Giorgio Gualberti
Development Co-operation Directorate, OECD
Jonas Wilcks
Development Co-operation Directorate, OECD
Giorgio Gualberti
Development Co-operation Directorate, OECD
Jonas Wilcks
Development Co-operation Directorate, OECD
Reflecting the cross-sectoral nature of digitalisation and digital transformation, development finance supports a range of activities and investments in digital infrastructure. However, the absence of explicit reporting guidance or a policy marker in the OECD Creditor Reporting System makes measuring and tracking official development finance for digitalisation more difficult. This chapter provides the first estimates of multilateral, bilateral and philanthropic development finance for digitalisation from 2015 to 2019, based on a methodology that combines relevant sector codes, keyword searches and Sustainable Development Goal tagging. The database suggests that this type of finance increased dramatically in recent years, though a few providers constitute the bulk of finance for digitalisation. The chapter discusses options for increasing the transparency of finance for digitalisation through better reporting guidance and statistical measures.
While some DAC members are developing markers to help track development finance in support of digitalisation, an agreed statistical method would be needed to better measure, co-ordinate and account for these investments.
Development finance for digital activities more than tripled between 2015 and 2019, with providers investing a total of USD 18.6 billion and mobilising another USD 4.2 billion in private finance, according to initial estimates based on the Creditor Reporting System.
Finance for digitalisation is increasing in volume, accounting for 1% of bilateral development finance, 2.7% of multilateral development finance and 4.6% of philanthropic development finance in 2018-19.
Measuring development finance for digitalisation is important to track the overall level of financial investment in digital transformation,1 to map and understand the different roles financing actors play, and to assess whether financing aligns with stated development objectives. With this information and insight, development co-operation providers will be better placed to strategically target finance for digital needs and gaps that impact development results, and to tailor their financing to the digital readiness of partner countries.
The gaps in financing for digitalisation in low- and middle-income countries affect all aspects of the digital transformation from capital and infrastructure investments to enable access, to digitalising government, services, the economy and industry, to equipping people and users with the right digital skills and literacy and many other areas as identified in chapters through-out this report. There is scope for much more transparency of finance for digital transformation by all relevant actors – public and private, domestic and international. There are also definitional and technical aspects to resolve, notably on how to measure funding for such a cross-sectoral phenomenon. This chapter offers initial answers and reflections on next steps. The first section describes the methodology the authors used to compile a dataset extracted from the OECD Creditor Reporting System (CRS) that estimates trends in development finance for digitalisation and discusses general measurement challenges. These estimates, presented in the second section, cover 2015‑19 for several bilateral, multilateral and philanthropic providers, and look at geographic and sectoral trends. The last section examines options to improve measurement and tracking of development finance in this space.
The OECD Development Assistance Committee statistical system does not have specific guidance or markers for reporting and tracking finance for digitalisation. To estimate development finance for digitalisation and digital transformation, this report used the same method as for other themes and issues where guidance does not exist. While the findings are robust, the process of analysing the CRS database to calculate the level of financing for digital transformation nevertheless raises methodological and analytical questions. For more accurate tracking and transparency, DAC members and other development finance actors should consider examining and agreeing on the most appropriate methods.
Accurately tracking financing for digitalisation is challenging. First, there is no standardised, general definition. Digitalisation is the adoption of new digital technologies, and its economic and societal impacts (see Reader’s Guide). This financing can take many forms and support a range of activities: introducing digital infrastructure such as networks, computing and communication tools; developing (through training, education etc.) the broad set of skills and technical abilities required to take advantage of digital technologies; and implementing organisational changes that take advantage of new technologies and enable new activities based on digital technologies.
Second, apart from the clear-cut investments in hard digital infrastructure that appear to be reported under the information and communication technology (ICT) sector code,2 most support for digitalisation and digital transformation is cross-sectoral. Related activities might well be in any sector – education (e.g. curricula development), health (telemedicine and diagnostic tools), banking (mobile banking), government (digitalisation of public institutions and e-government), and energy (smart grids and distributed renewable energies), among others. Furthermore, it is difficult to identify spending on digital capacity and skills-building, and support for digital policy reform, accountability and knowledge sharing, as they are not necessarily large budget expenditures and tend to be integrated within larger programmes.
Finally, CRS data collection does not include a specific tool to track financing for digitalisation. While some activities could be isolated through a series of sector codes (notably in the communication sector), activities that support digitalisation in other sectors can be identified only through a series of tailored techniques. These include looking at digitalisation-related Sustainable Development Goal (SDG) targets and at the descriptive information of activities by text-mining keywords, complemented by manual screening.
In 2020, the European Union (EU) developed a marker to track investment in digitalisation (Box 40.1). This marker identified a yearly average of USD 340 million in digitalisation-related commitments from EU institutions for 2020-21. The methodology used by in this report identified USD 205 million in digitalisation-related support by EU institutions for 2018-19. While they cover different time periods and are not directly comparable, the results of the two methodologies are quite similar. This suggests that the method developed to track DAC digitalisation support does not overestimate digitalisation-related finance and that the estimates presented in this chapter are robust.
By authors with input from EU colleagues
The European Commission developed a marker to track digitalisation activities and began implementing it with 2020 data. It was presented at the 2020 meeting of the Working Party on Development Finance Statistics (WP-STAT) (OECD, 2020[1]). The internal EU marker is designed to track actions that promote the following digital transformation objectives:
governance, policy and regulatory frameworks relevant to digitalisation and the digital economy
access to affordable and secure broadband connectivity, and digital infrastructures
digital literacy and skills
digital entrepreneurship and job creation
use of digital technologies as enablers for sustainable development (e.g. digital and e-services, including e-governance).
The EU policy marker uses the same three-value scoring system as DAC policy markers (OECD, 2020[2]). An activity can be scored as “2” when digitalisation is its principal (or primary) objective, as “1” when digitalisation is a significant (or secondary) objective among others and as “0” when the activity is evaluated as unrelated to digitalisation.
The methodology developed also outlines three steps to determine if an action should be considered related to digitalisation:
1. Analyse the digitalisation context to facilitate identification and articulation of the action’s digital component and inform future steps.
2. Identify the existence of a digitalisation context, specific objective or result.
3. Disaggregate indicators and data by sex, age, socio-economic status and region, where appropriate and applicable.
The data informing the estimates provided in this chapter were selected from commitments reported in the CRS by bilateral and multilateral providers and private philanthropic institutions. The CRS activity-level data were supplemented by aggregated data on private finance mobilised by official interventions, to obtain a broader picture of the finance that supports digitalisation.
The following criteria were used to identify the data:
sector codes in communications (communications policy and administrative management, and telecommunications and ICT)
keywords in the title or description of the activity reported (e-governance, e-health, telemedicine, blockchain, artificial intelligence, machine learning, digital, Internet, electron, ICT, online, telecom, software, e-commerce)
SDG targets (2.a, 5.b, 8.2, 8.3, 9.b, 9.c, 17.6, 17.7, 17.8).
Activities that matched at least one of these criteria were included in the dataset without double-counting activities that corresponded to multiple criteria. Manual screening of the largest activities selected (accounting for 88% of total development finance) was used to exclude activities not related to digitalisation. Figure 40.1 illustrates the steps taken to select the data. Figure 40.2 shows the share of activities matching each criteria.
Source: Authors' illustration.
The authors tested various combinations of keywords and SDGs to complement data selected through a set of purpose codes in the communication sector. This empirical approach also led to the elimination of some keywords in building the final sample. For example, keywords such as technology and communication, were tested in the manual search but not included for estimating development financing for digitalisation because the results showed a high number of activities unrelated to digitalisation. Including them risked inflating the results. For the same reason, only activities exclusively marked with SDG targets of interest were included as data sources.3 Activities marked with multiple SDGs showed activities with weaker and/or limited digitalisation focus.
Finally, the largest activities in terms of budget were manually checked to ensure they invested in digitalisation. Slightly more than 1 100 records were checked, representing about 88% (in value terms) of the activities identified by sectors, keywords and SDGs. Activities accounting for about 8% of the finance considered were excluded through manual screening because they were not closely related to digitalisation.
The data on private finance mobilised by official interventions are partially confidential and thus treated separately. The manual screening of this dimension was limited to activities reported by some bilateral donors that publicly disclose mobilisation data in the CRS database. Some of the mobilisation data were obtained in aggregated form by sector codes alone.
Approximately 15 000 digitalisation-related development finance activities are included in the estimates for 2015-19. Of these, bilateral providers reported about 10 766 activities (totalling USD 6.3 billion), multilateral providers reported 2 457 activities (totalling USD 10.3 billion), and philanthropic institutions reported 1 903 (totalling USD 1.2 billion). These figures exclude private finance mobilised. While records in the CRS database do not necessarily correspond to projects, the data indicate that multilateral organisations report larger programmes than bilateral providers, which is to be expected.
Of the three criteria used to create the estimates, keywords identified the largest proportion (75%) of digitalisation-related financing. The three that tagged most financing were digital, ICT and telecom, while other keywords on newer digital technologies that are promoted as potential accelerators for developing countries, such as blockchain or (Sirimanne and Freire, 2021[3]; Deshmukh, 2020[4]), receive far smaller shares. Communication and ICT as a sector accounted for 42% of bilateral development finance for digitalisation.
Note: A single activity can be identified by more than one criterion. Thus, the sum of shares indicated by all the bars exceeds 100%. Keywords are in blue, SDGs in green and the communication sector in orange for clarity.
Source: Authors’ estimates based on data available in the CRS database.
Development finance for digitalisation grew significantly over 2015-19. Using this methodology, the data indicate that development co-operation providers and philanthropic institutions have been investing increasing volumes of development finance in activities related to digitalisation and digital transformation.4
Over the five-year period, the authors estimate digital-related official development finance from bilateral and multilateral donors and philanthropic foundations totalled USD 18.6 billion. Bilateral and multilateral organisations mobilised more than USD 4.2 billion in additional private finance.
Bilateral and multilateral development finance and finance from philanthropic institutions more than tripled over 2015-19, increasing from USD 2 billion in 2015 to USD 6.8 billion in 2019 (Figure 40.3). The volume in the latest two years taken into consideration – 2018‑195 – represents 1.8% of the total bilateral, multilateral and philanthropic commitments. To put these figures in perspective, these institutions’ financing for digitalisation in 2019 is of the same order of magnitude as their commitments to the industry sector (USD 7.0 billion) and to renewable energy sources (USD 7.7 billion).
Development finance for digitalisation from multilateral institutions alone more than quadrupled, rising from USD 1.0 billion in 2015 to USD 4.2 billion in 2019. Multilateral institutions represented 62% of the total committed by multilateral and bilateral providers and philanthropic institutions in 2019.
Bilateral providers’ commitments to digitalisation-related activities also increased over the period analysed, more than doubling from USD 908 million in 2015 to USD 2.1 billion in 2019. DAC members account for 96.5% of the bilateral finance covered by this analysis.
Private philanthropic institutions’ support to digitalisation also grew, reaching USD 491 million in 2019, doubling the value recorded in 2017. Data collected from philanthropic institutions grew in recent years, so values prior to 2017 are surely underestimated.
Source: Authors’ estimates based on the data available in the CRS database.
In relative terms, philanthropic institutions devote a greater share of their investments to support digitalisation than do bilateral and multilateral providers. Digitalisation-related activities accounted for 4.6% of the 2018-196 portfolio of philanthropic institutions, compared to 2.7% for multilateral institutions and 1% for bilateral providers (Figure 40.4).
Source: Authors’ estimates based on the data available in the CRS database.
According to the data, bilateral and multilateral institutions also mobilised additional private finance in the amount of USD 700 million in 2019, divided roughly equally between the two (Figure 40.5). With a large share benefiting the financial sector, such activities can foster innovative banking services, including through digitalisation. However, confidentiality restrictions on multilateral development banks’ data on mobilisation prevent more granular analysis.
Note: Amounts are in current USD. Data from 2017 to 2019 are provided by a smaller set of reporters than in 2015-16.
Source: Authors’ estimates based on the data available in the CRS database and on data submitted confidentially to the OECD.
The data analysed reflect activities related to digitalisation reported by more than 100 bilateral, multilateral and philanthropic institutions over 2015-19. However, just ten providers account for 68% of the total estimated digital-related development finance over the period. Multilateral organisations finance was primarily (72%) non-concessional. Bilateral providers, mainly members of the DAC, provided 92% of concessional flows, or official development assistance (ODA), provided between 2015 and 2019. Philanthropic finance is exclusively grant based. Figure 40.6 breaks down the concessionality of multilateral and bilateral development finance for digitalisation.
Note: Concessional finance refers to grants, soft loans (provided with more favourable conditions than market terms) and other funding options. Non-concessional loans are provided at or near market terms.
Source: Authors’ estimates based on the data available in the CRS database.
Estimates indicate that, of the 40 bilateral providers in the dataset (30 DAC members plus ten other countries reporting their development finance to the CRS), five – EU Institutions, France, Germany, Korea and the United States – collectively provided over 60% of total bilateral development finance for digitalisation over 2015-19 (Figure 40.7). Three bilateral providers are estimated to having committed 10% or more of their portfolio to activities in support of digitalisation: Kazakhstan (17%), Estonia (15%) and Korea (10%).
Source: Authors’ estimates based on the data available in the CRS database.
The same trend can be seen among multilateral providers. Collectively, the estimates show that five institutions accounted for 78% (USD 8.6 billion) of the USD 11.1 billion committed by multilateral organisations over the five years, as reported to the CRS. In descending order, these are the Inter‑American Development Bank (IADB), the International Development Association, the International Bank for Reconstruction and Development, the Asian Development Bank, and the International Finance Corporation (Figure 40.8). Among multilateral institutions, the IADB is estimated to have the highest share of digitalisation-related commitments in its portfolio (10%), followed by the Inter-American Investment Corporation (7%), an affiliate of the IADB, and the World Tourism Organization (7%).
Source: Authors’ estimates based on the data available in the CRS database.
Development finance for digitalisation is also concentrated among a few philanthropic institutions. In absolute terms, the Bill & Melinda Gates Foundation is estimated to be the largest philanthropic provider of digital-related finance, committing 4% of investment, or USD 556 million, over 2015-19. The MasterCard Foundation was the second largest philanthropic provider, committing 19% of its portfolio, or USD 161.7 million, to digital projects over the period, and the Wellcome Trust was third with over USD 80 million, or 10% of its portfolio. In relative terms, some foundations provide a very large share of their total commitments to digitalisation-related activities: La Caixa Banking Foundation, 37%; Fondation Botnar, 27%; MasterCard Foundation, 19%; and MetLife Foundation, 17%, according to the estimates.
Africa received the most bilateral development finance for digitalisation of any region (37.9%), with sub‑Saharan countries alone receiving 27.5% of the total (USD 1.7 billion) in 2015-2019. Asia received 25.0% of bilateral development finance for digitalisation activities and the Americas, Europe, the Middle East and Oceania each received around 5%. The breakdown is different for multilateral development finance. The Americas received the biggest share of total multilateral finance for digitalisation – 36.6%, or USD 4.1 billion Figure 40.9. This is due to investments by the IADB, which emerged in the estimates as the largest provider of digitalisation‑related development finance. Bilateral providers appear to be investing more in digital projects in Africa, followed by Asia.
Source: Authors’ estimates based on the data available in the CRS database.
Development finance activities reported to the CRS database are categorised under various social and economic sectors. Finance for digitalisation is concentrated in the communications sector, which in the CRS taxonomy includes activities in information and communication technology, telecommunications, and related policy interventions. This sector accounted for 42% of all bilateral development finance activities and 65% of multilateral activities related to digitalisation. Bilateral providers also focus on the government and civil society and education sectors, while the banking and financial services sector appears to be an important focus area for multilateral providers (Figure 40.10).
Source: Authors’ estimates based on the data available in the CRS database.
There are several limitations and challenges to the methodology used to estimate support for digital transformation, and scope to define a clearer method.7 Ex-ante identification is more reliable than the ex-post identification system used based on data submitted to the CRS. As data providers have in-depth knowledge of their operations, they could produce information about a project from the design to approval phase, when data disclosure is required.
If the international development focus on digitalisation increases, so will the need for transparency and accountability. The statistical method for tracking this finance would need to be agreed. Given that fewer than half of DAC members explicitly focus on digital transformation in their strategies (see Chapter 33), a pragmatic, comprehensive and feasible approach to tracking official development finance for digitalisation could be the voluntary reporting of agreed digitalisation keywords complemented by analysis of reported activities through machine learning.
Policy markers are precise tools to track if reported activities promote a policy objective and, if so, to what extent. Policy markers are agreed by consensus in the DAC WP-STAT. They tend to be lengthy to negotiate and implement, and require adding a new data field and making changes in data collection and reporting processes of data providers and at the OECD. The OECD DAC statistical reporting template already contains many fields, and members might be unwilling to add further complexity. Some members expressed concerns about their capacity to provide additional dimensions to CRS reporting (OECD, 2020[5]); the newest fields in the CRS reporting template (SDGs and the policy markers on nutrition and disability) were added as voluntary fields. A recent review of the OECD policy marker system found that, generally, markers work better when the policy objective is truly cross-sectorial and when the topic is of great policy interest, eventually linked to an international agreement or a strong stakeholder community (OECD, 2020[5]).
In 2020, DAC members decided to track support to COVID-19 response and recovery through a new keyword. A keyword field was created in the CRS to allow members to flag with #COVID19 all activities that contributed to these objectives using a common definition. Some members expressed a willingness to expand the use of the keyword field for other topics.
WP-STAT is discussing the modalities for introducing this keyword approach to reporting for other cross-cutting themes. Digitalisation could be a strong candidate.
Introducing a digitalisation keyword would not necessitate adding a new field and changing data processing structures, pending agreement on keyword governance. Reporting keywords is also voluntary; reporting entities could also use different keywords to highlight different digitalisation aspects or other innovations in development co-operation.
Machine learning tools can extract information from large bodies of text and are increasingly used for data analysis and to check data quality. They are, however, complex to set up and fine tune and depend on the quality of reporting and details provided in the programme and project descriptions. Developing appropriate machine learning tools is another possible option to track digitalisation-related development finance. The OECD Secretariat is working on machine learning exercises for both purpose codes and the SDGs. These tools have the capacity to analyse large amounts of information but need appropriate resources to be developed and trained.
[4] Deshmukh, S. (2020), “3 ways blockchain can accelerate sustainable development”, World Economic Forum Agenda blog, https://www.weforum.org/agenda/2020/09/3-ways-blockchain-can-contribute-to-sustainable-development/ (accessed on 19 October 2021).
[5] OECD (2020), “Assessing the policy objectives of development co-operation activities: Review of the reporting status, use and relevance of Rio and policy markers”, DAC Working Party on Development Finance Statistics, https://one.oecd.org/document/DCD/DAC/STAT(2020)27/en/pdf (accessed on 22 November 2021).
[2] OECD (2020), Converged Statistical Reporting Directives for the Creditor Reporting System (CRS) and the Annual DAC Questionnaire: Annexes - Modules D & E, DAC Working Party on Development Finance Statistics, https://one.oecd.org/document/DCD/DAC/STAT(2020)44/ADD2/FINAL/en/pdf (accessed on 10 November 2021).
[1] OECD (2020), “Guidelines on the European Commission International Digitalisation Marker”, DAC Working Party on Development Finance Statistics, https://one.oecd.org/document/DCD/DAC/STAT/RD(2020)2/en/pdf (accessed on 10 November 2021).
[3] Sirimanne, S. and C. Freire (2021), How Blockchain Can Power Sustainable Development, United Nations Conference on Trade and Development, Geneva, https://unctad.org/news/how-blockchain-can-power-sustainable-development (accessed on 19 October 2021).
← 1. Digitalisation is understood as the use of digital technologies and data that results in new activities or changes to existing activities. Digitisation is the conversion of analogue data and processes into a machine-readable format. Digital transformation refers to the economic and societal effects of digitalisation and digitisation.
← 2. The database code for the communication sector includes activities for financing digital infrastructure such as large networks as well as ICT tools and related activities.
← 3. To avoid inflating the results, activities that were reported with several SDGs but only partially matched the list of digitalisation-related SDGs were not included as data sources unless the activity was also identified by other criteria such as keywords or sector codes.
← 4. Bilateral, multilateral and philanthropic finance is expressed in constant USD 2019 prices. Data on private finance mobilised by official intervention are only available at current prices and have some further limitations.
← 5. A restriction to 2018-19 data for this share was made to ensure coherent reporting for philanthropic institutions, which are more limited for previous years, and smooth any year-to-year fluctuation
← 6. See note 5.
← 7. For example, while keyword searches can be tested for robustness, keywords can be arbitrary. If providers do not consider these keywords when submitting project descriptions to the database, a keyword search would likely miss the projects. Another example concerns the communication sector code, which can include activities that do not strictly promote digitalisation. The same is true for the SDGs focus field. Furthermore, manual checking of activities based on their descriptions is difficult and time-consuming.