How ready is Africa to join the knowledge economy?

“In times past, we searched for gold, precious stones, minerals, and ore. Today, it is knowledge that makes us rich and access to information is all-powerful in enabling individual and collective success.” Lesotho Ministry of Communications, Science and Technology, 2005.

Changes in the ways knowledge is created and used — and how this is enabled by new information technologies — are driving economic and social development worldwide. Discussions of the “knowledge economy” see knowledge both as an economic output in itself, and as an input that strengthens economic processes; with developing countries tending to be described as planning or embarking on a journey of transformation into knowledge economies to bring about economic gain. Indeed, increasing connectivity has sparked many hopes for the democratization of knowledge production in sub-Saharan Africa.

Despite the centrality of digital connectivity to the knowledge economy, there are few studies of the geographies of digital knowledge and information. In their article “Engagement in the Knowledge Economy: Regional Patterns of Content Creation with a Focus on Sub-Saharan Africa”, published in Information Technologies & International Development, Sanna Ojanperä, Mark Graham, Ralph K. Straumann, Stefano De Sabbata, and Matthew Zook investigate the patterns of knowledge creation in the region. They examine three key metrics: spatial distributions of academic articles (i.e. traditional knowledge production), collaborative software development, and Internet domain registrations (i.e. digitally mediated knowledge production).

Contrary to expectations, they find distribution patterns of digital content (measured by collaborative coding and domain registrations) to be more geographically uneven than those of academic articles: despite the hopes for the democratizing power of the information revolution. This suggests that the factors often framed as catalysts for a knowledge economy do not relate to these three metrics uniformly.

Connectivity is an important enabler of digital content creation, but it seems to be only a necessary, not a sufficient, condition; wealth, innovation capacity, and public spending on education are also important factors. While the growth in telecommunications might be reducing the continent’s reliance on extractive industries and agriculture, transformation into a knowledge economy will require far more concentrated effort than simply increasing Internet connectivity.

We caught up with Sanna to discuss the article’s findings:

Ed.: You chose three indices (articles, domain registration, collaborative coding) to explore the question of Africa’s “readiness” to join the knowledge economy. Are these standard measures for the (digital) knowledge economy?

Sanna: Academic articles is a measure often used to estimate knowledge-rich activity, so you could consider it a traditional variable in this area. Other previous work measuring the geographies of codified knowledge have focused on particular aspects or segments of it, such as patents, citations, and innovation systems.

What we found to be an interesting gap in the estimation of knowledge economies is that even if digital connectivity is central to the knowledge economy discourse, studies of the current geographies of digital knowledge and information on online platforms are rare. We argue that digitally mediated participation in information- and knowledge-intensive activities offers a metric that closely measures human capacity and skills. An analysis of digitally mediated traces of skills and information might thus complement the knowledge economy discussion and offer a way to better detect the boundaries of contemporary knowledge economies. To address the gap of research on digital content, we examine the geography of activities in collaborative software development (using the GitHub platform), and the registration of top-level domains. While there are other indicators that we could have included in the analysis, we selected these two, as they have a global reach and because they measure two distinct, but important segments of knowledge economy.

Ed.: To what extent do the drivers commonly associated with knowledge economies (e.g., GDP, broadband Internet, education, innovation) explain the patterns you found? And what were these patterns?

Sanna: While connectivity plays a role in all three categories, it seems to have a strong effect only on digital content creation. Conversely, the production of academic articles is more strongly related to GDP than to connectivity. Innovation capacity appears to have a positive relationship to all three content types. Education as a topically narrower variable appears, perhaps unexpectedly, to be related only to variance in academic articles.

In terms of the patterns of these variables, we find that the geographies of collaborative coding and domain registrations are more uneven than the spatial distribution of academic authoring. Sub-Saharan Africa contributes the smallest share of content to all three categories, providing only 1.1% of academic articles. With 0.5% of collaborative coding and 0.7% of domain registrations, SSA produces an even smaller share of digital content.

While comparison across absolute numbers informs us of the total volume of content creation, it is useful to pair that with a standardized measure that informs us of the propensity of content creation across the populations. Considering the most productive countries in terms of their per capita content creation suggests geographies even more clustered in Europe than looking at total numbers. In SSA, the level of individual countries’ content production falls within the two lowest quintiles more often in the case of collaborative coding and domain registrations than with academic articles. This runs contrary to the expectation of contemporary digitally mediated content being more evenly geographically distributed than traditional content.

Ed.: You measured “articles” by looking at author affiliations. Could you just as well have used “universities” as the measure? Or is there an assumption that connectivity will somehow encourage international co-authorship (does it?) — or that maybe “articles” is a better measure of knowledge quality than presence of universities per se?

Sanna: We chose this indicator, because we consider scientific output in the form of academic articles to represent the progress of science. Publication of academic articles and the permanent scientific record they form are central for the codification of knowledge and are a key enabler of knowledge-intensive processes. Beyond being an indicator often included in the knowledge economy indices, we believe that academic articles offer a relatively uniform measure of knowledge-intensive output, as the process of peer-reviewed publishing and the way in which it constructs a permanent scientific record are rather similar around the world. In contrast, other systems for knowledge-intensive outputs such as registering patents and innovation systems are known to be vary greatly between countries and regions (Griliches, 1990).

We didn’t use the number of universities as an indicator of the knowledge economy, because we wanted to look at measures of knowledge-intensive content-creation. While universities educate individuals and increase the nation’s human capital, this ‘output’ is very diverse and looks very different between universities. Further, we wanted to assess whether education in fact drives the development of knowledge economy, and used a measure of enrollment rates in secondary and tertiary education as an explanatory variable in our analysis.

Ed.: There’s a lot of talk of Africa entering the “global” marketplace: but how global is the knowledge economy — particularly given differences in language and culture? I imagine most cultural and knowledge production remains pretty local?

Sanna: The knowledge economy could be seen as a new dynamic stage in the global economic restructuring, where economic processes and practices that place greater emphasis of intellectual abilities take place in the context of an increasingly interconnected world. To the extent that African information- and knowledge-intensive goods and services compete in these global markets, one could consider the region entering the global knowledge economy. While the markets for knowledge-based goods and services may be smaller in various African countries, many produce regionally or nationally and locally targeted knowledge-rich products. However, this understanding of the concept of knowledge economy tends to focus on commercial activities and scientific and technical knowledge and neglect indigenous, local or cultural knowledge. These types of knowledge have a higher tendency of being tacit rather than codifiable in nature. Unlike codified knowledge, which can be recorded and transmitted through symbols or become materialized in concrete form such as tools or machinery, tacit knowledge takes time to obtain and is not as easily diffused. While these tacit types of knowledge are prevalent and carry significant value for their users and producers, they are less easily converted to commercial value. This makes their measurement more challenging and as a result the discourse of knowledge economies tends to focus less on these types of knowledge production.

Ed.: Is the knowledge economy always going to be a “good” thing, or could it lead to (more) economic exploitation of the region — for example if it got trapped into supplying a market for low-quality work? (I guess the digital equivalent of extractive, rather than productive industries .. call centres, gig-work etc.)

Sanna: As is the case with any type of economic activity, the distributional effects of knowledge economies are affected by a myriad of factors. On one hand, many of the knowledge- and information-rich economic activities require human capital and technological resources, and tend to yield goods and services with higher value added. The investment in and the greater presence of these resources may help nations and individuals to access more opportunities to increase their welfare. However, countries don’t access the global information- and knowledge-based markets as equal players and the benefits from knowledge economies are not distributed equally. It is possible that exploitative practices exist in particular where institutions and regulatory practices are not sufficiently powerful to ensure adequate working conditions. In a previous study on the Sub-Saharan African gig economy and digital labour – both areas that could be considered to form part of the knowledge economy – some of the authors found that while a range of workers in these domains enjoy important and tangible benefits, they also face risks and costs such as low bargaining power, limited economic inclusion, intermediated value chains leading to exploitation of less experienced workers, and restrictions in upgrading skills in order to move upwards in professional roles.

Ed.: I guess it’s difficult to unpack any causation in terms of Internet connectivity, economic development, and knowledge economies — despite hopes of the Internet “transforming” Sub-Saharan African economies. Is there anything in your study (or others) to hint at an answer to the question of causality?

Sanna: In order to discuss causality, we would need to study the effects of a given intervention or treatment, as measured in an ideal randomized controlled experiment. As we’re not investigating the effect of a particular intervention, but studying descriptive trends in the three dependent variables using the Ordinary Least Squares (OLS) method of estimation in multiple linear regression framework, we cannot make strong claims about causality. However, we find that both the descriptive study of RQ1 as well as the regression modeling and residual mapping for RQ2 offer statistically significant results, which lend themselves for interpretations with relevance to our RQs and that have important implications, which we discuss in the concluding section of the article.

Read the full article: Ojanperä, S., Graham, M., Straumann, R.K., De Sabbata, S., & Zook, M. (2017). Engagement in the knowledge economy: Regional patterns of content creation with a focus on sub-Saharan Africa. Information Technologies & International Development 13: 33–51.


Sanna Ojanperä was talking to blog editor David Sutcliffe.

How useful are volunteer crisis-mappers in a humanitarian crisis?

User-generated content can provide a useful source of information during humanitarian crises like armed conflict or natural disasters. With the rise of interactive websites, social media, and online mapping tools, volunteer crisis mappers are now able to compile geographic data as a humanitarian crisis unfolds, allowing individuals across the world to organize as ad hoc groups to participate in data collection. Crisis mappers have created maps of earthquake damage and trapped victims, analyzed satellite imagery for signs of armed conflict, and cleaned Twitter data sets to uncover useful information about unfolding extreme weather events like typhoons.

Although these volunteers provide useful technical assistance to humanitarian efforts (e.g. when maps and records don’t exist or are lost), their lack of affiliation with “formal” actors, such as the United Nations, and the very fact that they are volunteers, makes them a dubious data source. Indeed, concerns have been raised about the quality of amateur mapping and data efforts, and the uses to which they are put. Most of these concerns assume that volunteers have no professional training. And herein lies the contradiction: by doing the work for free and at their own will the volunteers make these efforts possible and innovative, but this is also why crisis mapping is doubted and questioned by experts.

By investigating crisis-mapping volunteers and organizations, Elizabeth Resor’s article “The Neo-Humanitarians: Assessing the Credibility of Organized Volunteer Crisis Mappers” published in Policy & Internet presents evidence of a more professional cadre of volunteers and a means to distinguish between different types of volunteer organizations. Given these organizations now play an increasingly integrated role in humanitarian responses, it’s crucial that their differences are understood and that concerns about the volunteers are answered.

We caught up with Elizabeth to discuss her findings:

Ed.: We have seen from Citizen Science (and Wikipedia) that large crowds of non-professional volunteers can produce work of incredible value, if projects are set up right. Are the fears around non-professional crisis mappers valid? For example, is this an environment where everything “must be correct”, rather than “probably mostly correct”?

Elizabeth: Much of the fears around non-professional crisis mappers comes from a lack of understanding about who the volunteers are and why they are volunteering. As these questions are answered and professional humanitarian actors become more familiar with the concept of volunteer humanitarians, I think many of these fears are diminishing.

Due to the fast-paced and resource-constrained environments of humanitarian crises, traditional actors, like the UN, are used to working with “good enough” data, or data that are “probably mostly correct”. And as you point out, volunteers can often produce very high quality data. So when you combine these two facts, it stands to reason that volunteer crisis mappers can contribute necessary data that is most likely as good as (if not better) than the data that humanitarian actors are used to working with. Moreover, in my research I found that most of these volunteers are not amateurs in the full sense because they come from related professional fields (such as GIS).

Ed.: I suppose one way of assuaging fears is to maybe set up an umbrella body of volunteer crisis mapping organisations, and maybe offer training opportunities and certification of output. But then I suppose you just end up as professionals. How blurry are the lines between useful-not useful / professional-amateur in crisis mapping?

Elizabeth: There is an umbrella group for volunteer organizations set up exactly for that reason! It’s called the Digital Humanitarian Network. At the time that I was researching this article, the DHN was very new and so I wasn’t able to ask if actors were more comfortable working with volunteers contacted through the DHN, but that would be an interesting issue to look into.

The two crisis mapping organizations I researched — the Standby Task Force and the GIS Corps — both offer training and some structure to volunteer work. They take very different approaches to the volunteer work — the Standby Task Force work can include very simple micro-tasks (like classifying photographs), whereas the GIS Corps generally provides quite specialised technical assistance (like GIS analysis). However, both of these kinds of tasks can produce useful and needed data in a crisis.

Ed.: Another article in the journal examined the effective take-over of a Russian crisis volunteer website by the Government, i.e. by professionalising (and therefore controlling) the site and volunteer details they had control over who did / didn’t turn up in disaster areas (effectively meaning nonprofessionals were kept out). How do humanitarian organisations view volunteer crisis mappers: as useful organizations to be worked with in parallel, or as something to be controlled?

Elizabeth: I have seen examples of humanitarian and international development agencies trying to lead or create crowdsourcing responses to crises (for example, USAID “Mapping to End Malaria“). I take this as a sign that these agencies understand the value in volunteer contributions — something they wouldn’t have understood without the initial examples created by those volunteers.

Still, humanitarian organizations are large bureaucracies, and even in a crisis they function as bureaucracies, while volunteer organizations take a nimble and flexible approach. This structural difference is part of the value that volunteers can offer humanitarian organizations, so I don’t believe that it would be in the best interest of the humanitarian organizations to completely co-opt or absorb the volunteer organizations.

Ed.: How does liability work? Eg if crisis workers in a conflict zone are put in danger by their locations being revealed by well-meaning volunteers? Or mistakes being being made on the ground because of incorrect data — perhaps injected by hostile actors to create confusion (thinking of our current environment of hybrid warfare..).

Elizabeth: Unfortunately, all humanitarian crises are dangerous and involve threats to “on the ground” response teams as well as affected communities. I’m not sure how liability is handled. Incorrect data or revealed locations might not be immediately traced back to the source of the problem (i.e. volunteers) and the first concern would be minimizing the harm, not penalizing the cause.

Still, this is the greatest challenge to volunteer crisis mapping that I see. Volunteers don’t want to cause more harm than good, and to do this they must understand the context of the crisis in which they are getting involved (even if it is remotely). This is where relationships with organizations “on the ground” are key. Also, while I found that most volunteers had experience related to GIS and/or data analysis, very few had experience in humanitarian work. This seems like an area where training can help volunteers understand the gravity of their work, to ensure that they take it seriously and do their best work.

Ed.: Finally, have you ever participated as a volunteer crisis mapper? And also: how do you the think the phenomenon is evolving, and what do you think researchers ought to be looking at next?

Elizabeth: I haven’t participated in any active crises, although I’ve tried some of the tools and trainings to get a sense of the volunteer activities.

In terms of future research, you mentioned hybridized warfare and it would be interesting to see how this change in the location of a crisis (i.e. in online spaces as well as physical spaces) is changing the nature of volunteer responses. For example, how can many dispersed volunteers help monitor ISIS activity on YouTube and Twitter? Or are those tasks better suited for an algorithm? I would also be curious to see how the rise of isolationist politicians in Europe and the US has influenced volunteer crisis mapping. Has this caused more people to want to reach out and participate in international crises or is it making them more inward-looking? It’s certainly an interesting field to follow!

Read the full article: Resor, E. (2016) The Neo-Humanitarians: Assessing the Credibility of Organized Volunteer Crisis Mappers. Policy & Internet 8 (1) DOI:10.1002/poi3.112.

Elizabeth Resor was talking to blog editor David Sutcliffe.

How useful are volunteer crisis-mappers in a humanitarian crisis?

User-generated content can provide a useful source of information during humanitarian crises like armed conflict or natural disasters. With the rise of interactive websites, social media, and online mapping tools, volunteer crisis mappers are now able to compile geographic data as a humanitarian crisis unfolds, allowing individuals across the world to organize as ad hoc groups to participate in data collection. Crisis mappers have created maps of earthquake damage and trapped victims, analyzed satellite imagery for signs of armed conflict, and cleaned Twitter data sets to uncover useful information about unfolding extreme weather events like typhoons.

Although these volunteers provide useful technical assistance to humanitarian efforts (e.g. when maps and records don’t exist or are lost), their lack of affiliation with “formal” actors, such as the United Nations, and the very fact that they are volunteers, makes them a dubious data source. Indeed, concerns have been raised about the quality of amateur mapping and data efforts, and the uses to which they are put. Most of these concerns assume that volunteers have no professional training. And herein lies the contradiction: by doing the work for free and at their own will the volunteers make these efforts possible and innovative, but this is also why crisis mapping is doubted and questioned by experts.

By investigating crisis-mapping volunteers and organizations, Elizabeth Resor’s article “The Neo-Humanitarians: Assessing the Credibility of Organized Volunteer Crisis Mappers” published in Policy & Internet presents evidence of a more professional cadre of volunteers and a means to distinguish between different types of volunteer organizations. Given these organizations now play an increasingly integrated role in humanitarian responses, it’s crucial that their differences are understood and that concerns about the volunteers are answered.

We caught up with Elizabeth to discuss her findings:

Ed.: We have seen from Citizen Science (and Wikipedia) that large crowds of non-professional volunteers can produce work of incredible value, if projects are set up right. Are the fears around non-professional crisis mappers valid? For example, is this an environment where everything “must be correct”, rather than “probably mostly correct”?

Elizabeth: Much of the fears around non-professional crisis mappers comes from a lack of understanding about who the volunteers are and why they are volunteering. As these questions are answered and professional humanitarian actors become more familiar with the concept of volunteer humanitarians, I think many of these fears are diminishing.

Due to the fast-paced and resource-constrained environments of humanitarian crises, traditional actors, like the UN, are used to working with “good enough” data, or data that are “probably mostly correct”. And as you point out, volunteers can often produce very high quality data. So when you combine these two facts, it stands to reason that volunteer crisis mappers can contribute necessary data that is most likely as good as (if not better) than the data that humanitarian actors are used to working with. Moreover, in my research I found that most of these volunteers are not amateurs in the full sense because they come from related professional fields (such as GIS).

Ed.: I suppose one way of assuaging fears is to maybe set up an umbrella body of volunteer crisis mapping organisations, and maybe offer training opportunities and certification of output. But then I suppose you just end up as professionals. How blurry are the lines between useful-not useful / professional-amateur in crisis mapping?

Elizabeth: There is an umbrella group for volunteer organizations set up exactly for that reason! It’s called the Digital Humanitarian Network. At the time that I was researching this article, the DHN was very new and so I wasn’t able to ask if actors were more comfortable working with volunteers contacted through the DHN, but that would be an interesting issue to look into.

The two crisis mapping organizations I researched — the Standby Task Force and the GIS Corps — both offer training and some structure to volunteer work. They take very different approaches to the volunteer work — the Standby Task Force work can include very simple micro-tasks (like classifying photographs), whereas the GIS Corps generally provides quite specialised technical assistance (like GIS analysis). However, both of these kinds of tasks can produce useful and needed data in a crisis.

Ed.: Another article in the journal examined the effective take-over of a Russian crisis volunteer website by the Government, i.e. by professionalising (and therefore controlling) the site and volunteer details they had control over who did / didn’t turn up in disaster areas (effectively meaning nonprofessionals were kept out). How do humanitarian organisations view volunteer crisis mappers: as useful organizations to be worked with in parallel, or as something to be controlled?

Elizabeth: I have seen examples of humanitarian and international development agencies trying to lead or create crowdsourcing responses to crises (for example, USAID “Mapping to End Malaria“). I take this as a sign that these agencies understand the value in volunteer contributions — something they wouldn’t have understood without the initial examples created by those volunteers.

Still, humanitarian organizations are large bureaucracies, and even in a crisis they function as bureaucracies, while volunteer organizations take a nimble and flexible approach. This structural difference is part of the value that volunteers can offer humanitarian organizations, so I don’t believe that it would be in the best interest of the humanitarian organizations to completely co-opt or absorb the volunteer organizations.

Ed.: How does liability work? Eg if crisis workers in a conflict zone are put in danger by their locations being revealed by well-meaning volunteers? Or mistakes being being made on the ground because of incorrect data — perhaps injected by hostile actors to create confusion (thinking of our current environment of hybrid warfare..).

Elizabeth: Unfortunately, all humanitarian crises are dangerous and involve threats to “on the ground” response teams as well as affected communities. I’m not sure how liability is handled. Incorrect data or revealed locations might not be immediately traced back to the source of the problem (i.e. volunteers) and the first concern would be minimizing the harm, not penalizing the cause.

Still, this is the greatest challenge to volunteer crisis mapping that I see. Volunteers don’t want to cause more harm than good, and to do this they must understand the context of the crisis in which they are getting involved (even if it is remotely). This is where relationships with organizations “on the ground” are key. Also, while I found that most volunteers had experience related to GIS and/or data analysis, very few had experience in humanitarian work. This seems like an area where training can help volunteers understand the gravity of their work, to ensure that they take it seriously and do their best work.

Ed.: Finally, have you ever participated as a volunteer crisis mapper? And also: how do you the think the phenomenon is evolving, and what do you think researchers ought to be looking at next?

Elizabeth: I haven’t participated in any active crises, although I’ve tried some of the tools and trainings to get a sense of the volunteer activities.

In terms of future research, you mentioned hybridized warfare and it would be interesting to see how this change in the location of a crisis (i.e. in online spaces as well as physical spaces) is changing the nature of volunteer responses. For example, how can many dispersed volunteers help monitor ISIS activity on YouTube and Twitter? Or are those tasks better suited for an algorithm? I would also be curious to see how the rise of isolationist politicians in Europe and the US has influenced volunteer crisis mapping. Has this caused more people to want to reach out and participate in international crises or is it making them more inward-looking? It’s certainly an interesting field to follow!

Read the full article: Resor, E. (2016) The Neo-Humanitarians: Assessing the Credibility of Organized Volunteer Crisis Mappers. Policy & Internet 8 (1) DOI:10.1002/poi3.112.

Elizabeth Resor was talking to blog editor David Sutcliffe.

Has Internet policy had any effect on Internet penetration in Sub-Saharan Africa?

There is a consensus among researchers that ICT is an engine for growth, and it’s also considered by the OECD to be a part of fundamental infrastructure, like electricity and roads. The last decade has seen a rapid growth of Internet access across Africa, although it has not been evenly distributed. Some African countries have an Internet penetration of over 50 percent (such as the Seychelles and South Africa) whereas some resemble digital deserts, not even reaching two percent. Even more surprisingly, countries that are seemingly comparable in terms of economic development often show considerable differences in terms of Internet access (e.g., Kenya and Ghana).

Being excluded from the Internet economy has negative economic and social implications; it is therefore important for policymakers to ask how policy can bridge this inequality. But does policy actually have an effect on these differences? And if so, which specific policy variables? In their Policy & Internet article “Crossing the Digital Desert in Sub-Saharan Africa: Does Policy Matter?”, Robert Wentrup, Xiangxuan Xu, H. Richard Nakamura, and Patrik Ström address the dearth of research assessing the interplay between policy and Internet penetration by identifying Internet penetration-related policy variables and institutional constructs in Sub-Saharan Africa. It is a first attempt to investigate whether Internet policy variables have any effect on Internet penetration in Sub-Saharan Africa, and to shed light on them.

Based on a literature review and the available data, they examine four variables: (i) free flow of information (e.g. level of censorship); (ii) market concentration (i.e. whether or not internet provision is monopolistic); (iii) the activity level of the Universal Service Fund (a public policy promoted by some governments and international telecom organizations to address digital inclusion); and (iv) total tax on computer equipment, including import tariffs on personal computers. The results show that only the activity level of the USF and low total tax on computer equipment are significantly positively related to Internet penetration in Sub-Saharan Africa. Free flow of information and market concentration show no impact on Internet penetration. The latter could be attributed to underdeveloped competition in most Sub-Saharan countries.

The authors argue that unless states pursue an inclusive policy intended to enhance Internet access for all its citizens, there is a risk that the skewed pattern between the “haves” and the “have nots” will persist, or worse, be reinforced. They recommend that policymakers promote the policy instrument of Universal Service and USF and consider substituting tax on computer equipment with other tax revenues (i.e. introduce consumption-friendly incentives), and not to blindly trust the market’s invisible hand to fix inequality in Internet diffusion.

We caught up with Robert to discuss the findings:

Ed.: I would assume that Internet penetration is rising (or possibly even booming) across the continent, and that therefore things will eventually sort themselves out — is that generally true? Or is it already stalling / plateauing, leaving a lot of people with no apparent hope of ever getting online?

Robert: Yes, generally we see a growth in Internet penetration across Africa. But it is very heterogeneous and unequal in its character, and thus country-specific. Some rich African countries are doing quite well whereas others are lagging behind. We have also seen known that Internet connectivity is vulnerable due to the political situation. The recent shutdown of the Internet in Cameroon demonstrates this vulnerability.

Ed.: You mention that “determining the causality between Internet penetration and [various] factors is problematic” – i.e. that the relation between Internet use and economic growth is complex. This presumably makes it difficult for effective and sweeping policy “solutions” to have a clear effect. How has this affected Internet-policy in the region, if at all?.

Robert: On the one hand one can say that if there is economic growth, there will be money to invest in Internet infrastructure and devices, and on the other hand if there are investments in Internet infrastructure, there will be economic growth. This resembles the chicken and egg problem. For many African countries, which lack large public investment funds and at the same time suffer from other more important socio-economic challenges, it might be tricky to put effort into Internet policy issues. But there are some good examples of countries that have actually managed to do this, for example Kenya. As a result these efforts can lead to positive effects on the society and economy as a whole. The local context, and a focus on instruments that disseminate Internet usage in unprivileged geographic areas and social groups (like the Universal Service Fund) are very important.

Ed.: How much of the low Internet perpetration in large parts of Africa is simply due to large rural populations — and therefore something that will either never be resolved properly, or that will naturally resolve itself with the ongoing shift to urban centres?

Robert: We did not see a clear causality between rural population and low Internet on a country level, mainly due to the fact that countries with a large rural population are often quite rich and thus also have money to invest in Internet infrastructure. Africa is very dependent on agriculture. Although the Internet connectivity issue might be “self-resolved” to some degree by urban migration, other issues would emerge from such a shift such as an increased socio-economic divide in the urban areas. Hence, it is more effective to make sure that the Internet reaches rural areas at an early stage.

Ed.: And how much does domestic policy (around things like telecoms) get set internally, as opposed to externally? Presumably some things (e.g. the continent-wide cables required to connect Africa to the rest of the world) are easier to bring about if there is a strong / stable regional policy around regulation of markets and competition — whether organised internally, or influenced by outside governments and industry?

Robert: The influence of telecom ministries and telecom operators is strong, but of course they are affected by intra-regional organisations, private companies etc. In the past Africa has had difficulties in developing pan-regional trade and policies. But such initiatives are encouraged, not least in order to facilitate cost-sharing of large Internet-related investments.

Ed.: Leaving aside the question of causality, you mention the strong correlation between economic activity and Internet penetration: are there any African countries that buck this trend — at either end of the economic scale?

Robert: We have seen that Kenya and Nigeria have had quite impressive rates of Internet penetration in relation to GDP. Gabon on the other hand is a relatively rich African country, but with quite low Internet penetration.

Read the full article: Wentrup, R., Xu, X., Nakamura, H.R., and Ström, P. (2016) Crossing the Digital Desert in Sub-Saharan Africa: Does Policy Matter? Policy & Internet 8 (3). doi:10.1002/poi3.123


Robert Wentrup was talking to blog editor David Sutcliffe.

Why we shouldn’t believe the hype about the Internet “creating” development

Vast sums of money have been invested in projects to connect the world’s remaining four billion people, with these ambitious schemes often presenting digital connectivity as a means to achieve a range of social and economic developmental goals. This is especially the case for Africa, where Internet penetration rates remain relatively low, while the need for effective development strategies continues to be pressing.

Development has always grappled with why some people and places have more than others, but much of that conversation is lost within contemporary discourses of ICTs and development. As states and organisations rush to develop policies and plans, build drones and balloons, and lay fibre-optic cables, much is said about the power of ICTs to positively transform the world’s most underprivileged people and places.

Despite the vigour of such claims, there is actually a lack of academic consensus about the impacts of digital connectivity on economic development. In their new article, Nicolas Friederici, Sanna Ojanperä and Mark Graham review claims made by African governments and large international institutions about the impacts of connectivity, showing that the evidence base to support them is thin.

It is indeed possible that contemporary grand visions of connectivity are truly reflective of a promising future, but it is equally possible that many of them are hugely overblown. The current evidence base is mixed and inconclusive. More worryingly, visions of rapid ICT-driven development might not only fail to achieve their goals — they could actively undermine development efforts in a world of scarce resources. We should therefore refuse to believe it is self-evident that ICTs will automatically bring about development, and should do more to ask the organisations and entities who produce these grand visions to justify their claims.

Read the full article: Friederici, N., Ojanperä, S., and Graham, M. (2017) The Impact of Connectivity in Africa: Grand Visions and the Mirage of Inclusive Digital Development. Electronic Journal of Information Systems in Developing Countries, 79(2), 1–20.

We caught up with the authors to discuss their findings.

Ed.: Who is paying for these IT-development projects: are they business and profit-led, or donor led: and do the donors (and businesses) attach strings?

Nicolas: Funding has become ever more mixed. Foundational infrastructure like fibre-optic cables have usually been put in place through public private partnerships, where private companies lay out the network while loans, subsidies, and policy support are provided by national governments and organizations like the World Bank. Development agencies have mostly funded more targeted connectivity projects, like health or agricultural information platforms.

Recently, philanthropic foundations and tech corporations have increased their footprint, for instance, the Rockefeller Foundation’s Digital Jobs project or Facebook’s Open Cellular Base stations. So we are seeing an increasingly complex web of financial channels. What discourse does is pave the way for funding to flow into such projects.

The problem is that, while private companies may stop investing when they don’t see returns, governments and development funders might continue to pour resources into an agenda as long as it suits their ideals or desirable and widely accepted narratives. Of course, these resources are scarce; so, at the minimum, we need to allow scrutiny and look for alternatives about how development funding could be used for maximum effect.

Ed.: Simple, aspirational messages are obviously how politicians get people excited about things (and to pay for them). What is the alternative?

Nicolas: We’re not saying that the rhetoric of politicians is the problem here. We’re saying that many of the actors who are calling the shots in development are stubbornly evading valid concerns that academics and some practitioners have brought forward. The documents that we analyze in the article — and these are very influential sources — pretend that it is an unquestionable fact that there is a causal, direct and wide-spread positive impact of Internet and ICTs on all facets of development, anywhere. This assertion is not only simplistic, it’s also problematic and maybe even dangerous to think about a complex and important topic like (human, social) development in this way.

The alternative is a more open and plural conversation where we openly admit that resources spent on one thing can’t be spent on another, and where we enable different and critical opinions to enter the fray. This is especially important when a nation’s public is disempowered or misinformed, or when regulators are weak. For example, in most countries in Europe, advocacy groups and strong telecoms regulators provide a counterforce to the interests of technology corporations. Such institutions are often absent in the Global South, so the onus is on development organizations to regulate themselves, either by engaging with people “on the ground” or with academics. For instance, the recent World Development Report by the World Bank did this, which led the report to, we think, much more reliable and balanced conclusions compared to the Bank’s earlier outputs.

Ed.: You say these visions are “modernist” and “techno-determinist” — why is that? Is it a quirk of the current development landscape, or does development policy naturally tend to attract fixers (rather than doubters and worriers..). And how do we get more doubt into policy?

Nicolas: Absolutely, development organizations are all about fixing development problems, and we do not take issue with that. However, these organizations also need to understand that “fixing development” is not like fixing a machine (that is, a device that functions according to mechanical principles). It’s not like one could input “technology” or “the Internet,” and get “development” as an output.

In a nutshell, that’s what we mean when we say that visions are modernist and techno-determinist: many development organizations, governments, and corporations make the implicit assumption that technological progress is fixing development, that this is an apolitical and unstoppable process, and that this is working out in the same way everywhere on earth. This assumption glances over contestation, political choices and trade-offs, and the cultural, economic, and social diversity of contexts.

Ed.: Presumably if things are very market-led: the market will decide if the internet “solves” everything: ie either it will, or it won’t. Has there been enough time yet to verify the outcomes of these projects (e.g. how has the one-laptop initiative worked out)?

Nicolas: I’m not sure I agree with the implication that markets can decide if the Internet solves everything. It’s us humans who are deciding, making choices, prioritizing, allocating resources, setting policies, etc. As humans, we might decide that we want a market (that is, supply and demand matched by a price mechanism) to regulate some array of transactions. This is exactly what is happening, for instance, with the spread of mobile money in Kenya or the worldwide rise of smartphones: people feel they benefit from using a product and are willing to pay money to a supplier.

The issue with technology and development is (a) that in many cases, markets are not the mechanism that achieves the best development outcomes (think about education or healthcare), (b) that even the freest of markets needs to be enabled by things like political stability, infrastructure, and basic institutions (think about contract law and property rights), and (c) that many markets need regulatory intervention or power-balancing institutions to prevent one side of the exchange to dominate and exploit the other (think about workers’ rights).

In each case, it is thus a matter of evaluating what mixture of technology, markets, and protections works best to achieve the best development outcomes, keeping in mind that development is multi-dimensional and goes far beyond economic growth. These evaluations and discussions are challenging, and it takes time to determine what works, where, and when, but ultimately we’re improving our knowledge and our practice if we keep the conversation open, critical, and diverse.

Ed.: Is there a consensus on ICT and development, or are there basically lots of camps, ranging from extreme optimists to extreme pessimists? I get the impression that basically “it’s complicated” — is that fair? And how much discussion or recognition (beyond yourselves) is there about the gap between these statements and reality?

Nicolas: ICT and development has seen a lot of soul-searching, and scholars and practitioners have spent over 20 years debating the field’s nature and purpose. There is certainly no consensus on what ICTD should do, or how ICTs effect/affect development, and maybe that is an unrealistic — and undesirable — goal. There are certainly optimistic and pessimistic voices, like you mention, but there is also a lot of wisdom that is not widely acknowledged, or not in the public domain at all. There are thousands of practitioners from the Global North and South who have been in the trenches, applied their critical and curious minds, and seen what makes an impact and what is a pipe dream.

So we’re far from the only ones who are aware that much of the ICTD rhetoric is out of touch with realities, and we’re also not the first ones to identify this problem. What we tried to point out in our article is that the currently most powerful, influential, and listened to sources tend to be the ones that are overly optimistic and overly simplistic, ignoring all the wisdom and nuance created through hard scholarly and practical work. These actors seem to be detached from the messy realities of ICTD.

This carries a risk, because it is these organizations (governments, global consultancies, multilateral development organizations, and international tech corporations) that are setting the agenda, distributing the funds, making the hiring decisions, etc. in development practice.

Read the full article: Friederici, N., Ojanperä, S., and Graham, M. (2017) The Impact of Connectivity in Africa: Grand Visions and the Mirage of Inclusive Digital Development. Electronic Journal of Information Systems in Developing Countries, 79(2), 1–20.


Nicolas Friederici was talking to blog editor David Sutcliffe.

New Report: Risks and Rewards of Online Gig Work at the Global Margins

The cartogram depicts countries as circles sized according to dollar inflow during March 2013 on a major online labour platform. The shading of the inner circle indicates the median hourly rate published by digital workers in that country. See the report for details.

The growth of online gig work — paid work allocated and delivered by way of internet platforms without a contract for long-term employment — has been welcomed by economic development experts, and the world’s largest global development network is promoting its potential to aid human development. There are hopes that online gig work, and the platforms that support it, might catalyse new, sustainable employment opportunities by addressing a mismatch in the supply and demand of labour globally.

Some of the world’s largest gig work platforms have also framed their business models as a revolution in labour markets, suggesting that they can help lift people out of poverty. Similarly, many policymakers expect that regions like Sub-Saharan Africa and Southeast Asia can capitalise on this digitally mediated work opportunity as youth-to-adult unemployment rates hit historic peaks. More broadly, it has been suggested that online gig work will have structural benefits on the global economy, such as raising labour force participation and improving productivity.

Against this background, a new report by Mark Graham, Vili Lehdonvirta, Alex Wood, Helena Barnard, Isis Hjorth, and David Peter Simon, “The Risks and Rewards of Online Gig Work At The Global Margins” [PDF] highlights the risks alongside the rewards of online gig work. It draws on interviews and surveys, together with transaction data from one of the world’s largest online gig work platforms, to reveal the complex and sometimes problematic reality of this “new world of work”.

While there are significant rewards to online gig work, there are also significant risks. Discrimination, low pay rates, overwork, and insecurity all need to be tackled head-on. The report encourages online gig work platforms to further develop their service, policymakers to revisit regulation, and labour activists to examine organising tactics if online gig work is to truly live up to its potential for human development, and become a sustainable situation for many more workers.

The final section of the report poses questions for all stakeholders regarding how to improve the conditions and livelihoods of online gig workers, particularly given how these platforms have become disembedded from the norms and laws that normally regulate labour intermediaries. Specific questions that are discussed include:

  • Is it necessary to list nationality on profile pages? Will online gig workers receive formal employment contracts in the future?
  • What formal channels could exist for workers to voice their issues? Where should governments regulate online gig work in the future?
  • Will governments need to limit online gig work monopolies? And how will governments support alternative forms of platform organisation?
  • What online forms of voice could emerge for workers, and in what ways can existing groups be leveraged to promote solidarity?
  • To what extent will companies be held accountable for poor working conditions? Do platforms need a Fairwork certification program?

The report also offers suggestions alongside the questions, drawing on relevant literature and referencing historical precedents.

Read the full report: Graham, M., Lehdonvirta, V., Wood, A., Barnard, H., Hjorth, I., Simon, D. P. (2017) The Risks and Rewards of Online Gig Work At The Global Margins [PDF]. Oxford: Oxford Internet Institute.

Read the article: Graham, M., Hjorth, I. and Lehdonvirta, V. (2017) Digital Labour and Development: Impacts of Global Digital Labour Platforms and the Gig Economy on Worker Livelihoods. Transfer. DOI: 10.1177/1024258916687250

The report is an output of the project “Microwork and Virtual Production Networks in Sub-Saharan Africa and Southeast Asia”, funded by the International Development Research Centre (IDRC), grant number: 107384-001.

New Report: Risks and Rewards of Online Gig Work at the Global Margins

The cartogram depicts countries as circles sized according to dollar inflow during March 2013 on a major online labour platform. The shading of the inner circle indicates the median hourly rate published by digital workers in that country. See the report for details.

The growth of online gig work — paid work allocated and delivered by way of internet platforms without a contract for long-term employment — has been welcomed by economic development experts, and the world’s largest global development network is promoting its potential to aid human development. There are hopes that online gig work, and the platforms that support it, might catalyse new, sustainable employment opportunities by addressing a mismatch in the supply and demand of labour globally.

Some of the world’s largest gig work platforms have also framed their business models as a revolution in labour markets, suggesting that they can help lift people out of poverty. Similarly, many policymakers expect that regions like Sub-Saharan Africa and Southeast Asia can capitalise on this digitally mediated work opportunity as youth-to-adult unemployment rates hit historic peaks. More broadly, it has been suggested that online gig work will have structural benefits on the global economy, such as raising labour force participation and improving productivity.

Against this background, a new report by Mark Graham, Vili Lehdonvirta, Alex Wood, Helena Barnard, Isis Hjorth, and David Peter Simon, “The Risks and Rewards of Online Gig Work At The Global Margins” [PDF] highlights the risks alongside the rewards of online gig work. It draws on interviews and surveys, together with transaction data from one of the world’s largest online gig work platforms, to reveal the complex and sometimes problematic reality of this “new world of work”.

While there are significant rewards to online gig work, there are also significant risks. Discrimination, low pay rates, overwork, and insecurity all need to be tackled head-on. The report encourages online gig work platforms to further develop their service, policymakers to revisit regulation, and labour activists to examine organising tactics if online gig work is to truly live up to its potential for human development, and become a sustainable situation for many more workers.

The final section of the report poses questions for all stakeholders regarding how to improve the conditions and livelihoods of online gig workers, particularly given how these platforms have become disembedded from the norms and laws that normally regulate labour intermediaries. Specific questions that are discussed include:

  • Is it necessary to list nationality on profile pages? Will online gig workers receive formal employment contracts in the future?
  • What formal channels could exist for workers to voice their issues? Where should governments regulate online gig work in the future?
  • Will governments need to limit online gig work monopolies? And how will governments support alternative forms of platform organisation?
  • What online forms of voice could emerge for workers, and in what ways can existing groups be leveraged to promote solidarity?
  • To what extent will companies be held accountable for poor working conditions? Do platforms need a Fairwork certification program?

The report also offers suggestions alongside the questions, drawing on relevant literature and referencing historical precedents.

Read the full report: Graham, M., Lehdonvirta, V., Wood, A., Barnard, H., Hjorth, I., Simon, D. P. (2017) The Risks and Rewards of Online Gig Work At The Global Margins [PDF]. Oxford: Oxford Internet Institute.

Read the article: Graham, M., Hjorth, I. and Lehdonvirta, V. (2017) Digital Labour and Development: Impacts of Global Digital Labour Platforms and the Gig Economy on Worker Livelihoods. Transfer. DOI: 10.1177/1024258916687250

The report is an output of the project “Microwork and Virtual Production Networks in Sub-Saharan Africa and Southeast Asia”, funded by the International Development Research Centre (IDRC), grant number: 107384-001.

What Impact is the Gig Economy Having on Development and Worker Livelihoods?

There are imbalances in the relationship between supply and demand of digital work, with the vast majority of buyers located in high-income countries (pictured). See the full article for details.

As David Harvey famously noted, workers are unavoidably place-based because “labor-power has to go home every night.” But the widespread use of the Internet has changed much of that. The confluence of rapidly spreading digital connectivity, skilled but under-employed workers, the existence of international markets for labour, and the ongoing search for new outsourcing destinations, has resulted in organisational, technological, and spatial fixes for virtual production networks of services and money. Clients, bosses, workers, and users of the end-products of work can all now be located in different corners of the planet.

A new article by Mark Graham, Isis Hjorth and Vili Lehdonvirta, “Digital labour and development: impacts of global digital labour platforms and the gig economy on worker livelihoods”, published in Transfer, discusses the implications of the spatial unfixing of work for workers in some of the world’s economic margins, and reflects on some of the key benefits and costs associated with these new digital regimes of work. Drawing on a multi-year study with digital workers in Sub-Saharan Africa and South-east Asia, it highlights four key concerns for workers: bargaining power, economic inclusion, intermediated value chains, and upgrading.

As ever more policy-makers, governments and organisations turn to the gig economy and digital labour as an economic development strategy to bring jobs to places that need them, it is important to understand how this might influence the livelihoods of workers. The authors show that although there are important and tangible benefits for a range of workers, there are also a range of risks and costs that could negatively affect the livelihoods of digital workers. They conclude with a discussion of four broad strategies – certification schemes, organising digital workers, regulatory strategies and democratic control of online labour platforms — that could improve conditions and livelihoods for digital workers.

We caught up with the authors to explore the implications of the study:

Ed.: Shouldn’t increased digitisation of work also increase transparency (i.e. tracking, auditing etc.) around this work — i.e. shouldn’t digitisation largely be a good thing?

Mark: It depends. One of the goals of our research is to ask who actually wins and loses from the digitalisation of work. A good thing for one group (e.g. employers in the Global North) isn’t necessarily automatically a good thing for another group (e.g. workers in the Global South).

Ed.: You mention market-based strategies as one possible way to improve transparency around working conditions along value chains: do you mean something like a “Fairtrade” certification for digital work, i.e. creating a market for “fair work”?

Mark: Exactly. At the moment, we can make sure that the coffee we drink or the chocolate we eat is made ethically. But we have no idea if the digital services we use are. A ‘fair work’ certification system could change that.

Ed.: And what sorts of work are these people doing? Is it the sort of stuff that could be very easily replaced by advances in automation (natural language processing, pattern recognition etc.)? i.e. is it doubly precarious, not just in terms of labour conditions, but also in terms of the very existence of the work itself?

Mark: Yes, some of it is. Ironically, some of the paid work that is done is training algorithms to do work that used to be done by humans.

Ed.: You say that “digital workers have been unable to build any large-scale or effective digital labour movements” — is that because (unlike e.g. farm work which is spatially constrained), employers can very easily find someone else anywhere in the world who is willing to do it? Can you envisage the creation of any effective online labour movement?

Mark: A key part of the problem for workers here is the economic geography of this work. A worker in Kenya knows that they can be easily replaced by workers on the other side of the planet. The potential pool of workers willing to take any job is massive. For digital workers to have any sort of effective movement in this context means looking to what I call geographic bottlenecks in the system. Places in which work isn’t solely in a global digital cloud. This can mean looking to things like organising and picketing the headquarters of firms, clusters of workers in particular places, or digital locations (the web-presence of firms). I’m currently working on a new publication that deals with these issues in a bit more detail.

Ed.: Are there any parallels between the online gig work you have studied and ongoing issues with “gig work” services like Uber and Deliveroo (e.g. undercutting of traditional jobs, lack of contracts, precarity)?

Mark: A commonality in all of those cases is that platforms become intermediaries in between clients and workers. This means that rather than being employees, workers tend to be self-employed: a situation that offers workers freedom and flexibility, but also comes with significant risks to the worker (e.g. no wages if they fall ill).

Read the full article: Graham, M., Hjorth, I. and Lehdonvirta, V. (2017) Digital Labour and Development: Impacts of Global Digital Labour Platforms and the Gig Economy on Worker Livelihoods. Transfer. DOI: 10.1177/1024258916687250

Read the full report: Graham, M., Lehdonvirta, V., Wood, A., Barnard, H., Hjorth, I., Simon, D. P. (2017) The Risks and Rewards of Online Gig Work At The Global Margins [PDF]. Oxford: Oxford Internet Institute.

The article draws on findings from the research project “Microwork and Virtual Production Networks in Sub-Saharan Africa and South-east Asia”, funded by the International Development Research Centre (IDRC), grant number: 107384-001.


Mark Graham was talking to blog editor David Sutcliffe.

What Impact is the Gig Economy Having on Development and Worker Livelihoods?

There are imbalances in the relationship between supply and demand of digital work, with the vast majority of buyers located in high-income countries (pictured). See the full article for details.

As David Harvey famously noted, workers are unavoidably place-based because “labor-power has to go home every night.” But the widespread use of the Internet has changed much of that. The confluence of rapidly spreading digital connectivity, skilled but under-employed workers, the existence of international markets for labour, and the ongoing search for new outsourcing destinations, has resulted in organisational, technological, and spatial fixes for virtual production networks of services and money. Clients, bosses, workers, and users of the end-products of work can all now be located in different corners of the planet.

A new article by Mark Graham, Isis Hjorth and Vili Lehdonvirta, “Digital labour and development: impacts of global digital labour platforms and the gig economy on worker livelihoods”, published in Transfer, discusses the implications of the spatial unfixing of work for workers in some of the world’s economic margins, and reflects on some of the key benefits and costs associated with these new digital regimes of work. Drawing on a multi-year study with digital workers in Sub-Saharan Africa and South-east Asia, it highlights four key concerns for workers: bargaining power, economic inclusion, intermediated value chains, and upgrading.

As ever more policy-makers, governments and organisations turn to the gig economy and digital labour as an economic development strategy to bring jobs to places that need them, it is important to understand how this might influence the livelihoods of workers. The authors show that although there are important and tangible benefits for a range of workers, there are also a range of risks and costs that could negatively affect the livelihoods of digital workers. They conclude with a discussion of four broad strategies – certification schemes, organising digital workers, regulatory strategies and democratic control of online labour platforms — that could improve conditions and livelihoods for digital workers.

We caught up with the authors to explore the implications of the study:

Ed.: Shouldn’t increased digitisation of work also increase transparency (i.e. tracking, auditing etc.) around this work — i.e. shouldn’t digitisation largely be a good thing?

Mark: It depends. One of the goals of our research is to ask who actually wins and loses from the digitalisation of work. A good thing for one group (e.g. employers in the Global North) isn’t necessarily automatically a good thing for another group (e.g. workers in the Global South).

Ed.: You mention market-based strategies as one possible way to improve transparency around working conditions along value chains: do you mean something like a “Fairtrade” certification for digital work, i.e. creating a market for “fair work”?

Mark: Exactly. At the moment, we can make sure that the coffee we drink or the chocolate we eat is made ethically. But we have no idea if the digital services we use are. A ‘fair work’ certification system could change that.

Ed.: And what sorts of work are these people doing? Is it the sort of stuff that could be very easily replaced by advances in automation (natural language processing, pattern recognition etc.)? i.e. is it doubly precarious, not just in terms of labour conditions, but also in terms of the very existence of the work itself?

Mark: Yes, some of it is. Ironically, some of the paid work that is done is training algorithms to do work that used to be done by humans.

Ed.: You say that “digital workers have been unable to build any large-scale or effective digital labour movements” — is that because (unlike e.g. farm work which is spatially constrained), employers can very easily find someone else anywhere in the world who is willing to do it? Can you envisage the creation of any effective online labour movement?

Mark: A key part of the problem for workers here is the economic geography of this work. A worker in Kenya knows that they can be easily replaced by workers on the other side of the planet. The potential pool of workers willing to take any job is massive. For digital workers to have any sort of effective movement in this context means looking to what I call geographic bottlenecks in the system. Places in which work isn’t solely in a global digital cloud. This can mean looking to things like organising and picketing the headquarters of firms, clusters of workers in particular places, or digital locations (the web-presence of firms). I’m currently working on a new publication that deals with these issues in a bit more detail.

Ed.: Are there any parallels between the online gig work you have studied and ongoing issues with “gig work” services like Uber and Deliveroo (e.g. undercutting of traditional jobs, lack of contracts, precarity)?

Mark: A commonality in all of those cases is that platforms become intermediaries in between clients and workers. This means that rather than being employees, workers tend to be self-employed: a situation that offers workers freedom and flexibility, but also comes with significant risks to the worker (e.g. no wages if they fall ill).

Read the full article: Graham, M., Hjorth, I. and Lehdonvirta, V. (2017) Digital Labour and Development: Impacts of Global Digital Labour Platforms and the Gig Economy on Worker Livelihoods. Transfer. DOI: 10.1177/1024258916687250

Read the full report: Graham, M., Lehdonvirta, V., Wood, A., Barnard, H., Hjorth, I., Simon, D. P. (2017) The Risks and Rewards of Online Gig Work At The Global Margins [PDF]. Oxford: Oxford Internet Institute.

The article draws on findings from the research project “Microwork and Virtual Production Networks in Sub-Saharan Africa and South-east Asia”, funded by the International Development Research Centre (IDRC), grant number: 107384-001.


Mark Graham was talking to blog editor David Sutcliffe.

Examining the data-driven value chains that are changing Rwanda’s tea sector

Behind the material movement that takes tea from the slopes of Rwanda’s ‘thousand hills’ to a box on a shelf in Tesco, is a growing set of less visible digital data flows. Image by pasunejen.
Production of export commodity goods like tea, coffee and chocolate is an important contributor to economies in Africa. Producers sell their goods into international markets, with the final products being sold in supermarkets, here in the UK and throughout the world. So what role is new Internet connectivity playing in changing these sectors — which are often seen as slow to adopt new technologies? As part of our work examining the impacts of growing Internet connectivity and new digital ICTs in East Africa we explored uses of the Internet and ICTs in the tea sector in Rwanda.

Tea is a sector with well-established practices and relations in the region, so we were curious if ICT might be changing it. Of course, one cannot ignore the movements of material goods when you research the tea sector. Tea is Rwanda’s main export by value, and in 2012 it moved over 21,000 tonnes of tea, accruing around $56m in value. During our fieldwork we interviewed cooperatives in remote offices surrounded by tea plantations in the temperate Southern highlands, tea processors in noisy tea factories heavy with the overpowering smell of fermenting tea leaves, and tea buyers and sellers surrounded by corridors piled high with sacks of tea.

But behind the material movement that takes tea from the slopes of Rwanda’s ‘thousand hills’ to a box on a shelf in Tesco, is a growing set of less visible digital data flows. Whilst the adoption of digital technologies is not comprehensive in the Rwandan tea sector (with, for example, very low Internet use among tea growers), we did find growing use of the Internet and ICTs. More importantly, where they were present, digital flows of information (such as tea-batch tracking, logistics and sales prices) were increasingly important to the ability of firms to improve production and ultimately to increase their profit share from tea. We have termed this a ‘data-driven value chain’ to highlight that these new digital information flows are becoming as important as the flows of material goods.

So why is tea production becoming increasingly ‘data-driven’? We found two principal drivers at work. Firstly, production of commodities like tea has shifted to private ownership. In Rwanda, tea processing factories are no longer owned by the government (as they were a decade ago) but by private firms, including several multinational tea firms. Prices for buying and selling tea are also no longer fixed by the government, but depend on the market — flat rate prices stopped at the end of 2012. Data on everything from international prices, tea quality and logistics has become increasingly important as Rwandan tea firms look to be part of the global market, by better coordinating production and improving the prices of their tea. For instance, privately owned tea factories (often in remote locations) connect via satellite or microwave Internet links to head offices, and systems integration allows multi-national tea firms the ability to track and monitor production at the touch of a button.

Secondly, we need to understand new product innovation in the tea sector. In recent years new products have particularly revolved around growing demand in the retail market for differentiated products — such as ‘environmental’, fair trade or high quality teas — for which the consumer is willing to pay more. This relates most obviously to the activities in the fields and tea processors, but digital information is also crucial in order to allow for ‘traceability’ of tea. As this guarantees that tea batches have satisfied conditions around location, food safety, chemical use, fair labour (etc.) a key component of new product innovation is therefore data — because it is integral to firms’ abilities to prove their value-added production or methods.

The idea of agricultural value chains — of analysing agricultural production from the perspective of a fragmented network of interconnected firms — has become increasingly influential in strategies and policy making supported by large donors such as the World Bank and the International Fund for Agriculture Development (IFAD), an agency of the UN.

These value chain approaches explore the amount of economic ‘value’ that different actors in the supply chain are able to capture in production. For instance, Rwandan tea farmers are only able to capture very small proportions of the final retail prices — we estimate they are paid less than 6% of the cost of the eventual retail product, and only 22% of the cost of the raw tea that is sold to retailers. Value chain analysis has been popular for policy makers and donors in that it helps them to formulate policies to support how firms in countries like Rwanda improve their value through innovation, improving processes of production, or reaching new customers.

Yet, at the moment it appears that the types of analysis being done by policy makers and donors pay very little attention to the importance of digital data, and so they are presenting an unclear picture of the ways to improve — with a tendency to focus on material matters such as machinery or business models.

Our research particularly highlighted the importance of considering how to adapt digital data flows. The ways that digital information is codified, digitised and accessed can be exclusionary, reducing the ability for smaller actors in Rwanda to compete. For instance, we found that lack of access to clear information about prices, tea quality and wider market information means that smallholders, small processors and cooperatives may not compete as well as they could, or be missing on wider innovations in tea production.

While we have focused here only on tea production, our discussions with those working in other agricultural sectors — and in other countries — suggest that our observations have significance across other agricultural sectors. In agricultural production, strategy, policy and researchers mainly focus on the material elements of production — those which are more visible and quantifiable. However, we suggest that often underlying such actions is a growing layer of digital data activity. It is only through more coherent analysis of the role of digital technologies and data that we can better analyse production — and build appropriate policy and strategies to support commodity producers in sectors like Rwandan tea.

Read the full report: Foster, C., and Graham, M. (2015) Connectivity and the Tea Sector in Rwanda. Value Chains and Networks of Connectivity-Based Enterprises in Rwanda. Project Report, Oxford Internet Institute, University of Oxford.


Chris Foster is a researcher at the Oxford Internet Institute. His research focus is on technologies and innovation in developing and emerging markets, with a particular interest on how ICTs can support development of low income groups.