Experimental governance? The emergence of public sector innovation labs in Latin America

Abstract Public sector innovation labs are becoming an increasingly visible instrument in public sector innovation and experimentation. Proponents of these labs claim they can play an important role in addressing pressing social challenges, changing government structures and thereby shaping ideas and practices of future governance. Whilst some research has been carried out on public innovation labs, the focus of inquiry has been primarily on the emergence, models and activities of labs in Europe and North America. This paper attempts to contribute to this growing body of research by bringing forth some of the particularities of this phenomenon as it emerges in Latin America. Using as starting point three experimental interests identified in the available literature, namely increasing flexibilization of public procedures, developing methods for citizen engagement and experimental development of public policies, the paper presents insights and observations from a study of ten public sector innovation labs in Latin America. In particular, our focus is on how these interests are confronted with different realities and therefore what kind of challenges the labs face. Experimentation in Latin America seems to concern not only flexibilization, engagement and public policies; it also includes juggling with the tensions arising from budgetary constraints, the need to weave networks of regional labs to collaborate and finally the need to align their agendas to those of other institutions, while being accountable to different levels of society. This places Latin American labs in a different light than their European and North American counterparts.


Introduction
In the last decade public sector innovation labs (PSI labs) have become an increasingly visible instrument in public sector innovation and experimentation globally. The proliferation of labs in public administrations around the globe can be depicted as a trend which promotes the idea that public innovation is an imperative (McGann, Blomkamp, interests are manifested in the Latin American PSI labs in the study, and identify some of the particular issues and challenges that emerge. Finally, we conclude with some remarks on problematizing the innovation imperative, the importance of individuals and the qualities they bring to the labs, implications for practitioners and further research in Section 5.

Labs for public sector innovation and their experimental interests
Our focus on governance is rooted in how the concept's broad and ambiguous usage can be used to promote either a distributed and horizontal or a more narrow technocratic way of governing (Pollitt and Hupe 2011;Rhodes 1996). In Latin America it has been primarily promoted through the rather normative concept of "good governance", frequently determined by international organizations that support development projects around public administration initiatives in the region (PSI labs included) (Zurbriggen 2011). In addition, public administration in Latin America has also been influenced by complex processes such as struggles in the 1990s to respond to economic crises by trying to decentralize government structures, and in places by efforts to democratize after periods of dictatorship in the 1980s (Zurbriggen 2011). In that context, participatory approaches constituted attempts to regain citizens' trust (e.g. participatory budget; Goldfrank 2011). In some countries reforms to decentralize have supported citizen participation and closeness to the state, but in others they have also resulted in different degrees of privatization of many public services. These and many other interlocking issues present a panorama in which the promises of flexibilization of procedures, citizen engagement and experimental policy making, emphasized in the lab phenomena, manifest particular challenges in the region.
Flexibilization of public procedures involves not only introducing new innovative processes following technological change combined with managerial ideas, but also fostering "citizen-centric governance" (Tõnurist, Kattel, and Lember 2017). With managerial ideas such as New Public Management and New Public Governance being key precedents of the public innovation imperative (Joy, Shields, and Cheng 2019;McGann, Blomkamp, and Lewis 2018), particularly in virtue of their emphasis on the production of cost-efficient services, PSI labs have been considered as one of the instruments for enabling the transition of public procedures toward more cost-efficient, agile and user-friendly services (Tõnurist, Kattel, and Lember 2017).
Methods for citizen engagement are closely tied with the experimental methods and approaches deployed by the labs. These approaches and methods are based on different models or interpretations about what public innovation is and what could be done toward it, and specially how citizens should be involved and what agency they mobilize. While some labs rely on interpretative and creative methods, emphasizing empathy and human centeredness (Blomkamp 2018;Liedtka et al. 2020;Mintrom and Luetjens 2016), others follow the notion that public innovation should rely on scientific knowledge to inform policies, which will warranty good representation of citizens . Irrespective of the approach, there are underlying experiments on ways to mobilize new actors.
Experimental development of public policies in PSI labs address new processes to policy making, thereby contributing to redefine the problems and seeking for more citizen-centred solutions through making the process more participatory (Lewis, McGann, and Blomkamp 2020). Moreover, labs contribute to introduce the idea about failure as a possibility in the policy arena through trial and error processes. By promoting the idea that policies (like products and services) can be prototyped and piloted (Kimbell & Bailey 2016) the labs propose that experiments can be done without major risks on investment.
Each of these interests relates to different aspects of governance. The first deals with bureaucracies, time and management of public services, changing it toward citizen-centred and more agile procedures. The second introduces new actors to the decision making processes by making practice more participatory, enabling for instance collaborative development public services. However, governance is more explicit with regard to the third interest about development of public policies with labs aiming to intervene directly in the making of public policies. Thus, labs are involved in fostering these three interrelated interests, and through them, in producing new ways to govern. This emphasizes the potential role instruments such as PSI labs can have in developing, or hindering, new ways of governing, thereby encouraging diverse understanding of governance itself.

Methods
Our reflections in this paper are based on a larger qualitative study aiming to understand PSI labs in Latin America by the first author as part of her PhD project. This paper uses only a part of the collected material. Methodologically, the research comprises a thematic analysis from material collected using several sources. The material includes practitioner reports of PSI labs, desk research on their online presence, and most importantly observations from informal conversations with practitioners from the labs and academics in events dealing with relevant topics (Ibero American Lab Meeting 1 , IX International meeting of public policies and design 2 , DRS2018 Conference conversation CN15 3 ).
By reviewing projects, initiatives ( For this paper we have limited our inquiry to only a subset including ten (10) PSI labs of which six (6) are city level labs, and four (4) are state level labs (Table 1). We have only considered labs that are part of public administration, either at a state, regional or city level and that have been identified and discussed already in one way or another in the scant literature available. While three of the city labs were initiated by the respective municipal authorities, the last three city labs were initiated through and follow the logic of the Iberomerican General Secretariat (SEGIB) Project Civic Innovation. State level labs are not as prolific as city labs; therefore the 4 state level labs in our subset are the ones we recorded as having been operative for some years and that explicitly refer to themselves as labs. We checked their online presence, as well as visibility or activity in regional events. We searched for materials produced by or about them (e.g. reports or online interviews to members of the labs in different media like "Diseño y Diaspora" podcast and Apolitical platform). To complement such materials, we also rely on semi-structured interviews the first author did either in person or through video-conferencing (from December 2017 to February 2020) with 14 practitioners from 9 of the labs.

Flexibilization of public procedures
In Latin America the discourse about flexibilization of public administration begins with technological drivers, digitizing and changing public procedures to be more costefficient while engaging citizens and other actors in the processes (Acevedo and Dassen 2016;Cejudo, Dussague Laguna, and Michel 2016). The growing, albeit unevenly distributed, availability of digital devices and connectivity for citizens combined with the citizens' lack of trust in their respective governments, has made some governments promote their commitment to transparency, collaboration and participation by pledging to open government agendas (Acevedo and Dassen 2016). Open government is a central term in public innovation experimentation discourse in the region, and a prominent mandate of Latin American labs (Acevedo and Dassen 2016; Ram ırez-Alujas 2016; Rodr ıguez 2018). From our study, the clearest example of a PSI lab engaged in the flexibilization of public procedures is the AGESICLab. This lab started in a new governmental agency promoting e-government with the initial aim to contribute to the digitization of public procedures as a step toward a unified state digital platform. The lab was for example involved in digitizing procedures for water access where it worked with civil servants from two different ministries and users to co-create a new online permit procedure. Five other labs in our study were also explicitly engaged with flexibilization of public procedures and in each case this was a consequence of an open government agenda. Some municipal governments also see labs as an instrument to develop such an agenda, for instance CISNA and SantaLab explicitly pledge to open government. However, SantaLab has been more engaged in transparency by developing digital projects promoting open data sharing from public administration in their region, while CISNA has put efforts in co-creating technologies to be used for agriculture as their context is mainly rural with challenging connectivity. All the labs in our study mentioned that their aim is to transform or change public administration in a way or the other through or with technology. Nevertheless, they also emphasize the importance of improving people's life.

Methods for citizen engagement
Beyond seeking to improve various public procedures and services, PSI labs are also engaged in developing ways to increase citizen engagement and introduce the idea that citizens can and should be included as a collaborator in for example the development of public services. To that purpose and as their general methodology, the PSI labs in Latin America employ co-creation methods (Ram ırez-Alujas 2016; Zurbriggen & Gonz alez Lago 2015) that are mainly represented in a strong design-led approach (Acevedo and Dassen 2016), thereby introducing to public administration experimental creative methods led by design, including design thinking (Cejudo, Dussague Laguna, and Michel 2016). However in practice it is evidence-based approaches that seem to have more concrete support (or encouragement) from public administration (Acevedo and Dassen 2016). Moreover, labs in Latin America are confronted by more actors (government, external funders, development cooperation) who expect the labs to offer rigor and precision to the experimental approaches and who repeatedly suggests the introduction of resources such as big data or behavioral insights pilots as routes to follow (Acevedo and Dassen 2016).
Every lab from our study mentioned that they develop their own methods, but maintained also that they rely on some aspects of design for their operations (ranging from design thinking, user-centred design or service design). In addition, they combine and adapt methods from varied reference sources. For example, MvdLab, SantaLab and CISNA started within residencies offered by MediaLab Prado (innovation lab from Madrid) and the project Civic Innovation funded by SEGIB. Therefore, these labs follow closely methods proposed by MediaLab Prado and other practices introduced to them through SEGIB's Civic Innovation project. In the case of the state labs, they are all influenced by the Danish MindLab in one way or another. While Gnova was initiated specifically through a cooperation involving MindLab with support from the Danish government in 2016, the other three labs have participated in training provided by MindLab supported by the Inter American Development Bank. Moreover, while our state labs include behavioral economics and ethnography, and even agile methodologies or change theory in their toolkit, the city labs in our study introduce creative methods to engage citizens and understand their stories through storytelling, participatory video and alternative ways of mappings.
Both the city labs and the state labs in our study engage citizens in their processes and thereby introduce new actors to public innovation processes. However, the city labs seem to be more active on this front by focusing explicitly on civil society and activists. For instance at the CISNA lab they always try to identify collaborators that belong to organized civil society to help them gain traction in its local context. Other labs also interact with various citizen participation instruments that the city already has in place. For example, LABcdmx helped the team working with participatory budget visualize the proposals and MvdLab takes part in a citizen participation platform comprising different digital and analog participation processes of the city.

Experimental development of public policies
One of the driving ideas behind PSI labs is to push public administration and their attitude toward the development of public policies, not only by introducing new approaches for policy development but also new ways of defining and understanding policy problems that are more fit to the user's needs. However, experimental development of public policies faces obstacles and challenges in Latin America. For instance, one of the obstacles identified to pursue change in policymaking is tied to lack of support from authorities in public administration, something that can hinder the implementation of solutions or pilots designed by the labs (Acevedo and Dassen 2016). Moreover, while a transition toward more experimental approaches to public policies requires endorsement by the authorities, the transition also relies on support or at least understanding by the civil servants involved in the processes. To overcome the latter obstacle, the labs provide training for civil servants to support innovative capacity building to thereby develop an innovation culture in public administration (Acevedo and Dassen 2016).
With regard to the labs in the study, we can find some attempts to participate in the development of public policies like a co-written piece of legislation between different public administration organisms of Mexico City and LABcdmx who used open government drivers to draft an "open city" law. There is also the experience of the social innovation policy adopted in Nariño, which was co-created and initiated by work at the CISNA lab. However, a general issue for the labs is that they need to adapt their approaches in order to accommodate legislative boundaries regarding the possibility of altering public administration procedures. Even if many of the PSI labs have in their discourse the importance of changing or developing public policies, in practice their possibilities to do so have been limited. Therefore, a recurring strategy we have found is to indirectly inform public policies through the problems they frame with their projects. For instance, LABcdmx aimed with their experimental project peatoniños to introduce "play streets" in poor and marginalized neighborhoods. The intervention entailed closing traffic streets to enable "play streets" in neighborhoods where the local children lacked public spaces to play. While this project achieved some success when the local authorities of one of the boroughs they were experimenting in developed policies to enable the "play streets" to become permanent, this was only after repeated experiments and even after the lab itself was closed down.
As a way to ensure some influence on the public administration, four of the labs stressed the importance of engaging with civil servants not only through projects, but also through training. For example in the case of LABCapital, they decided to shift from citizen focus to a civil servant one. As one of the interviewees explained, while the political authorities change, civil servants typically have longer tenure and will continue in their positions. Engaging with the civil servants and having them as allies could therefore warranty better continuation for their process. Some labs, such as LABgobCh, pursue the training on a more strategic level and in a structured way with clear networks, training and contests. However, all of them seem to do it on a more informal level.

Challenges for public sector innovation labs in Latin America
Observations from our study revealed that in order to pursue these experimental interests the PSI labs in Latin America face certain challenges. This was for example evident in the case of their work on experimental development of public policies and in general that the labs need to consider how the legislation in their respective context is simply not enabling the kind of experimentation they are supposed to pursue. Moreover, they also need to be attentive to transitions in public administration; such transitions could easily imply a change in their mandate or even closure of their operation. This also generates a need to consider how they document and communicate their work. The labs then, are experiencing significant pressure to find ways for their practices or projects to survive or at least ways to share their learnings with the new administrations. Over and above these challenges, we have also identified a series of issues that can be grouped into three broader topics: challenges arising from budgetary constraints, constant weaving and managing of networks and struggles to align agendas and being accountable to different levels of society. While the first challenge is something that practitioners explicitly expressed as obstacles, the other two are our own reflections from the observations and discussions.
Budgetary constraints. Within the rhetoric of an innovation imperative PSI labs have been marketed as cost-efficient entities, in Latin America this assumption sometimes turns problematic for practitioners who frequently mentioned issues regarding budget constraints of their units. In particular, they felt resources assigned for experimentation were minimal or even non existent, which creates a constant "need to do with less resources". Budgets are limited not only because public administrations in Latin america often have other urgent basic needs to cover, but also because it might be difficult to justify spending public money for reasons that are unclear to the public, like experimentation in PSI labs, especially in cases where corruption has led to an increased scrutiny of public spending. While navigating issues of transparency in public spending is a complex endeavor anywhere, in Latin American the credibility of public administration still needs strengthening. This means that labs, as new experimental organizations, have particular challenges to gain credibility from both public administration and citizens at the same time. The Latin American context seems thus to reflect a reality where resources for the PSI labs are scarcer than in the North American and European context because they are distributed to more urgent problems, making experimentation a "privilege" and "luxury" in Latin America (Yosif 2019). Moreover, if we see this challenge in combination with the various legislative and transitional challenges mentioned above, it is clear the context in which Latin American PSI labs operate is complex. Of course, while these challenges could be seen as challenges for Latin American PSI labs in general, it should be stressed that it is not our intention to make such a general claim since Latin American countries differ in legislation and public administration realities. Instead, uncovering and exploring contextual differences in this regard for Latin American PSI labs is a topic for further studies.
Weaving networks. PSI labs are particularly oriented to developing networks to share their experiences and understand their contributions to public innovation. In the case of the Latin American labs, this encompasses not only to establish intercontinental and regional discussions with peers, but also to build a community locally. Moreover, Latin American labs also promote such regional encounters with peers, enabling them to build a "regional reflection" about the challenges they encounter in their own language (as most of the countries speak Spanish, and the burden to understand Portuguese, or vice versa is not high). Labs support collaboration among peers in the Latin American region or in their country to share stories, gain feedback and validate their ideas. Even if some of the labs also belong to formal networks such as Gealc by AGESICLab, Gnova and LABgobAR or Innovalab by LABCapital, the collaborative networks in Latin America are mainly informal, initiated through events or by mutual exchange. One of the practitioners mentioned for example that "at the beginning things move through friendship", thus highlighting the relevance of individuals in shaping the interactions and collaborations the lab establishes. Furthermore, PSI labs also need to build networks to create a community in their context that understands what they do and how they can help change public administration. In this regard, they establish contact with civil servants, authorities, activists, other organized civil society and citizens in general. At least five of our interviewees even mentioned other activist networks they belong to and work with besides their paid work in the lab. PSI labs and people working in them strive to gain credibility at these various levels of society. In virtue of being novel entities that are both part of but also distinct from public administration, they have a need to be acknowledged by other actors because trust to public administration is weak in most of the region.
Aligning agendas and accountability. PSI labs juggle with aligning agendas and being accountable to different levels of society, including national governments with their own imperatives, experimenters with their mandates, local communities with their expectations and enthusiasts from the international circuit. Thus, PSI labs are entangled between expectations and frustrations from different levels of society, not just from public administration. As we mentioned previously, the labs interact with various groups and actors in their regional and local context. However, they also need to take international organizations into consideration. For example, in their analysis of what public innovations labs are, Tõnurist, Kattel, and Lember (2017) mention briefly, as one of the many coordination challenges of the labs, the need to align activities with supranational bodies and international organizations. In the case of Latin America, this interest is not just a passing challenge; many times it is actually a key strategy for existence. Many of the Latin American labs from our study get support from a variety of international organizations for a variety of purposes with different agendas related to the aim of developing innovation and exchange. This support can be economical, but it can also be in other shapes like consultancy for building and starting the labs, resources to exchange knowledge among other labs by setting up events or shared training, or sometimes even only in the form of validation. This juggling is important for the labs, and it also enables them to operate, collaborate and share experiences, especially since the budgets of most of the labs are restricted to the salaries of the core group and often only limited resources to organize workshops or interventions. However, even if the interaction with international organizations is therefore necessary, questions about how this support influences the agendas of the labs arise. Tracing the extent of external support is not always easy since budget sources are not public information and it is not always possible to discuss such information.

Conclusion: experimental governance?
Public innovation labs in Latin America are experimental institutional forms with complex interactions both internally in public administration and externally across existing actors and institutions. Such interactions and the general Latin American context place in foreground challenges that are not always evident in the discussions of the PSI labs in other regions.
The complexity of the context and the challenges provide material for problematizing various aspects of how the innovation imperative manifests itself in Latin America. For example, while labs are encouraged to experiment with flexibilization of public procedures and methods for citizen engagement, the resources and methodological choices are not always made by labs themselves. Instead, they more often than not respond to external need (or even pressure) to validate their actions or be aligned with administrational mandates and/or trends supported by their donors. Moreover, although development of experimental public policies is crucial to the narrative of PSI labs, the lack of support from authorities and constraints in legislation push them to barely inform public policies and seek the support of civil servants to develop an innovation culture in public administration.
Beyond these more general challenges, we also identified budgetary constraints, their need to develop networks and to align agendas and being accountable to different levels of society. For practitioners working in the region it would be important to be more aware of the complexities these challenges imply for their operations. For example, while the support of international organizations do seem to encourage and facilitate exchange among labs, it also sets up thematic agendas and directions that need to be understood better. Experiences in Latin America show that PSI labs and their diverse approaches to public innovation might lead to new forms of governance, however doing it more purposefully will require more self reflection, validation of their work and more support, not only from public administration authorities but also by other sectors of society.
Finally, we note that by bringing forth the relevance of informal networks individuals at the labs develop, observations in this paper also support the suggestion by Ferreira and Berglund (2019), to focus on the importance of individuals for the potential outreach and sustainability of the lab. It could thus be worthwhile to look into the role of the individuals within the labs, but also into the informal networks they develop and their activities outside the lab that could have a potential impact for pursuing experimental ways of governing. Ireland. The first author co-hosted a conversation with Federico Vaz: CN15 -"Smuggling ideologies? Inquiring into the underlying ideas embedded in design for public governance and policy making".