All firms and industries are affected by the digital transformation, although the pace and scale differs. While almost no business today is run without ICTs, their impact depends on the type and sophistication of ICT tools integrated into business processes. Using special tabulations of enterprise data, the OECD carried out an experiment to calculate indicators of digital maturity in business. Additionally, a new OECD taxonomy of digital intensive sectors provides insights into the characteristics and dynamics of those sectors most affected by the digital transformation. New measures of the diffusion of robots (including service robots) in companies are presented, reflecting their role in transforming manufacturing. New data on the perceived impacts of digital technologies in the workplace are also analysed. Online US job vacancies data are used to examine the types of skills required for computer-related jobs. More people are connected than ever before, and many younger people are adopting an “always-on” lifestyle. Digitalisation is also changing the ways in which research is conducted and disseminated. The first results of the OECD International Survey of Scientific Authors (ISSA) reveal scientists’ views on the impacts of digitalisation in their work.
Measuring the Digital Transformation
1.2. Digital transformations
Abstract
Fast adopters and the diffusion of technology
Digital transformation in industry
Digital maturity in industries
Business dynamics and the digital transformation
Mark-ups in the digital era
Transforming production
Transforming the world of work
Which skills for computer jobs?
Computer skills in growing demand
Sophisticated adopters and uptake
Mind the gap
Always-on lifestyle
Science going digital
Impacts on science: scientists’ views
Fast adopters and the diffusion of technology
Most organisations use digital tools, but often not to their full potential. A number of major transformations – often collectively referred to as the “next production revolution” – are anticipated over the coming decade. The technological drivers of this revolution include the development of digital infrastructure and applications, such as high-speed broadband, Big data, cloud computing, 3D printing and the Internet of Things (IoT). Such technologies are increasingly affordable for smaller businesses. However, for technology diffusion to lead to productivity gains, firms must integrate the technology into their business processes and make complementary investments in skills and business models.
Recent surveys of ICT technology show that broadband access has reached saturation in large businesses. However, on average, only 20% of businesses in OECD countries benefited from high-speed broadband (100 Mbps or greater) in 2018. The adoption of digital technologies in business value chains, whether for purchases, sales or the automation of back office functions (ERP), has progressed smoothly, albeit with large differences between countries and sectors. Cloud computing services has registered the fastest increase in uptake – 50% over the four years to 2018 – when, on average, 56% of large businesses and 27% of small businesses purchased cloud computing services. A recent OECD study (Galindo-Rueda et al., 2019) based on analysis of micro-data from the Statistics Canada Survey of Advanced Technologies finds that larger firms tend to make greater use of advanced technologies, especially automated production process technologies, for which scale appears to be very important. In contrast, software and infrastructure service technologies (including cloud computing) register similar rates of uptake in both small and large Canadian firms.
18. Diffusion of selected ICT tools and activities in enterprises, OECD, 2010 and 2018
Source: OECD, ICT Access and Usage by Businesses Database, http://oe.cd/bus, January 2019. See 1. StatLink contains more data.
1. Broadband includes fixed connections with an advertised download rate of at least 256 Mbps.
For the most recent year, data refer to 2018 for the majority of countries included in the sample with the following exceptions:
For ERP, CRM, SCM and RFID, data refer to 2017.
For the earlier year, data refer to 2010 for the majority of countries included in the sample with the following exceptions:
For cloud computing, data refer to 2014 for the majority of countries.
For Big data, data refer to 2016.
For RFID, data refer to 2009 for the majority of countries.
For high-speed broadband, data refer to 2011 for the majority of countries.
19. Diffusion of selected ICT tools and activities in large and small businesses, OECD, 2010 and 2018
Source: OECD, ICT Access and Usage by Businesses Database, http://oe.cd/bus, January 2019. See 1. StatLink contains more data.
1. Broadband includes fixed connections with an advertised download rate of at least 256 Mbps.
For each ICT tool or activity, based on data available for 2010 and 2018, a simple OECD average was calculated for large and small firms
For the most recent year, data refer to 2018 for the majority of countries, with the following exceptions:
For ERP, CRM, SCM and RFID, data refer to 2017.
For the earlier data year, data refer to 2010 for the majority of countries, with the following exceptions:
For cloud computing, data refer to 2014 for the majority of countries.
For Big data, data refer to 2016.
For RFID, data refer to 2009 for the majority of countries.
For high-speed broadband, data refer to 2011 for the majority of countries.
Digital transformation in industry
Every industry is affected by the digital transformation but no single metric is able to capture its pace and extent. Due to their pervasive nature, digital technologies are profoundly transforming economies and societies. The innumerable ways in which the digital transformation is affecting production activities, both manufacturing and services, impede efforts to provide an all-encompassing definition of this multifaceted phenomenon. Recent OECD work (Calvino et al., 2018) assesses the digital intensity of sectors by looking at the technological components of digitalisation (tangible and intangible ICT investment, purchases of intermediate ICT goods and services, robots), the human capital required to embed technology in production (ICT specialist intensity), and the ways in which digital technology impacts how firms interface with the market (online sales). While the digital transformation progressively touches all sectors in the economy, it does so with differing speeds and extents. Only one sector, ICT services, stands out as being the most digital-intensive, as measured by the seven different metrics of sector digital intensity (OECD, 2017). European data from ICT use in business surveys, which allows a granular look at uptake of digital technologies along business value chains, shows that ICT services is the most digital-intensive sector. The presence of websites is rather high for businesses in every sector, and hence does not explain sectoral variations, while the use of Big data analytics is still in its infancy in almost all industries. What really discriminates digital intensity across sectors is the use of more sophisticated digital tools such as cloud computing, enterprise resource planning (ERP), and customer relations management (CRM).
20. ICT uptake by industry, EU28, 2018
Source: OECD, based on Eurostat, Digital Economy and Society Statistics, January 2019. See 1. StatLink contains more data.
1. For ERP and CRM, data relate to 2017.
Calculating indexes of digital maturity: an experiment
To analyse different aspects of the diffusion of digitisation in business, the OECD and Eurostat worked with participating countries to produce special tabulations of data from the 2018 European Community Survey on ICT Usage and E-commerce in Enterprises. As presented on the following page, these focus on the co-occurrence of different items along three dimensions of maturity and sophistication. ICT capabilities: (i) ICT training of staff, (ii) the employment of ICT specialists, and (iii) in-house performance of ICT functions (as opposed to contracting-out). Advanced ICT functions: (i) ICT security and data protection activities, (ii) tailoring of business management software and (iii) the development of web solutions. Web maturity: (i) having a website which allows for product customisation or tracks orders/visitors, and (ii) whether or not the business uses online advertising services. Enterprises were assigned a score for each based on the number of items present - from 0 (no items) to the joint occurrence of all 3 items (2 for Web maturity).
Digital maturity in industries
Businesses in Europe still have yet to exploit the full potential of the digital transformation. On average, 50% of all enterprises in the business sector, excluding financial services, have no specific internal ICT capabilities as measured by the availability of specific human capital. In ICT industries, such as IT services and telecommunications, 40% to 80% of enterprises possess at least intermediate capabilities. This compares to an overall average of 20%, while in relatively low-tech areas such as textile and apparel manufacturing and transport and storage services have rates around 10%. ICT capabilities tend to be associated performing of advanced ICT functions, but the relationship of both with Web maturity is weaker. Based on these benchmarks, which give only a partial view of digitalisation in firms, there are leading sectors (information and communication, travel, wholesale trade) and relative laggards (construction services and food, textile, and metal manufacturing industries). Retail trade and accommodation score high for Web maturity, while medium-high tech manufacturing industries, such as machinery, ICT, and electrical manufacturing, as well as professional and technical services are more oriented towards the integration of ICT applications within business processes.
21. Enterprises with internal ICT capabilities, by industry, EU countries, 2018
Source: OECD calculations based on Eurostat, Digital Economy and Society Statistics, January 2019. See 1.
1. Industry coverage is as follows according to NACE Rev.2:
IT services: Computer programming, consultancy and related activities, information service activities (J62-J63);
Telecommunications: Telecommunications (J61);
Publishing and broadcasting: Publishing activities; motion picture, video and television programme production, sound recording and music publishing; programming and broadcasting (J58, J59, J60);
ICT and electronics: Manufacture of computer, electronic and optical products (C26);
Travel activities: Travel agency; tour operator reservation service and related activities (N79);
Professional and technical activities: Professional, scientific and technical activities (M69-M75);
Wholesale trade: Wholesale trade, except of motor vehicles and motorcycles (G46);
Machinery and electrical equipment: Manufacture of electrical equipment, machinery and equipment n.e.c. (C27-C28);
Transport and equipment: Manufacture of motor vehicles, trailers and semi-trailers, other transport equipment (C29-C30);
Motor vehicles trade: Trade of motor vehicles and motorcycles (G45);
Chemicals: Manufacture of coke, refined petroleum, chemical and basic pharmaceutical products, rubber and plastics, other non-metallic mineral products (C19-C23);
Utilities: Electricity, gas, steam, air conditioning and water supply (D35-D39);
All industries: All non-financial enterprises;
Retail trade: Retail trade, except of motor vehicles and motorcycles (G47);
Real estate: Real estate activities (L68);
Wood, paper and printing: Manufacture of wood and products of wood and cork, except furniture; articles of straw and plaiting materials; paper and paper products; printing and reproduction of recorded media (C16-C18);
Accommodation and food services: Accommodation, food and beverage service activities (I55-I56);
Other manufacturing: Manufacture of furniture and other manufacturing; repair and installation of machinery and equipment (C31-C33);
Metal products: Manufacture of basic metals and fabricated metal products excluding machines and equipment (C24-C25);
Transport and storage: Transportation and storage (H49-H53);
Food products: Manufacture of beverages, food and tobacco products (C10-C12);
Textiles and apparel: Manufacture of textiles, wearing apparel, leather and related products (C13-C15); and
Construction: Construction (F41-F43).
22. Web maturity and advanced ICT functions, by industry, EU countries, 2018
Source: OECD calculations based on Eurostat, Digital Economy and Society Statistics, January 2019. See 1. StatLink contains more data.
1. For industry definitions, see note 21 above.
How to read these charts
ICT capabilities are characterised as “high” when all 3 items are present, “medium” for 2 items and “low” when only 1 of the items occurs. The scatterplot presents a synthetic measure of uptake for Web maturity and Advanced ICT functions. It shows the sum of percentage shares of enterprises for each dimension divided by the theoretical maximum value (i.e. 2 or 3), to create a normalised indicator varying from 0 to 1. See “Calculating indexes of digital maturity: an experiment” on previous page.
Business dynamics and the digital transformation
Business dynamism in highly digital-intensive sectors is high but declining. While business dynamism is, on average, greater in highly digital-intensive sectors, these sectors have also experienced more significant declines in business dynamism over time – especially in terms of firm entry rates. Recent OECD work shows that highly digital-intensive sectors are more dynamic than other sectors on average – consistent with the idea that digital technologies lower entry barriers and tend to facilitate reallocation – but have experienced significant declines in dynamism since 2001, especially in terms of entry and job reallocation rates. This appears to be related, in part, to the fact that while diffusion of digital technologies continues everywhere, in highly digital-intensive sectors – where these technologies have particularly advanced application – it reaches a stage of higher technological maturity. This process is similar to past trends in other innovative sectors, and holds across countries, although there are significant differences between countries in the patterns and dynamics of highly digital-intensive sectors. In this context, institutional and policy factors, such as workers’ training, the availability of venture capital, and the efficiency of business and bankruptcy regulations, play an important role in business dynamism in these sectors (Calvino and Criscuolo, 2019).
23. Changes in business dynamism, entry and exit rates, 1998-2015
Note: Sectors are classified by digital intensity (high/medium-high/medium-low/low) using a number of dimensions (ICT investment and ICT intermediates, use of robots, online sales and ICT specialists) and then grouped by quartile.
Source: OECD calculations based on the DynEmp v.2 (USA) and the DynEmp3 Databases, http://oe.cd/dynemp, January 2019. See 1.
1. The figures are based on the year coefficients of regressions within the country-STAN a38 sector, focusing separately on sectors in the “Highly digital-intensive” and “Other sectors” groups. Average trends for highly digital-intensive sectors are reported with a solid line and for other sectors with a dashed line. The dependent variables of the regressions are, respectively, entry rates or exit rates. Confidence bands (95%) are also reported based on robust standard errors.
Figures are based on data covering manufacturing and non-financial market services, and exclude self-employment and the Coke and Real estate sectors. The countries covered are: Austria, Belgium, Brazil, Canada, Costa Rica, Finland, France, Hungary, Italy, Japan, the Netherlands, Norway, Portugal, Spain, Sweden, Turkey and the United States. Data for Japan are for manufacturing only. The classification of sectors according to digital intensity is based on Calvino et al. (2018) (top quartiles in either of the two periods considered in the study). Owing to methodological differences, figures may deviate from officially published national statistics. Data for some countries are still preliminary.
OECD project on employment dynamics, young businesses and allocative efficiency (DynEmp)
The approach adopted by the DynEmp project is based on a common statistical code developed by the OECD, which is run in a decentralised manner by national experts from statistical agencies, academia, ministries or other public institutions, who have access to the national micro-level data. The micro-aggregated data generated by the centrally designed but locally executed program codes are then sent back for comparative cross-country analysis to the OECD. This distributed micro-data approach reduces confidentiality concerns as it aggregates information at a sufficiently high level, and achieves a high degree of harmonisation as the definition of the extracted information is the same, ensured by the centrally written computer routine. The experts also implement country-specific disclosure procedures in order to ensure that confidentiality requirements are respected. The figures shown are based on the second (DynEmp v.2) and third wave of data collection (DynEmp3) in the DynEmp project.
Mark-ups in the digital era
In recent years, there has been growing concern that markets around the world are becoming more concentrated and less competitive. This is sometimes attributed to the increasingly digital and globalised nature of many markets and the firms that operate within them (OECD, 2018). Digital technologies allow firms to access multiple geographical and product markets almost instantaneously, sharing ideas and exploiting increasing returns to scale, especially from intangible assets. Digital technologies are generally associated with lower costs of operations and of entry into a market, even across borders, thus potentially increasing competition among firms for the market itself. They foster the emergence of new business models, such as platforms, which further facilitate entry into other, non-digital markets, as happened in the case of Airbnb in the accommodation industry or Amazon in the retail sector.
Digital technologies can also potentially increase the market power of some firms at the expense of others. As with other general-purpose technologies, digital technologies do not diffuse instantaneously, and require complementary investments in intangible assets (e.g. in human capital and organisational capabilities) to be adopted. These knowledge-based assets are costly at the start and can take time to integrate into business models and processes. This may open a gap between leading and laggard firms. Moreover, once knowledge has been accumulated, it can be re-used without cost, allowing companies to scale up faster and more easily, and to generate increasing returns to scale. In addition, digital-intensive companies can leverage Big data analytics for targeted marketing, thus better maximising their sales. Many digital services also increase in value when the number of people using them increases (network effects), such that a potential competitor cannot extract the same level of profit until it attracts a sizeable share of the market. Over time, these characteristics may help industry leaders sustain and advance their position, and slow down the entry or growth of competitors.
According to recent OECD analysis following Calligaris et al. (2018), firms in digital-intensive sectors enjoy on average 13% to 16% higher mark-ups – the wedge between the price a firm charges for its output and the cost the firm incurs to produce one extra unit of output – than firms in less digital-intensive sectors (everything else held constant). Additionally, the wedge between the average mark-up of firms in the two groupings has grown over time. Lastly, the gap is significantly larger (up to 55%) and has increased more over time when comparing firms operating in highly digital-intensive sectors versus those in other sectors. The same analysis shows that the digital gap in firm mark-ups decreases in magnitude but remains significant when differences in international competition, intensity in intangible assets and firm patenting are taken in consideration.
24. The increasing wedge in mark-ups between firms in digital-intensive and less digital-intensive industries, 2001-03 and 2013-14
Note: The figure reports the estimates of a pooled OLS regression explaining firm log-mark-ups in the period, on the basis of the firm’s capital intensity, age, productivity and country-year of operation, as well as a dummy variable with value 1 if the sector of operation is digital-intensive vs less digital-intensive (specifications on the left in the graph), or if the sector of operation is among the top 25% of digital-intensive sectors vs not (specifications on the right in the graph). Sectors are classified as “digital-intensive” or “highly digital-intensive” according to the taxonomy developed in Calvino et al. (2018). Mark-ups are estimated from a Cobb Douglas production function. With respect to Calligaris et al. (2018), in this elaboration the parameters of the production function have been estimated at the 3-digit industry level (rather than 2-digit), and including year dummies. Moreover, mark-ups lower than 1 but greater than 0.95 have been winsorized (rather than trimmed) to 1. Standard errors are clustered at the firm level. All coefficients are significant at the 1% level.
Source: OECD elaborations on Calligaris et al. (2018), based on Orbis® data, July 2018.
Transforming production
Robots, including service robots, are transforming manufacturing. Advances in fields such as Big data, 3D printing, machine-to-machine communication, and robots are transforming production. Comparable and representative data on the deployment of industrial robots in 2016 show that Korea and Japan lead in terms of robot density in manufacturing (i.e. the stock of robots relative to employment). Robot density in these economies is about three times that of the average OECD country. The average density in BRIICS (Brazil, the Russian Federation, India, Indonesia, China and South Africa) is significantly lower, but has increased at twice the pace of the average of the top 25 economies between 2007 and 2016. Sales of service robots are also on the rise. In 2018, the International Federation of Robotics (IFR) identified more than 700 service robot manufacturers, both for professional and personal use (IFR, 2018). For the first time, statistics on the use of both industrial and service robots, and of 3D printing have been collected within European surveys of business ICT usage. In 2018, on average, 7% of respondent enterprises with more than ten employees were deploying robots, and 4% used 3D printing. The highest penetration rates are observed in manufacturing of metal products, chemical products and machinery.
25. Top robot-intensive economies and BRIICS, 2016
Source: OECD calculations based on International Federation of Robotics (IFR); OECD Annual National Accounts Database; OECD Structural Analysis (STAN) Database, http://oe.cd/stan; OECD Trade in Employment (TiM) Database; ILO, Labour Force Estimates and Projections (LFEP) Database and national sources, December 2018.
1. Robot use collected by the International Federation of Robotics (IFR) is measured as the number of robots purchased by a given country/industry. The robot stock is constructed by taking the initial IFR stock starting value, then adding to it the purchases of robots from subsequent years with a 10% annual depreciation rate. The figure covers manufacturing sectors only.
For Australia, Greece, Estonia and Slovenia, data are extrapolated from 2013 due to missing data for subsequent years.
For Canada and Mexico, stocks of robots are constructed starting from 2011 due to data availability.
For Chile and India, data refer to 2015 due to missing robot data for 2016.
The density is obtained by dividing the stock by the number of employed persons. Employment data refer to employed person and are sourced from the OECD Annual National Accounts (SNA) Database, the OECD Structural Analysis (STAN) Database or the OECD Trade in Employment (TiM) Database.
For Chinese Taipei, data are sourced from the ILO Estimates and Projections series.
For Singapore, data are sourced from the Ministry of Manpower and include non-resident employed persons.
26. Diffusion of robots and 3D printing in enterprises, by sector and firm size, EU28, 2018
Source: OECD, based on Eurostat, Digital Economy and Society Statistics, January 2019. See 1.
1. An industrial robot is defined by ISO 8373:2012 as “an automatically controlled, reprogrammable, multipurpose manipulator programmable on three or more axes, which can be either fixed in place or mobile for use in industrial automation applications”. A “service robot” is a robot “that performs useful tasks for humans or equipment excluding industrial automation applications”. (ISO 8373). The International Federation of Robotics collects information on shipments (counts) of industrial robots from almost all existing robot suppliers worldwide. No information on service robots is currently made available. The measure of the stock of robots displayed above has been calculated by taking the first-year stock value from the IFR, adding the sales of robots for subsequent years and assuming a 10% annual depreciation. Consequently, these metrics do not capture increases in the quality of robots or their ability to perform tasks.
Industry coverage is as follows according to NACE Rev.2:
Metal products: Basic metals and fabricated metal products, except machinery and equipment (C24-C25);
Chemicals: Petroleum, chemical, pharmaceutical, rubber, plastic products and other non-metallic mineral products (C19-C23);
Machinery and electrical equipment: Computers, electric and optical products, electrical equipment, machinery and equipment n.e.c, motor vehicles, other transport equipment, furniture, other manufacturing, repair and installation of machinery and equipment (C26-C33);
Manufacturing: Total manufacturing (C10-C33);
Food, textile, printing: Food, beverages, tobacco, textile, leather, wood, pulp and paper; publishing and printing (C10-C18);
All industries: All non-financial enterprises;
Trade and repairs: Wholesale and retail trade; repair of motor vehicles and motorcycles (G45-G47);
ICT sector: ICT sector
Retail trade: Retail trade (G47);
Utilities: Electricity, gas, steam and air conditioning; water supply, sewerage, waste management and remediation activities (D35-D39);
Construction: Construction (F41-F43);
Transport and storage: Transport and storage (H49-H53);
Professional and technical services: Professional, scientific and technical activities, except veterinary activities (L69-M74);
Administrative and support services: Administrative and support service activities (N77-N82);
Accommodation: Accommodation (I55);
Information and communication: Information and Communication (J58-J63); and
Real estate activities: Real estate activities (L68).
What is a robot?
An industrial robot is defined by ISO 8373:2012 as “an automatically controlled, reprogrammable, multipurpose manipulator programmable on three or more axes, which can be either fixed in place or mobile for use in industrial automation applications”. A “service robot” is a robot “that performs useful tasks for humans or equipment excluding industrial automation applications”. (ISO 8373). The International Federation of Robotics collects information on shipments (counts) of industrial robots from almost all existing robot suppliers worldwide. No information on service robots is currently available. The measure of the stock of robots displayed above has been calculated by taking the first-year stock value from the IFR, adding the sales of robots for subsequent years and assuming a 10% annual depreciation. Consequently, these metrics do not capture increases in the quality of robots or their ability to perform tasks.
Transforming the world of work
Digital technologies are perceived as having diverse impacts in the workplace; in particular, their adoption is resulting in more time being spent on learning new tools and acquiring new skills. In 2018, more than half of workers in EU countries were using ICTs in their daily work. The introduction of digital tools in the workplace entails learning and adaptation and also affects workers’ tasks and work organisation. In 2018, 40% of workers in the EU had to learn to use new software or ICT tools, and about one-in-ten needed specific training to be able to cope with those changes. The percentage of workers who had to learn new digital tools and the percentage who perceived changes in their work tasks were highest in ICT and finance services and in manufacturing. About 20% of workers using digital tools perceived changes in their work tasks, with the majority of them experiencing greater autonomy in organising tasks. The introduction of new digital tools, on balance, resulted in a decrease in repetitive tasks, yet 15% of workers using ICT technologies report experiencing an increase in such tasks. Workers found it easier to collaborate with colleagues but also felt their performance was more closely monitored. They often found that they needed to devote more time to acquire new skills and increased working of irregular hours was also reported. Important differences exist across countries, in particular in relation to ease of collaboration and the need to devote more time for the acquisition of skills.
27. Impacts of new software or computerised equipment at work, by industry, EU countries, 2018
Source: OECD calculations based on Eurostat, Digital Economy and Society Statistics, January 2019. See 1.
1. Change in tasks refers to the survey item “Individual’s main job tasks changed as a result of the introduction of new software or computerised equipment”.
Had to learn how to use software/equipment refers to the survey item “Individuals had to learn how to use new software or computerised equipment for the job”.
Needed further training refers to the survey item “Individuals needed further training to cope well with the duties relating to the use of computers, software or applications at work”.
28. Perceived impacts of digital technologies on specific aspects of work, EU countries, 2018
Source: OECD based on Eurostat, Digital Economy and Society Statistics, January 2019. See 1.
1. Data refer to Austria, Denmark, Estonia, Finland, Germany, Greece, Hungary, Lithuania, Luxembourg, Norway, Poland, Portugal, Slovak Republic, Slovenia and Spain.
Increase and decrease values are computed as a weighted average (based on the number of worker who used a computer equipment at work) across countries included in the sample.
Net impact refers to increase minus decrease over the computed weighted average.
Which skills for computer jobs?
Computer specialists are sought after, but even this narrowly defined category of jobs requires a wide array of skills – some general, some specific and many changing over time. As jobs change, so do the skills workers need to perform them. This is true for all jobs, including those in high demand, such as computer-related jobs. Burning Glass Technologies’ data on online job postings shed light on the types of skills in demand and on how the skills profiles of occupations change over time. An analysis of 1.8 million job postings in the United States in 2018 for four computer-related occupations points to the types of skills in highest demand for each occupation and the types of skills for which demand has been growing fast over 2012-18. On average, one or more of up to about 500 different skills were demanded for job openings posted in these computer-related occupations, thus highlighting the heterogeneous nature of these jobs. Among the 30 skills in highest demand, some are relevant for all four occupational categories, such as knowledge of Structured Query Language (SQL) tools, system design and implementation, or software development principles. Skills related to cybersecurity are important for both computer and network specialists. The demand for some skills cuts across several of these occupations, including skills related to Java, JavaScript, and jQuery – a computer language that creates “applets” (applications designed to be transmitted over the Internet and executed by a java-compatible web browser). These skills are in high demand for computer programmers and developers, while skills related to basic customer service and help desk support are in high demand when posting vacancies for computer support specialists.
29. Top-demanded skills in computer-related jobs, United States, 2018
Note: The word clouds display the top 30 skills demanded in each of the occupational categories considered. The size of the words mirrors the relative frequency with which words appear. Words appearing in more than 2% of the cases in the category considered are displayed in blue.
Source: OECD calculations based on Burning Glass Technologies, www.burning-glass.com, January 2019. See 1.
1. Data on skill demand by occupation are sourced from Burning Glass Technologies, and refer to the skill categories demanded in online advertisements for job vacancies in the United States in 2018. Skills demand is calculated as the number of online vacancies requiring the job candidate to display a given skill category. Multiple skill categories in the same vacancy are allowed. The font size in the picture increases with the number of vacancies in the occupation demanding the skill. Each of the skill categories represented in blue font is demanded in at least 2% of vacancies for the occupation.
The computer occupations considered represent a subset of the computer occupations identified in the 2010 Standard Occupational Classification System (SOC 2010) of the United States Bureau of Labor Statistics. “Computer and information analysts” corresponds to SOC 2010 class 15-112, “Programmers and developers” to SOC 2010 class 15-113; “Database and network administrators” to SOC 2010 class 15-114; and “Computer support specialists” to SOC 2010 class 15-115.
Computer skills in growing demand
Computer-related occupations are at the heart of the development and adoption of digital technologies, but the computer-related jobs of today are likely to be different from those of tomorrow. In this rapidly changing environment, online job postings can point to fast-growing job titles or profiles in demand. For example, job postings for people working on data lakes (repositories holding vast amounts of raw data) grew rapidly in the United States between 2012 and 2018. Some of the most increasingly demanded skills are common across all computer-related occupations. Examples include “IT automation skills”, “machine learning”, and “Big data” or “software development methodologies”. Others can be identified as fast growing across three or more of these computer occupations. The growth in demand for such technical skills is often combined with an increase in demand for complementary skills, such as the ability to train employees or industry-specific skills, such as “fintech”, “medical procedure and regulation” or “brand management”.
30. Top 10 skills in high demand for computer-related jobs, United States, 2012-18
Note: Only skills categories which were present in more than 2 000 vacancies in each eight-digit Standard Occupational Classification (SOC) 2010 occupation were analysed, so as to minimise the probability that a few large employers drive the resulting growth rate. Growth is calculated over the entire period.
Source: OECD calculations based on Burning Glass Technologies, www.burning-glass.com, January 2019. See 1.
1. The computer occupations considered represent a subset of the computer occupations identified in the 2010 Standard Occupational Classification System (SOC 2010) of the United States’ Bureau of Labor Statistics. “Computer and information analysts” corresponds to SOC 2010 class 15-112, “Programmers and developers” to SOC 2010 class 15-113; “Database and network administrators” to SOC 2010 class 15-114; and “Computer support specialists” to SOC 2010 class 15-115.
Data on skill demand by occupation are sourced from Burning Glass Technologies, and refer to the skills demanded in online advertisements for job vacancies in the United States from 2012 to 2018. Skills demand is calculated as the number of online vacancies requiring the job candidate to display a given skill category. Only skills categories which were present in more than 2 000 vacancies in each 8-digit Standard Occupational Classification (SOC) 2010 occupation were analysed, so as to minimise the probability that a few large employers drive the resulting growth rate. Growth is calculated over the entire period. Skill categories are standardised version of the skills reported in job postings, as identified by Burning Glass Technologies. Dark blue bars in the figure represent skill categories present in at least three of the four considered occupations and display at least a 20% growth rate in each of them.
Computer-related jobs and online job vacancies
Burning Glass Technologies scans more than 40 000 sources and tracks about 3.4 million unique, currently active job openings in the United States and a number of other countries. As the same job vacancies are often posted multiple times, duplicate postings, equivalent to close to 80% of all the postings collected, are removed using sophisticated algorithms. Skill requirements in job postings and workers’ résumés can be expressed in different ways (e.g. “Microsoft Excel” vs. “MS Excel”), necessitating standardisation and categorisation. Occupational information is similarly derived from the reported job titles.
Computer and information analysts are individuals analysing science, engineering, business and other data processing problems to implement, improve, review and automate systems, computers and networks, also for security purposes. Programmers and developers create, modify and test the codes and scripts that allow computer applications to run; set operational specifications and formulate and analyse software requirements; analyse user needs to implement website content, graphics, performance and capacity, and may integrate websites with other computer applications. Database and network administrators administer, test, implement, maintain and safeguard computer databases, networks, Internet systems or segments thereof; monitor networks to ensure availability, performance and security, and may help co-ordinate network and data communications hardware and software. Support specialists provide technical assistance to computer users concerning the use of computer hardware and software; analyse, test, troubleshoot and evaluate existing network and Internet systems; and perform network maintenance to ensure networks operate correctly with minimal interruption.
Sophisticated adopters and uptake
Even in economies with almost universal Internet uptake, the activities many people carry out online are relatively basic and limited, pointing to a divide in digital usage. The types of activities carried out over the Internet vary widely across countries as a result of different institutional, cultural, and economic factors including age and educational attainment. Likewise, country uptake of more sophisticated activities also varies and is impacted by factors such as familiarity with online services, trust and skills. In 2017, almost 60% of Internet users carried out both online purchases and Internet banking, an almost twofold increase from around 35% in 2010. The diffusion of both these activities is strongly related to daily usage and to the overall variety of activities performed online. Controlling for Internet usage, uptake patterns differ in Germany, Switzerland, France and the United Kingdom, where individuals are relatively more likely to purchase online than to use Internet banking, while the opposite is true in the Baltic countries.
31. Sophistication of Internet use by individuals, 2018
Source: OECD, ICT Access and Usage by Households and Individuals Database, http://oe.cd/hhind and Eurostat, Digital Economy and Society Statistics, January 2019. See 1.
1. The activities considered are as follows:
Within the last 3 months: E-mailing for private (non-work) purposes, Accessing social networking sites, Telephoning/video calling, finding information about goods and services, reading/downloading online newspapers/news magazines, uploading self-created content on sharing websites (e.g. YouTube), Internet banking.
Within the last 12 months: Downloading and installing software from the Internet, purchasing online, using software for electronic presentations (slides).
The following series refer to 2017: reading/downloading online newspapers, downloading and installing software from the Internet, using software for electronic presentations and uploading self-created content on sharing websites.
For Brazil, data refer to 2016.
For Chile, Korea, Mexico and Switzerland, data refer to 2017.
How to read this figure
The Y axis represents the number of activities performed by a percentage share band of Internet users in a given country. For example, in Denmark, 60% or more of Internet users carry out at least 8 Internet activities, and between 45% and 60% carry out all the ten activities considered here. In Norway, instead, 60% or more of Internet users carry out at least seven activities; between 45% and 60% carry out more than 7 but fewer than 10 activities, and at least 30% but less than 45% carry out all ten activities. The ordering of the countries in the figure reflects the average number of activities weighted by the share of Internet users.
32. Diffusion of Internet banking and online purchasing, OECD, 2010-17
Source: OECD, ICT Access and Usage by Households and Individuals Database, http://oe.cd/hhind, January 2019. See 1.
1. Data presented on the left-hand panel are based on OECD estimates.
Canada and New Zealand are not included due to data availability.
Unless otherwise stated, Internet users are defined as individuals who accessed the Internet within the last 3 months for Internet banking and the last 12 months for online purchasing. For Australia, Israel and the United States, Internet users are defined with a recall period of 3 months for both variables. For Japan, Internet users are defined with a recall period of 12 months for both variables.
For Australia, data refer to the fiscal year 2016/17 ending on 30 June.
For Israel and Japan, data refer to 2016.
Mind the gap
While Internet uptake is reaching saturation for the younger generation, there remains room for older generations to catch up. Today’s digital economy is characterised by connectivity between users and devices, as well as the convergence of formerly distinct parts of communication ecosystems such as fixed and wireless networks, voice and data, and telecommunications and broadcasting. The Internet and connected devices have become a crucial part of everyday life for most individuals in OECD countries and emerging economies. The average share of Internet users in OECD countries grew by almost 30 percentage points between 2006 and 2018, from 56% to 85%, and more than doubled in Greece, Mexico, and Turkey. Over 50% of 16-74 year olds in Brazil, China and South Africa use the Internet nowadays, and the gap in comparison to OECD countries is narrowing. Some economies are approaching universal uptake, while there remains significant potential for catching-up in others with relatively lower income per person. There are also cross-country differences in the generational gap in usage. In the majority of OECD countries nearly all 16-24 year-olds use the Internet on a daily basis – the median value was 96% in 2018 – while for individuals in the 55-74 age bracket the median stood at 55%, with very wide differences (about 50 percentage points) between leading and lagging countries.
33. Internet users, G20 countries, 2018
Source: OECD, ICT Access and Usage by Households and Individuals Database, http://oe.cd/hhind; Eurostat, Digital Economy and Society Statistics; ITU, World Telecommunication/ICT indicators Database and national sources, December 2018. See 1. StatLink contains more data.
1. Unless otherwise stated, Internet users are defined as individuals who accessed the Internet within the last 3 months. For Canada and Japan, the recall period is 12 months. For the United States, the recall period is 6 months for 2017 and no time period is specified in 2006. For India, Indonesia, the Russian Federation, Saudi Arabia and South Africa, no time period is specified.
For Argentina, data refer to 2016 instead of 2018.
For Australia, data refer to the fiscal years 2006/07 and 2016/17 ending 30 June. The reference period is 12 months in 2006.
For Brazil, data refer to 2008 and 2016.
For Canada, data refer to 2007 and 2012. Data refer to individuals aged 16 and over instead of 16-74 in 2006. The reference period is 12 months.
For China, Korea, the Russian Federation and South Africa, data refer to 2017 instead of 2018.
For EU28, data refer to 2007 instead of 2006.
For India, data refer to 2016 instead of 2018.
For Indonesia, data refer to 2017 instead of 2018 and to individuals aged 5 or more.
For Japan, data refer to 2016 instead of 2018 and to individuals aged 15 to 69.
For Mexico, data refer to 2017 instead of 2018.
For Turkey, data refer to 2007 instead of 2006.
For the United States, data refer to 2007 and 2017.
For Argentina, China, India, Indonesia, the Russian Federation, Saudi Arabia and South Africa, data originate from the ITU World Telecommunication/ICT Indicators (WTI) Database 2018.
34. Generational gap in Internet diffusion, OECD, 2008-18
Source: OECD, ICT Access and Usage by Households and Individuals Database, http://oe.cd/hhind, January 2019. StatLink contains more data.
How to read this figure
The figures show the inter-country gap of Internet use for ages 16-24 and 55-74 years respectively, between 2008 and 2018. In 2018, on average across all OECD countries, close to 95% of individuals aged 16-24 were Internet users with half of the countries ranging between the first (94%) and the third (98%) quartiles of the distribution. Internet users in the country with the lowest uptake represented 79% of the population as opposed to 100% in the country with the highest uptake.
Always-on lifestyle
Many young adults spend at least a quarter of their day online, with instant messaging and social media enabling an “always-on lifestyle”. Improvements in mobile technologies have contributed substantially to the diffusion of Internet usage and broadband penetration. These advances have made online access possible for people who were previously unable to afford fixed broadband connections or found it difficult to use computers. Mobile connectivity contributes to always-on behaviour. From 2009 to 2017, the penetration of wireless subscriptions per 100 inhabitants more than tripled in the OECD area as a whole, increasing from 32 to 102. Country comparisons now range from about 50 subscriptions per 100 persons, to 160 (1.6 per person) in leading countries, with many people owning multiple independently connected mobile devices. The development of apps and increasing device sophistication favour this trend. According to ComScore (2017; 2018), mobile connections account for over half of all digital minutes in most surveyed countries, with app usage amounting to almost 90% of mobile time in 2017. Instant messaging and social networking account for the majority of time spent online. European Social Survey data reveal that the average individual aged 14 and above spent more than three hours per day on the Internet in 2016, while young people aged 14-24 spent 4.5 hours online – about 50% more. Constant connectivity is changing attitudes and behaviour in people’s personal lives, with many social relations now occurring online and the distinction between work and leisure time becoming increasingly blurred. According to the Deloitte 2018 Global Mobile Consumer Survey, US consumers check their smartphones more than 50 times per day, on average, and a large majority (70%) of working adults who have work-provided mobile devices also use these outside of work.
35. Wireless broadband in OECD countries, 2009-17
Source: OECD, Broadband Portal, http://oe.cd/broadband, January 2019. See 1.
1. For 2009, data exclude Canada, Germany, Lithuania, Mexico, the Netherlands and Slovenia.
For 2010, data exclude Lithuania.
36. Average time spent on the Internet daily, all individuals and 14-24 year-olds, 2016
Source: OECD calculations based on the European Social Survey micro-data (2016 edition), January 2019. StatLink contains more data.
Science going digital
Digitalisation is changing the way in which research is conducted and disseminated. In order to identify emerging patterns of digitalisation in science, a new OECD survey, the International Survey of Scientific Authors (ISSA), asks scientists questions on whether digital tools make them more productive. It also includes questions on the extent to which they rely on Big data analytics, share data and source codes developed through their research, or rely on a digital identity and presence to communicate their research. Preliminary survey results reveal contrasting patterns of digitalisation by field. The use of advanced digital tools, including those associated with Big data, is more widespread in computer and decision sciences, and engineering. The life sciences (with the exception of pharmaceutical) and the physical sciences (other than engineering) report the greatest effort to make data and/or code usable by others. There are smaller systematic differences in the reported use of productivity tools, which have much higher general adoption rates. Scholars in the engineering domains report using productivity tools less frequently. Interestingly, the fields making less use of advanced digital and data/code dissemination tools – namely social sciences, arts, and humanities – are more likely to engage in activities that enhance their digital presence and external communication (e.g. the use of social media).
37. Patterns of digitalisation in science across fields, 2018
Note: This is an experimental indicator. This figure presents the average of four distinct standardised factor scores representing latent digitalisation indicators for each scientific field. The factor analysis is based on responses by scientists to 36 questions relating to digital or digitally enabled practices, combined in four synthetic indicators. These indicators or factors have been interpreted and labelled based on how strongly they correlate with different questions.
Source: OECD, International Survey of Scientific Authors (ISSA) 2018, preliminary results, http://oe.cd/issa, January 2019. See 1.
1. The factor analysis was applied to a set of binary variables reporting information on whether or not digital tools were used in a range of scientific activities, or whether or not more advanced digital tools (e.g. Big data analytics) were used or developed as part of an author’s core scientific activities. In the initial step of the factor analysis, given the binary nature of the variables observed, tetrachoric correlations were calculated for each pair of variables. The principal-component factor method was then applied to the resulting pairwise correlation matrix to extract the factors. The number of factors selected was forced to be four at most based on an initial observation of the eigenvalues. In a successive step, to improve the interpretability of the factor loadings, factors were rotated by applying an orthogonal rotation method, which produced factors that are uncorrelated.
The four resulting factors were interpreted and labelled based on their loadings with the observed variables. The factor “Digital productivity tools” exhibits higher loadings with the question items on the use of digital tools in milestone scientific activities, including data collection and analysis, project management, search of research material, manuscript dissemination and fundraising. The observed variables related to the features of the data and codes that are shared and made available by researchers are more strongly correlated with the factor “Data/Code dissemination”. The factor “Advanced digital tools/Big data” exhibits higher loadings with the question items on the use of more advanced digital tools (e.g. Big data analytics, sensors and participative networks), whereas the factor “Online presence and communication” is more strongly correlated with the variables reporting on the use of digital tools for research findings communication, interaction with other researchers, or the use of online personal or team profiles to report on research-related activities or outputs. The factors “Digital productivity tools” and “Data/Code dissemination” explain individually around 14% of the overall variance of the observed variables, whereas the factors “Advanced digital tools/Big data” and “Online presence and communication” explain individually approximately 10% of the variance.
The OECD International Survey of Scientific Authors (ISSA)
During the last quarter of 2018, the OECD contacted a large, randomly selected group of corresponding authors of scholarly documents, asking them to respond to an online survey aimed at identifying patterns of digitalisation in scientific research and exploring its drivers and potential effects. This OECD International Survey of Scientific Authors (ISSA) obtained rich information from nearly 12 000 scholars worldwide about their use of a broad range of digital tools and related practices. In order to provide an overarching and interpretable view of digitalisation patterns in science, answers to 36 questions relating to digitally enabled practices were analysed to identify four major “latent” factors. These represent how likely scientists are to: (i) make use of productivity tools to carry out regular tasks such as retrieving information and collaborating with colleagues, (ii) make data and code outputs arising from research available to others, (iii) use or develop unconventional data and computational methods, and (iv) maintain a digital identity expanding their communication with peers and the public in general. More detailed results and analysis from this study are made available on the ISSA project website (http://oe.cd/issa).
Impacts on science: scientists’ views
Scientists’ views regarding the impacts of digitalisation are positive overall, especially among younger authors. How do scientists themselves view the digital transformation of scientific research and its impacts? Evidence from the 2018 OECD Survey of Scientific Authors (ISSA) suggests that scientists’ views are positive, on average, across several dimensions. There is strong sentiment that digitalisation has the potential to promote collaboration in general and particularly across borders, as well as to improve the efficiency of scientific research. While remaining positive, scientists appear to harbour more reservations regarding the impact that digitalisation may have on systems of incentives and rewards (e.g. the ratings of publications, citations and downloads that constitute the digital “footprint” of a scientific author). Likewise for the ability to bring together scientific communities and scientists with the public (inclusiveness), and the role of the private sector in providing digital solutions. Younger authors are more positive than their older peers, except with respect to the impacts of digitalisation on the incentive system, which may reflect concerns about their future careers. Across countries, the average sentiment towards the impacts of digitalisation seems consistent overall with results from broader population surveys on attitudes towards the impacts of science and technology (OECD, 2015). Scientists outside Europe, including those in emerging and transition economies, appear to be more positive on average regarding the impacts of digitalisation on science.
38. Scientific authors’ views on the digitalisation of science and its potential impacts, 2018
Note: This is an experimental indicator. Survey respondents were asked to rate opposing scenarios on different dimensions from (1 = fully agree with negative view) to (10 = fully agree with positive view). For interpretability, average unweighted scores on each dimension and general summary view (average across dimensions) are presented as percentage deviations from the mid-range point.
Source: OECD, International Survey of Scientific Authors (ISSA) 2018, preliminary results, http://oe.cd/issa, January 2019. See 1.
1. The dimension “Science across borders” includes responses to the question item with the positive scenario “The trend towards an increasing use of digital tools in science and research facilitates personal interactions with researchers and experts abroad”. The dimension “Efficiency of scientific work” includes responses to the question item with the positive scenario “The trend towards an increasing use of digital tools in science and research makes scientific and related work faster and more efficient”. “Collaborative and interactive nature of science” summarises responses to the question item with the positive scenario “The trend towards an increasing use of digital tools in science and research facilitates collaboration and interdisciplinary teamwork”. “Quality of scientific research” summarises responses to two question items with the positive scenarios “The trend towards an increasing use of digital tools in science and research allows to tackle problems that were previously intractable” and “The trend towards an increasing use of digital tools in science and research facilitates the verification and reproducibility of scientific findings”, whereas the dimension “Private sector engagement in digital solutions for science” includes responses to the question item with the positive scenario “The trend towards an increasing use of digital tools in science and research promotes innovation in the generation of new tools and solutions for use by researchers and research administrators”. “Inclusiveness of research opportunities and public engagement” includes responses to the question items with the positive scenarios “The trend towards an increasing use of digital tools in science and research helps bring science closer to the public and society at large” and “The trend towards an increasing use of digital tools in science and research provides more equal opportunities to researchers to pursue successful careers”. “Functioning of incentives and rewards in science” includes answers to the question item with the positive scenario “The trend towards an increasing use of digital tools in science and research makes it easier to assess the broad impact of scientific research and provide better incentives”.
39. Scientific authors’ views on the digitalisation of science, by country of residence, 2018
Note: This is an experimental indicator. Country results should be interpreted and compared with caution as the population of corresponding scientific authors is not uniformly representative of a country’s scientific community. Only values for countries with at least 75 responses have been reported.
Source: OECD, International Survey of Scientific Authors (ISSA) 2018, preliminary results, http://oe.cd/issa, January 2019. See 1.
1. Survey respondents were asked to rate opposing scenarios on different dimensions from (1 = fully agree with negative view) to (10 = fully agree with positive view). For interpretability, average unweighted scores on a general summary view (average across dimensions) are presented as percentage deviations from the mid-range point.