Hollywood in the world market – evidence from Australia in the mid-1930s

By the mid-1930s the major Hollywood studios had developed extensive networks of distribution subsidiaries across five continents. This article focuses on the operation of American film distributors in Australia – one of Hollywood's largest foreign markets. Drawing on two unique primary datasets, the article compares and investigates film distribution in Sydney's first-run and suburban-run markets. It finds that the subsidiaries of US film companies faced a greater liability of foreignness in the city centre market than in the suburban one. Our data support the argument that film audiences in local or suburban cinema markets were more receptive to Hollywood entertainment than those in metropolitan centres.


Introduction
A central question in the field of international business has been how firms overcome liabilities of foreignness as they enter and operate in new markets. 1 The literature recognises that some national markets are likely to be more 'foreign' than others (depending on cultural, institutional or 'psychic distance' from the home market), and this can affect the speed and process by which firms internationalise their operations. 2 Even casual observers would recognise, of course, that differences exist within national markets as well as between them. Local or regional identities are often pronounced, and urban/ rural, or metropolitan/provincial distinctions also matter. Yet much less research has been focused on the way in which multinational firms have confronted the heterogeneity of the foreign markets in which they operate. 3 Scholars are certainly aware of the value of studying liabilities of foreignness at the sub-national level, but availability of appropriate data has hampered progress in this regard. 4 This article, drawing upon two newly constructed datasets, provides a bottom-up examination of the performance of film companies through the films they marketed, not just within a country but within a specific urban region: Sydney, New South Wales. Specifically, we compare the performance of American distribution subsidiaries relative to British and Australian firms in both the City q 2014 Taylor & Francis *Corresponding author. Email: sedgwicj52@gmail.com This article was originally published with errors. This version has been corrected. Please see Erratum http://dx.doi.org /10.1080/00076791.2014.919129 centre (first-run) market and the suburban market. In doing this an examination is conducted of whether the liability of foreignness faced by US subsidiaries increased, or diminished, as it moved away from the metropolitan centre. The article adds to a (re) emerging body of work combining concepts and theory from the field of international business with sources and methods from business history. 5 In particular, it builds on recent business history literature exploring the evolution of multinational firms within the creative industries. 6 There are differing explanations of why US film distributors would have faced liabilities in foreign markets in this period, and why these liabilities may have varied between city centres and suburbs. Theory from the field of international business suggests that the foreign distribution subsidiaries of US companies would have faced disadvantages relative to local firms in terms of both their 'foreignness' and their position of 'outsidership'. 7 First, they would have been less familiar with the institutional environment, both in terms of formal rules and regulations and also informal codes of behaviour and social conventions. Studies have shown that even in industry sectors where products are completely undifferentiated, liabilities of foreignness created by lack of institution knowledge of this type are still significant. 8 In the context of the film industry, this would be reflected in the reduced ability of US distributors to understand, and appropriately respond to, different regulatory systems (e.g. censorship) and the subtle social distinctions which underpin them. US firms were also targeted by legislation in many countries specifically designed to limit their influence through such measures as quotas and tariffs. 9 Secondly, US firms would have faced 'liabilities of outsidership' due to their lack of integration into local business networks. As 'outsiders', for example, US distributors would have found it more difficult than well-connected domestic firms to obtain film bookings in the most important cinemas at the most valuable times.
'Liabilities of outsidership' would likely have been more significant for US distributors in metropolitan centres than suburban districts. 'First-run' film distribution in city centres followed a different pattern to that seen in the second-or third-run suburban markets. 10 First-run distribution was driven by the incentive to maximise the revenues (and exposure) of the most popular 'hit' films. Though few in number, first-run cinema venues were typically larger, and charged higher admission prices than suburban halls. They booked films for longer 'runs' than was the case in the suburbs. For an individual film to stand any chance of becoming a 'hit' a first-run release was essential, but access to this market was restricted. First-run venues required a limited volume of film product in any given season, and there was intense competition among distributors to supply this content. In such circumstances it is not difficult to see why foreign distributors, as outsiders in the local market, would be at a relative disadvantage. In the suburbs, on the other hand, where there were far more venues with a much more rapid turnover of films, cinema managers' required a much higher volume of films. Under these conditions US distributors were far less likely to be disadvantaged by their foreignness (or outsidership) as cinema chains were only too eager to access their extensive product catalogues. Existing studies of film-going in different local contexts would appear to confirm that these different patterns of film distribution created quite distinct film markets in which audience experiences and expectations were markedly different. 11 An alternative, cultural explanation for liabilities of foreignness in this industry relates not to the nationality of the distribution companies, but of the films themselves. US distribution subsidiaries almost exclusively handled product that was in-house and hence American in origin, and thus recognisably 'foreign' to local audiences. Studies of the international trade in cultural products such as television programmes have shown that the further these products travel from their home market, when measured in terms of 'cultural distance' the greater the discount at which they are typically sold. 12 Historians of the film industry have emphasised that such problems associated with the cultural 'otherness' of foreign films was significantly increased in the interwar period -in the words of Victoria De Grazia, 'sound appeared to have "nationalised the cinema"'. 13 The introduction of talkies heightened the cultural specificity of films. This effect was most pronounced, of course, as films entered foreign language markets, which caused European firms such as Albatross in France and British International Pictures in the UK to shift their business strategies towards a more specific focus on domestic markets. 14 The arrival of sound also sounded the death knell of the 'film Europe' movement, which had sought to create a Europe-wide market that would enable European producers to compete on more equal terms with Hollywood. 15 But talkies created challenges for US companies as well as European ones, and not just in foreign language markets. American accents, expressions and dialogue would have increased the 'foreignness' of Hollywood films even in other English speaking countries.
Why should the cultural 'foreignness' of American films have been more pronounced in the city centres than the suburbs? Some studies have suggested that the output of national film industries within Europe was often oriented towards the tastes of cultural elites within metropolitan centres. As such, these national film industries were able to differentiate their products from those of Hollywood studios, but the appeal of their domestically produced films declined as they travelled to more provincial areas. As John Trumpbour argues, 'frequently missing or misrepresented in domestic cultural works, the provinces tend to settle for Hollywood over national film producers. . . . Films of the European metropolis (Paris and London), then, are sometimes regarded in the provinces as more alien than the products of Hollywood'. 16 The prevalence of American over domestic films in provincial French cinemas is also noted by Ulff-Moller. 17 The proposition that films of domestic film industries might become less popular (and more 'alien') the further they travelled from metropolitan centres has implications for the US firms. Liability of foreignness is a relative concept, and in these circumstances US films would appear much less 'foreign' (in relation to domestic rivals) when operating in provincial areas than in metropolitan ones.
These competing explanations for the liabilities facing Hollywood in its international markets could be summarised thus. On the one side, a liability related to the foreignness of US firms (distribution subsidiaries) reflected in limited market knowledge and a position of outsidership in local business networks. On the other, a liability caused by the foreignness of American films, related to their cultural distance from the market and reduced appeal for local consumers. In this article we explore the extent to which these liabilities varied within a specific foreign market. In particular, we examine whether the relative foreignness (both of US firms and US films) was more pronounced in the metropolitan centre than the suburban districts.
The present study draws on the exhibition records of every film released into Sydney's city centre (first-run) and suburban markets in the year 1934 -a year in which the NSW legislature commissioned Mr F.W. Marks to conduct an inquiry into the film industry as it operated in the State. By this time US film distributors were experienced operators in the Australian market, the transition to talking pictures was complete and Australia was established as an important foreign market for American films. The data collected allow an evaluation to be made about the relative competitiveness of each distributor operating in these markets, and thus the relative performance of Hollywood, British and domestic film distributors in both market segments. In doing this, a microscopic light is thrown on the rivalry between the distributors and their (mostly) Australian exhibitors, as the distributors sought to get their product screened as widely as possible for the highest possible rental price. As important as the structural arrangements which shaped business practice are, this, however, is not the whole story. In recording the exhibition records of all films marketed, and in the absence of film revenue data on a film by film and cinema by cinema basis, we are able to establish an index of film popularity (POPSTAT) which reflects the extent to which consumers and their preferences impact upon the said business environment.

Context
Although the area of Australia is of the same order of magnitude as that of the USA or of Europe, west of the Urals, in terms of population Australia was a small country in the 1930s. The 1933 Census recorded 6,629,839 inhabitants ('exclusive of full-blood aboriginals'), of which more than a third lived in the two great metropolitan centres of Sydney and Melbourne. Such was the size of their populations in the mid-1930s that both were among the world's largest 40 cities, with Sydney placed 19 th and Melbourne 36 th . When the comparison is confined to the British Empire (including Dominions), Sydney was the second largest city in the British Empire, after London (population over 8 million), ranking above Calcutta, Bombay, Glasgow, and Birmingham, followed by Melbourne in 7 th place. 18 With just under 80% of the population not born in Australia originating from the British Isles, the 1933 Census found that 99% of Australians identified themselves as being of British 'race' -a factor that is likely to be significant in understanding the positive reception afforded British films, particularly in first-run markets. 19 This differs markedly from the US experience, where immigration, although overwhelmingly European, was much more heterogeneous, culturally and linguistically. 20 As Jill Julius Matthews has argued 'Australia rejected the American experiment of the melting pot and aspired instead to create a racially, linguistically and culturally homogeneous and egalitarian nation. As a result, Australia's population remained small, compact and British. 21 As in Great Britain and the US, there is ample evidence to suggest that film-going in Australia during the 1930s was extremely popular, particularly with young people, even allowing for the social and economic context of large scale unemployment and social deprivation. 22 A Royal Commission on the Moving Picture Industry in Australia in 1928 reported that 'The annual attendances at picture shows in Australia are estimated to be 110 million', a figure that indicates that on average Australians went to the cinema 17 times a year, which is comparable to the frequency of attendances in Britain and the US in the late 1920s. 23 In fact, just prior to the Depression, US Department of Commerce reports indicate that Australia was Hollywood's third most important market -after North America and Britain. 24 However, a report based on 36,936 foreign accounts held by the distribution branches of Hollywood studios, published in the trade journal Variety in 1934, placed Australasia after Great Britain, France, Germany, Italy, and on a par with Czechoslovakia and Spain -Portugal, in terms of the revenue potential of these markets for a single 'outstanding film'. With regard to the Australasian market, this potential amounted to US $80,000, although when calculated by per capita expenditure, Australians and New Zealanders were ranked second only to the British in terms of film consumption. 25 The international earnings of the US film distributor United Artists that are represented in Table 1 and Figure 1, provide further evidence of Australia's significance for Hollywood firms. In the mid-1930s, Australasia was the second largest foreign market for United Artists (after the UK), and it remained an important regional market (comparable with Europe, Asia and South America) throughout the 1930s and 1940s Kristin Thompson, Exporting Entertainment, has shown that during the 1920s the major Hollywood studios set up in-house distribution networks in Australia, with headquarters in Sydney and Melbourne. At this time, two national cinema chains -Hoyts  and Union Theatres -emerged, coming to dominate the exhibition sector in the State capitals, and co-existing with a set of much smaller suburban and rural chains and a multitude of independent exhibitors (Shirley and Adams, Australian Cinema). By the mid-1920s the industrial structure of the Australian film industry was similar to that of the UK, comprising an indigenously owned exhibition sector, a distribution network dominated by the distribution arms of the major studios that practiced block and blind booking practices, and an amorphous, undeveloped production sector. Industrial organisation was tailored to the imperative of making those films that were revealed to be popular with audiences more available than those films that were not so popular, at the nexus of which was the complex relationship between distributors and exhibitors (as chains, small chains and independents): complex because of the nature of the intra and inter-rivalry between the two sets of agents. Ten years on, US firms maintained their dominant position in production and distribution, but the question of whether to introduce protective legislation was a live issue for both Federal and State Parliaments. In Britain, the 1927 Cinematograph Films Act had stimulated production, while at the same time the market power of American distributors was lessened as a result of the practice of block and blind booking being made illegal. These interventions in the British market led to a more highly capitalised production sector. 26 Better films resulted in British producers taking an increasing share of the home market. Furthermore, many of the films that found favour with British audiences were also widely screened in Australia by purpose-formed distribution companies. As will be shown, British studios were significant players in the Australian market by the mid-1930s, particularly in the first-run sector.
In contrast the Australian industry received no such support. The issue of protecting home film producers was one of the awkward, multifaceted problems investigated by the Royal Commission on the Moving Picture Industry in Australia. Reporting in 1928, the Commission proposed that an Empire quota (10% after two years, rising to 15% in the third year) be imposed on exhibitors. However, the Federal Government ultimately left this to the individual States to decide upon and no such decision was undertaken. 27 In New South Wales, the State Legislature set up an Inquiry in 1934, chaired by F.W. Marks, at which much of the evidence heard by the Royal Commission was replicated. 28 Marks, while rejecting calls for an end to block and blind booking, also proposed a quota starting at 4% and 5% of all feature length films for exhibitors and distributors respectively, rising to 12.5% and 15% over five years, leading to State legislation in 1935. 29 One of the more striking aspects of the 'Inquiry' perhaps not surprisingly, was that the testimonies made by the various interested parties were very similar to those made by their counterparts elsewhere in Ottowa (1931); Wellington NZ (1934);London (1935);and Washington DC (1936). 30 It would appear that the concerns felt by US distributors and domestically based exhibitors were common throughout the English-speaking world, albeit with distinctive local characteristics.
This study focuses on the city of Sydney, which is taken to be indicative of urban cinema-going in Australia, and is derived from a comparison of the City centre first-run market for films with the subsequent-run market for films in Sydney's suburbs. The firstrun market is analysed through the exhibition records of all 11 City-centre first-run cinemas. 31 Derived from a listing on page 2 of the daily Sydney Morning Herald, a second, suburban, dataset records the twice/thrice-weekly change double-bill film programmes of 69 cinemas -65 of these are drawn from the population of 160 suburban cinemas (see Table 2) located in the Sydney conurbation, together with four second-run cinemas located in the city. 32 These cinemas were supplied by one of nine distributors: the vertically integrated distribution arms of the Hollywood studios (Fox, MGM, Paramount, RKO, United Artists, Universal, and Warner Bros.) plus two local distributors -BEF (British Empire Films), which distributed films emanating from various British studios including British International Pictures, and GAF, responsible for the films of the Hollywood studio Columbia, as well as a number of British studios. 33 The organisation of exhibition in Sydney conforms to the expected run-hierarchy (cascade) pattern outlined earlier, in which City centre first-run cinemas screened films earlier and charged higher prices than City centre second-run cinemas, which in turn screened films earlier than suburban cinemas. Furthermore, while the principal City centre cinemas operated with weekly or less frequent changes in programme, most suburban cinemas changed programmes on a twice-or thrice-weekly basis.
These two sets of cinemas served different functions within the broad goal of getting as many consumers as possible to pay at the box office: the first-run was geared to attracting audiences to particular films at particular cinemas, with exhibitors screening films for as long as target audience levels/box office receipts were reached, while the second-and suburban-run was concerned with distributing particular films as widely as possible among geographically dispersed cinemas, with limited screening time at any one cinema. The temporal element in this process saw films transfer over time from box office rich to box office poor cinemas, through specified clearance zones of one month between first-run and suburban-run. The spatial element had films transfer from City centre cinemas with wide attachment areas, to suburban cinemas, which catered to local audiences only. 34 Less than a tenth of the population of Sydney lived in the city of Sydney. However, it was the location of the Central Business District into which tens of thousands of workers commuted daily. Gregory's Street Directory of 1936 shows Sydney had a highly developed transport system (train, ferry, tram, bus and trolley bus) linking the City to the suburbs, with some 500 million public transit passenger trips made in the City annually during the 1930s (Cosgrove, 'Long-term patterns of Australian public transport use'). Even with this caveat, the supposition made is that they represent a more accurate estimate than that provided by GTC, on the grounds that: (a) the cinemas are actually listed in the Yearbook; and (b) the Film Daily's reputation among exhibitors required it to furnish accurate up-to-the minute information.
Hence, in all likelihood, many of the patrons of city centre cinemas lived in the suburbs and either caught films after working hours, or came into the City especially over the weekend in search of entertainment (Matthews, Dance Hall and Picture Palace). Table 2 shows the General Theatres Corporation (GTC) combine was a dominant presence in Sydney's city centre cinemas, controlling seven out of the 11 first-run cinemas, and 13 out of the 19 first-and second-run cinemas, (which made up 72% of seats in the City). It was formed in 1933 from the amalgamation of the two most significant cinema chains in Australia at that time -Greater Union Theatres, a reincarnation of Union Cinemas (liquidated in 1931) and Hoyts, a controlling interest in which had been bought by Fox in 1930. 35 While GTC had a sizeable presence in the Sydney suburbs, Table 2 shows that it was much less prominent in the country areas. 36 Fox retained a significant financial interest in the new combine. 37 During the mid 1930s a major concern for GTC was that if the in-house distributors of the Hollywood majors came into the exhibition sector, building new large-scale cinemasas in fact happened in Melbourne with the opening of the Metro (MGM) in Melbourne in 1934 -they would naturally privilege their own in-house films, thereby taking audiences away from established first-run cinemas. Indeed, this potential scarcity was one of the terms of reference for the 1934 NSW Inquiry set up to investigate the State's film industry. 38 For US film distributors, the strategic question was whether to rely on local cinema operators to handle the first-run release of their films, or to establish their own city centre exhibition outlets, thereby completing a vertically integrated chain from production through to exhibition. Only MGM adopted the latter course of action, although as we have seen, Fox also possessed a financial stake in GTC. On this matter, it is instructive to quote from MGM's Managing Director in Australia Mr N.B. Freeman who gave the following reasons to the Marks inquiry in support of erecting a new theatre in Sydney: (a) 'To obtain an assured release for our products and to maintain continuity in our suburban releases. (b) To be in a position to deal with the trade on a competitive and fair footing with the distributors allied to the Combine. (c) To resist monopolising tendencies on the part of the General Theatres Corporation and to maintain the independent exhibitors in business. (d) To handle our exhibition of pictures in as attractive a manner as possible'. 39 In a nutshell, this testimony captures the strategic concerns facing a number of American distributors, faced by, as they perceived it, an aggressive monopsonist that wished to limit their access to, and exploitation of, first-run audiences. Indeed, the fact that US firms, but not Australian ones, were identified with the perceived problem of oversupply of cinemas is a classic example of the liability of foreignness.
A noteworthy element to Table 2 is the density of persons per seat and persons per cinema, which appears to be greater than in Britain at the time, where Simon Rowson, 'A statistical survey of the cinema industry in Great Britain in 1934', reported that there was one cinema for every 10,600 persons. 40 Remarkably, given the spread of the population in rural areas, there was little difference in population per seat between Sydney and Suburbs and the remainder of NSW, which means that as rural cinemas were significantly smaller than those in the city and suburbs, there were more of them per head. 41 Perhaps the most striking feature of the table is the sheer size of the suburban and rural markets in comparison to the metropolitan one. If we compare just the city centre and suburban sectors, we see that there were more than seven times the number of cinema seats in the suburbs compared to the city. Given that film programmes changed more frequently in suburban cinemas it seems unlikely that seating utilisation rates would have been much lower here than in the city centres. 42 Prices, of course, would have varied very considerably, but even if we assume that average ticket prices were three times higher in the city centre, this would still mean that the suburban market would have been almost two-and-a-half times the size of the metropolitan one. It would probably be reasonable to estimate that the suburban and rural markets combined would have been at least three times larger than the city centre one. Indeed, the trade journal Everyones [sic] carried an editorial that suggested that the distributors essentially make their money from suburban release, arguing: ' . . . few long runs on percentages are really profitable to the distributor, whose principal return, as a matter of fact, is the prestige the city run builds for his picture in the subsequent release' (Everyones, 28 March 1934). Table 3 provides summary statistics of the two datasets. Three observations are pertinent. First, large numbers of films are involved, almost all of which came from the United States or Great Britain (see Tables 7 and 10 below). Indeed, most films released onto the American and British markets were also distributed in Australia, implying that neither Table 3. Summary data of first-run, and second and suburban-run datasets of films screened in 1934 a . Source: Everyones, Sydney Morning Herald, The Telegraph. a In order to capture the temporal progression of films from first-to second-to suburban-run exhibition, all films that were released in the city centre cinemas in the later part of 1933 but received their suburban release in 1934, as well as films that opened in a city centre cinema in 1934 whose suburban-run was completed by 31 May 1935 are included. b Unlike the suburban-run, it was common for films in the first-run to be held-over for a second week, or more. Table 5 records the longest runs for each of the first-run cinemas American nor the British studios held back films from the Australian market as a matter of course on the grounds of unsuitability. 43 Second, the double-bill programme dominated absolutely in first-, second-, and suburban-run markets. 44 Third, an indication of the velocity with which films circulated once having had their premiership in a city centre cinema can be ascertained from the mean number of film bookings of 18.9, a figure that rises to 23.2 if we remove the 142 films that received no more than one booking. Thus, on completing its city centre first-run, the typical film would move through approximately 20 of the sample 65 suburban cinemas within a three-to-six month period. Of course, with so many suburban theatres in operation and double-bill twice/thrice-weekly change programmes the norm, the velocity of film circulation needed to be high to service such a large number of programmes. Details of the film programmes shown at the first-run and a number of second-run cinemas in the centre of Sydney were collected from the daily advertisements found in the daily newspapers The Telegraph and the Sydney Morning Herald, and the listings found in the weekly trade journal Everyones. Everyones is a particularly important source of information. In addition to the cinema seating capacities, screenings per day and admission prices (reported in Table 4), Everyones rated the films screened at each cinema on their predicted exhibition strength, applying, in order of merit, the categories Big, Excellent, Good Feature, Feature, Good Support, Support, Poor. Annually it awarded points on a decreasing scale, from six to zero, to each of these categories and compared the average film scores of those films handled by each of the distributors during that year. 45 In both 1933 and 1934 MGM topped the Everyones list of the major Hollywood and British studios. 46 The trade journal also recorded the weekly box office takings of each cinema for each week of a run as a percentage of the cinema's mean weekly takings over the previous 52 weeks. This provides the pattern of earnings of any single programme over the course of its run, but does not tell us actual box office performance. In the absence of such data a Cinema Weight Index (CWI) has been constructed, which was derived for each cinema by multiplying its seating capacity, its mid-range price (in pennies), and the number of programmes screened daily, to give an estimate of each cinema's daily revenue potential, expressed as a proportion of the mean revenue potential of all cinemas in the set, thereby generating a reference point of CWI ¼ 1. 47 The index is presented in Table 5.

The datasets
Combining the cinema weight index with the weekly record of relative box office performance for each film screened at one of Sydney's 11 first-run cinemas found in Everyones, together with a film's billing status, provides the basis for estimating an index of box office revenues of the 491 films that form the sample of first-run films. This index (POPSTAT) takes the form: where POPSTAT i ¼ Film Popularity Index for the ith film screened over the duration of the investigation; cwi j ¼ the cinema weight index of the jth cinema (of 11), (shown in Table 5); bo ikj ¼ the box office performance of the ith film during the kth week of its exhibition at cinema j expressed as a percentage of the average weekly revenue achieved by cinema j over the previous 52 weeks (published in Everyones); n ij ¼ the number of weeks film i was shown at cinema j     (3) Mean length of run per film (weeks) (4) No. of single-bill programmes (5) Average percentage box office b Median film quality c (7) Description of film booking policy (8) Longest runs (weeks) (9)  Squatter's Daughter (9) Hayseeds (8) Hayseeds (1355)  Masquerader (7) House of Rothschild (5) Masquerader ( Derived from the Everyones quality ratings, in which films were rated from 0 (Poor) to 6 (Big).
d The values in parenthesis represent the aggregation of the weekly box office performance expressed as percentage points, where 100 represents the average box office performance.
bs ij ¼ the billing weight of the ith film at cinema j. In particular, a single bill film received a weight of 1, films receiving joint billing received a weight of 0.5 each, a film receiving top billing of a double bill received a weight of 0.8 and the secondary film on such a bill received a weight of 0.2.
As explained earlier, the sample of second-run and suburban cinemas found in the Appendix comprises those cinemas whose programmes were listed daily on page 2 of the Sydney Morning Herald. The 69 cinemas in the list had between them 105,933 seats, giving a mean seating density of 1581 seats per cinema. 48 Table 2 shows the sample set of cinemas to be drawn from a suburban population of 160 cinemas, supplying 205,398 seats altogether, with a whole-population average of 1284 seats per cinema. 49 The sample thus constitutes 43% of suburban second-run cinemas, but 52% of seats.
The locations of the suburban cinemas have been identified using Gregory's Street Directory of 1936. Other than the six second-run cinemas found in the City, the cinemas are split 40 to 25 between inner and outer suburbs, with 25 cinemas located in the southern, 18 in the western, 15 in the northern, and 7 in the eastern suburbs. The main chains were Broadway Theatres with six cinemas; GTC-Hoyts with 13; and Western Suburban Cinemas with 10. A single owner owned most of the remainder, although the exhibition records of these cinemas suggest that many of them booked through one of the chains. With the exceptions of the Arcadia, Chatswood, the Kinema, Mosman, and the Star, Bondi Junction, all of the cinemas listed had twice-or (occasionally) thrice-weekly double-bill programmes.
With the seating capacities of these cinemas ranging from 540 to 2599, the cinemas in the sample, like those in the first-run, will have generated quite different levels of revenue for their owners and the film distributors. As with the first-run sample, the solution is to give cinemas weights based upon each cinema's revenue potential expressed as an average of the set. These are also to be found in the Appendix. Unfortunately, as admission prices and the number of screenings at these cinemas are not known, the weights are, in effect, based upon the relative seating capacity, requiring the assumptions to be made that all cinemas in the suburban sample: (1) ran a common number of screenings per day/week; (2) charged a common range of admission prices; and (3) (the assumption also made for the first-run houses) experienced the same seat utilisation ratios, irrespective of what film programme was being screened. The biases implicit in these assumptions will tend to underestimate the relative earnings of films being screened in those better accoutred and larger cinemas such as the Star, Bondi Junction, which got their films earlier and probably had higher seat utilisation statistics, than films being shown in less well furnished, and smaller, cinemas such as Mortdale Theatre, Mortdale, that screened films towards the end of their runs. Thus the POPSTAT Index for films being screened among the set of suburban cinemas is calculated for each film by aggregating the cinema weight, the billing status and the length of run of each programme in which it was recorded. 50 The first-run market First-run film distribution was driven exclusively by the objective of maximising film revenues, in cinemas charging premium admission prices, and hence was characterised by the extended distribution of popular films, and the ruthless weeding out of films that attracted relatively limited audience responses. This process necessarily produced highly skewed revenue distributions, in which the long right-hand tail contained the relatively small number of universally popular films that generated substantial profits for the distributor and exhibitor. Figure 2 (a) depicts the frequency distribution of the Sydney first-run POPSTAT values in which 75% of the films fall into the first decile class, causing the measures of location of the distribution to also fall in the same decile group, hence giving rise to a distribution that is highly positively skewed. Furthermore, as can be seen in Figure 2 (b) and (c), this distribution is very similar to those generated in a study of first-run cinemas in Great Britain and the United Sates during the mid-1930s. 51 For distributors, the ability to secure an extended release in the first-run market not only enabled them to maximise revenues from their most popular hit films, it also helped them to raise market awareness for films prior to their wider national release. However, competition among distributors to achieve the maximum possible exposure for their films was intense. In the Australian market, US films were in competition not just with each other, but with films from both domestic and British producers for access to first-run cinemas.
As shown in Table 4, of the 11 cinemas in the first-run, seven were owned by GTC. Of the remainder the Mayfair had a British only booking policy, while the Liberty took top product from the Columbia and Universal studios, Prince Edward showed a high proportion (31 out of 40) of Paramount films, and St James, as is known, screened MGM product exclusively. This state of affairs draws attention to the differences of interest between American distributors and the GTC organisation, referred to earlier, in that GTC had no access to the films of MGM and Paramount, and had only restricted, weak-product access to those of Columbia, Universal and Warner Bros. For a chain that controlled seven of the eleven cinemas in the City this was a major constraint on their operation. As has been indicated, the prospect of even more city centre first-run cinemas, particularly if they were owned or leased by the 'major' US distributors, was a cause of great concern for GTC management. 52 However, from the American distributors' point of view, the fact that GTC had such a stranglehold on first-run exhibition meant that they were strategically vulnerable: if they distributed to GTC they faced a monopsonist buyer that could exert its market power by negotiating a higher fixed guarantee and/or lower rental prices (the proportion of the box office commanded by the distributors), yet failure to secure bookings with the dominant circuit restricted their access to the first-run city centre market and its function to showcase films for wider distribution. GTC's booking policy appears to have been less than generous in its allocation of screen time to some US firms, suggesting a liability of outsidership faced by US distributors, who found if difficult to secure bookings in the most important first-run cinemas. 53 US distributors embarked on different strategies to overcome this liability in the city centre market. MGM and Warner Bros. actually trialled a straight-tothe-suburbs distribution policy for some of their product in the months of April and May 1934. 54 Ultimately MGM secured its own cinema by acquiring the lease of the St James in June 1934. 55 Columbia, on the other hand, opted not to distribute its own films in Australia, but instead contracted with GAF to distribute its pictures in this market. Alternatively Fox, while maintaining its own distribution subsidiary, also acquired a minority shareholding in GTC. This provided Fox with more favourable access to GTC cinemas than most other US distributors. RKO also struck a deal with GTC to supply films to its cinemas. Warner Bros. and Paramount, however, with their extensive film portfolios, appear to have been most at risk of being squeezed out of the first-run market. The range of strategies adopted by US firms to access this highly competitive market were very similar to those seen in Britain where national distribution hinged on a booking with one of the three dominant cinema circuits. 56 A notable aspect of first-run exhibition in Sydney was the screen time allocated to Australian and British films. The entry for the Civic in Table 5 Tatts (Efftee Film Productions), and The Man They Could Not Hang (Invicta Productions). The presence of British films in this first-run market was unquestionably distinctive. Three of the 11 cinemas were dedicated to the screening of British films, the Embassy and the Mayfair exclusively, with the Lyceum breaking ranks by showing three non-British films. This is noteworthy because in Great Britain itself there is no record of any cinema pursuing a British-only screening policy. Unlike their American counterparts, British film distributors appear not to have encountered a noticeable liability of outsidership in this first-run market. Table 6 presents the top 30 films at Sydney's first-run cinemas, ranked according to their POPSTAT scores. The results are remarkable in that three Australian films -The Squatter's Daughter, The Silence of Dean Maitland and The Hayseeds are ranked in the Top 10, ahead of a luminous parade of top Hollywood films such as Dinner at Eight, and Gold Diggers of 1933, with a fourth film, Ticket in Tatts, ranked at 19. 57 In the British and North American markets these films were either not released or received hardly any distribution. While Hollywood films were unquestionably highly prominent in Sydney's first-run market, it is notable that almost half of the top 20 films were British or Australian in origin. With these films occupying such valuable screen time, the competition among US distributors for film bookings in city centre cinemas was all the more intense. Table 7 presents information about the performance of distributors in the Sydney firstrun market. The shares of the five major Hollywood distributors (MGM, Fox, RKO, Warner Bros. and Paramount) are quite evenly matched, lying respectively between 11.4% and 8.8%. Interestingly, the higher market share performances of Fox and RKO relative to those of Paramount and Warner Bros. in the Sydney first-run market is the reverse of that found in both the US and the British markets, a fact that is likely to have been connected to the contractual ties of Fox and RKO with the GTC combine, with the consequent probability that their films received relatively greater exhibition exposure in Australia, while Paramount and Warner Bros. were constrained in their access to first-run cinemas. 58 The poor market share performance of Paramount, especially given the large number of films it marketed during the year relative to the other four Hollywood majors, is at least partially explained by its dependence on the Prince Edward cinema -a venue with only average revenue potential (CWI) -to showcase its films. Having the lowest market share of the five majors is in marked contrast to its position in the North American and British markets, where it was placed second only to MGM in both cases. 59 British films appeared to be well received, accounting for 22% of all pictures released in first-run cinemas but holding a market share of just over 24%. As Table 7 shows, the median POPSTAT score of all British films was higher than that of all American movies. Those distribution companies that handled both British and American pictures achieved notably better results with their British films. In the case of GAF, the median POPSTAT achieved by its British films was more than double that of its American output (97.9 compared to 48.3). The median POPSTAT of the British films distributed by Fox (26 of which were produced by the Gaumont British and Gainsborough Studios) was higher than for any distributor of American films apart from United Artists -a specialist distributor of small numbers of high profile films by leading Hollywood independent producers (Miskell, 'Selling America to the World?'). Fox's British films achieved a market share comparable with those of the major US film distributors, despite there being only 35 of them in total.
Finally, we must direct our attention back to the extraordinary performance of the set of 10 Australian films released in 1934. Table 7 indicates that if treated collectively, the Everyones' mean rating anticipates a box office reception for these films second only to the films distributed by United Artists. The median POPSTAT performance of these films is considerably greater than that of any single studio, Hollywood or otherwise, providing clear evidence of the extensive exposure granted to these films in the first-run market coupled with considerable audience interest.
Our comparison of the median POPSTAT scores in Table 7 has been informal to this point and we will complete this section with a number of formal tests of the equality of the median POPSTATs of national producers and distributors in the Australian first-run market. Given the highly skewed nature of the sample of POPSTATs (see Figure 2) the median, as we have argued, is the appropriate measure of central tendency, and nonparametric testing procedures will therefore be required. Table 8 presents these tests, derived from the Kruskal-Wallis procedure. The null hypothesis in all cases is that the median POPSTATs are equal. The first row of the table compares the median POPSTATs of US, British and Australian films in the first-run market (67. 6, 86.8 and 311.9, respectively, from Table 7). The second row compares the median POPSTATS of just US and British films. The third row tests for the equality of median POPSTATs across the eight US distributors (ignoring 'Other'), the fourth row drops United Artists from this comparison, given its atypical characteristics as a distributor. The final row compares the median POPSTATS across the four distributors of British films (again ignoring the 'Other' category). Also presented are the results of the Brown-Forsythe test for the equality of the variances, a test that is appropriate for the case of non-normal distributions -the strict Table 6. Top 30 Films screened in Sydney's first-run cinemas. application of the Kruskal-Wallis test requires that the populations from which the samples are drawn have equal variances (and shapes).

Rank
In terms of the country comparisons the Kruskal-Wallis test provides strong evidence for different population medians across the three countries, although the significant Brown-Forsythe test suggests that equal population variances cannot be assumed, thus questioning the strict validity of the Kruskal-Wallis test here. Limiting the comparison to just the US and Britain (row 2) suggests similar median performance levels of US and British films in the first-run market, with the Brown-Forsythe test supporting the assumption of equal population variances.
Comparing the performance of US distributors (row 3) differing median POPSTATs are implied, although the Brown-Forsythe test is significant. Dropping United Artists (row 4) a conclusion of differing performance levels across US distributors is still supported, with an insignificant Brown-Forsythe test now produced. Distributors of British films, on the other hand (row 5), achieved broadly consistent levels of performance.
In terms of examining just country differences, a refinement of these testing procedures can be derived within a regression context by first attempting to standardise for film quality. We have two measures of film quality -the Everyones quality rating, ranging from six for a 'Big' film to zero for a 'poor' film, and the POPSTAT value achieved in the British market. 60 These two variables are denoted by EVERYONES and BRPOP, respectively. We then define two binary variables -AUSTRAL which takes on the value 1 if the film is Australian, 0 otherwise, and BRITISH which takes of the value 1 if the film is British, 0 otherwise. The comparator country is therefore the US. The dependent variable is the POPSTAT value achieved by a film on the completion of its run in Sydney's first-run cinemas -FRSTPOP. Table 9 presents the results of the regression analysis. Two regression equations are presented, Equation (1) derived from the least squares procedure, thus focusing on the mean of the dependent variable, and Equation (2) derived from the quantile regression procedure, focusing on the median of the dependent variable, and hence directly comparable to the tests in Table 8. Initial estimates of both equations implied the presence of a non-linearity which was resolved by using the natural logarithm of the dependent variable -Log (FRSTPOP). Of the 491 films released in the Sydney first-run market over the data period, data observations for BRPOP and EVERYONES were available for 462 of these films.
The two equations produce consistent results with regard to the significance of the independent variables -the coefficients on both BRPOP and EVERYONES are positive (as expected) and significant, and the positive and significant coefficients on AUSTRAL and BRITISH implies a preference for Australian and British films over US films, once quality has been accounted for (but in the case of Australian films this derives from just two films which were released in both the British and Australian markets -The Silence of Dean Maitland and Ticket in Tatts). Thus in contrast to the results in Table 8, the regression results imply a preference for British over US films, albeit with significance at just the 5% level. However, as already suggested, a number of additional strategic factors were also at play in the first-run market, in terms of the pro-British booking policies of some of the first-run cinemas and the differential access that US distributors had to some cinemas. A fuller exploration of these issues would require a more complex analytical framework than that presented in Tables 8 and 9, and hence beyond the scope of this article. Thus the conclusion that can be drawn from Tables 8 and 9 is that British and US films achieved broadly comparable performance in the first-run market, with a suggestion of perhaps marginally superior performance of British films over US films, with the small number of Australian films achieving outstanding performance.

The suburban market
The suburbs were where Sydney's inhabitants lived. Comprising the film bookings of 69 cinemas (listed in the Appendix) -65 suburban cinemas, two second-run city cinemas (the Empire, and the Empress), and two suburban-run cinemas located in the city (the Broadway and King's Cross) - Table 3 shows suburban cinema to have been considerably larger in scope than the first-run. Altogether these cinemas screened 733 films over the sample period, for the most part on double-bill, twice weekly change programmes. Of these, 595 films received more than one booking. In contrast to the objective of maximising revenues per film of the first-run market, the suburban-run market was conditioned by the objective of stimulating habitual weekly film-going on the part of suburban audiences, via the mechanism of regular changes in film programmes, which typically occurred twice weekly, and sometimes thrice weekly. Thus the popularity of films released in suburban cinemas, and hence their POPSTAT values, was reflected in the number of cinemas in which they were shown, rather than extended exhibition in any given cinema. Figure 3 presents the decile distribution of the Sydney suburban POPSTAT values for the 733 films. Clearly, as indicated above, a great many films received little in the way of distribution. Nevertheless, with the mean and median falling within the third decile class, and a much lower coefficient of variation, the histogram in Figure 3 differs considerably from the long tail distribution generated by the first-run market, found in Figure 2, with the suburban market exhibiting a much shorter right tail. The two distributions reflect the different functions served by the two exhibitions sectors.
To obtain the throughput necessary for a twice-weekly change programme, exhibitors rented widely from various distributors. For instance the Acme, Rockdale (an outer southern suburb) took no films from BEF and United Artists, and one film only from Warner Bros., but 44 films from Paramount, 42 films from Fox, 39 films from MGM, 36 films from RKO and 23 films apiece from GAF and Universal, while the Crow's Nest, in the inner northern suburb of the same name, took no films from Paramount, three films from MGM, while screening 67 films from Fox, 44 films from RKO, 39 films from GAF, 23 films from Warner Bros and 18 films apiece from BEF and United Artists. Such multisourcing was the norm for exhibitors. Hence, while block booking contracts were entered into by exhibitors, as a general rule they were not dependent on any one distributor for their supply. For their part, distributors needed to organise the planned-for number of prints and publicity materials expeditiously, moving their supply out in steps from box office rich inner suburban cinemas screening four programmes a day, to smaller cinemas on the fringes of the suburbs where perhaps only one screening a day took place. By this time the potential audience for most films had been exhausted. In addition to the suburbanruns, these same prints were also required to service cinemas in rural areas.
By scaling-up the pattern of supply evident in the two cinemas given above, the suburban market appears to have absorbed virtually all of the output of the US studios, and this was likely to have been the case for British output. Evidence for this can be found in the studio ledgers of MGM, RKO and Warner Bros. 61 Between them they released 247 films onto the Australian market over the sample period, which consisted of 73% of the films that these three studios produced during the 1932/33 US season, and 87% of the films produced in the 1933/34 season, with a further 11 films released from earlier seasons, and seven films from the early 1934/35 season. Thus the relative symmetry of Figure 3 as compared to Figure 2 can be explained in terms of the different nature of initial film distribution in the suburban-run market, as compared to the first-run market. The over-riding objective of suburban and subsequentrun film distribution is the rapid throughput of films, rather than the prolonged exhibition of popular films. Access to this second-run market was clearly less restricted. Whereas US distributors often struggled to secure film bookings in the leading metropolitan cinemas, exhibitors in the suburbs were only too eager to access their extensive product offerings. 62 Tables 10,11 and 12 present analyses of the suburban-run market comparable to those presented in Tables 7,8 and 9 for the first-run market.
A direct comparison between Tables 7 and 10 highlights the key differences between the two exhibition sectors. Thus even though the proportions of US and British films released in the two markets are virtually identical, there are distinctive differences between the performances of individual studios and between the national affiliations of studios in the two markets. Taking national cinema first, it is clear that the very strong performance of British films in Sydney's first-run market, taking 28 of the top 100 berths, is not replicated in the suburban market, where British studios managed just eight places, with their market share shrinking by 10 percentage points to slightly less than 14%. In terms of particular distributors, the comparatively weak showing of Fox, distributor of Gaumont British and Gainsborough films, contributed over half of this reduction. This reduced performance is not likely to have been an issue of distribution, given that Fox owned not only 42% of equity in Gaumont British, but also had a substantial equity stake in the Hoyt's circuit -it seems improbable that Fox would have held back in its efforts to expose these films to suburban audiences if high levels of demand had existed. Rather, the explanation is likely to be found in differences in the nature and tastes of first-run and suburban audiences. Australian films, similarly, achieved lower market shares in the suburban market than the first-run one.
In comparison, the performance of the Hollywood studios is considerably improved in the suburban market, with their share improving by nearly 15 percentage points when compared to the first-run market. Whereas in the first-run market the median POPSTAT score for British films was higher than for American movies, in the suburbs the median US film was approximately four-and-a-half times more popular than its British equivalent. Of the films made in Hollywood, only the market shares of those films distributed by United Artists perform less well in the suburban market, with significant gains recorded by Paramount, Warner Bros. and Fox. The pattern of British films performing less well, in relative terms, as they moved from the city centre first-run cinemas to the suburban market, and of American films doing the opposite, is also clearly evident for those distributors which handled both British and American films. GAF's British films saw their market share decline by two percentage points as they moved from the city to the suburbs, while its American product saw a half point improvement in share performance. The British films of Fox witnessed a market share decline of nearly six percentage points, while its American pictures gained more than three points.
Our analysis thus far has been consistent with the proposition that Hollywood studios encountered a reduced liability of foreignness as their pictures moved from the first-run metropolitan cinemas to the suburban market. Further, it would seem that the liability of foreignness effect operated more at the level of the individual film than at the level of the distribution company. Hence Fox, as a foreign distributor, encountered little difficulty releasing British films in the first-run market, while the local firm GAF achieved more success with its American films in the suburbs than in the city. A closer examination of the figures in Tables 7 and 10, however, suggests that some US distributors may have been more successful than others in overcoming liability of foreignness in the first-run market.
The US firms which underperformed most noticeably in the first-run market (relative to the suburbs) were Paramount and Warner Bros. The difference in market share performance was around five and a half percentage points in both cases. The films of MGM and Columbia, in contrast, saw market share decline by less than a single percentage point. Columbia, of course, did not face a liability of foreignness as a distributor because its films were released by GAF. MGM's acquisition of the lease of a city centre cinema seems to have enabled it to overcome liability of foreignness in the firstrun market to a large extent. Fox's partial acquisition of the local GTC combine seemed to place it in a position in the first-run market that was stronger than Paramount or Warner Bros., but not as strong as MGM. RKO's market share performance ahead of Paramount and Warners is very likely to have been the result of the contract it signed with GTC for the exhibitor to take its product. 63 Tables 11 and 12 present formal statistical tests of country and distributor performance in the suburban market. In allowing for a direct comparison between first-run and suburban performance, these tests have three objectives: first, to verify the validity of the claim that American entertainment faced a reduced liability of foreignness as it moved from the   metropolitan centre to the suburbs; second, to analyse whether such a liability was created more by the national origin of the films or the nationality of the distributor; thirdly, to examine whether all US firms were affected equally by such a liability of foreignness effect. Thus in contrast to first-run performance, median performance levels differed widely between US and British films (although population variances were unequal). The formal Kruskal-Wallis tests appear to support the proposition that the liability of foreignness faced by US films diminished as they moved from the city centre to the suburbs -the broadly equivalent performance of US and British films in the first-run market, was transformed into markedly superior performance of US films in the suburbs. However, the unequal variances suggest that nationality of the films was not the only factor affecting performance in these two markets.
Comparing the performance of US distributors, a similar result to that in the first-run market is produced in the suburban market where we see differing performance levels between US distributors, again with an insignificant Brown-Forsythe result. Distributors of British films, on the other hand, achieved broadly consistent levels of performance in both the city and the suburbs. This conclusion of differing performance levels among distributors of American films (but not British ones) is consistent with the idea that some US firms were able to overcome liabilities of foreignness more effectively than others.
The regression results in Table 12 reinforce these conclusions with regard to country effects (the original sample size of 733 films is reduced to 460 films, the films that have an observation on both BRPOP and EVERYONES). Thus the highly significant and negative coefficient on BRITISH would imply markedly inferior performance of British films compared to US films in the suburban market, in contrast to the results in Table 9 implying that British films marginally outperformed US films in the first-run market.

Conclusion
Sydney was one of the largest cities in the world in the 1930s and the appetite of Australians for films was second to none. This study has investigated the system of film provision in the city, tracing films out from their first release dates and subsequent runs in highly capitalised city centre cinemas, through a myriad of suburban cinemas. It provides a unique insight (albeit a snapshot) into the distribution of Hollywood film to international audiences during the studio era. Three main findings emerge from the study.
First, as far as particular films were concerned, success in the first-run market was by no means a guarantee of success in the suburban one. Films that proved a hit with metropolitan audiences did not always receive a positive reception as they continued their run. Films from British and Australian producers, for example, were typically more popular in the city centre than in the suburbs. Pictures from the major US studios, on the other hand, appeared to be significantly more popular outside the metropolitan centre than within it. Thus, strong evidence is presented to support the conjecture that American distributors faced a reduced liability of foreignness once away from the cultural centre of Sydney. This finding is consistent with the observations of cultural historians looking at the reception of Hollywood films in other international markets. The Sydney datasets, however, provide valuable quantitative evidence to support an argument that has previously been presented as a qualitative assessment.
The second key finding, in keeping with existing literature on the economics of the industry in this period, is that first-run and the suburban market are distinguished in terms of the function each served and the pattern of film distribution and exhibition that each gave rise to. Whereas the first-run was concerned with drawing audiences from a wide geographic attachment area for specific film attractions which remained in exhibition for as long as a threshold audience was achieved, the suburban market attracted audiences more narrowly from specific localities to attractions that were commonly shared with lesser films and screened over three days. The first-run market was clearly vital in enabling 'hit' films to maximise their revenue potential. The suburban cinemas, however, constituted the bulk of the market, and this was most important for film distributors seeking to maximise overall revenues.
A third theme to emerge from this study is the variation between US firms in terms of their relative performance in the different market sectors. United Artists, as a specialist distributor of films by high profile 'independent' producers, was understandably atypical. Among the major vertically integrated American producer -distributors we also see interesting variations. MGM, for example, was the most successful US firm in the first-run market in terms of market share, yet it ranked fourth among the Hollywood studios in the suburban one. Paramount, on the other hand, ranked fifth in the first-run market but second in the suburbs.
Differences between US distributors are evident not just in their performance in different market sectors, but also in their strategy. Viewing US firms collectively (as Hollywood), we can see that they, and the leading Australian cinema chain (GTC), were dependent on each other to some extent, but that there was an asymmetry in the relationship. GTC required access to most (though not all) Hollywood output, though US firms did not necessarily need to go through GTC to reach the mass market. Whereas Fox tied its interests with GTC by making a financial investment in the firm, and RKO was awarded a contract by GTC to supply films to the chain, in the process securing access to its first-run cinemas, firms such as MGM and Warner Bros. were far more aggressive in challenging GTC's position. The contrasting strategies adopted by these firms allow us to scrutinise the idea that the reason for the difference in popularity of US films in city centre and suburban markets can be attributed to US distributors' restricted access to the first-run market.
The US firms which were most active in developing strategies to access the first-run market were Fox and RKO (through their relationship with GTC) and MGM (which leased its own city centre venue). The fact that these firms, in keeping with other US distributors, underperformed in the first-run market relative to the second-run one suggests that a cultural preference for Hollywood (over domestic) entertainment may well have been more evident in the suburbs than the metropolis. However, the extent of this underperformance was much more pronounced for Warner Bros. and Paramount than for the other three major US firms. Warners and Paramount both saw market share increase by around 5.5 percentage points as they moved from the city to the suburbs. The equivalent increases for Fox (3.3), RKO (1.5) and MGM (0.9) were noticeably lower. The business strategies adopted by these latter firms appear to have helped them (partially) overcome the liabilities of outsidership in the first-run market.
This article is built upon micro data encapsulating the manifold contracts made between exhibitors and distributors over the course of a long 1934 to supply city centre and suburban audiences with film entertainment. From these data, we are able to build up a framework of analysis of the market for films that is bottom-up and maintain that this snapshot can be scaled up to represent the Australian urban market and be projected for inferential purposes to the years immediately before and after 1934. In particular, this approach allows us to capture some of the dynamics evident in the rivalry (intra and inter) between exhibitors and distributors, as well as providing summary information about the statistical distribution of box office revenues and the market shares of distributors and, by association, the producers that they represent.
The ability of US firms to produce films that appealed to local, suburban audiences was perhaps the most remarkable feature of their international success. This, it would seem, was not the result of sophisticated attempts to tailor product to specific local markets around the world, but rather by encouraging the development of pictures that would appeal to local (as opposed to metropolitan) audiences in the United States itself. We end by giving the last word to Gradwell Sears, head of United Artists' domestic distribution who offered the following advice to a London-based colleague: I looked at the Ealing Production, San Demetrio, and I understand from Ed Raftery it carried a mild recommendation from you. Davey, my boy, this production is not for the American market. It is true the exploits set forth were very heroic, but they were done in such a way as to render the film completely useless as far as we are concerned. Whatever you do, don't lose your American perspective. Every time you look at one of these opus which you think might have a chance in the American market just picture yourself living at the Warwick Hotel, and showing the picture to the mob around Toots Shor's or Lindy's, or even Joe's Diner in Kansas City, Kas. I realise your viewpoint, because of your association in Canada and now in England, is not as typically American as mine, but gear your thinking with mine whenever you are considering anything for the American market.' 64 attractions. United Artists tended to distribute films from small-output producers operating in the high quality end of the market, like Samuel Goldwyn, 20 th Century Films and London Films. See Everyones, 12 December 1934. 47. It must be remembered that admission prices to cinemas were invariant with respect to popularity of the film being shown. Hence given this price, the revenue generated by any single film at a particular film was determined alone by the size of the audience. 48. The Broadway and King's Cross cinemas were located in the City Centre, but operated on the suburban distribution run, while the Empress and Empire operated as second-run cinemassee Appendix. 49. The seating data are drawn from the earliest available industry directory, published by Film Weekly for the year 1938. 50. In the case of billing status, it is not possible to discern from the table on page 2 of the Sydney Morning Herald the relative attractions of films. Rather than make a personal assessment, each film on double-bill programmes is treated as a joint attraction and assigned the billing status weight 0.5 to each. 51. Sedgwick and Pokorny 'The Film Business',[94][95]. This contrasts with United States where geographic concentration of ownership led first-run cinemas owned by the major studios to frequently screen the films of rival studios. See Pokorny 'The Film Business', 90 -93. 53. Everyones, 17 January 1934. 54. Everyones 11 April, 1934;Film Daily, 3 May 1934 and7 June 1934. 55 Glancy,'MGM',Glancy,'Warner Bros.',and Jewell,'RKO'. 62. Popular films would no doubt be shown at a larger number of suburban cinemas than relatively unpopular films. However, the frequency distribution in Figure 2 is best interpreted as reflecting the initial subsequent-run distribution of films from the first-run market, and the relative symmetry of this frequency distribution reflects the rapid turnover of films in this market rather than the relative popularity of films, which would emerge from the completed distribution of films nationwide. That is, once films have completed their nation-wide distribution we would expect a much more skewed distribution to emerge. 63. See Note 36 above. 64. (Gradwell Sears to David Coplan, 2 March, 1944