The scope of any project measuring New Zealand's electronic commerce performance must fulfil many objectives. Principally, it must enable both a measurement of "what is being done" and an assessment of "how those things that are being done" impacts upon the wider economic and social performance of the country. A simple set of measures counting a predefined range of activities will be of limited use unless it is integrated into a wider framework of economic and social performance measures and analysis. That is, how do these measures reflect or facilitate predictions of changes in firms and markets, individualism and globalisation, etc.? Furthermore, if empirical measurement is to be undertaken, and the factors one wishes to measure are intangible (such as information value), then proxies must be identified that can both be measured and appropriately capture the elements for which they stand as proxies.

These issues have been the subject of much research, and are addressed from a theoretical perspective in publications such as the OECD's (1999) Economic and Social Impacts of Electronic Commerce, Colecchia (1999), Blair and Wallman (2001) and Lev (2001). The added demand for adoption of consistent international benchmarking, however, has resulted in support for the set of measures developed by the OECD becoming widespread, with national measurement exercises reflecting the OECD criteria becoming a de facto standard for international electronic commerce and "knowledge economy" performance measurement reporting. For example, Australia's quarterly State of Play reporting and Singapore's statistical collecting methodologies (Wong and Lam (1999)) draw heavily on this work.

The OECD framework (Colecchia (1999) - summarised in Appendix 2) derives a set of measures based upon the premise of recording the introduction and maturing of a new segment of the economy. It recommends tracking measures of e-commerce readiness, intensity and impact over time, as illustrated in figure 2. The National Office for the Information Economy's (NOIE's) translation of this methodology into a series of practical measures which track performance is amply demonstrated in the frequent publication of The Current State of Play. A summary of the data collected for the Australian performance measurement exercise is contained in Appendix 3.

Figure 2: Availability of E-Commerce Indicators across OECD Countries

View full size (808x433 pixels)

Source: OECD

The defining feature of the Australian exercise is the comprehensive nature with which it collects into one document data measuring not only business readiness and intensity, but also individual and household indicators. This recognises that the impacts of e-commerce are economy-wide, and not limited in scope or effect merely to the business sector. It acknowledges the interaction of firms, markets and individuals, and the participation of citizens in a modern economy (as per Section 1).

Figure 3: Measuring E-Commerce Size or Its Impact in Relation to Overall Transactions/Activities

Source: OECD

While the focus is still on collecting statistics that measure the extent of e-commerce activities, there is a recognition that these statistics provide information and insights relevant to impact analysis. Thus it provides a base to move beyond the mere numbers, as represented in Figure 3 to a set of policy-based research priorities, as illustrated in figure 4.

Figure 4: Translating Policy Needs into Research Priorities

View full size (808x366 pixels)

Source: OECD

How Effective Are These Precedents?

While it is acknowledged that these frameworks have played an important role in both defining and measuring technology-specific activity in what has been identified as a new segment of economic activity, it is becoming increasingly evident that there is still a gap¹⁶ with respect to how these measures translate into overall economic and social performance in an environment where "old" and "new" technologies exist side-by-side (Power (1999), Colecchia (1999)). Furthermore, the emphasis on "electronic" has focused the measurement effort on electronic technology indicators without necessarily addressing the underlying issues of how these electronic technologies are changing the ways in which we are using information, and consequently non-electronic production, dissemination and consumption technologies (Borenstein and Saloner (2001)).

The existence of the gap is recognised principally by those charged with tracking national performance (such as national statistics agencies) and policy-making. For example, Poon and Avenell (1999) comment from the Western Australian State perspective that while:

"many of the statistics currently available are measures like network traffic, website hits or user demographics based upon self-selected participation ... are useful indicators, they reveal little about the true economic and social benefits of e-Commerce".

However, it would appear that measurement exercises such as that promoted by the OECD have gained favour because the pace of technological change leading to electronically based ways of transacting has generated an urgency to measure something. Resources have been devoted to measuring the tangible, such as technology uptake, because it is visible and measurable, rather than the "too hard to quantify" benefits of "convenience, improved lifestyle, or even being able to reach out to the community" (OECD (1997) Measuring Electronic Commerce, cited in Poon and Avenell (1999)). The danger is that in so doing, we have risked becoming captured by a "microchip myth" that only changes in electronic technology are relevant and justify measurement, while changes in human and other technologies, because they are difficult to measure, do not carry the same justification for allocation of measurement and analysis resource¹⁷.

The United States Department of Commerce publications The Emerging Digital Economy (1999) and The Digital Economy (2000) both highlight how the pace of adoption of electronic technologies is necessarily increasing the urgency with which new ways of tracking the growth and impact of the "modern economy" must be developed. The Emerging Digital Economy (1999) highlights as issues where direction is required:

The shape and size of the key components of the evolving digital economy such as e-commerce, specifically, and, more generally, the introduction of computers and related technology in the workplace.
The process through which firms develop and apply advances in Information Technology (IT) and the use of e-commerce.
The change in the structure and functioning of markets, including changes in the distribution of goods and services and changes in the nature of international and domestic competition.
The social and economic implications of the IT revolution and e-commerce such as the effects of investments in IT on productivity.
Demographic characteristics of user populations.

Yet while the need to identify causal links between e-commerce adoption measures and macroeconomic indicators such as productivity and demographic changes has been noted (points 1 to 3 and 5), the focus still rests upon the "technology" component of "IT" (point 4). The extent of this focus is evident in Haltiwanger and Jarmin (1999) who, in their attempt to address measurement issues from the perspective of gathering and analysing national statistics, identify three broad areas for consideration:

the impact of IT on aggregate activity such as productivity and living standards;
the impact of IT on labour markets and income distribution; and
the impact of IT on the way production is organised.

Although these three areas are broad enough to capture both electronic and non-electronic technological changes, the five categories of data that they identify as enabling these areas to be addressed are still strongly biased towards measurement of physical, and predominantly electronic, technologies:

Measures of the IT infrastructure (including)

investments in physical infrastructure
investments in software infrastructure
depreciation of infrastructure
"non-IT" equipment such as computer-controlled machines

Measures of electronic commerce

magnitude and type of business-to-business (B2B) and business-to-consumer (B2C) electronic transactions
separate measurement of digital and non-digital goods and services
use of e-commerce for transactions and non-transaction purposes (This has been presumed to represent "production" and "non-production" or ancillary purposes)

Measures of firm and industry organisation

improvements in IT, software, Internet on market structures
changes in location, industry size, organisational structure of businesses
changes in input mix (e.g. capital, labour, inventories) and relationships with other businesses (e.g. outsourcing)

Demographic and labour market characteristics of individuals using IT

those participating and those not participating in the digital economy
computer use at school, work and home related to measures of economic outcomes such as wages and assets, and demographic characteristics such as occupation, gender, race, age and place of residence

Price behaviour

quality-adjusted price deflators
price differentials across goods and services sold by different methods
price dispersion across producers using the same method.

We wish to acknowledge that all of the data collection frameworks identified and analysed above provide useful information about specific behaviours. To the extent that they change the outcomes of production and consumption activities, these technologies do represent a fundamental change and warrant measuring. For a country such as New Zealand that currently undertakes little systematic measurement of these activities at a national level, adopting such a measurement framework would offer a considerable increase in understanding "what is being done", and how "what is being done" in this country compares with "what is being done" in other countries. However, it is in linking "what is being done" with interpretation and analysis offering understanding of whether "what is being done" actually results in improved macroeconomic performance indicators, that knowledge is lacking. Hence, while we may know if country A is "doing more" of a specific indicator (such as average hours connected to the Internet), there is less certainty about whether "doing more" of this indicator is really driving changes in other indicators, such as productivity growth.

It is our contention that the emphasis given to the "electronic technology" component of "IT" has masked the role played by the "information" component, and thus obfuscated the resolution of how electronic technology uptake at the firm and individual level has contributed to economic and social changes at the macro level. If, as identified in section 1, information is becoming the key resource of the "modern economy", then it will be by analysing changes in the ways information, as opposed to technology, is used that the cause-and-effect relationships between electronic technology uptake and macroeconomic statistics, such as productivity measurement and demographic changes, may be given rational foundation.

An effective measurement exercise, therefore, needs to collect statistics not just of which technologies are used, but also statistics that reveal changes in the ways in which information is used. Furthermore, these statistics need also to capture information changes as a result of human and other non-electronic technology changes. Emphasis must move from collecting statistics based solely upon communication methods and technologies (the pipes that carry the information) to include statistics based upon the information these communication technologies facilitate the use of (akin to the water or gas transferred through the pipes) and the end products made using that information (akin to goods or services made using the gas or water as an input to production). In addition, these statistics must be analysed within a framework that recognises the role of information in the value-creation process.

An Alternative Framework

We strongly recommend that any statistical research effort must be focused upon how individual technologies change the way we use information, specifically:

what are the effects of a specific new technology on the ways in which we use information, and
how do we measure these effects (for example, what proxies reflect these changes).

Such research will provide a strong analytical base for understanding the ways in which economic value is created in an economy where information is a key capital resource, input factor and output of production, and we believe it will facilitate learning and understanding of the conceptual links between the firm and individual level technology measurements and national statistics (such as GDP and productivity growth).

Developing such a framework performance poses two challenges:

defining what is meant by the "modern economy"
defining what measures will provide valid indicators of performance in this economy

An important component in defining the "modern economy" addressed in section 1 above is the need to recognize that the changing structure of the economy has been, to a large extent, brought about by changes in the ways in which information is produced, transmitted, stored, processed and distributed. This has been a long-term trend. The changes brought about by recent developments in "information technologies" and the advent of electronic commerce form only a subset of the wider change in information usage¹⁸. An important challenge is to develop some conceptual clarity around the separation of the changes brought about by different uses of information from the physical technologies that make these different uses possible. Physical technologies will continue to develop, providing new and different ways to use information. However, the core product - information - and its physical and economic properties do not change as a consequence.

It is important that the measures developed reflect the underlying economic changes brought about by changing uses of information rather than merely counting the users and usage of specific technologies. The difficulties inherent in measuring variations underpinned by changes in information usage are compounded, however, by the fact that information is an intangible good. It is difficult to measure, and consequently difficult to value. Proxy measures must be developed to measure the extent of changes in economic inputs and outputs (and hence production processes¹⁹) wrought by changes in the ways in which information is used.

Again, it is stressed that this is not a new phenomenon. Defining and measuring performance of the service sector has always been problematical due to the large number of intangible inputs and outputs, and much productivity literature has been devoted to the measurement of this sector²⁰. Similarly, there has been much debate about the ability of performance measurement systems to recognize and capture the effects of changes to product quality, again due to the intangible nature of this element²¹. The advent of computerisation, digital information transmission and processing technologies has added significantly to the range and volume of intangible inputs and outputs which must be measured in order to assess the performance of an economy²².

At the same time, it is recognised that development of methodologies to capture the "digital" component of the "modern economy" should not be at the expense of improving the measurement and understanding of the "traditional economy" processes. For these processes are, and have always been, consumers and producers of information. There is much to be learnt still from relating measures of the changes in the quality and use of information and information technologies to changes in traditional measures. From this, we can learn about the impact of information and information technologies on the entire economy.

While macroeconomic frameworks such as that of Haltiwanger and Jarmin (1999) provide a useful "top-down" framework for analysing comparative performance, and offer valuable information on the performance of an economy in transition, we contend that they are largely "symptom-tracking" metrics based upon traditional economic performance instruments which do not necessarily recognise the emergence of information as a key good in the value-creation process. We believe that the core challenge of this brief is to develop a "bottom-up" approach for measuring New Zealand's performance, capturing:

the economic effects of new transaction forms made possible by new technologies (new technologies - as per the Solow definition)
the changes in information creation, flows and values made possible by new technologies (capital issues)
value created by the interaction of transactors, not just value creation within firms.

These prerequisites indicate that meaningful analysis must be conducted at the level not just of an industry or a firm, but also within the firm²³, and down to the level of all individual transactions. By gaining an understanding of performance at the transaction level, and in particular how changing information treatment affects transaction efficiency, we should also gain insights into how these changes impact ultimately upon the organisation of transactions into firms and markets (note that this also includes exchange of non-purchased goods via governments, families and other social distribution networks). That is not to say that some of the same information indicated necessary by Haltiwanger and Jarmin may not be useful and relevant in our analyses. The important elements become the assumptions under which these statistics are interpreted, and the conclusions that they support.

To that end, the starting point for this scoping exercise must be the key changes to the uses of information brought about by the digitisation of information (see the definition of Electronic Commerce in Section 1). Our framework therefore hinges on two questions:

Where is economic value being created in the "modern economy"?
What do we want/need to measure to assess the extent of value creation in the new economy?

Where Is Economic Value Being Created in the "Modern Economy"?

Many studies recognise that the sources of value creation have changed as a result of changes in information costs and uses brought about by electronic technologies specifically. While we recognise that these may not fully address the concomitant changes in non-electronic technologies, they provide a useful starting point to analyse where new value is being created, what we might want to measure and how we might go about measuring it.

Boston Consulting Group

The Boston Consulting Group (BCG)²⁴ classifies changes brought about by electronic technologies into two sets of activities,

value-shift activities; and
value-creation activities.

Value-shift activities are defined as those "activities that take value from one party and transfer it to another (zero-sum game)" and include aspects of aggregation, process automation and greater information transparency in processes such as on-line auctions. These activities may result in improved financial performance for the unit which uses electronic technologies to shift value-creation out of other entities and into itself, or which uses technologies to shift costs to other entities. While performance metrics for the observed unit (e.g. profit, market share, productivity) may improve, these are offset on an economy-wide basis by increases in the costs of, and reduction in value created by, other units. If the majority of new technologies are applied to these activities, then while large fluctuations may be evidenced at the firm and industry level, changes to overall performance (such as national statistics) will be negligible.

Value-creation activities are defined as those "activities that create new value through improved efficiencies or productivity (win/win scenario)". BCG includes lower marketing/sales costs, decreased transaction costs, lower cost in use, reduced inventory costs, lower cycle time and improved asset utilisation as examples of value-creation activities. These activities will result unequivocally in improved performance at both the measured unit and all subsequent aggregation levels.

Care must be given in analysis of changes using this framework to the extent of cost-shifting into, and value-shifting out of, those sectors not traditionally measured or measurable. Measured costs can appear to fall if costs have been shifted into an unable-to-be-measured area (such as product quality or unpaid time). While these may appear as "value-creation" on the basis of measured metrics, the reality may be a reduction of overall welfare as non-measured resources are substituted. Likewise, unmeasured benefits may be misconstrued as value-shift when they are really value-creation.

This approach provides a useful starting point, as it offers a conceptual framework that is technology-independent. In philosophy, it links with the value-added concepts which underpin national accounting, while recognising that value in a modern economy is created by transactions across a range of players, rather than contained within traditional sectors of the economy.

University of Texas Internet Indicators

The Center for Research in Electronic Commerce (CREC)²⁵ based at the University of Texas at Austin Business School has taken a much more technology-specific route to identifying where value is created in the modern economy. The CREC framework presupposes that the Internet is the catalyst for structural changes in the economy that have fundamentally changed the ways in which value is created.

This framework uses the University of Texas Internet Indicators²⁶ to decompose the Internet economy into four layers:

Layer 1: Internet Infrastructure
Including backbone service provision, service providers, network equipment providers, conduit manufacturers, server and client hardware;
Layer 2: Applications
Including Internet consultants, Internet commerce applications providers, multimedia applications, web development software, search engine software, online training, web-enabled databases, network operating systems, web hosting and support services, transaction processing companies;
Layer 3: Intermediaries
Including market-makers in vertical industries, online travel agencies, online brokerages, content aggregators, portals/content providers, internet advertising brokers, online advertising, web-based virtual malls;
Layer 4: Online Transactions
Including e-tailers selling over the web, manufacturers selling products direct (e.g. computers, software), transportation and ticketing service providers, online entertainment and professional services, shipping services.

This approach uses a functional basis to segment the value-creation activities of an Internet-dominated economy, rather than the product basis used in traditional national accounting. While this framework recognises the pervasiveness of Internet technologies across all traditional sectors, the focus is limited by its concentration on one specific technology. It does not neatly acknowledge the fact that Internet-based transactions form only a part of an entire production process for most transactors, and must be analytically integrated into a much wider measurement process. For example, it does not recognise the value created using other electronic technologies such as EDI and stand-alone computers, which form by far the greater volume of transactions currently undertaken (Galbi (2000))²⁷. Neither does it explicitly recognise that the Internet serves as only one of the "pipes" along which information is communicated, as it ignores the role of other networks such as person-to-person, observation and voice communications technologies in the information transmission process.

Inasmuch as it presupposes a technology that enables ubiquitous, free form network-based exchange of information between transacting entities, however, the CREC framework does embody an approach to how performance may be categorised for measurement in other such network technologies, such as person-to-person interaction.

Barua, Whinston and Yin

The shortcomings of the CREC indicators with respect to integrating the performance indicators of the Internet economy with traditional performance measures have been subsequently addressed by three University of Texas researchers. Barua, Whinston and Yin²⁸ define five categories of the "modern economy" and integrate Internet and non-Internet statistics using an import/export approach into a set of overall performance measures.

Category 1:
Pure digital-products businesses that offer content, knowledge or services directly over the Internet;
Category 2:
Internet-based companies that deal with physical products, importing goods to be sold from the physical economy;
Category 3:
Traditional businesses that sell some of their products or services directly over the Internet;
Category 4:
Content developers, Internet service providers, web and applications hosting services; and
Category 5:
Companies that do not sell directly over the Internet.

While this framework recognises the need to integrate "new" segments of the economy with existing ones, it is again limited in scope to addressing only transactions taking place over the Internet. Further, the emphasis is on companies that sell and distribute over the Internet, thus ignoring use of the Internet by producers who assemble crucial information inputs using the Internet, but then produce and sell their products via other channels. The focus on "technology" (the Internet) and "commerce" (buying and selling) has resulted in the almost total neglect from consideration of pure information products constructed using the Internet as an information transmission mechanism, electronic (but non-Internet production technologies) but sold using traditional methods and delivered in non-digital form.

For example, this research report was assembled using word-processing software and image-scanning equipment (information-processing technologies) from information obtained from a wide variety of public (and hence free to the user and neither bought nor sold) and proprietary (some of which a subscription was paid for off-line) sources obtained almost exclusively over the Internet. Other physical sources were located and ordered from the University library using the internal local area network (LAN) (internally accounted-for in a budget charge). Without all of these information inputs, the quality of the report would have been substantially less. An electronic version of this report will be produced along with a physical hard copy, and will be transmitted to the client on email via the Internet. Payment will take place by cheque or possibly using a direct debit and credit between bank accounts via the proprietary bank clearing system. Yet none of this Internet-enabled value-added would be recorded as attributable to the Internet under this methodology as the buying, selling and value exchange processes occur off-line, and all other uses would be cancelled out in classical productivity methodologies as intermediary components of the final value-adding process. Furthermore, other electronically-generated value, such as the use of the LAN, word-processing software and the central banking network, classed as electronic commerce using the OECD methodology, is ignored. By viewing information as the key input resource, and electronic technologies as enablers of transmitting and processing that information (including human processing technology - the authors' intellectual capital and labour), information-generated value is recognised and the role of technologies viewed impartially as a means and not an end in itself.

Other Management Consultancy Approaches

Given that the most appropriate analytical framework of the three examined thus far is that of the Boston Consulting Group, it is apposite to evaluate the frameworks of other consultancies. Specifically relevant to the New Zealand environment are those of Deloitte Research and PricewaterhouseCoopers, both of whom regularly poll the New Zealand marketplace for electronic commerce indicators.

We note that the focus of these consultancy reports is firmly grounded in business needs and applications. Thus, they do not (and cannot be expected to) give consideration to either the consumer/citizen perspective, or the social consequences resulting from changes in the usage of technologies.

Deloitte Research

The Deloitte Research framework is based upon the 1999 document The New Economics of Transactions²⁹. This framework appears to take a technology-independent approach, by identifying that all firms are attempting to maximise profits subject to their production technology. Gains in productivity are attributable to new ways of gathering information, reducing costs, gaining new customers, customer service improvements and sales force automation. The role of information in this is recognised, with a need for both passive and interactive information to support the production process. It also acknowledges the demand side of commercial activities by recognising that customers attempt to maximise satisfaction subject to their available budgets. Customer service increases, cost reductions are passed on and new product choices are specifically identified as benefits available to customers as a result of changes brought about by changes in information technologies.

The Deloittes analysis sets much store on joint outcomes yielded by producers and consumers working co-operatively and interactively to achieve transaction cost reductions, create new intermediaries and new (digital) products and services. Mass customisation, markets for trust and cyberisation of markets are cited as examples of how this is being achieved as a result of new technologies. However, it does not propose a framework for measuring jointly-created products.

Yet once again, the scope of this analysis is narrowed by the almost total emphasis in data collection and analysis upon Internet technologies. Despite the sound, technologically-independent and information-focussed economic foundation laid in The New Economics of Transactions, this document still states clearly:

"The message to all concerned with systems integration and enterprise-wide software solutions should be clear: the "I" in "IT" means Internet" (p5)

and this focus is reflected in the statistics collected in the annual survey conducted by this consultancy of New Zealand businesses (Deloitte e-Business Survey: Insights and Issues facing New Zealand Businesses)³⁰.

PricewaterhouseCoopers

The approach by PricewaterhouseCoopers is embodied in the 1999 document Electronic Business Outlook for the New Millenium³¹. As the title suggests, the focus again is on electronic technologies in isolation from the concomitant human technology and process developments, and focuses more upon strategic issues resulting from technological changes rather than quantifying the consequences or apportioning responsibility for them. Thus, it identifies outcomes of changes to information generation, transmission and use rather than causes of value changes. The key outcomes identified include the ability for firms to:

act strategically to leverage a superior position in their market;
increase global sales opportunities;
strengthen relationships with customers;
streamline supply chain management;
enhance operational efficiency;
reduce transactional and overhead costs; and
optimise human resource utilisation.

While there is acknowledgement within the report that this will require a restructuring of processes within firms in order to capture the full benefits offered by new technologies, the emphasis appears to focus on adapting firm activities to meet the requirements dictated by technological considerations, rather than starting from an information-based needs assessment or new product development which the use of specific technologies (both electronic and non-electronic) would enhance.

A Synthesised Approach

While we recognise that the fundamental issue underpinning the interpretation of electronic commerce performance statistics is a thorough understanding of the ways in which new technologies (both electronic and non-electronic) impact upon information use in value-creation, we also acknowledge that, appropriately interpreted, the statistics currently being collected do contain relevant pieces in the jigsaw puzzle of determining the impacts of specific new technologies on wider economic performance. We contend that an information-centric approach to analysis rather than a technology-centric approach allows additional information to be revealed.

Firstly, we recognise that electronic technologies have radically changed the costs and ways in which information is created, transmitted, accessed, stored, processed, utilised and communicated. Changes in the ways in which economic value is created are consequences principally of these changes in the use of information. The economic extent of these changes is significant and warrants informed measuring. But the techniques for analysing these changes need not be new or radical.

Although electronic technologies have been vehicles for recently-observed changes (for example, computerisation and the Internet have changed the form and vastly increased the quantity of information a firm, such as an insurance company, can maintain about its clients, the ease with which it can be accessed and the range of new products which can be created utilising this information), and the pricing of these technologies has enabled them to become widespread, the economic precedents for analysing such a change are not new. Similar changes to the form in which information was recorded, stored, reproduced, utilised and communicated occurred with the introduction of the printing press, and even earlier when written words replaced verbal exchange as an external immutable form of recording. Changes resulting from increased speed, frequency and reliability of information transfer occurred following the introduction of steamships³², a technology that simultaneously enabled the development of new marketplaces and demand for and supply of the information necessary to support them (for example, truly international trade and the emergence of international currencies). And in this framework, while the advent of the telephone and broadcast television have resulted in an increase in reach and timeliness for person-to-person voice and visual communication, the network effects accrued from additional connectivity were not new, but rather more widespread manifestations of network effects that had already existed from person-to-person connectivity and information communication in one physical location (e.g. communities of interest, sharing stories around the campfire).

Thus, the fundamental issue in understanding observed economic behaviours becomes the recognition that information can be analysed as an input and product of the production process. In this way, information can be modelled as a capital stock, much like capital in the traditional productivity modelling process (albeit one with different returns to scale). While it may not always be quantifiable, conceptually it is still an input whose existence must be recognised. Changing the ways in which information is used will change the ways in which other stocks are used (substitution of traditional capital and labour - particularly the information components which have been "tied" to (embodied in) human vehicles in the past), and other raw materials embodied in capital stocks (e.g. land, human capital) are employed.

This approach leads to the perspective that a change from a traditional economy based upon capital and labour, to a "modern economy" where information is recognised as a key resource in addition to capital and labour, requires analysis of how technologies of all descriptions are successively enabling information to be substituted for other inputs. Further, it is recognised that in almost all production processes, information has always been an important component - it just has not been explicitly acknowledged, recognised or measured due to its intangible properties.

For example, information is required to monitor and manage any production process. Changes in the costs of acquiring and processing this information have fundamental effects upon the efficiency of a production process, as can changes in the types of information used. These have traditionally been bundled into transaction costs. Yet new technologies are enabling specific information previously embodied in manual processes (e.g. observation) to be extracted out electronically, and stored separately from its human (or other) vehicle for future analysis in order to improve production processes (e.g. sourcing raw materials from a different source on the basis of detected error rates from given sources). This process replicates the efficiency gains yielded when the development of written words enabled separation of information from the vehicle (human) that carried it. And just as we have been able in the past to place values on information recorded on paper (e.g. value of information in books via copyrights and in processes via patents) due to their separation from the original bearer/creator, we can now start placing values on information captured and stored in electronic form as it becomes separated from the individual/process that bore/created it. While these values may as yet be uncertain or hard to verify, their existence is becoming harder to ignore as electronic technologies make the application of information-separating and revealing processes far more common and widespread.

This approach leads to a conceptual framework founded on the economic principles such as those enunciated in the BCG report. But instead of focusing on electronic technologies as capital stocks, investment in which yields productivity improvements by substituting capital for labour, the focus is on information as an input to and output of value-creating processes. The substitution activity and values of information must be incorporated in any productivity analysis, if it is to validly represent changes in levels of outputs for given levels of inputs. For if an electronic-based process substitutes for a physical one of equal costs, and the information outputs include information not available previously, an efficiency gain (more outputs for given inputs) has been achieved. But if the same substitution results in information previously available no longer being created (and that information is necessary to a subsequent process), then this must be considered in the assessment of whether the substitution has resulted in a real productivity improvement. In a cost-based model, such factors have seldom been considered directly (although we acknowledge that their effects are measurable subsequently in aggregated statistics). It is our contention that in an information-based economy, they are fundamental, and must be analysed at a process (or transaction) level.

Applying this methodology to the emergence of digital technologies and electronic commerce processes, two key information-based substitution activities emerge:

"digitisation"³³ of existing products, processes and procedures etc. for the purpose of transaction cost reduction
creation of new (innovative) products and services based upon digital processing formats in order to capture value not available previously³⁴.

"Digitisation" of existing products, processes and procedures substitutes new technologies and product forms for old. If the substitution results in replication of the same tasks and functionality that have always existed, then standard analysis techniques ought to effectively capture productivity outcomes and consequences. Real benefits will accrue as lowered transaction costs, with gains and losses accrued across groups at different levels of aggregation. However, it is only by analysing the information outcomes of each individual substitution that it can be adequately assessed if any specific information gain or loss has occurred. Aggregation may indicate if a gain or loss has occurred across firms, industries or sectors (by net productivity gain/loss), but only a detailed information analysis will identify the sources of those losses or gains due to information form or process substitution at the decision-making level. Replacing physical marketplaces with electronic B2B ones, manual supply chain management processes with electronically-linked ones, landlines with mobile phones, EDI applications with Internet-based ones and e-tailing of existing products are examples which all contain elements of significant "digital" substitution with downstream information consequences, the extent of which has potentially been evident in productivity figures, but for which explanations of aberrant or unexpected behaviour³⁵ have been elusive³⁶.

By contrast, creation of new products and services using electronic technologies to capture value not previously available is a simple new product generation process. Substitution of new electronic products for old, and demand for new electronic products is analytically no different to any other consumer substitution process. Products and services that offer improved benefits to consumers will prevail and replace those of lower value. Information value will form part of this assessment process (either explicitly or implicitly), but it is recognised that failure to adequately analyse the information consequences of this substitution process may result in unwise investments, and distortions in aggregate outcomes, just as when information consequences are not recognised in producer substitution choices.

Separation of these two issues thus provides clarity on whether gains and/or losses are from productivity improvements within existing production processes, or completely new "value creation propositions" enabled by new technologies. The framework also "demystifies" the extent to which electronic technologies are reportedly changing the economic basis for decision-making³⁷. Likening it to the change from steam-powered production to electric-powered production: the new product was electricity, the new machines that lowered production costs also increased productivity but did not fundamentally change the objectives of the production process - the product was unchanged. But it did create new industries - for example, electric machinery production, and electricity generating - and decimated others - steam engine making, coal mining. Changes in the information generation and utilisation process accompanied all of these changes, leading to delays in the accrual of net benefits across industries and societies as new information supplies and processes were developed (Greenwood and Yorukoglu (1997)), but were analytically unaccounted-for.

The key distinction between the introduction of electronic information technologies and the introduction of other technologies such as electricity is that the product information technologies produce and process - information - has not previously been analytically separated as a production input from the vehicles that embody it, in ways that other products such as motive power, have been. Whereas technology introductions such as that of electricity generated information that could be analysed separately from the technology itself, information technologies generate both information products and information to analyse the introduction of the technologies that are difficult to separate. The similarity of form of both outputs has clouded the analytic effort. Have we been confused by the fact that the very technology, which by its creation has enabled us to separate out the information effects of all technologies, is requiring us to analyse its impact first?

While this distinction challenges some of the models that have been used to analyse economies, such as the aggregations over which economic activity has traditionally been analysed, the concepts of production are essentially unchanged. Changes can, we believe, be accommodated by recognising that previously embodied and hard-to-measure information is becoming separate and measurable. The growing ubiquity of specific electronic technologies merely draws emphasis to the fact that this analytical challenge must be addressed.

Drawing this philosophical underpinning into analysis necessitates recognition that the "modern economy" is not a "new" economy, but an economy in transition. A new technology is enabling measurement of an asset of value which was previously not explicitly valued. While information is probably potentially a more valuable asset than others we have witnessed (such as electricity) and has different inherent properties (e.g. reusability), there is considerable understanding to be gained from measuring:

the share of new products and processes using information differently;
substitution between new and old products and processes using information; and
measuring the losses resulting from redundant and discarded products and processes.

To reap the full benefit, however, this understanding must be technology-neutral, as all technological changes have information consequences. Monitoring electronic technologies in isolation in order to determine indicators of relative performance ignores the flow-on informational consequences into other processes. Tracing information product flows as we hav e traditionally traced physical product flows allows the analytical links to be drawn between firm and technology-specific measures such as those of the University of Texas, and the wider economic indicators identified by Haltiwanger and Jarmin. Interpreting the measures in this context is, we believe, where real learning and understanding will come.

What Do We Want to, or Need to, Measure?

All of the literature-based measurement taxonomies above identify statistics that offer insights into how electronic technologies are influencing economic performance. However, we propose aligning them into a framework that facilitates translation of measurements and observations of what is being done into an analysis of economic performance indicators in an integrated "modern" economy where information is a key element.

From this point, we will limit our consideration to electronic technologies and information that emerges from them, but recognise that this same taxonomy can be used to analyse other information-based technology changes. Our taxonomy addresses four key questions:

What digital products/processes are we using and where are we using them?
What can we reasonably expect we could be doing with the resources available?
What are we using digital products/processes for? (What information products/processes are we utilising, are they new or substitute products/processes?)
How well/effectively are we using these digital products/processes?

from which we derive four key classifications under which we can collect and analyse the statistics collected by all of the other frameworks identified above:

CONNECTIVITY;
CAPABILITY;
UPTAKE; and
PERFORMANCE.

Connectivity

Connectivity recognises the extent to which communication of information can occur over networks. Network externalities recognise that in most networks, connecting one more node adds not only the benefit to the connecting node of connecting with all other nodes already on the network, but that all other nodes also gain benefit from the ability to connect with the node that has just joined. For most networks, if adding more nodes generates proportionately more benefit than that acquired by the joiner, then the more nodes (or more people, businesses etc.) connected, the greater the value for all concerned (although it is recognised that technological effects and costs may successively limit the size of the additional benefit per node as the network grows).

As most communication infrastructures (including person-to-person contact) are networks, and the purpose of all communication infrastructures is to facilitate the exchange of information, then up to the limits of size for efficient operation of the network, the more nodes connected, the better. The more individuals in a personal interest network (such as a stamp-swapping club), the better, up to the point at which the costs of maintaining the network (e.g. costs of keeping all members apprised of contact information, inventories etc.) outweigh the collective benefits.

Electronic communication infrastructures such as digital telephone networks enable interpersonal interconnectivity, and act as a conduit for interpersonal communication by providing a vehicle for connecting computers, and hence the people using them, via the Internet. Thus growth in measures such as telephone and Internet connections indicate growing value of these networks, as they indicate growth in the numbers of people who have the potential exchange information using them. Knowledge not only of how many people are connected, but where these people are connected from, provides information of where the benefits of use may accrue.

However, connectivity alone indicates only potential to accrue benefits due to the presence of the infrastructures that enable connectivity. It is from actual use that these networks result in benefits being accrued. For communication networks to yield benefits, information must be communicated (exchanged). This will only occur if the benefits of information exchange outweigh the costs (both tangible and intangible). When one communication technology begins substituting for another, then there will be direct costs and loss of network externality effects as the substituted technology loses nodes. If the substitution is complete, then value (both individual and externality benefits) will be transferred; however if the substitution is only partial, and some information transfer functionality is lost as a consequence (for example, the telephone substitutes for the voice communication aspect of person-to-person contact, but does not substitute for exchange of information contained in visual signals), then there may be compensating costs incurred to obtain the necessary information lost by the substitution process.

Thus, more connectivity to new networks is better only if the net value generated exceeds that of the networks that are replaced (including the costs of lost benefits). This is dependent upon both functionality and frequency of use. For information exchange networks, this requires an understanding of the volume, direction and form of the information flows that these networks support. Hence, connectivity measures must be examined in conjunction with uptake figures in order to understand how these substitution processes translate into wider economic performance measures. The analytical link is usage value of the information actually transferred.

Connectivity measures in the existing methodologies (e.g. OECD, NOIE, University of Texas) have focused almost exclusively upon person-to-person connectivity to new communication infrastructures³⁸ - that is, the telephone network and the Internet, as these are seen as the electronic facilitators in the "modern economy". These have been analysed by many demographic and business group categories, including:

geography;
firm type (individual, small, medium, large, government);
business/industry sector;
income-related characteristics of users; and
skill characteristics of users.

However, connectivity indicators must also recognise the potential of person-to-technology and technology-to-technology networks that have also been developed for the exchange of information. These uses are surfaced by asking questions such as:

what other electronic information processing technologies are being used?
where are they being used? and
who is using them?

Person-to-machine technologies include electronic calculators and stand-alone computers used to exchange information. Stand-alone computers may also be linked to other computers via internal networks such as LANs, terminals may link to mainframe computers, computer links between firms may bypass telephone and Internet connections by using proprietary cabling, all with consequent network effects generated by information being transferred across them and made available to a wider number of recipient people and processes. In the context of an "information-based" economy, these technologies at the current point in time are used far more frequently, and for much larger volumes of data, than voice telephony or the Internet. Comprehensive connectivity figures require that these statistics be collected.

Similarly, integration of computer-controlled equipment into utilising processes also represents a class of information exchange connectivity that for completeness must also be measured. Automatic teller and electronic funds transfer machines, bar-coding systems, "smart" domestic appliances such as fridges and washing machines and computer-controlled production equipment all collect information and via connectivity mechanisms store, process and transmit it for subsequent use. It is the integration of these sorts of information connectivity mechanisms within the business, and between businesses, that reports such as those undertaken by Deloittes and PricewaterhouseCoopers identify as imperative if the benefits of technologies such as supply chain management are to yield their potential. The growing use of such technologies in domestic environments means that analysis of integration of information into other processes should not be limited to the commercial environment.

Any analysis of connectivity from an information perspective is incomplete, however, if it does not also include measures of declining networks (the substituted ones) in addition to the growing ones. For it is in measuring both in tandem that confirmation of information substitution effects can be analysed, and greater understanding of incentives and impediments to usage gained. For example, if the assumption that Internet connectivity is the only connectivity metric important for future information exchange is correct, then a decline in EDI utilisation should have been witnessed. That EDI utilisation has not declined (although its rate of growth has slowed) and that it is still by far the largest medium of business-to-business data exchange should have alerted the designers of these methodologies to the fact that some of their key assumptions of the substitutability of the Internet for EDI may have been flawed. While the Internet enables cheaper interconnectivity than EDI for many businesses, it does not yet ensure complete security of data. Many firms still prefer to maintain private networks to ensure control, security and confidentiality, as their data will never be "in the public domain" (Galbi (2000)), while the lack of certainty in maximum information transmission time renders it unreliable for utilisation by firms for which information timeliness is mission-critical (Lehr and McKnight (2000)). The Internet is thus not the functional equivalent of EDI for these firms, and observed EDI and Internet uptake figures confirm this³⁹. Similar partial substitution arguments also underpin the decline in the number of bank branches resulting from the growth of connectivity to EFTPOS and ATM technologies (The State of E-New Zealand), and the partial substitution of websites for email by firms offering free-phone and centralised telephone ordering services (The Rural-Urban Digital Divide).

Broadcast media must also be considered in any analysis of connectivity, due to changes in formats of distribution (digital media, cable television etc.) resulting from the development of electronic technologies. Connectivity to these networks (albeit one-way transmission) also represents a dimension of information transmission that warrants consideration.

Thus, connectivity is an important classification of statistics, but all connectivity statistics must be interpreted in conjunction with capability, uptake and performance indicators if their full importance is to be realised. Furthermore, care must be taken, as it cannot be assumed even in light of this joint analysis that the same incentives and impediments to connectivity and uptake are relevant explanators for observations in different categories of application, even for the same benchmark. It is utilisation of the underlying information that is the definitive explanatory.

Capability

While measures of connectivity provide an indication of which electronic information communication technologies individuals and businesses have the capacity to communicate with, direct comparisons of levels of connectivity need to be interpreted not in isolation, but in conjunction with measures of capability to both become connected and to utilise that connectivity. Thus, connectivity and capability together should provide indications of the level of uptake (utilisation) of specific information generating and processing technologies.

Measures of capability generally indicate resource levels available that enable an entity to take advantage of the opportunities new technologies offer. However, like connectivity measures, capabilities measure potential to yield value rather than actual realisation. Actual realisation comes when capability, connectivity and need coincide, resulting in uptake - that is, utilisation - from which value is generated. Thus, connectivity and capability measures predominate in the Readiness category of the OECD methodology. The limited focus on these measures in all of the other taxonomies, given that they focus upon firm, industry or technology indicators, means that wider economic capability measures are absent, and even within-firm capabilities are given scant consideration.

The starting point for capability assessment thus is the level of resources available. Traditionally, these have been based upon measures of physical and human capital. Physical capital has generally been measured as either depreciated dollars invested in capital plant, equipment and inventories, hence the emphasis given to measuring numbers of specific items of equipment (e.g. computers, mobile telephone handsets) or investment in computers and communications equipment in order to "trap" capital investment resulting from electronic technologies. Comparison of these measures over time have been problematical, given the difficulties of adequately measuring the impact of rapid quality improvements and price reductions in computers (Lawrence and Diewert (1999)) and the much more frequent replacement cycle of computer equipment (generally three years) compared to other longer-lasting capital stock (average depreciation of ten years). In addition, emphasis on measuring physical equipment has been to the exclusion of measuring other intangible stocks associated with their use, such as software and information stocks, and human capital investments such as education and training of people using the equipment (Brynjolfsson and Hitt (2000)).

Human capital measures, due to their intangible nature and the difficulties in separating ownership of the capital from the individual bearing it, have typically been population and industry sector-based rather than firm-based, and focus upon demographic characteristics, such as occupation class and level of education (Mankiw, Romer and Weill (1992)) rather than dollars (or other resources) invested. As the associations between core demographic details and information transformation potential and ability are few, and those generally proven reliant upon difficult-to-measure characteristics, the emphasis given to the few tangible indicators is understandable.

Emergence of new technologies is, however, enabling the extraction and separate storage of much information previously considered embedded in human capital. For example, data warehousing systems are largely justified by extracting industry or firm-specific information from individuals and storing it electronically in order to minimise the vulnerability of firms to information loss due to staff turnover. Such substitution of information capital for human capital results from the fundamental conceptualisation of information as the key resource of the "modern economy". Similarly, if human beings can be thought of as "highly specialised information processing technologies" yielding different returns on information inputs than computerised information processing technologies, then the comparative effectiveness of computers-for-human substitutions can be measured on a functional as well as a financial basis.

Capability measures, therefore, need to address far more than just financial investment, population educational levels and preparedness or capacity to use new technologies, if they are to address the needs of a "modern economy". Much more needs to be understood about the nature of the information processing tasks individuals undertake, in both work and other capacities, in order to ensure that the match of technologies with needs is appropriate, and that the appropriate information resources are available as inputs to the processes that technologies of all sorts are purchased or developed to undertake. Thus, clear separation is required between counts of and the skill, knowledge and information input levels required of:

producers, creators and maintainers of electronic technologies;
users of electronic technologies; and
consumers and producers of information products.

Existing "e-statistics" focus principally on the first two. While these are vital for the development and uptake of new technologies, they do not address the fact that, in economies increasingly dominated by service provision, utilisation of information is the primary generator of value (e.g. Dunt and Harper (2001)⁴⁰). Technologies facilitate, but value is created by the utilisation of information. Developing measures of human capabilities (by volume and category) to extract this value are paramount. NOIE estimates by 2005, the most information-intensive sectors for electronic data transmission by input and output will be health and education (Bandwidth Inquiry). It is our contention that they are already the most information-intensive, as their inputs and outputs are almost exclusively information transfers. All that is happening is successively more of the transfers are being captured electronically.

Human capital capacity statistics thus need to measure the "knowledge worker" concept - the worker who takes information inputs and creates more valuable information outputs by processing that information with human technology⁴¹. Skill and capability measurement should also reflect the fact that not all knowledge workers will source their inputs or generate and transmit their outputs electronically. Measures of capability must be considered relative to the processing technologies utilised. Likewise, consideration must also be given to developing and measuring the potential of the "knowledge consumer". Statistics gathered should support information bases for both commercial and social decision-making processes that influence the balance of supply and demand for information products, "knowledge workers" and "knowledge consumers".

Key issues for capability measures include, therefore, the need to record a wide range of human and physical capital resources at a variety of levels, including:

international;
firm;
individual; and
transaction

with the acknowledgement that new and more appropriate proxies may need to be developed to capture many of the as yet uncaptured intangible characteristics. The basis for these proxies should be founded not on technological capabilities, but information needs and processing capabilities. For example, while scientists and researchers are important for developing new technologies and opportunities, the vast bulk of economic activity is generated not by inventors, but by routine utilisers of the technologies developed by these researchers. Levels of pure information workers, such as teachers, doctors, nurses and business process researchers may be more crucial to the sustainability of an information-based economy as scientists inventing new electronic technologies⁴² and technicians implementing them. This justifies collection of human capital statistics based upon information users and their capabilities to lever value out of information (just as farmers lever value out of land and stock) as well as technology generators.

Further, changing associations between transactors may mean that traditional bases for measuring capabilities may be less valid in an information-based economy. Traditional industry associations based on product creation may be valid for physical products, but an additional level of aggregation based upon information transfer value may also be relevant. Thus, international aggregation based upon "communities of interest", form or industry activities, and aggregation based upon transaction type may also be informative.

We acknowledge also the need for measuring environmental capabilities - in particular the legal and economic frameworks to encourage value creation based upon information. However, comparative measures must focus upon relevant indicators: for example, frameworks that address issues such as property rights to information itself may become just as important as those that support the framework that legalises electronic transactions using that information. This again draws attention to the need to separate the policy and legislation frameworks supporting technologies (both human and electronic) from those supporting the information those technologies use and create.

Finally, capability statistics must be interpreted not in isolation, but in relation to the needs that they meet. While we have addressed the necessity of separating technology demands from information demands, it is also important that these reflect the processes that are being utilised, and are sensitive to the substitution processes identified in the Connectivity section above. While this is an ongoing concern for all performance measurement processes, this is reiterated with special reference to e-commerce performance measurement due to the pace at which substitution is occurring. Care must be taken to ensure that capability measures in fact measure preparedness and actual changes relative to population bases, rather than just comparative to specific technology users, especially where functional substitution is not complete. This is an issue of interpretation methodology insofar as many surveys are based upon voluntary participation by committed users of specific technologies, so therefore may not be truly reflective of wider population behaviours.

Uptake

While measures of Connectivity and Capability will indicate potential economic and social performance available from new technologies, it is in actual uptake that the extent of this potential is translated into subsequently measurable wider economic performance indicators. Uptake, therefore, must be measured in the context in which it is applied. This requires measurement of both the new products and processes being used, and those that they are replacing (substituting).

Again, the key contexts of new technology uptake in a modern economy should focus not solely on electronic technologies, but all technologies as they relate to changing uses of information. Measuring uptake of new and substitute human information processing procedures is as important as measuring uptake of electronic and mechanical ones. Further, while commercial measurement may focus upon measuring uptake of business-related processes based upon changed information utilisation within and between firms, such as outsourcing and producer-controlled inventory management, measuring uptake across non-business divides is also important. Such uptake includes consumer utilisation of information technologies such as satellite television broadcasting and "smart" domestic appliances, and information-based processes such as EFTPOS.

Uptake analysis requires not only measurement of product and process utilisation, but also the concomitant understanding of which end performance indicators can be expected to be impacted by the utilisation of specific technologies. Thus, Connectivity, Capability and Uptake measures in conjunction should provide the rationale for observed wider economic and social performance indicators - discussed in the Performance section. While much work has been done in the taxonomies discussed above to record specific technology uptake measures, as indicated by the OECD discussion document (Colecchia (1999)) much is still not understood about how these link ultimately to performance measures. This is an area where further research effort is indicated.

Uptake indicators measured by the studies identified above focus principally upon one or the other of business or social factors. For example, the NOIE Australian indicators measure business and individual statistics separately. Furthermore, uptake figures within business use the firm as the unit of analysis, or aggregate upward into industries. While these units of analysis conform to traditional aggregations based upon product and production flows, such analysis is relevant in an information-based economy only in respect of end products that are physical in nature.

The dimension of analysis made possible by changing information-processing technologies is the ability to track flows of newly-created information products. These may contribute to the creation of physical products, but may also transcend traditional aggregations. Tracking information flows independent from physical product flows, and the consequences of utilising the information passed in these flows, becomes an important dimension of value-generating activity if the information so transferred is applied separately from the physical product flow process. Recording "electronic commerce" activity by the value of goods bought and sold on the Internet may measure the information flow associated with movement of physical goods through a value chain, but it leaves unaccounted the value of all intermediate information flows required for all other processes undertaken by both those adding value to and those consuming the physical good. Comprehensive uptake figures necessitate measures of information flow activities independent of physical flows if the full impact of the "modern economy" is to be recognised, especially if changed purchasing habits result in value-shift activities. For example, the costs of information search are moved from the producer to the consumer when mailed brochures (passive consumer response) are replaced with a website, requiring active consumer response and greater investment of consumer time to locate the same information.

Full information flow uptake measurement requires analysis of information moving both within firms and between firms. While analyses such as those of the consulting firms poll the types of activities where electronic information transfer is occurring (e.g. email, website utilisation, electronic banking, purchasing, selling), these capture only a "doing it/not doing it" indicator. No uptake measures found in the course of this research indicated any measures within firms to assess the volume of data transferred to support individual activities. While NOIE measures the amount of information transferred (in terabytes) over the Internet in Australia, this cannot be allocated over specific industries or activities. While individual firms may have the capacity to record electronically-transferred data volumes over applications, currently there are few incentives for firms to collect, utilise or analyse this information, as information is not typically viewed as a product input or output. Yet it is at the transaction level that uptake of electronic information movement can now be monitored, at least on a bit-volume basis. Relative information-intensity of differing transactions can be measured, potentially leading to greater understanding of the relative information needs of different activities. Aggregating these transactions up into different levels theoretically leads to the ability to assess the information-efficiency of different transactions, individuals, firms, networks, industries, nations or other such divisions, and ultimately, relative valuations of information inputs, outputs and processing activities. This model creates a framework to conceptualise information and begin measuring information of all forms - both electronic and other - as input into the production process.

Analysis of electronic technology uptake figures needs also to recognise the distinction between adoption of technologies to assist traditional production and consumption functions (substitution), and adoption of those supporting information flows not previously possible. These flows need to be measured with respect to the changing social and business structures that are emerging as a consequence of new information flows. Just as uptake of previous technologies such as the steam ship spurred new business and societal forms (multi-national companies, colonies, empires) and associated requirements to measure uptake of these new processes, so does uptake of new forms of organisational and societal technologies need to be measured for the "modern economy"

New organisational forms have already emerged as a consequence of changing costs of information transfer, including networks, clusters, alliances, and supply chains. Comprehensive uptake statistics ought to measure not only the extent of creation of these new forms and the patterns in which they substitute for existing forms, but also the patterns of information ownership and transfer they stimulate. For example, while information transfer costs have lowered, enabling perhaps even greater levels of specialisation than previously (e.g. outsourcing), this has resulted in increased risk of network failure arising from breakdown of one link in the more extensive value chains created (Yang and Ng (2001)). If individual optimisation of firms entering into such arrangements does not include a valuation of this increased network cost (network externality) due to no one entity having an incentive to optimise the outcomes for the entire network, then the result may be inefficient levels of over-specialisation, with costly flow-on effects.

Mitigation of such risks is already becoming evident in the ownership of B2B marketplaces. An Economist Intelligence Unit report⁴³ evidences a rationalisation in ownership of these exchanges towards private ownership with domination by one large player (or cartel of players - e.g. Covisint, the US car parts exchange) or ownership by industry associations (e.g. the New Zealand wholesale electricity market). This enables network balancing of the reduced costs of information transfer with the increased risks incurred by moving to such structures. However, it also results in different incidence of costs, as network members must pay a premium for risk-shifting - such as the price of membership of the industry association, restrictions on the parties with whom an agreement can be reached, or lower prices paid to the seller by a dominant purchaser who controls the exchange, in recognition of the additional risk assumed by the buyer. The emergence of accreditation agencies for B2C transactors, with associated costs of accreditation being factored into the price of goods, and consumer preparedness to pay a premium for purchases from trusted branded suppliers, reflect the need for similar risk-managing strategies for individual consumers. Uptake figures thus need to capture measures of the growth of these new product forms as well as growth of equipment usage.

Technology uptake figures must also be interpreted in recognition of the fact that quantity and price are not independent, and that different relationships between price and uptake may prevail in different environments. Thus, caution must be applied when comparing uptake rates in different countries, regions, industries, ownership structures, etc. Further, technology uptakes may be influenced by the range of options available within different environments. While links between electronic infrastructure availability and uptake of specific technologies are well enunciated in the literature (e.g. NOIE (1999)), the need to also interpret technology uptake in light of policy and legal infrastructures is also increasingly evident. The role of unmetered telephony pricing in internet uptake in New Zealand is documented in OECD (2000) Local Access Pricing and E-Commerce, but increasingly other pricing structures (Varian (1999)), service quality (Rood (2000)) and service options (Lehr and McKnight (2000)) are implicated in the rate of broadband uptake. Trade patterns (The State of e-New Zealand) and geographic location (the Rural-Urban Digital Divide) can also influence uptake patterns. Similarly, legal and government policy can also influence statistics. While international standardisation may be helpful for comparisons of visible counts, caution is required when comparing analyses that do not take these environmental and contextual considerations into account.

Finally, it must be remembered that uptake figures, while useful historic measures, are not necessarily good predictors of future performance in environments of rapid technological change. Concurrent understanding of technological change at the business and societal level must be iteratively integrated with expectation and performance measurement to determine the predictive value (if any) of uptake measures⁴⁴. Until more is understood about these relationships, and in particular the network costs and benefits not necessarily accounted for in individual and firm decision-making, it is difficult to justify infrastructure and investment policy on the basis that higher uptake of any specific application is necessarily "better".

Performance

While measures of Connectivity and Capability indicate potential uses, and uptake actual utilisation of specific information technologies, it is only when linked through uses of information that their ultimate effects can be translated into wider measures of economic performance and social impact. Thus, it could be interpreted that Connectivity, Capability and Uptake represent lower-level transaction, firm and individual measures, while Performance focuses on the wider system outcomes. For example, Lawrence and Diewert (1999) use this approach in an attempt to assess New Zealand's economic performance as a result of the structural changes occurring as a result of successive deregulation during the 1980s and early 1990s. If changes brought about by the introduction of electronic technologies could have been expected to result in significant changes in economic outcomes, then similar methodologies should be able to be applied to measure the resultant changes in performance.

Theoretically, productivity remains arguably the most comprehensive indicator currently available to measure cumulative performance of an economy as it enables the end products of multiple interacting systems (such as within and between firms and individuals) to be aggregated together to create a series of economy-wide snapshots enabling comparative assessment of economic performance over time. Depending upon where one chooses to draw the boundaries, productivity can be measured at a variety of levels: division, firm, industry, country for example. Furthermore, productivity can be measured using a variety of output measures, such as dollars of value-added, or units of specified output per units of specified inputs (for example units per man-hour).

However, reliable productivity measurement requires constant comparability of inputs and outputs over the time periods being analysed. Changes in both price and quality of inputs and outputs require constant reassessment of comparability (price indexing) to maintain accurate productivity measurement. When factors are changing rapidly (such as in the computer technology industry), the factors that are changing are intangible (e.g. information, service quality), or changes are occurring in sectors of the economy that are not routinely measured (e.g. the unpaid and self-production sectors), the reliability of economy-wide productivity measurement declines. The effectiveness of productivity measurement is further limited by the extent to which the composition of most first-world economies are moving from dominance of physical product manufacture, where inputs and outputs are physical, and therefore it is easier to measure and assess their quantity and quality, to dominance of service production, where inputs and outputs are almost always intangible and difficult to measure and assess (Triplett (2000)).

Ordinarily, one would expect that improvements in productivity would result in either increased output for given inputs, or the same level of outputs for less input. However, if benefits are accrued as intangibles, while costs expended on physical products remain unchanged, productivity will be unaffected, despite net gain to consumers. Haltiwanger and Jarmin (1999), for example, identify many gains accruing to customers of the US banking industry, such as 24 hour banking via telephone, Internet and ATMs, which have occurred at the same time as the banking sector has recorded a net drop in productivity, despite price index adjustments for computer technology having been made. As a New Zealand example, productivity measures resulting from changes in the telecommunications industry are incomplete as they do not include as tangible benefits the reduction in customer in waiting time for the installation of a telephone from several weeks in the 1980s to less than 24 hours in 2001.

Using productivity measures to assess the performance of a "modern economy" is fraught with difficulty because the product whose utilisation is driving the fundamental changes - information - is intangible. Thus, productivity measurers have resorted, not unexpectedly, to measuring the one tangible thing that does exist in this economic sea-change - connectivity to, capability to utilise, and uptake of, electronic information processing equipment. That this has not translated into productivity figures (Solow: "You can see the computer age everywhere except in the productivity statistics") is not surprising if the net result has been an increasing demand for new information-based intangible products and substitution of information inputs and outputs for physical ones. If, at the same time computer-controlled physical product creation processes have undergone real productivity improvements, leaving more disposable income available for the purchase of intangible products (that is, the rising demand for service products), then the so-called "productivity paradox" will in fact be self-reinforcing.

While we acknowledge the difficulties of measuring intangible inputs and outputs such as information and product quality, we contend that unless analysis can be expanded to recognise the role and value of these information-based intangibles, the figures derived from such analysis will become, at best, increasingly less relevant, and at worst, erroneous and misleading. Furthermore, performance targets set using such measures⁴⁵ may result in disproportionately high investment in technologies that can be physically observed at the expense of potentially more efficient investment in technologies that cannot. In the spirit of the incentive approach, to paraphrase Holmstrom and Milgrom (1991), what gets measured will be done, and in the cycle of immeasurability, what gets seen to be done gets measured, to the exclusion of that which cannot. This is exacerbated when network effects of information investment influence outcomes of actors whose behaviour is both unobservable and exempt from the decision-making process - such as the zero-sum cost-shifting (or at worst cost-incurring) activities identified in the BCG report.

If quantitative performance measurement cannot accurately measure all factors, then at least some indicative analysis is required to determine the direction of net impacts brought about by changes in the ways in which intangible inputs and outputs are treated. Thus, it is our recommendation that analytical effort be expanded to investigate the ways in which information flows and uses are changing, even though to date the effects may not be perfectly quantifiable. For it is only in this sort of analysis that the extent of changes can be traced, and new and better proxies can be found for the intangible items. This must also include the currently unmeasured non-traded (that is, self-production) sector if all economic and social performance measures relevant in the information-based "modern economy" are to be captured.

Key questions in defining the extent of this analysis must include the levels of aggregation to be considered. This entails setting the "boundaries" of systems that interact. This is an issue not just for government and social statisticians, but also for individuals, businesses and other transacting units. For example, which businesses in B2B chains will be linked for the purpose of such analyses? Who has the incentive to undertake timely statistical collection and analysis where ownership is distributed and membership of networks and collectives is unstable? And how widely will the results be communicated? Who will take responsibility for collecting of information relating to B2C and C2C exchanges, when incidental or non-standard exchanges occur? It is only the inclusion of all levels of activity that will enable the total consequences of the move to an "information society" to be recognised. Daunting though the task may appear, it is not necessarily insurmountable. Work by Kaplan and Norton (1996) on Balanced Scorecard methodologies for business measurement of intangibles has shown that well-designed proxy measures can be incorporated into wider performance measurement frameworks. Further research and understanding of how information is exchanged should ultimately result in successively improved approximations.

¹⁶Anecdotes and recent informal discussions with employees of two other national performance measurement agencies carry a common theme: that while preparedness indicators and technology uptake figures are useful measures of progress along a time continuum for the specific things identified as useful proxies, there is less certainty about what these figures actually mean in terms of longer-term economic and social benefit. Both of these agencies indicated that further research is warranted to develop a greater understanding of these relationships.

¹⁷"Thus, it is possible to conclude that we have succumbed to a `microchip myth' - that the `technology' that leads to productivity growth is embodied in the machinery that knowledge and learning has created. In succumbing to this myth, we appear to have lost sight of the ideals enunciated in Solow's model: that technology relates to "the application of practical or mechanical sciences to industry or commerce", the "methods, theory and practices governing such application", and thus is an embodiment of "the total knowledge and skills available to any human society". By focussing almost exclusively upon the investment process, we have been less cognisant of the fact that computers, while indisputably a technological development yielding productivity gains from the substitution of capital for labour (the capital deepening identified in Jorgenson and Stiroh (2000), Triplett (2000) etc.) have also been fundamental in changing many of the human and organisational processes on which our productivity functions are based. And although some research has been undertaken in this area (notably Brynjolfsson and Hitt (2000)), this work still focuses principally upon the need for complementary capital investments in human and organisational systems alongside computer purchases - that is, capital deepening from which the `manna from heaven' will emerge.

Romer (1996) ascribes this phenomenon to the fact that, once analysis moves beyond the split between technology and all other inputs (physical capital, human capital and labour), "technology" in the model "does not correspond to anything in the real world. You can understand capital in terms of things like machine tools that you can observe, but if you ask for a description of technology, neoclassical theory will tell you only about things that live in models, shifting production possibility frontiers and the like." (Romer (1996): 5). While embodied technological change in the 1960s implicitly modelled technology as designs for machines, and Nelson and Winter (1982) represented technology as routines within firms, recent generations of neoclassical growth theorists have, according to Romer, "contented themselves with a `force locomotif' explanation: `Technological change causes economic growth'." Thus, it is not surprising to find that, in the interface between productivity models and the real world effects these models seek to explain, analytic emphasis has been placed upon measuring returns to the visible fruits of technological endeavour - machines and processes (e.g. computers and Just-in-Time inventory systems) - over the less visible (e.g. improvements in service quality or reduction in risks faced)."

Extract from Howell (2001) Information, Technology, and the Productivity Paradox - ISCR Working Paper, forthcoming.

¹⁸Haltiwanger and Jarmin (1999).

¹⁹That is, productivity improvements as per the classic productivity function analysis - see (for example) Mankiw, Romer and Weill (1992).

²⁰For example, Triplett (2000) and the work of the Brookings Institute project on Service Sector measurement.

²¹We note here the work of Griliches and others on hedonic pricing methodologies as a technique to capture product quality as a component of productivity improvement.

²²Haltiwanger and Jarmin (1999) cite the example of the banking industry, where increased service levels from Automated Teller Machines (ATMs) and electronic banking have increased the service offered to customers from 6 hours a day 5 days a week to 24 hours a day 7 days a week. But the value of this is not recorded in any official statistics, thus resulting in significant understatement of productivity increases in banking resulting from investments in IT.

²³Brynjolfsson and Hitt (2000).

²⁴Boston Consulting Group (2000) After the Land Grab: B2B E-Commerce in Australia and New Zealand

²⁵Sponsored by Cisco Systems

²⁶University of Texas (2001) Measuring the Internet Economy

²⁷While the Internet is the fastest growing communications technology (313% growth per annum in bandwidth used for Internet applications), it is still dwarfed by the huge use of bandwidth for direct inter-office EDI (97.9% of all bandwidth deployed in 1999).

²⁸Barua, Whinston and Yin (1999) Not all Dot Coms are Created Equal: an Investigation of the Productivity of Internet-Based Companies.

²⁹Deloitte Research (1999). The New Economics of Transactions: Evolution of unique e-business Internet market spaces

³⁰However, it is noted that research manager Alasdair MacLeod has acknowledged that New Zealand business managers do not see the Internet necessarily as separate from other business tools in their failure to record an increase in active "Internet Business Strategy" development between 2000 and 2001 (Infotech (or was it the Dominion) sometime in June). He questions whether "having an Internet strategy" is any more sensible or relevant to most businesses than "having a mobile phone strategy". The critical issue should be having and developing an information strategy (personal interview (date)).

³¹PricewaterhouseCoopers (2000) Electronic Business Outlook for the New Millenium

³²More frequent, reliable message transmission (independent of the trade winds on which sailing ships were reliant) enabled by steamships facilitated not only more frequent and reliable trade (minimising the risks and subsequent costs of delays) but also the administrative message communication system that underpinned the effective operation of the British and all other European colonies. (Cunard Steamship Society http://www.scotweb.co.uk)

³³That is, substituting non-electronic forms or products and processes with electronically based ones.

³⁴This distinction recognises different uses of information as opposed to different uses of technology (Section 1). Using a new technology to undertake the same information transfer as existed previously would fit into the first category (e.g. replacing landlines with mobile phones to convey voice messages, replacing paper invoices with a digitised transfer containing the same information, shifting the locus of ownership of a supplier's inventory information to the manufacturer who needs it for production scheduling purposes) whereas new uses of information (e.g. creating new information products - risk profiles - and trading them as in the creation of the Insurance industry; digitisation of music into MP3 files - an information product - distributed via the Internet rather than pressing physical CDs for distribution via a wholesale or retail network) would constitute the creation of a new digital product.

³⁵For example, the so-called Productivity Paradox - if computers have made processes at the firm level apparently more productive (e.g. banking, as evidenced in Haltiwanger and Jarmin), why have overall productivity figures for industries and economies not matched firm-level expectations?

³⁶Thus, Brynjolfsson and Hitt's (2000) assertions that investments in computers only yield productivity improvements with concomitant investments in organisational systems can be interpreted as an acknowledgement that the computer process on its own is not a perfect substitute for the production processes it replaces, and full benefit is gained only if the informational inputs of computer technologies and organisational process technologies are jointly optimised alongside capital and labour substitution effects. By focussing productivity measures solely on investments in computer technology and investments in computer software, we have been overlooking investments in all other forms of information processing technologies, and in information as a stock.

³⁷Deloittes reports the creation of "a new economics of transactions", and PWC highlights that 44% of businesses undertake investment in electronic technologies without putting either quantitative or qualitative measures in place to assess their value comparative to the investment made.

³⁸Infrastructure appears the logical place to start in a performance measurement exercise. For example, The State of e-New Zealand starts here by assuming that, in the absence of any other firm statistics to the contrary, high uptake of infrastructure implies a demand for the applications which underpin the application of the technologies.

³⁹This is reinforced by reports such as Bright Planet (http://www.brightplanet.com/deep_content), which asserts that even within the context of the web, information contained in the "deep web" - that is, Internet content unobtainable using current search engines, either by accident or design - is between 400 and 550 times larger than that available on the commonly available World Wide Web.

⁴⁰Dunt, Emily and Ian Harper. 2001. E-Commerce and the Australian Economy. Melbourne Business School.

⁴¹We have always had knowledge workers - they just have not been classified as such. The authors of this report, as researchers, are pure knowledge workers, taking input information, processing it to add value by creating research reports. Likewise, doctors, nurses and similar health-care providers take information form patients (descriptions of symptoms), process it using human technologies to create an output - the diagnosis and treatment recommendation, which is consumed by the patient. Proportionately a very small amount of medical sector resources are expended on physical products (pharmaceuticals, bandages etc.).

⁴²Noting that "technology" is not limited to machinery - it can also include business method and practice innovators (creators), managers capable of rolling out business re-engineering processes (reproducer/replicator) - indeed, this category which develops practical applications of mechanical technologies and then "institutionalises" them is arguably the most important resource for growth (e.g. Romer's (1986) typology separating "things" and "ideas", etc.).

⁴³The Economist. B2B Exchanges - time to rebuild. 17 May 2001.

⁴⁴PWC states that over 40% of businesses using Internet-based technologies have no specific measures in place to measure quantitative and/or qualitative success.

⁴⁵For example, national economic growth targets based upon measurable GDP figures.

Jump to main page content.

Printed: 25/11/2009