Knowledge and trust: the regulation of cooperation in industrial districts. Birmingham (UK) and Providence (USA)

The Economic History Conference
Reading 31 March - 2 April 2006

KNOWLEDGE AND TRUST: THE REGULATION OF COOPERATION IN INDUSTRIAL DISTRICTS. BIRMINGHAM (UK) and PROVIDENCE (USA)
Article and Presentation

By Francesca Carnevali
About the Author

December 2006
©2006

I. The Conceptual Problem

When in 1890 Alfred Marshall observed the persistence of localized economies, made of small firms specializing in the production of one good, such as cutlery (Sheffield), metalwares (Birmingham), furniture (High Wycombe) and pottery (North Staffordshire) he identified local external economies as the factor that allowed these firms not to internalize inputs and therefore remain small¹. Since Marshall's analysis, local economies such as industrial districts have been the object of intense study, to try to understand how they work and what regulates the relations between the small firms within these economies. Much of the literature written on this subject has identified cooperation between small firms as one of the factors that has allowed industrial districts to survive and prosper. As SMEs are strongly dependent on external resources, resource sharing arrangements can promote competitiveness. Small firms' lack of scale does not allow them to internalize market transactions and as a consequence transaction costs are potentially high. However, by cooperating in sharing, for example, machinery, the training of workers and information about customers and processes small firms can reduce these costs.

This is where the conceptual problem lies: most of the existing literature assumes cooperation as an inherent characteristic of industrial districts and treats it as unproblematic². However, there can be nothing simple about cooperation between competing agents. Cooperation requires a mutual assumption of reciprocation and the absence of free riding. Why should competing agents make such assumptions? Economic relations are plagued by problems of adverse selection and moral hazard. Conflict and opportunism are the consequences of bounded rationality and finding out who to trust is a time consuming and costly activity. These costs and hazards are one of the reasons why firms grow to absorb inputs and create governance structures within which relations are controlled internally³. These problems have been the reason why contract law has developed. Public ordering has become the way to regulate transactions outside the firm. While economists (and historians) are well aware of the extent to which human beings will go to further their material gains, most of the literature on industrial districts depicts these spaces as happy islands, untainted by opportunism, where cooperation reigns, making public ordering unnecessary. The sharing of a defined geographical space is seen as enough to generate trust spontaneously. The explanation given for the existence of trust and cooperation is that industrial districts are not just economic organizations but also social spaces in which economic actions are embedded in the social structure. Embeddedness locks together economic agents and means that moral sanctions can be applied which discourage opportunistic behaviour. According to this logic SMEs in industrial districts know that they are locked together by these social relations and this generates the trust necessary to cooperate⁴.

The concept of embeddedness is a powerful one and provides a different way of reading capitalism, a more human way than that based on the rational pursuit of utility maximization⁵. However, I am not sure that it can always be applied. Even in communities that are socially and culturally homogeneous, like those found in Italian industrial districts (those that have inspired most of the literature on industrial districts) relations between firms are not always unproblematic. The control of opportunism and conflict is exercised by local institutions such as chambers of commerce, service centers, local authorities and local banks. These are institutions that have a strong stake in the economic wellbeing of the area and therefore have a strong incentive in making firms develop competitive relations tempered by cooperation⁶. While these institutions are observable in Continental Europe they are less visible in 'Anglo-Saxon' business environments. In fact, the relative weakness or even absence in Britain of local and sectoral associations of firms providing collective services to members has been identified as one of the reasons why industrial districts are no longer a feature of the British economy⁷.

The comparison between the histories of the jewellery making districts of Birmingham in Britain and Providence in the United States, allows us to analyse in depth the nature of cooperation within localised economies and in doing so answer some important questions:

How is knowledge shared between agents who are also competing with each other?
How do small firms learn to trust each other and why is this trust important in reducing contractual hazards?
Can governance structures help set up trust based relations and reduce transaction costs by reducing the impact of opportunistic behaviour?

Governance is defined here as the establishment of mechanisms to craft order, thereby mitigating conflict and realising mutual gains for all the actors involved. Conflict and opportunism between economic agents are the consequence of bounded rationality and the reason for transaction costs. Hence governance can be said to be the establishment of mechanisms to reduce transaction costs⁸. If small firms can set up governance institutions placed between themselves and the market they will not need to grow in size - thus preserving the structure of the local economy.

The jewellers of Providence, the capital of the state of Rhode Island, on the East coast of the United States, have been defined as 'networked specialists' competing and cooperating in a well-defined area. The prosperous 1880s-90s were followed by decades of decline as the trade collapsed under the pressure of 'opportunism, design copying, interfirm suspicion and price shaving…defenseless against the flow of workers-entrepreneurs who fuelled these abuses'⁹. The paper details how these evils destroyed the craft nature of the industry and in the long run generated negative externalities, even while the industry was growing in terms of output and of the number of firms. This pattern of spiraling, ruinous competition has been explained as the consequence of 'the evils of overcompetition, derived from the jewelry sector's own structure and technical capabilities'¹⁰. Structure and technological capabilities, however, might not be the reason for Providence's decline. By looking at the Birmingham jewellery trade it is possible to see that, though the structure of the industry and its technical capabilities encouraged competition, they did not make ruin inevitable. Despite suffering from the same contractual hazards as Providence in the late nineteenth century, Birmingham's jewelers were able to create governance structures to regulate competition (without price fixing) and promote cooperation. From the late 19th century through to the troubled inter-war years and the economic recovery of the 1950s and 60s the Birmingham Jewellers' Association (BJA, later British Jewellers Association) created mechanisms to allow firms to reduce the risks associated with the fragmented structure of the industry. The activities of the Association ranged from credit checks, design theft prosecution, debt collecting, workers' training, political lobbying at a national level, to schemes for firms to share machinery. Thanks also to the activities of the BJA the district prospered, maintaining its distinctive productive structure. Though after the Second World War the number of firms declined, there was no concentration and the district maintained its distinctive productive structure. Jewellery is the Midland trade which has survived longer than most of the region's manufacturing industries¹¹.

In Providence similar attempts at associating failed. Following from this failure to establish internal forms of governance, the district started losing its structure and the craft based character of its production. While after 1945 the Birmingham district responded successfully to changes in the consumer market the Providence district, instead, went into terminal decline. Some established large firms continued to expand but outwork and start-ups continued to proliferate. By the 1960s the number of firms had decreased while the surviving firms had become bigger but trapped in a low-wage, low-quality equilibrium, under the pressure of foreign competition. By the 1970s the trade had become the symbol of the general collapse of light manufacturing in New England¹².

This paper compares the historical development of the two industrial districts to identify what factors allowed cooperation to flourish in Birmingham and not in Providence and what impact this had on the industrial structure of the two districts.

II. Birmingham and Providence

Both Birmingham and Providence started producing jewellery around the 1790s and by the mid-nineteenth century had become their countries' centers for the production of costume and low carat gold jewellery, with London and New York specialising in the upper end of the market. In both towns, mechanisation did not always lead to vertical integration but more often to the increasing specialisation of tasks carried out by small and medium-sized firms clustered in well-defined urban spaces, the Jewellery Quarter in Birmingham and the Jewelry District in Providence¹³. Many of the early settlers in Rhode Island were English and the first recorded link between the two towns was established in 1846 when at the age of 36 Thomas Lowe moved from Birmingham to Providence bringing to the town a new method of making rolled plate by 'sweating' a thin sheet of gold on base metal and then rolling the ingot into thin sheets. This method had been used by the Birmingham firm of Lutwich and Green. This was a simpler and more effective method than soldering and made the production of even cheaper jewelery possible¹⁴. Lowe was followed by other Birmingham jewelers like William Clark who in 1889 was manufacturing rings in Providence. Clark had moved to America from Birmingham with his father in the first part of the century but had originally settled in New York. Clark senior had been the first to introduce the process of burnishing in the United States and had set up as a burnisher and finisher for a large silverware manufacturer. It was probably the opportunities offered by the sheer size of the American market that led these craftsmen to move to the U.S. and take new processes with them¹⁵.

In 1860 about 75 jewelery manufacturers operated in Providence, employing 1,750 workers and making products worth $2.2 million (22% of what was produced nationally). Although the Civil War depressed the industry for two years, the increase in demand for ornaments with a patriotic, martial or funerary style fashioned in brass and other base metals revived employment. By 1870, following strong post-war recovery, auxiliary specialty services developed with an increase in the number of refiners, platers, engravers, gemstone cutters and tool producers (led by Brown and Sharpe). By 1875 there were 133 work shops in Providence employing 2667 workers. Wages were considered good with males under the age of fifteen earning 73cents a day and over fifteen $2.63. Girls earned 56cents a day and women $1.13. These wages were higher than those paid by other industries in the area, such as textiles. Some of the firms that prospered well into the twentieth century were formed in this period, such as P.A. Linton Co. (1857), Irons and Russell (1861), A. Ballou and Co. (1868) and Clark and Coombs (1872)¹⁶. By 1880 the Providence jewellery trade included 142 firms with about 3,300 workers (mostly men at this stage) and 32 companies offering specialty services and employing 300 workers¹⁷.

As in Birmingham there were few barriers to entry to the trade. Craftsmen who had served five to seven years of apprenticeship and with some design skills, a set of tools, some savings and a good reputation could easily become proprietors. Casting had been replaced by die presswork to produce brooches, cufflinks, pins and lockets, decorated with stones, wirework or enamel and fitted with clasps and chains, all in different workshops. Machinery such as presses and lathes were widely used and mostly foot operated, as in Birmingham¹⁸. In 1895 in the Providence-Attleboros area there were 350 companies, of which 205 sold finished goods (but did not necessarily make them all). About 140 sold women's jewelry, rings, bracelets, necklaces, earrings, hair ornaments and brooches. The rest specialized in goods more commonly worn by men, such as watch chains, tie pins, cuff, shirt and collar buttons and studs. A dozen firms specialized in badges for fraternal and religious organizations and for unions. Other firms specialized in pearl, shell and stone work, while others only made rings. Of the total number of firms 68 made findings and provided jewelry manufacturers with components from machine made brass, to plated chain to customized settings. In addition there were a dozen platers, 22 die-cutting shops, 6 enamelers, 11 tool and machinery specialists and 14 refiners. Jewelry manufacturers could also buy jewelry-boxes and sample cases made locally.

In 1881 one of Birmingham's trade directories listed 386 manufacturing jewelers and goldsmiths, 167 gilt jewelry manufacturers, 111 plated jewelers, 37 stone setters, 17 diamond mounters, 23 jewelers' stampers and 31 jewelers' materials manufacturers and dealers together with jewelers' blowpipe makers, box and case makers, alloy makers, draw and gauge plate makers and enamel makers. The census returns for 1881 show that the jewelery trade, including watchmakers, employed almost 20,000 people, making it the biggest employer in the region, alongside the brass and copper trade¹⁹. Although no data detailing the size distribution of these firms exists there is some evidence that suggests that most of these firms were small and that jewelery makers employing 30-40 people would have been considered large concerns. Capital and energy intensive firms such as stampers which used heavy power presses and planning machines, were larger, employing about 150 people. These figures show that at this point the structure of the two districts was similar, with small and medium sized firms specializing in one phase of production and a number of firms selling the finished product.

III. Cooperation

In Providence until the 1873 crash goods were sold to wholesalers, called 'jobbers', for cash. But after the mid-seventies jobbers were able to buy jewelry from manufacturers on credit, to sell on a consignment basis or with long delays in settling accounts. Providence manufacturers, desperate for business, accepted small orders, returns of unsold goods, cancellation of confirmed orders and the cost of huge inventories of all styles. Jobbers also showed one firm's samples to another manufacturer inviting them to duplicate them at a lower price. Copying was made easy even for the smaller firms thanks to the presence in Providence of a large number of diesinkers and 'findings' firms (makers of chains, clasps and other small ornamental parts to decorate lockets, brooches, earrings etc). Copies could be made so quickly, in less than two weeks, that the more established firms were always chasing novelties, while their stocks were devalued. Design patenting was ineffectual and established firms either agreed to lower their prices or the jobber would go the imitator's firm, usually a newer, smaller concern. The trade's extreme seasonality meant that there was always a pool of idle skilled workers ready to set-up on their own²⁰.

Birmingham suffered from very similar problems, compounded by fraud as goods were sold for gold when they were not. The collapse in the price of commodities (and of jewelery) in 1885-86 brought matters to a head and in 1887 the Birmingham Jewelers Association was created. At the same time as the BJA was created the jewelers of Providence created the New England Manufacturing Jewelers Association (NEMJA) and the Jewelers' Board of Trade. The NEMJA was not able to address the problems of price regulation, time of selling, rating of concerns, theft, litigation between bankrupt jobbers and retailers and schooling for craftsmen. The NEMJA and the Jewelers' Board of Trade were only able to establish a credit-rating system but not set common seasonal opening dates, stop design piracy, establish standard trade terms and establish uniform cost-accounting procedures. In other words the jewelers were unsuccessful in regulating competition²¹. Conversely, the Birmingham jewellers' association brought together most of the medium and large firms in the district and, crucially, it was open to manufacturers and wholesalers alike. Regulating the trade was of interest to both groups for reasons that are explored in detail elsewhere²². Instead, Providence manufacturers and jobbers organized themselves in two different organizations. Despite Providence's problems the trade grew in the 1880s. The number of firms increased to 170, the number of workers to more than 3,900 (about 20% increase). Product value rose by 43 percent, value added in manufacturing by 55 percent and workers earnings also increased. There gains were also due to technical changes (such as electroplating and gold rolled seamless wire), increasing levels of mechanization and the falling cost of materials²³. Developments in the machine tool industry, in which Providence was a national leader, also gave the costume jewelery manufacturers access to an ever increasing array of stamping and cutting machines, while the machine tool industry also provided the mechanical production basis for the expanding findings industry.

Between 1909 and 1912 the relationship between jobbers and manufacturers deteriorated again, as jobbers instead of ordering seasonal stock decided to slim their operations even further by only ordering samples. This forced manufacturers to hold inventories of their entire lines, instead of just their more successful styles. In order to get hold of the newer styles ahead of their competitors jobbers were visiting manufacturers earlier than the informal opening dates of the season 1 May and 1 December, also to get the copying of lines done. The NEMJA and the jobbers association agreed that new lines would open only on those dates but members and other promptly broke the agreement re-starting the spiral of design piracy, inventory piracy and price shaving. In 1912 S.O. Bigley placed a series of trade-journal adds in the Manufacturing Jeweler denouncing the fall in quality, devaluation of styles and skills that beset the district. The Association denounced these articles as self-promoting and did nothing²⁴.

Having to carry large stocks and depending on jobbers to sell their goods (on credit) meant that when demand slowed down, as during the economic recession of 1913, manufacturers suffered financial constraints. In 1914 thirteen Providence jewelers filed for bankruptcy and the 1913-14 crash was the worst crisis since that of the early 1890s.

From its inception until at least the First World War, the jewelry industry in Providence and in Birmingham developed along similar lines, with a similar industrial structure thanks to the existence of a network of small producers, with a few large manufacturers. The main difference seems to have been in the relationship between manufacturers and jobbers. This was adversarial, and the existence of two separate trade associations did not foster cooperation. Birmingham jewelers and wholesalers instead were able to associate together and create mechanisms of enforcement and incentive that were effective in regulating competition and avoid the excesses that ravaged Providence. Contingency and chance might have played a part in explaining why cooperation did not emerge in Providence. However, we should also consider one crucial difference. Jobbers were hardly ever local as most of them operated from New York²⁵ while most of the wholesalers that belonged to the BJA were Birmingham firms who acted as factors for the smaller manufacturing firms. Jobbers could be portrayed as predatory outsiders while wholesalers as stakeholders with a strong long-term incentive to keep the local economy prosperous²⁶.

The other difference between the two communities was that as a consequence of the lack of cooperation between manufacturers and jobbers, the larger firms started selling directly to retailers. Smaller firms did not have the internal resources necessary to do this and therefore had to adapt their production to the model imposed by the jobbers' demands. According to Scranton the 'market struggles of the 1880s divided the industry into groups of firms either selling direct or dependent on jobbers, groups whose interests were opposed on most trade practices'²⁷. The larger firms hired roadmen and the rising department stores provided them with the first outlets, as these bought in large quantities and paid promptly. Some companies printed catalogues to be sent to the larger independent jewelers and to national fraternal, business, religious and sporting associations. Direct selling was complex and costly, especially when it came to collect payments from smaller stores and although the threat of direct selling was useful in stopping the jobbers worst excesses it was not a long-term solution, especially for the smaller manufacturers. Different strategies created different interests and made unlikely the possibility of creating an association that could act as a governance structure. This can be seen also in the NEMJA failure to organize and mobilize the jewelers in establishing a training school to create craftsmen.

IV. Knowledge

The original apprenticeship system, lasting 5-7 years, prepared skilled craftsmen who could take on any job. By the end of the 19th century this system had been replaced by vocational training that created specialized pressmen, solderers, polishers, stonesetters and so on. Machinery made such specialization much easier. In 1919 the Jewelers' Circular Weekly reckoned that of the more than 10,000 people employed as jewelry workers only a very small portion would have been able to produce a single finished article²⁸. The training of skilled craftsmen was a costly and time consuming activity and one of those activities that was subject to free riding, as firms could 'poach' workers without having taken on any of the costs of training them. Such a hazard was particularly harmful to the smaller firms. In such a situation the incentive for firms was to move to a system of production based on semi-skilled and unskilled workers, operating simple machinery. The scarcity of skilled workers was compounded by the fact that those with skills could easily find employment with the large firms in the region who offered better working conditions and constant employment, as opposed to the seasonal work provided by most manufacturing jewelers²⁹.

The scarcity of skilled workers meant that outwork (done in homes by women and children) increased leading to progressive deskilling. By 1906 some firms were moving small machinery, such as foot-powered die presses to the homes of married women who also did more routine piecework, like assembling pendants and earrings, or attaching pieces to cards. Free riding could have been minimized by making training a collective good by pooling knowledge and resources in a school. As happened in Birmingham.

The scarcity of skilled workers was one of the problems that affected Birmingham as well, for very similar reasons. At the first AGM of the BJA a number of rules were approved. The first one stated that one of the objects of the Association was: "The advancement of taste in the manufacture of jewellery and personal ornaments by developing the art education of employers and employed and by instructing apprentices and young persons in the true principles of decorative and constructive art"³⁰. As a consequence of this resolution, in 1890 the Birmingham School of Jewellers and Silversmiths was created. The school had no equivalent in the country and its promotion of technical and artistic standards ensured the steady supply of skilled workers needed by the trade.

In Providence, although there was discussion in the trade about creating a school, or linking with the Rhode Island School of Design, the NEMJA was not able to mobilise enough interest and money. During the prosperous years of the early 1900s the Association was able to get the jewelers in the area to donate to the School only small amounts of money³¹.

In Providence firms' decision to shape production around outwork and semiskilled workers would, in the long run, lead to deskilling and to the rise of negative externalities. It made, however, perfect sense in the context of the changes that were affecting the city, mostly in terms of its population. The growth of the population of Providence matched its industrial expansion. Between 1865 and 1925 the number of Providence residents increased almost five times, from 54,595 to 267,918 and only a small part of this growth was due to natural increase. Most came from in-migration from other towns and states and immigration from abroad. Over that period the source of immigration changed significantly. In 1800 nine tenths of the immigrant fathers of Providence school children came from British Isles and of these the larges proportion was Irish. By 1900 only half did. The source of the new immigrants was Southern and Eastern Europe, principally from Italy and to a lesser extent, from Russia, Poland, Armenia and the Azores. When in 1911 the Fabre Line started passenger service between Naples, Palermo, Lisbon, the Azores and Providence Italian and Portuguese women and men flooded the city and the state. By 1915, Italian immigrants constituted fourteen percent of the city's population. The textiles mills were the first source of employment as their high level of mechanization required un-skilled and semi-skilled workers. After the turn of the century many of these workers became employed in the jewelry industry as the manufacture of costume jewelery became increasingly mechanized. Mechanisation created a labour force stratified by function and by skill. At the top were precision toolmakers, die cutters and mold makers who made the machines, dies and molds that produced the jewelry components. Also at the top were the senior platers who regulated the chemical process involved in electroplating components. At the next level were journeymen platers, polishers and tool setters. Below them were the bench workers who made up the majority of the work force. Bench workers soldered findings, linked chains and generally assembled the machine cut or cast pieces into finished jewellery. Some bench jobs required more skills than others such as setting stones and enameling. Because of the low level of skills required bench workers were frequently women and children and their earnings were among the lowest in the manufacturing sector. In any case mill work and jewelry work offered women employment that although poorly paid was flexible and as a result a large percentage of Providence's work force was female. By 1910 one third of the women over sixteen in Providence were gainfully employed. By 1905 women accounted for thirty-five percent of jewelry manufacturing employees and for fifty-four percent by 1930. Conversely the manufacture of precious metal goods retained an almost exclusively male, native or Northern European workforce, skilled and well paid. The 1905 Rhode Island census listed 1,149 gold and silver workers in Providence, of whom all but forty-eight were men and all but twenty-two were born in North America of northern Europe. These were skilled workers whose wages were on average thirty-six percent higher than those earned by Providence's costume jewelry workers³². Birmingham instead experienced a migration wave earlier, in the second part of the nineteenth century and it attracted mostly skilled workers from Northern Europe. Therefore the pool of workers available to the jewellery industry was different than in Providence. Although outwork and women's labour were used by the trade in Birmingham as well, the absence of a large pool of unskilled workers meant that Birmingham jewellery manufacturers had fewer incentives to substitute machinery for labour, and dilute the craft nature of the district.

The Birmingham jewellers, thanks to the BJA, developed cooperation in areas such as artistic and technical training for workers and the development of machinery (while continuing to pursue the application of 'rules of the game' in commercial relations). While in Providence the lack of cooperation in the sharing of knowledge led to deskilling and impacted on the structure of the district while at the same time some firms (which then became the leaders in the post-1945 period) grew in size to internalise these processes.

V. CONCLUSION

By comparing Providence and Birmingham it is clear that it was not the structure of the jewellery industry that led to the evils of overcompetition. The comparison also shows that such evils could be controlled through cooperation, in the case of Birmingham through the creation of a formal governance structure. Providence's jewelers could not find a structre of incentives that would unite them. The reasons for this could be many, however, comparison with Birmingham suggests looking at the composition of the population of firms, at the distribution of small and large firms and at where manufacturers and wholesalers/jobbers were located.

What effect did all these problems really have on the trade? In Providence the number of firms continued to grow, production did not collapse although many firms continued to seek short-term gains through the use of outwork and design copying but the industry did not disappear, also because there was no other American center that possessed Providence's advantages for disaggregated production. The United States census for 1910 reported that the three Rhode Island industries of jewelry, silverware and platedware employed 11,887 people and turned out products valued at $31,325,000 leading all other States in the manufacture of jewelry. The Federal statistics for 1910 show that there were 296 jewelry establishments employing 9,511 people of whom 56.8 percent were male and 40.6 percent female, and 2.6 percent were children below the age of sixteen. The total value of the product was $20,685,100 making it the fifth largest industry in the State. Of the 296 establishments returned, 58 produced less than $5,000 each with a total of 234 people employed. There were 66 firms that produced between $5,000 and $20,000 each with 602 people; 108 firms that produced between $20,000 and $100,000 each and employed 3,225 people and 64 concerns each doing a business of between $100,000 and $1,000,000 employing 5,450 people and producing an aggregate of $14,516,375 or two-thirds of the entire production of the industry of Rhode Island. Silverware gave employment to 2,294 in 11 establishments whose product amounted to $6,198,000; 15 gold and silver refiners employed 82 people with product valued at $4,442,000; 18 brass and bronze firms with 225 employees and $828,000 of product and 21 enameling and japanning firms with 615 employees and $570,000 of product³³. By the time World War One ended the industry had become concentrated in an area bounded by four streets, Pine, Chestnut, Dorrance and Eddy, an area know as the Jewelry District, within which about 14,000 people worked for about three hundred manufacturers and jobbers. For the most part the industry was made of concerns employing no more than thirty to forty people. Among these manufacturers were a few big concerns such as the manufacturers Ostby and Barton ( located at 118 Richmond St.) who employed close to five hundred workers during rush periods.

'Overcompetition' did drive down prices, as novelties were imitated and made more cheaply by the smaller enterprises, using less skilled labour. Phil Scranton is critical of this process as it meant that fashion was recoded as self-destructing, seasonal commodities³⁴. This criticism, however, leaves out the part played by consumers' demand. This was also driving the trade, as novelties, by definition have to be new. Providence's productive structure allowed for extreme flexibility, adapting to constant changes in demand. The public and therefore retailers and dealers were constantly searching for something new and original and each Providence manufacturer believed that they could exploit this. Turning out new things, even copying them with a new twist was what Providence jewelry was about. Although we can see where deskilling would eventually lead the industry, a comparison of the styles made in Providence and in Birmingham between the 1880s and the 1930s shows that novelty, innovation and variety were Providence's hallmark, not Birmingham's³⁵. Despite its many problems Providence's jewellery trade continued to grow and it is only through historical narrative and comparison that we can identify its weaknesses.

Presentation

The conceptual problem that I am interested in is how and why small firms decide to cooperate.

As SMEs are strongly dependent on external resources, resource sharing arrangements can promote competitiveness. Small firms' lack of scale does not allow them to internalize market transactions and as a consequence transaction costs are potentially high. However, by cooperating in sharing, for example, machinery, the training of workers and information about customers and processes small firms can reduce these costs.

Much of the literature on industrial districts has taken cooperation between the firms within the district for granted.

However, there can be nothing simple about cooperation between competing agents. Cooperation requires a mutual assumption of reciprocation and the absence of free riding. Why should competing agents make such assumptions? Economic relations are plagued by problems of adverse selection and moral hazard. Conflict and opportunism are the consequences of bounded rationality and finding out who to trust is a time consuming and costly activity. These costs and hazards are one of the reasons why firms grow to absorb inputs and create governance structures within which relations are controlled internally³⁶.

These problems have been the reason why contract law has developed. Public ordering has become the way to regulate transactions outside the firm. While economists (and historians) are well aware of the extent to which human beings will go to further their material gains, most of the literature on industrial districts depicts these spaces as happy islands, untainted by opportunism, where cooperation reigns, making public ordering unnecessary. The sharing of a defined geographical space is seen as enough to generate trust spontaneously. The explanation given for the existence of trust and cooperation is that industrial districts are not just economic organizations but also social spaces in which economic actions are embedded in the social structure. Embeddedness locks together economic agents and means that moral sanctions can be applied which discourage opportunistic behaviour. According to this logic SMEs in industrial districts know that they are locked together by these social relations and this generates the trust necessary to cooperate³⁷.

The concept of embeddedness is a powerful one and provides a different way of reading capitalism, a more human way than that based on the rational pursuit of utility maximization³⁸. However, I am not sure that it can always be applied. Even in communities that are socially and culturally homogeneous, like those found in Italian industrial districts (those that have inspired most of the literature on industrial districts) relations between firms are not always unproblematic. The control of opportunism and conflict is exercised by local institutions such as chambers of commerce, service centers, local authorities and local banks. These are institutions that have a strong stake in the economic wellbeing of the area and therefore have a strong incentive in making firms develop competitive relations tempered by cooperation³⁹. While these institutions are observable in Continental Europe they are less visible in 'Anglo-Saxon' business environments. In fact, the relative weakness or even absence in Britain of local and sectoral associations of firms providing collective services to members has been identified as one of the reasons why industrial districts are no longer a feature of the British economy⁴⁰.

How is knowledge shared between agents who are also competing with each other?
How do small firms learn to trust each other and why is this trust important in reducing contractual hazards?
Can governance structures help set up trust based relations and reduce transaction costs by reducing the impact of opportunistic behaviour?

Governance is defined here as the establishment of mechanisms to craft order, thereby mitigating conflict and realising mutual gains for all the actors involved. Conflict and opportunism between economic agents are the consequence of bounded rationality and the reason for transaction costs. Hence governance can be said to be the establishment of mechanisms to reduce transaction costs⁴¹. If small firms can set up governance institutions placed between themselves and the market they will not need to grow in size - thus preserving the structure of the local economy.

The jewellers of Providence, the capital of the state of Rhode Island, on the East coast of the United States, have been defined as 'networked specialists' competing and cooperating in a well-defined area. The prosperous 1880s-90s were followed by decades of decline as the trade collapsed under the pressure of 'opportunism, design copying, interfirm suspicion and price shaving…defenseless against the flow of workers-entrepreneurs who fuelled these abuses'⁴². The paper details how these evils destroyed the craft nature of the industry and in the long run generated negative externalities, even while the industry was growing in terms of output and of the number of firms. This pattern of spiraling, ruinous competition has been explained as the consequence of 'the evils of overcompetition, derived from the jewelry sector's own structure and technical capabilities'⁴³. Structure and technological capabilities, however, might not be the reason for Providence's decline. By looking at the Birmingham jewellery trade it is possible to see that, though the structure of the industry and its technical capabilities encouraged competition, they did not make ruin inevitable. Despite suffering from the same contractual hazards as Providence in the late nineteenth century, Birmingham's jewelers were able to create governance structures to regulate competition (without price fixing) and promote cooperation. From the late 19th century through to the troubled inter-war years and the economic recovery of the 1950s and 60s the Birmingham Jewellers' Association (BJA, later British Jewellers Association) created mechanisms to allow firms to reduce the risks associated with the fragmented structure of the industry. The activities of the Association ranged from credit checks, design theft prosecution, debt collecting, workers' training, political lobbying at a national level, to schemes for firms to share machinery. Thanks also to the activities of the BJA the district prospered, maintaining its distinctive productive structure. Though after the Second World War the number of firms declined, there was no concentration and the district maintained its distinctive productive structure. Jewellery is the Midland trade which has survived longer than most of the region's manufacturing industries⁴⁴.

By tracing the history of the two places from the early 19th century one can see that the jewelery industry developed in a similar way, and that production was organized within an industrial district, with small and medium sized firms specializing in one phase of production and a number of firms selling the finished product.

The comparison of the response of the two communities to crisis, such as the one that hit the districts at the end of the 19th century shows:

that proximity doesn't always generate cooperation
that we can understand why competing agents chose to cooperate by analysing the problem in terms of incentives

The Birmingham jewelers were able to create a governance structure to regulate competition, while the Providence jewelers created the NEMJA and this was ineffective as a regulatory body.

BJA: manufacturers and wholesalers

NEMJA: only manufacturers

The main difference seems to have been in the relationship between manufacturers and jobbers. This was adversarial, and the existence of two separate trade associations did not foster cooperation. Birmingham jewelers and wholesalers instead were able to associate together and create mechanisms of enforcement and incentive that were effective in regulating competition and avoid the excesses that ravaged Providence. Contingency and chance might have played a part in explaining why cooperation did not emerge in Providence. However, we should also consider one crucial difference. Jobbers were hardly ever local as most of them operated from New York⁴⁶ while most of the wholesalers that belonged to the BJA were Birmingham firms who acted as factors for the smaller manufacturing firms. Jobbers could be portrayed as predatory outsiders while wholesalers as stakeholders with a strong long-term incentive to keep the local economy prosperous⁴⁷.

Different strategies created different interests and made unlikely the possibility of creating an association that could act as a governance structure. This can be seen also in the NEMJA failure to organize and mobilize the jewelers in establishing a training school to create craftsmen.

IV. Knowledge

The training of skilled craftsmen was a costly and time consuming activity and one of those activities that was subject to free riding, as firms could 'poach' workers without having taken on any of the costs of training them. Such a hazard was particularly harmful to the smaller firms. In such a situation the incentive for firms was to move to a system of production based on semi-skilled and unskilled workers, operating simple machinery.

The scarcity of skilled workers meant that outwork (done in homes by women and children) increased leading to progressive deskilling.

Free riding could have been minimized by making training a collective good by pooling knowledge and resources in a school. As happened in Birmingham.

At the first AGM of the BJA a number of rules were approved. The first one stated that one of the objects of the Association was: "The advancement of taste in the manufacture of jewellery and personal ornaments by developing the art education of employers and employed and by instructing apprentices and young persons in the true principles of decorative and constructive art"⁴⁸. As a consequence of this resolution, in 1890 the Birmingham School of Jewellers and Silversmiths was created. The school had no equivalent in the country and its promotion of technical and artistic standards ensured the steady supply of skilled workers needed by the trade.

Birmingham instead experienced a migration wave earlier, in the second part of the nineteenth century and it attracted mostly skilled workers from Northern Europe. Therefore the pool of workers available to the jewellery industry was different than in Providence. Although outwork and women's labour were used by the trade in Birmingham as well, the absence of a large pool of unskilled workers meant that Birmingham jewellery manufacturers had fewer incentives to substitute machinery for labour, and dilute the craft nature of the district.

V. CONCLUSION

¹A. MARSHALL, Principles of Economics, Basinstoke, 1952:1890
²G. BECCATTINI, 'Sectors and/or districts: some remarks on the conceptual foundations of industrial economics', in E. GOODMAN (ed.), Small firms and industrial districts in Italy, London, 1989
³ O. WILLIAMSON, The economic institutions of capitalism, New York, 1985.
⁴G. DEI OTTATI, 'Trust, interlinking transactions and credit in industrial districts', Cambridge Journal of Economics, vol. 18, no. 6, 1994, pp.529-546.
⁵M. GRANOVETTER, 'Economic action and social structure: The problem of embeddedness', American Journal of Sociology, 91, 1985, pp.481-510.
⁶For a recent review of the relevant literature see: G. BECCATTINI, M. BELLANDI, G. DEI OTTATI, F. SFORZI(eds.), From industrial districts to local development, Cheltenham, 2003
⁷J. ZEITLIN, 'Why are there no industrial districts in the UK?', in A. BAGNASCO, C. SABEL (eds.), Small and medium-sized enterprises, London, 1995
⁸Though this definition of governance is normally applied to the large firm it is possible to apply it to other institutions, such as trade associations. This definition of governance derives from the literature on new institutional economics. This literature does not restrict its analysis to the firm; other economic institutions considered are corporate governance, regulation, antitrust and the public bureau. Nor does Williamson think that this list includes all possible forms of governance. WILLIAMSON, The economic institutions of capitalism, p.385 and WILLIAMSON, 'The New Institutional Economics', Journal of economic literature, 38, 2000 pp.603-604.
⁹P. SCRANTON, Endless novelty. Specialty production and American industrialization 1865-1925, Princeton, 1997, p. 326.
¹⁰SCANTON, Endless novelty, p.244.
¹¹The Birmingham jewellery district is still alive today and it maintains the same productive structure characteristic of the nineteenth century industrial district, producing almost half of the gold jewellery sold in the country. For an indepth analysis of the history of the Birmingham jewellery trade see: F. CARNEVALI, 'Golden opportunities: jewelry making in Birmingham between mass production and specialty', Enterprise and Society, 4, 2003, pp.272-298; F. CARNEVALI, '"Crooks, thieves and receivers"': transaction costs in nineteenth-century industrial Birmingham', Economic History Review, LVII, 3, 2004, pp.533-550.
¹²D. GAGGIO, 'Technologies of trust. The politics of small-scale industrialization in the jewelry districts of Providence and Northern Italy, 2002, unpublished paper; L.E. BROWNE, S. SASS, 'The transition from amill-based to a knowledge-based economy: New England 1940-2000', in P. TEMIN, Engines of Enterprise, Harvard, 2000
¹³J. CATELL, S. ELY, B. JONES, The Birmingham Jewellery Quarter, Swindon, 2002; G. FRANKOVICH, 'History of the Rhode Island jewelry and silverware industry', Rhode Island Yearbook, 1971
¹⁴FRANKOVICH, 'History of the Rhode Island', p. 84 and The Jewelers' Circular Weekly ¹⁵J.D. HALL, Biographical History of the Manufacturers and Businessmen of Rhode Island, Providence, 1901, p. 110.
¹⁶FRANKOVICH, 'History of the Rhode Island', p.84
¹⁷P. SCRANTON, 'The horrors of competition: Innovation and Paradox in Rhode Island's jewelry industry 1860-1914', Rhode Island History, 55, 2, p.47
¹⁸SCRANTON, 'The horrors of competition', p.48
¹⁹G.C. ALLEN., The Industrial Development of Birmingham and the Black Country, 1860-1927, London, 1929, Table 5, p.459.
²⁰SCRANTON, 'The horrors of competition', pp.48-50.
²¹SCRANTON, 'The horrors of competition', pp. 51-52.
²²CARNEVALI, 'Crooks, thieves and receivers'
²³SCRANTON, 'The horrors of competition', p.52
²⁴SCRANTON, 'The horrors of competition' p.64
²⁵U.G. DIETZ, The glitter and the gold. Fashioning America's jewelry, Newark, 1997
²⁶New York jobbers could also go elsewhere to buy jewelry, while Birmingham's wholesalers could not, as the only other centre of cheap jewellery manufacture was London, firmly in the hands of the London wholesalers.
²⁷SCRANTON, 'The horrors of competition', pp.51-52
²⁸Jewelers' Circular Weekly, p.285
²⁹It is important to remember that jewelry making was not the only large employer in the state and there were other large firms in the area, such as the Corliss Steam Engine Company who employed more than 1000 workers in 1888 at the turn of the century and was the biggest enterprise of its kind in the world; Brown and Sharpe, makers of machine tools and employing more than 2,000 people in 1902; Nicholson File Company, largest produce of files in the country and the American Screw Company also a national leader, and Gorham which in 1890 had the largest silverware factory in the world and in 1907 employed almost 6,000 workers. The textile industry employed 22,600 in 1875 in Rhode Island and 25,000 in 1895. Although Providence had only 25 percent of the state's total textile industry it had some very large mills such as the giant Riverside, Providence and National Worsted and the Weybosset and Atlantic Mills.
³⁰BJA, The Birmingham Jewellers and Silversmiths Association, Rules, Booklet, MS 1646/1.
³¹Collectively the Providence/Attleboro jewelers donated only $4,000 against the $5,000 donated by Brown and Sharpe and $2,000 by Gorham, SCRANTON, 'The horrors of competition', note 40,
³²R.A. MECKLE, 'The Jewelry Industry. Industrial Development and Immigration in Providence, 1790-1993', in N.D. WEISBERG, Diamonds are Forever, but Rhinestones are for Everyone. An Oral History of the Costume Jewelry Industry of Rhode Island, Providence, 1999, pp. 2-6
³³The Jewelers' Circular Weekly, February 5, 1919, pp.283-84.
³⁴SCRANTON, 'The horrors of competition', p.56
³⁵R. ETTINGER, Popular jewelry. 1840-1940, Atglen, 2002; S. BURY, Jewellery 1789-1910: The international era, Woodbridge, 1997
³⁶O. WILLIAMSON, The economic institutions of capitalism, New York, 1985.
³⁷G. DEI OTTATI, 'Trust, interlinking transactions and credit in industrial districts', Cambridge Journal of Economics, vol. 18, no. 6, 1994, pp.529-546.
³⁸M. GRANOVETTER, 'Economic action and social structure: The problem of embeddedness', American Journal of Sociology, 91, 1985, pp.481-510.
³⁹For a recent review of the relevant literature see: G. BECCATTINI, M. BELLANDI, G. DEI OTTATI, F. SFORZI(eds.), From industrial districts to local development, Cheltenham, 2003
⁴⁰J. ZEITLIN, 'Why are there no industrial districts in the UK?', in A. BAGNASCO, C. SABEL (eds.), Small and medium-sized enterprises, London, 1995
⁴¹Though this definition of governance is normally applied to the large firm it is possible to apply it to other institutions, such as trade associations. This definition of governance derives from the literature on new institutional economics. This literature does not restrict its analysis to the firm; other economic institutions considered are corporate governance, regulation, antitrust and the public bureau. Nor does Williamson think that this list includes all possible forms of governance. WILLIAMSON, The economic institutions of capitalism, p.385 and WILLIAMSON, 'The New Institutional Economics', Journal of economic literature, 38, 2000 pp.603-604.
⁴²P. SCRANTON, Endless novelty. Specialty production and American industrialization 1865-1925, Princeton, 1997, p. 326.
⁴³SCANTON, Endless novelty ⁴⁴The Birmingham jewellery district is still alive today and it maintains the same productive structure characteristic of the nineteenth century industrial district, producing almost half of the gold jewellery sold in the country. For an in depth analysis of the history of the Birmingham jewellery trade see: F. CARNEVALI, 'Golden opportunities: jewelry making in Birmingham between mass production and specialty', Enterprise and Society, 4, 2003, pp.272-298; F. CARNEVALI, '"Crooks, thieves and receivers"': transaction costs in nineteenth-century industrial Birmingham', Economic History Review, LVII, 3, 2004, pp.533-550.
⁴⁵D. GAGGIO, 'Technologies of trust. The politics of small-scale industrialization in the jewelry districts of Providence and Northern Italy, 2002, unpublished paper; L.E. BROWNE, S. SASS, 'The transition from amill-based to a knowledge-based economy: New England 1940-2000', in P. TEMIN, Engines of Enterprise, Harvard, 2000
⁴⁶U.G. DIETZ, The glitter and the gold. Fashioning America's jewelry, Newark, 1997
⁴⁷New York jobbers could also go elsewhere to buy jewelry, while Birmingham's wholesalers could not, as the only other centre of cheap jewellery manufacture was London, firmly in the hands of the London wholesalers.
⁴⁸BJA, The Birmingham Jewellers and Silversmiths Association, Rules, Booklet, MS 1646/1.
⁴⁹Collectively the Providence/Attleboro jewelers donated only $4,000 against the $5,000 donated by Brown and Sharpe and $2,000 by Gorham, SCRANTON, 'The horrors of competition', note 40,