Chapter 2. Tracking E-Commerce Trends

As digital transformation has accelerated, the e-commerce landscape has become increasingly dynamic. New players have emerged at the same time that established actors have taken on new roles; some barriers to e-commerce at the firm, individual and country levels have been overcome, while new barriers have emerged. This chapter tracks e-commerce trends along several dimensions, presenting new evidence on broad e-commerce trends, developments for firms and consumers, and cross-border e-commerce trends. It concludes by identifying key areas for policy action.

    

As digital transformation has accelerated, the e-commerce landscape has become increasingly dynamic. New players have emerged at the same time that established actors have taken on new roles; some barriers to e-commerce at the firm, individual and country levels have been overcome, while new barriers have emerged; and new opportunities have emerged to unlock the potential of e-commerce to potentially boost growth and consumer welfare.

E-commerce takes place through a variety of seller-buyer relationships that have been evolving. At the same time, increasing fixed and mobile broadband penetration – in addition to the uptake of mobile devices (e.g. the smartphone) – has changed the way in which buyers and sellers participate in e-commerce transactions. To track trends in e-commerce from a firm perspective, two measures are useful to consider:

  • Firm participation refers to the percentage of firms engaging in e-commerce. This measure captures the extensive margin of e-commerce, but yields no direct information about the percentage of e-commerce firms contributing to the overall size of the e-commerce market.

  • Turnover reflects the percentage of total turnover (i.e. value of sales) that e-commerce firms contribute to the market. This measure provides information on the relative size of the e-commerce market. As it is an aggregate measure, it does not provide information on the intensity of e-commerce activities at the firm level. Accordingly, the share of e-commerce in total turnover can be large if: 1) there is a large number of (potentially small) e-commerce firms in the market or 2) a moderate number of e-commerce firms account for a large share of the total market (e.g. because e-commerce firms are large).

Both measures are used in this chapter. The advantage of the former (firm participation) is that its interpretation is relatively straightforward, while the latter (turnover) mixes several dimensions that cannot be disentangled easily (e.g. the relative size of e-commerce firms, share of e-commerce in individual firms’ sales, and firm participation). The advantage of the latter (turnover) is that it captures a value dimension. This chapter analyses e-commerce trends along several dimensions: broad e-commerce trends, developments for firms and consumers, and cross-border e-commerce trends. It concludes with key areas for policy action.

Broad e-commerce trends

E-commerce developed primarily as a means to facilitate repeated transactions between large firms, and it relied on custom networks for the electronic exchange of data (see Chapter 1). With the expansion of open networks like the Internet, e-commerce is now spreading to smaller firms and it is increasingly used for transactions between firms and consumers. While transactions between firms still dominate the e-commerce landscape in absolute terms, the current speed of uptake is on average faster in sectors like accommodation or retail where consumers are a major player. Ubiquitous access to the Internet via mobile devices as well as new payment methods are supporting these dynamics.

B2B transactions dominate the e-commerce landscape, but B2C is on the rise

Current trends confirm that business-to-business (B2B) transactions still account for the lion’s share of e-commerce transactions. With close to EUR 1.8 trillion, the manufacturing sector alone accounted for 43% of total EU28 e-commerce (EUR 4.1 trillion, up from EUR 3 trillion in 2013) in 2016 (Figure 2.1). In the same year, US manufacturing turnover from e-commerce represented 51% (USD 3.5 trillion) of total e-commerce turnover (USD 6.8 trillion, up from USD 6.1 trillion in 2013).1 In both geographic regions, the manufacturing sector was followed by wholesale trade, accounting for EUR 1 trillion in the EU28 and over USD 2.3 trillion in the US. Together, these two sectors accounted for 67% and 85% of the captured e-commerce turnover in Europe and the United States, respectively.

On the other hand, sectors with a strong focus on end consumers accounted for much smaller shares of total e-commerce. Specifically, retail accounted for only about 5% of total e-commerce turnover in the EU28 (EUR 217.6 billion), accommodation for 1% (EUR 42.9 billion) and real estate for 0.1% (EUR 4.7 billion). The remaining turnover was captured by sectors with an intermediate share of business-to-consumer (B2C) sales in e-commerce, including electricity, gas, steam, air conditioning and water supply (EUR 234.3 billion or 6%) and transportation (EUR 318.4 billion or 8%). Similarly low shares from consumer-facing sectors can be observed in the United States, where the retail sector (excluding motor vehicles and parts dealers) accounted for USD 389.1 billion, or 6% of the total turnover from e-commerce. Service activities beyond retail and wholesale in the United States jointly captured close to 9% of total e-commerce turnover shown (see Figure 2.1).2 Overall, the contribution of the manufacturing and wholesale sectors to the total e-commerce transaction value has been diminishing in both regions.3

2.1. Value of e-commerce in the US and EU28 by sector, 2016
Absolute values in terms of turnover
 2.1. Value of e-commerce in the US and EU28 by sector, 2016

Note: See Chapter notes.1

1. Figure 2.1: US: Professional services include scientific and technical services. Administrative and support include waste management and remedies. Real estate et al. includes rental and leasing. Other services excludes public administration. Arts, entertainment et al. includes recreation. Health care et al. includes social assistance. EU28: Firms with more than 10 employees. Total turnover is calculated multiplying turnover or gross premiums written from Eurostat Annual Enterprises Statistics with enterprises’ turnover from EDI-type sales and web sales as percentage of turnover from the Digital Economy and Society database. Wholesale and retail trade exclude motor vehicles and motorcycles; electricity et al. includes gas, steam, air conditioning and water supply; professional activities include scientific and technical activities; publishing et al. includes motion picture, video and television programme production, sound recording, and music publishing, programming and broadcasting.

Source: USA: OECD calculations based on US Census (NAICS): Annual Survey of Manufacturers (database), https://www.census.gov/programs-surveys/asm.html (accessed January 2019); Annual Wholesale Trade Survey (database) https://www.census.gov/awts (accessed January 2019); Service Annual Survey (database) https://www.census.gov/programs-surveys/sas.html (accessed January 2019), and Annual Retail Trade Survey (database) https://www.census.gov/programs-surveys/arts.html (accessed January 2019). EU28: OECD calculations based on Eurostat, Annual Enterprises Statistics by size class for special aggregates of activities (NACE Rev. 2) (database) https://data.europa.eu/euodp/data/dataset/e1USD9juizbBvCXBVLjQ (accessed January 2019) and Eurostat, Digital Economy and Society Statistics (database) https://ec.europa.eu/eurostat/web/digital-economy-and-society/data/comprehensive-database (accessed January 2019).

 StatLink https://doi.org/10.1787/888933922766

B2C sectors are not much behind in terms of e-commerce intensity, and tend to be more dynamic in terms of e-commerce uptake

For 2017, Figure 2.2 depicts the minimum, maximum, median and mean share of sales via computer mediated networks in total turnover for the OECD and some additional countries. Sectors are ordered according to the share of B2C sales in total e-commerce. The figure shows that differences between sectors are far less striking when it comes to the share of turnover that arises from e-commerce transactions. Accordingly, the figure suggests that the overall dominance of B2B e-commerce is explained to a large extent by the absolute size of sectors engaging heavily in B2B transactions, rather than a particular relevance of e-commerce for these sectors.

2.2. E-commerce intensity by sector, 2017
As a percentage of total turnover
 2.2. E-commerce intensity by sector, 2017

Note: See Chapter notes.1

1. Figure 2.2: Retail trade excludes motor vehicles and motorcycles; professional et al. includes scientific and technical activities; accommodation et al. include food and beverage service activities. Based on available data for Australia, Austria, Belgium, Colombia, the Czech Republic, Denmark, Estonia, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Netherlands, Norway, Poland, Portugal, Slovak Republic, Slovenia, Spain, Sweden, the United States and the United Kingdom. Some countries do not provide information on all sectors and only reported data is used to calculate averages. Data from the 2018 survey was not available for some countries and the most recent data was used instead (earliest: 2012). For Australia, data for a given year refers to the fiscal year, ending in June of that year. Data are from 2016 instead of 2018. The Australian definition of e-commerce includes any transaction where the commitment to purchase was made via the Internet, including via email. Data has large standard errors. For Austria, data for manufacturing are from 2017. For Colombia, data are from 2016 and data for construction are not available. For Iceland, data are from 2012 and data for real estate services are not available. For Italy, data for wholesale are not available. For Luxembourg, data for administrative and support services are from 2017, data for manufacturing are from 2012, and data for professional services are from 2016. Data for real estate activities, retail and wholesale trade are not available. For Slovenia, data for information and communication services are from 2016, data for professional services are from 2012, and data for construction, real estate activities and wholesale are not available. For the United States, data by sector are not limited to firms with 5 or more employees. They relate to all employment firms, regardless of size. The North American Industry Classification System (NAICS) was used instead of ISIC Rev.4. Data for administrative and support services are from 2015. For other sectors, data are from 2016. Data for construction are not available.

Source: OECD calculations based on OECD, ICT Access and Usage by Businesses (database), http://oe.cd/bus (accessed March 2019).

 StatLink https://doi.org/10.1787/888933922785

Figure 2.3 shows the absolute change in percentage points over the years 2010 to 2017 in terms of both firm participation in e-commerce (%) and the share of e-commerce in total turnover (%) for the EU28.4 Sectors are grouped according to whether the share of B2C transactions in total e-commerce is above or below 50%. Figure 2.3 reveals that sectors heavily relying on B2C transactions have grown faster on average on both margins. The average percentage point increase in terms of firm participation (turnover) was 6 (8) percentage points in B2C sectors relative to 3 (4) percentage points in sectors heavily relying on B2B transactions.

The largest increases in terms of firm participation for sectors strongly engaged in e-commerce with end consumers occurred in retail (from 17% to 28%) and accommodation (from 58% to 68%). In terms of turnover, travel agency services (from 23% to 35%) and accommodation services (from 19% to 31%) experienced the largest increases. Among B2B sectors, the wholesale sector (from 26% to 35%) and trade of motor vehicles and motorcycles (from 18% to 26%) have seen the largest percentage point increases in terms of firm participation. The share in turnover grew strongest in the sector aggregate comprising electricity, gas, steam, air conditioning and water supply, namely 10 percentage points from 6% to 16%. The construction sector saw the share of e-commerce firms slightly reduced from 5% to 4%, and the percentage of e-commerce in turnover diminished from 3% to 2% in the real estate sector.

Across all sectors, the share of e-commerce turnover resulting from B2C transactions in the EU28 increased from 12% to 16% between 2012 and 2017.5 These dynamics underscore the importance of ensuring that appropriate consumer policy frameworks are in place, and that they evolve as needed with changes in technologies and business models.

2.3. Growth in EU28 e-commerce − B2C- vs. B2B-intensive sectors, 2010-17
Percentage point change
 2.3. Growth in EU28 e-commerce − B2C- vs. B2B-intensive sectors, 2010-17

Note: See Chapter notes.1

1. Figure 2.3: Data for enterprises with 10 or more employees. Sectors are grouped by the average (2010-17) share of B2C web sales in total e-commerce (survey waves 2011-18). Telecommunications (50%), real estate (54%), accommodation (59%), travel agencies (59%) and retail (65%) obtain equal to or more than half of e-commerce orders from private customers. The figure shows absolute changes in percentage points. Averages represent the average absolute percentage point change for each group. Wholesale and retail trade exclude motor vehicles and motorcycles. Computer programming et al. includes consultancy and related activities and information service activities, rental and leasing et al. includes activities for employment, security and investigation, services to buildings and landscape, office administrative, office support and other business support. Professional et al. includes scientific and technical activities. Electricity et al. includes gas, steam, air conditioning and water supply. Publishing activities et al. includes motion picture, video and television programme production, sound recording and music publishing, programming and broadcasting.

Source: OECD calculations based on OECD, ICT Access and Usage by Businesses (database), http://oe.cd/bus (accessed April 2019).

 StatLink https://doi.org/10.1787/888933922804

Mobile e-commerce and alternative payment methods are increasing in importance

Estimates of the rise in e-commerce conducted through mobile devices for the retail sector suggest that by 2021, worldwide mobile commerce will account for USD 3.6 trillion, corresponding to 73% of total e-commerce in the retail sector (eMarketer, 2018[1]). This would result in an increase by over 20 percentage points in only five years, up from 52% in 2016. For 2016, these data imply a total volume of retail e-commerce of USD 1.8 trillion. However, according to GlobalData (2017[2]), the rise of mobile commerce is happening faster in some countries than in others. While an estimated 56% of e-commerce transactions in the People’s Republic of China (hereafter “China”) were carried out on smartphones and/or tablets in 2017, the corresponding share was only 23% in the United States and 26% in the United Kingdom.

Evidence from Paypal (2018[3]) confirms the regional differences for cross-border purchases. These data from 2018 suggest that Asian Pacific users relied most heavily on mobile devices for such purchases, realising about 41% of cross-border purchases in the past 12 months through either a smartphone or a tablet. The region is followed by the Middle East (37%), Africa (37%), Latin America (34%), North America (33%), Western Europe (30%) and Eastern Europe (27%). According to these data, between 56% (Asia Pacific) and 71% (Eastern Europe) of all cross-border purchases were realised through a desktop computer, laptop or notebook.

A rise in alternative payment methods is accompanying the rise in mobile e-commerce

Capgemini (2017[4]) estimates that the number of digital payments that were made over the Internet for e-commerce activities (e-payments) globally reached 40 billion in 2015, and they are estimated to increase by 18% by 2019. M-payments, defined as payments where a mobile phone is used as a payment method and not just an alternative channel to send payment instructions, had reached 49.5 billion transactions according to the same source and is estimated to grow by 22% until 2019, indicating a shift towards mobile payments. Juniper Research (2017[5]) predicts that the total annual transaction value of online, mobile and contactless payments will reach USD 3.6 trillion in 2016, up from USD 3.0 trillion in 2015.

Also with regard to payment methods, Worldpay (2017[6]) estimates that in 2016, credit card payments accounted for 29% of total e-commerce payments, followed by eWallets (18%), bank transfers (17%), debit cards (13%), cash-on-delivery (9%), charge or deferred debit cards (6%) and others (8%). By 2021, eWallets are expected to reach a market share of 46%, whereas credit cards are expected to lose ground. Worldpay also predicts that bank transfers will gain market share due to convenience and increasing access, particularly in the developing world.

Using payment data, a Payvision study estimates that around 8% of global commercial transactions involve digital payment, i.e. they are paid neither by cash nor by a physically present payment card (McDermott, 2016[7]). About 11% of these commercial transactions are accounted for by mobile sales points, e.g. using near field communication technologies. Another 36% of transactions use mobile devices from a distance (about 63% of which are SMS payments) and the remaining 53% is non-mobile. Data suggest that consumers in less developed markets have a stronger willingness to use mobile payments, and mobile technology is thriving in many middle-income economies including Indonesia. Kenya, Mexico, Turkey and Ukraine.

In line with the findings for overall mobile commerce, a GlobalData (2017[2]) report finds that the Asian Pacific markets lead when it comes to alternative payment methods, which often rely on mobile technologies. Their findings reveal that digital and mobile wallets accounted for about 47% of total e-commerce transactions in Asian Pacific countries, followed by payment cards (28%), bank transfers (13%) and cash/checking accounts (11%). Chapter 3 takes a closer look at changes in payment methods from a business model perspective.

E-commerce trends: A firm perspective

Firms drive e-commerce developments, including through innovative business models (see Chapter 3), and play a key role in spurring e-commerce innovations. The Internet and digital technologies enable firms of all sizes to enter new markets and expand their reach, allowing them to grow, scale and benefit from knowledge spillovers as they engage more easily in global value chains (GVCs).

2.1. Defining sellers in the e-commerce ecosystem

E-commerce surveys and data collections measure e-commerce depending on how the transaction takes place. E-commerce may take place over electronic data interchange (EDI), the Internet, or some combination of the two (OECD, 2011[8]). Throughout this report, the following terms distinguish which type of e-commerce is being referred to so that appropriate comparisons may be made.

  • E-commerce firms: All firms engaging in e-commerce, either through EDI or the Internet.

  • E-commerce firms using EDI: Firms that use EDI to participate in e-commerce (B2B). EDI may rely on online channels.

  • Online sellers: Firms that use the Internet to participate in e-commerce, not counting sales via EDI (B2B or B2C).

Firms increasingly participate in e-commerce, with large variations across countries and by firm size

The share of firms that participate in e-commerce sales has grown in most of the OECD area and some other countries, from an average of 16% in 2008 to about 23% in 2017 (Figure 2.4). Absolute increases of more than 10 percentage points are evident in Australia, the Czech Republic, Ireland, Slovenia and Sweden. Starting from relatively low levels in 2008, e-commerce participation in Hungary, Italy, Latvia, Poland, the Slovak Republic and Slovenia more than doubled between 2008 and 2017.

New Zealand and Australia had a particularly high share of e-commerce firms (above 40%) in 2017, explained in part by a large geographical distance from other countries, high Internet penetration and high-quality communications infrastructures. On the other hand, firm participation remained around 11% in the most recent data for Korea and slightly below 10% for Turkey.

2.4. Firm participation in e-commerce by size, 2017
As a percentage of enterprises with ten or more persons employed
 2.4. Firm participation in e-commerce by size, 2017

Note: See Chapter notes.1

1. Figure 2.4: Firm participation is the percentage of all businesses employing more than 10 employees receiving orders over computer networks. Data from for 2017 comes from the 2018 survey (and equally for other years). Unless otherwise stated, only enterprises with 10 or more employees are considered, small firms are defined as companies with between 10 and 49 employees, and large firms as companies with 250 or more employees. For Australia, data for a given year refers to the fiscal year, ending in June of that year. Data are from 2016 and 2010. The Australian definition of e-commerce includes any transaction where the commitment to purchase was made via the Internet, including via email. For Canada, data are from 2013 and 2012; medium-sized enterprises have 50-299 employees and large ones have 300 or more employees. Sales online over the Internet may include EDI sales over the Internet as well as website sales, but do not include sales via manually typed e-mail or leads. For Colombia, data are from 2016 and 2012. For Iceland, data are from 2010 instead of 2009. For Japan, data are from 2016 instead of 2018. Data refer to businesses with 100 or more employees instead of 10 or more. Medium-sized enterprises have 100-299 employees and large firms have 300 or more employees. For Korea, data are from 2016 instead of 2018. For New Zealand, data are from 2016 and 2008. For Switzerland, data are from 2011 instead of 2018. For Turkey, data are from 2010 instead of 2009. For Brazil, data are from 2017 instead of 2018. Data do not exclude manually typed emails or any other such channels after 2010.

Source: OECD calculations based on OECD,ICT Access and Usage by Businesses (database), http://oe.cd/bus (accessed February 2019).

 StatLink https://doi.org/10.1787/888933922823

Large firms participate in e-commerce more than small firms, and the absolute gap is widening

There is a large difference in e-commerce participation between large and small firms (Figure 2.4). In Austria, for example, the relatively low average share of 18% masks significant variation in the participation of large (54%) and small (16%) firms. While the average across firms is largely determined by the high share of small firms in the firm population, it is noteworthy that the participation rate for large firms is higher than for small firms in every single country for which data are available, and it is more than double the rate for small firms in over 60% of the countries studied.

Comparing the gap in e-commerce participation rates between large firms and small firms over time, i.e. between 2017 and 2008 (not shown), reveals that for many countries this gap has widened in recent years. In absolute terms, the gap only diminished in Brazil, Canada, Germany, Korea, Luxembourg and the United Kingdom, and widened in all other countries depicted in Figure 2.4. In relative terms, the gap is still closing for most countries, because small firms in many country started from low adoption rates and adoption thus grew faster among these firms in relative terms. However, in Australia, Colombia, Finland, France, Latvia, Lithuania, the Netherlands, Norway, Portugal and the Slovak Republic, small firms lost ground relative to large firms in both absolute and relative terms, highlighting substantial and persisting imbalances across the firm-size distribution.

The general trend of rising e-commerce can also be observed when looking at e-commerce turnover. Data show that the share of e-commerce in total turnover increased from 13% to 19% between 2008 and 2017, again with a large variation across countries (Figure 2.5). In addition, there are large differences between small firms (9%) and large firms (24%) with regard to the weight of e-commerce in turnover. The largest absolute changes in terms of the e-commerce share in turnover are observed for Belgium and the Czech Republic, with an increase of 21 and 13 percentage points, respectively. Australia, France, Hungary, Ireland and the Slovak Republic also had percentage point increases of more than 8 percentage points each.

Changes over time in the share of e-commerce turnover in total turnover can be driven by two factors that are not easily disentangled without a closer look at the micro-data: the variation in the share of firms participating in e-commerce (shown in Figure 2.4) and the share of turnover that the sub-set of firms participating in e-commerce obtained from electronic sales rather than traditional sales channels.

2.5. E-commerce intensity, 2017
As a percentage of total turnover
 2.5. E-commerce intensity, 2017

Note: See Chapter notes.1

1. Figure 2.5: Percentage of e-commerce in total turnover is the percentage of orders received over computer networks. Data from for 2017 comes from the 2018 survey (and equally for other years). Unless otherwise stated, only enterprises with 10 or more employees are considered, small firms are defined as companies with between 10 and 49 employees and large firms as companies with 250 or more employees. For Australia, data for a given year refer to the fiscal year, ending in June of that year. Data are from 2016 instead of 2018. The Australian definition of e-commerce includes any transaction where the commitment to purchase was made via the Internet, including via email. For Belgium, data for large firms is from 2012 instead of 2018. For Colombia, data are from 2016 and 2012. For Denmark and Estonia, data are from 2010 instead of 2009. For Greece, data are from 2008 instead of 2009. For Iceland, data is from 2012 and 2010. For Luxembourg, data are from 2012 instead of 2009.

Source: OECD calculations based on OECD, ICT Access and Usage by Businesses (database), http://oe.cd/bus (accessed March 2019).

 StatLink https://doi.org/10.1787/888933922842

2.2. Comparison of e-commerce participation and turnover contribution suggests mixed business models and rising e-commerce intensity at the firm level

Additional insights into the relationship between e-commerce participation (the extensive margin) and the share of e-commerce sales in total sales can be gleaned from Figure 2.6. The figure reflects the share of firms participating in e-commerce and the share of e-commerce in total turnover for the EU28 average in three different size groups.

2.6. Firm participation and e-commerce turnover by size class, 2008-17
As a percentage of enterprises with ten or more people employed and as a percentage of total turnover, EU28
 2.6. Firm participation and e-commerce turnover by size class, 2008-17

Note: See Chapter notes.1

1. Figure 2.6: Percentage of e-commerce in total turnover is the percentage of orders received over computer networks. Firm participation is the percentage of firms receiving orders over computer networks. Data for 2017 comes from the 2018 survey (and equally for other years). Data for 2017 were added from Eurostat and include a break in series for Germany. Only enterprises with 10 or more employees are considered, small firms are defined as companies with between 10 and 49 employees and large firms as companies with 250 or more employees.

Source: OECD calculations based on OECD, ICT Access and Usage by Businesses (database), http://oe.cd/bus (accessed April 2019) and Eurostat, Digital Economy and Society Statistics (database), https://ec.europa.eu/eurostat/web/digital-economy-and-society/data/comprehensive-database (accessed March 2019).

 StatLink https://doi.org/10.1787/888933922861

In all three size groups, the share of firms participating in e-commerce tends to be higher than the share of turnover resulting from e-commerce. There are two possible explanations for this: either the average e-commerce firm does not derive its entire turnover from e-commerce, or firms participating in e-commerce tend to be smaller in terms of turnover than the average firm in each size group. Given that larger firms tend to be faster in the uptake of new technologies, the first interpretation seems more likely. Accordingly, Figure 2.6 reflects the fact that many e-commerce business models still depend on a mix of electronic and more traditional sales channels, as discussed in more detail below.

In addition, comparing across groups, it is noteworthy that from 2008 to 2017 both the share of firms participating in e-commerce and the share of turnover resulting from e-commerce sales grew faster in large firms (10 and 7 percentage points, respectively) than in medium sized firms (9 and 3 percentage points, respectively) and small firms (6 and 3 percentage points, respectively), confirming a widening gap on average. However, as a large majority of firms in OECD countries are small firms, the group of small firms is still likely to account for the largest entry into the e-commerce market in absolute numbers.

Comparing relative changes in both time trends, the ratio of participation share over turnover share increased for all size-groups between 2008 and 2012. In simple terms, this suggests that by 2012 a larger share of e-commerce firms accounted for a relatively lower share of e-commerce in total turnover (from both e-commerce and offline sales). This dynamic holds within each size group and could be driven by two effects: either the average share of e-commerce in turnover has fallen per e-commerce firm or, what seems more likely, e-commerce firms have on average become smaller (e.g. because online platforms facilitated market access for small firms). After 2012, the ratio between participation and turnover share decreased on average, likely explained by rising shares of e-commerce in total turnover for the average e-commerce firm. In addition, there could have been an increase in the average size of e-commerce firms.

Small e-commerce firms are significantly more likely than large firms to participate in web sales

As argued in OECD (1999[9]), the rise of networks using non-proprietary protocols, like the Internet, enables new forms of e-commerce that no longer rely on costly and customised private networks and EDI, which enabled the first wave of B2B e-commerce between large firms.6 Accordingly, the share of e-commerce turnover that firms attributed to web sales rather than sales via EDI increased from 26% to 39% between 2010 and 2017 in EU28 countries.7 This trend reflects technological change, as the expansion of the Internet allows SMEs to more easily participate in the market and removes the need to set up a costly EDI network.

Over the years 2010 to 2017, small e-commerce firms on average tended to have significantly higher shares of e-commerce turnover resulting from web sales (53%), rather than EDI, both compared to medium sized (35%) and large firms (30%) in the EU28 (Figure 2.7). Within web sales, small firms further derived a slightly higher share of turnover (43%) from end consumers (B2C), compared to 41% for large firms and 34% for medium sized firms. These data suggest that overall small firms obtain a significantly higher share of e-commerce turnover from consumers. The share of B2C transactions for small firms is about 23%, or almost double the share for both medium sized and large firms (around 12%).8

The overall finding that web sales, and particularly B2C sales, are more common among small firms is supported by data at the extensive margin, i.e. the share of firms that participate in a specific sales channel. About 84% of small e-commerce firms in Europe participate in e-commerce through web sales, of which 79% report sales to consumers. This implies that around 66% of all small firms are using e-commerce to sell to end consumers (B2C). Less than a third (28%) was involved in traditional B2B sales via EDI.9 Among large firms, 62% used the Internet for e-commerce, of which 66% engaged in sales to end consumers, implying a B2C share in e-commerce firms of only 41%. The corresponding share was 49% for medium sized sellers. The fact that among large firms both the share of firms using EDI and the share of firms using web sales are relatively high suggests that adding web sales to an existing EDI setup might not be a major challenge for most large firms.

2.7. E-commerce engagement in web sales and EDI by firm size, 2010-17
As a percentage, averages over time, EU28
 2.7. E-commerce engagement in web sales and EDI by firm size, 2010-17

Note: Values represent averages over the survey years 2011 to 2018 to identify structural differences. See Chapter notes.1

1. Figure 2.7: Data for 2017 comes from the 2018 survey (and equally for other years). Values are averaged over the survey years 2011 to 2018 to identify structural differences. Only enterprises with 10 or more employees are considered. Small firms are defined as companies with between 10 and 49 employees and large firms as companies with 250 or more employees. Web sales capture all e-commerce turnover other than EDI-type sales. Firms with web sales are a subset of e-commerce firms. B2C sales are a subset of web sales. Selling via the Internet/apps (% e-commerce firms) is the share of enterprises having received orders via a website or apps (web sales) in all enterprises receiving e-commerce orders over the last calendar year (accordingly for EDI). E-commerce marketplaces are websites or apps used by several enterprises for trading products. Data for B2C as percent of firms and percent of turnover are averaged over the years 2013 to 2018 due to limited data availability. Data for sales via marketplaces are averaged over the years 2017 and 2018 due to limited data availability.

Source: OECD calculations based on Eurostat, Digital Economy and Society Statistics (database) https://ec.europa.eu/eurostat/web/digital-economy-and-society/data/comprehensive-database (accessed April 2019).

 StatLink https://doi.org/10.1787/888933922880

Small firms engaging in web sales also more frequently sell to consumers via online platforms (35%), compared to 29% for medium sized and 23% for large firms. Considering all e-commerce sellers (including those selling via EDI), the share of small e-commerce firms selling via an electronic marketplace to end consumers is 29%, i.e. more than twice the share for large firms (14%), both because small firms rely on the Internet for sales more frequently and because a larger share of these sales goes to end consumers via online platforms.10

The Internet has enabled B2C transactions in all sectors and provides business opportunities for SMEs in some classical B2B sectors

Figure 2.8 shows several sectors in the order of their average share of B2C sales in total turnover from e-commerce based on data from Eurostat (2012-17).11 These data provide information on e-commerce transactions only and are not representative of offline sales in each sector. Almost all sectors posted at least some sales to final consumers, with the share being lowest for the manufacturing sector (2%), the wholesale sector (5%) and the construction sector (10%). The sectors with the highest share of B2C transactions in e-commerce were retail (excluding motor vehicles and motorcycles) (65%), accommodation (59%) and travel agency services (59%).

While information on firm size within sectors is unavailable in the Eurostat data, the previous findings of the importance of B2C transactions and the low attractiveness of EDI technology for small firms (see Figure 2.7) suggest that sectors heavily engaging in transactions with end consumers currently provide the best opportunities for SMEs.

Figure 2.8 also shows that web sales have replaced EDI sales in many sectors with a high share of B2B transactions. In particular, some B2B sectors dealing with services seem to use web technologies more than others, and thus provide additional opportunities for SMEs. For example, web sales account for 54% of B2B transactions in the sector comprising computer programming, consultancy and related activities and information services activities. Other B2B-intensive sectors that rely heavily on web sales include the aggregate of activities related to rental and leasing, employment, security and investigation, services to buildings and landscape, office administration, office support and other business support (48%) as well as trade of motor vehicles and motorcycles (44%).12

2.8. Sales channels and B2C sales by sector, 2012-17
As a percentage of all e-commerce firms, averages over time, EU28
 2.8. Sales channels and B2C sales by sector, 2012-17

Note: Values represent averages over the years 2012 to 2017 to identify structural differences. See Chapter notes.1

1. Figure 2.8: Data for enterprises with 10 or more employees. Values are averaged over the years 2012 to 2017 (survey waves 2013-18) to pick up structural differences. Wholesale and retail trade exclude motor vehicles and motorcycles. Trade of motor vehicles et al. includes trade of motorcycles. Computer programming et al. includes consultancy and related activities and information service activities, rental and leasing et al. includes activities for employment, security and investigation, services to buildings and landscape, office administrative, office support and other business support. Professional et al. includes scientific and technical activities, electricity et al. includes gas, steam, air conditioning and water supply, publishing activities et al. includes motion picture, video and television programme production, sound recording and music publishing, programming and broadcasting.

Source: OECD calculations based on Eurostat, Digital Economy and Society Statistics (database) https://ec.europa.eu/eurostat/web/digital-economy-and-society/data/comprehensive-database (accessed April 2019).

 StatLink https://doi.org/10.1787/888933922899

Factors that influence firms’ participation in e-commerce

This section provides evidence on factors that firms have identified as obstacles that limit or prevent sales via a website or an app.13 The data mostly stems from Eurostat and is representative for the EU28 countries. Although there is little to no representative data for other countries, patterns that stand out in the large diversity of countries within the EU are likely to provide some useful insights for policy makers in other developed countries. By the same token, changes over time are often driven by the same business model innovations affecting several countries, such as the proliferation of online platforms (see Chapter 3).14

Product suitability is a major e-commerce challenge, particularly for large firms and B2B sectors

Figure 2.9 shows the percentage of firms that indicated that they faced specific obstacles that limited or prevented sales via a website in 2015. Panel A shows the responses from firms that received orders via websites or apps and thus have participated in sales via e-commerce. Panel B shows the responses from firms that did not receive orders via websites or apps.15

The figure, but particularly Panel B, indicates that the major perceived obstacle limiting or preventing sales via a website appears to be the suitability of the product offered for sale online. According to these data, 57% of offline sellers mentioned product suitability as an obstacle to e-commerce (down from 60% in 2012), while the percentage was significantly lower, though still sizeable, among online sellers (24%, up from 20% in 2012).

Interestingly, this challenge was mentioned more often by large firms, whereas all other obstacles, including costs, logistics, payments, the legal framework or ICT security were more of an issue for small firms. This finding could be explained by evidence presented earlier that shows that large firms tend to participate in e-commerce activities more intensively overall. Assuming that firms active in markets that lend themselves more easily to e-commerce transactions are the first to enter the online market, the overall higher participation of large firms would imply that remaining large offline sellers are not selling online because they perceive that their products are less suitable for e-commerce, while relatively more small firms remain offline due to firm rather than market conditions.

2.9. Obstacles to selling on the web cited by firms, 2015
 2.9. Obstacles to selling on the web cited by firms, 2015

Note: See Chapter notes.1

1. Figure 2.9: Products not suitable refers to goods and services. Costs are too high: the costs of introducing web sales too high compared to the benefits. Logistics problems apply to shipping of goods or delivery of services. ICT security problems also relate to data protection. Responses are from the 2016 and the 2013 survey respectively and relate to sales activities in the previous year, i.e. 2015 and 2012.

Source: OECD calculations based on Eurostat, Digital Economy and Society Statistics (database) https://ec.europa.eu/eurostat/web/digital-economy-and-society/data/comprehensive-database (last accessed April 2018).

 StatLink https://doi.org/10.1787/888933922918

Entry barriers appear to have fallen, but challenges related to high costs of delivery and returns persist

The fact that SMEs consider the suitability of products to be less of an issue for e-commerce could also be related to the sectors these firms are particularly active in, namely sectors with a high share of B2C transactions. Thus, while the percentage of offline sellers that referred to suitability as an obstacle to e-commerce participation was lowest in the accommodation (33%), travel agency (40%) and retail sector (42%), it was significantly higher in sectors with high shares of B2B customers, including wholesale (49%), manufacturing (58%) and the construction sector (64%) (not shown).

Nevertheless, it is noteworthy that even in sectors like retail trade, travel agencies and accommodation, which are rapidly opening up to e-commerce, the share of offline sellers that cite product suitability as an obstacle to e-commerce participation is not negligible. Contrary to the overall trend depicted in Panel B of Figure 2.9, this share actually increased in four sectors: travel agencies (4 percentage points), real estate (2 percentage points), trade of motor vehicles and motorcycles (2 percentage points) and retail (1 percentage point).

While it might seem trivial that product suitability is a major challenge to e-commerce, this finding is interesting for two reasons. First, it illustrates again that products vary in terms of the ease with which they can be sold online. Second, it highlights the additional potential to unlock e-commerce at the extensive margin, namely through technological or business model innovations that allow new types of products to be sold online. As this area is often overlooked in the literature, Chapter 3 takes a closer look at business models that help firms to increase e-commerce at the extensive margin.16

Figure 2.9, Panel B further reveals that firms that do not sell online perceived most obstacles slightly less important than in 2015. Only logistics were perceived as problematic in 2015 as in 2012. While the changes are relatively small and reductions for the most part did not exceed one percentage point, these changes are nevertheless noteworthy, given that the share of e-commerce firms selling via the Internet increased from 14% to 16% over the 2012-15 period.

It is also likely that those firms that entered e-commerce markets during these years faced relatively lower entry barriers to begin with, implying that the pool of remaining offline sellers would have perceived obstacles at the extensive margin as more challenging on average. Considering this potential selection effect, entry conditions might actually have improved by more than the absolute changes suggest. To some extent, this could reflect how online platforms and other business model innovations have reduced entry barriers, e.g. by offering new solutions like e-payment methods or e-fulfilment (OECD, 2013[10]). A longer time horizon is needed to better evaluate these trends.

At the same time, all obstacles (except for payment issues) were mentioned more frequently in 2016 compared to 2013 by those firms that actually engaged in web sales during in the previous year. Again, this could be explained by a selection effect regarding the firms that entered the online market between 2012 and 2015. Specifically, some new entrants might have entered the market with products less suitable for e-commerce, facing more difficulties to flourish online. Furthermore, increasing competition from new entrants might have reduced the profit margin of incumbent online sellers, implying increasing vulnerability of these firms. Whether due to these or other reasons, the percentage of e-commerce firms that identified high costs as an obstacle increased significantly from 13% to 18%, surpassing the share of firms that identified logistical problems or issues related to payments as obstacles.

A Eurobarometer (2015[11]) report provides some more detailed information about the difficulties that offline sellers in 26 countries in the European Union (EU) expected to face when selling products online. The survey (8 705 respondents) allows firms to identify obstacles according to their relevancy; Figure 2.10 depicts the aggregate responses across countries.

The data reveal that almost a third of firms (33%) expected high delivery costs to cause major problems with respect to e-commerce participation, with an additional 24% expecting minor problems in this regard. Delivery costs were followed by expensive returns and guarantees as a major (32%) or a minor (23%) concern. Insecurity about the rules that have to be followed represented another major (29%) and minor (25%) concern, as did a slow Internet connection, which was a major concern for 29% of firms and a minor concern to another 17%. Other problems that were frequently identified as major problems related to the risk that online sales would bring down product prices (a major concern for 27% and a minor concern for 22%) and a lack of the necessary information and communication technology (ICT) skills (a major concern for 23% and a minor concern for another 23%).

Overall, over a third of all firms expected additional problems (major and minor) related to suppliers charging higher prices for products sold online (39%) and the risk that online sales would damage the image of the company or its trademarks (35%). In addition, 31% of respondents feared that suppliers would not allow them to use third party platforms to sell their products and/or services and 30% feared that suppliers would restrict or forbid the sales of products online.17

2.10. Problems that EU28 firms expect if they were to sell products online, 2015
As a percentage of all firms not selling online
 2.10. Problems that EU28 firms expect if they were to sell products online, 2015

Note: See Chapter notes.1

1. Figure 2.10: Questions: If you were to sell your products and/or services online, tell me if each of the following difficulties would be a major problem, a minor problem or not a problem at all for your company? Base: Companies that do not sell online (N = 5122).

Source: OECD calculations based on European Commission (2015) “Flash Eurobarometer 413: Companies engaged in online activities”, http://ec.europa.eu/commfrontoffice/publicopinion/flash/fl_413_en.pdf.

 StatLink https://doi.org/10.1787/888933922937

Cross-border disputes, language skills, tax rules and other regulations create additional cross-border e-commerce challenges

The Eurobarometer (2015[11]) survey asks similar questions to EU firms that sell online to other EU countries or that used to do so in the past. In this case, responses relate to actual problems encountered in cross-border e-commerce transactions within the EU. In line with the expectations of offline sellers, Figure 2.11 shows that high delivery costs and expensive guarantees and returns are the most frequently mentioned challenges for online exporters, with delivery costs resulting in problems for over half of all surveyed firms; major problems for 27% and minor problems for 24% of all firms. Guarantees and return costs also generated problems for 42% of all firms (19% major problems). Both problems tend to be related to exports of physical products and highlight the importance of improving the logistics infrastructure connecting European countries.

These problems were followed by the high cost of having to deal with cross-border complaints and disputes (41%), a lack of language skills (39%), having to deal with foreign tax rules (38%) and insecurity about which rules must be followed (37%). Specifically, the latter two imply that over a third of all European exporters might benefit from a further harmonisation of the European Single Market. Issues related to suppliers that requested different prices to be charged abroad (20%), prohibited sales through online platforms (17%) or more generally restricted sales abroad (16%), were mentioned less frequently.

High delivery and return costs particularly affect SMEs, while product labelling and restrictions from business partners are more important for large firms

Figure 2.12 provides less detailed but more recent and representative data on difficulties faced by European firms with web sales to other European countries. It also allows a closer look at challenges of particular relevance to SMEs.

2.11. Difficulties that EU28 firms encountered when selling to other European countries, 2014
As a percentage of firms that sold, used to or tried to sell products online in another EU country
 2.11. Difficulties that EU28 firms encountered when selling to other European countries, 2014

Note: See Chapter notes.1

1. Figure 2.11: Questions: for each of the following difficulties that may present itself when selling or trying to sell online to other EU countries, can you tell me if it has been a major problem or not a problem at all? Base: Companies that sold their products and/or services online in another EU country in 2014 and those that used to do so or tried to do so (N = 1903).

Source: OECD calculations based on European Commission (2015) “Flash Eurobarometer 413: Companies engaged in online activities”, http://ec.europa.eu/commfrontoffice/publicopinion/flash/fl_413_en.pdf.

 StatLink https://doi.org/10.1787/888933922956

2.12. Difficulties experienced when making web sales to other EU countries, 2016
As a percentage of firms with web sales to other EU countries, EU28
 2.12. Difficulties experienced when making web sales to other EU countries, 2016

Note: See Chapter notes.1

1. Figure 2.12: Percentage of enterprises with web sales to other EU countries. Responses are from the 2017 survey and relate to sales activities in the previous year, i.e. 2016.

Source: OECD calculations based on Eurostat, Digital Economy and Society Statistics (database) https://ec.europa.eu/eurostat/web/digital-economy-and-society/data/comprehensive-database (last accessed April 2018).

 StatLink https://doi.org/10.1787/888933922975

While the relative importance of problems is broadly in line with Figure 2.11, the more recent survey responses provide a more optimistic picture of challenges faced by European exporters. Thus, while high delivery and return costs are still the challenge most frequently faced by exporting firms, the absolute share of firms facing these challenges is significantly lower. Furthermore, 59% of all firms indicate that they did not face any of the selected difficulties when engaging in e-commerce exports to other EU countries. To some extent, this may reflect the limited number of response options, but it could also be indicative of the improved functioning of the Digital Single Market in areas such as product labelling and the increasing efficiency of online dispute resolution mechanisms put in place by the European Commission (EC).

The data further reveals that fewer small firms (58%) than large firms (63%) encountered none of the listed difficulties. In relative terms, the difficulties more important to small firms were high delivery and return costs (28% compared to 21% for large firms) and the lack of knowledge of foreign languages (14% compared to 11% for large firms). Adapting product labelling or restrictions from business partners were more often encountered by large firms (12%) than small firms (8%).18

E-commerce trends: A consumer perspective

This section considers the consumer side of B2C e-commerce, although it is important to highlight that this captures only part of the demand side for the transactions covered in the previous sections. It is also important to underscore that the distinction between B2C and B2B e-commerce is increasingly blurring, as businesses begin to make purchases through consumer-oriented online platforms and manufacturers can engage with consumers directly through the same channels. As a result, this section considers sectors beyond retail, potentially including more business-oriented sectors (e.g. manufacturing).

Consumer participation is rising, but varies by age, gender, income and education

It is clear that consumer participation in e-commerce continues to increase. Figure 2.13 shows that by 2018, the percentage of individuals in OECD countries that had participated in online purchases during the last 12 months increased by about 61% compared to 2009 (from 35% to 57%). At least three quarters of individuals recently purchased online in Denmark (84%), the United Kingdom (83%), the Netherlands (80%), Norway (79%), Sweden (78%), Switzerland (77%), Germany (77%) and Iceland (75%). Of the countries where data are available, only Mexico (13%) and Colombia (8%) recorded participation shares of less than 15%.

Compared to 2009, the largest increase of e-commerce participation occurred in Estonia, where the percentage of individuals participating in e-commerce increased by 44 percentage points from 17% to 61%. Other countries with percentage point increases of over 30 percentage points include the Czech Republic and Lithuania (35 percentage points each) as well as Iceland, the Slovak Republic and Spain (31 percentage points each). The lowest increases in absolute terms were observed in Colombia (6 percentage points), Canada (8 percentage points), Korea and Norway (9 percentage points each). Participation in Japan diminished by 4 percentage points.19 Considering relative changes over time reveals that the percentage of e-commerce participation by individuals increased close to eight-fold in Turkey, close to seven-fold in Mexico and close to six-fold in Chile and Brazil.

Older people are significantly less likely to participate in e-commerce and the gender gap persists in some countries

Overall consumer participation in e-commerce is rising, but participation still varies widely by age. Comparing the sub-sample of individuals aged 16 to 24 in 2018 (“Generation Z”) with the group of individuals aged 55 to 74 (“Baby Boomers”) shows that a person from the younger cohort is roughly twice as likely to have purchased a good or a service online during the past 12 months than a person from the older cohort.20 Thus, while the percentage of e-commerce users was only 34% among older users (slightly below the average level of 2009), approximately 70% of younger users had made purchases online.21 The absolute age gap was above 50 percentage points in Estonia, Ireland, Korea and the Slovak Republic. It was lowest in Colombia, Israel and the United States (below 10 percentage points).22 Comparing the age gap in absolute terms over time reveals a substantial increase from 26 percentage points in 2009 to 37 percentage points in 2018, indicating that the older cohort is falling behind in terms of usage when compared to the young cohort.23

2.13. Individuals who participated in e-commerce by age and gender, 2018
As a percentage of all individuals aged 16 to 74
 2.13. Individuals who participated in e-commerce by age and gender, 2018

Note: See Chapter notes.1

1. Figure 2.13: Unless otherwise stated, data refer to the percentage of individuals (age 16 to 74) that purchased online over the last 12 months. OECD average by age group excludes Canada and New Zealand due to missing data for the age group 55 to 74. For Australia, data refer to the fiscal years 2016/17 and 2008/09 ending on 30 June. In 2016/17, the information provided is taken from a question wording that differs slightly from other countries: “In the last 3 months, did you personally access the Internet for any of the following reasons: Purchasing goods or services?” In 2008/09, the recall period is 12 months. For Canada, data are from 2012 instead of 2018. Data for 2009 refers to all individuals aged 16 and over instead of 16 to 74. Data for 2012 refers to all individuals aged 16 to 74. For Chile, data are from 2017 instead of 2018. In 2014 there has been a break in series. Source and methodology differ. For Colombia, data are from 2017 instead of 2018. Between the years 2007 to 2011, the survey module was inserted into the Great Integrated Household Survey (GEIH) and as of 2012 it was inserted into the Quality of Life Survey (ECV). The information is not comparable between 2009 and 2017. For Israel, data are from 2016 instead of 2018. Data refer to individuals aged 20 and over instead of 16 to 74 and 20 to 24 instead of 16 to 24, having used the Internet for purchasing goods or services in the last three months. This includes all types of goods and services. For Japan, data are from 2016 instead of 2018. Until 2010 included, data received by OECD include individuals aged 6 and over (instead of individuals aged 16 to 74). From 2011 onwards, data received by OECD include individuals aged 15 to 69 (instead of individuals aged 16 to 74). Age brackets 16 to 24 refer to 15 to 19. Age bracket 55 to 74 refers to 50 to 69. For Korea, data are from 2017 instead of 2018. For Mexico, data are from 2017 instead of 2018. From 2015, the information is collected through an independent thematic survey, unlike previous years in which the information was obtained through a module raised in various surveys. This methodological change must be taken into account when comparing data prior to 2015. For New Zealand, data are from 2012 and 2006. In 2012, data include individuals who have made a purchase through the Internet for personal use, which required an online payment in the last 12 months. For Switzerland, data are from 2017 and 2010. For 2017, data originate from Eurostat. Before 2017, data have been provided by the Swiss Federal Statistical Office (OFS). For 2010, data originate from Enquêtes Omnibus TIC. For the United States, data are from 2017 and 2013. Supplements to the Current Population Survey (CPS) have been conducted during varying months. The 2013 survey went into the field in July, and the 2017 survey was conducted in November. It is unknown to what extent seasonal variations may play a role in survey responses. The CPS Supplement uses the previous six months as the reference period. Prior to 2015, no reference period was specified. For Brazil, data are from 2016 instead of 2018. From 2008 to 2010: quota sampling approach for the selection of the respondent at the household level. For Costa Rica, data are from 2017 instead of 2018. Data for 2017 correspond to individuals aged 18 to 74.

Source: OECD calculations based on OECD, ICT Access and Usage by Households and Individuals (database), http://oe.cd/hhind (accessed April 2019).

 StatLink https://doi.org/10.1787/888933922994

Recent results from a survey by the EC (2017[12]) suggest that there might be a role for policy makers when it comes to establishing trust and consumer confidence in e-commerce among older individuals. Thus, while in 2016 the percentage of young consumers (aged 18 to 34) feeling confident about purchasing goods or services via the Internet from retailers or service providers in their country was relatively high (86%), this confidence significantly decreases with age, namely to 67% for 55 to 64 year-olds to 43% for people age 65 or older.

There also remains a moderate gender gap across OECD countries on average, but in contrast to the age gap it is narrowing in most countries. Specifically, with a participation rate of 58% in 2018, men still tend to be slightly more likely to participate in e-commerce than women (56%). But there has been some improvement over time, with the absolute gap narrowing from 4 percentage points in 2009 to 2 percentage points in 2018. However, the adoption of e-commerce was higher among women than among men in only 14 out of the 39 countries surveyed. In Estonia, Ireland, Latvia, Lithuania and the United States, women participated over 5% more frequently in e-commerce than men. Nevertheless, in a number of countries the gender gap is still striking, particularly in some Latin American countries. In Mexico, the average man was almost 45% more likely than the average woman to participate in e-commerce in 2018 (16% vs. 11%). In relative terms, Mexico was followed by Turkey (30%), Costa Rica (29%), Brazil (26%), Italy (24%) and Chile (21%). In absolute terms, the gap in participation rates between men and women was above 6 percentage points in Austria, Brazil, Chile, Costa Rica, Hungary, Italy, Luxembourg and Turkey.24

E-commerce participation varies substantially by levels of income and education

There are much higher gaps when it comes to different income groups and levels of educational attainment. Panel B of Figure 2.14 shows the percentage of individuals who have purchased online in the last 12 months by income quartile. Across OECD countries, the data reveal that individuals that live in households at the top income quartile are about 79% more likely to have participated in online purchases over the last 12 months, with an absolute gap between the two groups of over 30 percentage points (73% vs. 41%). This implies that low income can be a very strong deterrent to e-commerce participation of an order of magnitude that is comparable to the role of age.

2.14. Individuals who participated in e-commerce by household income and educational level, 2018
As a percentage of all individuals aged 16 to 74
 2.14. Individuals who participated in e-commerce by household income and educational level, 2018

Note: See Chapter notes.1

1. Figure 2.14: Ordering of countries follows Figure 2.13. Unless otherwise stated, data by educational achievement and household income refer to the percentage of individuals (age 16 to 74) that purchased online over the last 12 months. Educational attainment is measured according to the International Standard Classification of Education (ISCED) maintained by UNESCO. The OECD Model Survey breaks down education into three groups: High education refers to tertiary education (ISCED 5 or above), middle education refers to upper or post-secondary, but not tertiary education (ISCED 3 or 4), and low education refers to at most lower secondary education. Data was unavailable for many countries. For Australia, data refer to the fiscal years 2016/17 ending on 30 June. In 2016/17, the information provided is taken from a question wording that differs slightly from other countries: “In the last 3 months, did you personally access the Internet for any of the following reasons: Purchasing goods or services?”. In 2008/09, the recall period is 12 months. There was a methodology change between 2014 and 2016 in the low level of education category. In 2016, persons with no educational attainment and an undefined vocational certificate were included in this category and totals for the first time. This had a negligible impact on data. For Canada, data are from 2012 instead of 2018. For the 2012 data, the lowest quartile is defined as less than or equal to CAD 30 000, the second quartile is from CAD 30 000 to CAD 55 000, the third quartile is from CAD 55 000 to CAD 94 000 and the highest quartile is CAD 94 000 or higher. In order to obtain equal weighted counts in each category, cases with incomes equal to the category cut-offs were randomly assigned to one of the two categories on either side of the cut-off. For Chile, data are from 2017 instead of 2018. In 2014 there was a break in series. Source and methodology differ. For 2009, low or middle level of education have been surveyed in a single category (which refers to “No formal education completed, primary or lower secondary education” and “Upper secondary education”) and are therefore not available separately. High level of education (ISCED-97: 5 or 6) refers to “tertiary education”. For Colombia, data are from 2017 instead of 2018. Between the years 2007 to 2011, the survey module was inserted into the Great Integrated Household Survey (GEIH) and as of 2012 it was inserted into the Quality of Life Survey (ECV). The information is not comparable between 2009 and 2017. For Iceland, data by income refers to 2017 instead of 2018. For Israel, data are from 2016 instead of 2018. Data refers to individuals aged 20 and over instead of 16 to 74, having used the Internet for purchasing goods or services in the last three months. This include all types of goods and services. For Italy, data on income are from 2013 instead of 2018. For Korea, data are from 2017 instead of 2018. For Mexico, data are from 2017 instead of 2018. From 2015, the information is collected through an independent thematic survey, unlike previous years in which the information was obtained through a module raised in various surveys. This methodological change must be taken into account when comparing data prior to 2015. For New Zealand, data are from 2012 and 2006. In 2012, data include individuals who have made a purchase through the Internet for personal use, which required an online payment in the last 12 months. For Switzerland, data are from 2017 and 2010. For 2017, data originate from Eurostat. Before 2017, data have been provided by the OFS. For 2010, data originate from Enquêtes Omnibus TIC. For the United Kingdom, data on income are from 2008 instead of 2018. For the United States, data are from 2017 and 2013. CPS Supplements have been conducted during varying months. The 2013 survey went into the field in July, and the 2017 survey was conducted in November. It is unknown to what extent seasonal variations may play a role in survey responses. The CPS Supplement uses the previous six months as the reference period. Prior to 2015, no reference period was specified. Income quartiles are approximate because family income is a categorical variable. The fully allocated income variable is unavailable prior to 2010. For Brazil, data are from 2016 instead of 2018. From 2008 to 2010: quota sampling approach for the selection of the respondent at the household level. Income data are collected using ranges in minimum wages. Reported data were aggregated in a way that best fit the distribution into 4 balanced categories. Income nonresponse are not included. From 2008 to 2009: Respondents were provided no reference period, indicator refers to having ever purchased online. For Costa Rica, data on education are from 2015 instead of 2018.

Source: OECD calculations based on OECD, ICT Access and Usage by Households and Individuals (database), http://oe.cd/hhind (accessed March 2019).

 StatLink https://doi.org/10.1787/888933923013

The absolute percentage point gap between the highest and the lowest income percentile group was larger than 50 percentage points in Brazil (59 percentage points) and Lithuania (52 percentage points). In Iceland, the percentage of individuals who made online purchases was slightly higher in the lowest income percentile than in the top percentile (75% vs. 73%). In a number of other countries, including Austria, Colombia, Denmark, the Netherlands and Norway, the gap was below 20 percentage points. It should be noted, however, that in Colombia the overall participation is very low, implying that individuals in the top income quartile were still more than 5 times as likely to participate in e-commerce than individuals in the bottom quartile.25

Data for 2009 show that in absolute terms the income gap increased over time in OECD countries, from about 28 percentage points in 2009 to 32 percentage points in 2018 (not shown).26 This increase in the income gap in absolute terms is particularly worrisome because it occurred despite decreasing costs of digital technologies and connectivity and despite the increasing scope of products that can be purchased online. These products include an increasing number of everyday products like clothing, medicine or groceries, which tend to be relatively more relevant to households at the bottom of the income distribution than classical e-commerce products like books or computer games (see Chapter 3). The fact that these dynamics did not help to reduce the income gap likely implies that there are additional forces at play that interfere with the participation rates of low-income households.

One candidate in this regard is educational attainment, which is closely related to income and thus could explain the persistence in the gap across income quartiles despite falling prices and increasing product variety. Panel A of Figure 2.14 shows the percentage of individuals that made online purchases in the past 12 months for individuals with high, middle or low education levels.27 Across OECD countries, the data reveal that individuals with high educational attainment are about 33% more likely to participate in e-commerce than those with a medium level of educational attainment (76% vs. 57%) and more than twice as likely as individuals with no or only a low level of education (37%). This implies that individuals with low levels of education in 2018 still had e-commerce participation rates close to the aggregate participation rate of 2009 (35%).

The gap between individuals with high and low education levels was particularly large in Latin American countries, and especially Colombia, where less than 1% of individuals with low education had participated in e-commerce in 2018. The share was more than 8 times higher for individuals with medium levels of education (7%) and 29 times higher for the highly educated (26%). In Brazil, 67% of individuals with high educational attainment participated in e-commerce, as compared to only 33% of individuals with medium educational attainment and 6% of individuals with low levels of educational attainment. Large gaps are also noticeable in Costa Rica (51%, 23% and 5%) as well as in Mexico (33%, 16% and 4%). In Chile, individuals with high educational attainment (57%) were about three times as likely to have participated in e-commerce than individuals with low levels of education (18%).28 But even in the countries with the most equal distribution like Denmark, Estonia and Norway (i.e. where the gap in usage between the highest and lowest educational attainment group is less than 20 percentage points), highly educated individuals were on average about 25% more likely to have participated in e-commerce during the past 12 months.

These data suggest that compared to income, educational attainment has an even larger effect on e-commerce participation in many countries. As with income, the gap in e-commerce participation between individuals with high and low education respectively further increased over time in absolute terms, from 35 percentage points in 2009 to 38 percentage points in 2018.29 Education itself can be a powerful policy tool to diminish the large and persistent gaps in e-commerce uptake discussed in this section. However, governments should carefully analyse the specific situation in their country to assess whether additional factors are hindering certain subgroups of the population to engage in e-commerce. While challenges related to a lack of payment mechanisms and consumer trust have improved overall in OECD countries (see Chapter 3), they may still be important inhibitors of e-commerce at the bottom end of the income distribution or for the older cohort of individuals.

A range of factors contribute to the urban-rural e-commerce divide

E-commerce usage varies between rural and urban areas. In particular, available data for the EU28 show that individuals in areas with a greater degree of urbanisation are more likely to participate in e-commerce than their counterparts in more rural areas (Figure 2.15). According to the 2018 Eurostat survey, of those who live in densely populated areas (at least 500 inhabitants per square kilometre), 52% completed at least one online purchase within the past three months. This compares to 50% for intermediate urbanised areas (between 100 and 499 inhabitants per square kilometre) and 46% for sparsely populated (fewer than 100 inhabitants per square kilometre) on average. Among the EU28, Poland (15 percentage points), Lithuania, Luxembourg and the Slovak Republic (12 percentage points each) show the largest gap between e-commerce usage in densely populated and rural areas.

2.15. Broadband access and participation in e-commerce by rural and urban areas, 2018
 2.15. Broadband access and participation in e-commerce by rural and urban areas, 2018

Note: See Chapter notes.1

1. Figure 2.15: Highly urbanised are areas with at least 500 inhabitants per square kilometre. Intermediate urbanised are areas with 100 to 499 inhabitants per square kilometre. Rural areas have fewer than 100 inhabitants per square kilometre. Panel A shows the percentage of individuals that had purchased online during the last 3 months. Data for the United Kingdom is from 2017. Panel B shows the share of households with a broadband connection. A broadband connection implies Internet access for both desktop and mobile devices at speeds at or greater than 256 Kbps. Data for Luxemburg contains break in series due to change in the survey methodology. This figure contains data for “Cyprus”.

Note by Turkey:

The information in this document with reference to “Cyprus” relates to the southern part of the Island. There is no single authority representing both Turkish and Greek Cypriot people on the Island. Turkey recognises the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Turkey shall preserve its position concerning the “Cyprus issue”.

Note by all the European Union Member States of the OECD and the European Union:

The Republic of Cyprus is recognised by all members of the United Nations with the exception of Turkey. The information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.

Source: OECD calculations based on Eurostat, Digital Economy and Society Statistics (database) https://ec.europa.eu/eurostat/web/digital-economy-and-society/data/comprehensive-database (accessed March 2019).

 StatLink https://doi.org/10.1787/888933923032

Over time (not shown), it is further noteworthy that for the EU28, the absolute size of the gap with respect to e-commerce usage between densely populated and sparsely populated areas has only marginally fallen between 2009 and 2017 (from 11 percentage points in 2009 to 10 percentage points in 2017). Between 2017 and 2018, the survey indicates a more significant drop in the gap to only 6 percentage points.

One potential explanation for this urban-rural divide is broadband access. Access to the Internet is a prerequisite for e-commerce participation, and as such one should expect a correlation between broadband access and e-commerce participation. Indeed, broadband access in EU countries is highest in densely populated areas, second-highest in intermediate urbanised areas and lowest in sparsely populated areas for 19 out of 28 countries, as Panel B of Figure 2.15 shows.

However, a closer look at the country level reveals that broadband access cannot fully account for patterns of e-commerce participation. For example, Poland, which has the highest gap in terms of e-commerce participation between households living in densely populated and households living in sparsely populated areas (15 percentage points), has a comparatively modest gap between urban and rural areas when it comes to broadband access. With 7 percentage points, this gap is slightly below the cross-country average of 8 percentage points. For Norway, there is almost no gap between households living in densely and sparsely populated areas in terms of broadband access (1 percentage point), but e-commerce is significantly more common for individuals from households in densely populated areas (69% vs. 59%). These data suggest that additional factors need to be considered when trying to understand rural-urban divides, such as the availability of postal services in rural areas or socio-economic factors, including age, education or income, which are likely to differ across rural and urban areas.

An additional factor might be at play in the cases of Belgium, Malta and the United Kingdom, where residents in sparsely populated areas are in fact more likely than residents in densely populated areas to participate in e-commerce. When looking at the differences between intermediate urbanised and sparsely populated areas, this list of countries also includes Estonia, France and the Netherlands. These patterns could be explained by the so-called “efficiency hypothesis” proposed by Farag et al. (2006[14]), which argues that individuals living in rural areas participate in e-commerce at a higher rate than those in urban areas because of a lack of access to brick-and-mortar stores as well as a more limited product variety in rural areas. As these limitations are likely to be less severe for intermediate urbanised areas, the efficiency hypothesis could also explain why these areas have fallen behind when compared to rural areas.

Outside of the EU, this efficiency hypothesis appears particularly relevant in contexts such as rural China, where e-commerce is becoming increasingly prevalent. In rural China, e-commerce firms are taking a hybrid approach to reaching rural populations, combining online ordering with brick-and-mortar drop-off points where e-commerce customers can retrieve packages. This method allows rural dwellers access to greater product variety while cutting costs for e-commerce firms by centralising delivery points. As disposable income in rural China has grown at a higher rate than that of urban China – between 2012 and 2016 disposable income grew by 36% in rural areas as compared to 29% in urban areas – firms are increasingly looking to rural areas as the next high potential market (Liao, 2018[15]). Similar hybrid models with brick-and-mortar drop-off points combined with online ordering are also being developed in other rural regions, such as rural Pakistan and India, suggesting that the urban-rural divide may represent potential business opportunities in certain contexts (Ranipeta, 2017[16]).

What consumer are buying online is changing, with clothing, footwear and sporting goods in high demand

In 2018, about 64% of individuals who purchased online in the EU28 purchased in the category of clothing, footwear and sporting goods. These goods were also most frequently bought in many other OECD countries, and in particular Korea, where 83% of online buyers have purchased clothing and footwear.30 The second most frequently purchased product category across EU28 countries was travel products, including tickets, accommodation or vehicle hires among other things (53%), followed by tickets for entertainment events (39%) and books, magazines and newspapers (35%). A quarter of all e-commerce customers purchased movies, films, images and music products; photographic, telecommunications or optical equipment; and food, groceries, alcohol, tobacco and cosmetics. The least purchased product categories were computer games or video games, and computer software (23%), ICT services (20%), computer equipment (17%) and medicine (14%).

The overall averages mask variation across countries. In particular, less than 50% of online purchasers ordered clothing and related goods in Italy (44%), Latvia (47%) and Slovenia (46%). Beyond EU borders, clothing and related products were also relatively less popular in Australia (48%), where travel products and related services, including tickets or vehicle hires (51%), but also movies, images and music products (50%) were in higher demand. Travel services were the most frequently purchased product category in several countries, but in particular in Norway (79%), Switzerland (78%), Denmark (75%), Iceland (74%), Sweden (73%) and Ireland (72%). On the other hand, travel services were purchased by less than 25% of online buyers in Latvia (23%), Poland (22%), Costa Rica (20%) and Chile (9%). In Chile, the most purchased product categories were computer equipment, movies, films, images and music products, as well as online tickets for entertainment events (each 12%). In Japan, the most purchased product categories were movies, films, images and music products (54%), as well as books, magazines and newspapers (43%).

Consumption patterns are shifting towards new types of products, reflecting new business models and a larger consumer base

Comparisons over time are difficult because the availability of response categories varies across countries and years. Focusing only on the average for EU28 countries, for which data is more consistent across time, reveals several noteworthy changes between 2009 and 2018. First, the share of individuals reporting online purchases of clothing, footwear and sporting goods rose from 45% to 64%, significantly outstripping online purchases of travel products, which experienced only a small increase from 50% to 53%. Furthermore, the share of online sellers buying food, groceries, alcohol, tobacco or cosmetics almost doubled, reaching 25% in 2018, up from 13% in 2009. Large relative increases were also observed for medicine, with the share of online purchasers increasing from 9% to 14% between 2009 and 2018. Small absolute decreases in the average share of online shoppers were observed for movies, films, images and music products (7 percentage points), computer and video games and computer software (6 percentage points), computer equipment (1 percentage point) as well as photographic, telecommunication and optical equipment (1 percentage point).

These changes indicate a broadening of the scope of goods and services purchased online. In relative terms, products whose fit is relatively more difficult to evaluate from a distance, such as clothing, have become more important relative to items that were first sold via online transactions, including through innovative businesses models like Amazon (books) or Booking.com (travel products). The changes observed are also in line with a new generation of online buyers. Goods of general interest, including food, groceries and medicine, have become relatively more important in the e-commerce landscape. Chapter 3 shows how new business models are supporting the expansion of e-commerce at the extensive margin by fostering the entry of new products and by attracting new types of customers.

Convenience, price and availability explain why many individuals participate in e-commerce, but certain impediments persist

When the EC (2015[21]) asked consumers to select up to five main reasons for shopping online, 49% referred to being able to order at any time of the day, 49% said they were finding cheaper products online, 42% said they were saving time, 37% found it easier to compare prices online, 36% referred to more choice online, 25% found it easier to find certain products online, and about 24% referred to delivery being more convenient. Information related reasons, such as the availability of reviews (21%), product comparisons (20%) or general information (18%) seemed to be slightly less important.31

Evidence from a Paypal (2018[3]) survey seems to confirm these benefits for cross-border e-commerce relative to domestic e-commerce. Asked why they had chosen to buy from a foreign website, about 72% of consumers referred to better prices, 49% referred to access for items that were not available in their home country, 34% referred to the discovery of new and interesting products, 29% mentioned higher product quality and 24% indicated that shipping was more affordable. This confirms earlier findings from the Google Consumer Barometer, suggesting that the major reasons for individuals to purchase from abroad rather than from within their country relate to appealing offers (36%), better availability (33%) and better conditions (33%), followed by a broader range of products (24%), better quality of products (11%), recommendations by others (10%) and the least on trustworthiness of the (online) shop (8%).32

2.3. Tapping into new data sources: BBVA credit card data

The OECD and the Spanish Bank BBVA have analysed credit card transactions of Spanish customers, providing novel insights into the consumption patterns of consumers online (OECD, 2019[18]) and the determinants of domestic and cross-border expenditure flows (OECD, 2018[20]).

In absolute terms, the number of online credit card transactions from BBVA has grown from 11 million in 2012 to 76 million in 2016, an almost seven-fold increase. As total online and offline transactions only slightly more than doubled, this implies that the share of online transactions increased from 2% to 8% over this time period. In value terms, rarely available in official statistics, online transactions increased from EUR 814.8 million in 2012 to EUR 4.6 billion in 2016. This represents an increase in online expenditures as a share of total expenditures from 3% to 8% and a decrease in the average value per transaction from EUR 73 to EUR 61, potentially due to an increasing share of low-cost digital products, including apps (OECD, 2019[18]).

Considering only private customers, the data reveal that the online share in total credit card usage reached between 15% (women) and 20% (men) for individuals aged 25 or younger, but remained as low as 4% (women) and 5% (men) for individuals aged 56 or older. On the other hand, the average amount spent per transaction increases with age (from EUR 42 for young individuals irrespective of gender, to EUR 68 and EUR 74 for older women and men, respectively. Individuals aged 26 to 35 are the biggest spenders in absolute terms, with an average yearly expenditure of EUR 1 218 (men) and EUR 968 (women).

2.16. Domestic online credit card expenditure in Spain by product category, 2016
As a percentage of total domestic online credit card expenditure, BBVA customers
 2.16. Domestic online credit card expenditure in Spain by product category, 2016

Note: “Hyper” refers to large supermarkets. See Chapter notes.1

1. Figure 2.16: Domestic online expenditure is the sum of all “card not present” transactions using a BBVA credit card by customers in Spain. Product categories are assigned to merchants (online sellers) by BBVA. “Hyper” relates to large supermarkets.

Source: OECD (2019[18]), “BBVA big data on online credit card transactions: The patterns of domestic and cross-border e-commerce”, https://doi.org/10.1787/8c408f92-en.

 StatLink https://doi.org/10.1787/888933923051

The average amount spent varies widely across sectors, with potentially important implications for consumer protection. For example, the highest average amounts were spent in the travel sector (EUR 248), implying relatively high financial risks in case of disputes. In contrast, in the fashion sector transactions had a much lower value of EUR 62 on average. This is below the average value across all domestic transactions (EUR 70). The corresponding value was only EUR 46 for international transactions, suggesting that customers are relatively more cautious when making purchases from abroad.

As the data also capture refunds from merchants to customers, additional insights can be gained with respect to product returns or failed transactions. For almost every 20 purchases, there is one refund transaction in the data (5%). The ratio is only slightly lower for international transactions (4%) and implies a potentially substantial amount of deadweight loss in terms of transportation, both domestically and across borders. Approximately 57% of all domestic refunds were related to the fashion sector, highlighting how important the right fit is for this sector. This is clearly an online phenomenon, given that the data reveals a corresponding share of only 6% for offline transactions, where the customer can try on clothes before buying.

Furthermore, BBVA identified and categorised the individual merchant for 75% of all transactions that involve a merchant with fiscal headquarters outside of Spain – transactions for which merchant information is not available in the original data. Of the transaction value that these identified merchants represent (i.e. the explained transactions), BBVA found close to 38% to be accounted for by B2C retail online platforms like Amazon and AliExpress. Relative to all cross-border transactions (explained and unexplained) these online platforms accounted for at least 29% of the number of transactions and close to 22% of the total transaction value. Another 24% of the explained transaction value is accounted for by car rental websites, airlines or travel booking portals, whereas major high-tech firms, including Apple, iTunes and Google, accounted for more than 7% of the explained value. Overall, it is noteworthy that the contribution of a handful of prominent online players (excluding payment intermediaries such as Paypal) can account for at least one third of the total online cross-border expenditure (explained and unexplained) by Spanish BBVA customers in the given year.

Recent OECD (2019[18]) work applies a gravity model of trade to the same data and takes a closer look at some of the determinants of intra-regional, inter-regional and cross-border expenditure flows. The analysis reveals that free trade agreements, borders and distance are still important determinants of trade flows in the era of e-commerce. Nevertheless, it appears that the role of distance is less pronounced than for classical trade flows and, as the analysis of domestic purchases reveals, in particular for products like clothing, where a limited regional offer might drive people to purchases from other regions. The results also suggest that customers are more willing to purchase from other Spanish regions if they live in regions with a high average level of education, high consumer prices or a high dissemination and use of digital technologies.

The analysis of cross-border purchases further reveals that customers still tend to be significantly less likely to purchase from other countries. Nevertheless, the results also confirm that the enabling environment for e-commerce in a potential partner country, measured by credit card penetration, the number of secure servers or quality of the logistical system, is an important determinant of the bilateral transaction value. Additionally, factors like the regulatory quality or the existence of a legal framework for e-commerce (e.g. cybercrime prevention) turn out to be good predictors of e-commerce transactions. Finally, the explanatory power of the gravity model increases when large online players are excluded from the data, illustrating how digital multinationals might have an impact on the applicability of well-established trade models.

Preferences, habits and skills are important barriers to individuals’ participation in e-commerce

There is limited information from official statistics that assesses why individuals do not purchase through the Internet. The most recent data available from Eurostat is presented in Figure 2.17. According to these data, the primary reason individuals do not engage in e-commerce relates to preferences. An estimated 69% of individuals who have not ordered goods or services online during the past 12 months mentioned that they prefer to shop in person, see the product, remain loyal to brick-and-mortar shops, or simply were not willing or able to change their habits. Even in 2009, this was the single most important reason (among those mentioned) for purchasing offline (61%) and indicates that the transformation of habits remains one of the biggest challenges for B2C e-commerce. These data suggest that relative to other factors, habits are becoming increasingly important as determinants of e-commerce participation. It also highlights the increasingly important role that innovative business models will likely play in convincing remaining offline consumers of the benefits of e-commerce.

2.17. Reasons for not shopping online, 2017
As a percentage of individuals who ordered products online more than a year ago or never did, EU28
 2.17. Reasons for not shopping online, 2017

Note: See Chapter notes.1

1. Figure 2.17: Reasons for not shopping online, as percentage of individuals who ordered goods or services, over the Internet, for private use, more than a year ago or who never did. Prefer to shop in person also relates to: like to see product and loyalty to shops or force of habit. Trust concerns relate to receiving or returning goods or complaint or redress concerns. Reception of goods relates to problems receiving the ordered goods at home.

Source: OECD calculations based on Eurostat, Digital Economy and Society Statistics (database) https://ec.europa.eu/eurostat/web/digital-economy-and-society/data/comprehensive-database (accessed April 2018).

 StatLink https://doi.org/10.1787/888933923070

Apart from the unspecified category, there is only one other reason for not engaging in e-commerce that was more frequently mentioned in 2017 than in 2009, namely the lack of necessary skills (19% vs. 17%). Furthermore, the share of individuals that mentioned a lack of a payment mechanisms, associated in most OECD countries with income and other socio-economic variables, has not decreased from its 2009 level of 12%. The fact that reasons related to skills and available payment mechanisms have become relatively more important impediments for those currently not engaging in e-commerce indicates how crucial it might become for policy makers to address the gaps highlighted in Chapter 3 on the demand side of the market. The rise of new payment mechanisms, discussed in more detail in Chapter 3 might provide viable market solutions in this regard.

Other factors, including payment security concerns (from 35% in 2009 to 25% in 2017), trust concerns about receiving or returning goods as well as redress concerns (from 26% in 2009 to 16% in 2017) and concerns related to long delivery times or problematic reception of products ordered at home (from 11% in 2009 to 6% in 2017) appear to have become considerably less important over time. To some extent successful policy interventions (e.g. with respect to consumer protection) and business innovations, including same-day delivery and alternative payment mechanisms, may help to explain these results.

Cross-border e-commerce trends

E-commerce enables an increasing number of firms to sell across borders, including many small firms, and increases the product choice for consumers. While e-commerce firms are more likely to sell across borders than average offline firms, cross-border e-commerce is still often limited to geographically close trading partners.

B2C e-commerce is essential for SME exports and has spurred the rise of the Chinese market for e-commerce

While it is difficult to estimate global B2C e-commerce because of a lack of detailed official statistics, several sources suggest that global B2C e-commerce reached around USD 2.3 to 3.9 trillion in 2015, with the Asia Pacific region growing fastest (UNCTAD, 2017[22]); (GlobalData, 2017[2]); (E-Commerce Foundation, 2016[23]). Indeed, China appears to be the largest e-commerce market (accounting for about 75% of e-commerce in the Asian Pacific region), followed by North America, Latin America and the Middle East and North Africa (MENA) region (GlobalData, 2017[2]).

The World Customs Organization (2015[24]) indicates that cross-border e-commerce in goods may have accounted for 10% to 15% of the total goods e-commerce volume. It is estimated that by 2025, Asia may well account for about 40% of this volume, with Europe potentially accounting for 25% and North America for 20%. The United Postal Union (2018[25]) provides further evidence of the increasing role that the Asian Pacific region plays for cross-border e-commerce, in particular with respect to border crossings of small packages up to 2 kilogrammes.33

Furthermore, so-called ICT-enabled services (i.e. services that can potentially be delivered remotely over ICT networks (UNCTAD, 2015[26])), were estimated to have accounted for 56% of all services exports from EU member states to non-EU countries and 52% of all service imports from non-EU countries in 2014. The corresponding numbers for the United States are 54% of services exports and 48% of imports (US Department of Commerce, 2016[27]). In this context, it should be noted, however, that while e-commerce as understood in this report captures all digitally-delivered services that have been ordered online (e.g. a Netflix subscription), not all potentially ICT-enabled services lend themselves easily to e-commerce, given that they might be difficult to specify in an online order.34

Many European e-commerce firms export, but the share has been decreasing in some countries and large gaps remain between large firms and small firms

In 2016, 45% of all online sellers in EU28 countries received orders from other EU countries or the rest of the world, compared to 42% in 2010 (Figure 2.18).35 The countries with the highest percentage point increases for the export share were Croatia (from 35% to 49%), Sweden (from 35% to 48%), and the Czech Republic (from 43% to 54%).

2.18. Enterprises that participated in e-commerce sales to other countries, 2016
As a percentage of enterprises that received e-commerce orders over the last calendar year, EU28
 2.18. Enterprises that participated in e-commerce sales to other countries, 2016

Note: See Chapter notes.1

1. Figure 2.18: Enterprises having undertaken electronic sales to other EU countries or the rest of the world as percentage of enterprises receiving e-commerce orders over the last calendar year. This figure contains data for “Cyprus”.

Note by Turkey:

The information in this document with reference to “Cyprus” relates to the southern part of the Island. There is no single authority representing both Turkish and Greek Cypriot people on the Island. Turkey recognises the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Turkey shall preserve its position concerning the “Cyprus issue”.

Note by all the European Union Member States of the OECD and the European Union:

The Republic of Cyprus is recognised by all members of the United Nations with the exception of Turkey. The information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.

Source: OECD calculations based on Eurostat, Digital Economy and Society Statistics (database) https://ec.europa.eu/eurostat/web/digital-economy-and-society/data/comprehensive-database (last accessed April 2018).

 StatLink https://doi.org/10.1787/888933923089

Norway had the lowest share of cross-border e-commerce sellers, with only 20% making sales in other countries, the same level as in 2010. Other countries experienced significantly higher reductions in the share of firms participating in cross-border e-commerce, such as Ireland (from 84% to 57%), Malta (from 70% to 58%), Romania (33% to 24%), Latvia (from 49% to 41%), France (44% to 38%), Bulgaria (37% to 33%) and Estonia (from 49% to 45%).36

At the same time, the overall average is largely determined by small firms (42%), while the export share among large firms participating in e-commerce was significantly higher (55%). The difference was highest for Sweden (27 percentage points) as well as Denmark, Finland and Hungary (23 percentage points each), including some of the most advanced economies in terms of e-commerce uptake. The share of exporters was higher among small firms than among large firms for Cyprus,37 Greece, Lithuania and Latvia, which might be driven by sectoral differences.

Importantly, exporting activity among e-commerce firms was much less frequent beyond the borders of the European Union. On average, only 26% of all firms analysed exported to countries other than EU countries. The share was relatively high for economies like Cyprus,38 where roughly 85% of all exporters indicated to have exported to countries beyond the EU. The corresponding share was still above 65% for Greece, Ireland, Latvia, Malta, Norway, Portugal and the United Kingdom. For other countries, including the Czech Republic, Romania and Slovenia, the share of exporting e-commerce firms that went beyond European borders remained below 33%.39

Sectoral differences drive exports from European e-commerce firms

The share of exporters is depicted again in Figure 2.19 as a cross-country average across sectors. The figure distinguishes exporters according to whether they export to other EU countries only; to the rest of the world (RoW) but not to EU countries; or to both other EU countries and the RoW. Across sectors, this division reveals that of the 45% of exporters, 19% export to the EU only, 24% to both EU and the RoW, and 2% only to the RoW.40

2.19. Enterprises that participated in e-commerce sales to other countries by sector, 2016
As a percentage of enterprises that received e-commerce orders over the last calendar year, EU28
 2.19. Enterprises that participated in e-commerce sales to other countries by sector, 2016

Note: RoW = rest of the world. See Chapter notes.

1. Figure 2.19: RoW = Rest of world. Wholesale and retail trade exclude motor vehicles and motorcycles. Computer programming et al. includes consultancy and related activities and information service activities, rental and leasing et al. includes activities for employment, security and investigation, services to buildings and landscape, office administrative, office support and other business support. Professional et al. includes scientific and technical activities, trade of motor vehicles et al. includes motorcycles, electricity et al. includes gas, steam, air conditioning and water supply, publishing activities et al. includes motion picture, video and television programme production, sound recording and music publishing, programming and broadcasting.

Source: OECD calculations based on Eurostat, Digital Economy and Society Statistics (database) https://ec.europa.eu/eurostat/web/digital-economy-and-society/data/comprehensive-database (last accessed April 2018).

 StatLink https://doi.org/10.1787/888933923108

There is significant variation between sectors in terms of exporting behaviour. Accommodation, a sector where about 57% of total e-commerce is directed to end consumers, showed the highest trade intensity across all sectors, with 91% of all active e-commerce firms exporting their services to customers in other countries. Of these, more than 82% (75% of total) reported sales to customers in countries other than the EU. The travel agency sector follows with 65% of exporters, of which two-thirds exported to RoW countries (43% of all e-commerce firms). Both the manufacturing and the wholesale sector had lower shares of exporters, 47% and 39% respectively, indicating how open some consumer-facing sectors have become relative to B2B sectors that are at the core of GVCs. The highest share of exporters to RoW among all exporting e-commerce firms are found in the real estate sector, where 86% of exporters sell beyond European borders, even if the share of exporters in all e-commerce firms remained below average (37%). With only 38% of firms exporting and only 21% exporting to countries other than EU members, the retail sector seems to remain below potential when it comes to export performance even though the export intensity is relatively high when compared to offline firms as mentioned before.

EU consumers increasingly purchase online from abroad, but it has become more difficult to determine the origin of online goods and services

While there is relatively little evidence as to whether and how much individuals purchase online from other countries, Figure 2.20 provides insights into the percentage of individuals in EU countries who indicated they had purchased online from abroad in 2018. For the EU28, the data reveal that the percentage of individuals who made purchases from e-commerce firms in other EU countries had reached 21% (16% for RoW) by 2018.41

2.20. 2.20. Individuals who purchased online from sellers in other countries, 2018
As a percentage of all individuals aged 16 to 74
 2.20. 2.20. Individuals who purchased online from sellers in other countries, 2018

Note: See Chapter notes.1

1. Figure 2.20: Percentage of individuals who have purchased online from domestic sellers, sellers in other EU or sellers in non EU countries, i.e. rest of world (RoW). Data for Latvia is from 2017. This figure includes data for Kosovo. This designation is without prejudice to positions on status, and is in line with United Nations Security Council Resolution 1244/99 and the Advisory Opinion of the International Court of Justice on Kosovo’s declaration of independence. This figure contains data for “Cyprus”.

Note by Turkey:

The information in this document with reference to “Cyprus” relates to the southern part of the Island. There is no single authority representing both Turkish and Greek Cypriot people on the Island. Turkey recognises the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Turkey shall preserve its position concerning the “Cyprus issue”.

Note by all the European Union Member States of the OECD and the European Union:

The Republic of Cyprus is recognised by all members of the United Nations with the exception of Turkey. The information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.

Source: OECD calculations based on Eurostat, Digital Economy and Society Statistics (database) https://ec.europa.eu/eurostat/web/digital-economy-and-society/data/comprehensive-database (accessed March 2019).

 StatLink https://doi.org/10.1787/888933923127

The share of individuals who made online purchases from abroad (either countries in the EU or non-EU countries) was highest in Iceland (65%), Luxembourg (63%) and Malta (53%), probably due to the relatively small domestic market in these countries. In all three countries, the share of individuals who had purchased from sellers abroad was significantly higher than the share of individuals who had purchased domestically (49% for Iceland, 22% for Luxembourg, and 20% for Malta), underscoring the important role that e-commerce plays in accessing products unavailable in the domestic market. More than 45% of individuals report buying abroad in Austria, Denmark and Norway. The countries with the lowest shares of importing individuals were Turkey (3%) and Romania (3%).

The figure reveals that imports from the RoW were equally or more important than imports from other European countries to consumers in Bosnia and Herzegovina, Croatia, Iceland, Kosovo, Montenegro, the Republic of North Macedonia, Serbia and the United Kingdom, whereas RoW imports were relatively less important in Austria, Belgium and Romania.

Although overall these numbers seem to suggest significant potential for improvements in the international scope of online purchases, it is noteworthy that across the EU28 about 27% of individuals had purchased from e-commerce firms in other countries in 2018. This is surprising, given that imports at the individual level were relatively rare before the digital transformation.

The underlying data can also be used to analyse time trends (not shown).42 In particular, the data reveal a significant increase in cross-border purchases from only 6% for sellers from other EU countries (4% RoW) in 2008 to 21% in 2018 (16%). While these increases in the percentage of individuals who completed foreign purchases fell behind the increase in domestic purchases in absolute terms (i.e. percentage point changes), the relative increase of individuals that imported via e-commerce from other countries is nevertheless astonishing, with the share of importing individuals more than tripling for both the EU28 and the RoW.

Comparing the relative size groups over time implies that while there were roughly 4.7 domestic purchasers for every EU importer in 2008 and 7 domestic purchasers for every individual purchasing from the RoW, these relationships had fallen to 2.5 for EU imports and 3.3 for RoW imports only 10 years later.43

Importantly, though, the data also reveal a significant increase over time in the percentage of individuals reporting purchases from sellers with an unknown country of origin (from 2% to 7% over the 2008-18 time period). Accordingly, insecurity about a seller’s origin is rising at more or less the same rate at which individuals are increasingly purchasing from other countries. This has important implications for consumer protection and shows that multinational business models, often involving local websites or warehouses, are contributing to an increase in the complexity of the e-commerce landscape.

The products most frequently purchased across borders tend to be physical goods

Figure 2.21 provides some insights into the type of goods that consumers in the EU28 purchased from abroad in 2017. The figure shows the percentage of all individuals engaging in e-commerce imports that have purchased a certain type of product. The data reveal that 80% of individuals purchased physical goods from abroad. Interestingly, the corresponding numbers are significantly lower for intangible products including travel arrangements and accommodation (34%) and downloaded products including e-books, videos or music (25%). The relatively high percentage of individuals purchasing physical goods is surprising, given that they involve transportation costs. Several of the less frequently imported services on the other hand can be digitally purchased, even if they might require travel to other countries in order to be consumed (e.g. accommodation services). What this potentially indicates is that consumers might encounter difficulties in finding domestic providers for certain physical products, whereas digital content tends to be more readily available from domestic suppliers. One explanation for this could relate to the fact that it is significantly cheaper to set up a website in several countries to distribute digital content domestically than it is to set up a local distribution centre for physical goods, which often requires storage facilities and warehousing.44

2.21. Products that individuals purchased online from sellers abroad, 2017
As a percentage of individuals that purchased online from sellers abroad, EU28
 2.21. Products that individuals purchased online from sellers abroad, 2017

Note: See Chapter notes.1

1. Figure 2.21: Percentage of individuals that has purchased online from sellers abroad (other EU or non-EU countries). Physical goods: e.g. electronics, clothes, toys, food, groceries, books and CDs/DVDs. Travel, accommodation or holiday arrangements (e.g. tickets and documents by mail or printed by oneself). Products downloaded or accessed from websites or apps (e.g. films, music, e-books, e-newspapers, games and paid applications). Other services (e.g. tickets for events received by mail and telecom subscriptions).

Source: OECD calculations based on Eurostat, Digital Economy and Society Statistics (database) https://ec.europa.eu/eurostat/web/digital-economy-and-society/data/comprehensive-database (database) (accessed March 2019).

 StatLink https://doi.org/10.1787/888933923146

There are very few official sources for statistics on countries outside of the EU, but private sector survey data from the Google Consumer Barometer 2014/15 (Google, 2015[28]) suggest that the overall ranking of product categories that the official statistics indicate for domestic sales also apply to international sales. Unfortunately, the categories included in the Google data relate almost exclusively to goods (not services).45

Key areas for policy action

As digital transformation progresses, the economy will continue to digitalise, with positive impacts on economic efficiency and convenience in many cases. As a result, those who do not engage in e-commerce may find themselves on the wrong side of a potentially persistent and harmful digital divide. Policy can help foster the participation of actors, from SMEs to older individuals, that still do not buy or sell online.

Innovative business models can help address the specific needs of some of these groups, for example by offering solutions that help SMEs to sell online or by providing alternative payment mechanisms. However, there are a number of challenges to business innovation due to either low incentives to invest or cumbersome regulation. For example, complex rules can leave customers responsible for unforeseen duties, taxes or burdensome return requirements. Transparent rules, consistently applied at the border across both digital and brick-and-mortar firms, can reduce some of the resulting uncertainties.

With respect to individuals, significant gaps remain with respect to education, income and age, but also gender and for households in rural areas. Factors that inhibit participation of these groups are often related to economic and social conditions that reach far beyond e-commerce, including urban-rural divides, income distribution, unequal access to education or an aging society. With regard to e-commerce, these conditions may manifest themselves in low connectivity, a lack of ICT skills, low levels of trust or a lack of viable payment options – factors that can all be addressed by policy action. Relevant measures in this regard include targeted information campaigns, trust building initiatives, adult training, as well as public-private partnerships that target the participation of low-income households and those in rural areas.

In the case of firms, data suggest that SMEs still lag behind in terms of e-commerce participation. This is true despite the emergence of web-based and standardised solutions specifically targeting these firms. In many cases, this is related to high costs of delivery and returns, a problem that SMEs face significantly more often than other firms (European Commission, 2015[11]). Some business models have emerged that aim to boost firm participation in e-commerce (see Chapter 3). Updating regulations to overcome bottlenecks in areas such as postal services or custom clearance may help in this respect.

In addition, SMEs are likely to struggle more with regulatory uncertainty, as they often lack the financial means to obtain the required legal expertise. This carries over to the relationships between SMEs and larger service providers, such as online platforms. In particular, the EC recently proposed new rules on transparency and fairness to foster a predictable and trusted business environment for both SMEs and online platforms.46 With respect to cross-border e-commerce, governments should also be open to emerging multi-stakeholder initiatives such as the Electronic World Trade Platform (eWTP), a public-private sector dialogue that could foster a more effective policy environment, providing better access to online trading opportunities in particular for SMEs.

Notes

Israel

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

1. Evidence for the US shows that credit cards are disproportionately prevalent among households with higher levels of income, more education and among non-hispanic white adults (Federal Reserve Board, 2017[32]).

References

Allinson, G. et al. (2015), “Digital capabilities in SMEs: Evidence review and re-survey of 2014 small business survey respondents”, Department for Business, Innovation and Skills, BIS Research Paper, No. 247, London. [58]

Capgemini (2017), “World payments report 2017”, https://www.capgemini.com/fr-fr/wp-content/uploads/sites/2/2017/10/world-payments-report-2017_year-end_final_web-002.pdf (accessed on 19 April 2018). [4]

CRC (2017), “El comercio electronico en Colombia”, https://www.crcom.gov.co/recursos_user/2017/ComElecPtd_0.pdf (accessed on 23 April 2019). [34]

E-Commerce Foundation (2016), “Global B2C e-commerce report 2016”, https://www.ecommercewiki.org/wikis/www.ecommercewiki.org/images/5/56/Global_B2C_Ecommerce_Report_2016.pdf (accessed on 19 April 2018). [23]

eMarketer (2018), “Worldwide retail and ecommerce sales: EMarketer’s updated forecast and new mcommerce estimates for 2016-2012”, https://www.emarketer.com/Report/Worldwide-Retail-Ecommerce-Sales-eMarketers-Updated-Forecast-New-Mcommerce-Estimates-20162021/2002182 (accessed on 19 April 2018). [1]

European Commission (2017), “Final report on the e-commerce sector inquiry”, http://ec.europa.eu/competition/antitrust/sector_inquiry_final_report_en.pdf (accessed on 22 April 2019). [12]

European Commission (2015), “Flash Eurobarometer 413: Companies engaged in online activities”, http://ec.europa.eu/commfrontoffice/publicopinion/flash/fl_413_en.pdf. [11]

European Commission (2015), “Provision of two online consumer surveys as support and evidence base to a Commission study: Identifying the main cross-border obstacles to the Digital Single Market and where they matter most”, https://ec.europa.eu/info/sites/info/files/digital-single-market-final-report_en.pdf. [21]

Eurostat (2017), “Digital economy and society: ICT usage in households and by individuals”, https://ec.europa.eu/eurostat/web/digital-economy-and-society/data/comprehensive-database. [13]

Farag, S. et al. (2006), “E-shopping in the Netherlands: Does geography matter?” Environment and Planning B: Planning and Design, Vol. 33, pp. 59-74, https://dx.doi.org/ 10.1068/b31083. [14]

Federal Reserve Board (2017), “Report on the economic well-being of U.S. households in 2016”, https://www.federalreserve.gov/publications/2017-economic-well-being-of-us-households-in-2016-banking-credit.htm. [30]

GlobalData (2017), “Online consumer payments: Seizing the opportunity”, Report Code: GDF0009IA, GlobalData UK Ltd, https://www.globaldata.com/store/report/gdf0009ia--online-consumer-payments-sizing-the-opportunity. [2]

Google (2015), “Consumer barometer from Google”, https://www.consumerbarometer.com/en/insights/?countryCode=IE (accessed on 20 April 2019). [28]

Jabri, P. (2016), “Low broadband connectivity areas to have e-commerce easy shops soon”, Business Recorder, https://www.thenewsminute.com/article/ecommerce-goes-rural-startup-helping-villagers-buy-amazon-70580 (accessed on 13 July 2018). [17]

Liao, R. (2018), “Journey to the West: How China’s online retailers are taking the battle to the rural front”, Techinasia, https://www.techinasia.com/china-ecommerce-rural-small-cities (accessed on 13 July 2018). [15]

McDermott, K. (2016), “The mobile payments report 2016 – An omnichannel evolution”, https://www.payvision.com/downloads/mobile-payments-report-2016-an-omnichannel-evolution.pdf (accessed on 20 April 2019). [7]

METI Japan (2016), “Results compiled of the e-commerce market survey”, http://www.meti.go.jp/english/press/2016/0614_02.html. [31]

METI Japan (2018), “Internet buying boom: METI releases results of FY2017 e-commerce market survey”, https://www.meti.go.jp/english/press/2018/0425_002.html. [33]

OECD (2019), “BBVA big data on online credit card transactions: The patterns of domestic and cross-border e-commerce”, OECD Digital Economy Papers, No. 278, OECD Publishing, Paris, https://doi.org/10.1787/8c408f92-en. [18]

OECD (2018), “Consumer protection enforcement in a global digital marketplace”, OECD Digital Economy Papers, No. 266, OECD Publishing, Paris, http://dx.doi.org/10.1787/f041eead-en. [20]

OECD (2013), “Electronic and mobile commerce”, OECD Digital Economy Papers, No. 228, OECD Publishing, Paris, http://dx.doi.org/10.1787/5k437p2gxw6g-en. [10]

OECD (2011), OECD Guide to Measuring the Information Society 2011, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264113541-en. [8]

OECD (1999), The Economic and Social Impact of Electronic Commerce – Preliminary Findings and Research Agenda, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264172081-en. [9]

Paypal (2018), “PayPal cross-border consumer research 2018”, PayPal, https://www.paypalobjects.com/digitalassets/c/website/marketing/global/shared/global/media-resources/documents/PayPal_Insights_2018_Global_Report.pdf (accessed on 19 April 2019). [3]

Payvision (2017), “Payvision’s annual report reveals cross-border e-commerce trends - In 2017, with online marketplaces and consumer technology leading the way for growth”, https://www.payvision.com/edit-press-release-payvision%E2%80%99s-annual-report-reveals-cross-border-ecommerce-trends-2017 (accessed on 19 April 2018). [5]

Ranipeta, S. (2017), “Ecommerce goes rural: This startup is helping villagers buy on Amazon”, The News Minute, https://www.thenewsminute.com/article/ecommerce-goes-rural-startup-helping-villagers-buy-amazon-70580 (accessed on 13 July 2018). [16]

UNCTAD (2017), “Consumer protection in electronic commerce”, TD/B/C.I/CPLP7, Note by the UNCTAD Secretariat, http://unctad.org/meetings/en/SessionalDocuments/cicplpd7_en.pdf (accessed on 19 April 2018). [22]

UNCTAD (2017), ICT Policy Review: National E-commerce Strategy for Egypt, UNCTAD/DTL/STICT/2017/3, United Nations Publishing, Geneva, https://unctad.org/en/pages/PublicationWebflyer.aspx?publicationid=1956. [29]

UNCTAD (2016), “In search of cross-border e-commerce trade data”, UNCTAD Technical Notes on ICT for Development, United Nations Publishing, Geneva, https://unctad.org/en/PublicationsLibrary/tn_unctad_ict4d06_en.pdf. [32]

UNCTAD (2015), “International trade in ICT Services and ICT-enabled services: Proposed indicators from the Partnership on Measuring ICT for Development”, UNCTAD Technical Notes on ICT for Development, https://unctad.org/en/PublicationsLibrary/tn_unctad_ict4d03_en.pdf. [26]

United Postal Union (2018), “Postal economic outlook 2018”, http://www.upu.int/uploads/tx_sbdownloader/postalEconomicOutlook2018En.pdf. [25]

US Department of Commerce (2016), “New BEA estimates of international trade in digitally enabled services”, http://www.esa.doc.gov/economic-briefings/new-bea-estimates-international-trade-digitally-enabled-services (accessed on 22 April 2018). [27]

World Customs Organization (2015), “Annual Report 2014-2015”, http://www.wcoomd.org/-/media/wco/public/global/pdf/media/annual-reports/wco_annrep_en_2014_2015.pdf. [24]

Worldpay (2017), “Global payments report 2017”, Worldpay LLC, https://www.worldpay.com/global/insight/articles/2017-11/global-payments-report-2017 (accessed on 22 April 2019). [6]

Notes

← 1. Both percentages and total only relate to the sectors depicted in Figure 2.1. As the sectors included and the classification of sectors are not identical for both regions, numbers should not be directly compared. European figures relate to firms with 10 or more employees. Total sector level turnover was not available for some sectors. 2016 was the last available year in Eurostat’s Structural Business Statistics and for US census data at the time of drafting. Data from the Digital Economy and Society database come from the 2017 survey and relate to e-commerce turnover in 2016.

← 2. The service sector in the US data captures slightly more activities than the European data. Activities related to electricity, gas, steam, air conditioning and water supply on the other hand are not captured in the US data.

← 3. For Japan, the total e-commerce data is broken down into B2C and B2B transactions. Data from the Ministry of Economy, Trade and Industry suggest that B2C e-commerce in 2016 accounted for JPY 15.1 trillion, while B2B transactions (including web sales and EDI) reached JPY 291 trillion in the same year. By 2017, B2C e-commerce had risen to JPY 16.5 trillion and B2B sales to JPY 317.2 trillion, implying a B2C contribution of 5.2%, up from 4.6% in 2015. See METI Japan (2016[31], 2018[33]).

← 4. Data for the EU28 is used because of the greater consistency in terms of coverage and reporting over time.

← 5. OECD calculations based on Eurostat, Digital Economy and Society: ICT Usage in Enterprises (database) (last accessed April 2019). Data across industries is available from 2013 onwards only. Survey data on e-commerce activities by firms relates to the previous year (e.g. data from the 2013 survey relates to activities in 2012).

← 6. See also Chapter 1.

← 7. OECD calculations based on Eurostat data. Eurostat data is used for greater coherence.

← 8. The Eurostat survey design attributes sales via EDI in general to either B2B or B2G transactions. All B2C transactions are therefore stemming from web sales. See Box 1.1.

← 9. Firms can engage in both EDI and web sales at the same time.

← 10. While 94% of large EU28 firms with web sales also realised sales through their own website in 2017, the corresponding percentage is 86% for small firms.

← 11. The average across time is used to highlight more structural sector level differences. Data is not available for most OECD countries outside of the European Union.

← 12. Among the sectors that focus mostly on final consumers (share of B2C transaction is larger than 50% of total e-commerce transactions), the retail sector sticks out with a relatively high share of EDI transactions (76%) in total B2B transactions, probably reflecting a higher share of physical goods compared to sectors like accommodation or travel agencies, where the corresponding EDI shares in B2B sales are around 30% and 38%, respectively.

← 13. An important limitation of survey evidence on e-commerce challenges is that in most cases respondents can choose from a closed set of possible responses only, restricting the scope of challenges that are represented in the data. In addition, perceived challenged might be different from actual challenges encountered.

← 14. See also UNCTAD (2015[26]).

← 15. As the web is specified as the sales channel for these survey questions, Panel B of Figure 2.9 may contain some responses from firms that have received sales via EDI.

← 16. The data also confirm earlier findings for SMEs in the United Kingdom, where 80% of the sampled firms with a website but no plans to introduce any form of e-commerce (521 firms) indicated that they were not selling goods or services that can be ordered directly as a reason. The number was particularly high for the business services sector (92%) but still identified by 60% of SMEs active in the retail, transport and the food sector. Other reasons identified were lack of relevancy (4%), high cost (3%), other reasons (6%) or no particular reason (5%), see Allinson et al. (2015[58]).

← 17. See European Commission (2015[11]) for significant cross-country differences.

← 18. Restrictions from business partners could involve for example producers that prohibit sales via online platforms for downstream retailers.

← 19. In Canada, Colombia and Japan there was a break in survey methodology over time, which might be partly responsible for this result.

← 20. There is no generally acknowledged definition of the terms “Baby Boomers” or “Generation Z”. As the age groups captured in the figure roughly coincide with the birth years typically associated with both groups the terms are used for illustrative purpose. In a statistical sense, all findings should be understood as related to the stated age groups only.

← 21. Average for younger users excludes data for Canada and New Zealand to allow for comparison with older users.

← 22. In Colombia, Israel, New Zealand and the United States, e-commerce penetration was below the country average for the youngest generation, driven largely by a higher penetration for intermediate age groups.

← 23. Due to missing data for 2009, the comparison over time excludes data for Australia, Canada, Chile, Israel, Japan and New Zealand.

← 24. Recent data from the Ministry of Communications and Information Technology in Egypt in co-operation with UNCTAD (UNCTAD, 2017[32]) suggest that the gender gap is potentially more striking in other countries. In Egypt, about 5% of individuals age 15 or older that had used the Internet in the last three months had participated in the purchase or ordering of goods or services via the Internet over the past year. Of these purchasers, more than two-thirds were male (69%) whereas less than one third was female (31%).

← 25. A detailed analysis of the Colombian e-commerce market is provided by the national telecommunications regulator, highlighting factors like transportation infrastructure, a lack of payment methods and trust as the main barriers to e-commerce. See (CRC, 2017[34]).

← 26. Due to missing data for 2009, comparison over time excludes data for France and Ireland.

← 27. Education is grouped by individuals with at most lower secondary (ISCED 0, 1 or 2), individuals with upper or post-secondary, but not tertiary (ISCED 3 or 4) and individuals with tertiary (ISCED 5 or above) education.

← 28. Data for Costa Rica relates to 2015, Brazil to 2016, Chile, Colombia and Mexico to 2017.

← 29. Due to missing data for 2009, comparison over time excludes data for New Zealand.

← 30. OECD calculations based on OECD, ICT Access and Usage by Households and Individuals (database). Data from the 2018 survey. For Chile, Costa Rica, Korea, Mexico and Switzerland, data are from 2017. For Australia, data are from 2016 and wording of some questions differs slightly to that requested. For Japan, data are from 2015. For Japan the age range is 15 to 69 and data is “in the last 12 months”. In Korea, medicine can’t be traded online and alcohol and tobacco are not included. For Mexico, wording of some questions differs from that requested. For Switzerland, the data provided correspond to individuals aged 18 to 74.

← 31. The remaining categories were: dislike going to shops (12%), ability to return products easily (9%) and the ability to find better quality products online (5%).

← 32. See note 67 for details on the survey.

← 33. International parcel deliveries have been proposed as a proxy for the increase in cross-border e-commerce although the data does usually not allow distinguishing deliveries by ordering mode. See (UNCTAD, 2016[32]) for a discussion.

← 34. For instance, while an architectural design file can in principle be provided over ICT networks, the ordering process typically involves complex demand specifications that often require ongoing communication via e-mails, phone calls, video conferences or actual meetings – and thus would not comply with the characteristics of an e-commerce transaction. Nevertheless, innovations in both technology and business models might increasingly lead to an overlap between services that can be delivered and ordered online.

← 35. Includes 43% of firms that have undertaken electronic sales to other EU countries and 2% that have undertaken electronic sales to the rest of the world but not to other EU countries.

← 36. From the available data it is impossible to infer what drives these results at the country level. Structural changes, including at the sectoral level as well as issues related to data consolidation might be responsible for the results.

← 37. Note by Turkey:

The information in this document with reference to “Cyprus” relates to the southern part of the Island. There is no single authority representing both Turkish and Greek Cypriot people on the Island. Turkey recognises the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Turkey shall preserve its position concerning the “Cyprus issue”.

Note by all the European Union Member States of the OECD and the European Union:

The Republic of Cyprus is recognised by all members of the United Nations with the exception of Turkey. The information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.

← 38. Note by Turkey:

The information in this document with reference to “Cyprus” relates to the southern part of the Island. There is no single authority representing both Turkish and Greek Cypriot people on the Island. Turkey recognises the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Turkey shall preserve its position concerning the “Cyprus issue”.

Note by all the European Union Member States of the OECD and the European Union:

The Republic of Cyprus is recognised by all members of the United Nations with the exception of Turkey. The information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.

← 39. For a proper interpretation of these numbers a comparison with offline firms would be required.

← 40. The breakdown can deliver interesting results from a country perspective. In Sweden for example, the share of exporters among e-commerce firms increased from 36% to 48% between 2010 and 2016. This increase was exclusively driven by exporters that only sold to RoW (from 2% to 15%), while the share of exporters to the EU slightly diminished from 34% to 33%.

← 41. Surveyed individuals were asked whether they purchased goods or services from sellers in any of four possible regions: domestic sellers, sellers in other EU countries, sellers in the rest of the world (RoW) or sellers abroad (other EU or RoW). The former three are depicted in Figure 2.20. Additionally, respondents were able to select a separate option when they were not sure about the seller’s location. The responses are not mutually exclusive and thus individuals could choose several options simultaneously.

← 42. In this context, it is important to highlight the extent to which dynamics that are true for the EU might also be true for other OECD countries. A comparison with regard to domestic e-commerce can provide some insights in this regard. Overall, the trends for European and OECD countries seem to be quite aligned, driven not least by the significant weight that EU countries have in the OECD average. Thus, according to the Eurostat data, the percentage of domestic online customers in the EU28 increased from 36% in 2009 to 60% in 2018. Figure 2.13 shows that the corresponding value was 35% across OECD countries and confirms a similar increase to 57% by 2018. Considering only the non-EU countries in Figure 2.13 (Australia, Brazil, Canada, Chile, Colombia, Iceland, Israel, Japan, Korea, Mexico, New Zealand, Norway, Switzerland, Turkey and the United States) implies an increase from 30% to 45%. These numbers are lower than for the EU partly due to the relatively low uptake in countries like Brazil, Chile, Colombia, Mexico and Turkey and partly because data for non-EU countries is often less up-to-data than the data provided by Eurostat, dating back to 2012 in some cases (e.g. for Canada and New Zealand).

← 43. This comparison is not precise and illustrative only. In particular, while most individuals with imports from RoW will also show up in the group of domestic purchasers or EU importers, others might not. Thus, the Eurostat data reveal that while in 2018 roughly 21% had purchased from other EU countries, the total percentage of individual importers from anywhere was as high as 27%, implying at least 6% of individuals that had purchased from countries outside of the European Union but not from within.

← 44. In the light of the findings presented earlier, suggesting that individuals are increasingly less confident about the location of an online seller, it is noteworthy that transportation costs can be an important indicator of transactions involving foreign sellers. It is therefore likely that customers relatively more often mistake cross-border purchases for domestic purchases, when the purchases involves digital services rather than goods. This would imply that cross-border purchases of digital imports have a higher probability of being underrepresented in the data than goods.

← 45. See Consumer Barometer Survey 2014/15 with Google (www.consumerbarometer.com). Survey evidence based on 56 391 respondents in Australia, Austria, Belgium, Canada, Chile, the Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Israel, Italy, Japan, Korea, Latvia, Mexico, the Netherlands, New Zealand, Norway, Poland, Portugal, the Slovak Republic, Slovenia, Spain, Sweden, Switzerland, Turkey, the United Kingdom and the United States. (Question: What type of product have you ever purchased online from abroad?)

← 46. https://ec.europa.eu/growth/content/online-platforms-commission-sets-new-standards-transparency-and-fairness_en.

End of the section – Back to iLibrary publication page