THE ADOPTION OF NEW TECHNOLOGIES: UNDERSTANDING HOLLYWOOD’S (SLOW) CONVERSION TO COLOR

THE ADOPTION OF NEW TECHNOLOGIES: UNDERSTANDING HOLLYWOOD’S (SLOW) CONVERSION TO COLOR

Order Description
Your review should do the following:
– Motivate the topic discussed by the article
– Discuss the relevant literature
– Discuss the article’s methods and results
– Provide a substantial, valid criticism of the article
– Provide substantial suggestions for improvement or future research
Reviews must be typed and submitted via www.turnitin.com.

Electronic copy available at: http://ssrn.com/abstract=2176802
1
THE ADOPTION OF NEW TECHNOLOGIES:
UNDERSTANDING HOLLYWOOD’S (SLOW) CONVERSION TO COLOR,
1940-70*
RICARD GIL, JOHNS HOPKINS CAREY BUSINESS SCHOOL AND
RYAN LAMPE, DEPAUL UNIVERSITY
November 14, 2012
Hollywood fully converted to sound within three years. In comparison, Hollywood’s
conversion to color film lasted more than three decades, and included a three-year period
between 1954 and 1957 in which the share of color movies declined from 58 to 31
percent. We investigate this puzzling pattern of adoption using detailed data on 10,385
movies released between 1940 and 1970. The data indicate that cost-saving innovations
and television’s own conversion to color explain large shifts to color in the early 1950s
and mid-1960s. Movie-level data on profits from one of Hollywood’s largest studios,
MGM, as well as box office receipts from 393 theaters across 26 cities between 1945 and
1955 suggest that studios’ resistance to color was driven by consumer tastes. Movies
filmed in color did not attract sufficiently large audiences to cover additional expenses. In
1955, MGM’s color features generated profits that were $3 million lower than
comparable black-and-white movies. In the same year, weekly ticket sales of color
features were 10 percent lower.
*We wish to thank Shaun McRae, Petra Moser, Anthony Niblett, Laura Owen, Matt Pyle, Jonah Rockoff,
Zhu Wang, seminar participants at the Bureau of Economic Analysis, and librarians at The Margaret
Herrick Library of the Academy of Motion Picture Arts and Sciences. Xiao Xiao Peng and Asad Khan
provided excellent research assistance. All errors are our own.
Electronic copy available at: http://ssrn.com/abstract=2176802
2
In a 7 November 1938 memo to a production executive on Gone with the Wind,
producer David O. Selznick commented:
“…MGM, our partners in the enterprise, are still far from being sold on color as an
economic proposition; and, on the contrary, I have spent hours…debating the value of
color by comparison with its extra cost” (Behlmer 2000, pp. 185-6).
Principal photography on Gone with the Wind began on 21 January 1939. Despite
studio apprehension, the Civil War epic was shot using Technicolor’s unique three-strip
color film process. The movie premiered on 15 December 1939 and amassed box office
revenues of $176.7 million based on a production budget of only $4.2 million (Block and
Wilson 2010, p. 220). It also earned art director William Menzies an Honory Academy
Award for outstanding achievement in the use of color.
Despite the film’s success, however, the use of color film remained below 20
percent for the next 10 years and would only pass 90 percent three decades later (Figure
1). Furthermore, the adoption of color film did not follow the traditional S-shape curve
(e.g. Rogers 2003). Abrupt spikes in the early 1950s and mid-1960s punctuate the
adoption of color. Between 1954 and 1957, the proportion of color movies even fell from
58 percent to 31 percent. By comparison, Hollywood fully converted to sound in just
three years.1 In 1927, few films included sound; by the end of 1929 the major studios
ceased production of silent films (Salt 2009, p. 197).
We analyze this puzzling pattern of technology adoption using detailed data on
10,385 feature-length movies released between 1940 and 1970. We complement these
data with detailed financial information on 700 movies released by MGM between 1940
1 See Hanssen (2002) for an account of the changes occurring in the industry as a consequence of the
transition to sound.
3
and 1964, as well as box office receipts from 393 theaters across 26 cities between 1945
and 1955.
The motion picture industry provides an ideal setting for studying technology
adoption since the industry witnessed significant changes in both technological
innovation and competition. In 1950, a low-cost alternative to Technicolor’s expensive
three-strip process, Eastman Color, was introduced, and subsequent improvements
followed at regular intervals between 1952 and 1968. In the aftermath of the Second
World War, average weekly theater attendance began a sharp decline from a peak of 90
million in 1946 to 18 million in 1970. This decline was partly driven by the adoption of
television, which increased from 0 percent to 95 percent between 1948 and 1970.
Interestingly, color was introduced in a setting similar to the one currently faced
by advocates of 3-D technology. As with color, studios are promoting widespread
adoption of 3-D to combat a decline in annual audiences. 3-D is also closely associated
with genres such as action, science fiction, and horror, in the same way color was initially
associated with musicals, comedies, and romance (Mordden 1990, p. 115). Finally, the
benefits of 3-D were forcefully demonstrated in a smash-hit movie, Avatar. That film’s
director, James Cameron, predicted in 2010 that the switch to 3-D will be fast: “it took
only two years for sound movies to become conventional and replace silent films, but 25
years for color to replace black-and-white films after ‘Gone with the Wind’ in 1939…the
3-D transition won’t take 25 years – too many market forces are involved now” (Korea
Times, 13 May 2010).2
2 http://www.koreatimes.co.kr/www/news/biz/2010/05/123_65843.html, accessed 11 October 2012.
Hennig-Thurau and Knapp (2012) show that the adoption of 3-D follows a similar pattern to the adoption
of color since the cost of 3-D is much higher than its benefits for the average movie.
4
Our data highlight the importance of expected costs and benefits in explaining
Hollywood’s slow conversion to color. These data indicate that the introduction of
Eastman Color, which lowered the costs of filming in color, was an important factor
behind the initial surge in color movies between 1950 and 1954. During this time, the
annual output of color films increased from 20 to 58 percent. Network television’s
switch to color broadcasting in 1965, which created a lucrative market for color films,
also prompted a large increase in color adoption. Between 1965 and 1967, color
production increased from 59 to 88 percent.
Movie-level regressions that estimate the likelihood of a film being shot in color
confirm the importance of cost-saving innovations and the emergence of color television.
The introduction of Eastman Color and subsequent improvements in its speed explain 25
percent of the increase in the use of color. The adoption of color television increased
color production by 40 percent. Finally, estimates indicate that studios were motivated to
shoot in color to attract audiences back to the cinema after theater attendance began to
plummet in the post-war period; a drop in weekly attendance of 10 million increased the
likelihood of color by 4 percent. These estimates are robust to including controls for
studio, genre, the use of special widescreen formats, and running time.
Despite the introduction of a low-cost color film process suitable for
commercial production in 1950, Hollywood did not fully convert to color until television
networks switched to color broadcasting in 1965. The puzzle then remains: What
explains this slow conversion to color, and what explains the decline in color production
from a peak of 58 percent in 1954 to 31 percent in 1957?
5
Contemporary accounts suggest that color movies did not attract sufficiently
larger audiences in the 1950s to cover the added expense of shooting in color. Financial
data on MGM releases between 1940 and 1964 confirm that color movies were not more
profitable than black-and-white movies. In 1955, when 87 percent of MGM’s movies
were filmed in color, the median movie made a loss of $454,000. Regressions that
control for movies characteristics indicate that in 1955 MGM’s profits on color movies
were $3 million dollars lower on average than comparable black-and-white movies. In
the next two years, MGM cut the share of color releases to 47 percent.
Theater-level data on the weekly box office performance of 4,300 movies
between 1945 and 1955 indicate that lower returns to shooting in color in the mid-1950s
were not confined to MGM. Regressions of weekly box office receipts that control for
movie and theater characteristics indicate that revenues for color movies were only 1.3
percent larger than revenues for black-and-white movies. Estimates of annual
coefficients indicate that the premium from shooting in color declined during the first
half of the 1950s. In 1950, weekly revenues were 6.7 percent higher for color movies.
However, by 1955 revenues were 9.6 percent lower.
Our results contribute to an established literature that emphasizes the importance
of expected costs and benefits in determining the rate of diffusion of a new technology.
In his study on the adoption of hybrid corn, Griliches (1957) stresses the importance of
expected profits in explaining variation in diffusion rates across Midwestern states after
1940. Similarly David (1975) argues that nineteenth century farms adopted the
mechanical reaper only when the cost of labor exceeded the cost of the reaper. More
6
recent work on diffusion finds a significant role for network effects as a driver of
expected benefits (e.g. Goolsbee and Klenow 2002).
Our findings also contribute to a growing literature on the U.S. film industry in
the post-war period. Sedgwick (2002) finds that studios responded to declining
audiences and changing patterns in the riskiness of movies after 1946 by focusing on
“hit” movies with higher production values. A large body of research also focuses on
the aftermath of the 1948 Paramount decision (e.g. Conant 1994; DeVany and Eckert
1991; Gil 2010, 2012). Prior to 1948 the majors owned theaters in addition to production
and distribution means. The Supreme Court of 1948 judgment mandated that these
studios be separated from their theatrical branches and sell 25 to 50 percent of their
theater holdings. It also required that majors and minors end the use of block booking
that “compelled an exhibitor to buy sight unseen a whole line of mediocre and poor films
along with the few features in which he is interested” (Katz 2008, p. 148).
The remainder of this paper is structured as follows. In section 2 we describe
technological innovation in the U.S. movie industry between 1940 and 1970. In section 3
we describe the movie- and theater-level data used in this paper. Section 4 analyzes the
determinants of color adoption. Section 5 discusses the decline in color use between
1954 and 1957. A final section concludes.
BACKGROUND, 1940-1970
During the 1930s and 1940s, the Technicolor Motion Picture Corporation
maintained an effective monopoly over high-quality color cinematography. Their threecolor
process used special cameras that captured red, green, and blue light on three
7
separate black-and-white negatives. Each of these negatives was then combined on a
single three-color release print (Cook 1996, p. 257). The three-strip process gained
prestige in 1939 with the release of such films as The Wizard of Oz and Gone with the
Wind.
Shooting in Technicolor was expensive because the company controlled every
stage of the color process. Rather than sell its cameras, it would rent them as part of a
package that included specialized cameramen and lighting, a color supervisor (usually
Natalie Kalmus, wife of Technicolor’s inventor, Hebert Kalmus) and laboratory
processing and printing (Basten 2005 p. 55). In 1935, shooting in Technicolor increased
production costs by 30 percent based on average production costs of $300,000 per
feature; in 1949 Technicolor increased production costs by 10 percent based on average
production costs of $1,000,000 (Kindem 1979, p. 34).
On 21 December 1951, the National Film Board of Canada released Royal
Journey, a documentary that became one of the first professional films to use Eastman
Color, a new, low-cost color film introduced by the Eastman Kodak Company.3 The
advent of Eastman Color, announced the previous year, ended Technicolor’s twenty-year
monopoly of the market for color cinematography based on its costly three-strip process.
Eastman Color combined the three strips of color sensitive film required for the
production of color images onto a single “monopack” roll.4 The simplicity of the single
3 Other titles released in 1951 that used Eastman Color film stock include Hurricane Island, The Sword of
Monte Cristo, and Sunny Side of the Street.
4 According to a Government complaint in 1947, Eastman Kodak and Technicolor entered into an
agreement on June 25, 1934 that prevented Eastman Kodak from commercial processing of monopack film;
prior to 1934 both Technicolor and Kodak were independently attempting to develop a single-strip color
film technology. Consent Judgments on 1948 and 1950 ended their agreement and required the two
companies to license relevant patents (United States v. Technicolor, Inc., CCH 1950-51 Trade Cases ¶2,586
(S.D. Calif. Civil No. 7507-M; Complaint, 1947; Consent Judgments, 1948 and 1950)). See Lampe and
Moser (2012).
8
roll meant that color movies could be shot on ordinary black-and-white cameras and
processed by conventional processing facilities (Neale 1985, p. 142). In 1952, Eastman
Kodak introduced an improved monopack technology, type 5248, with higher emulsion
speed and better grain structure that allowed the technology to be adopted by major
Hollywood studios.5 In 1954, Technicolor responded by cutting rental charges for its
three-strip cameras by 50 percent to $120 per week and $20 per day. It also reduced the
price of color release prints from 6.4 to 4.3 cents per foot (Daily Variety, 3 February
1954, p. 3). Despite these price cuts, Foxfire (1955), filmed in 1954 by Universal,
became the last American-made feature to be photographed with a Technicolor threestrip
camera (Basten 2005, p. 129).
Eastman Kodak responded to price reductions by Technicolor and lower-price
black-and-white film stock with continued improvements to Eastman Color throughout
the 1950s and 1960s. Improved stocks were made available in 1959 (Negative Film
#5250), 1962 (Negative Film #5251), and 1968 (Negative Film #5254) (Ryan 1978). In
addition to improving the sharpness of the image, these new stocks increased the speed of
Eastman Color. Film speed, measured by its exposure index (EI) with artificial Tungsten
lighting, increased from 16 to 100 between 1950 and 1968 (Figure A1). Film speed was
important because faster film stocks reduced production costs since they required less
lighting for proper exposure (Lev 2006, p. 109). The new Eastman Color in 1959, for
example, “photographed more clearly scenes with dim or drab illumination” (Limbacher
1968, p. 69).
5 Records of the Eastman Kodak Company, accessed on May 25, 2012 at:
http://motion.kodak.com/motion/About/Chronology_Of_Film/1940-1959/index.htm.
9
Diffusion of television and declining audiences
Shooting in color offered studios a way to differentiate their product from
predominantly black-and-white television programming, and attract audiences back to the
cinema after attendance began to plummet in the post-war period. In 1948, less than 1
percent of households had a television set at home; by 1955, the percentage had risen to
65 percent (Figure 1). This rapid increase in the availability of cheaper substitutes for
movies is often cited as a major factor behind the massive post-war decline in theater
attendance (e.g. Cook 1996, p. 460). Increased migration to the suburbs, away from
theater locations in downtown neighborhoods, is another reason (e.g. Gomery 1992, 83-
88).6 Between 1946 and 1950, average weekly theater attendance declined from a peak
of 90 million to 60 million. By 1970, average weekly attendance was just 17.7 million
(Block and Wilson 2010).
The technological advantage of color was diminished by the widespread adoption
of color television in the late 1960s. Between 1964 and 1970, the percentage of homes
with color television rose from 3 percent to 39 percent. In 1965 NBC announced that its
fall prime time schedule would be broadcast almost entirely in color; ABC and CBS
quickly embraced color to limit NBC’s competitive advantage. Though the switch to
color broadcasting eliminated one of Hollywood’s key technological advantages, it also
created a new market for color features and allowed amortization of any film project
(Mordden 1990, p. 116).7 In 1967, Variety estimated that a “good” feature could earn
6 Sedgwick (2002, p. 679) notes that while the number of Americans living in city centers “grew by 32
percent [between 1950 and 1970], the suburban population grew by 161 percent.”
7 Mordden (1990, p. 116) notes that “black-and-white movies generated smaller sales.”
10
$1.75 million through a television release (Variety, 4 January 1967, p. 98).8 By the
following year, license fees were consistently above $800,000 per movie (Block and
Wilson 2010, p. 431).9 This fee compares to average domestic theatrical revenues of
$1,590,391 for MGM’s releases in the 1960s.
Widescreen technologies
In addition to color, studios sought to differentiate their product from television
programming by stretching the size of the screen image (Gomery and Parafort-Overduin
2011, p. 240). The first widescreen process, Cinerama, debuted in 1952. It required
three special cameras and the resulting three strips were projected in sync onto large
deeply curved screens (Block and Wilson 2010, p. 332).10
In 1953, 20th Century-Fox introduced CinemaScope, a cheaper and less
cumbersome “anamorphic” widescreen process. Unlike Cinerama, CinemaScope
squeezed a widescreen image onto a single 35mm filmstrip during photography and then,
using anamorphic lenses, returned the image to widescreen specifications during
projection. In the late 1950s, Robert E. Gottschalk introduced an improved anamorphic
process, the Panavision lens, which produced a distortion free image and, unlike
CinemaScope, “allowed a change in the anamorphic power of the lens with a simple turn
8 In 1966, the color feature The Bridge Over the River Kwai was licensed to ABC for $2 million and drew
60 million viewers (Block and Wilson 2010, p. 431).
9 In 1967, Broadcasting reported that approximately 80% of prime time ABC movies were in color
(Kindem 1979, p. 35).
10 The resulting image was three times as wide and nearly twice as tall as the standard 35mm image (Cook
1996, p. 464). The process was first demonstrated in 1952 with the movie This Is Cinerama. Although
popular, Cinerama was expensive for filmmakers and exhibitors; only theaters in large cities could afford to
install the large three-panel screens necessary for projection (Cook 1996, p. 465).
11
of a knob” (Gomery and Parafort-Overduin 2011, p. 243).11 Panavision eventually
replaced CinemaScope for 35mm widescreen cinematography (Cook 1996, p. 471).
These new anamorphic formats were made possible by the introduction of Eastman
Color, which further encouraged its adoption (Haines 2003).12
Studios that did not develop or license a widescreen process generated a wide
image by using an ordinary spherical lens with a removable plate that cropped the top and
bottom of the frame.13 Unlike true widescreen processes such as CinemaScope and
Panavision, however, images that were cropped and enlarged in this way were grainier
and fuzzier since only a portion of the image was used for projection (Haines 2003, p.
86).14 Nevertheless, cropped (or “flat”) widescreen was popular in the 1950s and
diminished the impact of CinemaScope (Carr and Hayes 1988, p. 63). Since 1954,
almost all movies have been projected with a widescreen image, generated either by
cropping or by using a special widescreen process (Salt 2009, p. 275). Our data indicate
that the use of special widescreen systems peaked at 37.6 percent in 1959 before
declining to 18.2 percent in 1970 (Figure 2). The decline in 1959 coincides with the
11 Panavision provides a better-illuminated image than CinemaScope because the system uses a smaller
picture area for its squeeze and reduces its aspect ratio to 2.25:1. Furthermore, Panavision eliminated
CinemaScope’s wide-angle distortion almost entirely through the use of highly curved spherical lenses
(Cook 1996, p. 471).
12 The three-strip Technicolor camera could not be immediately adapted for anamorphic widescreen
processes (Daily Variety, 11 November 1953, p. 11). In the meantime, Technicolor continued to operate its
processing facilities and supplied release prints for movies filmed using Eastman Color stock (Neale 1985,
p. 144). Technicolor eventually released its own widescreen process in 1957, Super Technirama, based on
modifications to the company’s obsolete three-strip cameras. However, the process was expensive and not
widely adopted (Carr and Hayes 1988, pp. 152-153).
13 One of the first features to use this approach was Universal’s It Came from Outer Space in 1953 which
was presented with a 1.85×1 aspect ratio compared to the standard 1.33×1 ratio. MGM used a 1.66 × 1
masking for a 1954 widescreen reissue of Gone with the Wind though this resulted in the actors’ heads and
feet being chopped off (Block and Wilson 2010, p. 221; Haines 2003, p. 86).
14 Studios also presented movies in 3-D to compete with television. However, use of the technology was
short-lived. Haines (2003, p. 74) lists only 25 3-D features between 1953 and 1954. The first 3-D feature
was the 1953 horror film House of Wax.
12
introduction of Eastman Color Type 5250, which improved the clarity of cropped films
and diminished the need for specialized widescreen processes.
Shooting in color was more expensive
Color, like widescreen, had the potential to lure audiences back to cinemas.
However, shooting in color increased both production and distribution costs. Production
costs were higher because of technical difficulties related to lighting. For exteriors,
variation in natural light during different hours of the day created problems during editing
since shots taken at different time would not “match.” This limited the number of hours
of the day for consistent photography (Katz 2008, p. 291). For interiors, shooting in
color required very high light levels compared to black-and-white film (Lev 2006, p.
109).15 Eastman Color type 5250, released in 1959, enabled “filmmakers to use color
film where never before possible and in as little as half the light previously required”
(Weekly Variety, 13 May 1959, p. 17).
Distributing color movies was also more costly because processing color film
stock was more than twice as expensive as black-and-white stock. For example, in 1949
the cost of color release prints was seven cents per foot compared to three cents per foot
for black-and-white prints (Kindem 1979, p. 34). In 1956, the cost differential was still
four cents (Variety, 11 April 1956, p. 4). For a 90-minute 35mm movie, this difference
amounted to $324 per release print.16 Since the major and minor studios usually had print
runs of 300, distribution costs were therefore $97,200 larger per movie on average
15 For example, the lighting cost on Gone with the Wind was still more than $100,000, even though
Technicolor introduced a new film that cut lighting requirements by 50 percent in 1939 (Basten 2005, p.
78).
16 A 90-minute 35mm film requires 8,100 feet of film. Based on conversions from
http://www.intervalometers.com/resource/index.php, accessed 11 September 2012.
13
(Haines 2003, p. 33). Compared to an average budget of $1.1 million for movies released
in the 1950s, this suggests that shooting in color raised distribution costs alone by 8.8
percent (Block and Wilson 2010, p. 320).
DATA ON AMERICAN FILM PRODUCTION
The main data for this study consists of detailed information on 10,385 featurelength
movies released in the United States between 1940 and 1970. We complement
these data with information on the profits of 700 MGM releases between 1940 and 1964,
as well as data on the weekly box office performance of 4,300 movies released in 393
theaters in 26 U.S. cities between 1945 and 1955.17
Feature films, 1940-1970
Detailed information on 10,385 feature-length movies released in the United
States between 1940 and 1970 comes from the American Film Institute (AFI) movie
catalog (www.afi.com). The Catalog provides online information on nearly 60,000
American feature-length films produced in America or financed by American production
companies from 1893 to 2012. The information included in the data set includes details
on cast, crew, plot summaries, subjects, genres, technical information, and historical
notes for each film.
17 A total of 12,959 films are listed on the AFI website between 1940 and 1970. We exclude 2,308 non-
U.S. movies, 123 movies without information on running time, 5 movies that are not feature-length (less
than 40 minutes in length), and 138 movies for which we cannot determine if they were filmed in color.
14
For all of 10,385 movies, we are able to determine if the film was shot in blackand-
white or color.18 Between 1940 and 1970, 3,511 movies (34 percent) were filmed in
color, including 88 black-and-white films that include color sequences, e.g. The Wizard
of Oz (Table 1). Data on 687 movies filmed using Technicolor’s 3-strip Technicolor
camera come from Haines (2003), data on other color processes come from Limbacher
(1968), and data on the speed of Eastman Color comes from Ryan (1978). The timing of
movies is based on release years.19
TV saturation and box office revenue, 1940-1970
We obtained data from the Television Factbook on yearly TV saturation
(percentage of homes in the US with a TV). Even though TV saturation was 0 until 1945
it increased to 95.2 percent in 1970. In 1970, the proportion of homes with color
television was 39.3 percent. Data on weekly box office attendance comes from Brown
(1995). Between 1940 and 1970, average weekly attendance declined from 80.0 million
to 17.7 million; attendance peaked in 1946 at 90 million.
Controls for other movie characteristics
To compete with television and lure audiences back to the cinema in the post-war
period, studios attempted to produce “hit” films with enhanced production values
(Sedgwick 2002, p. 678). Historical epics such as The Robe (1953) and Cleopatra (1963)
exemplify the characteristics of such movies: color, wider images, and longer running
18 Information on color choice for 9,398 movies was obtained from the AFI website. We obtained
information on color choice for another 987 movies from IMDB.
19 Information on production years is only available for 7,254 of 10,385 movies (69.9 percent). We run
regressions that use these production years as a robustness check. The average difference between
production and release years is 0.6 years.
15
times. To control for correlation between these characteristics, we collected data on the
use of a special widescreen format from Carr and Hayes (1988) as well as data on
running times from the AFI and IMDB.20 1,296 of 5,227 (24.8 percent) movies released
between 1952 and 1970 used a special widescreen system. Of these movies, 515 (39.7
percent) were filmed in CinemaScope (Figure 2). The average movie in our data has a
running length of 83 minutes. Color movies have an average running time of 94 minutes
compared to an average running time of 77 minutes for black-and-white movies (Table
1).
To finance increased production costs in the post-war period, studios also began
to share production costs. Although only 1,249 of 10,385 movies (12.0 percent) list more
than one producer; the maximum number of producers is 9. The five major studios are
listed as a producer on 2,696 (26.0 percent) of movies while the two “minor” studios,
Universal and Columbia, are listed as a producer on 1,447 (13.9 percent) of movies.21
Both majors and minors produced larger shares of black-and-white movies than color
movies, which is consistent with the major studios distributing more movies after the
1948 Paramount decision (Sedgwick 2002) (Table 1).
The decision to shoot in color is likely to have been influenced by the type of
movie being produced. Kindem (1979, p. 36) notes that “color film’s association with
fantasy and black-and-white’s ‘illusion of reality’ affected the types of feature films
which were made in color and black-and-white, until television’s conversion to color
broadcasting and color news film in 1965-1966 gave color added realism in comparison
20 We incorporate changes from Sherlock (2004) that provides corrections to Carr and Hayes (1988).
21 United Artists is often grouped with Columbia and Universal. Since United Artists was a distributor and
did not produce any movies, we do not include it as a minor.
16
to black-and-white.” The most popular genres in our data are drama (38 percent),
comedy (18 percent), westerns (17 percent), and romance (4 percent).22
Profits for MGM movies, 1940-1964
We also obtained data on production costs, domestic and foreign revenues, and
profits for 700 movies released by MGM between 1940 and 1964. The data were
obtained from the ledger of an administrative vice-president at MGM, Edgar J. Mannix.23
MGM was one of the five majors, known for its expensive, high-quality productions.24
The data indicate that average production costs, excluding distribution costs, were more
than $2 million for color movies and $1 million for black-and-white movies (Table 1).25
However, these costs were not dissimilar to the costs of rival studios. In 1949, average
production costs at Paramount and Fox were $1.4 million and $1.8 million (Lev 2003, p.
14).
Theater-level box office, 1945-1955
We further complement these data with weekly box office information from
Variety magazine for 393 theaters across 26 different cities between 3 January 1945 and
22 We use the following set of genres: adventure, comedy, drama, documentary, film noir, horror, musical,
mystery, romance, science-fiction, western, and “other.”
23 The Eddie Mannix Ledger is available at The Margaret Herrick Library of the Academy of Motion
Picture Arts and Sciences in Los Angeles. See Glancy (1992) for a thorough discussion of the ledger’s
origin and accuracy. The profit figures match closely financial statistics obtained from other sources at
MGM. Glancy (1992) suggests the revenue figures represent theater rentals – the share of box-office
revenues that are received by the producer or distributor – rather than gross receipts.
24 Unlike the four other major studios, MGM did not completely divest itself of its theater chain (Loews)
until 1959 (Lev 2003, p. 14).
25 Profits are calculated by subtracting production costs and additional distribution and publicity costs (not
separately reported in the ledger) from total revenues (Glancy 1992). Production costs are available for 698
of 700 movies (99.7 percent) with data on profits.
17
28 December 1955.26 As the number of theaters and cities reported by Variety changes
week by week, the data set is an unbalanced panel data set (Table 2). The end of each
row offers the total number of theaters for each city that Variety offered information.
Though there is turnover of theaters, there is also a lot of repetition in reporting which
allows us to exploit high frequency panel data. We are able to match box office receipts
to film characteristics for 4,300 movies released between 1945 and 1955. Data on the
years that television was available in each of the 26 cities in our dataset comes from
Gentzkow (2006).
DECOMPOSING THE DETERMINANTS OF COLOR ADOPTION
Large increases in color production after 1950 and 1965
Summary statistics for color production reveal large increases in the use of color
between 1950 and 1954, and after TV networks switched to color broadcasting in 1965
(Figure 1). Between 1950 and 1954, the annual proportion of color movies increased
from 20.0 to 57.6 percent. Between 1954 and 1957, color production declined to 30.5
percent before slowly rising again. After networks converted to color broadcasting in
1965, color production surged again. Between 1965 and 1970, the proportion of color
movies spiked from 59.3 to 93.9 percent.
Summary statistics that control for studio size indicate that the above pattern of
adoption was true for the major and minor studios, as well as smaller and independent
producers. However, these data indicate that the major studios were more likely to shoot
in color and were early adopters of Eastman Color (Figure 3). Between 1940 and 1970,
the average annual proportion of color movies was 51.3 percent for the majors compared
26 This data set is fully described in Gil (2012).
18
to 46.8 percent for Universal and Columbia, and 35.6 for smaller studios and
independents. In 1955, 87.4 percent of movies produced by the five major studios were
filmed in color compared to 54.5 percent for the Universal and Columbia, and 39.5
percent for smaller studios and independents. In that same year, 87.0 percent of MGM’s
features were shot in color.27
Data on the type of color process used in production suggest that the introduction
of Eastman Color in the early 1950s was an important factor in the rapid increase in color
production after 1950 (Figure 4). The use of Eastman film surged after the introduction
of its improved film stock in 1952: from 2.2 percent of all movies in 1952 to 43.2 percent
of movies in 1956. In comparison, use of Technicolor’s 3-strip camera collapsed from
27.1 percent of all releases in 1953 to 1.2 percent in 1956.28 Similarly, lower-quality
color processes that competed with Technicolor in the 1940s, such as CineColor, Ansco
Color, and TruColor, were phased out in favor of Eastman Color.
Least squares regressions using MGM’s financial data indicate that Eastman
Color reduced the cost of color features. Shooting in color with Technicolor’s 3-strip
camera increased production costs by 23.2 percent while shooting on Eastman Color
increased costs by 15.3 percent (significant at 5 percent, Table 3, Column 4).29
27 The data also suggest that the Paramount case had a limited impact on the adoption of color film. The
major studios did not change their pattern of color use immediately after 1948 even though they were most
affected by the resolution of the Paramount case. The data, however, suggest that Columbia and Universal
increased their use of color after the Paramount decision granted them access to larger theaters. Universal
and Columbia dramatically increased the number of color movies prior to 1951. In 1948 and 1949, the two
studios produced a combined output of 12 color features; in 1950 and 1951 the studios produced 37 color
features. Examples include Universal’s The Glenn Miller Story and Columbia’s From Here to Eternity
(Thomson and Bordwell 1994 p. 375).
28 Studios that used Eastman Color stock but processed the film themselves adopted trade names such as
Metrocolor (MGM), Warner Color (Warner Bros.) and Color by Deluxe (20th Century Fox) (Gomery and
Parafort-Overduin 2011, p. 244).
29 An F-test statistic 0.79 and p-value of 0.4 does not reject the null hypothesis the coefficients on Eastman
Color and 3-Strip Technicolor are equal.
19
Probit estimates for the choice of color
To separate the impact of Eastman Color and color broadcasting from the
diffusion of television and declining box office attendance, we examine the adoption of
color at the level of individual movies. A movie-level analysis is appropriate in this
setting since studios were not locked into using color and could opt to shoot in color
depending on market conditions and the characteristics of the movie. Probit regressions
predict the choice of color, controlling for other movie characteristics such as the use of a
widescreen process, running time, the studio’s annual output, and genre:
(1) Pr (colori) = a + ß1 majori + ß2 minori + ß3 exposure indexi +ß4 b&w tv saturationi
+ ß5 color tv saturationi + ß6 weekly attendancei + ß7 widescreeni
+ ß8 durationi + ß9 moviesi + ß10 coproductioni + gi + e
i
where the dependent variable colori equals 1 if movie i released in year t was filmed in
color. Binary variables majori and minori equal 1 if movie i was produced by one of the
five majors (Paramount, RKO, Warner Bros., 20th Century Fox, and MGM) or one of the
two minors (Columbia and Universal) studio. The coefficient on exposure indexi
estimates the impact of Eastman Color as well as improvements in the film stock. It
measures the speed of Eastman Color in the year movie i was released. Coefficients on
b&w tv saturationi, color tv saturationi, and weekly attendancei estimate the impact of
television and declining box office attendance on the choice of color.
Additional regressors control for movie and studio characteristics. The variable
widescreeni controls for the use of a special widescreen process such as CinemaScope,
and durationi controls for movie i’s running time. Finally, the variables moviesi and
coproductioni control for the number of other movies produced by the same studio in the
20
year movie i was released, and whether movie i had multiple producers. gi are genre
fixed-effects. Standard errors are clustered on the producer.30
Estimates in Table 4 confirm that the adoption of color was strongly correlated with
changes in expected benefits and costs. Increases in the speed of Eastman Color, which
reduced the amount of lighting needed for proper exposure, are strongly correlated with
color production. A 10-point rise in the exposure index increased the likelihood of
shooting in color by 3 percent (significant at 1 percent, Table 4, Column 3). This implies
that technological improvements in Eastman Color between 1950 and 1968 that increased
the exposure index from 16 to 100 increased the use of color by 25.2 percent.
Filmmakers were also more likely to use color as households adopted color
television, which created lucrative markets for color features. A 1 percent increase in the
number of households with a color television increased the likelihood of color by 1.0
percent (significant at 1 percent, Table 4, Column 3). The estimates also indicate that the
five major and two minor studios were 12.4 and 14.6 percent more likely to shoot in color
than smaller studios (significant at 5 percent, Table 4, Column 3).31
The choice of color was strongly correlated with box office attendance. A drop in
average weekly attendance of 10 million is associated with a 6 percent increase in the use
of color (significant at 1 percent, Table 4, Column 3). The diffusion of black-and-white
television, however, is not positively correlated with the production of color films. This
finding is consistent with steep declines in box office attendance and increased use of
30 Specifically, we cluster on the producer listed first in the AFI data. Only 1,249 of 10,385 movies (12.0
percent) list more than one producer. The variable coproduction controls for multiple producers.
31 To model persistance of technology adoption, we estimated specifications that include a dummy variable
equal to one if movie i’s producer had already produced a color movie. The variable is strongly correlated
with major and minor (which comprise 40 percent of the observations in our data set), and insignificant
when we include producer fixed effects. The main results are unchanged.
21
color after 1946, before the widespread adoption of television between 1950 and 1955
(e.g. Gomery 1992, p. 84).
Regression results also indicate that the use of color was positively correlated
with the use of a special widescreen process and running length, but negatively correlated
with the number of productions.32 Films shot in a widescreen process such as
CinemaScope were 27.6 percent more likely to be filmed in color (significant at 1
percent, Table 4, Column 3). Every 10 minutes in running time is associated with a 6
percent increase in the likelihood of color (significant at 1 percent, Table 4, Column 3).
There is also evidence of studios cutting back on the number of movies to enhance
production values; the likelihood of color increases by 4.0 percent when a studio
produced 10 fewer movies in a given year (significant at 1 percent, Table 4, Column 3).
Regressions results in Table 5 indicate that certain genres were more closely
associated with color than others. Coefficients on genre dummies indicate that comedy,
drama, film noir and mystery genres were more likely to be filmed in black-and-white
while adventure, documentary, musical and western genres were more likely to be shot in
color during this 30-year period. Dramas, mysteries and film noir were more than 18
percent less likely to be filmed in color than movies compared to the excluded category
“other.”
Results are robust to including producer fixed effects (Table 4, Columns 4, 5 and
7), and using data on production years for 7,254 movies (Table 4, Columns 8-9).33 An
32 Results for seemingly-unrelated regressions that allow for unobserved correlation in the choice of color,
use of a widescreen process, and running length are presented in Table A1.
33 We include fixed effects for the the producer listed first in the AFI data. Since 1,250 of 10,385 movies
(12.0 percent) list more than one producer, coefficients on major and minor are still estimated. 1,508
producers were only listed as producer on one movie; these are dropped in regressions incorporating
producer fixed effects.
22
additional robustness check includes dummy variables for years after the introduction of
Eastman Color (post 1950i) and NBC’s conversion to color broadcasting (post 1965i).
The coefficient on post 1950i is associated with a 13.2 percent increase in color movies
(significant at 1 percent, Table 4, Column 5). Although it is clear that Eastman Color had
a causal effect, we cannot pin down the actual magnitude of this effect. The coefficient
on post 1965i is large but not significant (Table 4, Column 5). This is not surprising since
NBC’s announcement was correlated with an increase in color TV adoption; the
coefficient on color tv saturationi drops slightly to 1.20 (significant at 10 percent, Table
4, Column 5).
Though the above estimates can explain large and long-term upward swings
towards color, they cannot, however, explain the rapid decline in color that occurred
during the mid-1950s. Predicted shares of color between 1940 and 1970 indicate that
estimates of equation (1) capture the rapid increases in color adoption after 1950 and
1965, but do not explain the rapid drop in color between 1954 and 1957 (Figure 5).
Regression coefficients predict a mild decline in color of 4.0 percent between 1955 and
1957 compared to the sharp decline of 27.1 percent that actually occurred between 1954
and 1957. In the next section we discuss the factors underlying this retreat from color
features that cannot be captured by our regressions in Tables 4 and 5.
WHY DID COLOR PRODUCTIONS DECLINE AFTER 1954?
Though Eastman Color substantially reduced the costs of shooting color movies,
Hollywood did not fully embrace color movies until households began to adopt color
television and networks began all-color broadcasting in 1965. What explains this
23
resistance to color and, in particular, what explains the decline in color production from a
peak of 58 percent in 1954 to 31 percent in 1957?
Contemporary accounts indicate the decline in color was caused by economic
concerns:
“Return to b&w has been prompted by a general desire for economy plus the
realization that color does not always provide the extra value that can be translated to
additional dollars at the boxoffice. Some of the recent b&w pictures have been the best
grossers…the film companies, however, are taking a more realistic approach, figuring
that the extra cost for color is a gamble in today’s uncertain market” (Weekly Variety,
25 July 1956, pp. 125).34
We explore this explanation using movie-level data on profits from MGM
between 1940 and 1964, as well as box office receipts from 393 theaters across 26 cities
between 1945 and 1955.
MGM profits, 1940-1964
MGM’s financial data confirm that color movies did not always attract
sufficiently large audiences to cover the added expense of shooting in color. The average
color movie received domestic revenues of $2.3 million and profits of $79,750 compared
to domestic revenues of $1.2 million and profits of $164,740 for black-and-white movies
(Table 1).35 The data also indicate that in 1955, when 87 percent of MGM’s movies were
filmed in color, the average movie lost $513,000. In that year, MGM produced three
34 One common explanation for the decline in color features is that studios cut back on color productions
because they began selling their movie catalogues to the networks in 1955, and since television was still
black-and-white, there was little reason to add the expense of color (e.g. Kindem 1979, p. 34; Cook 1996,
p. 462). However, this explanation is unlikely since studios had a tacit agreement to limit television films
to features released no later than 1948 (Brooks and Marsh 1981, p. 510). Network schedules also indicate
little or no use of Hollywood feature films between 1956 and 1961 (Hilmes 1990, p. 163). NBC only
began premiering recent films in prime time in 1961 with How to Marry a Millionaire (Gomery 1992, p.
250).
35 “Of the top 10 grossing films for each year from 1940 to 1949, 5.5 were in black-and-white and 4.5 were
in Technicolor” (Kindem 1979, p. 34).
24
black-and-white movies (Trial, I’ll Cry Tomorrow, and Blackboard Jungle) that each
generated profits between $518,000 and $4,392,000.
Least squares regressions estimate the difference in profits from color movies,
controlling for movie characteristics and year fixed-effects:
(2) profiti = a + ß1iI(1940-1964)i *colori +ß2 widescreeni +ß3 color*widescreeni
+ß5 durationi + ß6 moviesi + ß7 coproductioni + gi + yi + e
i
where the indicator variable I(1940-1964)i is equal to 1 for any release year between
1940 and 1964.
Year-specific estimates indicate that the returns to color movies were declining
during the early 1950s. In 1951, shooting color movies received profits that were
$953,059 larger than black-and-white movies, controlling for characteristics such as the
use of widescreen, running time and genre (significant at 5 percent, Figure 6). However,
in 1955, color movies received profits that were $2.9 million lower on average compared
to similar black-and-white movies (significant at 5 percent, Figure 6). Box office returns
therefore provided little evidence to studios that shooting in color had a positive effect on
profit. In the next two years, the share of MGM’s releases that were filmed in color
dropped from 87 percent to 47 percent.36
Theater-level receipts 1945-1955
36 Financial data during Hollywood’s conversion to sound indicate that silent pictures brought in
significantly lower profits after sound was introduced. In the 1928-29 season, 15 of 53 MGM releases had
some form of sound. Average profits for these movies were $381,200 compared to average profits for that
seaon’s silent movies of $125,454. In the 1929-30 season, all but 2 of MGM’s 44 releases contained sound
(Glancy 1992, p. 133).
25
Data on weekly box office receipts indicate that modest or negative returns to
shooting in color were not confined to MGM. Between 1945 and 1955, the average color
feature received $15,250 per week and theater, while the average black-and-white feature
received $13,400 (Table 1). Least squares regressions estimate additional revenues from
shooting in color, controlling for movie and theater characteristics:
(3) log (revenueijt) = a + ß1 colori +ß2 color*TV in city?ijt +ß3 TV in city?ijt
+ß4 widescreeni +ß5 durationi +ß6 week numberijt +ß7 reissueijt
+ß8 number of runijt +gi + yt + fi + tj + eijt
where the dependent variable is the log of revenue of movie i in theater j during week t.
The binary variable colori distinguishes between color and black-and-white movies. The
binary variable TV in city? equals 1 if TV was available in the same city as theater j in
week t. The main specification also controls for movie characteristics such as its
duration, whether the movie was filmed in widescreen, the number of weeks playing in
theater j, and whether it was a reissue or a second- or higher-run movie. 37 gi, yt, fi, and tj,
are genre, year, studio, and theater fixed effects. Standard errors are clustered at the
theater level.
Estimates indicate that color movies did not attract significantly larger audiences
than black-and-white movies. Weekly box office revenues were only 1.3 percent larger
for color movies after controlling for genre, year, studio, and theater fixed effects
(significant at 10 percent, Table 6, Column 9). The introduction of (black-and-white)
television reduced revenues by 13.7 percent for black-and-white movies and 11.2 percent
37 Only four films in our data between 1952 and 1955 were filmed in widescreen but not in color (The Rose
Tattoo, The Return of Jack Slade, Bengazi, and The Desperate Hours). We therefore exclude an interaction
between color and widescreen.
26
for color movies (significant at 5 percent, Table 6, Column 9). Movies that were filmed
in widescreen did not receive significantly larger revenues after controlling for city fixed
effects suggesting that certain theaters were more likely to show widescreen movies
(Table 6, Column 9).38
Estimates of annual coefficients indicate that additional revenues from shooting in
color declined during the first half of the 1950s. In 1950, weekly revenues were 6.7
percent higher for color movies compared to black-and-white movies with similar
characteristics (significant at 5 percent, Figure 7). However, by 1955 revenues were 9.6
percent lower (significant at 5 percent, Figure 7). Alternative specifications without
studio and theater fixed effects indicate that revenues for color movies were significantly
lower between 1952 and 1954 (significant at 10 percent, Table 7, Columns 3-4).
Black-and-white CinemaScope features after 1955
The insufficient returns from shooting in color are reflected in Twentieth-Century
Fox’s decision in 1956 to allow CinemaScope productions to be filmed in black-andwhite.
Fox initially required that CinemaScope productions be filmed in color (Weekly
Variety, 23 May 1956, p. 7). This pledge was thought to be necessary to convince theater
owners to retrofit their cinemas for CinemaScope features (Lev 2003, p. 116). The
enhancement included lenses costing $2,875 each, and screens for $5 to $5.50 per square
foot (Daily Variety, 8 July 1953, p. 3).39 However, beginning with MGM’s The Power
38 Estimates also indicate that longer movies collected more revenues while older movies received lower
revenues. Every additional minute is associated with 0.4 percent larger revenues (significant at 1 percent,
Table 6, Column 6). Reissues received 17.0 percent less revenue while each week in released is associated
with 1.8 percent less revenue (significant at 1 percent, Table 6, Column 6). Movies in their second and
third runs received 20.6 and 30.9 percent less revenue (significant at 1 percent, Table 6, Column 6).
39 In 1956 there were 17,058 theaters in the U.S. and Canada that were equipped for CinemaScope (Weekly
Variety, 25 May 1956, p. 7). For a screen 22 feet high and 50 feet wide, such as the Miracle Mirror screen
27
and the Prize in 1956, Fox allowed monochrome CinemaScope productions. The studio
itself saved $250,000 by releasing Teenage Rebel in black-and-white instead of color
(Weekly Variety, 27 June, 1956). As a result of this change, between the 1955 and 1957,
the share of color CinemaScope productions fell from 25.1 to 13.8 percent while the
share of black-and-white CinemaScope productions rose from 0 to 5.7 percent (Figure 2).
Color and genre
The swing back towards black-and-white also reflected a realization that certain
genres were better suited to black-and-white. In 1956, Elia Kazan, director of A Streetcar
Named Desire and On the Waterfront, commented that color had artistic drawbacks and
that he had “yet to find a drama filmed in color where the tinting didn’t soften the impact
of the story” (Daily Variety, 5 December 1956, p. 8). Kazan’s first color film, East of
Eden in 1955, was followed up in 1956 and 1957 by two black-and-white features, Baby
Doll and A Face in the Crowd. Consistent with a rise in color during the 1960s, between
1960 and 1969, three of his four movies were shot in color. Kindem (1979, pp. 35-36)
argues that color became more closely associated with “realism” in the 1960s when news
film was broadcast in color; before network TV’s conversion to color, information on
contemporary events was primarily found in black-and-white newspapers and magazines.
A return to filming dramas in black-and-white was an important factor in the
decline in color films between 1954 and 1957, as well as slow growth afterwards,
because the share of dramas increased significantly after 1954. Between 1954 and 1959,
the share of dramas across all movies increased from 35.4 to 50.4 percent. During the
at RKO’s 86th Street Theater in New York, the additional cost would be between $5,500 and $6,050
(Weekly Variety, 13 May 1953, p. 5).
28
same period, the share of dramas that were shot in color declined from 43.9 to 29.5
percent (Figure 8).
The popularity of dramas in the second half of the 1950s partly reflected an
increase in independent productions, made possible by the breakdown of the studio
system. Between 1955 and 1965 the share of films produced by the major studios
was just 15.7 percent compared 31.1 percent between 1945 and 1955. Since
independent producers did not share the larger studios’ sense of its audience or its
close ties to the Motion Picture Production Code that sought to homogenize film
content, independent productions reflected more variety and experimentation. The
more diverse output of independent studios during the “postclassical Hollywood”
period of the late 1950s is reflected in black-and-white films such as Vertigo (color)
and Touch of Evil (black-and-white) (Lev 2003, p. 217).40 Between 1955 and 1960,
the share of musicals and westerns, genres associated with both color and the major
studios, declined from 21.8 to 10.6 percent.
SUMMARY AND DISCUSSION
This paper highlights the importance of expected costs and benefits from adopting
new technologies. Hollywood’s reluctance to embrace color appears to have been driven
by consumer tastes. During the 1950s color movies did not attract sufficiently larger
audiences to justify increased production and distribution costs. Data from 393 theaters
across 26 cities between 1945 and 1955 indicate revenues were only slightly larger for
40 A shift away from musicals in the 1950s was principally driven by MGM’s reduction in musical
production. Between 1950 and 1955, MGM was produced 30 musicals; between 1956 and 1960, the studio
produced only 8. Financial records indicate that musicals were not MGM’s most profitable features in the
early 1950s (Lev 2003, p. 15).
29
color movies. In 1955, weekly revenues were 10 percent smaller, controlling for movie
characteristics. In the same year, profits on MGM’s color movies were $3 million lower.
Studios responded by slashing the share of color movies to 31 percent by 1957.
Improvements in Eastman Color, that lowered the production costs of shooting in color,
helped to close the gap before networks TV’s own conversion to color programming
virtually eliminated black-and-white cinematography.
Our results have particular relevance to Hollywood’s current efforts to combat
flagging box office revenues. Between 2002 and 2011, theater admissions declined
steadily from 1.6 to 1.3 billion (MPAA 2011). Part of Hollywood’s strategy to lure
audiences back to the cinema involves the use of improved 3-D technologies. In the
same way Gone with the Wind highlighted the benefits of shooting in color, the 2009
science-fiction movie Avatar highlighted the benefits of filming in 3-D. However,
despite the latter’s record-breaking grosses, audience reactions to 3-D have been mixed.
In particular audiences do not value 3-D enough to justify higher ticket prices (e.g. New
York Times, 29 May 2011).41 As a result, only 45 of 610 movies (7.3 percent) released in
2011 were in 3-D (MPAA 2011). Our results suggest that improvements in 3-D (such as
technologies that do not require special glasses) and the emergence of new markets for
content such as broadcast and on-demand TV may prove crucial to widespread use of the
technology.42
41 “3-D Starts to Fizzle, and Hollywood Frets,” available at:
http://www.nytimes.com/2011/05/30/business/media/30panda.html, accessed 17 September 2012.
42 In the beginning of 2012, worldwide adoption of 3D television was less than 2 percent (Wired, 5 January
2012, avaliable at: http://www.wired.com/gadgetlab/2012/01/state-of-3-d-technology/, accessed 17
September 2012). See Hennig-Thurau and Knapp (2012) for recent evidence detailing a benefit-cost
analysis of 3-D movie production.
30
This paper has described the many factors playing a role in a studio’s movie-level
decision to shoot in color. We leave for future research the study of technology adoption
at the studio level as well as the impact of head to head competition in the use of new
technologies. Understanding how competition shapes technology adoption is important
for the design of future competition policy guidelines as creative industries continue to
increase their contribution to the economy.
REFERENCES
Basten, Fred. Glorious Technicolor: The Moves’ Magic Rainbow. Camarillo, CA:
Technicolor, 2005.
Baum-Snow, Nathaniel. “Did Highways Cause Suburbanization?,” Quarterly Journal of
Economics, 2007, Vol. 122, No. 4, pp. 775-805.
Behlmer, Rudy. Memo from David O. Selznick. New York, NY: Modern Library, 1972,
2000.
Block, Alex, and Lucy Wilson. George Lucas’s Blockbusting. New York, NY:
HarperCollins, 2010.
Brooks, Tim, and Earle Marsh. The Compete Directory to Prime Time Network TV
Shows: 1946-Present. New York, NY: Ballantine, 1979, 1981.
Brown, Gene. Movie Time. New York, NY: Macmillan, 1995.
Campbell, Jeffrey, and Thomas Hubbard. “The Economics of ‘Radiator Springs:’
Industry Dynamics, Sunk Costs, and Spatial Demand Shifts”, March 2010, mimeograph.
Carr, Robert, and R. M. Hayes. Wide Screen Movies: A History and Filmography of
Wide Gauge Filmmaking. Jefferson, NC: McFarland & Company, 1988.
Conant, Michael. Antitrust in the Motion Picture Industry. Berkeley, CA: University of
California Press, 1960.
Cook, David. A History of Narrative Film. New York, NY: W. W. Norton Company,
1981, 1996.
31
David, Paul. “The Mechanization of Reaping in the Ante-bellum Midwest,” in P. A.
David, Technical Choice, Innovation, and Economic Growth. Cambridge, Cambridge
University Press, 1975, pp. 195-232.
DeVany, Albert and R. Eckert. “Motion Picture Antitrust: the Paramount Cases
Revisited.” Research in Law and Economics, 1991, Vol. 14, No. 1, pp. 51-112.
Finler, Joel. The Hollywood Story. Chippenham, UK: Wallflower, 1988, 2003.
Gentzkow, Matthew. “Television and Voter Turnout.” Quarterly Journal of Economics,
2006, Vol. 121, No. 3, pp. 931-972.
Gil, Ricard. “An Empirical Investigation of the Paramount Antitrust Case.” Applied
Economics, Vol. 42, No. 2, 2010, pp. 171–183.
Gil, Ricard. “Does Vertical Integration Decrease Prices? Evidence from the Paramount
Antitrust Case of 1948.” Working paper, Johns Hopkins University, 2012.
Glancy, H. Mark. “MGM Film Grosses, 1924-1948: The Eddie Mannix Ledger.”
Historical Journal of Film, Radio and Television, Vol. 12, No. 2, pp. 127-144.
Gomery, Douglas. Shared Pleasures: A History of Movie Presentation in the United
States. Madison, WI: University of Wisconsin Press, 1992.
Gomery, Douglas, and Clara Pafort-Overduin. Movie History: A Survey. New York,
NY: Routledge, 1991, 2011.
Goolsbee, Austan, and Klenow, Peter. “Evidence on Learning and Network Externalities
in the Diffusion of Home Computers.” Journal of Law and Economics, Vol. 45, No. 2,
2002, pp. 317-343.
Griliches, Zvi. “Hybrid Corn: An Exploration in the Economics of Technological
Change.” Econometrica, Vol. 25, No. 4, 1957, pp. 501-522.
Haines, Richard. Technicolor Movies: The History of Dye Transfer Printing. Jefferson,
NC: McFarland & Company, 1993, 2003.
Hanssen, F. Andrew. “Revenue Sharing in Movie Exhibition and the Arrival of Sound.”
Economic Inquiry, Vol. 40, No. 3, 2002, pp. 380-402.
Hennig-Thurau, Thorsten, and Ann-Kristin Knapp. “The Economic Implications of
Adding a Third Dimension.” Working Paper, University of Muenster, 2012.
Hilmes, Michele. Hollywood and Broadcasting. Chicago, IL: University of Illinois
Press, 1990.
32
Katz, Ephraim. The Film Encyclopedia 6e: The Complete Guide to Film and the Film
Industry. New York, NY: Harper Collins, 2008.
Kindem, Gorham. “Hollywood’s Conversion to Color: The Technological, Economics
and Aesthetic Factors.” Journal of the University Film Association, Vol. 31, No. 2,
Economic and Industry History of the American Film (Spring 1979), pp. 29-36.
Lev, Peter. The Fifties: Transforming the Screen, 1950-1959. Berkeley, CA: University
of California Press, 2006.
Limbacher, James. Four Aspects of the Film. 1968.
Motion Picture Association of America. Theatrical Market Statistics, 2011. Available
at: http://www.mpaa.org/resources/5bec4ac9-a95e-443b-987b-bff6fb5455a9.pdf.
Mordden, Ethan. Medium Cool: The Movies of the 1960s. New York, NY: Alfred A.
Knope, 1990.
Neale, Steve. Cinema and Technology: Image, Sound, Colour. Bloomington, IN: Indian
University Press, 1985.
Rogers, Everett. Diffusion of Innovations. New York, NY: Free Press, 1962, 2003.
Rosenberg, Nathan. “Learning by Using,” in N. Rosenberg, Inside the Black Box.
Cambridge, Cambridge University Press, 1982, pp. 120-140.
Ryan, R. T. A History of Motion Picture Color Technology. New York, NY: Focal
Press, 1978.
Salt, Barry. Film Style and Technology: History and Analysis. London, UK: Starwood,
1983, 2009.
Sedgwick, John. “Product Differentiation at the Movies: Hollywood, 1946-65. Journal
of Economic History, Vol. 63, No. 3, 2002, pp. 676-705.
Sherlock, Daniel. Wide Screen Movies Corrections. 1994, 2004.
Television Digest. Television Factbook.
Thompson, Kristin, and David Bordwell. Film History: An Introduction. McGraw-Hill,
1994.
33
TABLE 1: SUMMARY DATA
Panel A: Movie-level Observations, 1940-70
Color Black-and-White All
Major 0.25 0.27 0.26
Minor 0.10 0.16 0.14
Eastman Exposure Index 41.53 11.87 21.90
B&W TV Saturation 0.61 0.25 0.37
Color TV Saturation 0.09 0.01 0.03
Weekly Attendance (Millions) 42.57 69.30 60.26
Widescreen 0.29 0.04 0.12
Duration (Minutes) 94.00 76.91 82.69
Movies 8.06 19.89 15.89
Coproduction 0.18 0.09 0.12
N 3,511 6,874 10,385
Notes: Data from the AFI (www.afi.com), IMDB (www.imdb.com), and Carr and Hayes (1988). An
observation is a movie. Major equals 1 if the film’s producer is Paramount, RKO, Warner Bros., Fox, or
MGM. Minor equals 1 if the film’s producer is Columbia or Universal. Eastman exposure index measures
the speed of Eastman Color film (a measure of technological progress). Widescreen equals 1 if the movie
was shot with a special widescreen process (e.g. CinemaScope or Panavision). Duration is the movie’s
length (in minutes). Movies is the number of other movies produced by the same studio in the same year.
Coproduction equals 1 if the movie has multiple producers.
Panel B: Theater-level Observations, 1945-55
Color Black-and-White All
Weekly Box Office Revenue $15,250 $13,400 $14,017
Duration (Minutes) 96.85 88 90.95
Widescreen 0.17 0.01 0.06
Week Number 2.23 2.12 2.16
Reissue 0.02 0.04 0.04
Run number 1.39 1.52 1.48
N 39,282 78,456 117,738
Notes: Data on weekly box office revenue from Variety magazine (3 January 1945-28 December 1955).
An observation is a movie-theater-week triple. Week number is the number of weeks in release. Reissue
equals 1 if the movie is a reissue. Run number indicates the number of times the movie has been released.
Panel C: MGM Financial Data (in thousands), 1940-64
Color Black-and-White All
Production Cost 2,083.35 1,050.52 1,416.10
Domestic Revenues 2,296.22 1,205.43 1,591.42
Foreign Revenues 1,682.15 649.07 1,017.08
Profits 79.75 164.74 134.51
N 249 451 700
Notes: Data from Eddie Mannix Ledger. Domestic and foreign revenues represent rentals – the share of
box-office revenues that are received by the producer or distributor – rather than gross receipts. Profits are
calculated by subtracting production costs and additional distribution costs from total revenues (Glancy
1992).
34
TABLE 2: CITY AND YEAR STRUCTURE OF THEATER-LEVEL DATA SET, 1945-55
City 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 Theaters
Baltimore, MD 8 9 8 8 8 7 8 9 9 10 10 12
Birmingham, AL 0 0 0 0 0 0 5 0 0 0 0 5
Boston, MA 12 15 17 15 13 12 12 12 14 13 13 22
Buffalo, NY 5 6 6 6 6 5 5 5 7 6 6 7
Chicago, IL 11 11 13 14 15 14 12 12 16 16 16 21
Cincinnati, OH 9 8 8 7 8 7 8 6 6 6 5 13
Cleveland, OH 7 7 9 8 7 8 10 8 8 8 7 11
Columbus, OH 4 0 0 0 0 0 0 0 0 0 0 4
Denver, CO 9 9 11 10 14 11 11 12 12 11 14 23
Detroit, MI 8 8 9 9 7 7 8 6 9 9 9 12
Indianapolis, IN 5 5 5 5 5 5 5 5 5 5 5 5
Kansas City, MO 7 7 11 9 11 10 12 14 13 13 12 18
Lincoln, NE 0 5 0 0 0 0 0 0 0 0 0 5
Los Angeles, CA 27 31 38 36 39 33 33 28 33 34 32 55
Louisville, KY 7 7 8 8 8 4 5 5 4 4 4 8
Minneapolis, MN 9 10 12 12 11 10 9 10 9 8 8 12
New York, NY 18 21 26 25 21 21 24 26 27 28 24 42
Omaha, NE 5 5 5 5 6 5 5 5 4 4 7 8
Philadelphia, PA 11 11 13 13 14 13 12 13 13 12 13 17
Pittsburgh, PA 8 7 8 7 7 8 6 6 6 6 7 12
Portland, OR 8 8 9 9 9 6 7 9 9 8 6 11
Providence, RI 8 7 7 7 7 7 6 5 5 5 4 8
Saint Louis, MO 6 6 10 8 9 6 8 9 8 9 10 13
San Francisco, CA 8 10 15 16 13 14 11 11 13 14 13 19
Seattle, WA 11 10 11 11 10 9 9 10 9 9 8 12
Washington, DC 6 8 10 11 11 11 9 10 10 10 11 18
Theaters/Year 217 231 269 259 259 233 240 236 249 248 244 393
Cities/Year 24 24 23 23 23 23 24 23 23 23 23 26
Notes: Data on weekly box office revenue from Variety magazine (3 January 1945-28 December 1955). Table
indicates the number of theaters per city and year for which the data set used in this paper.
35
TABLE 3: OLS – DEPENDENT VARIABLE IS LOG PRODUCTION COST, 1940-64
(1) (2) (3) (4)
Eastman Color 0.662*** 0.015 0.153* 0.151**
(0.072) (0.084) (0.079) (0.076)
3-Strip Technicolor 0.809*** 0.529*** 0.318*** 0.231***
(0.073) (0.056) (0.052) (0.054)
Other Color Process 0.333* 0.275** 0.086 0.040
(0.182) (0.137) (0.131) (0.125)
Widescreen 0.305*** 0.333*** 0.246***
(0.083) (0.079) (0.076)
Duration (Minutes) 0.026*** 0.026*** 0.025***
(0.001) (0.001) (0.001)
Movies -0.005*** 0.020*** 0.020***
(0.002) (0.005) (0.005)
Coproduction -0.362*** -0.137* -0.235***
(0.083) (0.080) (0.078)
Constant 6.735*** 4.503*** 2.499*** 3.599***
(0.033) (0.136) (0.453) (0.443)
Year Fixed Effects – – Y Y
Genre Fixed Effects – – – Y
Observations 687 687 687 687
R-squared 0.202 0.559 0.690 0.732
Notes: Data on MGM production costs from The Eddie Mannix Ledger. Data on movie
characteristics from AFI (www.afi.com), IMDB (www.imdb.com), and Carr and Hayes (1988).
Widescreen equals 1 if the movie was shot with a special widescreen process (e.g. CinemaScope
or Panavision). Duration is the movie’s length (in minutes). Movies is the number of other
movies produced by MGM in the same year. Coproduction equals 1 if the movie has multiple
producers. *** significant at 1 percent, ** significant at 5 percent, * significant at 10 percent.
36
TABLE 4: PROBIT – DEPENDENT VARIABLE IS COLOR, 1940-70
Release Years (1)-(7) Production Years (8)-(9)
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Major 0.189*** 0.111** 0.124*** -0.039 -0.045 0.107*** -0.070 0.125*** -0.040
(0.041) (0.048) (0.043) (0.090) (0.088) (0.035) (0.092) (0.042) (0.091)
Minor 0.089 0.142* 0.146* -0.160 -0.176 0.101 -0.186 0.156** -0.155
(0.063) (0.082) (0.079) (0.147) (0.140) (0.070) (0.141) (0.077) (0.150)
Exposure Index 0.002** 0.003*** 0.003*** 0.004*** 0.003** 0.004*** 0.006***
(0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.002)
B&W TV Saturation 0.266*** -0.155** -0.038 -0.260*** -0.267*** -0.173** -0.451***
(0.070) (0.066) (0.065) (0.086) (0.084) (0.072) (0.110)
Color TV Saturation 1.289*** 0.755*** 0.965*** 1.693*** 1.200* 0.651*** 1.392***
(0.223) (0.220) (0.215) (0.481) (0.729) (0.221) (0.498)
Weekly Attendance -0.005*** -0.007*** -0.006*** -0.007*** -0.004*** -0.007*** -0.008***
(0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001)
Widescreen 0.302*** 0.276*** 0.299*** 0.306*** 0.317*** 0.371*** 0.288*** 0.335***
(0.036) (0.038) (0.049) (0.049) (0.037) (0.051) (0.037) (0.049)
Duration (Minutes) 0.005*** 0.006*** 0.006*** 0.006*** 0.006*** 0.007*** 0.006*** 0.007***
(0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001)
Movies -0.004*** -0.004*** -0.006*** -0.006*** -0.002* -0.000 -0.004*** -0.007***
(0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001)
Coproduction -0.015 -0.019 0.015 0.014 -0.016 0.033 -0.018 0.025
(0.024) (0.022) (0.041) (0.040) (0.021) (0.038) (0.022) (0.041)
Post 1950 0.132***
(0.043)
Post 1965 0.249
(0.162)
Genre Fixed Effects – – Y Y Y Y Y Y Y
Producer Fixed Effects – – – Y Y – Y – Y
Year Fixed Effects – – – – – Y Y – –
Observations 10,283 10,283 10,283 7,341 7,341 10,283 7,341 10,150 7,230
Notes: Data from AFI (www.afi.com), IMDB (www.imdb.com), and Carr and Hayes (1988). Marginal effects are reported. An observation is a movie. The dependent
variable equals 1 if the movie was shot in color. Major equals 1 if the film’s producer is Paramount, RKO, Warner Bros., Fox, or MGM. Minor equals 1 if the film’s
producer is Columbia or Universal. Exposure index measures the speed of Eastman Color film. Widescreen equals 1 if the movie was shot with a special widescreen
process (e.g. CinemaScope or Panavision). Columns 7 and 8 use production years for 7,254 movies (69.9 percent) instead of release years. Robust standard errors
clustered on producer in parentheses. *** significant at 1 percent, ** significant at 5 percent, * significant at 10 percent.
TABLE 5: PROBIT – COEFFICIENTS ON GENRE DUMMIES, 1940-70
Adventure 0.265*** Musical 0.308***
(0.087) (0.094)
Comedy -0.117*** Mystery -0.200***
(0.039) (0.021)
Documentary 0.299** Romance 0.054
(0.129) (0.054)
Drama -0.186*** Science-Fiction -0.109
(0.038) (0.068)
Film Noir -0.223*** Western 0.193***
(0.012) (0.068)
Horror -0.087
(0.057)
Notes: Coefficients on genre dummies from Table 4, Column 7. Regression include year and producer
fixed effects as well as controls for other characteristics. Marginal effects are reported. An observation
is a movie. The dependent variable color equals 1 if the movie was shot in color. Excluded genre
category is “other.” Robust standard errors clustered on producer in parentheses. *** significant at 1
percent, ** significant at 5 percent, * significant at 10 percent.
38
TABLE 6: OLS – DEPENDENT VARIABLE IS LOG OF WEEKLY BOX OFFICE REVENUE, 1945-55
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Color 0.104*** 0.070*** 0.006 0.006 -0.017 0.025 0.022 0.017** 0.013*
(0.022) (0.025) (0.022) (0.021) (0.022) (0.015) (0.015) (0.008) (0.007)
Color*TV in City? 0.009 0.008 -0.036 -0.029 -0.026 -0.001 -0.007 0.025**
(0.030) (0.029) (0.028) (0.028) (0.021) (0.021) (0.010) (0.010)
TV in City? 0.130*** 0.127*** 0.162*** 0.170*** -0.055* -0.091** -0.137*** -0.137***
(0.046) (0.045) (0.045) (0.046) (0.028) (0.038) (0.028) (0.028)
Widescreen 0.079*** 0.101*** 0.109*** 0.207*** 0.293*** 0.055*** 0.025
(0.028) (0.032) (0.031) (0.027) (0.033) (0.017) (0.018)
Duration (minutes) 0.006*** 0.005*** 0.005*** 0.005*** 0.005*** 0.003*** 0.004***
(0.001) (0.001) (0.001) (0.001) (0.001) (0.000) (0.000)
Week number -0.012*** -0.011*** -0.020*** -0.020*** -0.012*** -0.018***
(0.003) (0.003) (0.005) (0.005) (0.002) (0.003)
Reissue -0.397*** -0.391*** -0.448*** -0.453*** -0.161*** -0.170***
(0.038) (0.038) (0.035) (0.035) (0.017) (0.017)
Number of run -0.183*** -0.178*** -0.082*** -0.080*** -0.114*** -0.103***
(0.042) (0.041) (0.025) (0.026) (0.014) (0.014)
Constant 9.226*** 9.153*** 8.638*** 9.024*** 9.087*** 9.141*** 9.114*** 9.303*** 9.237***
(0.037) (0.049) (0.097) (0.090) (0.171) (0.114) (0.115) (0.071) (0.068)
Genre Fixed Effects – – – – Yes Yes Yes Yes Yes
City Fixed Effects – – – – – Yes Yes – –
Year Fixed Effects – – – – – – Yes Yes Yes
Theater Fixed Effects – – – – – – – Yes Yes
Studio Fixed Effects – – – – – – – – Yes
Observations 117,738 117,738 117,710 117,710 117,710 117,710 117,710 117,710 117,699
R-squared 0.00 0.01 0.04 0.07 0.07 0.27 0.28 0.66 0.68
Notes: Data on weekly box office revenue from Variety magazine (3 January 1945-28 December 1955). Data on movie characteristics from the AFI
(www.afi.com), IMDB (www.imdb.com), and Carr and Hayes (1988). An observation is box office of movie i in theater j in week t. Color equals 1 if movie
was shot in color. TV in city? equals 1 if TV was available in theater j’s city. Widescreen equals 1 if the movie uses a special widescreen. Duration is the
movie’s length in minutes. Week number is the number of weeks in release. Reissue equals 1 if the movie is a reissue. Number of run indicates the number of
times the movie has been released. Robust standard errors in parentheses, clustered at the theater level. *** significant at 1 percent, ** significant at 5 percent,
* significant at 10 percent.
39
TABLE 7: OLS – DEPENDENT VARIABLE IS LOG OF WEEKLY BOX OFFICE REVENUE, 1945-55
(1) (2) (3) (4) (5) (6)
Color 0.128*** 0.118*** 0.069*** 0.048* 0.018 -0.007
(0.044) (0.030) (0.027) (0.026) (0.024) (0.014)
Color*Year 1946 0.0113 0.003 0.010 0.015 0.035 0.042**
! (0.0373) (0.031) (0.030) (0.030) (0.030) (0.018)
Color*Year 1947 -0.0153 -0.014 -0.057* -0.045 -0.039 0.027
! (0.043) (0.033) (0.033) (0.032) (0.032) (0.020)
Color*Year 1948 -0.044 -0.032 -0.047 -0.045 -0.035 -0.003
! (0.045) (0.032) (0.032) (0.031) (0.031) (0.018)
Color*Year 1949 -0.071 -0.036 -0.035 -0.022 -0.009 0.014
! (0.052) (0.040) (0.038) (0.037) (0.033) (0.020)
Color*Year 1950 -0.034 -0.015 -0.042 -0.037 -0.027 0.067***
! (0.064) (0.047) (0.045) (0.043) (0.040) (0.024)
Color*Year 1951 -0.063 -0.048 -0.046 -0.040 0.002 0.046**
! (0.055) (0.044) (0.043) (0.041) (0.039) (0.023)
Color*Year 1952 -0.079 -0.056 -0.089* -0.078* -0.030 0.025
! (0.060) (0.047) (0.046) (0.044) (0.041) (0.022)
Color*Year 1953 -0.108* -0.096* -0.112** -0.090** -0.007 0.051**
(0.061) (0.049) (0.045) (0.044) (0.042) (0.025)
Color*Year 1954 -0.067 -0.031 -0.219*** -0.202*** -0.009 0.015
! (0.066) (0.052) (0.051) (0.049) (0.045) (0.027)
Color*Year 1955 -0.045 -0.030 -0.391*** -0.379*** -0.235*** -0.096***
(0.062) (0.049) (0.061) (0.059) (0.052) (0.031)
Color*TV in City? 0.079** 0.075** 0.086*** 0.085*** 0.074*** 0.021
(0.039) (0.030) (0.029) (0.029) (0.026) (0.013)
TV in City? 0.383*** -0.091** -0.113*** -0.111*** -0.101*** -0.136***
(0.076) (0.042) (0.040) (0.039) (0.038) (0.028)
Widescreen 0.005*** 0.005*** 0.004*** 0.004***
(0.001) (0.001) (0.001) (0.000)
Duration (minutes) 0.390*** 0.388*** 0.202*** 0.062***
(0.040) (0.040) (0.030) (0.019)
Week number -0.022*** -0.022*** -0.027*** -0.018***
(0.005) (0.005) (0.005) (0.003)
Reissue -0.458*** -0.454*** -0.444*** -0.170***
(0.035) (0.035) (0.033) (0.017)
Number of run -0.084*** -0.081*** -0.095*** -0.102***
(0.027) (0.026) (0.024) (0.014)
Constant 8.975*** 9.274*** 8.987*** 9.098*** 8.832*** 9.240***
(0.057) (0.044) (0.075) (0.116) (0.153) (0.068)
Year FE Yes Yes Yes Yes Yes Yes
City FE – Yes Yes Yes Yes –
Genre FE – – – Yes Yes Yes
Studio FE – – – – Yes Yes
Theater FE – – – – – Yes
Observations 117,738 117,738 117,710 117,710 117,699 117,699
R-squared 0.04 0.23 0.28 0.28 0.34 0.68
Notes: Data on weekly box office revenue from Variety magazine (3 January 1945-28 December 1955).
Data on movie characteristics from the AFI (www.afi.com), IMDB (www.imdb.com), and Carr and Hayes
(1988). An observation is box office of movie i in theater j in week t. Color equals 1 if movie was shot in
color. TV in city? equals 1 if TV was available in theater j’s city. Robust standard errors in parentheses,
clustered at the theater level. *** significant at 1 percent, ** significant at 5 percent, * significant at 10
percent.
40
FIGURE 1 – SHARE OF COLOR MOVIES AND HOUSEHOLDS WITH TV, 1940-1970
Notes: Data on color share from AFI (www.afi.com) and IMDB (www.imdb.com). Data on TV
penetration from Television Factbook. Eastman Kodak introduced its monopack film in 1950. In 1965,
NBC announced that its fall prime time schedule would be broadcast almost entirely in color.
FIGURE 2 – SHARES OF CINEMASCOPE AND ALL WIDESCREEN FILMS, 1950-1970
Notes: Data from the AFI (www.afi.com) and Carr and Hayes (1988). Prior to 1956, Twentieth-Century
Fox required all CinemaScope films to be film in color (Weekly Variety, 27 June 1956, p. 10). The first
CinemaScope film to be shot in black-and-white was The Power and the Prize in 1956. All Widescreen
includes alternative widescreen processes such as Cinerama (1952-1958), VistaVision (1954-1963), and
Panavision (1957-present).
41
FIGURE 3 – SHARE OF COLOR MOVIES BY STUDIO TYPE, 1940-1970
Notes: Data from the AFI (www.afi.com) and IMDB (www.imdb.com). The major studio producers are
Paramount, RKO, Warner Bros., Fox, and MGM. The minor studio producers are Columbia and Universal.
Eastman Kodak introduced its monopack film, Eastman Color, in 1950. In 1965, NBC announced that its
fall prime time schedule would be broadcast almost entirely in color.
FIGURE 4 – SHARE OF MOVIES SHOT IN TECHNICOLOR AND EASTMAN COLOR, 1940-1970
Notes: Data from Haines (2003), Limbacher (1968), and the AFI (www.afi.com). Technicolor films are
films shot with Technicolor’s three-strip camera. Eastman Kodak introduced its monopack film (type
5247) in 1950, and an improved version suitable for studio work (type 5248) in 1952. In 1965, NBC
announced that its fall prime time schedule would be broadcast almost entirely in color.
42
FIGURE 5 – PREDICTED AND ACTUAL SHARE OF COLOR MOVIES, 1940-1970
Notes: Data on movies and color from AFI (www.afi.com) and IMDB (www.imdb.com). Predicted values
based on regression results from equation (1) (Table 4, Column 3).
FIGURE 6 –YEAR-SPECIFIC ESTIMATES OF ADDITIONAL
MGM PROFITS FROM SHOOTING IN COLOR, 1943-1961
Notes: Data from The Eddie Mannix Ledger, AFI (www.afi.com) and IMDB (www.imdb.com). An
observation is an MGM movie released between 1940 and 1964. Annual estimates of ß1i in the OLS
regression profiti = a + ß1iI(1940-1964)i *colori +ß2’Xi + e
i where the indicator variable I(1940-1964)i is
equal to 1 for any release year between 1940 and 1964, and colori equals 1 if movie i was filmed in color.
Xi includes controls the use of a widescreen process, an interaction between the use of widescreen and
color, running time, the number other released by MGM in the same year, whether the movie is a
coproduction, and year and genre fixed effects.
43
FIGURE 7 – YEAR-SPECIFIC ESTIMATES OF ADDITIONAL THEATER
REVENUES FROM COLOR MOVIES, 1946-1955
Notes: Data on weekly box office revenue from Variety magazine (3 January 1945-28 December 1955).
An observation is a movie-theater-week triple. Annual estimates of ß1t in the OLS regression log
(revenueijt) = a + ß1t colori + +ß2’Xijt + eijt where revenueijt is the box office revenue of movie i at theater
j in week t, and colori equals 1 if movie i was filmed in color. Xijt includes controls for the use of a
widescreen process, running time, the number of weeks in release, whether the movie is a reissue, and the
number of runs. Regressions also include year, theater, producer, and genre fixed effects.
FIGURE 8 – SHARE OF ALL DRAMAS AND DRAMAS SHOT IN COLOR, 1940-1970
Notes: Data from the AFI (www.afi.com) and IMDB (www.imdb.com).
44
APPENDIX
TABLE A1: SEEMINGLY UNRELATED REGRESSION, 1952-70
DEPENDENT VARIABLES ARE COLOR, WIDESCREEN AND DURATION
Model I (1-3) Model II (4-6)
(1) (2) (3) (4) (5) (6)
Color Widescreen Duration Color Widescreen Duration
Major 0.241*** 0.326*** 17.476*** 0.236*** 0.293*** 17.457***
(0.017) (0.015) (0.741) (0.019) (0.018) (0.848)
Minor 0.187*** 0.115*** 6.671*** 0.186*** 0.071*** 9.363***
(0.023) (0.021) (1.009) (0.026) (0.024) (1.156)
Exposure Index 0.001* -0.004*** 0.121*** 0.001 -0.004*** 0.105***
(0.001) (0.001) (0.029) (0.001) (0.001) (0.029)
B&W TV Saturation 0.215*** 0.934*** 26.337*** 0.294*** 1.007*** 23.155***
(0.057) (0.051) (2.473) (0.057) (0.053) (2.525)
Color TV Saturation 1.160*** 1.131*** 21.286** 1.263*** 1.186*** 22.629**
(0.242) (0.215) (10.426) (0.231) (0.212) (10.165)
Weekly Attendance -0.004* 0.002 -0.008 -0.004** 0.002 0.005
(0.002) (0.002) (0.078) (0.002) (0.002) (0.076)
Movies -0.002** 0.002* -0.203***
(0.001) (0.001) (0.042)
Coproduction 0.030* 0.054*** 2.850***
(0.016) (0.014) (0.686)
Constant 0.362*** -0.482*** 61.047*** 0.363*** -0.571*** 67.634***
(0.104) (0.092) (4.459) (0.103) (0.094) (4.520)
Genre Fixed Effects – – – Y Y Y
Observations 5,227 5,227 5,227 5,227 5,227 5,227
R-squared 0.167 0.137 0.147 0.247 0.167 0.193
CORRELATION MATRIX OF RESIDUALS
Model I Model II
Color Widescreen Duration Color Widescreen Duration
Color 1.000 1.000
Widescreen 0.252 1.000 0.231 1.000
Duration 0.230 0.284 1.000 0.237 0.283 1.000
Notes: Data from AFI (www.afi.com), IMDB (www.imdb.com), and Carr and Hayes (1988). An observation is a
movie. The dependent variable color equals 1 if the movie was shot in color. The dependent variable widescreen
equals 1 if the movie was shot with a special widescreen process (e.g. CinemaScope or Panavision). The dependant
variable duration is the movie’s length in minutes. Major equals 1 if the film’s producer is Paramount, RKO,
Warner Bros., Fox, or MGM. Minor equals 1 if the film’s producer is Columbia or Universal. Exposure index
measures the speed of Eastman Color film (a measure of technological progress). Movies is the number of other
movies produced by the same studio in the same year. Coproduction equals 1 if the movie had multiple producers.
Robust standard errors clustered on producer in parentheses. *** significant at 1 percent, ** significant at 5
percent, * significant at 10 percent.
45
FIGURE A1 – IMPROVEMENTS IN EASTMAN COLOR, 1940-1970
Notes: Data from Ryan (1978). The exposure index measures the speed of the film: the faster the film, the
less lighting (either artificial Tungsten lighting or natural daylight) that is required for proper exposure. Each
point represents a new Eastman Color film stock released by Eastman Kodak. A doubling of a film’s