Economics of ICT

The role of ICT in the economy

ICT is potentially a positive driver of economic growth. However, there is a concept known as the Productivity Paradox. We see the computer age everywhere, except in the productivity statistics. The productivity paradox, also referred to as the Solow paradox, refers to the actual impact of information technology on economic productivity. Despite significant advancements in computer technology and communications during the "computer age," there has not been a proportional increase in productivity at the macroeconomic level. The productivity paradox emerged in the 1980s and 1990s when businesses began investing in information technology, such as computers, software, and networks. It was expected that these investments would lead to substantial increases in economic productivity. However, initially, the anticipated benefits were not realized.

Is the productivity paradox resolved?

The Solow Paradox was temporarily resolved in the mid-to-late 1990s. Explained by ICT implementation lag: the 1990s productivity surge was preceded by multiple decades of firm investments in ICT. But productivity growth slowed down again from the mid-2000s. The temporal lag can be reasonably explained by the lack of complementary assets.

AI as a general-purpose technology

General purpose technology (GPT) is characterized by (Helpman, 1998):

• Rapid technical improvement
• Pervasive application across sectors
• Spillover effects: indirect and external impacts of an economic activity on other activities or sectors

Examples of AI: Electricity, Robots, Autonomous Vehicles, ICT.

The impact of GPTs

Typical effects:

• Productivity growth [ + effect ]
• Employment substitution [ - effect ]

But, only a subset of GPTs change the invention methods (new ways to generate innovation): GPT as invention of a method of invention (IMI).

What is an IMI?

Deep learning enables a new approach to undertaking (do) scientific and technical research:

• Shift from labor-intensive research towards research that takes advantage of large datasets and prediction algorithms
• No more huge R&D laboratories with expensive machinery and large numbers of scientists
• Atomwise is an example of AI applied to drugs

IMIs are the key drivers of the Industrial Revolution:

• Stem -> 1° Industrial Revolution
• Electricity -> 2° Industrial Revolution
• ICT -> 3° Industrial Revolution
• AI -> 4° Industrial Revolution (likely)

IMIs have strong positive effects on productivity growth (key drivers of Industrial Revolutions). We are observing a shift in the key driver of economic growth (GPT-IMI), and this interval period is traditionally characterized (as in the case of the previous IMIs) by a productivity slowdown.

Research productivity

The production of new knowledge is central to sustaining economic growth:

Economic Growth = Research Productivity * Number of Researchers

AI & research productivity

The production of new ideas is fundamentally a combinatorial process. Example: incandescent light bulb = electricity + heated filament + inert gas + glass bulb. Technologies that predict what combinations of existing knowledge will yield useful new knowledge hold out the promise of improving growth prospects. Breakthroughs in AI (deep learning) represent a potential step change in this direction.

AI technologies impact research productivity both in the search and discovery phase:

• Search phase: Access to the potential relevant knowledge is more difficult due to the explosion of data. AI makes predictions of the most relevant pieces of knowledge.
• Discovery phase: AI predicts which combinations of existing knowledge will yield valuable new knowledge across a large number of domains.

Type and patterns of innovation

Innovations affect the evolution of industries over time. (Radical vs incremental innovations). Innovations are also crucial in determining the competitive advantage of a firm (product differentiation and cost advantage). Innovations are strictly related to ICT technologies. ICT increases the probability to innovate, to introduce new products or new processes in the market. AI, which can be considered as an evolution of ICT technologies, solves the decreasing trend in research productivity. So, innovations matter and ICT positively affects innovations.

We focus on two aspects of innovations:

1. Types of innovations: help clarify how different types of innovations offer different opportunities for firms, producers, users, and regulators.
2. Technology trajectory: the path a technology follows through time. Several factors affect the trajectory; however, some scholars demonstrate that technology trajectory in many industries follows an S-shape curve. It allows firms to make adequate strategies, in particular allows firms to decide whether and when to adopt a new technology.

Types of innovations

We focus on four dimensions:

Product and process innovation

The first dimension allows to distinguish between product and process innovation. Product innovation is embodied in the outputs of an organization, meaning that it is a new product or a new process that is introduced in the market. Process innovations are innovations in the way an organization conducts its business, such as new techniques for manufacturing or producing or marketing goods or services. A characteristic of all process innovations is that they are directed to improve efficiency or effectiveness of production. New processes may enable the production of new products and vice versa. New products may enable the development of new processes. The distinction between the two is relative because what is a product innovation for one organization might be a process innovation for another. For example, UPS created a new distribution service (product innovation) that enables its customers to distribute their goods more widely or more easily (process innovation).

Radical and incremental innovation

One important dimension regards the distinction between radical and incremental innovation. The radicalness of an innovation is the degree to which it is new and different from previously existing products and processes. Radicalness can also be defined in terms of risk. Radical innovations are high-risk innovations because they require relevant investments in capital and other resources. An example is the 3G wireless technology that required:

• Investment in new networking equipment and infrastructure
• Development of new phones with greater display and memory capabilities as well as a stronger battery and/or better power utilization
• Degree of user acceptance of the technology was unknown

Incremental innovations may involve only a minor change or adjustment to existing practices. The radicalness of an innovation is relative: it may change over time or with respect to different observers. For example, digital photography is considered a more radical innovation for Kodak, a company known for its expertise in chemical photography, compared to Sony, a company with a background in electronics. This difference in perception arises from their respective areas of expertise and the extent to which the innovation disrupts their existing knowledge and practices.

Competence-enhancing vs competence-destroying innovation

These concepts relate to how innovations impact a firm’s existing knowledge and competencies. Competence-enhancing innovations are based on a firm's current knowledge base and capabilities. These innovations leverage and extend the firm's existing expertise, technologies, and processes. Competence-destroying innovations render a firm's existing competencies obsolete. These innovations introduce new technologies or approaches that fundamentally change the competitive landscape, making the firm's current knowledge and capabilities irrelevant or outdated. The example provided is Keuffel & Esser, a company known for manufacturing slide rules. The emergence of electronic calculators rendered their expertise in slide rules obsolete because electronic calculators offered superior functionality and convenience.

Architectural vs component innovation

These are concepts that relate to the process of innovation and the nature of the changes introduced. Architectural innovation refers to a change in the way the different parts of a system or product are organized or interconnected. It involves reconfiguring the overall structure or design of a product or system. Examples of architectural innovation include the shift from traditional landline telephones to mobile phones or the transition from physical music albums to digital streaming services. Component innovation focuses on improving or introducing new individual components or elements within a system or product. It involves advancements in specific technologies, materials, or features that contribute to the overall functionality or performance of the product. Examples of component innovation include the development of more efficient computer processors or the introduction of higher-resolution camera sensors in smartphones. The two types of innovation are often interrelated.

ICT and innovation

ICTs positively affect innovations:

• Speeds up the diffusion of information
• Favours networking among firms
• Enables closer links between businesses and customers

ICT provides organizations with tools and capabilities to enhance their innovation efforts, resulting in improved products, services, and operational processes.

Technology S-Curves

The rate of a technology’s improvement, diffusion, and its rate of adoption to the market typically follow an s-shaped curve. At the initial stage of a technology's development, progress may be slow as researchers and innovators work on refining and perfecting the technology. This phase is often characterized by incremental improvements and limited market penetration. As the technology matures and key discoveries occur, the rate of improvement starts to accelerate. This marks the ascending part of the S-shaped curve. As the technology reaches a certain point, it gains momentum and experiences rapid growth and adoption. This phase is often referred to as the "takeoff" phase or the steep slope of the S-curve. During this period, more users and customers become aware of the technology's benefits and start adopting it. The rate of diffusion increases rapidly as the technology becomes more accessible, affordable, and user-friendly. Eventually, as the technology saturates the market and reaches widespread adoption, the rate of growth starts to slow down. This represents the leveling off or saturation phase of the S-shaped curve. At this stage, most potential users or customers have already adopted the technology, and the remaining market consists of late adopters or those who are more resistant to change. Technology trajectory can be used to analyze both the rate of performance improvement and the rate of adoption by firms. These two phenomena are related: a greater adoption stimulates investments in order to improve the performance of technologies, a better performance means a faster adoption of technologies.

Why it’s important to have the amount of effort and not the time period? Because if the amount of effort is not constant over time, the resulting s-curve when you use the time period on the axis can be misleading. Technologies do not always get to reach the final stage of the technology s-curve. That’s because they may be displaced by a completely new technology that fulfills a similar market need. This kind of technology is called in the literature a discontinuous technology. When a discontinuous technology emerges, it may initially have lower performance or capabilities compared to the incumbent technology. For instance, in the early days of automobiles, they were slower and less reliable than horse-drawn carriages. However, these new technologies often have the potential for rapid improvement and development, eventually surpassing the performance of the previous technology. If the returns to effort invested in new technology are much higher than the effort invested in the incumbent technology, in the long-run it is more likely to displace the incumbent technology.

S-Curves in technology diffusion

The process of technology adoption follows a pattern where the initial adoption is slow due to the unfamiliarity of the technology. As people become more familiar, the rate of adoption accelerates. As the technology becomes better understood and its benefits become more apparent, the rate of adoption increases. This acceleration occurs as early adopters and opinion leaders embrace the technology and showcase its advantages, creating a positive perception. However, there is a point where the market becomes saturated, and the rate of new adoptions naturally slows down because there are fewer remaining users who have not yet adopted the technology. It is important to note that technology diffusion typically takes longer than information diffusion. This is because technology adoption often involves acquiring complex knowledge or skills, and sometimes requires complementary resources or infrastructure to realize its full value. For example, electric lights did not become practical until the development of efficient bulbs, which complemented the technology and made it more valuable to users. The relationship between these curves can be explained by the concept of the learning curve. The learning curve suggests that as the production or use of a technology increases, individuals and organizations involved in its development become more experienced and proficient. This leads to improvements in efficiency, cost reduction, and increased quality. These improvements often follow an S-shaped curve, where initial progress is slow, followed by a rapid acceleration, and finally reaching a plateau as maximum efficiency is approached. This reduction in costs makes the technology more affordable and accessible to a larger number of users, leading to a faster rate of adoption. It's worth noting that the alignment of S-curves of technology improvement and diffusion is not always perfect. Other factors such as market demand, regulatory hurdles, and societal acceptance can influence the rate and pattern of technology diffusion.

S-Curves as a prescriptive tool

While mapping the technology’s s-curve is useful for gaining a deeper understanding of its rate of improvement or limits, its use as a prescriptive tool is limited. That’s because the shape of s-curve can be influenced by changes in the market, component technologies, or complementary technologies, and firms that follow the s-curve model too closely could end up switching technologies too soon or too late. So a company must also consider: new technology’s fit with the company’s current abilities and the firm’s position in complementary resources.

Diffusion of innovation & adopter categories

Technology trajectories

A well-known example of this phenomenon is Intel's focus on the low-end personal computer market, also known as "segment zero." In the early stages, the profit margins in the low-end market may be unattractive. However, Intel recognized the potential of this market as technology improved and costs decreased. By meeting the needs of the mass market at a lower price point than high-end technologies, Intel was able to capture a significant market share and establish a dominant position.

Technology cycle

Technological change tends to be cyclical, meaning that each new s-curve starts with an initial period of turbulence, followed by rapid improvement, then diminishing returns, and ultimately is displaced by a new technological discontinuity. Some scholars characterized the technology cycle into two phases:

• Fluid phase: When there is considerable uncertainty about the technology and its market, and essentially firms compete to introduce a successful technology, experimenting with different product designs.
• Dominant design: After a dominant design emerges (bringing a stable architecture to the technology), the specific phase begins when we have a shift from product to process innovations. In this phase, firms are interested in introducing new processes directed to increase efficiency.

The difference is that they have four phases:

• Era of Ferment: The era of ferment is a period of intense competition and experimentation. It is a time of uncertainty and volatility as companies strive to gain market share and establish their position. This phase is marked by rapid advancements, diversity of offerings, and frequent changes in the competitive landscape.
• Era of Incremental Change: As the technology cycle progresses, a dominant design or approach emerges and becomes widely accepted as the standard. This marks the era of incremental change, where the focus shifts from radical innovations to incremental enhancements, cost reductions to meet customer needs. The rate of change slows down compared to the previous phase.
• Technological Discontinuity: After a period of incremental change, a technological discontinuity may occur. A technological discontinuity is an innovation that fundamentally alters the existing dominant design or creates a new paradigm. It disrupts the established order and opens up new possibilities. This phase initiates a new cycle of exploration, competition, and evolution.

Timing of entry

Increasing returns suggest that timing of entry can be very important. First movers don’t always have the advantage. There are a number of advantages and disadvantages to being a first mover, early follower, or late entrant. These categories are defined as follows:

• First movers: The first entrants to sell in a new product or service category (“pioneers”).
• Early followers: Early to market but not first.
• Late entrants: Do not enter the market until the product begins to penetrate the mass market or later.

First-mover advantages and disadvantages

Advantages:

• Brand loyalty and technological leadership: First movers can build a reputation as a leader in that area of technology, which can help it maintain a competitive advantage.