Theorizing media production: the poverty of political economy

The problem of theorizing media production

In their seminal essay calling for a political economy of communications (PEC), Murdock and Golding (1973) argued, ‘the obvious starting point for a political economy of mass communications is the recognition that the mass media are first and foremost industrial and commercial organizations which produce and distribute commodities’ (p. 205). In the second issue of Media Culture and Society, Garnham (1979) argued that ‘to examine the specifically capitalist mode of media production we need to see the ways in which capital uses the real process of media production in order to increase its value, in order to grow’ (p. 139).

Yet in a comprehensive review of the field, four decades later, Garnham (2011) concluded that PEC has had ‘a tendency to regard the system of production as a black box’ (p. 51), Mosco (2011) suggested that the field ‘has tended to situate its object within the sphere of consumption and this has contributed to a focus on the relationship of audiences to texts more than on the media labor process’ (p. 359). Mosco (2011) is also critical of accounts of media production which emphasize its creative rather than industrial character: ‘the emphasis on creativity only obscures a complex process of production, one that, however unevenly, has come to look more like a labor process in the general economy’ (p. 360).

In the absence of a theory of media production, academic study has been significantly influenced by labour process analysis. Thus, Caldwell’s (2008) study of film and TV production in contemporary Hollywood cites a key contrast: ‘whereas the rationalized Taylorist efficiencies of the classic studio or network assembly line safeguarded profits, the new world of contract production sought profits through greater forms of corporate “flexibility”’ (p. 156). A recent British collection theorizing cultural work notes, ‘the story of declining union representation in the cultural industries as they reconstructed themselves along post-Fordist lines’ (Oakley, 2013: 61).

This theory of a historic transition in modes of media production owes most to a series of articles by Storper and Christopherson (1986, 1987, 1989) and Storper (1989, 1993) whose study of film production in Hollywood identified a period of mass production followed by a transformation to flexible specialization. Storper and Christopherson’s theory of two production paradigms has been broadly influential in accounts of other media industries including UK television (Saundry and Nolan, 1998; Starkey and Barnatt 1997; Starkey et al., 2000), newsrooms (Cottle, 2007), games (Champion, 2013) and the creative industries in general (Caves, 2000; Davies and Sigthorsson, 2013; Lash and Urry, 1994). Detailed observational studies of TV productions have also found evidence to support the mass production thesis: ‘soap opera is the genre most suited to routinised production with workers performing limited repetitive tasks’ (Brennan (2004:8)). Brennan (2004) describes an RTE soap opera as a ‘standardised product’ with a production process, ‘resembling many of the productive aspects of Fordism’ (p. 8–9). Critics of the theory (see Aksoy and Robins, 1992; Hesmondhalgh, 1996; Wasko, 1994) objected to Storper and Christopherson’s argument that flexible specialization weakened the majors’ control of distribution, but did not significantly challenge their underlying theory of production. In studying UK film, Blair and Rainnie (2000) accepted that ‘in the 1920s scripts were produced by journeyman writers, cut and pasted on predictable and interchangeable lots using a semi or unskilled workforce’ (p. 189).

The influence of Storper and Christopherson’s study of Hollywood has contributed to a situation where Hollywood film production has become paradigmatic of media production in much the same way as Ford was for manufacturing (see Aglietta, 1979). DeFillippi et al. (2007) recently noted that ‘Hollywood is not only a major hub of movie production; it is also a production model’ (p. 513). If Hollywood is paradigmatic of media production, then any theory of media production must correctly analyse this model. In this article, I try to do this by returning to Storper and Christopherson’s analysis of Hollywood film production and also to the much more detailed analysis by Staiger (1985), and comparing their analysis of mass production in Hollywood with the classic case of Ford Motors. Next, Storper and Christopherson’s thesis of transition to flexible specialization in production is tested against the Hollywood case. Having argued that the mass production/flexible specialization thesis is flawed, I briefly examine an alternative, that media production is characterized by a pre-industrial, craft paradigm. The next step is to see if organization theory and media economics may provide a better foundation for theory, and test this against the Hollywood case. This draws into question the extent to which Hollywood is genuinely paradigmatic of media production. An alternative is to think of media production processes as responses both to common economic and organizational challenges and to media-specific creative and cultural influences. A full statement of this theoretical approach is the subject of a future article; however, a brief attempt is made to show the relationships between market and organizational variables and media-specific aspects of production.

Mass production in Hollywood and Michigan

To state Storper and Christopherson’s argument baldly, we are asked to accept that film production in Hollywood’s ‘studio system’ was organized in the same way as mass production in manufacturing (the same ‘mode of production’). Although Storper and Christopherson do not significantly reference her work, the first, and most detailed attempt to make this case is Staiger’s (1985) contribution to Bordwell et al.’s (1985) The Classic Hollywood Cinema (CHC). ¹ CHC is a major work of historical scholarship, hugely influential in the study of film style and as I shall try to show in the concluding section, CHC’s attempt to sketch the relationship between film styles and production systems remains a fruitful starting place for a theory of media production. My criticisms concern the theoretical interpretation (rather than the historical detail) of CHC’s account of the development of film production.

The only attempt I am aware of, to test the theory of film mass production by comparing Hollywood with Ford, is Dawson (2012). His account, although helpful, comprises only a few paragraphs and discusses neither Staiger’s (1985) detailed account nor empirical accounts of mass production (at Ford or elsewhere) and so cannot be considered a fair test. Here, I attempt a systematic comparison, the primary source for which will be Hounshell’s (1985) classic account of Ford’s development of the mass production system.

For her theory of film production, Staiger adopts Braverman’s (1974) critique of ‘Taylorism’. Taylor (1914) believed the key to Scientific Management was separation of the conception and execution of production work – so managers could design the most efficient tasks. Braverman argued that Taylorism ignored the moral distinction between social and detailed division of labour: ‘between the practice of farming, cabinetmaking, or blacksmithing, and the repeated tightening of a single set of bolts hundreds of times each day’ (p. 51). While the former ‘may enhance the individual and the species’, Braverman argued that the latter ‘renders the worker inadequate to carry through any complete production process … (and) is a crime against the person and against humanity (ibid.)’.

Staiger (1985) argues that the studios instituted Taylor’s separation of conception and execution in film production such that an ‘artisanal mode of production’, where ‘the cameraman for all intents and purposes created the entire product on his own’ (p. 231), was broken down into a detailed division of labour. Storper and Christopherson’s analysis of production is fundamentally similar: ‘it began as a craft but, with the creation of a large, assured market, the product was standardised and the production process rationalised’ (Storper, 1989: 277).

The strongest element of this theory of ‘mass production’ in Hollywood is the vertical integration of production processes on one ‘factory’ site. Staiger (1985: 92) and Storper (1989: 276) both make the point that the studios explicitly used Henry Ford as their role model. The construction of large studios facilitated a ‘factory’ system which ‘allowed an ease in controlling labor time, eliminating irregularity of production, and permitting detailed division of labor as well as reducing material and labor cost’ (Staiger, 1985: 207). Staiger shows that the studios were keen to promote their facilities as factories. A description of the studios of Twentieth Century Fox in 1919 read:

It has been laid out along scientific lines by authorities on factory and office construction, with a view to speed, economy, and concentration in every possible phase of efficient motion-picture production, from the filming to bookkeeping to stenography to starring. (Scientific American cited in Staiger, 1985: 213)

However, factories do not necessitate a separation of conception and execution, and complex tasks may remain under the control of craft workers. The purpose of Fordism was to move from factories, which enable manufacture, to mass production. Hounshell (1985) shows that it was not the construction of Ford’s Highland Park factory, but the revolution inside which transformed production: ‘Until about two years before the introduction of the Model T the factory of the Ford Motor Company resembled more closely a poorly equipped job shop than a well-planned manufacturing establishment’ (p. 220).

Staiger and Storper both identify a central role for the continuity script – ‘a paper record to coordinate the assembly of the product’ (Staiger, 1985: 215) – in Taylorizing film production. For Storper (1989) the continuity script facilitated ‘a management-orientated model that strictly separated conception from execution’ (p. 278). Staiger (1985) argues that ‘the continuities included meticulous technical instructions such as special effects and tinting directions and notes to the cutters … once the story was divided into shots … the filmmakers could tick off the shots one by one as they were completed’ (p. 228). This labour process theory of the continuity script as a design blueprint is now central to accounts of the history of screenwriting (Maras, 2009) and the geographic clustering of Hollywood (Scott, 2005).

Dawson (2012) challenges this argument: ‘mass production industries, such as the automobile industry, have no need to issue such elaborate instructions, as the assembly line or belt provides physical control and determines the pace and order of work’ (p. 29). However, this view unduly minimizes the role of manufacturing blueprints. The difference between film and car production is that while each continuity script could be used to plan production of a single film, Ford’s single blueprint was used to plan production of millions of cars. So crucial was the Model T blueprint, that he would allow no changes: ‘Ford … seemed to worship his original product design and his chief engineer stated “in all the years with the Model T no one worried or bothered Mr Ford with design changes”’ (cited in Best, 1998: 11).

Bordwell et al. (1985) argue that the rationalizing effect of the continuity script depended on standardization of the film product as the fiction feature film. CHC (Bordwell et al. 1985: 99) analyses 100 films of the period and identifies ‘repetition of characteristics considered desirable in the film’. Storper (1989) describes standardized, ‘formulaic’ films (‘the formula picture’) which were ‘sold by the foot rather than on the basis of content’ (p. 277). Staiger also shows that rise of the ‘factory’ was accompanied by a narrowing of the product range, from non-fiction scenics and topicals to fictional narrative films. Staiger (1985) argues that standardization of the Hollywood film was ‘a move to uniformity to allow mass production’ (p. 112).

Comparing this account of standardization in Hollywood with standardization at Ford reveals that the same word is being used to mean qualitatively different things. Staiger (1985) argues that standardization also allowed flexibility: ‘just as the specifications for a machine allow a degree of tolerance for individual parts, so too we might think of the standardised film’ (p. 98). Similarly while the focus was on producing feature films, she notes that studios: ‘also planned various shorts, comedies, serials, and newsreels’ (Staiger, 1985: 233). Although Dawson (2012) disputes Storper’s description of standardized products, he advances no evidence to support this view: ‘for all the sameness of Hollywood products in the 1930s, feature films were sufficiently differentiated by studio style, genre and cast to convince audiences of their individuality’ (p. 29).

By contrast, Hounshell (1985) shows how ‘Ford’s determination to produce only the Model T provided his engineers the perfect opportunity to install single-purpose machine tools’ which could produce large volumes of standardised, interchangeable parts (p. 233). ‘Every critical part of the Model T was machined in standard fixtures and checked by standard gauges both during and after the operation sequence’ (Hounshell, 1985: 229). Ford had a product range of one. Standardization meant producing vast numbers of identical products. Ford’s engineers made product standardization measurable and permitted no tolerance or margin of variation. At Ford, standardization meant absolute and measurably uniform parts and products, while Bordwell et al., and Storper and Christopherson use the term to describe subjective, stylistic similarities among unique products.

The mass production theory suggests that the combined effect of continuity scripts and product standardization was to achieve a detailed division of labour and routinize tasks: ‘a similar blueprint confronted the workers time after time, making its use routine and fast’ (Staiger, 1985:236). Film production was Taylorized so that workers were ‘completing relatively repetitive, predictable tasks’, for example, ‘make-up artists had a standard tool kit of techniques that they applied repeatedly’ (Storper and Christopherson, 1989: 340). Similarly, Staiger (1985) argues ‘sets could be lit basically the same way for illumination of the action no matter what happened in the narrative or who was playing the lead’ (p. 211). Finally, work could be planned sequentially, completing the comparison with Ford: ‘a product would move from department to department in assembly line fashion’ (Storper, 1989: 278). The result, Staiger (1985) argues, was the transformation of the work of the artisanal filmmaker until a ‘detailed division of labor facilitated mass production’ (p. 244). Storper (1989) states that ‘contrary to its glamorous and artisanal image, large firms once manufactured motion pictures via a production process organized along mass production principles’ (p. 276).

But did filmmaking involve dehumanizing, routinized, detailed labour? Describing the emergence of specialist occupations, Staiger (1985) notes that ‘it was not uncommon for a person to hold several (of these roles) simultaneously or successively’ and director and cameraman were interdependent and collaborative roles rather than rigidly prescribed and routinized (p. 212). Storper and Christopherson (1989) show that ‘crafts were handed down, father-to-son, and apprenticeships represented admission to a social world as well as preparation for a livelihood’ (p. 333).

The revolution at Ford bears much closer similarities to Braverman’s thesis. At the beginning of the 20th century, like other car manufacturers, Ford was reliant on skilled workers: ‘what tools the company possessed were general machines, operated by hard-to-find skilled machinists’ (Hounshell, 1985: 220). As Wilson and McKinlay (2010) note, this form of production ‘left the choice of technique, task sequencing and pace in the hands of skilled artisans, anathema to the new engineering logic’ (p. 761). One of Ford’s senior engineers recalled the importance of transforming this work organization: ‘to create great quantity of production … to accomplish the rapid assembly of units. There can’t be much hand work or fitting’ (cited in Hounshell, 1985: 221).

Hounshell (1985) shows how Ford’s engineers used planning, control and specific technical choices to replace the skills and control of craft workers with management control and unskilled, detailed labour: ‘the Ford tool experts designed almost all of the fixtures and gauges so that they could be used by unskilled machine tenders’ (p. 230). Henry Ford rejoiced in this complete separation of conception and execution. ‘The net result’ he announced ‘is the reduction of the necessity for thought on the part of the worker and the reduction of his movements to a minimum. He does as nearly as possible, only one thing with only one movement’ (cited in Chinoy, 1982: 87). It is here that we find a detailed division of labour which meets Marx’s (1976 [1867]) definition: ‘the worker’s continued repetition of the same narrowly defined act’ (p. 458).

Staiger, Caves and Storper all contend that Hollywood used ‘assembly line’ techniques. But again the term is being used to describe qualitatively different processes. Ford’s assembly line controlled the intensity and especially the regularity of work, enabling the coordination of thousands of workers in producing hundreds of thousands of cars (Hounshell, 1985: 237). In Hollywood, lack of product standardization made it impossible to create a single assembly line able to produce films in volume. Instead, sequential elements of film production had to be organized for each film. Effectively, when the studios wished to produce another unit of output, they had to create another dedicated ‘assembly line’. As Staiger shows, the ‘director unit’ system enabled an increase from 1 film a week to 4 films, but only by multiplying the number of production units. Far from being used to produce huge numbers of films, these units often specialized in producing bespoke sequences (such as montages).

Finally, we may quantify ‘mass’ production in the two contexts. Storper (1989: 280) shows that Hollywood’s annual output reached a peak of 497 films in 1941 and declined thereafter. ² He notes that in 1918 Universal’s studio produced 250 films – equal to the total output of the industry in 1989. By comparison as, Dawson (2012) notes, in the early 1930s, Ford produced over 1 million ‘Model A’ cars per year (p. 29). However, the crucial point about mass production is its ability to deliver economies of scale. Hounshell (1985: 224) shows that between 1908–09 its first year of production, and 1916, Ford was able to increase Model T output from 13,840 to 585,388, and thereby reduce the unit price from US$950 to US$360. Hollywood achieved no such reduction in cost per film and, as I shall argue below, this reflects a key difference between the economics of media and manufacturing production.

Alternative paradigms: flexible specialization and craft production

Around the time of the publication of CHC, a new theory of capitalist production suggested that Taylorism, Fordism and mass production had been supplanted by another paradigmatic production system, variously described as ‘flexible specialization’ (Piore and Sabel, 1984) and post-Fordism (Murray, 1987). The theory suggested that, roughly since the late 1960s, mass consumer markets had become saturated, forcing manufacturers to abandon mass production of standardized products. New flexible production systems, based on the use of programmable manufacturing technologies (rather than dedicated assembly lines) and a variable mix of independent supplier firms (as opposed to vertically integrated corporations) enabled production of a constantly changing variety of products, often for niche markets.

Storper and Christopherson adopt Piore and Sabel’s thesis to explain the end of Hollywood’s mass production system. The studios, they argue, experienced a crisis ‘typical of Fordist industries’ – a decline in consumption of their standardized products – precipitated by two ‘shocks’. The 1948 Paramount court verdict ended the ‘assured market’ created by the studios’ control of cinema chains and the diffusion of television ownership brought competition from new content suppliers (Storper, 1989: 279). The studios’ revenues began to decline, this argument runs, because their mass production system was incapable of producing the product variety and innovation necessary to succeed in these newly competitive markets. Because of their ‘curious form of standardisation of the film product … the integrated studios found, in their 1950s and 1960s profit crises, that production bureaucracies were less effective at continuously revolutionizing the product than a disintegrated, externalized structure’ (Storper, 1993: 481).

Over a long period, he argues, the studios abandoned mass production of formulaic films and developed a flexible production system able to provide the innovation and variety required for blockbuster films (‘spectaculars’). This involved the studios closing many of their own production facilities and working with a proliferation of small sub-contractors with specialist skills and technologies (like 3D). Mass production was no more; in place of repetitive, predictable tasks, production of blockbusters required ‘increasing skill specialization and flexibility’ (Storper and Christopherson, 1989: 341).

Unlike Staiger, Storper and Christopherson do not conduct a detailed analysis of film production techniques. Instead, their theory of production is part of a broader theory of the role of production in the transition to the post-studio era. Storper (1989) makes a convincing case that this transition was a response to a long-term decline in audiences and studio revenues. He shows that there was a fall in film output, a rise in the number of independent films produced and an increase in average film budgets. He also demonstrates that the fall in film output, the increase in budgets and the studios’ increasing use of independent producers and contractors was part of a decision to make qualitative changes in the type of films produced – the transition to blockbuster movies.

It is the specific role of the production system in Storper and Christopherson’s argument which appears unsustainable. Storper (1989) argues that the fall in studio revenues resulted from the inability of formulaic Hollywood films to compete, once the studios lost control of the cinemas and with the emergence of TV. However, as Sedgwick (2002: 679) and others have shown, the decline in movie audiences occurred before the ‘shocks’ of the 1948 Paramount judgment and the emergence of widespread TV ownership in the late 1950s. Studio revenues fell, not because their formulaic films were unable to compete with content from other providers, but as a result of the same post-war social trends which were causing mass markets in manufacturing to expand. As Americans were buying cars, homes full of consumer durables, and a range of domestic recreational goods and services, they were also abandoning cinemas:

When middle-class Americans moved to the suburbs in record numbers after the Second World War, they also abandoned propinquity to the matrix of downtown and neighbourhood movie theatres. In addition these young adults, previously the most loyal fans, concentrated on raising families. (Gomery, cited in Sedgwick 2002)

The fall in studios’ revenues, rather than resulting from the inflexibility of their production system, was a consequence of these changes in consumer habits which decreased the total size of the market for films. Furthermore, this change could not be related to a crisis in the Fordist mode of production since, at the time, manufacturing was enjoying a post-war boom.

Below, I shall try to develop a theoretical explanation for changes and continuities in media production (including the Hollywood case), which is more convincing than mass production/flexible specialization theories. First, however, we must deal with a contrasting theory, also influential in PEC, that media production resisted Taylorism and mass production and remains a pre-industrial, craft. Garnham (1979) originally saw in media production ‘a pre-capitalist artisanal mode of economic organisation’ (p. 149). Banks (2010) argued that Ryan’s (1992) conceptualization of the craft workshop is the media industry standard (p. 309). To explain the persistence of craft production, Garnham (1979) developed a theory of inherent barriers to industrialization of media production:

Historically the sphere of mental production or non-material production presented and continues to present important barriers to this process and the forms and dynamics of the mass media can in part be understood as resulting from a continuous attempt to surmount those barriers and from the concretely various successes and failures of this attempt. (p. 140)

Garnham (1979) proposed two potential barriers to industrialization – labour resistance and the strategies of capital. While labour resistance, via media craft unions, clearly did limit management ability to rationalize production for many years (see Mosco and McKercher, 2009), in conditions of extensive workforce casualization, this seems unlikely to be a sufficient explanation. Garnham (1979) argues that capitalists may retain artisanal production because it ‘ensures the necessary production of a range of heterogeneous cultural artefacts’ (p. 149). Ryan (1992) suggests that capital’s ability to replace skilled, craft labour is constrained by social attitudes: ‘art is conventionally held to be a product of the imagination and talents of identifiable individuals’ (p. 108). Testing this theory against the Hollywood case, we find production of a variety of films, in studio ‘factories’ with a degree of specialization and scale which sets them apart from Marx’s and Braverman’s accounts of craft production.

Recently, Garnham (2011) has shown how PEC has drawn on neo-classical economics:

Rather than attempting to deploy a very generalized theory of industrialization and commodification, (PEC) has needed to look at both specific markets and specific industry structures and dynamics as highly differentiated. (p. 49)

The next section follows this approach, drawing on media economics and organization theory to try to develop a more convincing theoretical explanation of the Hollywood production model, which might then be applied to other forms of media production.

A new paradigm? Projects and markets

Baumol (2006) suggests that differences between creative and manufacturing production generate different economics of the two industries. He argues that creative production does present inherent barriers to standardisation and rationalization, citing the example of a choreographer:

Unlike the production of automobiles or shoes, in which identical products can be turned out in apparently endless succession, the choreographer’s efforts (like the research and development division of a firm) must provide creations, each of which differs significantly from each and every one of the other products in the arena. (p. 348)

Because creative organizations produce unique products – each is a ‘first copy’ (Shapiro and Varian, 1998: 20) – they have limited opportunities to achieve economies of scale. However, as Rosenberg (1963) shows, economies of specialization may be achieved

not from the production of a completely homogeneous product but from the concentration upon a relatively narrow (heterogeneous) product range, where each small production run differs only slightly from the others and where all outputs draw upon a homogeneous collection of resources. (p. 220)

Rather than craft production, therefore, a growing literature suggests much media production is organized as a series of more-or-less similar projects (see DeFillippi et al., 2007, for a review). This theory suggests that media production is inherently flexible (and thus has not required a transition to flexible specialization) and should, over time, exhibit continuities and changes around the fundamental project model. As I shall argue below, however, this project model fails to capture differences in production which are specific to media cultural forms.

Secondly, media economics identifies the different economics of mass distribution for media firms. As we have seen, Ford achieved a mass market for the Model T by rationalizing production and lowering unit costs and thus prices. By contrast, media firms can reach mass markets without lowering production costs. If Baumol’s choreographer’s performance was filmed (i.e. it became media production) then, compared to its ‘first copy’ costs, the costs of mass reproduction (as a DVD or TV broadcast) would be negligible. Exceptional profits would be obtainable if it could be sold to a mass audience: ‘to achieve cost leadership by scale economies and experience, a tried and true competition strategy in various manufacturing contexts, is tame in comparison’ (Shapiro and Varian, 1998). So important is this factor, that Wasko (1994) and Aksoy and Robins (1992) argue that the key to PEC is distribution, not production.

While the project production model helps identify continuities over time, it does not explain change in production systems. Here we may turn to another distinguishing feature of media economics – risk. The high first copy costs already noted make media production inherently risky. These risks are compounded by the exceptionally risky nature of most media product markets. Caves (2000) argues media product markets are distinguished by unusually high information uncertainty - audiences don’t know if they will like a product until they have consumed it. Thus, although mass markets are relatively inexpensive to reach, they are also very uncertain. Our choreographer faces the problem that (contra Ryan and Banks) originality and authenticity are no guide to achieving sales. To a significant degree, ‘nobody knows’ what makes a successful media product (Caves, 2000: 3).

Based on the argument thus far, we may hypothesize that, other things being equal, media producers will attempt to reduce information uncertainty by restricting product variety and will produce recognizable, popular styles, genres, formats and franchises. Much media criticism since the Frankfurt School has evidenced this phenomenon (Adorno, 2001: 100). However, rather than being driven by the economics of standardization, as in manufacturing (as the Frankfurt theorists suggested) this relative standardization (of an infinitely variable product) enables media firms first to realize economies of specialization and second to reduce the ‘nobody knows’ risk, by increasing the information available to consumers. Lorenzen (2009:19) shows how producers who reuse film brands, stars ‘but also writers, directors, producers, and sometimes also designers and some technical staff’ increase brand value, at the expense of variety in stories or aesthetics.

However, the history of media formats and franchises demonstrates this strategy can only control risks to a degree:

The prototype issue of a new magazine or the pilot program for a new TV series can be tested, but their results can only vaguely forecast the outcome of the launching of the new product … Once in the market, each new magazine issue and each new episode of a series is subject to creativity and innovation; each one is a new project. (Reca, 2006: 195)

Yet, unlike Ford, media producers’ project production system means they are not bound to a standardized product. An alternative product strategy is the ‘hit model’ (practised in R&D intensive organizations, see Gassmann and Reepmeyer, 2005) where firms accept that risky product markets make failures inevitable. Rather than trying to spread risk, firms increase the scale and differentiation of their products in the hope one will be a ‘blockbuster’, achieving mass distribution and exceptional profits and more than recouping the costs of the failures. Despite its questionable efficacy, this continues to be the Hollywood studios’ modern strategy (see De Vany and Walls, 1999).

Our theory suggests that economies of specialization and strategy towards risky media product markets are key factors in explaining the nature of and changes in the project production systems of media firms. Translating this into the Hollywood case, we would hypothesize that, other things being equal, the studios would maximize economies of specialization in project production and minimize product market risk by producing recognizable genres and styles. And this is what Lampel and Shamsie’s (2003) study suggests. Their conception of film production accords well with Rosenberg’s idea of specialization around ‘a homogeneous collection of resources’. They define the key skills of filmmaking not in terms of technique or creativity, but as the transformation of a ‘loose bundle of resources’ (script, director, stars, etc.) into a finished product. The skills required are ‘evaluating talent’, ‘negotiating contracts’ and ‘recruiting and working with producers, writers, directors’. They show that in the studio era, Hollywood firms generally retained these key resources on long-term contracts and – once committed to these high, fixed costs – sought to achieve economies of scale by vertical integration. This logic is consistent with Staiger’s and Storper and Christopherson’s finding of relative standardization of the product via familiar genres and styles.

The theory would predict that studios would change the production system only when their product market strategies changed. As we have seen, post-war suburbanization reduced the size of the box-office market for films. Downtown cinemas faced new competition, not initially, from TV or independents, but from leisure activities in the suburbs. Producers like Twentieth Century Fox’s Daryl F. Zanuck initiated a new product market strategy, aiming to reduce market risk by making a smaller number of more expensive and more differentiated films in the hope of finding a blockbuster:

Theatregoers are more selective than ever before … This does not mean that every picture we make must be a freak attraction completely off the beaten path but it does mean that it must have at least an idea that will lift it out of the commonplace. (Cited in Sedgwick, 2002)

This change in strategy both increased ‘first copy’ costs and removed the opportunity for economies of specialization by reusing the resource bundle. This undermined the logic of maintaining the high, fixed costs of long-term contracts with creative labour and so studios began to acquire resource bundles on a project-by-project basis. As Lampel and Shamsie show, this in turn, undermined the economies achievable through vertical integration and so it was logical to reduce other fixed costs (sound stages, post-production, etc.) through ‘outsourcing’.

However, Lampel and Shamsie’s (2003) study suggests that these changes did not require a fundamental change to the project model of production: ‘the craft of movie making today is not fundamentally different than it was during the studio era’ (p. 2205). Filmmaking, they say, never involved an ‘assembly line’ but has always been ‘not strictly sequential but iterative … (and) tightly interdependent’ (Lampel and Shamsie, 2003: 2194). They conclude that, from the beginning of the studio era, the studios were well aware (despite wishing to show the parallels with Ford) that they were operating project production:

When the Hollywood movie industry achieved maturity in the 1920s, the focus shifted to the feature film … The shift coincided with the recognition in Hollywood that movie making is intrinsically a project-based enterprise. (Lampel and Shamsie, 2003)

This theory of a fundamentally unchanging project method of film production, altered by changes in strategies towards product market risk, appears to provide a better explanation of the Hollywood case than the theory of mass production/flexible specialization. However, this conception of ‘project’ production provides an explanation only at an abstract level. To provide a genuine alternative to Staiger’s detailed account of the division of labour in film production requires a theory which can articulate the relationships between market risk, project production and the division of labour and thus provide a more convincing explanation than theories founded on Taylorism (succeeded by flexible specialization) or craft control of production. Organization theory suggests that project and matrix organizations pose a fundamental challenge to Taylorism, by creating two competing sources of authority – specialist departments and project managers (Mintzberg, 1979: 170). The theory suggests that in conventional matrix organizations, project managers resolve this dilemma by controlling conception and leaving detailed execution to the specialists.

Lampel (2006) suggests that the studios developed the ‘central producer’ system as a variation of this approach (p. 46). Rather than a separation of conception and execution, the central producer was deeply involved ‘in the conceptualization and detail of the project … bringing top management downward into the process’. Hesmondhalgh (2002) follows Ryan (1992) in suggesting that a similar role – a ‘creative manager’ closely involved in the work of a project team of relatively autonomous specialists – is typical of media production (p. 50). And Hesmondhalgh (2006: 227) notes, this idea has similarities to Bourdieu’s (1984) idea of cultural intermediaries.

We can thus begin to show the relationship between product markets, media firm strategies, project production and the division of labour. Our theory would hypothesize that, as project production, media production also poses a challenge to Taylorist management methods by creating two competing sources of authority – on the one hand, the particular strategies for managing product market uncertainty and on the other, the creative autonomy required for media specialists to produce differentiated products designed to attract a mass audience. Rather than a Taylorist division of labour, therefore, we would expect creative managers to play a key role in resolving this dilemma on a daily basis.

Drawing on media economics and organization theory thus appears to provide a more convincing explanation of the Hollywood case than the mass production/flexible specialization paradigm. But to provide an explanation at the level of detail of Staiger’s account requires attention to the cultural as well as the market and organizational characteristics of media production. Mintzberg and McHugh’s (1985) study of the Canadian National Film Board (NFB) provides a good example of this problem:

The dilemma of leadership in managing adhocracy lies in trying to exercise influence without being able to rely on formal controls. NFB managers had their hands on some levers of decision, such as staffing levels and the design of the structure itself, but not on others, notably the content of specific films. Trying to manage in this situation is a little like trying to drive an automobile without controlling the steering wheel. You can accelerate and brake, but not determine direction. (p. 192)

Mintzberg and McHugh are able to identify the importance of the ‘creative manager’ but they cannot explain how such managers resolve the dilemmas presented by media production. Such analyses produce, in reverse, the problem Mosco identified with theories focused on creativity – the focus on complex processes obscures the role of creativity and culture. This is also a failing of ‘production of culture’ approaches (see Peterson, 1976). As noted above, the project model captures the similarities between film, TV, print and game production at the expense of the differences. To an extent, ‘project’ has replaced mass production/flexible specialization as the paradigm of media production. The next section makes a first step in developing a theory capable of explaining the differences between media production systems by attempting to identify the relationships between media markets, project organization, creative management and specific media styles and genres. Space prevents more than an outline of these relationships, which will be the subject of a future article.

Media-specific theories of production

Attention to cultural influences on production raises the question of whether political economy alone can provide the basis for a theory of media production. The strength of PEC has been its ability to enlighten by comparing the media to other industries. However, this has been achieved at the expense of an ability to understand and explain the role of creativity and culture in media production. This would suggest that Murdock and Golding’s (1973) conception of PEC is unduly limiting. A theory of media production needs to depart from the assumption that the media should be analysed, ‘first and foremost’, as industrial and commercial organizations.

Many writers, especially Kellner (2002), have argued that this problem should be addressed by a theoretical reconciliation, and integration, of PEC and cultural studies. However, an attempt to incorporate the cultural influences on production would suggest that there can be no general theory (or paradigm) of media production, since these cultural influences are media specific. Instead, the theory would show the relationships between common economic and organizational variables (product market risk, dilemmas of project production, creative managers) and culturally specific aspects of producing particular media content. Miège’s (1987) approach proposed that ‘social logics’ generated five types of media production (p. 274). Although he (Miège, 2011) recognized problems with these ideal types, Altheide (2004:4) provides an alternative conception, connecting media logics to the concepts of format, genre and style:

Media logic refers to the assumptions and processes for constructing messages within a particular medium. This includes rhythm, grammar, and format. Format, while a feature of media logic, is singularly important because it refers to the rules or ‘codes’ for defining, selecting, organising, presenting, and recognizing information as one thing rather than another (e.g. ‘the evening news’ and not a ‘situation comedy,’ or a ‘parody of news’).

Tracing the articulation of cultural variables with project production and economies of specialization enables analysis of the specific economics of media content genres (see also Dwyer, 2015). For example, Doyle (2002) has shown how ‘first copy’ costs vary significantly both between genres of TV (from expensive drama to low cost chat shows) and within genres (single dramas to serials and series). This suggests that the economies of specialization achievable by project production are influenced by the conventions of genre, style and format. Thus: ‘notwithstanding the spread of digital technologies, the skills, techniques and equipment involved in newspaper production and distribution are, in fact, generally still quite different from those required in the television industry, (Doyle, 2002: 32). As an alternative to rationalizing production, media firms, such as broadcasters, may change their product market strategies choosing between a broad range of content genres: ‘for broadcasters, reducing costs involves replacing expensive genres with less expensive ones’ (Magder, 2009 [2004]: 143).

Theorizing the relationship between cultural variables and the division of labour, between creative managers and other production workers, requires particular care. There is a tendency to return to Staiger, Storper and Christopherson’s approach, suggesting that genres, formats and styles act (like continuity scripts) as Fordist blueprints, enabling detailed control of execution. For example, Ryan (1992) argues that media formats mean: ‘creative work is performed to a management plan. Specific, fixed cultural rules are formulated as company policy by its creative managers and applied to members of the project team’ (p. 150). Kellner (2002) (among others) suggests that ‘production bibles’ also enable Taylorist management of production:

Within the genre, each series has its own codes and formats which are followed according to the dictates of the production company; each series, for instance, uses a manual (or ‘story bible’) that tells writers and production teams what to do and not to do, defines characters and plot lines, and the conventions of the series; continuity experts enforce the following of these codes rigorously. (p. 47)

Our analysis of the Hollywood case has illustrated the limits to Taylorism in media production. As Garnham (2011) notes: ‘while capitalist owners and managers might want to control production, they (have) few tools for so doing’ (p. 48). A more convincing articulation of the role of creative managers in the division of labour is suggested by post-Braverman theories of the labour process. Knights and McCabe (1999), for example, have proposed a conception of authority which reflects the ‘plurality of disciplinary mechanisms, techniques of surveillance and power knowledge strategies’ (p. 203). In this respect, as indicated at the start of this article, some sections of CHC still provide a good foundation for an attempt to integrate theories of film style and production. Analysing trade documents and the films themselves, Bordwell et al. (1985) explain the emergence of the Hollywood style and its production system as a product of ‘specifiable discourses discussing, describing, and validating these practices’ (p. 90). Mittell (2004) adapted Bordwell’s ‘historical poetics’ to analyse TV genres. Empirical ethnographic studies (e.g. Dornfeld, 1998) offer an alternative to the historical method, but tend to provide ‘thick description’, rather than a theoretical explanation, of the ways media producers and managers exercise autonomy and supervision in media production.