Is the Visual Effect Industry Unstable?
When the VFX studio Rhythm and Hues went bankrupt shortly after winning an academy award for its work on Life of Pi (2012) a group of VFX workers protested the ceremony. The protesters wanted to draw attention to the race-to-the bottom economics of the industry and to advocate for unionization. They even hired a plane to pull a banner overhead that read “Box Office + Bankrupt = Visual Effects.” Since then it has become a commonplace that the VFX industry is unsustainable volatile and that studios are in constant threat of bankruptcy. While there is good anecdotal evidence of this volatility, I wondered if there is numerical evidence.
To find out I used a 2017 dataset from IMDB that lists all credits for “special effects companies” in their database. This list of about 85,000 entries includes visual effects, practical effects, animation work, and a range of other jobs they categorize as special effects. In addition to analyzing the whole dataset (henceforth referred to as the SFX dataset) I have also created a sub-list that includes only the digital animation and visual effects work on the list (which I will refer to as the VFX dataset).
Findings
An interactive version of all the following graphs can be accessed by clicking the link below the image.
This Gantt chart draws a line between the first and last recorded credit for each studio in the dataset. This operational lifespan gives us an approximation of when these studios were formed and when they went out of business, as we can assume that a studio could not stay in business more than a year or two without a contract. Clearly there is too much information here to tell us much. But a quick scroll through this list does make a few things apparent. First, that there are a lot of special effects studios, second, that most of them existed between 2000 and the present, and third, that most of these studios only existed for a single year.
1994 to 2008 Saw Unprecedented Growth, Followed by a Downturn for Less Well Established Studios
These graphs show how many studios were working each year and they break them down into three categories of age: The oldest companies that have been around for 15 years or over, the medium age companies that have existed for 2 to 14 years, and companies that have only existed for a year. For SFX studios we can see a period of relatively strong growth between the 1970s and 1990s, with the lower and medium aged studios growing more than the oldest. This trend accelerates massively in the 1990s, peaking around 2008. After 2008 things level off, and the shorter and medium aged studios begin to shrink, with the smallest suffering the worst. The graph of VFX studios starts later, seeing even more dramatic growth in the 1990s, except this time medium aged studios are greater in number. The VFX industry slows in 2008 with the global financial crisis, just like the SFX industry, and just like the SFX industry the low and medium aged studios shrink in numbers, with the number of older studios leveling off. Despite talk of the volatility of the VFX industry in 2012, the VFX industry had a much lower proportion of short-lived studios than SFX. If you only looked at the proportion of one year old studios, you would conclude that that SFX industry was more volatile during years of growth than the VFX industry was during the learner post-2008 years.
Average Studio Age Has Grown Since the 1990s and VFX Has Longer-Lived Studios
This mixture of different aged companies has the unexpected result of pulling up the average age of studios in both industries. Indeed, the median age of VFX studios increased after 2008.
For the SFX dataset the charts shows a general upward trend over all time, and a consistent upward trend staring in the late-1990s. The median is considerably lower, suggesting either a smaller number of long-lived studios are pulling up the average, and/or a large number of shorter-lived studios are dragging down the median.
For the VFX dataset the mean and median age are high in the 80s, dip in the mid-1990s, and gradually climb back up. The ages of studios are generally higher for VFX, and there is a clearer upward trend in median age after the mid-90s. As with the SFX dataset, the average is higher than the median, and both have generally trended upward over time. On their own these data suggest the closure of Rhythm and Hues was an anomaly and more established studios have generally fared well.
The Number of SFX Studios Employed Per Film Flattened After 2005
There had been a trend up to 2005 where movies and television shows would get more and more complicated and production studios would contract work to more and more SFX studios. While this trend continued after 2005 for the top blockbusters, the scale of the average project stagnated.
Studio Closures Have Been Growing Since the 1990s and SFX Has Seen Higher Rates
These two graphs show what proportion of studios that had a credit in a given year had their last credit in that year. In other words, it is a measure of how many studios went out of business in a given year expressed as a percentage rather than a count. There are two lines, one smoothed over three years to show trends. This helps average some of the noise in the early years and it does not contort the data too much because this is only an approximate measure of when studios actually stopped operation.
For the whole SFX dataset, there seems to have been a great deal of instability before 1950. This is in part due to the relative lack of credits before then. 1945 shows 100% of studios going out of business, while 1947 has 0%. After 1950 the rate settles into a consistent 25%. The rate begins to increase in 2000, even as there are more total studios year-over-year, and the trend accelerates after 2008, peaking at over 50% in 2016. In other words, half of the studios in operation in 2016 went out of business.
The VFX dataset does not go back as far because the term visual effects only became a common in the 1980s. The 1980s are the first decade to have a visual effects credit for every year and there are only 50 credits for all the preceding years combined.
There is a period of calm in the 1990s where average closures dip, and the late 1990s and 2000s see a steady rate of about 20%, much better than the SFX dataset. Every year since 2008 has seen greater and greater closure rates, peaking at 47% in 2016. In sum, while the VFX and SFX datasets show the same trends, the rate of closures for SFX was worse.
The Oldest Studios Suffer After 2008, But Not As Bad As Others
This graph shows a count of how many studios ceased production every year and divides them into categories based on their age. After 2008 we can see the most well established studios doing considerably worse than before. Between 2008 and 2015, the oldest SFX studios’ closure rate more than quadrupled, and for VFX it increased even more. This would seem to be the phenomenon Rhythm and Hues was an example of. This is a greater increase in closures than the shorter-lived studios, but that is because those studios had already been losing many times more for at least a decade, even as the industry was growing.
As the graphs of Average Years Active showed, the oldest category of studios still did the best in this period, keeping the average and median age of studios high. Although more larger studios were closing during this period relative to the past, the lower two age categories suffered the worst. It is also important to remember that studios age over time, and that the relatively small population of old studios is constantly being fed by younger studios growing old, and the oldest studios that do survive are constantly getting older, pulling up the average.
This final graph introduces a new dataset. It takes a list of contemporary top visual effects and animation studios from LinkedIn and color codes the studios by size. These data suggest that larger companies last longer. Only one studio with fewer than 500 workers has lasted fewer than 11 years. But there are also many long-lived studios that employ fewer than 500 workers. This is only a glimpse of the contemporary state of affairs, but it supports the other findings here.
Conclusion
There is a great deal of evidence here supporting the idea that the VFX and SFX industries are highly volatile. The fact that both industries have seen between 50% and 25% attrition rates over most of their histories is harrowing in itself. Yet there is also a great deal of complicating information. For one, the SFX industry as a whole was more volatile than VFX. Also, even years of growth can see high rates of closure for smaller studios.
The downturn that followed 2008 also needs to be put in context of the period of wild unsustainable growth that preceded it. This boom and bust is a synecdoche for the global economy, but it was also exacerbated by the film industry’s increasing desire for spectacular and often cost-saving effects. The total number of SFX and VFX jobs has also not recovered, even as the global economy has grown robustly for many years.
The closure of Rhythm and Hues has proven to be a less than ideal example for its time. Founded in 1987, it was already in a rarefied category of well-established studios when it went bankrupt in 2013. Although studios such as these certainly suffered after 2008, they suffered the least, and the average age of studios has actually gone up. A better example would be the scores of precarious new companies competing with the well-established ones. Also, the story about the “vanishing piece of the Pi” tended to conflate the fortunes of studios with the fortunes of workers. While the risk of razor thin profit margins and unfair contracts was certainly being passed on to workers through 6-month employment contracts and extreme working hours, it is important to remember that sometimes companies thrive while workers still suffer. Many of the cost and risk cutting tricks adopted by employers in the wake of 2008 have been slow to disappear, and wages in the US economy have stagnated even as stock market indexes have grown and companies have made record profits.
Data Sources and Methods
The data used in this study came from a dataset IMDB published up until 2017, which I converted to a CSV using a combination of Pandas and Excel before analyzing and visualizing it in Tableau. I removed the data for 2017 and 2018 because they are incomplete and could cause studios to appear to have ceased operation.
There are a few potential weaknesses in the dataset that are worth noting. There are a handful of studios that have credits from before their official start date because they did work on a restoration or a re-release of a film. This should not affect the aggregated data, but anomalies do show up in the Gantt chart that lists all studio life spans. Using credit dates is an imperfect way of measuring a studio’s economic health. Rhythm and Hues, for example, would not show up as ever having gone bankrupt because it was bought and continues to do credited work. Certain studios, like Technicolor, also have a “ship of Theseus” problem. Although the photochemical company has little to do with the French media company that now bears its name, they are technically the same company. There is further a related data entry error problem where studios are either credited with different names (adding an “inc.” for example) or they change their names despite being the same company. Finally, television shows complicate some statistics because they are credited based on the year they started rather than by season. This is visible in the graph on SFX Studios Credited per Movie, where there is a massive spike in the line that represents the top films of the year. This was caused by Game of Thrones.
This article is part of a series on the special effects, visual effects, and animation industries. It is based on research supported by:
Le Fonds de Recherche Québécois sur la Société et la Culture Postdoctoral Fellowship
The University of California, Berkeley’s Department of Film & Media
Title Image by Chris Murray
