Sources: Ferranti Ltd. and Ferranti plc annual reports, 1907-87.

The development of this silicon diode provides a fascinating insight into the technology transfer process in action, because apart from securing a licence from Bell Lab's Ferranti also sent several of its semiconductor development team to work at the US pioneer's facilities. This experience proved vital in perfecting what was one of the crucial aspects of the production process, namely, fabricating the silicon crystals. At that time, electronics firms were required to make their own crystals, providing a substantial challenge given the need for purity and consistency. It is also important to note that at an early stage the Ferranti had chosen to use silicon, rather than the material that almost all the early semiconductors used, germanium. The reasons behind this move were linked to the perceived need to establish a competitive niche in a market that US firms already dominated. At the same time, as silicon was more reliable and worked at temperatures of up to 150°C, twice that of germanium, the resulting devices would be more suitable to the exacting requirements of a guided weapon. On the other hand, as semiconductor engineers were more accustomed at that time to working with germanium, the Ferranti team was faced with a considerable learning curve. It was consequently necessary to work closely with Bell Lab's engineers in perfecting the crystal-drawing process. Bell Lab's even sent its leading authority on silicon to the Ferranti laboratory, indicating how the first-mover and imitator worked intimately in producing a solution.
Even though the Ferranti semiconductor team spent considerably more than the initial budget allocated from internal funds,[43] once the silicon diode had been launched the firm established a strong presence in what was increasingly recognised as a strategic important sector. Ferranti were actually the only British producer of these devices up to the mid-1960s, providing an opportunity to charge lucrative prices to military and civil customers who had not yet been approached by the more competitive US suppliers. The internal R&D budget was consequently increased in the mid-1950s to £53,000 per annum, resulting in the production of the ZT20 Mesa transistor by 1960. As Ferranti representatives had attended the 1956 symposium organised by Bell Lab's, where the diffusion process was announced to the world, this radical new approach was being incorporated into the expanded production facilities at a disused cotton mill in Chadderton, Gem Mill. By 1960-61, capital employed at Gem Mill stood atalmost £1 million, while R&D costs since 1953 totalled £823,000, indicating how in less than a decade Ferranti had committed serious resources to this technology. Again, it is impossible to calculate how much of this had been provided by CVD, but with Bloodhound moving into production and a Mark II emerging in the laboratories, it was clear that Ferranti was strongly encouraged by the defence establishment to continue with this work. Returning to an earlier point, though, one should also stress how this support focused on fundamental research, rather than product development, indicating how Britain's leading semiconductor firm was committed to a strategy of technological leadership that would become increasingly difficult to sustain.
The difficulties associated with this strategy can be amply illustrated by looking at how in the 1960s Ferranti was obliged to commit ever-larger resources to a division that was commercially unsuccessful. Of course, the de Ferranti family was willing to sustain this commitment because of the firm's financial structure. Even though other directors and divisional managers encouraged Sebastian de Ferranti to sell off the loss-making Gem Mill activity, he argued that it was an essential feature of the firm's technology-led strategy. He was especially convinced that given the role microelectronics was beginning to play in the design and development of all types of electronic equipment, Ferranti should maintain a significant presence. The role of microelectronics had also been given a significant fillip in 1961, when Texas Instruments and Fairchild launched the IC. Ferranti had responded to this development by producing its own IC, marketed as Micronor I. This programme had been funded by the Royal Navy, given its need to miniaturise on-board computing. It also provided the firm's automation systems department with the opportunity to build on this success by developing Micronor II, a DTL IC that was much faster and more reliable than anything being produced in Europe at that time. While as we shall see later Micronor II was later superseded by Texas Instruments' much cheaper and faster 54/74 TTL IC, it again demonstrated the firm's ability to innovate in a technology that was advancing rapidly. As Golding notes: 'Measured by its ability to develop new products based on external innovations, Ferranti must surely rank as the foremost British semiconductor company. On a number of occasions ... Ferranti has demonstrated a rapid and flexible response to new technology resulting in a short lead time to commercial production'.[44]
Of course, in assessing this highly innovative record in microelectronics one should stress how reliance upon American licences was sustained throughout the 1950s and 1960s. Starting with the early experiments with silicon right through to the production of a technically successful IC, Ferranti could not avoid acquiring the basic licences for patents of firms like Bell Lab's, Texas Instruments and Fairchild. In addition, government support continued unabated, sustaining the initial stimulus provided by the CVD in 1952-52. In this respect, however, as Table 2 reveals the 1960s contrasted starkly with the 1950s. While one should remember that Ferranti provided some development funds for Dr. Shepherd and his team, the stimulus provided by both CVD and the Bloodhound guided weapon project was crucial. Similarly, the funds for producing Micronor I were provided by the Royal Navy, indicating how up to the early-1960s Ferranti were being well-supported by the defence establishment. On the other hand, as the final column in Table 2 indicates, the proportion of IC R&D funding provided by the government fell drastically after 1961. This trend was all the more disturbing to Ferranti management because of the rapid escalation in IC R&D, from just £32,000 in 1961 to £570,000 by 1971. In addition, Ferranti invested £1.2 million in new production facilities at Gem Mill, creating the capacity to produce 40,000 IC's per week. This increased the capital employed at Gem Mill to £3.4 million by 1970, compared to just £1 million at the beginning of the decade. Even though IC sales (see Table 2) did rise significantly, however, substantial losses of almost £300,000 per annum were being made on this activity by the late-1960s and early-1970s.

Table 2: Ferranti Integrated Circuit (IC) Sales and R&D, 1961-71.

Year to March	Sales £'000	Total IC R&D £'000	Sponsored IC R&D £000(& as a % of Total)
1961	-	32	16 (50)
1962	-	58	27 (47)
1963	1	78	21 (27)
1964	40	135	32 (24)
1965	85	200	58 (29)
1966	194	234	70 (30)
1967	497	308	83 (27)
1968	688	404	89 (22)
1969	756	461	111 (24)
1970	1351	461	92 (20)
1971	1680	570	103 (18)

Source: Ferranti internal accounts.

The commercial problems besetting Gem Mill after the significant technological successes associated with the silicon diode, Mesa transistor and Micronor I and II were clearly worrying for Ferranti management. It was also not difficult to see why this highly progressive division was struggling, because once US firms like Texas Instruments, Fairchild and Motorola started to market and produce their IC's in Britain competition intensified enormously. We have already noted that the 54/74 TTL IC of Texas Instruments was both cheaper and faster than Micronor II,[45] significantly eating into the potential market that Ferranti had anticipated. For example, Britain's leading mainframe computer producer, ICT (later, ICL) purchased their IC's from Texas Instruments for the 1906 range. Ferranti were also informed by potential defence customers that orders would only increase once a second IC factory had been built, thereby guaranteeing regular deliveries. While this problem was overcome in 1967 by the acquisition of a factory in Barrow, with the severe cut-backs in defence spending (especially on aerospace) in the late-1960s the market remained difficult.
It was during the late-1960s that Sebastian de Ferranti launched a major campaign to awaken British policy-makers to what he regarded as a startling admission - that Britain was losing the 'Electronics War'. Although it was clear that the Texas Instruments IC was superior to Micronor II, Sebastian argued forcefully that this competitive advantage had been achieved as a direct result of enormous governmental backing through NASA and the USAF. As we noted earlier, US corporations like Texas Instruments received substantial R&D funds, as well as guaranteed orders from the public sector, that encouraged the construction of extensive production facilities. In Britain, on the other hand, returning to a constant theme of this paper, public money was devoted to fundamental research in microelectronics, rather than trying to find a competitive niche for indigenous suppliers. It was a technology leadership strategy, rather than one based on followership. Given the enormous technical and commercial lead gained by US microelectronics firms, it is difficult to see how the British industry could attempt to compete. In this context, one must examine how successive governments responded to the demands of Sebastian de Ferranti and his contemporaries for greater public support of such a strategic industry.
The first scheme to examine is the Advanced Computer Techniques Project (ACTP) launched by the Ministry of Technology (MinTech) in 1966. Although this was principally aimed at boosting the technical proficiency of ICT,[46] IC firms were able to apply for funds. However, these funds were only provided in the form of loans, rather than grants. Furthermore, as these loans had to be repaid from a levy on sales, rather than on profits, it was apparent that the government was hardly acting as a partner in this venture. At the same time, Dr. Shepherd was grateful for this support, because up to 1970 Ferranti received almost £1.4 million, or 37% of Gem Mill's negative cash-flow. It is also important to stress that these funds were used in the successful development of a production process, collector diffusion isolation (CDI), that not even the US giants like Fairchild had been able to master. Even though Fairchild had patented the CDI process, Ferranti were the world's first microelectronics producer to apply this highly sophisticated technology to the efficient manufacture of VLSI IC's. This reflected yet again the firm's ability to compete technologically with the best microelectronics talent. It was also a peculiar form of technology transfer, because while Ferranti paid Fairchild £150,000 for the use of its CDI patents, the licensee utilised the technology much more effectively than the licensor.
Just as work on the CDI process was nearing completion, however, there was a change of government, from interventionist Labour to free-market Conservative. This new government was so keen to reduce the role of the state in business matters that initiatives and schemes like MinTech and ACTP were closed down almost immediately. Of course, this ignored what was happening in virtually every other developed and developing economy, where the state performed vital roles in stimulating the development of new technologies. One need only read what we have said earlier about the support for microelectronics provided by NASA and the USAF, not to mention how the Japanese government funded and nurtured indigenous efforts in this sector. At Ferranti, deprived of ACTP funding, yet with losses mounting on IC sales and the chairman continuing to insist that his firm would always remain in this vital sector, it was clear that a radical solution was required. Dr. Shepherd was especially aware that Ferranti would never be able to compete in price with his US rivals in the market for standard IC's, even if his team was capable of applying highly advanced technologies like CDI. At the same time, by utilising CDI for the manufacture of non-standard IC's, there was a possible niche to be exploited in serving equipment manufacturers that required custom-designed components. It was this realisation that spawned the birth of the ULA (uncommitted logic array), a customised bipolar IC capable of being adapted to each individual customer's requirements. Out of the adversity created by a sudden change in governmental attitudes, Ferranti had produced a remarkable innovation that over the following decade was to become a great commercial success.
While Sciberras has argued that in leaving the mass IC market Ferranti had adopted a highly defensive strategy,[47] in the circumstances it is difficult to see how the firm could have competed head-on against the likes of Texas Instruments. As we saw earlier, even Britain's largest electrical-electronics firm, GEC, was withdrawing from microelectronics, given the chief executive's rational aversion to risking money on technologies that were already dominated by rival suppliers. While Ferranti could compete technologically with the US firms, acquiring their licences and introducing a constant stream of innovative products, in commercial terms it was incapable of producing at the same price or marketing goods as aggressively. With Ferranti and GEC consigned to relatively minor roles in microelectronics, it is consequently no surprise to read that by 1977 85% of all UK semiconductor sales emanated from multinational subsidiaries.[48] In the Ferranti case, it was a rational response to the hiatus on government support that prompted the highly realistic strategy of searching for a niche market that could be exploited by the successful application of US technology. With losses mounting throughout the late-1960s and early-1970s, other than dispose of the factory there was very little else that management could have done.
Having stressed the role played by government policy in fashioning the Ferranti ULA strategy, it is also important to report that in 1973, as part of a much broader U-turn prompted by rising unemployment, the Department of Trade & Industry (DTI) was provided with funds to reintroduce interventionist policies. With specific reference to the industry covered in this paper, in 1973 the DTI launched the Micro-Electronics Support Scheme (MESS). Apart from bolstering the late-1960s ACTP scheme, another crucial dimension of MESS was the repayment schedule: all loan repayments were based on the profits generated by supported firms, making the state a partner in the relationship, rather than a banker. MESS was also more generous than previous schemes, in that firms were able to claim 50% of all R&D costs, boosting significantly the range and depth of work that firms like Ferranti could undertake. This proved decisive, because Ferranti were able to secure sufficient support from MESS to perfect the CDI process even further and radically improve the reliability of production. While one can only hazard a guess at whether Ferranti would have clung to the standardised IC market had MESS been introduced in 1973, it is clear that after 1973 the DTI scheme ensured that the ULA strategy would eventually become a great commercial success.

Conclusion.

Given the technological successes achieved by Ferranti up to the early-1970s, it is clear that British firms were capable of keeping up with the American pioneers like Fairchild, Texas Instruments and Motorola. Indeed, with specific regard to the CDI process, Ferranti was ahead of the firm that had originally developed this highly sophisticated IC production process, giving the firm a significant competitive advantage over its European rivals. On the other hand, Ferranti were not only forced to concentrate on the specialised chip market, bringing out its highly novel ULA concept as an alternative to a standardised product, but also little money was ever made by this operation up to the 1970s. This reflects the difficulties British microelectronics firms were having in keeping pace with an industry that was being driven by American governmental and corporate forces. Furthermore, as we have already stressed, apart from Ferranti and Plessey by the late-1970s there was only a token British presence in the European microelectronics market, American-owned operations having come to dominate the sector, especially after they had established production facilities in several overseas markets.
It is consequently apparent that British industry had failed to grasp the opportunities offered by the new microelectronics age. This view was given official sanction by the 1978 report of the Advisory Council for Applied Research and Development (ACARD), which noted that:

It was also in 1978 that the Labour government's National Enterprise Board (NEB) created and funded a British 'national champion' in the microelectronics field, Inmos. Using American technology, as well as American managers, Inmos was regarded as the most effective way of sustaining Britain's presence in the mass-production of standardised chips. It was provided with £50 million of public money, offering the prospect of a sustained existence. However, in 1984 the Thatcher government sold Inmos to Thorn-EMI, as part of its dismantling of the NEB's interventionist policies. Neither did Inmos ever achieve any of its initiators' aims, especially in terms of market share, because by that time the market for microprocessors and memory chips had also been invaded by highly competitive Japanese firms like NEC, Hitachi and Toshiba. While the Thatcher government was also responsible for starting the ALVEY programme, a scheme launched in 1984, this microelectronics support scheme can only be regarded as a case of 'too little, too late'.
Or was it? Was ALVEY simply an extension of the earlier programmes? Like many of the government-sponsored microelectronics schemes of the previous thirty-five years, ALVEY was aimed at establishing a British presence in a fundamental area of research, in this case the so-called 'fifth generation' of artificial intelligence, and catch up on American and Japanese firms in vital production techniques. Certainly, it was also of some benefit to the British firms (Ferranti, Plessey and GEC) that retained a presence in this highly competitive area. Just like its predecessors, however, it failed to create the right kind of conditions that might have stimulated a revival in British microelectronics. This failure again reflects a theme that has been sustained throughout this paper, namely, that the orientation of government-inspired programmes towards fundamental research ensured that British firms devoted most of their microelectronics resources (technical and managerial) to imitating what the American pioneers had already achieved. Admittedly, it was vital to understand the basic production processes involved, especially the diffusion and planar techniques. On the other hand, American firms like Texas Instruments and Fairchild achieved such a competitive advantage in manufacturing cheap, reliable IC's that one might regard it as folly for British licensees to try and take on these powerful corporations. In simple terms, while the US firms would always been capable of cost leadership in this sector, Porter's alternative generic strategies associated with differentiation, linked with a dedicated policy of technology followership, would have been more successful.
In discussing this issue, one inevitably must note that market factors have frequently been identified as the key reason why British firms were never capable of achieving the kind of economies of scale that American firms enjoyed up to the 1970s. At the same time, while the British market was undoubtedly much smaller than its American counterpart, there were niche markets that could have been effectively exploited. To substantiate this point, one need only refer to the Ferranti experience, because having launched the ULA in the early-1970s and struggled to create a new market for customised chips, it is a matter of record that by the early-1980s this division was the most profitable microelectronics activity in the Western world. While by the mid-1980s Japanese and American rivals had swamped this sector with their own products, undermining Ferranti profitability to such an extent that the microelectronics operation was sold to Plessey in 1988, their earlier success indicates that had a more focused approach been taken to this sector from the 1950s, then British firms could well have been more successful. Product differentiation and technology followership would have been a more effective use of public money, rather than buying American technology largely with the aim of achieving exactly what the licensees had already done. In addition, much greater effort should have been expended in creating the kind of clustering that clearly benefited American microelectronics efforts, given the enormous advantages derived from having Silicon Valley. Instead, until recently the British effort has remained highly dispersed, while only large, diversified firms were encouraged to venture into microelectronics, in contrast to the smaller, dynamically-focused operations that characterised the US industry. This reveals how public money was largely dissipated in establishing a British microelectronics industry that could never have competed against its much larger rivals in the USA and Japan.
The transfer of American microelectronics technology was consequently wasted on an excessive desire to pursue fundamental research, ignoring the alternative strategies that could well have done more for the firms that ventured into this technology in search of a radical approach towards componentry and circuit design. This failure was also part of a general British problem that was linked to successive governments' desire to maintain a strong military-political-industrial position on the world stage, a mission that distracted planners from the need to be more realistic when fashioning macro- and micro-economic strategy.