As Hollywood tallies its year-end totals, a new study released Feb. 5 offers insight into how gender and race/ethnicity of the lead/co-lead character is related to economic success.

The report entitled “The Ticket to Inclusion,” was commissioned by ReFrame, a collaboration between Sundance Institute and Women In Film, and is authored by Associate Professor Stacy L. Smith and the Annenberg Inclusion Initiative (AII) at USC Annenberg, and by Professor René Weber and the Media Neuroscience Lab at UCSB’s Department of Communication. The study was co-funded by ReFrame and the AII and examines the economic performance of 1,200 top films from 2007 to 2018.

Many industry decision-makers base hiring, financing and production choices on the belief that films with female leads/co-leads and those from underrepresented racial/ethnic groups earn less money than films with male or white leads/co-leads at the box office both domestically and internationally. But in accepting this myth as truth, they ignore that these same films also receive lower median production budgets, less median marketing support in dollars, and are distributed in fewer median theaters than films with male or white leads/co-leads. This research asks whether it is the gender and race/ethnicity of leading/co-leading characters and/or the support these films receive that contributes to differences in box office revenue.

The study reveals that when controlling for a set of production, distribution, and exhibition factors, films with female lead or co-lead characters do not perform significantly differently than those with male lead characters domestically and internationally. The story for underrepresented leads and co-leads is more nuanced; domestically and internationally, these films were significantly more likely to out earn films with white leads and co-leads. However, a significant interaction between the percentage of underrepresented cast and lead/co-lead character race/ethnicity offers preliminary evidence that films with a large percentage of underrepresented characters with a white lead/co-lead are less successful than those with a large percentage of underrepresented characters and an underrepresented lead/co-lead character.

The study applied sophisticated statistical controls and path modeling to answer this query, going beyond the limited simple analyses employed in other research. The authors examined more than a dozen production, distribution, and exhibition factors that could influence economic performance domestically and internationally. Along with these indicators, the lead/co-lead character’s gender or race/ethnicity, as well as the percentage of cast that was female or from an underrepresented racial/ethnic group, were included in the models.

The findings showed that domestically, the strongest predictors of economic success were the number of theaters in which a movie was released, story strength (as measured by Metacritic), marketing spend, and production costs. As these factors increased, so did revenue. Films with female leads/co-leads were not associated with greater earnings than those with male leads/co-leads. Additionally, the presence of an underrepresented lead/co-lead was a significant and positive predictor of performance. Simply put, films with leads/co-leads from underrepresented groups earned more revenue when other influences were accounted for in the model. A weaker, but still significant predictor of success was the percentage of female characters in a film.

“This study confirms our previous work indicating that the gender of the lead/co-lead character is not a significant predictor of box office performance,” Smith said. “Rather, it is other factors that are within the control of executives — production costs, promotion, distribution density, and the story itself that play a key role in a movie’s success. Moreover, films with leading/co-leading characters from underrepresented racial/ethnic backgrounds are a significant and notable predictor of economic revenue domestically. This is a finding that cannot be ignored and is consistent with what activists, advocates and artists have been saying for years. Stories with underrepresented leads/co-leads make money. Period.”

ReFrame co-founders Cathy Schulman and Keri Putnam added, “Now that scientists have worked diligently to compare box office performance by adjusting for multiple economic and strategic variables, the industry should be mindful to model films without relying upon unfounded predictors. The research allows for a more nuanced understanding of the factors that relate to economic performance and demonstrate that the beliefs about gender and race that have been used to guide industry practices do not hold up to scrutiny.”

Alison Emilio, Director of ReFrame, added a call to action, "We call on the industry to level the playing field by providing the same production and marketing support to all films (of a similar genre) without regard to gender or race/ethnicity #leveltheplayingfield.”

The findings regarding international revenue were more complex. Here, production costs, international marketing, whether a film was released in China, the number of international territories in which a film was released, and story strength were the strongest positive predictors of success. Having a cast with a higher percentage of females positively influenced revenue, while a greater proportion of characters from underrepresented racial/ethnic groups was associated with a decrease in earnings. However, the gender or race/ethnicity of the lead/co-lead character did not significantly predict earnings.

“These results contradict long-standing beliefs about the economic viability of underrepresented leads overseas,” Smith said. “Instead, the findings demonstrate that bias and exclusionary production and marketing practices are driving decision-making about leading roles rather than data regarding return on investment.”

The finding regarding underrepresented characters and international revenue prompted further scrutiny. The researchers examined the interaction between underrepresented characters and leads/co-leads to understand their unique and combined impact on international financial performance. The results demonstrated that films with a high proportion of underrepresented characters (81% or more) in stories with an underrepresented lead/co-lead were associated with the greatest international revenue. In contrast, films with a high proportion of underrepresented characters (81% or more) but with only white leads/co-leads were associated with the lowest international revenue. The researchers indicate that this finding suggests audiences respond positively to authentic storytelling about diverse communities and are less inclined toward inauthentic films. However, due to the small sample sizes of both of these types of films, further analyses are needed, and this interaction should be interpreted with caution.

“The in-depth analysis of this unique dataset enabled us to better understand the factors that truly explain the economic success of films,” Weber said. “In addition to the impact of inclusion variables we have learned that a good story still plays a major role in a film’s financial success and that factors such as animation as a storytelling style indirectly increase financial performance through stronger stories. Interestingly, our analyses show that star power exerts no significant influence in the U.S. and on international receipts. What exactly are the driving ingredients of a good story is the goal of our ongoing projects. Decision-makers can utilize these findings to understand how best to position their movies for success.”

Few films each year feature women of color in leading roles, which required the researchers to explore the success of movies about underrepresented females in a different way. In a supplementary report, the authors take aim at how films with underrepresented female leads were supported in comparison to those starring white females, underrepresented males, and white males. Using sample of 817 live-action films with solo leading characters, the analysis explored the production, marketing, and release of films starring women of color. The results demonstrate that these movies receive the least amount of support across all groups.

Films starring women of color received the lowest funds in terms of median production costs and median marketing spend (domestic or international). These movies were also distributed in fewer median international territories than films with white female, underrepresented male, or white male leads. Only 3 films with a solo underrepresented female lead were released in China across the 12 years studied, whereas 48 starring white females, 24 featuring underrepresented males, and 174 with white male leads were shown there. The number of movies with underrepresented female leads distributed in Japan, South Korea, and India were equally problematic. Films starring underrepresented female leads were also far less likely to be action films. Of the 167 action movies in the sample, 128 had white male leads, 27 had underrepresented male leads, 10 had white female leads, and only 2 featured women of color as leads.

“Films with women of color as the lead are not only the least likely to be released, but the least likely to be supported by producers and distributors,” Smith said. “We know the factors that increase the likelihood that a film will be financially successful, and movies with women of color in the lead are systematically disadvantaged in these areas. We need studios to change the way they do business when it comes to supporting these films.”

Schulman and Putnam announced, “ReFrame, committed to actionable solutions, will expand its 14-point roadmap for mitigating bias to include recommended changes to finance models based on the findings of this study, so that studios and production companies do not inadvertently underfund movies and shows based on mythological presumptions.”

The authors note that the study, while offering a comprehensive look at the factors that drive the economic performance of films, is limited to the movies included in the sample and does not account for other variables that were not measured or included. For that reason, the authors caution against generalizing the findings to films outside the sample.