Democracy Lab

Redrawing the Map for Democracy

How South Africa's post-apartheid government tried to do away with the territorial legacy of racial segregation.

Note: This article is an abridged version of a longer historical case study produced by Innovations for Successful Societies, a research program at Princeton University. 

From 1948 to 1994, the ruling National Party government carved apartheid into South Africa's political map, dividing the country's internal administration based on the principle of racial segregation. To escape the legacy of discrimination, the framers of post-apartheid South Africa had to redraw the country's internal boundaries while navigating potentially explosive competing interests. They called upon representatives of each faction to reach a political settlement, albeit an imperfect one, that survives today. The process was challenging and political, but it serves as an example for other transitioning governments dealing with similar mandates.

By 1990, faced with mass rioting, international sanctions, and a determined underground resistance, South Africa's President F.W. de Klerk of the National Party had unbanned the African National Congress (ANC) -- the anti-apartheid movement led by Nelson Mandela -- and began secret negotiations to formulate a post-apartheid government.

Multiparty talks proceeded fitfully, stalling amid civilian fighting and an increase in violence throughout the country, but the specter of mass uprising ultimately forced compromise in 1993. "If we didn't [work for a settlement], the masses would have taken to the streets and demanded immediate takeover," Roelf Meyer, a chief negotiator for the National Party government, said. Formally known as the Multi-Party Negotiating Forum, the negotiators set April 1994 as the deadline for both a new constitution and elections -- South Africa's first with universal adult suffrage.

But with elections looming, the post-apartheid map of South Africa was still largely undetermined. The new, unified South Africa with equal citizenship for all would require an end to the practice of segregated homeland territories, or Bantustans, that divided the black African majority from the white and other minority populations. The multiparty negotiators were faced with a difficult task: dividing the country into integrated provinces without reinforcing old wounds from segregation or opening new ones by dividing existing communities. Any discussion about new borders would open the door to groups that wanted semi-independent ethnic enclaves separate from the rest of South Africa. Yet ensuring a peaceful transition from minority to majority rule required balancing a wide range of moderate and extremist positions. The two center parties had to weigh voices from the fringes that threatened to act as political spoilers in the process, without rewarding them for using violence to advance their agendas.

South Africa was previously divided into four provinces (two former British colonies and two previously independent Boer republics) and 10 homeland areas. Under apartheid, the government did not recognize residents of the homelands as full citizens of South Africa, as part of a broader strategy of racial segregation and discrimination. Residents of the homelands received markedly inferior social services, had fewer economic opportunities, and were considered guest workers if they sought employment in one of the provinces.

During the transition era, homeland leaders argued for divisions along ethnic lines that would allow them to retain power; in an ethnically diverse community, they stood little chance of winning election to political office. Long-serving homeland leaders were strong advocates of such ethnically defined provinces. In KwaZulu, Mangosuthu Buthelezi commanded considerable influence through the Inkatha Freedom Party, a Zulu cultural movement turned political party. Other negotiators could not ignore Buthelezi's demands due to ongoing violence between ANC and Inkatha supporters in South Africa's urban slums, where migrant workers from KwaZulu often lived in cramped hostels. Tension between the groups had been growing as apartheid came to an end. In July 1991, newspapers had reported that members of the apartheid government's security forces had funded and trained Inkatha supporters to provoke violence between the two groups in order to undermine the ANC.

Buthelezi was not alone; other homeland leaders, such as Lucas Mangope of Bophuthatswana, also sought to keep power by preserving the boundaries of the homelands they controlled. To bolster their bargaining positions, the homeland leaders teamed up with another group advocating segregation: conservative Afrikaners who demanded an ethnic Afrikaner homeland (volkstaat). This faction, which included leaders of railway workers, farmers, and mine workers, threatened to use their collective power to damage the country's economy if the ANC did not set aside a territory for white speakers of Afrikaans. Buthelezi, Mangope, and leaders of the conservative Afrikaners formed the Concerned South Africans Group (COSAG), later called the Freedom Alliance, through which they sought to preserve or augment the power they held under apartheid.

More radical Afrikaner groups also used threats of violence to gain more influence than their level of national support warranted; media reports estimate that these groups represented only a minority of South Africa's roughly 3 million Afrikaners. During the negotiating forum's constitutional discussions, members of the far-right paramilitary Afrikaner Resistance Movement rampaged through the conference center. The vandalism and intimidation of participants made it clear that extremists were willing to use violence to achieve their objectives.

Even though the ANC was vehemently opposed to the plans advocated by these factions, they still needed the conservative Afrikaners and the homeland leaders to "buy in" to the border-drawing process, in order to avoid further violence and isolate groups like the Afrikaner Resistance Movement.

Recognizing that they needed both technical expertise and representation of the competing viewpoints to draw up a mutually acceptable map, the negotiating forum's leaders set up the Commission on the Demarcation/Delimitation of States/Provinces/Regions (CDDR) in May 1993, which would hear and evaluate proposals for the new boundaries from interested parties around the country. The commission would then report its conclusions and recommendations to the negotiating forum.

The CDDR was made up of 15 commissioners, with wide representation among the negotiating parties, though the ANC and the National Party retained significant influence over the process and each nominated one of the commission's co-chairmen. Members were chosen to reflect the political, gender, and racial makeup of South Africa. Although the commissioners had different political affiliations, they were united by the desire to find a solution that would enable the elections to move forward.

"You had to get a balance, because if you had somebody that was so hell-bent on their own political view, then you would never have been able to come up with a consensus document," Renosi Mokate, head of the technical committee, explained. "Because we were all trying to arrive at a workable solution for the country ... people also tempered their own ideological and political agendas."

The commissioners' first task was to decide how to evaluate boundary proposals. The negotiating forum mandated that the CDDR take into account historical boundaries, the availability of infrastructure and service delivery, existing government structures, demographics, economic viability, potential for development, "cultural and language realities," and that it limit financial costs, inconvenience to citizens, and dislocation of service. Interpretation of those criteria, however, was in the hands of the commissioners themselves. They organized the criteria into four categories -- economic aspects, geographic coherence, institutional and administrative capacity, and socio-cultural issues -- and agreed to evaluate each proposed boundary on its merits within each of those four categories. The "cultural and linguistic realities" criterion posed a particular challenge for the commission. "We had to take into account that in demarcating the provinces, we must not reinforce the legacy of apartheid," Mokate said.

With the April 1994 deadline for elections looming, time was critical. The commission first met in May 1993 with only three months to gather data, hold public hearings, and complete its work. Moreover, its technical committee lacked basic resources to carry out its mandate. The government of South Africa had never produced a census of the entire population, and other official statistics did not include the homelands, leaving the commission searching for other sources for the country's demographic information. As a starting point for their analysis, the commissioners decided to base the provincial borders largely on a map created by the government-funded Development Bank of South Africa in the 1980s. In order to plan its own investments, the Development Bank had divided the country into nine economic regions, disregarding existing homeland and provincial boundaries. The ANC representatives also used information from World Bank missions to supplement their knowledge of South Africa's demographics and infrastructure, since World Bank officials had the resources to conduct aerial surveys and extract information from institutions such as the army.

To give South Africans a voice in the process, the commission asked the public to propose ideas. Individuals could send their suggestions directly to the commission or attend public forums held around the country, which were attended by delegations from the commission and advertised over radio and in newspapers. Commissioners recalled engaging in lively discussions with the public during those events. "Our job was to be there, to listen, to take notes, but to also interrogate," Mokate said. For example, the commission grilled a leader of the Xhosa people when he proposed an ethnically Xhosa-centered province, asking whether the province would be economically viable and asking him to consider the implications of creating provinces elsewhere on an ethnic basis.

Although public consultation aimed to make the process inclusive, the commissioners and their researchers recognized that many people could not participate. Discussions were often dominated by people who were well organized and who had the money and education to engage with the issues. By contrast, people in poor, far-flung, and weakly organized communities -- often ones in the homelands -- were less involved.

Once the public consultation phase was complete, the technical committee summarized and cataloged the arguments, and compiled a report for the commissioners, highlighting the major issues for consideration. The requirement that each province be contiguous eliminated many proposals centered on linguistic majorities, such as the submission from the Bophuthatswana leadership, which included several physically unconnected regions. Initial proposals to set up an all white, Afrikaans-speaking volkstaat also did not meet tests of administrative rationality. Paul Daphne, a commissioner and ANC party leader, recalled: "The people proposing a volkstaat outcome were battling to find a map which would show any part of the country with a majority of whites in it." COSAG, the conservative Afrikaners group, was particularly persistent in to its push for majority single-language communities.

Commissioners hotly debated whether to split the Eastern Cape into two provinces. Those favoring the split pointed to the demands of two former homeland areas: the Transkei and the Ciskei. The Transkei leader, an important supporter of the ANC, urged the party to create a separate province based on his homeland, arguing that the Transkei would receive more development attention as a separate province. Some commissioners argued that it would be better to unite the relatively underdeveloped former homelands in a single province that included the economically vibrant coastal cities of Port Elizabeth and East London. In the final proposal, economic considerations trumped political pressures and the Eastern Cape remained a single province. 

In parallel discussions, the commissioners considered creating a Northern Cape province by splitting up the Western Cape. Both the National Party and segments of the conservative Afrikaner contingent supported the proposal for the Northern Cape. The National Party was convinced that the demographics of the proposed territory (which would have a majority Afrikaans-speaking population) would give it a better chance of winning provincial elections there. Conservative Afrikaners, for their part, thought that a Northern Cape province might vote in favor of hosting an Afrikaner homeland. Opponents argued that such a province would not be economically viable.

The night before it delivered its report to the negotiating forum, the commission put the Northern Cape decision to a vote. At first, Afrikaner representative Koos Reyneke declined to vote, despite his instrumental role in the original volkstaat proposal. Reyneke knew that many of the region's residents wished to be included in the Western Cape province, and he did not want to go against their wishes. After the initial vote ended in a tie, however, Reyneke was convinced to vote in favor of the new province because of a bargain he had previously brokered with other conservative Afrikaners. In exchange for his vote, the conservative Afrikaners promised to support his proposal for an Afrikaner homeland near Pretoria, the country's administrative capital. The Northern Cape proposal won by a single vote, but might have deadlocked had an ANC-nominated commissioner not been absent. In this instance, politics triumphed over economics.

The CDDR submitted its report to the negotiating forum on July 31, 1993, recommending in broad strokes a nine-province map, acknowledging that certain highly contested border towns and communities might warrant further investigation. It also discussed a volkstaat, but noted that the groups advocating one had been unable to unite behind a single proposal for its location. The report was signed by thirteen of the commissioners, while two, Reyneke and Ann Bernstein, a development expert, submitted dissenting reports.

"I would strongly urge the negotiating parties not to impose an undemocratic map on the country," Bernstein wrote in her dissent. "To try and actually produce a regional map for the country in such a short time and think that this will resolve the differences that exist between all the many interests on this matter is to my mind totally unrealistic and dangerous." Bernstein argued that the negotiating forum's original criteria were insufficient: They did not include a position on small versus large provinces and did not ask the CDDR to consider the electoral implications of the regional boundaries. She argued that the proposed map represented a political settlement, brokered in order to move negotiations forward as quickly as possible, and therefore did not represent a viable blueprint for regional administration and development.

The Multi-Party Negotiating Forum accepted the CDDR's report but identified eight "sensitive areas" that had not been fully resolved. Apparently acceding to some of Bernstein's concerns, the negotiators extended the CDDR's deadline, sending the commissioners back to gather additional citizen input on the "sensitive areas," including whether the Eastern Cape should be split, whether a Northern Cape province should be created, and where Pretoria should fall. After reissuing a call for proposals and commentary, the CDDR submitted its second report in October 1993, effectively concluding the commissioners' work.

In their second report, the CDDR's two chairmen argued that the responsibility "rests with the political leaders" to negotiate boundaries and convince their constituents to accept the subsequent outcome. "The demarcation of regions ... deals with the wishes, fears, and emotions of human beings," they wrote, "and therefore requires a forum capable of reaching consensus and agreement through a process of compromise and 'give' and 'take.' [The CDDR's mandate was] not to find compromise between conflicting historical, political, and often emotional interests held by various groups."

The report met resistance from communities strongly opposed to their placement in certain boundaries. Political parties had to work to ensure that their supporters accepted the proposed boundaries so that the disputes wouldn't disrupt the upcoming elections. For instance, after working on the commission's technical committee, Trevor Fowler joined the ANC's public relations team in the region that became Gauteng province, receiving phone calls from unhappy constituents asking that their areas be moved to different provinces. Fowler urged callers to table their concerns until after the elections were held. (The constitution contained a provision that allowed communities to dispute a new boundary within 30 days of it coming into effect.) "These comrades agreed that they would not voice their concerns now; they would first go through the elections," said Fowler. "Well, the day after the elections, they called."

The issue of an Afrikaner homeland was also deflected until after the elections, to be considered by a volkstaat board (which ultimately failed to bring the different pro-volkstaat factions together on a single proposal). In Bophuthatswana, however, opposition manifested in a more dangerous and dramatic way. As the country geared up for elections, Lucas Mangope refused to join his homeland with the territory of South Africa or to permit elections to take place. Facing a strike by civil servants and a mutiny by the homeland army in February 1994, Mangope asked heavily armed conservative Afrikaner paramilitaries to secure the territory; members of the extremist Afrikaner Resistance Movement also invaded. In the ensuing conflict between the extremists and Bophuthatswana's mutinying security forces, as many as 100 civilians and combatants were killed. The national army moved in and the government replaced Mangope with a caretaker leader. Both Mangope and the extremist group were discredited by the affair.

On April 27, 1994, South Africa voted as one country for the first time. The CDDR had aimed to create a non-segregated map of the country -- one that would allow elections to proceed without violence and would begin to undo the physical segregation of the homeland system. On that front, it succeeded.

The CDDR had also aimed to create provinces that would be economically sustainable and logically governable. Nearly two decades later, South Africa's provincial boundaries remain largely unchanged from the original demarcations. Local disputes, however, continue, often over provincial boundaries that cut through municipalities or that separate border towns from nearby economic centers. According to Mokate, the current border disputes occur where communities did not have a strong say in the original process in the early 1990s. "Those people are now feeling much more empowered, much more organized, and are saying 'But how did this happen? I don't like this. Now I've got a government in power, I must tell them I don't want this.'"

As political nominees, the members of the CDDR could not cast aside their party affiliations and loyalties. Still, their common commitment to a unified South Africa allowed them to work beyond their interests to create viable provincial boundaries for the new South Africa. "Nobody got exactly what it is that they wanted, but the people, as a collective, pulled together to the extent that they could," Mokate said. "Even the people who disagreed with each other fundamentally, they still talked to each other.... They recognized each other as South Africans who participated in a process that contributed to building this country."

The CDDR's experience illustrates how careful consultation and negotiation can bring together factions with seemingly irreconcilable interests. The CDDR brought together individuals who represented the groups vying to control South Africa's post-apartheid map, but who were willing to find a common settlement. The commission's members remained behind the scenes, which allowed them to avoid public backlash for making concessions. This consensus was vital to ensuring South Africa's peaceful transition.

But the story of the CDDR also warns of the dangers of a map based heavily on political settlement at the expense of administrative considerations. Meyer, who was appointed minister of provincial affairs and constitutional development after the 1994 elections, said the negotiators did not focus enough on creating viable provincial administrations. "We failed ourselves by only doing this within the last period of time before the transition," he said. "I think we were so consumed, all of us, in our negotiations, in finding a constitutional settlement, that we didn't think about the bigger consequences of what we were doing about creating a new state."

Tom Stoddart/Getty Images

Democracy Lab

Bridging Kenya's Data Divide

How Africa’s tech leader made it from PDFs to paperless.

Note: This article is an abridged version of a longer historical case study produced by Innovations for Successful Societies, a research program at Princeton University. 

On July 8, 2011, President Mwai Kibaki officially launched the Kenya Open Data Initiative at a public event attended by more than 3,000 people. The new Internet portal compiled previously scattered or hard-to-access government information and made it available to the public for free. 

Kenya launched its initiative at a time when a number of countries were also putting government information online. The United States led the global open data movement, launching its site in 2009. Britain's official data site went online a year later. Within three years, 31 countries -- including Australia, Canada, Chile, Denmark, Estonia, Italy, Moldova, Morocco, the United Arab Emirates, and Uruguay -- had created similar portals. Multilateral organizations embraced the concept as well; The World Bank inaugurated its own open data website in 2010. 

Kenya's Open Data Initiative was an important milestone for the country. In the past, citizens had found it difficult to obtain reliable information from the government. During President Daniel arap Moi's tenure (1978 - 2002), the government restricted the free flow of information and clamped tight restrictions on Kenya's few private radio and television networks. In addition to stifling the media, the Moi government barred civil servants from sharing data outside government. The Official Secrets Act, a holdover from Kenya's colonial era, gave the government the ostensible authority to withhold data. 

When President Kibaki took office in 2002, he did not earmark freedom of information as a high priority; he was instead primarily concerned with job growth. The first three years of his tenure witnessed few advances for open data. Entrenched in a silo mentality, civil servants continued to guard their information. Citizens could buy official reports from the government printing office, but it was a costly and inconvenient process. Other information was available only through appeal to the permanent secretaries who headed ministries. The system created ample opportunities for mismanagement, misappropriation of funds, and bribery. 

When Bitange Ndemo joined Kibaki's Cabinet as permanent secretary of the Ministry of Information and Communications in 2005, he was fully aware of the challenges of accessing government data. As a senior lecturer and head of research and consultancy at the University of Nairobi's faculty of commerce, Ndemo had firsthand knowledge of the difficulty in securing data for academic research. "Getting information from the government was a big problem when you did research from the outside," he recalled. "For example, one area that I really wanted to study was the reasons for the collapse of firms in Kenya. But I couldn't get the information. And I certainly couldn't get several years' worth of information."

Kibaki's government did make incremental moves to improve the availability of information. A significant step came in 2006, when the administration enacted a structural reform to facilitate streamlined access to information. It created the Kenya National Bureau of Statistics (KNBS), replacing the outdated and ineffective Central Bureau of Statistics. The new bureau assumed responsibility for gathering, storing, and analyzing government information, as well as publishing and disseminating the data for public use. 

For Kenya's citizens, however, complete government data remained elusive. Although the KNBS regularly updated its website with the monthly consumer price index, quarterly gross domestic product aggregates, and basic economic indicators such as coffee or tea prices, other statistics were limited to summaries in PDF files. It published paper copies of some of the information in the forms of books and reports, selling them for small fees. 

The government's moderate efforts to widen information access contrasted with the technological explosion that took place in Kenya at the same time. Information and communications technology (ICT) became Kenya's fastest-growing sector, thanks largely to deregulation and privatization following years of government monopoly. After 2000, the sector grew at an average rate of 20 percent annually and contributed an average of 1 percentage point to Kenya's gross domestic product.

Pointing to the fast-growing ICT sector, Kenya's technology community started clamoring for easier access to government information. It claimed that access to government data could help reduce waste, spur development of mobile or Web applications, and promote technology start-ups. Athman Mohamed, director at Trademark East Africa, an agency that promotes regional trade, offered transportation as an example. "Up to 40 percent of the cost of anything in a landlocked country is due to transportation," he said. "You can have a direct impact on the lives [of people] if you can use data to save a day here and there [on goods being transported]." 

The ICT community argued that the new software applications entrepreneurs would create using open data would help encourage innovation and fuel growth. Al Kags, an entrepreneur and founder of the Open Institute, an organization dedicated to open data and open governments, said, "A lot of young people are driven by innovation. They will develop applications if you provide them with data. And they will provide important service to others." For instance, in 2007 - 08, developers had created the Ushahidi platform that helped monitor incidents of violence following Kenya's 2007 elections. It used crowd sourcing, an online tool that rapidly collected and disseminated data from contributors, to track the violence. 

Kenyans could point to another recent example. In 2007, cellular company Safaricom launched M-PESA, a service that enabled subscribers to pay bills and transfer money by using their cell phones. In 2011 - 12, the company earned 16.9 billion Kenyan shillings ($200 million) in revenues from M-PESA alone, with 14.9 million users in a country of 43 million. 

Open data advocates found a champion in Ndemo. Soon after assuming office, Ndemo had identified five priorities for the ICT sector: infrastructure through fiber cables, content and application development, public-private partnerships, capacity building, and employment. The need for open data cut through the last four. 

Early in his tenure, Ndemo looked to build relationships that would be conducive to innovation. He allied himself closely with the minister of information and communications, Mutahi Kagwe, who helped him gain the president's support. Ndemo also developed professional relationships and coalitions by reaching out to ICT people in the private sector. In 2007, he created an ICT Board, a state corporation for implementing the ministry's policies and projects. "In government, you don't get a choice of people you work with," he said. "So, we actually created a small, semiautonomous agency called the ICT Board to quickly implement some of the programs we had started." 

By 2009, Ndemo had developed a track record for carrying out successful reforms by liberalizing the telecommunications sector, supporting M-PESA, and helping bring high-speed fiber-optic cables to Kenya. Tackling the thorny issue of open data was a logical next step. 

Ndemo knew that the success of open data reforms would hinge largely on approvals from ministries to release data through a public website; obtaining such approvals would be costly in terms of time and energy. 

In early 2011, Ndemo made a strategic decision. He opted to launch the government's open data site with information that was already in the public domain, but not yet broadly available or in a usable form. In doing so, he sidestepped the need to confront government agencies about their release of nonpublic information. As the legislative backbone of his efforts, Ndemo pointed to the 2010 constitution, which called for the government to "publish and publicize any important information affecting the nation." 

Although Ndemo's open data initiative operated within existing parameters, he still found it difficult to loosen ministries' tight grip on government information. Needing high-level backing, he made his case during a personal visit with President Kibaki in June 2011. "I went to the president and told him we have a lot of data in government, which we can use and convert into businesses for the youth and for more employment," Ndemo said. "He is an economist, and he understood all this very quickly." Kibaki gave his blessing to the open data project and accepted Ndemo's invitation to preside at a launch event a few weeks later. 

The tight timeline put pressure on Ndemo to move quickly but also worked in his favor. Before approaching the president, Ndemo had met with planning, finance, health, and education officials to get data or secure permission for using data the officials had already supplied to the World Bank. 

To expedite the data-gathering process and lessen the workload on ministries, Ndemo accepted data in any format, printed or digital. Paul Kukubo, chief executive officer of the ICT Board explained, "The idea was that any data is good. We will do the hard work in making that data relevant and cleaning it on our end." 

Ndemo formed a task force of 23 members comprising public officials, developers, and World Bank data experts. The group had teams responsible for (1) solving technical and usability issues, and getting the website up and running; (2) cleaning and formatting data for presentation on the site; (3) addressing legal and policy matters, including terms and conditions for data usage; and (4) organizing the launch-day event. The task force, especially the data team, received technical support from data experts at the World Bank headquarters in Washington, D.C.. 

The task force settled a number of strategic questions at its first meeting. The team would focus on data visualization -- visual representation of information through graphs, maps, applications, and easily downloadable files -- instead of just posting PDF files or tables on the website. It also decided to develop applications to showcase how the data could be used. For building the actual portal, Ndemo and other team members agreed to contract with an outside firm to secure the software platform, and customize it for Kenya's needs. 

Starting on June 17, 2011, the four teams began work in earnest. The group worked under extreme pressure. Mohamed of Trademark East Africa, who led the technical team, said, "The project management was minute to minute. There was no detailed plan apart from the overall plan, which was [that] we needed to do this by this day." The team did not have time to write out technical documents or specifications. 

Coordination was important, especially between the technical team and the data team. For instance, a technical team member who found missing or faulty data would contact the data group to remedy the error, no matter how small the mistake. That simple step ensured quality and accountability. To vet software applications before the launch, the task force set up a peer review procedure. For instance, when Erik Hersman, founder of the software developer community iHub and a member of the task force, submitted an application, the technical team had two days to respond with comments. Mohamed said, "We had to enforce quality control [and] didn't want delays, because the president [was coming to the launch]." 

Not surprisingly, agencies with national security concerns were wary of the open data concept. With two days left before the launch, President Kibaki summoned Ndemo to his office. Several ministers, including the minister for internal security, had expressed concerns about the Kenya Open Data Initiative, comparing it to WikiLeaks. The president's office was poised to cancel the portal. 

At a July 7 meeting, the day before the proposed public launch, Ndemo and other task force members assured the president and cabinet ministers that the initiative was controlled locally and that the data was already in the public domain -- in electronic or paper format. The 20-minute meeting turned into a two-hour session as task force members demonstrated through charts and graphs how visualizing open data could help allocate resources. Ndemo noted that the group present was particularly interested in ways that open data could spur employment. Kibaki assured Ndemo that he would inaugurate the portal the next day.                                                        

On July 8, 2011, Kibaki officially launched the site, which showcased 200 data sets organized into six categories: education, energy, health, population, poverty, and water and sanitation. The data included the 2009 census; seven years of detailed government expenditure data, including national and county public expenditure; national budgets; the 2005 household income survey; and information on health care and education. 

Kenya's portal made international headlines and lit up the blogosphere. Newspapers such as The Guardian and The New York Times carried articles about the site. The Star, a national Kenyan newspaper, rated it number one in a list of "Kenya's biggest ICT stories of 2011." 

Proponents lauded the portal as a giant step forward for Kenya, but some observers were less effusive. They insisted that the project's value ultimately depended on citizen use. Davis Adieno, former national coordinator at the National Taxpayers Association, a government accountability organization, commented on the disparity between the praise the portal received and its actual usage: "It is being celebrated internationally, but very few Kenyans know about the portal or what it is about." In 2012, a year after the launch, the media, and the public had not used the open data portal as widely as proponents had anticipated. As Ndemo related, "Right now, we have dealt with just the supply side of data. The challenge now is to build the demand side of data." 

The lack of Internet availability outside Kenya's major cities sharply curtailed the number of citizens who could use the new website. In 2012, only 6.5 million Kenyans out of a population of 43 million subscribed to an Internet connection, although 11.8 million had access to the Internet. But about 29.2 million people had mobile phones, with 98.8 percent of total Internet subscriptions through mobile phones -- one reason the ministry planned to encourage software developers to focus on phone applications that could reach a wider population. 

A second challenge was timeliness. Ndemo related that after launch, "The pressure I have now is the need for real-time data." Hersman of iHub underscored the issue, asking, "How can we get updated data and engage ordinary people to make their lives better and talk about services?" An important ally in facilitating data flow was the KNBS. The bureau was Kenya's largest repository of government data. Crucially, it had staff in key ministries to gather data or monitor data collection. Ndemo tapped into this expertise, putting in place a plan to allocate responsibility for portal updates to the KNBS, while the ICT Board would continue to project-manage the site. 

A third challenge arose from the muted reception the portal received within government. Sharing of government data required a greater commitment on the part of the government and a stronger legislative framework. In 2012, Ndemo and his supporters were still pushing for greater acceptance of the portal across ministries. 

Despite its challenges, the open data portal provided one avenue of access to government information in Kenya. At the time of the initial launch, the portal had 200 data sets. The number stood at 434 in June 2012. The ICT Board reported 50,515 site visits as of June 2012. When the board opened a Twitter account in February 2012, 429 followers started following the portal, which had had 634 tweets by June 2012. It also reported that users had viewed 29,081 data sets and downloaded 2,600 data sets by June 2012. 

The portal promoted Kenya's reputation as a progressive nation. Kenya was the first country in sub-Saharan Africa to launch an open data portal, and the second one on the continent, after Morocco. The portal also facilitated Kenya's membership in the Open Government Partnership, a global effort by eight founding members -- Brazil, Indonesia, Mexico, Norway, the Philippines, South Africa, Britain, and the United States -- to promote transparency and accountability in government. 

Although several factors contributed to the successful launch of Kenya's open data portal in 2011, top-level political support and the speed with which Ndemo pushed the portal forward were the two most significant. Ndemo summed up his approach to the reforms: "In government, you seize the moment and the opportunity when you get it. How do you get it? You do the end first, and then you can put the rest in place later. You simply must deal with the why you need something, then think about [the] how later."    

Brent Stirton/Getty Images