The aim of this working paper series is to disseminate widely papers and short reports produced by the European Commission’s FAIR (Forecast and Assessment of Socio-Economic Impact of Advanced Communications and Recommendations) project. Through this series we intend to make a variety of types of material available ranging from discussion documents to analytical reports and papers. These are intended to stimulate discussion on the important social, economic, political and technological factors which are affecting businesses and people in the transition towards an ‘Information Society’.

FAIR is a horizontal action project in the Advanced Communication Technologies and Services (ACTS) programme of the European Commission. Work began in September 1995 and is expected to continue throughout the ACTS Programme. The project is led by Databank Consulting, Milan and collaborators include Technology Investment Partners, Paris, and the Graduate Research Centre in Culture and Communication, University of Sussex.

This contribution to the working paper series is a joint initiative of the SPRU (Science Policy Research Unit) Centre for Information and Communication Technologies, University of Sussex in the United Kingdom and MERIT (Maastricht Economic Research Institute on Innovation and Technology), University of Limburg in the Netherlands.

SPRU and MERIT are responsible for socio-political and techno-economic analyses of the impact of developments in advanced communication technologies and services.

The views expressed in these papers are those of the authors and do not necessarily reflect those of any institution or organisation.

Additional copies of this paper and others in the series can be obtained by contacting:

The SPRU Publications Office

University of Sussex

Falmer, Brighton BN1 9RF

Sussex, UK

Tel: +44 (0)1273 678176

Fax: +44 (0)1273 685865

While there is increasing emphasis on examining the institutional context in which ICTs are applied and used, most writers still regard these technologies as essentially a variant form of process automation whose application will lead to corresponding reductions in cost. Underlying this perspective is an implicit view of commercial organisations as ‘information processors’ (IPs) where knowledge and information can be given commodity status, subject to the same economic laws of production and exchange as are other physical objects. Such an approach permits information to be treated as simply an undifferentiated input into the production process subject only to certain asymmetries that increase the processing costs involved. Consequently, issues concerned with technological change are largely restricted to examining how quickly and effectively ICTs are applied. In this regard, information and communication technologies (ICTs) are seen "to substantially cheapen the cost of handling information as a universal commodity by permitting it to be processed, stored, transmitted and marketed in its electronic form" .

The ‘information processing’ (IP) model sees firms and their role in the economy in terms of the economies of scope and scale which they enjoy in the processing of information. In itself, the model represents an important step forward. It overcomes the difficulties of some earlier approaches where these issues were either effectively sidelined or considered unimportant. In particular, the problems which arise in assuming that actors have access to ‘perfect’ information at all times are avoided. By highlighting the importance of informational issues it focuses attention on the potential importance of ICTs in transforming some of the basic conditions in which economic relations will be taking place. In particular, the IP model highlights the value of examining how firms will be able to adapt to the technological opportunities which rapid advances in data communication, data processing and data storage are able to engender. In this way, the IP model provides a common basis from which to assess technological change. Innovations in ICTs can all be considered from the same perspective, namely in terms of resulting reductions in cost in the handling of ‘information’. Since the services sector and specifically commercial banking is identified as being particularly ‘information’ intensive, it provides a clearly delineated framework within which their impact can be assessed in these particular areas.

However, a model which views information as a ubiquitous commodity that already exists, suffers from inherent deficiencies. The model starts to lose some of its interpretative power once an attempt is made to develop a richer picture of the relationship between ICTs and ‘information’. In particular, the model becomes restricted because issues concerned with how ‘information’ comes into existence are not specifically addressed. Once ‘information’ is regarded as the result of a socially constructed process, the role which ICTs play in mediating its production and exchange in organisations no longer enjoys the same simplicity. ICTs then have to be considered in terms of their specific impact on the information production and exchange processes (IPEPs) themselves. Their role can no longer be restricted to bringing about a progressive cost reduction which serves to cheapen ‘information’ as an input into either economic or social processes.

A recent DPhil research project completed by the author (Crede 1997) examined the use of ICTs by commercial banks and to what extent it fits the IP model. It confirmed the considerable difficulties that arise when ICTs are considered without any attempt being made to examine the nature of the information and knowledge production itself. In particular, the IP model was found to be deficient when attempting to explain the high degree of selectivity which can be observed with respect to the implementation and use of ICTs within the specific institutional setting of commercial banks. For example, the study found that while commercial banks have made very large investments in computer systems and networks which facilitate intra-organisational communication the usage of these networks and systems to automate the manipulation of data was significantly constrained. Confidential customer data although in principle available electronically, was strictly controlled with availability dependent on ‘negotiated access’ taking place at the inter-personal level.

An IP model faces difficulties in generating insights to explain these apparent anomalies and where investment in ICTs needs to be justified in terms of a lowering the commodity cost of ‘information’. Since ‘information’ already exists ready for processing, ICTs are primarily considered on the basis of achieving reductions in the cost of ‘information’ as a commodity. Variations in how ICTs are applied and used have to be explained in terms of ‘non-rational’ or unexpected institutional or organisational factors. In the longer term, an IP based model would therefore expect resistance to the smooth adoption of these cost reducing technologies to be overcome, particularly when competitive forces apply. It is therefore presented with a problem of interpretation when considering the commercial banking sector where ICTs have been used extensively for more than 30 years.

One feature emerging from the IP model is that ICTs are seen as intrinsically transformative in their effect on economic institutions. By viewing certain organisations such as commercial banks specifically as IPs, ICTs are seen to transform both institutions and processes, as the costs of producing information are substantially reduced. It is assumed that simply because information can be communicated more quickly and processed more rapidly, this will automatically create corresponding institutional changes. Thus, by dramatically reducing the processing cost of information, ICTs are seen to bring about a redefinition of existing institutions, dominating any other considerations which may apply. Furthermore, through ‘digitisation’, information is seen to lose its intangible qualities, thereby removing any asymmetries which may have been associated with localised knowledge.

Digitisation is seen to permit codified information to become more readily manipulated as advances in ICTs produce ever cheaper storage, transmission and processing. Eventually, this allows information to become a fully traded and marketed electronic commodity in its own right. For example, Barras considers that "…the progression from manufacturing to information economy lifts information from merely being an essential factor of production in the economy into being its most important product… the essential contribution of information technology is both to cheapen substantially the cost of producing information as a universal commodity, and also to provide a means by which this previously largely intangible commodity can be processed, stored, transmitted and marketed in electronic form as a database or software program" .

The IP model leads to a perspective where, providing a suitable electronic communication infrastructure is implemented, new information products can move from process to product innovation based on the sophisticated networks which become available. Implicit is the notion that the information is already there and all that is required is to deliver it to the customer at the lowest possible cost. This working paper argues that the relationship between ICTs and of information is more complex. For example, has specifically criticised such an approach for making telecommunication into the new 'black box' of economic analysis. It ignores the dynamic relationship which is increasingly apparent between the institutional forms which deliver this technology and the forms of network which emerge. Implementation of ICTs involves selections and choices. The move to network communications then becomes a product of how the bundle of technological innovations incorporated into communication technologies is deployed within a specific set of institutional arrangements. Such an approach may be more helpful in explaining the distinctive features of, for example, ATM networks, and the process by which they have evolved rather than assuming like (Barras 1990) that there is a linear relationship between what is referred to as the ICT/telecommunication infrastructure and specific information products.

In order to move beyond these limitations, this working paper argues for the need to complement the IP model with a theoretical approach which examines the nature of information production and exchange. This places emphasis on the observation that individuals have first to create information which in turn is the product of codification, abstraction and diffusion of otherwise undifferentiated sensory data. The production and exchange of information becomes particularly important when one considers the limitations which apply to the ability of human beings to process raw data. Individuals constantly economise to overcome these constraints but this means that these data are reduced to abstract codes (i.e. information) which can only be diffused when there is a shared (largely tacit) context which will permit their interpretation.

Such a perspective takes as its starting point that individuals and organisations process data, not information. With this point of reference, information is understood as the result of a complex production process whereby undifferentiated data are codified, abstracted and diffused so that meanings, interpretations and significance can be associated with them. Far from being a ubiquitous commodity, information must be located in a much wider and largely tacit context in order to be usable at both the cognitive and organisational levels. In particular, the relevance and significance of the data have to be established and confirmed by the social actors who are involved.

Following Boisot (1995), it is necessary to distinguish between data, information and knowledge. Information is defined as the product, rather than the input to informational processes. Human beings create intelligibility from the large amounts of data which bombard their senses by a process of abstraction and codification. Information is therefore a response to the need to economise on limited human capacity to store and process data. Similarly, because human communication always requires a degree of codification and abstraction, the receiver of the message is still only receiving data. Interpretation and successful receipt will therefore be dependent on a mutually shared and largely tacit context. Within this framework, knowledge represents a disposition to act on the basis of data made intelligible through information production and exchange processes (IPEPs).

An IPEPs based approach permits the impact of ICTs to be examined in terms of how they mediate production of information, with equal emphasis given to their automating capability and their ability to re-establish a richer, more ‘human’ environment for the communication of less codified data. By focusing on how ICTs can enhance the production of information from data, their application and use can be properly differentiated. Thus, these technologies are no longer viewed as simply a set of process technologies which bring about substantial reductions in the cost of handling information as a universal commodity but are seen as having an impact in several different ways, depending on the conditions that are required for producing information from data.

This provides the basis for considering the mediated impact of ICTs to be examined in three different ways rather than simply as a form of ‘information’ process automation. Firstly in terms of the automation of data processing. ICTs permit a hugely increased volume of data to be processed in its raw state thus limiting the need for data codification, abstraction and diffusion. Secondly, in terms of the capability of ICTs to ‘informate’, that is to create new data as a result of applying ICTs. This feature is unique to ICTs as a group of technologies and was first described in detail by . Thirdly, the effect of ICTs on the process of communication itself and in particular, the ability to ‘re-personalise’ communication. . This latter feature was found particularly important in reviewing the use of ICTs in commercial banks. It focuses on the way ICTs add incremental communication channels which can help to enrich human communication that would otherwise be restricted to higher levels of codification.

While ICTs clearly have the potential to automate processes which involve the handling of data, failure to focus on how information is actually produced results in all ICTs being given equal treatment and in a failure to differentiate in their effects. This leads to specific applications of ICTs becoming misunderstood and apparent anomalies in implementing specific technologies being simply reduced to a failure to achieve some hypothetical ‘best practice’. Thus rather than viewing ICTs as having a single unitary affect on information costs, it is argued is that the evidence points to a different picture, one in which ICTs are selected in a way that reflects the specific characteristics of IPEPs at the institutional level. Rather than being seen as a digitised commodity, the experience of ICT selection in commercial banks points to the need to examine a wider institutional environment within which information and ultimately knowledge is created. In particular, it suggests that there should be focus on those technological features which permit communication to take place under conditions which mimic direct human interaction, where the need for codification is minimised and which thus provides a richer tacit environment within which data can be ‘contextualised’. Thus, contrary to the view that ICTs lead to progressive codification, such a perspective looks for technologies which are being adopted to facilitate the greater richness of multi-channel, ‘re-personalised’ communication. In this context, rather than reducing or eliminating the ‘intangible’ qualities of information, ICTs give these tacit elements greater rather than reduced significance.

For example Boisot suggests that rather than increasing the level of codification, ICTs permit a 're-personalisation' of information exchange relationships by virtue of their ability to communicate large volumes of data, without the need for codification. In other words, human communication no longer needs to be reduced to the dots and dashes of the telegraph in order to escape its spatial and temporal constraints. New digitally based communication technologies permit much greater richness to be preserved, initially using voice, data and facsimile, but culminating in full multi-media contact. Boisot explains that by massively increasing the amount of data which can be transmitted, digital communication technologies effectively reduce the pressure to codify. Communication, therefore, can retain more tacit components directly rather than being dependent on a prior understanding of the context in which the transmitted data is received. In this respect it is worth noting that two of the most successful ICTs in terms of both their impact and the revenues which they generate, have been the telephone and the facsimile machine. Both permit human communication to take place on a richer, less codified level. Despite being more than 100 years old, voice telephony continues to experience rapid growth when one includes the most recent variants like mobile telecommunications. In the case of the facsimile machine, the technology released written telex communication from the straitjacket of the telex. The facsimile permitted typed documents to be annotated, to be supplemented by drawings and, if required, to be transmitted in hand-written form. In many commercial sectors including banking sector, time sensitive communication which was once almost exclusively telex based has switched almost entirely to facsimile based transmission.

The next section will go on to examine the elements which make up a broader context for examining informational issues at the institutional level, specifically how information is produced and exchanged. The resulting theoretical model is then compared by examining the experience of implementing ICTs in commercial banks which was the basis for a recent study of technological change in this sector.

Boisot, in a recently published book, has made an important contribution to understanding the nature of information, how it is processed and becomes usable knowledge in organisations . Boisot's starting point is that economics in its present essentially neo-classical guise implicitly takes energy as the primary source of wealth, with information only providing a supporting role. By keeping information issues out of the equation, its conceptual tools become increasingly inappropriate in a modern economy where innovation, both in the form of knowledge and artefacts, becomes a form of wealth in its own right.

Boisot argues that a new theoretical model is required when looking at information and knowledge since neo-classical economics is not helpful in developing a theory of how useful information is produced and exchanged. In order to develop such a model, Boisot starts with a fundamental premise, namely that in order to avoid being overloaded with data, human beings (and ultimately nature itself) have to achieve economies in the handling of raw data. In other words, raw data can never be processed in its originally sensory state but instead, various strategies are employed which ensure that there is not an overload. However, this process of data economising creates a dynamic environment in which the codes and abstract concepts used to ‘chunk’ the data into more manageable elements, will themselves become part of the 'data' shared with other human beings.

Boisot's work is particularly helpful in eliminating the sharp distinctions which are made between codified and tacit knowledge. He avoids these by arguing that both represent components in a larger multi-dimensional information or ‘I-space’. In this respect, tacit knowledge plays a key role by providing the context within which more codified information can be rendered meaningful. Boisot develops a dynamic picture of human beings constantly interacting with their social and physical environments by codifying and abstracting from the raw data which their senses pick up. However, since this is happening in a social context, the way abstraction and codification takes place becomes determined, and is in itself determined, by the institutional structures which are created.

Boisot's theoretical model identifies two key strategies for managing data - codification and abstraction. Although the two operate in harness, codification simplifies the raw data by putting it into defined categories, while abstraction sets the boundaries around the data which is included for consideration. comments that "Simplifying somewhat, we might say that codification reduces the complexity of forms whereas abstraction reduces the complexity of content. The first proceeds by differentiation, and allows us to enumerate finite sets of discrete elements; the second aims for integration, and brings elements so created into some limited and specifiable relationship with each other".

Boisot goes on to argue that these two processes of codification and abstraction operating at the individual level, in turn determine the way the resulting information can be communicated. Codified and abstracted data can, in principle, be transmitted or diffused to a much wider audience because it is no longer necessary for human beings to be in physical proximity for communication to occur. However, unless the meanings can be properly decoded, accurate interpretation cannot take place. Codification therefore results in an element of the richness of the message being sacrificed. Furthermore, codification is dependent on the existence of a shared context in which the message can be interpreted. This context is dependent on tacitly based knowledge without which the message would be rendered meaningless. Boisot argues that by increasing the richness of the message, potential ambiguities in interpretation are removed since more of the context associated with the message can then be retained. ICTs can be used to facilitate this process, which he refers to as ‘re-personalisation’,

Boisot's approach, therefore, extends beyond just regarding information as something that already exists and that becomes diffused. He moves the focus to an examination of how codification and abstraction in turn determine the level of information within an institutional/organisational context. In other words, IPEPs become the key issue, rather than asymmetries in information processing, accordingly "institutions are themselves the outcome of information production and exchange processes and not merely devices for their governance" . Boisot argues that writers like Williamson and others within the ‘new’ institutional economic tradition, miss the point when they focus on transactions as the principal unit of analysis. In common with other critics, he says that a focus on transactions serves largely to reinstate the equilibrium analysis of neo-classical economics but within the boundaries of the firm. Informational issues become relegated to considering the asymmetries which determine whether a particular transaction is subject to markets or is internalised.

Applying Boisot’s I-space model to ICTs has a number implications. By increasing both the volume of data which can be transmitted as well as the volume of data which can be processed in a given unit of time ICTs are seen to have an important impact on the I-space. However, the relationships are complex and Boisot's model demonstrates that considering ICTs in a deterministic fashion is not very helpful. In order to be used, ICTs initially require certain minimum levels of codification and abstraction. This in turn has implications for how the message is communicated and how it is received. However, as ICTs advance, they have the effect of increasing the diffusion of information at given levels of codification and abstraction while also reducing the cognitive pressures to economise on data processing. In other words, a larger audience can be reached in a given unit of time but, more important, communication can take place at a less impersonal level, requiring fewer 'communication relays'.

Elements of the literature on the effect of ICTs assumes that these technologies are part of a trend towards greater codification of information. In part, this is a result of an emphasis on IP at the expense of IPEP issues. In contrast, Boisot argues that ICTs impact on the I-space by enabling a richer context in which the communication will occur. This is achieved by removing spatial obstacles which previously would have demanded greater levels of codification and abstraction. He observes that "the new computing and communication technologies are avid consumers of data and they can digest great quantities of it in an unrefined form. Thus for a specified diffusion population with a given degree of spatial scattering, communication will be increasingly re-personalised. Many of the qualities of face-to-face exchange that had been lost through codification and abstraction will now be restored independently of spatial distance". .

As evidence to support this argument, Boisot cites the de-layering which has been taking place in the large Chandlerian economic organisations over the past two decades. This has resulted in fewer communication relays between the top and the base combined with a greater capacity at the lower level of the organisation "to absorb and master at an implicit level, the abstract codes used at the top" . In technological terms, this translates into a preference for using the telephone rather than written communications, the facsimile machine rather than the telex and, ultimately, multi-media based communication over currently based technologies.

In this context it is interesting to examine the growth in the use of electronic mail. The introduction of this technology has been considered to offer evidence of a broader trend towards digitisation which is assumed by much of the ICT literature. However, on closer examination, electronic mail can be understood as offering a new channel for communication that appears to confirm Boisot's emphasis on the capability of ICTs to ‘re-personalise’ human communications rather than increasing the level of codification. Electronic mail functions in this way by overcoming some of the formalities of paper based communications and while also overcoming the temporal constraints experienced with telephony based communications. This observation was confirmed in the commercial banking study (Crede 1997) where a rapid and successful introduction of internal e-mail was observed. However, rather than displacing other more direct methods of communication, e-mail was found to complement existing forms of internal intra-organisational communication and in particular, to reduce some of the previous artificial formalities of paper based internal documentation.

Key to Boisot’s conceptual model, is the view that IPEPs occur within a multi-dimensional I-space made up different components which change dynamically and reflecting what Boisot terms the ‘distribution of prior cognitive investments in the I-space’. Information ceases to be an undifferentiated factor of production but is itself both a product and a process in its own production and exchange. Boisot’s model offers a basis for understanding how information is produced and exchanged within a particular institutional and organisational environment. It overcomes the limitations of the IP model which is implicit in much of the existing literature on innovation in financial services. Boisot offers a way of understanding how cognitive data is turned into information and knowledge. This paper argues that when applied to specific institutional settings, such as commercial banks, it provides a better model for understanding the way in which ICTs have been adopted and how ICTs in themselves have stimulated organisational change.

A growing element of the literature on innovation is concerned with the role of tacit knowledge in technological development. It has been recognised by a number of writers to represent a key component for examining the nature of innovation. Senker argues that although there has been substantial codification of knowledge in the 20th century, this has not diminished the contribution of tacit knowledge to innovation. The importance of tacit knowledge is based on the fact that heuristic, experiential and internalised knowledge is difficult to communicate and is best learnt through practical examples .

Where the innovation literature has acknowledged issues arising from tacit knowledge, this has principally been in the context of what skills are required for firms to create and exploit technological innovations. The effort has focused on achieving recognition that these uncodified components need to be given consideration and that models of the innovation process which rely entirely on formalised or codified knowledge are therefore deficient. However, once we accept Polanyi’s starting point for all human knowledge - "we know more than we can tell" - a tacit context for producing meaningful information from data assumes importance in its own right. Drawing on Plato, Polanyi points out that if all knowledge were explicit, that is capable of being clearly stated, all problems would either already be known or unknowable. This paradox can only be accepted by acknowledging that there is a tacit dimension to human knowledge which enables both the original problem and its solution to be determined using knowledge which is not explicit. In other words, tacit thought becomes "an indispensable element of all knowing" and the "ultimate mental power by which all explicit knowledge is endowed with meaning", .

While tacit knowledge can be held in its own right, information representing abstracted and codified data, requires a tacit context in which it can be understood and applied. Thus tacit components are always integral and can never be entirely separated. For Boisot, codification is only one element of a much richer I-space which by definition comprises other more tacit forms. Furthermore, codification is not something that has developed as a result of ICTs; it is viewed by Boisot as an integral part of the cognitive process which allows human beings to make sense of their surroundings and to communicate this with their fellows. .

The tacit dimension, therefore, must be seen as an indispensable element of information production and exchange. As Boisot points out, tacit knowledge will always define the context within which information can be understood. A linearity between codified and tacit knowledge is therefore misleading and simplistic since there is a constant dynamic inter-play between the way in which codified and tacit knowledge is used. This can best be illustrated by an example quoted by Boisot, that the map which we carry of our local neighbourhood is in our head and not in our pockets. A street map represents a highly codified representation of a particular geographical space which, in its actuality, is handled in a much richer form. Once we become familiar with a new location, the codified guide is no longer required. However, if we are giving someone local directions or if we need to calculate the time it will take to reach a certain destination, we may once again wish to consult the street map as a codified representation that will facilitate the giving of directions or the calculation of time for travelling.

While the existence of tacit knowledge fits well with an approach which sees information as arising from a production process, at the same time it seriously undermines the acceptability of the IP model. In order for organisations like banks to be IPs, information has to be able to acquire the characteristics of a ubiquitous commodity, that is to assume a form which in theory excludes any tacit dimension. However, as Polanyi points out this approach encounters immediate difficulties. He observes that "…a mathematical theory can be constructed only be relying on prior tacit knowing and can function as a theory only within an act of tacit knowing, which consists in our attending from it to the previously established experience on which it bears. Thus the ideal of a comprehensive mathematical theory of experience which would eliminate all tacit knowing is provided to be self-contradictory and logically unsound" .

A key component of tacit knowledge is what Polanyi terms ‘indwelling’, that is the ability to maintain awareness of the broader context while considering the particular. Polanyi describes it as tacit knowledge that "dwells in our awareness of particulars while bearing on an entity which the particulars jointly constitute". In order to gain this tacit knowledge it is necessary to learn the component parts, confident in the knowledge that "a teaching which appears meaningless to start with has in fact a meaning which can be discovered" by achieving a similar level of understanding to that of the teacher . Although Polanyi is primarily concerned with the implications this has for ‘scientific rationalism’, he acknowledges that the process has a broader significance and is closely associated with trust. Polanyi quotes St. Augustine who taught, "Unless you believe, you shall not understand".

Based on Polanyi’s concept of ‘indwelling’, this paper introduces a new concept of ‘tacit authentication’ based on the notion of trust which is associated with the authority of a teacher imparting tacit knowledge. This concept attempts to capture the role of the tacit dimension in the attribution of significance and relevance to data. Polanyi points out that the development of science has required that "scientists must rely for their facts on the authority of fellow scientists". Similarly, within any commercial institution or social organisation, there is a requirement to rely on the authority of more senior executive ‘teachers’ in the process of seeking tacit authentication, so that usable information can be derived from unverified and often privileged and confidential data.

The relationship between trust and the production of information is also discussed by Boisot. He suggests that trust plays an important role in defining the nature of information exchange at the level of the institution. "Uncertain and complex transactions pose the problem of trust in an acute way and the greater the uncertainty, the greater the amount of trust between the parties needed to overcome it" . Boisot identifies the need for greater levels of trust when data is at its most abstract, least codified and most diffuse. He stresses the importance of prior familiarity as "a particularly important requirement in fostering attitudes of trust towards expected outcomes" . ‘Tacit authentication’, therefore, can be understood as an effective mechanism for overcoming the limitations imposed by the particular characteristics associated with a given set of data and the associated IPEPs.

Cultural issues associated with trust have been a focus in sociology for some time and this literature can give additional depth to a discussion of ‘tacit authentication’. Sztompa, in examining the implications of a deficiency of trust in Polish culture, made a recent attempt to develop a ‘sociological theory of trust’ . While Sztompa addresses cultural rather than informational issues, he also highlights how ‘tacit authentication’ can apply to a range of institutional settings. Sztompka defines trust as "a bet on the future contingent actions of others". Trust removes the need constantly to monitor, cross-check and watch and, therefore, "raises the overall level of mobilisation, activism and freedom". When trust breaks down, it often does so irrevocably with potentially very serious consequences. Sztompka quotes data on Polish public opinion which reveals that over 60% of the population is convinced that economic data released by official sources is actually false and similarly over half the population is fundamentally sceptical of any data they receive from the Polish state-run media, be it television or newspapers. Sztompka points out that when a society like Poland becomes pervasively distrustful, functional substitutes emerge to fill the vacuum. Corruption and vigilance create the illusion of empowerment, either through a network of reciprocal favours or through the direct supervision and control of others. Alternatively, society can become fragmented or ghettoised, reaching a state where "a diffused distrust in the wider society is compensated for by strong loyalty to tribal, ethnic or familial groups".

However, the role of trust in an informational context should be distinguished from a discussion of trust as part of wider system of governance. is largely concerned with the latter where the author argues for the need to maintain a ‘social capital’ of trust as the basis for economic prosperity, comparing the economies of several major industrialised countries and how arguably the nature of economic development is linked to the level of trust inherent in the respective society. However, while Fukuyama is primarily concerned with trust as a critical component of all economic relations, he confirms the importance of tacit authentication in establishing the informational components of economic exchange. Unlike the ‘new’ institutional economics, he argues that the relationship between buyer and seller has to comprise more than a ‘bundle of contracts’ since no legal agreement is ever able to compensate for the absence of the minimal level of informal trust required. In this way Fukuyama acknowledges the role of ‘tacit authentication’ as means by which data relevance and significance can be determined in order to permit the underlying economic transaction to proceed.

Finally, a further feature of tacit authentication is related to learning. Individuals are constrained by their ability to interpret complex reality (bounded rationality). points out that learning is socially constructed inasmuch as what is learned is profoundly connected to the conditions in which it is learned. The learning that takes place by an individual in an organisation is very much dependent on what is believed or already known by other members of the organisation. Learning is therefore contextually constrained and a function of a complex web of tacitly based shared meaning.

Tacit knowledge, therefore, is a key social component of the context in which new information can be evaluated, a process that can be described as 'tacit authentication'. In other words, tacit knowledge provides the context for the communication and diffusion of information. The concept of ‘tacit authentication’ is at variance with the IP model and, in particular, the view taken by those writers who argue that progressive digitisation will ensure the progressive removal of existing information asymmetries. Writers like David and Foray argue that on a theoretical level, tacit and codified knowledge is in effect interchangeable, providing the ‘price is right’. They argue that the "…codification of knowledge is a step in the process of reduction and conversion which renders the transmission, verification, storage and reproduction of information all the less costly. Whether or not knowledge is put in codified form is in part a question of how costly it is to do that" . This paper argues that the process of ‘tacit authentication’ assumes particular importance when both the quantity and source of data, place special requirements on the economising processes associated with the production of information from data. In this context, a commodity view of information, particularly one which views the ‘digitisation’ capabilities of ICTs as unlimited in principle, has to be seen to be incomplete.

In summary, this section has introduced the concept of ‘tacit authentication’ which aims to capture the information processes taking place at the organisational level, particularly as data comes to be given meaning, significance and relevance. The next section will examine this concept and the underlying IPEPs based approach to ICTs in the context of the commercial banking sector. Drawing on a recently completed DPhil research project by the author (Crede 1997), it will argues that the use of ICTs by commercial banks can be more effectively understood using an IPEPs based approach which overcomes some of the basic limitations of the IP model. The research suggests that there is evidence that commercial banks rely heavily on special kinds of data, namely privileged, private and confidential customer data, whose particular characteristics require special forms of information production in which ‘tacit authentication’ plays a key role.

A recent study of the commercial banking sector (Crede 1997) examined how advances in ICTs have changed the way in which commercial banks are incorporating these technologies in some of their core businesses processes. The study focus was the impact of technological change in credit assessment procedures which was then contrasted with the way in which ICTs are used in the implementation of Automated Teller Machine (ATM) networks, when these are viewed within the broader context of payment and settlement systems. This comparison was intended to draw out some of the institutional characteristics of commercial banks, illustrating the need to view IPEPs in a broader context. It was also intended to compare the information obtained from payment data with the more tacitly-based credit related information that is the product of the credit risk assessment process. The research was based on a a case study methodology to examine these areas of ICT application. The empirical data was based on twelve major world class commercial banks which were selected by virtue of their size, global presence and ability to provide representation from all the major industrial regions. The objective was to identify key common features in the production and exchange of information that transcended other specific cultural and historical differences.

The main research findings were that banks perform their role as savings intermediators, principally by ‘adding trust to data’. This has important consequences for key bank processes, the selection of ICTs and the design as well as operation of digital networks. Commercial banks depend largely on the confidential and proprietary data which is made available by their clients. This data is provided on the basis of trust but also requires internally generated ‘tacit authentication’ before it can become information in a usable form. This process can be mediated by ICTs by ensuring that a rich, highly tacit context for human interaction can be preserved. The resulting selectivity has meant favouring ICTs that ‘re-personalise’ and placing correspondingly less emphasis on the storage and manipulation of data in its more codified form. While banks have successfully adopted the use of digital networking technologies, issues concerned with trust and security have been paramount. Directly controlled, proprietary networks continue to be preferred even when these result in a higher cost of handling data.

Despite the growing variety and sophistication of financial services, all the major commercial banks continue to derive more than half their operating revenues from their roles as intermediators between savers/investors and borrowers. In this respect, their core activities have remained unchanged for 100 years or more. Furthermore, while technological change normally brings with it changes in market share and the arrival of new entrants, the commercial banking sector is characterised by major institutions which all have long histories during which they have maintained a more or less dominant position in their respective markets. As financial markets have become more global, each bank has built on its existing level of international exposure. An IP model would have anticipated that a much greater level of upheaval would have taken place in the global banking industry. The longevity of some of commercial banks (for example Deutsche Bank, HSBC and NatWest which were all included in the study) stands in stark contrast to the leading elements of the manufacturing sector where many of the major Fortune 500/ FT 500 companies were not even in existence 25 years ago. If banks were predominately IPs, it could be expected that the degree of impact on industry structure would be greater. Banks have been affected by ICTs in terms of new competitors at the margins but continue to dominate the sector as a whole.

Similarly, if ICTs were as transformative in their effect on institutional structures as the IP model would lead us to believe, we would expect to find a progressive automation of the functions of skilled banking staff, resulting in lower overall employment levels, reducing staffing costs and increasing incremental expenditures on ICTs. However, in contrast the study found that commercial banks continue to spend a major part of their resources on staff. While there has been rationalisation in domestic branch networks, major banks are expanding staffing in their foreign operations as well as specialist functions. Overall, staff numbers have either risen or remained constant. This is in sharp contrast to other major companies which have dramatically reduced staff numbers over the past 10 years. For example, in the period 1989-94, IBM reduced its world-wide staff from 383,220 to 219,839 (a decline of 43%), while Philips reduced its staff from 304,800 to 249,759 (-18%). In this same period, Deutsche Bank’s staff increased by 30% (56,580 to 73,450), HSBC’s by 91% (including acquisitions form 53,375 to 101,737) and Bank of America staff increased by 50% (including acquisitions from 54,779 to 82,100).

The study also found that the proportion of total operating costs allocated to salaries and wages in each of the banks has been stable, over the past five years. This is difficult to reconcile with the notion that banks are progressively introducing process automation in order to automate IPs. If banks were IPs, automation would have been expected to have had a much more widespread effect. In all the CSBs, staff costs represent around 50-60% or more of total operating costs. This compares with an information service company like Reuters (whose services are primarily sold to commercial and investment banks) where staffing costs represent around 33% of the total cost base. Furthermore, when compared to national and world averages, employee remuneration in the CSBs is very high. Based on information in their published accounts average salaries are often in excess of $50-$60,000. If a major proportion of these employees is simply processing information which already exists, this raises the question of why such skills have to be so highly rewarded. To put the figures in perspective, Deutsche Bank paid out more than $6 billion to its 73,450 staff members in 1995. This figure is only slightly less than the Gross National Product for a developing country like Kenya, with a population of more than 25 million people. At the same time, the average salary level appears to be rising. For example, for the three year period 1992-95, Deutsche Bank increased its average pay per employee by 47%. These findings, while not offering proof as such, are nevertheless very difficult to reconcile with a model which sees information as being the object of progressive investment in process innovation. Given average annual increases which exceed both inflation and average wage increases, what appears to be taking place is a further upgrading of professional and specialist skills and not the partial de-skilling which an IP based model would have led us to expect.

Economies of scale and specialist knowledge, are traditionally seen as giving banks the necessary comparative advantage to perform credit intermediation. However, a detailed examination of the credit review process in the study confirmed that commercial banks are engaged in producing bank information from confidential customer data rather than processing information which is already ‘out there’. It confirmed that the credit review process is not a simple case of collating different elements of the argument for or against making a loan and then making a collective decision. In order for the information to be accepted in the first place, tacit authentication has to take place and this forms an integral element of the review process. Credit applications are prepared, then reviewed and analysed in depth by the approval hierarchy. Key elements of the presentation are filtered through the credit review process and different values attributed. The credibility of the proposal is established through the series of personal contacts which make up this process all involving degrees of personal trust. The credit review process was therefore not an exercise in IP but was found to be built around a complex process of giving meaning, significance and relevance to this data with much less emphasis on the more mechanical exercise of weighing up existing information which is favourable or unfavourable to going ahead with the proposed transaction. Key bank processes, particularly those associated with credit assessment, thereby involve complex processes of authentication of data obtained as a result of privileged access to customers.

The study found that credit decision-making procedures emphasise personal experience and acquired skills rather than more formalised forms of knowledge. This was reflected in the organisation of credit approval processes. Although attempts had been made periodically to separate out credit decision-making and to give it greater standing as a separate function, de facto control is held by those managers with the most seniority within the overall management hierarchy. The same emphasis on personal experience and judgement in contrast to more formalised forms of knowledge is supported by other empirical research. An examination of the German banking system by , also based on in-depth interviews found that credit decisions made by bank loan officers in German banks placed the greatest emphasis on the experience and judgement of loan officers. Although German banks have acquired a reputation for superior industry evaluation skills, Edwards’ research provided no evidence that loan officers in German banks possessed specialist industry knowledge or that there was a specialised technical department. Other studies have been carried out which confirm that similar credit procedures operate in UK and Japanese banks. Similarly, points out that historically major US banks have had contrasting bad loan experiences while operating in the same sector. He specifically cites the example of Chase Manhattan which had major write-offs in the property sector in the early 1970s only to repeat the same experience with loss making property loans in the 1980s. In this case, organisational learning clearly failed, yet competitors like Morgan Guaranty were able to achieve a much lower level of write-offs. These differentials are rather difficult to explain within the terms of the IP model, where one would expect similar sized organisations to achieve equivalent results in information processing.

The study also found that very little automation is applied to the critical process of credit evaluation, which remains essentially a highly interactive, human process involving a large amount of communication. Tacit authentication plays an important role in co-ordinating data which cannot otherwise be made freely available. The credit approval process makes little use of external electronic data and places important restrictions on data that are internally generated in a way which serves to inhibit organisational learning. If commercial banks really did act as information processors in the broader sense, it could be expected that there would be large departments whose role it was to provide access to external data sources. These sources could then be combined with information supplied by customers and knowledge generated within the organisation. Instead access to external data was found to be often relatively primitive and artificially constrained on financial grounds. While some external information in the form of newspaper clippings or trade press articles were utilised, on-line database would generally be employed to assist with marketing activities rather than to support the credit process. Even when electronic sources were available, these were generally very restricted in scope and compared unfavourably with the resources one would expect to find in a research based organisation.

There was also evidence to suggest that IPEPs in banks in their currently constituted form have difficulty handling external data of this kind. The subtle, complex and tacitly based authentication processes which are used to filter customer provided data cannot be readily applied to these external information sources. On-line electronic data often created confusion since its origin could not be verified tacitly. For instance, a photo-copy of a Financial Times clipping has distinctive typographical style and familiarity giving it an immediately shared label that everyone can understand. Text from an electronically based data base on the other hand lacks many of these tacit cues and is therefore much more difficult to authenticate.

The study also found that the internal data which commercial banks have is not widely distributed but controlled through a process of negotiated access. Data in customer files are strictly controlled and remain the responsibility of the respective account manager. Its release can only occur with the express permission of this individual so effectively access to the data has to be negotiated in every individual instance. Banks placed relatively low emphasis on digitisation of their own information resources and have retained paper based systems rather than manipulating data in its electronic form. This applies particularly to what is arguably their core process, the processing of credit decisions. The emphasis is on paper based storage systems, serves to reinforce negotiated data access. Access to paper files locked in a filing cabinet is inevitably easier to control than if the data is stored in its electronic form on a file server. In contrast to these findings, an IP model would predict that such a key process would become progressively automated/codified as advances in ICTs reduce the cost of information processing and as the benefits and feasibility of wider data access become reflected in lower costs and therefore provide greater competitive advantage. In this way, despite having the electronic means for greater sharing of data, issues concerned with confidentiality and security require that data has to be compartmentalised, remaining specific to each department with access obtained by going through a human filter.

Similarly, although extensive investments have been made in PC-based networks they function largely to improve communication and provide office automation rather than common access to data. In many of the CSB interviews, these inconsistencies were acknowledged but it was argued that inefficiencies in data dissemination had to be tolerated since if data was made more widely available using electronic means, it could not longer be controlled. Implicit was the threat to mutual trust which would occur, should customer data be compromised as a result of such open access. Similar issues have been at the fore of the development of ATM networks. The emphasis has been on establishing proprietary networks, thereby ensuring that security, control and trust can be maintained. While this approach has been very successful in achieving a steady increase in the use of ATMs for cash withdrawals, it has been achieved at relatively high cost which in turn defines the density and degree of functionality of the network for the user. In the UK, and in other countries where the banks operate similar ATM networks, the present selection of technologies results in a payment system where over 80% of all payments are either made in cash or are cash related . At an average ATM transaction cost of US$ 0.50, cash provides the only real alternative for small value transactions and establishes barriers to attempts to introduce electronic payments for low value transactions. The development of ATM networks can therefore be seen in a broader context in which information production takes place. Data transfers within ATM networks have been designed to retain the direct customer relationship and to maintain trust. However, this in turn defines the options available for other forms of electronic payments, effectively reinforcing the dependence on cash as the principal medium for payments.

Similarly conclusions may be drawn from the continuing process of concentration which has characterised the emergence of key global financial centres in London, New York, Tokyo and to a lesser extent, Hong Kong and Singapore. Despite advances in ICTs which in principle make it possible to deal and trade in financial securities from any remote geographical location, London as well as the other global financial centres continue to maintain and extend their hold. In London, close human contact is assured by the fact that the offices and personnel engaged in the global markets are all concentrated in and around the area referred to as the ‘City Square Mile’. It contains the highest concentration of international financial institutions in the world. London has become the world centre for foreign exchange trading, euro-bonds and international equities. It has also become the centre for international lending, with the number of foreign banks located in London steadily increasing over the past 20 years. There are now more than 540 banks represent compared to under 300 in the early 1970s.

In conclusion, this paper has attempted to develop a model of informational processes and the role of ICTs which moves beyond some of the limitations identified with a simpler IP model of organisations. An attempt has been made to demonstrate the value of such an approach by examining the experience of the implementation and use of ICTs in the commercial banking drawing on recent research in this sector. It has been argued that an IPEPs based approach benefits in providing a ‘richer’ picture and provides a greater understanding of how technology can mediate informational processes at the level of the individual organisation. A further strength is that it can be combined with other research which has examined organisational learning and core competencies. The specific skills, abilities and routines that organisations have developed for a particular form of information production and exchange can thus be more easily identified.

One potential weakness of the theoretical model is that by placing a stronger emphasis on institutional factors, the dynamic qualities of the technologies themselves may be under-emphasised. Earlier in the paper, three ways in which ICTs can mediate the production and exchange of information were identified, that is data automation, ‘informating’ and ‘re-personalisation’. The evidence obtained from the banking study emphasised the important role performed by ICTs in facilitating communication and specifically in establishing an environment in which ‘tacit authentication’ of data can operate most effectively. The study findings further suggested that ‘informating’ plays a much lesser role and that this is primarily due to the nature of the confidential customer data which places restrictions on intra-organisational data sharing.

In her study of of ICTs, Zuboff emphasised the need for new competencies before the benefits of ‘informating’ could be fully realised. In order to move beyond automation of data, individuals have to acquire skills which permit the new data which is generated (in itself highly abstract in form) to be properly interpreted. Zuboff makes a distinction between the ‘action’ based skills which can draw on a wide range of situational clues available in an inter-personal setting, compared to the more cognitive, analytical and procedural know-how which individuals require before data generated through automation can be attributed meaning and signficance. While the IPEPs model provides a very effective tool to understand how current information and production exchange processes are mediated by ICTs, it may be subject to more restrictions when one is trying to establish the extent to which technologies have the ability to restructure the underlying information production and exchange process itself. This raises the possibility that as advances in ICTs open up new ‘informating’ opportunities, they will in turn start to modify, rather than simply mediate, the information production and exchange processes themselves. New organisational forms may arise as a result, with consequent implications for the competitive position of incumbent providers of products and services. In particular, as vast amounts of data are becoming networked and linked to powerful data search engines, new competencies may start to displace some of the advantages associated ‘tacit authentication’, which to date have proved to be so effective in commercial banks for attributing relevance, meaning and significance to customer supplied data.