Code as Policy
The Political Process of Patching Development
India’s National Rural Employment Guarantee Act (NREGA) guarantees adults from rural households the right to work for one hundred days per year of employment doing “unskilled” manual labor on rural development projects. Yet, implementation of development programs remains a challenge. Political will and good policy design are critical, but they are often not enough because of resistance from local power systems. Such systems — dense nuclei of class, caste, gender, and bureaucratic power — are often what thwart development programs. Reformers do not possess the capacity to fundamentally change power systems through direct confrontation with local elites.
Maintenance of irrigation canals is a popular type of project undertaken through the NREGA. Local bureaucrats and politicians sometimes choose canal repair during the rainy season. Heavy rains make it impossible to audit a certain type of canal repair work because water covers up any evidence of repairs. In pre-digital days, upper-level bureaucrats would issue circulars disallowing canal work during a particular season in a particular district, but there was no way to ensure that their circulars were followed in the field. Today, a software patch deletes canal work from the list of approved projects in a particular region during the rainy season. Through patching, upper-level bureaucrats fine-tune the system on an ongoing basis and control the type of NREGA work to be performed based on the season and local conditions.
With its deep inequalities and a strong capacity in digital technology, India has emerged as an important country for digital innovation in socioeconomic development and governance. Many large technical systems, “e-government platforms,” have been rolled out under the justification that they would reduce corruption, improve accountability, and have a positive impact on the lives of the poor. Such platforms have also been confronted by strong opposition from civil society organizations as well as academics, who have argued that these big technology platforms worsen power imbalances, centralize power, and bring new inefficiencies.
There is a strong positive case for introducing technologies in governance. Indeed, well-designed technologies could improve the well-being of marginalized communities. I argue for a position where we neither throw technology out altogether, nor put our faith in a magical technological determinism. Instead, the effective use of technology requires continuous adjustment and reappraisal. Technology use does not need to come from a narrow technocratic vision, but instead from political imagination.
The use of technology should not be thought of as a one-time fix, but rather adapted to changes through a process I call Patching Development.
Patching development is a socio-technical process that is about altering power equations by directing attention to the mundane minutiae of processes. In the case that I examine in Patching Development, patches that were issued by bureaucrats were intentional responses to address challenges from local power systems in implementing an employment program. In particular, patching highlights the ability of upper-level bureaucrats to access local information about implementation details to address counter-strategy from powerful actors at the local last mile. Patches are attempts by higher-level state institutions to iteratively fix problems in the implementation of policies.
Patching development at the last mile requires political will to create an enabling environment by providing the necessary autonomy for bureaucrats to rewrite their relationship with marginalized people. Social movement organizations, with their history of collective action, are necessary to empower marginalized citizens to be able to contest the actions of the local state through social audits or other citizen-supported mechanisms. Crucially, upper-level bureaucrats must create mechanisms to ensure that those records will not be sabotaged at the local level. To do so, upper-level bureaucrats need to constantly innovate by embracing and adapting new technologies and evolving new processes to ensure that members of local political parties, local elites, and lower-level bureaucrats do not sabotage the delivery of programs at the last mile.
The process of patching development has three key features:
- First, patching is top-down, where the patch sender has jurisdiction over the patch receiver.
- Second, patching is about fine-grained changes, where patches are extremely specific and make focused alterations to policy.
- Third, patching is iterative, where patches are repeatedly authored, are sent based on new information, and become part of a continuous cycle of small, detailed changes to overall program implementation.
Patches are not one-time fixes, nor are they an indication that the policies themselves were patchwork or ad hoc. Patching is also different from “apolitical” technocratic fixes that reduce complex development issues to technical problems for which there are solutions (if only one could find the right technology!). Patches, again, are a series of incremental remedies rather than a one-time overhaul, which has implications for design and evaluation of these interventions. By patching, states can repeatedly change their practices and plans as information and challenges arise and change during the implementation process.
As in my opening example of irrigation canal maintenance, patching is a mechanism by which targeted adjustments that leverage local information can be made to existing digital governance reforms, which are often centralizing in their design and rollout. The need for patching points to the fallibility of state institutions and their information systems. Given the rapid pace of deploying technological “solutions” to public goods delivery, patching provides one way to access local information to help address problems in implementation. Patching development is neither a purely technological process nor an exclusively administrative one. It need not always result in “good” outcomes. Patching could misfire and degenerate into “solutionism,” the tendency to “render technical” socio-technical issues as problems with technical fixes.
Patching is ultimately a political process of exercising power.
We need to develop a model for exercising citizenship over governmental platforms that focuses on small details at the level of patches. Creating such democratic engagement would require practices of documentation, transparency, and citizen engagement. It would require civil society organizations to develop a form of politics that pays attention to the mundane minutiae of technology — drop-down boxes, links, reports, and other details in governmental platforms. And it is as much about institutional features, like who should be hired as a local field assistant, where you should conduct a public meeting, and what time and place the meeting should be held to maximize participation from marginalized citizens. Focusing on “technical” details is not a technocratic act, but a very political process that needs to be recognized as such.
The framework for creating greater democratic engagement would also reform the design process to include patching of non-technical aspects such as designing institutions and ultimately patching development. While each patch may have only limited local significance, the cumulative impact of continued engagement is potentially transformative in fundamental ways and is more likely to lead to effective and just programs.
Rajesh Veeraraghavan is an Assistant Professor of Science Technology and International Affairs (STIA) Program at Georgetown University’s School of Foreign Service.
Code as Policy is the fourth post in Towards a Mindful Digital Welfare State, a series investigating data-driven state services, commissioned by Researcher Ranjit Singh and Research Analyst Emnet Tafesse at the AI on the Ground Initiative at Data & Society, with editorial support from Seth Young.
To date, the series also includes:
- Towards a Mindful Digital Welfare State by Ranjit Singh
- Between Belonging and Individuating: Data, Identification, Diaspora by Emnet Tafesse
- Good Welfare, Bad Platforms? by Silvia Masiero
- Critical Digital Education for All! by Soledad Magnone
- Identifying Beneficiaries in Ghana’s Donor-Oriented Welfare State by Alena Thiel
- Datafying Children by Grace Mutung’u
