be black-boxed around a few assumptions, which are, firstly, that outcomes of movie participation in festivals, which we are just beginning to understand, reflect some of this dynamic and that, secondly, festivals strategically develop the capacity to attract certain types of art works in correspondence with their missions and their organizational reputations.
Until now, no systematic evidence about micro-level strategies by filmmakers with respect to creating their participations is available. The interviews undertaken for this book, however, suggest that these strategies are constrained by film factor and resources and that the behavior is ‘satisficing’ (Simon, 1957). It makes sense to assume that at the aggregate level (the circuit) any effects from micro-level submission strategy patterns of film makers are greatly diminished by the overwhelming power of devices.6 The major variable reflecting film festival participation is the ‘festival run’. The driving insight behind this dimension is that the sum of them operationalizes the collective impact of devices collected in film festivals. The following section introduces the five conjectures (denoted by ‘C’) for the festival effect, which are translated into quantitative models.
2.1 C1: Film Festival Participation Increases the Odds of a Movie Attaining Theatrical Revenue
Model 1 examines the impact of the festival in commercial terms. The hypothesis captures the wish of the festival-participating filmmaker to gain reputation leading to paying audiences, which permits remuneration for some part of past efforts. As audiences face ‘infinite diversity’, they will read movies with many festival logos (a sign of a substantial festival run) as proven quality and promise of entertainment. We can expect prestigious-event participation by
2.2 C2: Film Festival Participation Increases the Odds of a ‘Deal’ for the Movie
Model 2 examines whether festival participation affects future movie distribution and which genres are to benefit from festivals providing for distributor attention. More frequent festival participation, performing in more prestigious festivals, industry accreditation, and geographic proximity to Hollywood will positively affect the odds of distribution. Positive signals from participation in specialized circuits such as community-specific festivals as well as domestic exposure—concentration ratio of US American festivals in the festival run—are expected to raise the odds of a deal. Control for movie budget as investment into the project which instills trust in the ‘worth of the game’ is included as are control for past production achievements (past theatrical movies called ‘Wide Screen’, and television works, called ‘Small Screen’) and also for audience achievement (past prizes) for the filmmaker of the movie, as past achievements have been shown to influence movie revenue in Hadida’s review. The model includes a control for signals from the ancillary market for dvds, which may influence distributor attention either way. Similar to Model 1, control for film market participation is included to capture potential effects from the provided opportunity for contacts and facilitations for distribution deal-making.
2.3 C3: Film Festival Participation Increases the Odds of a Film Artist Making Future Movies
Model 3 captures the effect of festival participation on different types of future creative production (movies directed by the filmmaker). The dependent variables are set up so that directing future features (Model 3a) and future shorts and documentaries (Model 3b) acknowledge the different dynamics and identities associated with movie types. As most movies play in only one festival, it is hypothesized that circuit exposure can have an overall positive effect, but that more is gained by more frequent festival participation, leading to higher visibility, increasing the odds of reputation wins, and therefore creating more future project opportunities for perceived talent. One can also expect a modest positive effect on the filmmaker’s future productivity from having had their work screened in festivals with particular status identities, such as prestige, community (here the variable chosen is participation in lgbtq festivals, see the Appendix for more), festivals in the vicinity of Hollywood, US/domestic festivals and industry-accredited festivals. The models include controls for a number of organizational and film factors, such as budget, Small Screen-track record and Wide Screen-track record by genre as well as the movie director’s past prizes. Furthermore, a control for film market participation, as argued for Model 1, is included. The model tests the industry belief in a mobility pattern from shorts to features-making.
2.4 C4: Film Festival Participation Increases the Odds of a Film Artist Receiving Any Future Prizes
Model 4 hypothesizes that festival participation leads to future prizes for the filmmaker. For movies with smaller odds of future commercial exhibition (most of independent cinema by past experience), devices governing the authenticity and the expert-opinion regimes are crucial. This has been studied with peer and expert reviews (Holbrook, 1999). Festival signals bundle such device effects which potentially generate the ‘buzz’ needed for a clear ‘breakthrough’ in artist, but this effect is expected to be on the smaller side, being part of a longer-trend Matthew effect (Merton, 1968), and more likely to occur when the movie is presented in top-tier festival, or industry-accredited festivals as well as longer on the festival circuit.7 The festival effect is separated from other effects by analyzing those potentially stemming from past directed works as well as prizes for the artists received before and during the time of the movie’s festival participation. Finally, as gender hierarchy has been established for industry
2.5 C5: Film Festival Participation Increases the Odds of Prizes Awarded to a Movie
Model 5 hypothesizes a positive effect on the movie’s artistic recognition from circuit participation. Longer or more frequent festival participation can potentially result in more prizes of any type of recognition. Festivals are the places in which such opportunities are abundant and devices of the authenticity regime perform their diversity. One would expect a stronger positive effect from festival runs containing festivals with formal and media-attended competitive elements, like A-list and industry-accredited events. Positive effects from playing in community-based festivals and also, as in the previous models, from playing in festivals near Hollywood and domestically are expected. I control for filmmaker’s past directing achievements, past Small Screen-career influence as well as the effect of artists’ past prizes on a movie’s reputational future, and gender as in the previous model. Finally, because commercial outcomes and artistic outcomes can influence each other, the outcome variable of Model 2 is entered as a control for film-distribution outcome.
3 Statistical Results
This section presents the statistical results listed in Tables 10 and 11 below.
Four of the six outcome variables were significantly related to genre (feature, documentary, and short) in the bivariate analyses. More than twice as many features had a box-office income compared to documentaries, whereas shorts had no box-office income. Over twice as many features as documentaries have a listed distributor (64.3 percent vs. 30.8 percent), and only 8.5 percent of shorts found a distributor. Among future directed works, the genre difference is statistically significant for feature-directing. Artists with a feature in festivals directed four times more features in the future than their counterparts with
Characteristics of the 2006-movie sample by: movie genre, means, percentages and statistical significance (n = 299)
Overall |
Feature |
Short |
Documentary |
Test statistica |
Statistical significance |
|
|---|---|---|---|---|---|---|
Sample N |
299 |
70 |
185 |
44 |
||
100% |
23% |
62% |
15% |
|||
Outcomes: |
||||||
Box office revenueb |
6.0% |
18.6% |
0.0% |
9.1% |
˚ χ2 (2) = 33.78 |
p < 0.000 |
Future distributionc |
23.9% |
64.3% |
8.5% |
30.8% |
˚ χ2 (2) = 64.44 |
p < 0.000 |
Artist’s future movies—directed features |
0.16 |
0.36 |
0.09 |
0.14 |
F = 11.11 |
p < 0.000 |
Artist’s future movies—directed shorts & documentaries |
0.62 |
0.31 |
0.75 |
0.55 |
F = 4.65 |
p < 0.102 |
Artist’s future prizes |
0.21 |
0.31 |
0.15 |
0.30 |
F = 1.09 |
p < 0.338 |
Movie’s future prizes |
0.52 |
1.06 |
0.29 |
0.66 |
F = 6.40 |
p < 0.012 |
Festival Circuit Participation: |
||||||
One festival |
61.0% |
50.0% |
66.0% |
56.8% |
χ2 (2) = 5.77 |
p < 0.056 |
Two-three festivals |
20.0% |
24.3% |
19.5% |
18.2% |
χ2 (2) = 0.885 |
p < 0.642 |
Four or more festivals |
19.0% |
25.7% |
14.6% |
25.0% |
χ2 (2) = 5.548 |
p < 0.065 |
Sundance |
4.0% |
7.1% |
3.2% |
0.0% |
χ2 (2) = 4.150 |
p < 0.126 |
fiapf accredited festival |
6.0% |
17.1% |
2.7% |
4.6% |
χ2 (2) = 18.08 |
p < 0.000 |
Queer-community festival |
3.0% |
4.3% |
2.2% |
4.6% |
χ2 (2) = 1.201 |
p < 0.549 |
Festival near Hollywood |
12.0% |
12.9% |
13.5% |
6.8% |
χ2 (2) = 1.49 |
p < 0.475 |
Ratio US festivals to total festival run |
95.0% |
90.0% |
96.2% |
97.7% |
F = 4.81 |
p < 0.009 |
Film and Artist Factors: |
||||||
Movie’s budget ($million US) |
0.92 |
3.86 |
0.01 |
0.06 |
F = 20.71 |
p < 0.000 |
N of artist’s prior Widescreen works—directed shorts & docs |
0.66 |
0.51 |
0.71 |
0.64 |
F = 0.64 |
p < 0.527 |
N of artist’s prior Widescreen works—directed features |
0.40 |
1.00 |
0.11 |
0.66 |
F = 11.24 |
p < 0.000 |
Artist’s past Smallscreen works |
28.0% |
35.7% |
25.4% |
27.3% |
χ2 (2) = 2.689 |
p < 0.261 |
N of prizes before festival run |
0.42 |
1.07 |
0.22 |
0.23 |
F = 11.37 |
p < 0.000 |
N of prizes during festival run |
0.47 |
0.69 |
0.37 |
0.55 |
F = 1.38 |
p < 0.252 |
Controls: |
||||||
Ancillary market premiere (dvd) |
0.29% |
1.9% |
0.01% |
0.05% |
χ2 (2) = 56.92 |
p < 0.000 |
Participation in film market |
2.0% |
4.3% |
1.1% |
2.3% |
χ2 (2) = 2.67 |
p < 0.263 |
Female movie director |
20.0% |
12.9% |
18.4% |
36.4% |
χ2 (2) = 9.987 |
p < 0.007 |
Logistic regression of commercial outcomes and negative binomial regression of artistic outcomes for movies and filmmakers participating in film festivals on festival circuit, film and artistic factors, and controls (n = 299)
Commercial outcomes |
Artistic outcomes |
|||||
|---|---|---|---|---|---|---|
Model 1 |
Model 2 |
Model 3a |
Model 3b |
Model 4 |
Model 5 |
|
Box office revenue |
Future distribution |
Artist’s future features |
Artist’s future shorts/docs |
Artist’s future prizes |
Movie’s prizes |
|
Odds ratio (c.i.) |
Odds ratio (c.i.) |
irr (c.i.) |
irr (c.i.) |
irr (c.i.) |
irr (c.i.) |
|
Festival circuit participation: |
||||||
Playing in two or three festivals (Reference: one festival) |
0.49 (0.04–5.6) |
0.53 (0.2–1.7) |
0.42 (0.2–1.1) |
1.71* (1.09–2.7) |
0.59 (0.2–2.3) |
1.96 (0.9–4) |
Playing in four or more festivals (Reference: one festival) |
1.4 (0.2–10) |
1.93 (0.7–5.5) |
0.21** (0.07–0.7) |
1.44 (0.84–2.5) |
1.68 (0.4–7.3) |
4.98*** (2.4–10.4) |
Playing in Sundance Film Festival |
3.03 (0.2–60) |
0.28 (0.01–8.4) |
1.99 (0.4–10.4) |
0.20 (0.02–1.6) |
1.22 (0.05–28) |
1.74 (0.5–6) |
Playing in a fiapf-accredited festival |
14.46* (1.6–132) |
2.75 (0.2–41) |
1.28 (0.4–4.6) |
0.45 (0.13–1.6) |
1.30 (0.0–3.3) |
1.32 (0.4–4) |
Playing in a Queer-Community festival |
— |
8.94* (1.4–55.7) |
2.95 (0.87–10) |
1.66 (0.6–4.5) |
— |
— |
Playing in festivals near Hollywood |
5.01 (0.6–41.4) |
0.99 (0.3–3.4) |
2.31 (0.85–6.3) |
1.02 (0.6–3.8) |
0.32 (0.05–2.2) |
0.82 (0.3–2) |
Playing in U.S. American festivals |
0.98 (0.07–14) |
0.21 (0.04–1.2) |
0.42 (0.2–1.2) |
1.60 (0.7–3.8) |
0.63 (0.07–5.5) |
0.59 (0.2–1.9) |
Movie category: |
||||||
Short film (Reference: feature film) |
(a) |
0.07*** (0.03–0.2) |
0.28*** (0.14–0.6) |
1.82* (1.07–3.1) |
0.70 (0.2–2.6) |
0.61 (0.3–1.3) |
Documentary film (Reference: feature film) |
2.28 (0.4–13) |
0.27* (0.09–0.8) |
0.44 (0.2–1.4) |
1.25 (0.6–2.5) |
0.94 (0.2–5.1) |
0.72 (0.3–1.9) |
Film and artist factors: |
||||||
Movie’s budget (in million U.S. dollars) |
1.21*** (1.1–1.4) |
1.02 (0.95–1.1) |
— |
— |
— |
— |
Artist’s past Wide Screen works (All) |
— |
1.12 (0.9–1.4) |
— |
— |
— |
— |
Artist’s past Wide Screen works (Shorts/Docs) |
— |
— |
1.27** (1.07–1.5) |
1.35*** (1.18–1.6) |
1.49 (1–2.2) |
1.08 (0.9–1.4) |
Artist’s past Wide Screen works (Features) |
— |
— |
1.16* (1.01–1.34) |
1.08 (0.9–1.3) |
1.00 (0.6–1.6) |
0.90 (0.7–1.2) |
Artist’s past Small Screen works |
— |
1.00 (0.42–2.4) |
0.67 (0.31–1.42) |
0.78 (0.5–1.2) |
0.73 (0.24–2.3) |
1.28 (0.7–2.4) |
Any prizes received before festival run |
— |
1.19 (0.76–1.9) |
1.04 (0.86–1.27) |
0.61* (0.42–0.9) |
1.11 (0.74–1.7) |
1.06 (0.8–1.4) |
Any prizes received during time of festival run |
— |
— |
— |
— |
1.77* (1.1–2.8) |
— |
Model-specific controls: |
||||||
Ancillary market premieres (Model 2) |
— |
11.94* (1.3–110) |
— |
— |
— |
— |
Participated in film market (Model 2) |
4.24 (0.21–86) |
19.48* (1.4–276) |
— |
— |
— |
— |
Gender of movie director (Model 4) |
— |
— |
— |
— |
2.11 (0.6–8) |
0.77 (0.3–1.7) |
Distribution status outcome (Model 5) |
— |
— |
— |
— |
— |
1.37 (0.6–2.9) |
Pseudo R-Square |
0.40 |
0.37 |
0.17 |
0.07 |
0.10 |
0.10 |
N |
114 |
260 |
299 |
299 |
299 |
299 |
Among the film and artist factors, I find a statistically significant genre difference for movie budget, with the average budget for the feature being 3.86 US million dollars, compared to 10,000 for shorts and 60,000 US dollars for documentaries.9 Artists presenting their features in festivals had directed ten times more features in the past than those with shorts on the circuit, and one and a half times more past features than those with documentaries participating. More than a quarter of the film artists had prior experience in Small Screen work, a finding not significantly different by genre. Among artists’ past prizes, there is only a statistically significant result for prizes won before the movie’s festival participation started: directors presenting features had received roughly five times more past prizes than those who presented shorts or documentaries. Finally, among the model-specific controls are statistically significant genre differences for ancillary-market participation and director’s gender. Nearly all dvd premieres turn out to be features. Women directors turn up statistically significantly more frequently in the participation of documentaries (36.4 percent) than in shorts participation (18.4 percent) and even more so regarding features (12.9 percent).
Starting with the commercial outcome models, there is a positive effect on box-office performance from fiapf-festival participation (or = 14.46, p = .018) in Model 1. Movies exhibited in fiapf-accredited festivals have about 15 greater odds of making box office revenue than movies without fiapf-festival participation. A one-unit (or one million-dollar) increase in budget is associated with a 21-percent increase in the odds of making box-office income (or = 1.21, p = .001).
In Model 2, movies exhibited in lgbtq-community festivals have about 9 greater odds of finding a distributor than movies not participating in these festivals (or = 8.94, p = .019). Genre appears as an important predictor of distribution: compared to features, both shorts (or = .07, p = .000) and documentaries (or = .27, p = .016) have slimmer odds of distribution. Finally, movies with an ancillary-market premiere have about 12 greater odds of getting a distributor than those without the dvd premiere (or = 11.94, p = .029), while movies participating in film markets have nearly 20 greater odds of getting a distributor than those not presented in film markets (or = 19.48, p = .028). Model 2 also shows an effect with borderline statistical significance: more exposure in US festivals may be associated with lesser odds of getting distribution than festival movies showing in festivals abroad (or = .21, p = .08).
Turning to the findings for the artistic-outcome models, Model 3a indicates that filmmakers with comparatively longer circuit exposure (category: four or more festival participations) direct about 80 percent fewer features in the future than those who had their movie entered in just one festival (irr = .21, p = .008). Short-filmmakers with longer circuit exposure of their movies turn out to direct about 70 percent fewer features in the future than those with only one festival participation (irr = .28, p = .000). Furthermore, each additional short or documentary in the artist’s track record was associated with about a 30 percent-increase in future features (irr = 1.27, p = .005,) while each additional feature in the artist’s accomplished work history was associated with roughly a 20 percent-increase in future features (irr = 1.16, p = .03). There are also two effects with borderline statistical significance in this model: firstly, a large positive effect from artists screening their festival movie in one or more
In Model 3b, artists whose movie participates two or three times in festivals are found to have directed roughly 70 percent more future shorts or documentaries than those whose movie participates only once (irr = 1.71, p = .02). Similarly, artists who presented their shorts directed about 80 percent more shorts and documentaries in the future, compared to those whose features screened on the circuit (irr = 1.82, p = .028). Furthermore, each additional directed short or documentary on the artist track-record was associated with a 35 percent-increase in future shorts and documentary-directing (irr = 1.35, p = .000). Each additional prize garnered by the artist in the past was associated with a 40 percent-decrease in future shorts and documentary directing (irr = .61, p = .014).
In Model 4, the only effect on future artist prizes emerges from prizes the artist won during the movie’s festival run. Each additional prize won in that performance window was associated with a 77 percent-increase in future artist prizes (irr = 1.77, p = .014). An effect with borderline statistical significance for past shorts and documentary-directing, discerned from the raw result outputs, indicates that each additional past artist production was associated with about a 50 percent-increase in future artist prizes (irr = 1.49, p = .056).
In Model 5, movies with four or more festival participations won nearly five times more future prizes than movies with one festival performance only (irr = 4.98, p = .000). In addition, movies with two or three festival participations potentially garner double the prizes of movies shown in only one festival (irr = 1.96, p = .08).
Based on the statistically significant findings, which together reveal a number of independent effects associated with festival participation, C1, C2, and C5 conjectures can be partly accepted, whereas C3 can be partly rejected and C4 fully rejected. The next section offers an interpretation of the findings in relation to the other Chapters of this book. The last section offers thoughts on the limitations of this modeling approach in light of future research, which this chapter hopes to encourage.
4 The Festival Effect
This statistical analysis is based on a statistically representative sample of the large product world of movies made in the United States and performing in film festivals around the world at least one time. The goal was to test whether
Overall, there is strong evidence for a festival effect. The odds of a festival-participating movie to be shown to audiences is bigger when a festival performs in a fiapf-festival as part of its festival run. Regardless of that, Hollywood’s big-budget movie strategy paying off is also confirmed among these festival-participating movies, independent of the length of exposure to qualifying devices present in film festivals. As of 2015, fiapf as the umbrella organization of producer associations worldwide had accredited 47 festivals. In my survey ahead of Chapter 1, I summarized the ‘industry and media might’ collected in accredited festivals. Among all festivals in the world, this category of festivals is the only one that may be securing audiences for classical-format exhibition. The second model that measures the effect on commercial movie circulation, this time using ‘deals’ for distribution being made, shows that regardless of the festival type, features submitted to festivals have better odds of getting distribution deals than shorts or documentaries. Festival identity with respect to being US domestic vs. international, and lgbtq-community festival also increases the odds of a deal. Film market participation and premieres in the dvd market increase the odds for distribution. Together these two commercial-outcome models provide some first evidence that festival participation, if played strategically, can ‘add value to the chain’, the topic of Chapter 5. The models suggest a number of qualifications by which the movies as singularities can enter the market.
Participation in festivals also increases the odds of gaining reputation by critical devices (as phrased by Karpik, see Chapter 5), but only with a good amount of ‘staying power’ on the festival circuit. Of all prizes garnered by this sample of movies, 84 percent were festival prizes. More circuit visibility should predict more prizes (as suggested by film festival researchers), and the bivariate analysis suggests that it is the features and documentaries that get such prizes to greater extent than shorts do. Results from a follow-up analysis with only non-circuit prizes show that festival participation has no statistically significant, independent effect, but that there are independent significant effects from participation in lgbtq-community festivals (irr = 17.25, p = .012) and Sundance participation (irr = 5.88, p = .093), while the borderline significant effect for shorts (irr = .20, p = .065) suggests potentially
The fact that no other variable in the full Model 5 (all the movie’s prizes) gauging artist factors has a statistically significant effect could be interpreted as a sign for past achievements having potentially little to do with the ‘buzz’ which a long festival run can potentially create independent of other traits that could be equally evaluated. The finding may speak to a general strength of critical devices operating in film festivals, as distribution status makes no difference in this model. Similarly, and that concerns prizes for movies and filmmakers, the gender of the filmmaker makes no difference.
The notion that festival participation with a movie increases the odds of future movies being made (operationalizing creativity and productivity) is only partly borne out by the data, as a lengthy exposure to festival devices seems bad for future feature-filmmaking but good news for future shorts and documentaries-making. This result confirms industry knowledge with respect to feature financing and production being in a different world compared to shorts and documentary-making. But it also gives reason to understand more of the grant-making patterns of the film funds (Falicov, 2016), which at least in part work as mechanisms by which artists can build a career from small grants for small projects to big finance outside the festivals for bigger movies—a pattern of sequential merit competition typical for the grants economy inspected in Chapter 13. This result could potentially mean that a festival effect, modelled by grants-economic participation in festivals, might not carry through or under conditions that still need to be specified. Fully rejected by this modeling exercise is the notion that more festival participation leads to more prizes for the artist in the future. There seems to be no ‘simple’ exposure effect on the future reputation of the artist, and other mechanisms modelled here cannot be discerned other than that a reputational career accomplished prior to this festival participation has a positive effect on the future ability to garner prizes.
No effects appear to exist related to movies playing in the top-independent cinema film festival Sundance, from playing in festivals in the proximity of Hollywood, and the disproportionately higher participation in domestic festivals. The first null-finding is surprising given the productivity of this festival in artistic, commercial, and grants-economic terms.
Overall, the results show that the festival network for movies has consequences for movies, positive ones for all types with respect to prize-winning
Finally, identity effects pertaining to film festivals can be discerned not only for fiapf-festivals, the dynamics of which need further study, but also for lgbtq-community events, which have been studied already intensively (Loist, 2013; Richards, 2017; White, 1999). The finding of this lgbtq-festival effect for distribution confirms past research (Rhyne, 2007) and interviews done with festival organizers in preparation for this book. Especially my interviews with a community-support organization met in Hollywood, Los Angeles and Park City (the site of the Sundance festival) provides evidence for a well-structured niche field of cinema, the active market-making activities such as sponsor stewardship, networking summits at Sundance and Berlinale film festivals, web-based subscription networks, skilling of film makers, pitching arranged for filmmakers with distributors and financiers, creation of awareness for the cinematic good, reaching out to broader audiences at non-community festivals, a film library held by the oldest film festival of the community, grant-making constituencies, and specific awards (the most famous being the ‘Teddy’ which is awarded as a major prize at the Berlinale).11
By the start of the millennium, lgbtq-themed movies have become a sizeable product world, with many of them having found larger audiences and winning mainstream awards, as they can have so-called crossover potential (Leung, 2010). ‘Queer Cinema’, a film movement, circulates in over two-hundred film festival organizations belonging to the community (Loist, 2013; Richards, 2017). lgbtq-cinema festivals exist all over the world, combating stereotypes around the world and providing social and art worlds for people to join (Kim, 2007). Less visibility of Queer Cinema can be found in the regions of Central and Eastern Europe, Central Asia, Africa, and, with the lone exception of Israel, the entire Middle East. Such patterns speak, of course, to the negative sanctioning of alternative sexualities, state- and religion-side cultured homophobia, populist movements’ influences and existing heteronormative gender expectations.
Loist, who reflects on the achievements of Queer Cinema circulation into broader (mainstream) performance possibilities, notes that the lgbtq-festival circuit can be lucrative for low-budget movies and that the cinema moves from
Prior to this final analysis, exploratory search for Black (African-American) film festivals did not render a similar community effect (the corresponding categorical variable prepared was therefore excluded). This may be so because Black-community film festivals in the sample were world-wide incidences, with some festivals being on the African continent, or because there are different community dynamics, including market-organizational networks, which need to be discovered to make sensible hypotheses (Dovey, 2015). The lgbtq-community festivals were more concentrated in the United States, many of which are part of a well-established sub-circuit and highlight events in cosmopolitan cities (Loist, 2013). The findings for the lgbtq-film festival effect suggests that community-building and market-organization can go hand in hand, without associational forms having to give up their specific collective identity. Finally, historians on documentary filmmaking have commented on more uptake by commercial exhibition upon festival attention increasing in the early 2000s (Ellis & McLane, 2008, pp. 341–342). The odds for distribution are much slimmer than those for features, as Model 2 has revealed, and the lack of positive effects on artistic outcomes also warrants more conjecturing and systematic research.
5 Limitations and Methodological Challenges to Device Studies
Together with the literature used in Chapter 5, this study’s findings open up new theoretical, empirical, and methodological research avenues for
The limitations incurring from the sample size, with the small N rendering disproportionally large effects that may not mean much to filmmakers and other industry professionals who seek clear answers about the magnitude of the effect, owes to efficient research work with labor-intensive media data, which in film performance research, where large N gives more power to detect smaller effects, has probably been a deterrence to work with festival data. As this still appears to be the first statistical analysis aiming to understand how festivals of any prestige level and type intervene in movie and artist future success or failure, the small sample facilitated important explorations for modeling options and aided consultation with industry insiders of emerging data patterns. Future research should also oversample features and documentaries to improve statistical power for evaluating differences within and between these groups and understand the device agency of festivals in such product worlds subjected to many regimes of economic coordination.
Future research should include a control group of movies not participating in festivals. This first study translated scholarly concerns about festivals into models for their device power, seeking to understand how different art productions and differently successful artists and movies get a lift (or not) through festival participation. A challenging limitation, for which effective designs are needed, lies in the measurement of the film festival run as well as in the festival prestige variables, which may require sequence-analytic modeling to account for the various dimensions of differences in these participations.13 Here the currently available typologies discussed in Chapter 1 may not suffice as foundations for measurement ideas and more work on these conceptions arising from the research perspectives in communication studies is needed. Finally, having performed the analysis with a national sample, it is possible that for other national populations of movies this festival effect and its underlying micro dynamics vary. Moreover, such a study of national differences tied into an international division of labor concerning market-making could bolster the pursuits of post-colonial argument on cinema and the interdependencies of independent cinema with the Hollywood hegemony (Chan, 2008) and media more broadly (Kapur & Wagner, 2011; Wagner, 2015).
To ease the reading, I omitted most of the references to the relevant literature, which has already been provided in the previous chapters.
To add a second concept for third-party information to the models performed in Chapter 6, one of the most sensible measures of critical expert review external to film festivals, the RottenTomato.com’s t-meter (Holbrook & Addis, 2008), was collected but could not be used because of causal measurement conflict.
I use the term ‘stage actor’ when I refer to artists, as I use the term ‘actor’ for sociological formulations throughout the book.
The major distinctions are feature, documentary, and shorts. Rather than selecting from the ‘infinite variety’ (e.g., drama, fantasy film, comedy, thriller, etc.), the study aims to include the breadth of cinema by selecting on movie type.
The timing of events on the festival circuit is inextricably linked to the product cycle of the industry, which corresponds with the industry seasons for cinema and television. Premiere status follows from two attributes: the first public screening of a movie and the ability of the festival to claim an organizational status as premiere festival.
Even where cronyism or legitimate promotion of talent exists (see Falicov, 2016), the effects of such interventions may have, just like attention from first awards and first-time awards, an effect on early outcomes while eroding over time substantially.
The Matthew Effect is the well-known theorem of accumulated advantage.
In 2010, Kathryn Bigelow became the first woman in history to win an Oscar for best directing. In 2015–2016, Hollywood artists, both women and men, accomplished to create media awareness for the gender discrimination in the industry, which may lead to more opportunities of women filmmakers in the future.
Many budget values had to be imputed, as most were missing and could not be obtained from the movie producers in correspondence. Mezias and co-authors’ first version of their study reports a mean of four million U.S. dollars (see Mezias, et al., 2008). Montal’s suggested sample budget is 15,000 US dollars, recommended for an “indie film with the goal of going to ten major festivals” (2004, p. 323).
A first-ever lesbian and gay film market for distributors, sales agents, and independent film producers is associated with the San Francisco International Lesbian & Gay Film Festival and started in 1995.
An often cited example of a popular ‘crossover’ product is Brokeback Mountain by director Ang Lee (2006), which won a prestigious prize at Venice film festival and became a global audience success.
In further work, film festival sequences must also be hypothesized in terms of festivals’ relationship to the industry and big-awards calendars for strategic device work (Berra, 2008, pp. 26–77).