Technology for Transparency Review, Part III · Global Voices
David Sasaki

Throughout our mapping and analysis of the 36 case studies on the Technology for Transparency Network  we have tended to group projects by their geographic region. This is in large part because our team of researchers and research reviewers were hired explicitly to map and evaluate technology projects that promote transparency and accountability in the regions were they are based. However, when one steps back to take in all of the projects from a global perspective, it becomes readily apparent that there are thematic and strategic categories which apply across regions. This week's posts aim to tease out some of those trends and offer constructive criticism and concrete recommendations to funders, project leads, and researchers as to how such projects can become more effective, efficient, and sustainable.
On the Technology for Transparency Network platform we have organized case studies and project listings into ten different categories: budget monitoring, civic complaints, election monitoring, parliament informatics, extractive industries, private sector transparency, advocacy, crime, local government, and aid transparency. You can sort through each of the ten categories by clicking on the filters beneath our map interface. If you have suggestions for other categories that we should be tracking, please leave a comment below.
Today's post will focus on two of the largest categories we reviewed – budget and election monitoring.
Monitoring the budgets of local and national governments is a key instrument in the toolkit of any accountability activist. Active budget monitoring can both prevent and expose corruption. As an example, despite protests from privacy activists, the Mexican government decided to publish the salaries of elected officials. (At the time, Mexican governors were among the highest paid in the world.) The reasoning is that any elected official who is clearly spending more than his/her salary permits should be scrutinized closely to determine where that money is coming from, and whether it is linked to political misbehavior.
Budget monitoring can also lead to improved public services and infrastructure. For example, in the United States the government recently passed the largest economic stimulus program in the country's history. To track how that money was spent the government created Recovery.gov while ProPublica created Eye on the Stimulus, which also tracks how the money is spent. Kenya had its own stimulus program, called the Constituency Development Fund, which started in 2003 as a way to fund local governments to improve their infrastructure and services. Budget Tracking Tool is a way to see how that money is being spent and to leave comments to report on the progress of those projects. However, so far few people seem to use the tool, and even fewer comment on the progress of the development projects in their constituency.
In order to effectively monitor and evaluate any budget, the data must be available in a format which allows for analysis in a spreadsheet or database program. The information should be both granular, in order to evaluate as many variables as possible, and timely, in order to expose corruption and inefficiencies before it is too late. Unfortunately, most governments that do publicly release their budget information do so using Adobe's PDF format, which doesn't allow for data analysis. As Noam Hoffstater and Alon Padan of Our Budget pointed out in our interview, the accounting offices of governments obviously have their budget information in spreadsheet format, but they purposely publish it as a PDF document in order to discourage closer scrutiny.
Our Budget uses OCR technology to create an Excel spreadsheet version of the Tel Aviv municipal budget. Volunteers go over and check every entry, and then they make visualizations and graphs of how the municipality is spending taxpayers’ money. Importantly, while they have done this for two years running, in the end they decided that litigation (demanding that the municipality release the budget in spreadsheet format) was a better strategy than time-consuming, technological solutions.
Dinero y Politica uses a similar strategy to create more information about campaign financing in Argentinian elections so that voters can make fully informed choices. In Argentina, political parties must disclose all of the campaign contributions they received at least ten days before the election. But, once again, they only have to disclose those numbers in a PDF report, which, doesn't allow citizens analyze the data to see relationships between political interests and politicians. So the Dinery y Politica team has created an interactive database which maps donations and creates visualizations of which parties receive donations from which groups, unions, and companies.
Recommendations:
Both Our Budget and Dinero y Política use Many Eyes to visualize the data they collect. We recommend that other budget monitoring activists learn how to use Many Eyes and Many Eyes Wikified in order to dynamically visualize budget information. Google's Fusion Tables is another powerful online tool to both store and visualize complex information related to public budgets.
The information collected and analyzed from all three projects does not seem to be exploited well by civil society organizations, journalists, or bloggers in the countries where they are based. We recommend to project leaders that they do more outreach to train journalists, activists, and bloggers how to use the tools they have developed. A sample gallery of effective ways their information has been used could help inspire others to build on that work.
A number of traditional civil society organizations have been working in the realm of budget monitoring and open budgets for quite some time. The International Budget Partnership has a very useful world map which links to country profiles with related information and organizations working on budget monitoring. We recommend to donors that they support an international event that brings together technologists and budget monitoring activists to share best practices and strategies regarding the use of modern technology and information management systems to improve the efficiency of budget monitoring. This event should be followed by a three-day intense “book sprint” which leads to a open licensed, freely accessible book that explains the technical steps in order to: 1) extract financial information from PDF documents using OCR technology, 2) store budget information in structured, public databases, 3) verify and cross-reference information, 4) analyze and evaluate data using Many Eyes and Google Fusion Tables, and 5) use resulting findings and conclusions to partner with media, bloggers, civil society and government to seek greater accountability.
Following the publication of such a book – and its translation into relevant languages – we recommend that donors support barcamp-style events at the national level that bring together technologists, civil society organizations, government officials, investigative journalists, and bloggers to focus specifically on budget monitoring.
Finally, we recommend that budget monitoring platforms partner with accounting, statistics, and computer science professors at local universities so that the students of their classes can help improve the governance of their country while learning new skills and techniques.
Like budget monitoring, election observation is an activity of government accountability and transparency with a long history. According to Wikipedia, it dates back to at least to the 1866 plebiscite of Moldavia and Wallachia, which led to modern Romania. In more recent times, election monitoring has tended to focus on countries with weak democracies or democracies in transition, and has been organized by international organizations such as the Organization for Security and Co-operation in Europe, the European Union, the Commonwealth Secretariat, the Council of Europe and the African Union. Major international NGOs like the Carter Center, the International Foundation for Election Systems (IFES), the National Endowment for Democracy (NED), the Electoral Institute of Southern Africa, the National Democratic Institute, and the European Network for Election Monitoring have also become increasingly active, and often partner with local NGO's to take advantage of established national networks.
As if all of those many players weren't enough, a new generation of election monitoring websites now ask ordinary citizens to also become election observers by using their cell phones to report any voting irregularities they may encounter. If all goes to plan, such reports are then verified, categorized, and placed on a publicly accessible map.
One such project that we did not document because we felt that it has been discussed sufficiently in both mainstream and citizen media is Vote Report India. In fact, co-founder Gaurav Mishra has joked that perhaps there were more articles about the project than reports submitted to the platform. Mishra has also written an excellent and candid evaluation of the project, which lists successes, failures, and – most importantly – lessons learned for the next incarnation of the platform for the 2014 Indian elections.
Last month saw the first multiparty election to take place in Sudan in over 20 years. Sudan Vote Monitor is one of many Ushahidi-based websites we reviewed that allow voters to report irregularities by submitting text messages which are then verified by a partner NGO and placed on a map. To understand the context behind Sudan Vote Monitor and the difficulty in implementing a technology project in Sudan, I highly recommend Rebekah Heacock's post, “Sudan: Is ICT all it's cracked up to be?” Most of the attention given to the initiative during the election itself focused on the fact that the site was temporarily blocked by the government. Less reported was the website's impact on ensuring credible elections. According to the website itself, 257 reports of voting irregularities were received. In a followup interview with Heacock, project coordinator Fareed Zein says that number is probably somewhere between 300 – 500 if you count all SMS reports, which have yet to be added to the system. But few, if any, of these reports have been verified, and there has been no official response to any of the reports. Still, Zein suggests that the objective of this first implementation was simply to create more information, rather than necessarily holding anyone accountable:
In previous elections it was all a closed door affair — nobody knew exactly what went on on the ground. The intention of this was to be able to get the information out to the public, internally and externally, about what's going on. Just being able to get the word out was enough for us. We didn't set out to try to urge anybody to take any specific action. Our mission was to get the information out and then let people judge and act for themselves.
Zein says it is likely that Sudan Vote Monitor will be used again during the January 2011 Southern Sudanese independence referendum.
Like Vote Report India and Sudan Vote Monitor, Cuidemos el Voto is another national election monitoring platform that uses Ushahidi. Co-founder Oscar Salazar notes that, while Mexico transitioned to multiparty democracy in 2000 with the election of Vicente Fox, vote buying and conditional cash transfer programs are still corrupting the democratic process. Cuidemos el Voto managed to achieve something important that other, similar projects lacked: support and endorsement from an official government body, in this case, Mexico's Special Prosecutor's Office. Still, that endorsement did not apparently lead to any sort of accountability. For example, on July 5, 2009 someone reported that in Puebla they were offering 500 pesos to vote for the PAN party. But this report wasn't verified and we don't see any kind of followup. For such election monitoring projects to make an impact in terms of accountability, they need the staff and resources to verify all reports and ensure that the proper government body responds. Or, perhaps more appropriately, they must partner with other organizations who can invest in the long haul of followup work. It could even be the perfect semester-length project for a political science graduate-level class.
The African Elections Project differs slightly from other election monitoring projects we reviewed in that 1) it doesn't use Ushahidi and 2) it focuses on multiple countries throughout Sub-Saharan Africa. The project, funded by the Open Society Initiative for West Africa, uses new media tools to produce and disseminate more information about elections in ten African countries including Malawi, Namibia, Botswana, Niger, Guinea, Mozambique, Mauritania and Togo. Their hope is that more watchful eyes reporting on the electoral process will help prevent and expose vote fraud, and encourage clean elections. However, most of the countries where they work have broadband penetration levels between five and ten percent. Until there is greater connectivity, the impact of an online project like African Elections Project is inherently circumscribed, despite its relatively large budget.
VoteReport PH is the last Ushahidi-based election monitoring initiative we reviewed. Most of these projects were only attract the participation of very few users because there was not broad awareness that the websites exist at all. VoteReport PH is different in that its team spent six months prior to the election going around the country and giving voter education classes about how to use automated voting machines (which were used for the first time), and simultaneously, how to submit reports to VoteReporter PH by sending text messages. Such on-the-ground outreach work is necessary in order to create more awareness about citizen election monitoring platforms. In all, 654 reports of voting irregularities were submitted. For example, at 1 p.m. on May 10 “massive vote buying” was reported by an anonymous contributor via text message. We are told that the report was verified, but there is no further detail on what constitutes verification, or if anything was done to follow up on the report. One very useful strategy by the VoteReportPH team was to write a separate blog post highlighting the most urgent reports of vote fraud. They also published a helpful blog post summarizing some of their early experiences and conclusions. I recommend Mong Palatino's overview, “Monitoring Philippine Elections through Social Media“, to get a better idea of how Twitter and blogging played a distributed role in monitoring the Philippine elections.
Though Ushahidi was created to map reports of violence that occurred after Kenya's 2007 election, it was quickly adapted to monitor elections themselves. In addition to the above-mentioned case studies, Ushahidi was also used to map voting irregularities in Afghanistan and Lebanon.
Ushahidi implementations have already been set up for Colombia's presidential election later this month, the Puebla municipal elections in July, and Brazil's general election in October.
Recommendations:
We recommend to project implementers that they plan at least one year ahead of the elections they will monitor. In addition to the technical challenges of implementing and localizing the Ushahidi platform, they must also concentrate on outreach efforts to 1) establish an SMS shortcode, 2) hold on-the-ground training workshops, 3) partner with relevant civil society organizations, and 4) partner with media organizations to spread awareness. Such projects should also seek funding to cover the expenses of marketing the project via billboards, radio commercials, posters, and leaflets. Finally, and perhaps most importantly, election monitoring platforms should establish strong relationships with the governing electoral commission in order to agree upon a protocol for verifying and acting on reports of vote fraud. Without a signed agreement in place, the project is unlikely to make a concrete impact toward greater accountability and credibility of elections.
We recommend that online election monitoring projects partner with students at universities to verify and follow up on all submitted reports, similar to how Tufts University students worked to verify reports submitted to the Ushahidi implementation for the Haiti earthquake.
We recommend to funders that they fund specific election-related plugins that make Ushahidi a more functional platform for election monitoring. Oscar Salazar of Cuidemos el Voto has noted that Ushahidi lacks certain features for election monitoring:
Ushahidi's main design was to provide a common pool of reports. So if I started giving administrative access to everyone, everyone will see the same pools . What happen's if two NGOs that are associated with two different political parties get access to the same pool and start approving or disapproving the reports? I don't want everyone to have access to the same pool. i want to give special accounts to different NGOs, where they see only their own reports plus the citizen reports. Ushahidi wasnt designed for a lot of NGOs working together. So we are tweaking it for these local elections to make it work in this way.
The tweaks by the Cuidemos el Voto team should be packaged into a plugin that can be shared with other election monitoring initiatives.
We recommend to the Ushahidi team and to their funders, that increased emphasis be placed on documentation, especially in regard to best practices regarding election monitoring. Patrick Meier has written an introductory primer and Erik Hersman has made a forum posting to compare common election-related categories, but there is still a lack of needed documentation for any activist wanting to use Ushahidi to monitor elections. We suggest that Ushahidi aspire to a documentation resource as thorough and easily accessible as WordPress’ Codex.
We recommend that researchers do both more longitudinal and comparative research in order to better understand the impact and methodology of online, citizen election monitoring websites. How does verification compare across projects? What strategies bring about accountability? How do you increase the signal to noise ratio for submitted reports? How do you best visualize reports to inspire action? What are obstacles to collaboration with traditional election monitoring organizations? What are the pros and cons of anonymity in citizen election reporting? These are just a few open-ended questions that require more research.
Finally, we recommend that multilateral and civil society organizations focused on election monitoring organize an international event to bring together the coordinators and technologists behind the various online election monitoring websites we have listed above to share experiences and prepare improved documentation for future implementations. This has already happened at the regional level in East Africa, which led to calls for a “Tech Election Monitoring Toolbox”, but it should also happen at the international level to share skills, techniques, resources, and future plans across distinct tech communities.
In the next post we will summarize and offer our recommendations related to more categories of case studies from the Technology for Transparency Network.